mktables revamp

5 files changed, 14783 insertions, 2769 deletions

diff --git a/lib/unicore/Makefile b/lib/unicore/Makefile
index b73d3b8089..0dd77b7bb5 100644
--- a/lib/unicore/Makefile
+++ b/lib/unicore/Makefile
@@ -1,12 +1,6 @@
 all:
-	../../miniperl -I../../lib ./mktables
-
-TestProp.pl: mktables UnicodeData.txt Scripts.txt Blocks.txt PropList.txt
-	../../miniperl -I../../lib ./mktables -maketest
-
-test:   TestProp.pl
-	../../miniperl -I../../lib TestProp.pl
+	../../miniperl -I../../lib ./mktables -P ../../pod -T ../../t/re/uniprops.t -makelist
 
 clean:
-	rm -f *.pl */*.pl */*/*.pl
-	rm -f Properties
+	rm -fr *.pl To lib
+	rm -f ../../pod/perluniprops.pod ../../t/re/uniprops.t mktables.lst
diff --git a/lib/unicore/README.perl b/lib/unicore/README.perl
index 96b2caa2fc..7515825c6f 100644
--- a/lib/unicore/README.perl
+++ b/lib/unicore/README.perl
@@ -1,11 +1,17 @@
 The *.txt files were copied from
 
-	http://www.unicode.org/Public/5.1.0/ucd
+	ftp://www.unicode.org/Public/UNIDATA
 
-and subdirectories 'extracted' and 'auxiliary' as of Unicode 5.1.0 (March 2008).
+with subdirectories 'extracted' and 'auxiliary'
 
-The big file, Unihan.txt (28 MB, 5.8 MB zip) was not included due to space
-considerations.  Also NOT included were any *.html files and *Test.txt files.
+The Unihan files were not included due to space considerations.  Also NOT
+included were any *.html files and *Test.txt files.  It is possible to add the
+Unihan files, and edit mktables (see instructions near its beginning) to look
+at them.
+
+The file 'version' should exist and be a single line with the Unicode version,
+like:
+5.2.0
 
 To be 8.3 filesystem friendly, the names of some of the input files have been
 changed from the values that are in the Unicode DB:
@@ -27,42 +33,68 @@ mv extracted/DerivedLineBreak.txt extracted/DLineBreak.txt
 mv extracted/DerivedNumericType.txt extracted/DNumType.txt
 mv extracted/DerivedNumericValues.txt extracted/DNumValues.txt
 
-The names of files, such as test files, that are not used by mktables are not
-changed, and will not work correctly on 8.3 filesystems.
+If you have the Unihan database (5.2 and above), you should also do the
+following:
 
-The file 'version' should exist and be a single line with the Unicode version,
-like
-5.1.0
+mv Unihan_DictionaryIndices.txt UnihanIndicesDictionary.txt
+mv Unihan_DictionaryLikeData.txt UnihanDataDictionaryLike.txt
+mv Unihan_IRGSources.txt UnihanIRGSources.txt
+mv Unihan_NumericValues.txt UnihanNumericValues.txt
+mv Unihan_OtherMappings.txt UnihanOtherMappings.txt
+mv Unihan_RadicalStrokeCounts.txt UnihanRadicalStrokeCounts.txt
+mv Unihan_Readings.txt UnihanReadings.txt
+mv Unihan_Variants.txt UnihanVariants.txt
+
+If you download everything, the names of files, such as test files, that are
+not used by mktables are not changed by the above, and will not work correctly
+as-is on 8.3 filesystems.
+
+mktables is used to generate the tables used by the rest of Perl.  It will warn
+you about any *.txt files in the directory substructure that it doesn't know
+about.  You should remove any so-identified, or edit mktables to add them to
+its lists to process.  You can run
+
+    mktables -globlist
+
+to have it try to process these tables generically.
+
+If any files are added, deleted, or their names change, you must run
 
-NOTE: If you modify the input file set you should also run
- 
     mktables -makelist
-    
-which will recreate the mktables.lst file which is used to speed up
-the build process.    
+
+to generate a new list of all the files.
 
 FOR PUMPKINS
 
-The files are inter-related.  If you take the latest UnicodeData.txt, for example,
-but leave the older versions of other files, there can be subtle problems.
+The files are inter-related.  If you take the latest UnicodeData.txt, for
+example, but leave the older versions of other files, there can be subtle
+problems.
+
+When moving to a new version of Unicode, you need to update 'version' by hand
+
+	p4 edit version
+	...
+
+You should look in the Unicode release notes (which are probably towards the
+bottom of http://www.unicode.org/reports/tr44/) to see if any properties have
+newly been moved to be Obsolete, Deprecated, or Stabilized.  The full names for
+these should be added to the respective lists near the beginning of mktables,
+using an 'if' to add them for just this Unicode version going forward, so that
+mktables can continue to be used for earlier Unicode versions. 
+
+When putting out a new Perl release, think about if any of the Deprecated
+properties should be moved to Suppressed.
 
 The *.pl files are generated from the *.txt files by the mktables script,
 more recently done during the Perl build process, but if you want to try
 the old manual way:
 	
 	cd lib/unicore
-	cp .../UnicodeOriginal/*.txt .
-	rm NormalizationTest.txt Unihan.txt Derived*.txt
-	p4 edit Properties *.pl */*.pl
-	perl ./mktables
+	p4 edit *.pl */*.pl */*/*.pl
+	perl ./mktables -P ../../pod -T ../../t/re/uniprops.t -makelist
 	p4 revert -a
 	cd ../..
 	perl Porting/manicheck
-
-You need to update version by hand
-
-	p4 edit version
-	...
 	
 If any new (or deleted, unlikely but not impossible) *.pl files are indicated:
 
@@ -78,4 +110,4 @@ And finally:
 	p4 submit
 
 -- 
-jhi@iki.fi; updated by nick@ccl4.org
+jhi@iki.fi; updated by nick@ccl4.org, public@khwilliamson.com
diff --git a/lib/unicore/mktables b/lib/unicore/mktables
index 6ab4e70480..abd44f5cc4 100644
--- a/lib/unicore/mktables
+++ b/lib/unicore/mktables
@@ -1,2232 +1,13697 @@
-## !!!!!!!!!!!!!!       IF YOU MODIFY THIS FILE       !!!!!!!!!!!!!!!!!!!!!!!!!
-## Any files created or read by this program should be listed in 'mktables.lst'
-
 #!/usr/bin/perl -w
-require 5.008;	# Needs pack "U". Probably safest to run on 5.8.x
+
+# !!!!!!!!!!!!!!       IF YOU MODIFY THIS FILE       !!!!!!!!!!!!!!!!!!!!!!!!!
+# Any files created or read by this program should be listed in 'mktables.lst'
+# Use -makelist to regenerate it.
+
+require 5.008;        # Needs pack "U". Probably safest to run on 5.8.x
 use strict;
+use warnings;
 use Carp;
+use File::Find;
+use File::Path;
 use File::Spec;
-use Text::Tabs ();  ## using this makes the files about half the size
-
-##
-## mktables -- create the runtime Perl Unicode files (lib/unicore/**/*.pl)
-## from the Unicode database files (lib/unicore/*.txt).
-##
-
-## "Fuzzy" means this section in Unicode TR18:
-##
-##    The recommended names for UCD properties and property values are in
-##    PropertyAliases.txt [Prop] and PropertyValueAliases.txt
-##    [PropValue]. There are both abbreviated names and longer, more
-##    descriptive names. It is strongly recommended that both names be
-##    recognized, and that loose matching of property names be used,
-##    whereby the case distinctions, whitespace, hyphens, and underbar
-##    are ignored.
-
-## Base names already used in lib/gc_sc (for avoiding 8.3 conflicts)
-my %BaseNames;
-
-##
-## Process any args.
-##
-my $Verbose        = 0;
-my $MakeTestScript = 0;
-my $AlwaysWrite    = 0;
-my $UseDir         = "";
-my $FileList       = "$0.lst";
-my $MakeList       = 0;
-
-while (@ARGV)
-{
+use Text::Tabs;
+
+sub DEBUG () { 0 }  # Set to 0 for production; 1 for development
+
+##########################################################################
+#
+# mktables -- create the runtime Perl Unicode files (lib/unicore/.../*.pl),
+# from the Unicode database files (lib/unicore/.../*.txt),  It also generates
+# a pod file and a .t file
+#
+# The structure of this file is:
+#   First these introductory comments; then
+#   code needed for everywhere, such as debugging stuff; then
+#   code to handle input parameters; then
+#   data structures likely to be of external interest (some of which depend on
+#       the input parameters, so follows them; then
+#   more data structures and subroutine and package (class) definitions; then
+#   the small actual loop to process the input files and finish up; then
+#   a __DATA__ section, for the .t tests
+#
+# This program works on all releases of Unicode through at least 5.2.  The
+# outputs have been scrutinized most intently for release 5.1.  The others
+# have been checked for somewhat more than just sanity.  It can handle all
+# existing Unicode character properties in those releases.
+#
+# This program needs to be able to run under miniperl.  Therefore, it uses a
+# minimum of other modules, and hence implements some things itself that could
+# be gotten from CPAN
+#
+# This program uses inputs published by the Unicode Consortium.  These can
+# change incompatibly between releases without the Perl maintainers realizing
+# it.  Therefore this program is now designed to try to flag these.  It looks
+# at the directories where the inputs are, and flags any unrecognized files.
+# It keeps track of all the properties in the files it handles, and flags any
+# that it doesn't know how to handle.  It also flags any input lines that
+# don't match the expected syntax, among other checks.
+# It is also designed so if a new input file matches one of the known
+# templates, one hopefully just needs to add it to a list to have it
+# processed.
+#
+# It tries to keep fatal errors to a minimum, to generate something usable for
+# testing purposes.  It always looks for files that could be inputs, and will
+# warn about any that it doesn't know how to handle (the -q option suppresses
+# the warning).
+#
+# This program is mostly about Unicode character (or code point) properties.
+# A property describes some attribute or quality of a code point, like if it
+# is lowercase or not, its name, what version of Unicode it was first defined
+# in, or what its uppercase equivalent is.  Unicode deals with these disparate
+# possibilities by making all properties into mappings from each code point
+# into some corresponding value.  In the case of it being lowercase or not,
+# the mapping is either to 'Y' or 'N' (or various synonyms thereof).  Each
+# property maps each Unicode code point to a single value, called a "property
+# value".  (Hence each Unicode property is a true mathematical function with
+# exactly one value per code point.)
+#
+# When using a property in a regular expression, what is desired isn't the
+# mapping of the code point to its property's value, but the reverse (or the
+# mathematical "inverse relation"): starting with the property value, "Does a
+# code point map to it?"  These are written in a "compound" form:
+# \p{property=value}, e.g., \p{category=punctuation}.  This program generates
+# files containing the lists of code points that map to each such regular
+# expression property value, one file per list
+#
+# There is also a single form shortcut that Perl adds for many of the commonly
+# used properties.  This happens for all binary properties, plus script,
+# general_category, and block properties.
+#
+# Thus the outputs of this program are files.  There are map files, mostly in
+# the 'To' directory; and there are list files for use in regular expression
+# matching, all in subdirectories of the 'lib' directory, with each
+# subdirectory being named for the property that the lists in it are for.
+# Bookkeeping, test, and documentation files are also generated.
+
+my $matches_directory = 'lib';   # Where match (\p{}) files go.
+my $map_directory = 'To';        # Where map files go.
+
+# DATA STRUCTURES
+#
+# The major data structures of this program are Property, of course, but also
+# Table.  There are two kinds of tables, very similar to each other.
+# "Match_Table" is the data structure giving the list of code points that have
+# a particular property value, mentioned above.  There is also a "Map_Table"
+# data structure which gives the property's mapping from code point to value.
+# There are two structures because the match tables need to be combined in
+# various ways, such as constructing unions, intersections, complements, etc.,
+# and the map ones don't.  And there would be problems, perhaps subtle, if
+# a map table were inadvertently operated on in some of those ways.
+# The use of separate classes with operations defined on one but not the other
+# prevents accidentally confusing the two.
+#
+# At the heart of each table's data structure is a "Range_List", which is just
+# an ordered list of "Ranges", plus ancillary information, and methods to
+# operate on them.  A Range is a compact way to store property information.
+# Each range has a starting code point, an ending code point, and a value that
+# is meant to apply to all the code points between the two end points,
+# inclusive.  For a map table, this value is the property value for those
+# code points.  Two such ranges could be written like this:
+#   0x41 .. 0x5A, 'Upper',
+#   0x61 .. 0x7A, 'Lower'
+#
+# Each range also has a type used as a convenience to classify the values.
+# Most ranges in this program will be Type 0, or normal, but there are some
+# ranges that have a non-zero type.  These are used only in map tables, and
+# are for mappings that don't fit into the normal scheme of things.  Mappings
+# that require a hash entry to communicate with utf8.c are one example;
+# another example is mappings for charnames.pm to use which indicate a name
+# that is algorithmically determinable from its code point (and vice-versa).
+# These are used to significantly compact these tables, instead of listing
+# each one of the tens of thousands individually.
+#
+# In a match table, the value of a range is irrelevant (and hence the type as
+# well, which will always be 0), and arbitrarily set to the null string.
+# Using the example above, there would be two match tables for those two
+# entries, one named Upper would contain the 0x41..0x5A range, and the other
+# named Lower would contain 0x61..0x7A.
+#
+# Actually, there are two types of range lists, "Range_Map" is the one
+# associated with map tables, and "Range_List" with match tables.
+# Again, this is so that methods can be defined on one and not the other so as
+# to prevent operating on them in incorrect ways.
+#
+# Eventually, most tables are written out to files to be read by utf8_heavy.pl
+# in the perl core.  All tables could in theory be written, but some are
+# suppressed because there is no current practical use for them.  It is easy
+# to change which get written by changing various lists that are near the top
+# of the actual code in this file.  The table data structures contain enough
+# ancillary information to allow them to be treated as separate entities for
+# writing, such as the path to each one's file.  There is a heading in each
+# map table that gives the format of its entries, and what the map is for all
+# the code points missing from it.  (This allows tables to be more compact.)
+
+# The Property data structure contains one or more tables.  All properties
+# contain a map table (except the $perl property which is a
+# pseudo-property containing only match tables), and any properties that
+# are usable in regular expression matches also contain various matching
+# tables, one for each value the property can have.  A binary property can
+# have two values, True and False (or Y and N, which are preferred by Unicode
+# terminology).  Thus each of these properties will have a map table that
+# takes every code point and maps it to Y or N (but having ranges cuts the
+# number of entries in that table way down), and two match tables, one
+# which has a list of all the code points that map to Y, and one for all the
+# code points that map to N.  (For each of these, a third table is also
+# generated for the pseudo Perl property.  It contains the identical code
+# points as the Y table, but can be written, not in the compound form, but in
+# a "single" form like \p{IsUppercase}.)  Many properties are binary, but some
+# properties have several possible values, some have many, and properties like
+# Name have a different value for every named code point.  Those will not,
+# unless the controlling lists are changed, have their match tables written
+# out.  But all the ones which can be used in regular expression \p{} and \P{}
+# constructs will.  Generally a property will have either its map table or its
+# match tables written but not both.  Again, what gets written is controlled
+# by lists which can easily be changed.
+
+# For information about the Unicode properties, see Unicode's UAX44 document:
+
+my $unicode_reference_url = 'http://www.unicode.org/reports/tr44/';
+
+# As stated earlier, this program will work on any release of Unicode so far.
+# Most obvious problems in earlier data have NOT been corrected except when
+# necessary to make Perl or this program work reasonably.  For example, no
+# folding information was given in early releases, so this program uses the
+# substitute of lower case, just so that a regular expression with the /i
+# option will do something that actually gives the right results in many
+# cases.  There are also a couple other corrections for version 1.1.5,
+# commented at the point they are made.  As an example of corrections that
+# weren't made (but could be) is this statement from DerivedAge.txt: "The
+# supplementary private use code points and the non-character code points were
+# assigned in version 2.0, but not specifically listed in the UCD until
+# versions 3.0 and 3.1 respectively."  (To be precise it was 3.0.1 not 3.0.0)
+# More information on Unicode version glitches is further down in these
+# introductory comments.
+#
+# This program works on all properties as of 5.2, though the files for some
+# are suppressed from apparent lack of demand for.  You can change which are
+# output by changing lists in this program.
+
+# The old version of mktables emphasized the term "Fuzzy" to mean Unocde's
+# loose matchings rules (from Unicode TR18):
+#
+#    The recommended names for UCD properties and property values are in
+#    PropertyAliases.txt [Prop] and PropertyValueAliases.txt
+#    [PropValue]. There are both abbreviated names and longer, more
+#    descriptive names. It is strongly recommended that both names be
+#    recognized, and that loose matching of property names be used,
+#    whereby the case distinctions, whitespace, hyphens, and underbar
+#    are ignored.
+# The program still allows Fuzzy to override its determination of if loose
+# matching should be used, but it isn't currently used, as it is no longer
+# needed; the calculations it makes are good enough.
+
+# SUMMARY OF HOW IT WORKS:
+#
+#   Process arguments
+#
+#   A list is constructed containing each input file that is to be processed
+#
+#   Each file on the list is processed in a loop, using the associated handler
+#   code for each:
+#        The PropertyAliases.txt and PropValueAliases.txt files are processed
+#            first.  These files name the properties and property values.
+#            Objects are created of all the property and property value names
+#            that the rest of the input should expect, including all synonyms.
+#        The other input files give mappings from properties to property
+#           values.  That is, they list code points and say what the mapping
+#           is under the given property.  Some files give the mappings for
+#           just one property; and some for many.  This program goes through
+#           each file and populates the properties from them.  Some properties
+#           are listed in more than one file, and Unicode has set up a
+#           precedence as to which has priority if there is a conflict.  Thus
+#           the order of processing matters, and this program handles the
+#           conflict possibility by processing the overriding input files
+#           last, so that if necessary they replace earlier values.
+#        After this is all done, the program creates the property mappings not
+#            furnished by Unicode, but derivable from what it does give.
+#        The tables of code points that match each property value in each
+#            property that is accessible by regular expressions are created.
+#        The Perl-defined properties are created and populated.  Many of these
+#            require data determined from the earlier steps
+#        Any Perl-defined synonyms are created, and name clashes between Perl
+#            and Unicode are reconciled.
+#        All the properties are written to files
+#        Any other files are written, and final warnings issued.
+
+# As mentioned above, some properties are given in more than one file.  In
+# particular, the files in the extracted directory are supposedly just
+# reformattings of the others.  But they contain information not easily
+# derivable from the other files, including results for Unihan, which this
+# program doesn't ordinarily look at, and for unassigned code points.  They
+# also have historically had errors or been incomplete.  In an attempt to
+# create the best possible data, this program thus processes them first to
+# glean information missing from the other files; then processes those other
+# files to override any errors in the extracted ones.
+
+# For clarity, a number of operators have been overloaded to work on tables:
+#   ~ means invert (take all characters not in the set).  The more
+#       conventional '!' is not used because of the possibility of confusing
+#       it with the actual boolean operation.
+#   + means union
+#   - means subtraction
+#   & means intersection
+# The precedence of these is the order listed.  Parentheses should be
+# copiously used.  These are not a general scheme.  The operations aren't
+# defined for a number of things, deliberately, to avoid getting into trouble.
+# Operations are done on references and affect the underlying structures, so
+# that the copy constructors for them have been overloaded to not return a new
+# clone, but the input object itself.
+
+# The bool operator is deliberately not overloaded to avoid confusion with
+# "should it mean if the object merely exists, or also is non-empty?".
+
+#
+# WHY CERTAIN DESIGN DECISIONS WERE MADE
+
+# XXX These comments need more work.
+#
+# Why have files written out for binary 'N' matches?
+#   For binary properties, if you know the mapping for either Y or N; the
+#   other is trivial to construct, so could be done at Perl run-time instead
+#   of having a file for it.  That is, if someone types in \p{foo: N}, Perl
+#   could translate that to \P{foo: Y} and not need a file.   The problem is
+#   communicating to Perl that a given property is binary.  Perl can't figure
+#   it out from looking at the N (or No), as some non-binary properties have
+#   these as property values.
+# Why
+# There are several types of properties, based on what form their values can
+# take on.  These are described in more detail below in the DATA STRUCTURES
+# section of these comments, but for now, you should know that there are
+# string properties, whose values are strings of one or more code points (such
+# as the Uppercase_mapping property); every other property maps to some other
+# form, like true or false, or a number, or a name, etc.  The reason there are
+# two directories for map files is because of the way utf8.c works.  It
+# expects that any files there are string properties, that is that the
+# mappings are each to one code point, with mappings in multiple code points
+# handled specially in an extra hash data structure.  Digit.pl is a table that
+# is written there for historical reasons, even though it doesn't fit that
+# mold.  Thus it can't currently be looked at by the Perl core.
+#
+# There are no match tables generated for matches of the null string.  These
+# would like like \p{JSN=}.  Perhaps something like them could be added if
+# necessary.  The JSN does have a real code point U+110B that maps to the null
+# string, but it is a contributory property, and therefore not output by
+# default.
+#
+# FUTURE ISSUES
+#
+# The program would break if Unicode were to change its names so that
+# interior white space, underscores, or dashes differences were significant
+# within property and property value names.
+#
+# It might be easier to use the xml versions of the UCD if this program ever
+# would need heavy revision, and the ability to handle old versions was not
+# required.
+#
+# There is the potential for name collisions, in that Perl has chosen names
+# that Unicode could decide it also likes.  There have been such collisions in
+# the past, with mostly Perl deciding to adopt the Unicode definition of the
+# name.  However in the 5.2 Unicode beta testing, there were a number of such
+# collisions, which were withdrawn before the final release, because of Perl's
+# and other's protests.  These all involved new properties which began with
+# 'Is'.  Based on the protests, Unicode is unlikely to try that again.  Also,
+# many of the Perl-defined synonyms, like Any, Word, etc, are listed in a
+# Unicode document, so they are unlikely to be used by Unicode for another
+# purpose.  However, they might try something beginning with 'In', or use any
+# of the other Perl-defined properties.  This program will warn you of name
+# collisions, and refuse to generate tables with them, but manual intervention
+# will be required in this event.  One scheme that could be implemented, if
+# necessary, would be to have this program generate another file, or add a
+# field to mktables.lst that gives the date of first definition of a property.
+# Each new release of Unicode would use that file as a basis for the next
+# iteration.  And the Perl synonym addition code could sort based on the age
+# of the property, so older properties get priority, and newer ones that clash
+# would be refused; hence existing code would not be impacted, and some other
+# synonym would have to be used for the new property.  This is ugly, and
+# manual intervention would certainly be easier to do in the short run; lets
+# hope it never comes to this.
+
+# A NOTE ON UNIHAN
+#
+# This program can generate tables from the Unihan database.  But it doesn't
+# by default, letting the CPAN module Unicode::Unihan handle them.  Prior to
+# version 5.2, this database was in a single file, Unihan.txt.  In 5.2 the
+# database was split into 8 different files, all beginning with the letters
+# 'Unihan'.  This program will read those file(s) if present, but it needs to
+# know which of the many properties in the file(s) should have tables created
+# for them.  It will create tables for any properties listed in
+# PropertyAliases.txt and PropValueAliases.txt, plus any listed in the
+# @cjk_properties array and the @cjk_property_values array.  Thus, if a
+# property you want is not in those files of the release you are building
+# against, you must add it to those two arrays.  Starting in 4.0, the
+# Unicode_Radical_Stroke was listed in those files, so if the Unihan database
+# is present in the directory, a table will be generated for that property.
+# In 5.2, several more properties were added.  For your convenience, the two
+# arrays are initialized with all the 5.2 listed properties that are also in
+# earlier releases.  But these are commented out.  You can just uncomment the
+# ones you want, or use them as a template for adding entries for other
+# properties.
+#
+# You may need to adjust the entries to suit your purposes.  setup_unihan(),
+# and filter_unihan_line() are the functions where this is done.  This program
+# already does some adjusting to make the lines look more like the rest of the
+# Unicode DB;  You can see what that is in filter_unihan_line()
+#
+# There is a bug in the 3.2 data file in which some values for the
+# kPrimaryNumeric property have commas and an unexpected comment.  A filter
+# could be added for these; or for a particular installation, the Unihan.txt
+# file could be edited to fix them.
+# have to be
+#
+# HOW TO ADD A FILE
+
+# Unicode Versions Notes
+
+# alpha's numbers halve in 2.1.9, answer cjk block at 4E00 were removed from PropList; not changed, could add gc Letter, put back in in 3.1.0
+# Some versions of 2.1.x Jamo.txt have the wrong value for 1105, which causes
+# real problems for the algorithms for Jamo calculations, so it is changed
+# here.
+#   White space vs Space.  in 3.2 perl has +205F=medium math space, fixed in 4.0, and ok in 3.1.1 because not there in unicode. synonym introduced in 4.1
+# ATBL = 202.  202 changed to ATB, and all code points stayed there.  So if you were useing ATBL you were out of luck.
+# Hrkt Katakana_Or_Hiragana came in 4.01, before was Unknown.
+#
+# The default for missing code points for BidiClass is complicated.  Starting
+# in 3.1.1, the derived file DBidiClass.txt handles this, but this program
+# tries to do the best it can for earlier releases.  It is done in
+# process_PropertyAliases()
+#
+##############################################################################
+
+my $UNDEF = ':UNDEF:';  # String to print out for undefined values in tracing
+                        # and errors
+my $MAX_LINE_WIDTH = 78;
+
+# Debugging aid to skip most files so as to not be distracted by them when
+# concentrating on the ones being debugged.  Add
+# non_skip => 1,
+# to the constructor for those files you want processed when you set this.
+# Files with a first version number of 0 are special: they are always
+# processed regardless of the state of this flag.
+my $debug_skip = 0;
+
+# Set to 1 to enable tracing.
+our $to_trace = 0;
+
+{ # Closure for trace: debugging aid
+    my $print_caller = 1;        # ? Include calling subroutine name
+    my $main_with_colon = 'main::';
+    my $main_colon_length = length($main_with_colon);
+
+    sub trace {
+        return unless $to_trace;        # Do nothing if global flag not set
+
+        my @input = @_;
+
+        local $DB::trace = 0;
+        $DB::trace = 0;          # Quiet 'used only once' message
+
+        my $line_number;
+
+        # Loop looking up the stack to get the first non-trace caller
+        my $caller_line;
+        my $caller_name;
+        my $i = 0;
+        do {
+            $line_number = $caller_line;
+            (my $pkg, my $file, $caller_line, my $caller) = caller $i++;
+            $caller = $main_with_colon unless defined $caller;
+
+            $caller_name = $caller;
+
+            # get rid of pkg
+            $caller_name =~ s/.*:://;
+            if (substr($caller_name, 0, $main_colon_length)
+                eq $main_with_colon)
+            {
+                $caller_name = substr($caller_name, $main_colon_length);
+            }
+
+        } until ($caller_name ne 'trace');
+
+        # If the stack was empty, we were called from the top level
+        $caller_name = 'main' if ($caller_name eq ""
+                                    || $caller_name eq 'trace');
+
+        my $output = "";
+        foreach my $string (@input) {
+            #print STDERR __LINE__, ": ", join ", ", @input, "\n";
+            if (ref $string eq 'ARRAY' || ref $string eq 'HASH') {
+                $output .= simple_dumper($string);
+            }
+            else {
+                $string = "$string" if ref $string;
+                $string = $UNDEF unless defined $string;
+                chomp $string;
+                $string = '""' if $string eq "";
+                $output .= " " if $output ne ""
+                                && $string ne ""
+                                && substr($output, -1, 1) ne " "
+                                && substr($string, 0, 1) ne " ";
+                $output .= $string;
+            }
+        }
+
+        if ($print_caller) {
+            if (defined $line_number) {
+                    print STDERR sprintf "%4d: ", $line_number;
+            }
+            else {
+                    print STDERR "     ";
+            }
+            $caller_name .= ": ";
+            print STDERR $caller_name;
+        }
+
+        print STDERR $output, "\n";
+        return;
+    }
+}
+
+# This is for a rarely used development feature that allows you to compare two
+# versions of the Unicode standard without having to deal with changes caused
+# by the code points introduced in the later verson.  Change the 0 to a SINGLE
+# dotted Unicode release number (e.g. 2.1).  Only code points introduced in
+# that release and earlier will be used; later ones are thrown away.  You use
+# the version number of the earliest one you want to compare; then run this
+# program on directory structures containing each release, and compare the
+# outputs.  These outputs will therefore include only the code points common
+# to both releases, and you can see the changes caused just by the underlying
+# release semantic changes.  For versions earlier than 3.2, you must copy a
+# version of DAge.txt into the directory.
+my $string_compare_versions = DEBUG && 0; #  e.g., v2.1;
+my $compare_versions = DEBUG
+                       && $string_compare_versions
+                       && pack "C*", split /\./, $string_compare_versions;
+
+sub uniques {
+    # Returns non-duplicated input values.  From "Perl Best Practices:
+    # Encapsulated Cleverness".  p. 455 in first edition.
+
+    my %seen;
+    return grep { ! $seen{$_}++ } @_;
+}
+
+$0 = File::Spec->canonpath($0);
+
+my $make_test_script = 0;      # ? Should we output a test script
+my $write_unchanged_files = 0; # ? Should we update the output files even if
+                               #    we don't think they have changed
+my $use_directory = "";        # ? Should we chdir somewhere.
+my $pod_directory;             # input directory to store the pod file.
+my $pod_file = 'perluniprops';
+my $t_path;                     # Path to the .t test file
+my $file_list = 'mktables.lst'; # File to store input and output file names.
+                               # This is used to speed up the build, by not
+                               # executing the main body of the program if
+                               # nothing on the list has changed since the
+                               # previous build
+my $make_list = 1;             # ? Should we write $file_list.  Set to always
+                               # make a list so that when the pumpking is
+                               # preparing a release, s/he won't have to do
+                               # special things
+my $glob_list = 0;             # ? Should we try to include unknown .txt files
+                               # in the input.
+my $output_range_counts = 1;   # ? Should we include the number of code points
+                               # in ranges in the output
+# Verbosity levels; 0 is quiet
+my $NORMAL_VERBOSITY = 1;
+my $PROGRESS = 2;
+my $VERBOSE = 3;
+
+my $verbosity = $NORMAL_VERBOSITY;
+
+# Process arguments
+while (@ARGV) {
     my $arg = shift @ARGV;
     if ($arg eq '-v') {
-        $Verbose = 1;
-    } elsif ($arg eq '-q') {
-        $Verbose = 0;
-    } elsif ($arg eq '-w') {
-        $AlwaysWrite = 1;	# update the files even if they havent changed
-        $FileList = "";
-    } elsif ($arg eq '-check') {
+        $verbosity = $VERBOSE;
+    }
+    elsif ($arg eq '-p') {
+        $verbosity = $PROGRESS;
+        $| = 1;     # Flush buffers as we go.
+    }
+    elsif ($arg eq '-q') {
+        $verbosity = 0;
+    }
+    elsif ($arg eq '-w') {
+        $write_unchanged_files = 1; # update the files even if havent changed
+    }
+    elsif ($arg eq '-check') {
         my $this = shift @ARGV;
         my $ok = shift @ARGV;
         if ($this ne $ok) {
             print "Skipping as check params are not the same.\n";
             exit(0);
         }
-    } elsif ($arg eq '-maketest') {
-        $MakeTestScript = 1;
-    } elsif ($arg eq '-makelist') {
-        $MakeList = 1;        
-    } elsif ($arg eq '-C' && defined ($UseDir = shift)) {
-	-d $UseDir or die "Unknown directory '$UseDir'";
-    } elsif ($arg eq '-L' && defined ($FileList = shift)) {
-        -e $FileList or die "Filelist '$FileList' doesn't appear to exist!";
-    } else {
-        die "usage: $0 [-v|-q|-w|-C dir|-L filelist] [-maketest] [-makelist]\n",
-            "  -v          : Verbose Mode\n",
-            "  -q          : Quiet Mode\n",
-            "  -w          : Write files regardless\n",
-            "  -maketest   : Make test script\n",
-            "  -makelist   : Rewrite the file list based on current setup\n",
-            "  -L filelist : Use this file list, (defaults to $0.lst)\n",
-            "  -C dir      : Change to this directory before proceeding\n",
-            "  -check A B  : Executes only if A and B are the same\n";   
-    }
-}
-
-if ($FileList) {
-    print "Reading file list '$FileList'\n"
-        if $Verbose;
-    open my $fh,"<",$FileList or die "Failed to read '$FileList':$!";
-    my @input;
-    my @output;
-    for my $list ( \@input, \@output ) {
-        while (<$fh>) {
-            s/^ \s+ | \s+ $//xg;
-            next if /^ \s* (?: \# .* )? $/x;
-            last if /^ =+ $/x;
-            my ( $file ) = split /\t/, $_;
-            push @$list, $file;
-        }
-        my %dupe;
-        @$list = grep !$dupe{ $_ }++, @$list;
-    }
-    close $fh;
-    die "No input or output files in '$FileList'!"
-        if !@input or !@output;
-    if ( $MakeList ) {
-        foreach my $file (@output) {
-            unlink $file;
-        }
-    }            
-    if ( $Verbose ) {
-        print "Expecting ".scalar( @input )." input files. ",
-              "Checking ".scalar( @output )." output files.\n";
-    }
-    # we set maxtime to be the youngest input file, including $0 itself.
-    my $maxtime = -M $0; # do this before the chdir!
-    if ($UseDir) {
-        chdir $UseDir or die "Failed to chdir to '$UseDir':$!";
-    }
-    foreach my $in (@input) {
-        my $time = -M $in;
-        die "Missing input file '$in'" unless defined $time;
-        $maxtime = $time if $maxtime < $time;
-    }
-
-    # now we check to see if any output files are older than maxtime, if
-    # they are we need to continue on, otherwise we can presumably bail.
-    my $ok = 1;
-    foreach my $out (@output) {
-        if ( ! -e $out ) {
-            print "'$out' is missing.\n"
-                if $Verbose;
-            $ok = 0;
-            last;
-        }
-        if ( -M $out > $maxtime ) {
-            print "'$out' is too old.\n"
-                if $Verbose;
-            $ok = 0;
-            last;
-        }
     }
-    if ($ok) {
-        print "Files seem to be ok, not bothering to rebuild.\n";
-        exit(0);
+    elsif ($arg eq '-P' && defined ($pod_directory = shift)) {
+        -d $pod_directory or croak "Directory '$pod_directory' doesn't exist";
+    }
+    elsif ($arg eq '-maketest'
+             || ($arg eq '-T' && defined ($t_path = shift))) {
+        $make_test_script = 1;
+        $t_path = 'TestProp.pl' unless defined $t_path;
     }
-    print "Must rebuild tables.\n"
-        if $Verbose;
-} else {
-    if ($Verbose) {
-        print "Not checking filelist.\n";
+    elsif ($arg eq '-makelist') {
+        $make_list = 1;
     }
-    if ($UseDir) {
-        chdir $UseDir or die "Failed to chdir to '$UseDir':$!";
+    elsif ($arg eq '-C' && defined ($use_directory = shift)) {
+        -d $use_directory or croak "Unknown directory '$use_directory'";
+    }
+    elsif ($arg eq '-L') {
+
+        # Existence not tested until have chdir'd
+        $file_list = shift;
+    }
+    elsif ($arg eq '-globlist') {
+        $glob_list = 1;
+    }
+    elsif ($arg eq '-c') {
+        $output_range_counts = ! $output_range_counts
+    }
+    else {
+        my $with_c = 'with';
+        $with_c .= 'out' if $output_range_counts;   # Complements the state
+        croak <<END;
+usage: $0 [-c|-p|-q|-v|-w] [-C dir] [-L filelist] [ -P pod_dir ]
+          [ -T test_file_path ] [-globlist] [-makelist] [-maketest]
+          [-check A B ]
+  -c          : Output comments $with_c number of code points in ranges
+  -q          : Quiet Mode: Only output serious warnings.
+  -p          : Set verbosity level to normal plus show progress.
+  -v          : Set Verbosity level high:  Show progress and non-serious
+                warnings
+  -w          : Write files regardless
+  -C dir      : Change to this directory before proceeding. All relative paths
+                except those specified by the -P and -T options will be done
+                with respect to this directory.
+  -P dir      : Output $pod_file file to directory 'dir'.
+  -T path     : Create a .t test file as 'path'
+  -L filelist : Use alternate 'filelist' instead of standard one
+  -globlist   : Take as input all non-Test *.txt files in current and sub
+                directories
+  -maketest   : Make test script
+  -makelist   : Rewrite the file list $file_list based on current setup
+  -check A B  : Executes $0 only if A and B are the same
+END
     }
 }
 
-foreach my $lib ('To', 'lib',
-		 map {File::Spec->catdir("lib",$_)}
-		 qw(gc_sc dt bc hst ea jt lb nt ccc)) {
-  next if -d $lib;
-  mkdir $lib, 0755 or die "mkdir '$lib': $!";
+# Stores the most-recently changed file.  If none have changed, can skip the
+# build
+my $youngest = -M $0;   # Do this before the chdir!
+
+# Change directories now, because need to read 'version' early.
+if ($use_directory) {
+    if ($pod_directory
+        && ! File::Spec->file_name_is_absolute($pod_directory))
+    {
+        $pod_directory = File::Spec->rel2abs($pod_directory);
+    }
+    if ($t_path
+        && ! File::Spec->file_name_is_absolute($t_path))
+    {
+        $t_path = File::Spec->rel2abs($t_path);
+    }
+    chdir $use_directory or croak "Failed to chdir to '$use_directory':$!";
+    if ($pod_directory
+        && File::Spec->file_name_is_absolute($pod_directory))
+    {
+        $pod_directory = File::Spec->abs2rel($pod_directory);
+    }
+    if ($t_path
+        && File::Spec->file_name_is_absolute($t_path))
+    {
+        $t_path = File::Spec->abs2rel($t_path);
+    }
 }
 
-my $LastUnicodeCodepoint = 0x10FFFF; # As of Unicode 5.1.
+# Get Unicode version into regular and v-string.  This is done now because
+# various tables below get populated based on it.  These tables are populated
+# here to be near the top of the file, and so easily seeable by those needing
+# to modify things.
+open my $VERSION, "<", "version"
+                    or croak "$0: can't open required file 'version': $!\n";
+my $string_version = <$VERSION>;
+close $VERSION;
+chomp $string_version;
+my $v_version = pack "C*", split /\./, $string_version;        # v string
+
+# The following are the complete names of properties with property values that
+# are known to not match any code points in some versions of Unicode, but that
+# may change in the future so they should be matchable, hence an empty file is
+# generated for them.
+my @tables_that_may_be_empty = (
+                                'Joining_Type=Left_Joining',
+                                );
+push @tables_that_may_be_empty, 'Script=Common' if $v_version le v4.0.1;
+push @tables_that_may_be_empty, 'Title' if $v_version lt v2.0.0;
+push @tables_that_may_be_empty, 'Script=Katakana_Or_Hiragana'
+                                                    if $v_version ge v4.1.0;
+
+# The lists below are hashes, so the key is the item in the list, and the
+# value is the reason why it is in the list.  This makes generation of
+# documentation easier.
+
+my %why_suppressed;  # No file generated for these.
+
+# Files aren't generated for empty extraneous properties.  This is arguable.
+# Extraneous properties generally come about because a property is no longer
+# used in a newer version of Unicode.  If we generated a file without code
+# points, programs that used to work on that property will still execute
+# without errors.  It just won't ever match (or will always match, with \P{}).
+# This means that the logic is now likely wrong.  I (khw) think its better to
+# find this out by getting an error message.  Just move them to the table
+# above to change this behavior
+my %why_suppress_if_empty_warn_if_not = (
+
+   # It is the only property that has ever officially been removed from the
+   # Standard.  The database never contained any code points for it.
+   'Special_Case_Condition' => 'Obsolete',
+
+   # Apparently never official, but there were code points in some versions of
+   # old-style PropList.txt
+   'Non_Break' => 'Obsolete',
+);
 
-my $HEADER=<<"EOF";
-# !!!!!!!   DO NOT EDIT THIS FILE   !!!!!!! 
-# This file is built by $0 from e.g. UnicodeData.txt.
-# Any changes made here will be lost!
+# These would normally go in the warn table just above, but they were changed
+# a long time before this program was written, so warnings about them are
+# moot.
+if ($v_version gt v3.2.0) {
+    push @tables_that_may_be_empty,
+                                'Canonical_Combining_Class=Attached_Below_Left'
+}
 
+# These are listed in the Property aliases file in 5.2, but Unihan is ignored
+# unless explicitly added.
+if ($v_version ge v5.2.0) {
+    my $unihan = 'Unihan; remove from list if using Unihan';
+    foreach my $table qw ( 
+                           kAccountingNumeric
+                           kOtherNumeric
+                           kPrimaryNumeric
+                           kCompatibilityVariant
+                           kIICore
+                           kIRG_GSource
+                           kIRG_HSource
+                           kIRG_JSource
+                           kIRG_KPSource
+                           kIRG_MSource
+                           kIRG_KSource
+                           kIRG_TSource
+                           kIRG_USource
+                           kIRG_VSource
+                           kRSUnicode
+                        )
+    {
+        $why_suppress_if_empty_warn_if_not{$table} = $unihan;
+    }
+}
+
+# Properties that this program ignores.
+my @unimplemented_properties = (
+'Unicode_Radical_Stroke'    # Remove if changing to handle this one.
+);
+
+# There are several types of obsolete properties defined by Unicode.  These
+# must be hand-edited for every new Unicode release.
+my %why_deprecated;  # Generates a deprecated warning message if used.
+my %why_stabilized;  # Documentation only
+my %why_obsolete;    # Documentation only
+
+{   # Closure
+    my $simple = 'Perl uses the more complete version of this property';
+    my $unihan = 'Unihan properties are by default not enabled in the Perl core.  Instead use CPAN: Unicode::Unihan';
+
+    my $other_properties = 'other properties';
+    my $contributory = "Used by Unicode internally for generating $other_properties and not intended to be used stand-alone";
+    my $why_no_expand  = "Easily computed, and yet doesn't cover the common encoding forms (UTF-16/8)",
+
+    %why_deprecated = (
+        'Grapheme_Link' => 'Deprecated by Unicode.  Use ccc=vr (Canonical_Combining_Class=Virama) instead',
+        'Jamo_Short_Name' => $contributory,
+        'Line_Break=Surrogate' => 'Deprecated by Unicode because surrogates should never appear in well-formed text, and therefore shouldn\'t be the basis for line breaking',
+        'Other_Alphabetic' => $contributory,
+        'Other_Default_Ignorable_Code_Point' => $contributory,
+        'Other_Grapheme_Extend' => $contributory,
+        'Other_ID_Continue' => $contributory,
+        'Other_ID_Start' => $contributory,
+        'Other_Lowercase' => $contributory,
+        'Other_Math' => $contributory,
+        'Other_Uppercase' => $contributory,
+    );
+
+    %why_suppressed = (
+        # There is a lib/unicore/Decomposition.pl (used by normalize.pm) which
+        # contains the same information, but without the algorithmically
+        # determinable Hangul syllables'.  This file is not published, so it's
+        # existence is not noted in the comment.
+        'Decomposition_Mapping' => 'Accessible via Unicode::Normalize',
+
+        'ISO_Comment' => 'Apparently no demand for it, but can access it through Unicode::UCD::charinfo.  Obsoleted, and code points for it removed in Unicode 5.2',
+        'Unicode_1_Name' => "$simple, and no apparent demand for it, but can access it through Unicode::UCD::charinfo.  If there is no later name for a code point, then this one is used instead in charnames",
+
+        'Simple_Case_Folding' => "$simple.  Can access this through Unicode::UCD::casefold",
+        'Simple_Lowercase_Mapping' => "$simple.  Can access this through Unicode::UCD::charinfo",
+        'Simple_Titlecase_Mapping' => "$simple.  Can access this through Unicode::UCD::charinfo",
+        'Simple_Uppercase_Mapping' => "$simple.  Can access this through Unicode::UCD::charinfo",
+
+        'Name' => "Accessible via 'use charnames;'",
+        'Name_Alias' => "Accessible via 'use charnames;'",
+
+        # These are sort of jumping the gun; deprecation is proposed for
+        # Unicode version 6.0, but they have never been exposed by Perl, and
+        # likely are soon to be deprecated, so best not to expose them.
+        FC_NFKC_Closure => 'Use NFKC_Casefold instead',
+        Expands_On_NFC => $why_no_expand,
+        Expands_On_NFD => $why_no_expand,
+        Expands_On_NFKC => $why_no_expand,
+        Expands_On_NFKD => $why_no_expand,
+    );
+
+    # The following are suppressed because they were made contributory or
+    # deprecated by Unicode before Perl ever thought about supporting them.
+    foreach my $property ('Jamo_Short_Name', 'Grapheme_Link') {
+        $why_suppressed{$property} = $why_deprecated{$property};
+    }
+
+    # Customize the message for all the 'Other_' properties
+    foreach my $property (keys %why_deprecated) {
+        next if (my $main_property = $property) !~ s/^Other_//;
+        $why_deprecated{$property} =~ s/$other_properties/the $main_property property (which should be used instead)/;
+    }
+}
+
+if ($v_version ge 4.0.0) {
+    $why_stabilized{'Hyphen'} = 'Use the Line_Break property instead; see www.unicode.org/reports/tr14';
+}
+if ($v_version ge 5.2.0) {
+    $why_obsolete{'ISO_Comment'} = 'Code points for it have been removed';
+}
+
+# Probably obsolete forever
+if ($v_version ge v4.1.0) {
+    $why_suppressed{'Script=Katakana_Or_Hiragana'} = 'Obsolete.  All code points previously matched by this have been moved to "Script=Common"';
+}
+
+# This program can create files for enumerated-like properties, such as
+# 'Numeric_Type'.  This file would be the same format as for a string
+# property, with a mapping from code point to its value, so you could look up,
+# for example, the script a code point is in.  But no one so far wants this
+# mapping, or they have found another way to get it since this is a new
+# feature.  So no file is generated except if it is in this list.
+my @output_mapped_properties = split "\n", <<END;
+END
+
+# If you are using the Unihan database, you need to add the properties that
+# you want to extract from it to this table.  For your convenience, the
+# properties in the 5.2 PropertyAliases.txt file are listed, commented out
+my @cjk_properties = split "\n", <<'END';
+#cjkAccountingNumeric; kAccountingNumeric
+#cjkOtherNumeric; kOtherNumeric
+#cjkPrimaryNumeric; kPrimaryNumeric
+#cjkCompatibilityVariant; kCompatibilityVariant
+#cjkIICore ; kIICore
+#cjkIRG_GSource; kIRG_GSource
+#cjkIRG_HSource; kIRG_HSource
+#cjkIRG_JSource; kIRG_JSource
+#cjkIRG_KPSource; kIRG_KPSource
+#cjkIRG_KSource; kIRG_KSource
+#cjkIRG_TSource; kIRG_TSource
+#cjkIRG_USource; kIRG_USource
+#cjkIRG_VSource; kIRG_VSource
+#cjkRSUnicode; kRSUnicode                ; Unicode_Radical_Stroke; URS
+END
+
+# Similarly for the property values.  For your convenience, the lines in the
+# 5.2 PropertyAliases.txt file are listed.  Just remove the first BUT NOT both
+# '#' marks
+my @cjk_property_values = split "\n", <<'END';
+## @missing: 0000..10FFFF; cjkAccountingNumeric; NaN
+## @missing: 0000..10FFFF; cjkCompatibilityVariant; <code point>
+## @missing: 0000..10FFFF; cjkIICore; <none>
+## @missing: 0000..10FFFF; cjkIRG_GSource; <none>
+## @missing: 0000..10FFFF; cjkIRG_HSource; <none>
+## @missing: 0000..10FFFF; cjkIRG_JSource; <none>
+## @missing: 0000..10FFFF; cjkIRG_KPSource; <none>
+## @missing: 0000..10FFFF; cjkIRG_KSource; <none>
+## @missing: 0000..10FFFF; cjkIRG_TSource; <none>
+## @missing: 0000..10FFFF; cjkIRG_USource; <none>
+## @missing: 0000..10FFFF; cjkIRG_VSource; <none>
+## @missing: 0000..10FFFF; cjkOtherNumeric; NaN
+## @missing: 0000..10FFFF; cjkPrimaryNumeric; NaN
+## @missing: 0000..10FFFF; cjkRSUnicode; <none>
+END
+
+# The input files don't list every code point.  Those not listed are to be
+# defaulted to some value.  Below are hard-coded what those values are for
+# non-binary properties as of 5.1.  Starting in 5.0, there are
+# machine-parsable comment lines in the files the give the defaults; so this
+# list shouldn't have to be extended.  The claim is that all missing entries
+# for binary properties will default to 'N'.  Unicode tried to change that in
+# 5.2, but the beta period produced enough protest that they backed off.
+#
+# The defaults for the fields that appear in UnicodeData.txt in this hash must
+# be in the form that it expects.  The others may be synonyms.
+my $CODE_POINT = '<code point>';
+my %default_mapping = (
+    Age => "Unassigned",
+    # Bidi_Class => Complicated; set in code
+    Bidi_Mirroring_Glyph => "",
+    Block => 'No_Block',
+    Canonical_Combining_Class => 0,
+    Case_Folding => $CODE_POINT,
+    Decomposition_Mapping => $CODE_POINT,
+    Decomposition_Type => 'None',
+    East_Asian_Width => "Neutral",
+    FC_NFKC_Closure => $CODE_POINT,
+    General_Category => 'Cn',
+    Grapheme_Cluster_Break => 'Other',
+    Hangul_Syllable_Type => 'NA',
+    ISO_Comment => "",
+    Jamo_Short_Name => "",
+    Joining_Group => "No_Joining_Group",
+    # Joining_Type => Complicated; set in code
+    kIICore => 'N',   #                       Is converted to binary
+    #Line_Break => Complicated; set in code
+    Lowercase_Mapping => $CODE_POINT,
+    Name => "",
+    Name_Alias => "",
+    NFC_QC => 'Yes',
+    NFD_QC => 'Yes',
+    NFKC_QC => 'Yes',
+    NFKD_QC => 'Yes',
+    Numeric_Type => 'None',
+    Numeric_Value => 'NaN',
+    Script => ($v_version le 4.1.0) ? 'Common' : 'Unknown',
+    Sentence_Break => 'Other',
+    Simple_Case_Folding => $CODE_POINT,
+    Simple_Lowercase_Mapping => $CODE_POINT,
+    Simple_Titlecase_Mapping => $CODE_POINT,
+    Simple_Uppercase_Mapping => $CODE_POINT,
+    Titlecase_Mapping => $CODE_POINT,
+    Unicode_1_Name => "",
+    Unicode_Radical_Stroke => "",
+    Uppercase_Mapping => $CODE_POINT,
+    Word_Break => 'Other',
+);
+
+# Below are files that Unicode furnishes, but this program ignores, and why
+my %ignored_files = (
+    'CJKRadicals.txt' => 'Unihan data',
+    'Index.txt' => 'An index, not actual data',
+    'NamedSqProv.txt' => 'Not officially part of the Unicode standard; Append it to NamedSequences.txt if you want to process the contents.',
+    'NamesList.txt' => 'Just adds commentary',
+    'NormalizationCorrections.txt' => 'Data is already in other files.',
+    'Props.txt' => 'Adds nothing to PropList.txt; only in very early releases',
+    'ReadMe.txt' => 'Just comments',
+    'README.TXT' => 'Just comments',
+    'StandardizedVariants.txt' => 'Only for glyph changes, not a Unicode character property.  Does not fit into current scheme where one code point is mapped',
+);
+
+################ End of externally interesting definitions ###############
+
+my $HEADER=<<"EOF";
+# !!!!!!!   DO NOT EDIT THIS FILE   !!!!!!!
+# This file is machine-generated by $0 from the Unicode database,
+# Version $string_version.  Any changes made here will be lost!
 EOF
 
 my $INTERNAL_ONLY=<<"EOF";
+
+# !!!!!!!   INTERNAL PERL USE ONLY   !!!!!!!
 # This file is for internal use by the Perl program only.  The format and even
-# name or existence of this file are subject to change without notice.  Don't
-# use it directly.
+# the name or existence of this file are subject to change without notice.
+# Don't use it directly.
+EOF
+
+my $DEVELOPMENT_ONLY=<<"EOF";
+# !!!!!!!   DEVELOPMENT USE ONLY   !!!!!!!
+# This file contains information artificially constrained to code points
+# present in Unicode release $string_compare_versions.
+# IT CANNOT BE RELIED ON.  It is for use during development only and should
+# not be used for production.  
 
 EOF
 
-sub force_unlink {
-    my $filename = shift;
-    return unless -e $filename;
-    return if CORE::unlink($filename);
-    # We might need write permission
-    chmod 0777, $filename;
-    CORE::unlink($filename) or die "Couldn't unlink $filename: $!\n";
+my $LAST_UNICODE_CODEPOINT_STRING = "10FFFF";
+my $LAST_UNICODE_CODEPOINT = hex $LAST_UNICODE_CODEPOINT_STRING;
+my $MAX_UNICODE_CODEPOINTS = $LAST_UNICODE_CODEPOINT + 1;
+
+# Matches legal code point.  4-6 hex numbers, If there are 6, the first
+# two must be 10; if there are 5, the first must not be a 0.  Written this way
+# to decrease backtracking
+my $code_point_re =
+        qr/ \b (?: 10[0-9A-F]{4} | [1-9A-F][0-9A-F]{4} | [0-9A-F]{4} ) \b/x;
+
+# This matches the beginning of the line in the Unicode db files that give the
+# defaults for code points not listed (i.e., missing) in the file.  The code
+# depends on this ending with a semi-colon, so it can assume it is a valid
+# field when the line is split() by semi-colons
+my $missing_defaults_prefix =
+            qr/^#\s+\@missing:\s+0000\.\.$LAST_UNICODE_CODEPOINT_STRING\s*;/;
+
+# Property types.  Unicode has more types, but these are sufficient for our
+# purposes.
+my $UNKNOWN = -1;   # initialized to illegal value
+my $NON_STRING = 1; # Either binary or enum
+my $BINARY = 2;
+my $ENUM = 3;       # Include catalog
+my $STRING = 4;     # Anything else: string or misc
+
+# Some input files have lines that give default values for code points not
+# contained in the file.  Sometimes these should be ignored.
+my $NO_DEFAULTS = 0;        # Must evaluate to false
+my $NOT_IGNORED = 1;
+my $IGNORED = 2;
+
+# Range types.  Each range has a type.  Most ranges are type 0, for normal,
+# and will appear in the main body of the tables in the output files, but
+# there are other types of ranges as well, listed below, that are specially
+# handled.   There are pseudo-types as well that will never be stored as a
+# type, but will affect the calculation of the type.
+
+# 0 is for normal, non-specials
+my $MULTI_CP = 1;           # Sequence of more than code point
+my $HANGUL_SYLLABLE = 2;
+my $CP_IN_NAME = 3;         # The NAME contains the code point appended to it.
+my $NULL = 4;               # The map is to the null string; utf8.c can't
+                            # handle these, nor is there an accepted syntax
+                            # for them in \p{} constructs
+my $COMPUTE_NO_MULTI_CP = 5; # Pseudo-type; meanst that ranges that would
+                             # otherwise be $MULTI_CP type are instead type 0
+
+# process_generic_property_file() can accept certain overrides in its input.
+# Each of these must begin AND end with $CMD_DELIM.
+my $CMD_DELIM = "\a";
+my $REPLACE_CMD = 'replace';    # Override the Replace
+my $MAP_TYPE_CMD = 'map_type';  # Override the Type
+
+my $NO = 0;
+my $YES = 1;
+
+# Values for the Replace argument to add_range.
+# $NO                      # Don't replace; add only the code points not
+                           # already present.
+my $IF_NOT_EQUIVALENT = 1; # Replace only under certain conditions; details in
+                           # the comments at the subroutine definition.
+my $UNCONDITIONALLY = 2;   # Replace without conditions.
+my $MULTIPLE = 4;          # Don't replace, but add a duplicate record if
+                           # already there
+
+# Flags to give property statuses.  The phrases are to remind maintainers that
+# if the flag is changed, the indefinite article referring to it in the
+# documentation may need to be as well.
+my $NORMAL = "";
+my $SUPPRESSED = 'z';   # The character should never actually be seen, since
+                        # it is suppressed
+my $DEPRECATED = 'D';
+my $a_bold_deprecated = "a 'B<$DEPRECATED>'";
+my $A_bold_deprecated = "A 'B<$DEPRECATED>'";
+my $DISCOURAGED = 'X';
+my $a_bold_discouraged = "an 'B<$DISCOURAGED>'";
+my $A_bold_discouraged = "An 'B<$DISCOURAGED>'";
+my $STRICTER = 'T';
+my $a_bold_stricter = "a 'B<$STRICTER>'";
+my $A_bold_stricter = "A 'B<$STRICTER>'";
+my $STABILIZED = 'S';
+my $a_bold_stabilized = "an 'B<$STABILIZED>'";
+my $A_bold_stabilized = "An 'B<$STABILIZED>'";
+my $OBSOLETE = 'O';
+my $a_bold_obsolete = "an 'B<$OBSOLETE>'";
+my $A_bold_obsolete = "An 'B<$OBSOLETE>'";
+
+my %status_past_participles = (
+    $DISCOURAGED => 'discouraged',
+    $SUPPRESSED => 'should never be generated',
+    $STABILIZED => 'stabilized',
+    $OBSOLETE => 'obsolete',
+    $DEPRECATED => 'deprecated'
+);
+
+# The format of the values of the map tables:
+my $BINARY_FORMAT = 'b';
+my $DECIMAL_FORMAT = 'd';
+my $FLOAT_FORMAT = 'f';
+my $INTEGER_FORMAT = 'i';
+my $HEX_FORMAT = 'x';
+my $RATIONAL_FORMAT = 'r';
+my $STRING_FORMAT = 's';
+
+my %map_table_formats = (
+    $BINARY_FORMAT => 'binary',
+    $DECIMAL_FORMAT => 'single decimal digit',
+    $FLOAT_FORMAT => 'floating point number',
+    $INTEGER_FORMAT => 'integer',
+    $HEX_FORMAT => 'positive hex whole number; a code point',
+    $RATIONAL_FORMAT => 'rational: an integer or a fraction',
+    $STRING_FORMAT => 'arbitrary string',
+);
+
+# Unicode didn't put such derived files in a separate directory at first.
+my $EXTRACTED_DIR = (-d 'extracted') ? 'extracted' : "";
+my $EXTRACTED = ($EXTRACTED_DIR) ? "$EXTRACTED_DIR/" : "";
+my $AUXILIARY = 'auxiliary';
+
+# Hashes that will eventually go into Heavy.pl for the use of utf8_heavy.pl
+my %loose_to_file_of;       # loosely maps table names to their respective
+                            # files
+my %stricter_to_file_of;    # same; but for stricter mapping.
+my %nv_floating_to_rational; # maps numeric values floating point numbers to
+                             # their rational equivalent
+my %loose_property_name_of; # Loosely maps property names to standard form
+
+# These constants names and values were taken from the Unicode standard,
+# version 5.1, section 3.12.  They are used in conjunction with Hangul
+# syllables
+my $SBase = 0xAC00;
+my $LBase = 0x1100;
+my $VBase = 0x1161;
+my $TBase = 0x11A7;
+my $SCount = 11172;
+my $LCount = 19;
+my $VCount = 21;
+my $TCount = 28;
+my $NCount = $VCount * $TCount;
+
+# For Hangul syllables;  These store the numbers from Jamo.txt in conjunction
+# with the above published constants.
+my %Jamo;
+my %Jamo_L;     # Leading consonants
+my %Jamo_V;     # Vowels
+my %Jamo_T;     # Trailing consonants
+
+my @unhandled_properties;  # Will contain a list of properties found in
+                           # the input that we didn't process.
+my @match_properties;      # properties that have match tables, to be
+                           # listed in the pod
+my @map_properties;        # Properties that get map files written
+my @named_sequences;       # NamedSequences.txt contents.
+my %potential_files;       # Generated list of all .txt files in the directory
+                           # structure so we can warn if something is being
+                           # ignored.
+my @files_actually_output; # List of files we generated.
+my @more_Names;            # Some code point names are compound; this is used
+                           # to store the extra components of them.
+my $MIN_FRACTION_LENGTH = 3; # How many digits of a floating point number at
+                           # the minimum before we consider it equivalent to a
+                           # candidate rational
+my $MAX_FLOATING_SLOP = 10 ** - $MIN_FRACTION_LENGTH; # And in floating terms
+
+# These store references to certain commonly used property objects
+my $gc;
+my $perl;
+my $block;
+
+# Are there conflicting names because of beginning with 'In_', or 'Is_'
+my $has_In_conflicts = 0;
+my $has_Is_conflicts = 0;
+
+sub internal_file_to_platform ($) {
+    # Convert our file paths which have '/' separators to those of the
+    # platform.
+
+    my $file = shift;
+    return undef unless defined $file;
+
+    return File::Spec->join(split '/', $file);
 }
 
-##
-## Given a filename and a reference to an array of lines,
-## write the lines to the file only if the contents have not changed.
-## Filename can be given as an arrayref of directory names
-##
-sub WriteIfChanged($\@)
-{
-    my $file  = shift;
-    my $lines = shift;
+sub file_exists ($) {   # platform independent '-e'.  This program internally
+                        # uses slash as a path separator.
+    my $file = shift;
+    return 0 if ! defined $file;
+    return -e internal_file_to_platform($file);
+}
 
-    $file = File::Spec->catfile(@$file) if ref $file;
+# This 'require' doesn't necessarily work in miniperl, and even if it does,
+# the native perl version of it (which is what would operate under miniperl)
+# is extremely slow, as it does a string eval every call.
+my $has_fast_scalar_util = $ !~ /miniperl/
+                            && defined eval "require Scalar::Util";
 
-    my $TextToWrite = join '', @$lines;
-    if (open IN, $file) {
-        local($/) = undef;
-        my $PreviousText = <IN>;
-        close IN;
-        if ($PreviousText eq $TextToWrite) {
-            print "$file unchanged.\n" if $Verbose;
-            return unless $AlwaysWrite;
-        }
+sub objaddr($) {
+    # Returns the address of the blessed input object.  Uses the XS version if
+    # available.  It doesn't check for blessedness because that would do a
+    # string eval every call, and the program is structured so that this is
+    # never called for a non-blessed object.
+
+    return Scalar::Util::refaddr($_[0]) if $has_fast_scalar_util;
+
+    # Check at least that is a ref.
+    my $pkg = ref($_[0]) or return undef;
+
+    # Change to a fake package to defeat any overloaded stringify
+    bless $_[0], 'main::Fake';
+
+    # Numifying a ref gives its address.
+    my $addr = 0 + $_[0];
+
+    # Return to original class
+    bless $_[0], $pkg;
+    return $addr;
+}
+
+sub max ($$) {
+    my $a = shift;
+    my $b = shift;
+    return $a if $a >= $b;
+    return $b;
+}
+
+sub min ($$) {
+    my $a = shift;
+    my $b = shift;
+    return $a if $a <= $b;
+    return $b;
+}
+
+sub clarify_number ($) {
+    # This returns the input number with underscores inserted every 3 digits
+    # in large (5 digits or more) numbers.  Input must be entirely digits, not
+    # checked.
+
+    my $number = shift;
+    my $pos = length($number) - 3;
+    return $number if $pos <= 1;
+    while ($pos > 0) {
+        substr($number, $pos, 0) = '_';
+        $pos -= 3;
     }
-    force_unlink ($file);
-    if (not open OUT, ">$file") {
-        die "$0: can't open $file for output: $!\n";
-    }
-    print "$file written.\n" if $Verbose;
-
-    print OUT $TextToWrite;
-    close OUT;
-}
-
-##
-## The main datastructure (a "Table") represents a set of code points that
-## are part of a particular quality (that are part of \pL, \p{InGreek},
-## etc.). They are kept as ranges of code points (starting and ending of
-## each range).
-##
-## For example, a range ASCII LETTERS would be represented as:
-##   [ [ 0x41 => 0x5A, 'UPPER' ],
-##     [ 0x61 => 0x7A, 'LOWER, ] ]
-##
-sub RANGE_START() { 0 } ## index into range element
-sub RANGE_END()   { 1 } ## index into range element
-sub RANGE_NAME()  { 2 } ## index into range element
-
-## Conceptually, these should really be folded into the 'Table' objects
-my %TableInfo;
-my %TableDesc;
-my %FuzzyNames;
-my %AliasInfo;
-my %CanonicalToOrig;
-
-##
-## Turn something like
-##    OLD-ITALIC
-## into
-##    OldItalic
-##
-sub CanonicalName($)
-{
-    my $orig = shift;
-    my $name = lc $orig;
-    $name =~ s/(?<![a-z])(\w)/\u$1/g;
-    $name =~ s/[-_\s]+//g;
-
-    $CanonicalToOrig{$name} = $orig if not $CanonicalToOrig{$name};
-    return $name;
+    return $number;
 }
 
 
-##
-## Store the alias definitions for later use.
-##
-my %PropertyAlias;
-my %PropValueAlias;
-
-my %PA_reverse;
-my %PVA_reverse;
-
-sub Build_Aliases()
-{
-    ##
-    ## Most of the work with aliases doesn't occur here,
-    ## but rather in utf8_heavy.pl, which uses PVA.pl,
+package Carp;
 
-    # Placate the warnings about used only once. (They are used again, but
-    # via a typeglob lookup)
-    %utf8::PropertyAlias = ();
-    %utf8::PA_reverse = ();
-    %utf8::PropValueAlias = ();
-    %utf8::PVA_reverse = ();
-    %utf8::PVA_abbr_map = ();
+# These routines give a uniform treatment of messages in this program.  They
+# are placed in the Carp package to cause the stack trace to not include them,
+# although an alternative would be to use another package and set @CARP_NOT
+# for it.
 
-    open PA, "< PropertyAliases.txt"
-	or confess "Can't open PropertyAliases.txt: $!";
-    while (<PA>) {
-	s/#.*//;
-	s/\s+$//;
-	next if /^$/;
+our $Verbose = 1 if main::DEBUG;  # Useful info when debugging
 
-	my ($abbrev, $name) = split /\s*;\s*/;
-        next if $abbrev eq "n/a";
-	$PropertyAlias{$abbrev} = $name;
-        $PA_reverse{$name} = $abbrev;
+sub my_carp {
+    my $message = shift || "";
+    my $nofold = shift || 0;
 
-	# The %utf8::... versions use japhy's code originally from utf8_pva.pl
-	# However, it's moved here so that we build the tables at runtime.
-	tr/ _-//d for $abbrev, $name;
-	$utf8::PropertyAlias{lc $abbrev} = $name;
-	$utf8::PA_reverse{lc $name} = $abbrev;
-    }
-    close PA;
+    if ($message) {
+        $message = main::join_lines($message);
+        $message =~ s/^$0: *//;     # Remove initial program name
+        $message =~ s/[.;,]+$//;    # Remove certain ending punctuation
+        $message = "\n$0: $message;";
 
-    open PVA, "< PropValueAliases.txt"
-	or confess "Can't open PropValueAliases.txt: $!";
-    while (<PVA>) {
-	s/#.*//;
-	s/\s+$//;
-	next if /^$/;
+        # Fold the message with program name, semi-colon end punctuation
+        # (which looks good with the message that carp appends to it), and a
+        # hanging indent for continuation lines.
+        $message = main::simple_fold($message, "", 4) unless $nofold;
+        $message =~ s/\n$//;        # Remove the trailing nl so what carp
+                                    # appends is to the same line
+    }
 
-	my ($prop, @data) = split /\s*;\s*/;
+    return $message if defined wantarray;   # If a caller just wants the msg
 
-	if ($prop eq 'ccc') {
-	    $PropValueAlias{$prop}{$data[1]} = [ @data[0,2] ];
-	    $PVA_reverse{$prop}{$data[2]} = [ @data[0,1] ];
-	}
-	else {
-            next if $data[0] eq "n/a";
-	    $PropValueAlias{$prop}{$data[0]} = $data[1];
-            $PVA_reverse{$prop}{$data[1]} = $data[0];
-	}
+    carp $message;
+    return;
+}
 
-	shift @data if $prop eq 'ccc';
-	next if $data[0] eq "n/a";
+sub my_carp_bug {
+    # This is called when it is clear that the problem is caused by a bug in
+    # this program.
 
-	$data[1] =~ tr/ _-//d;
-	$utf8::PropValueAlias{$prop}{lc $data[0]} = $data[1];
-	$utf8::PVA_reverse{$prop}{lc $data[1]} = $data[0];
+    my $message = shift;
+    $message =~ s/^$0: *//;
+    $message = my_carp("Bug in $0.  Please report it by running perlbug or if that is unavailable, by sending email to perbug\@perl.org:\n$message");
+    carp $message;
+    return;
+}
 
-	my $abbr_class = ($prop eq 'gc' or $prop eq 'sc') ? 'gc_sc' : $prop;
-	$utf8::PVA_abbr_map{$abbr_class}{lc $data[0]} = $data[0];
+sub carp_too_few_args {
+    if (@_ != 2) {
+        my_carp_bug("Wrong number of arguments: to 'carp_too_few_arguments'.  No action taken.");
+        return;
     }
-    close PVA;
 
-    # backwards compatibility for L& -> LC
-    $utf8::PropValueAlias{gc}{'l&'} = $utf8::PropValueAlias{gc}{lc};
-    $utf8::PVA_abbr_map{gc_sc}{'l&'} = $utf8::PVA_abbr_map{gc_sc}{lc};
+    my $args_ref = shift;
+    my $count = shift;
 
+    my_carp_bug("Need at least $count arguments to "
+        . (caller 1)[3]
+        . ".  Instead got: '"
+        . join ', ', @$args_ref
+        . "'.  No action taken.");
+    return;
 }
 
+sub carp_extra_args {
+    my $args_ref = shift;
+    my_carp_bug("Too many arguments to 'carp_extra_args': (" . join(', ', @_) . ");  Extras ignored.") if @_;
 
-##
-## Associates a property ("Greek", "Lu", "Assigned",...) with a Table.
-##
-## Called like:
-##       New_Prop(In => 'Greek', $Table, Desc => 'Greek Block', Fuzzy => 1);
-##
-## Normally, these parameters are set when the Table is created (when the
-## Table->New constructor is called), but there are times when it needs to
-## be done after-the-fact...)
-##
-sub New_Prop($$$@)
-{
-    my $Type = shift; ## "Is" or "In";
-    my $Name = shift;
-    my $Table = shift;
+    unless (ref $args_ref) {
+        my_carp_bug("Argument to 'carp_extra_args' ($args_ref) must be a ref.  Not checking arguments.");
+        return;
+    }
+    my ($package, $file, $line) = caller;
+    my $subroutine = (caller 1)[3];
 
-    ## remaining args are optional key/val
-    my %Args = @_;
-
-    my $Fuzzy = delete $Args{Fuzzy};
-    my $Desc  = delete $Args{Desc}; # description
-
-    $Name = CanonicalName($Name) if $Fuzzy;
-
-    ## sanity check a few args
-    if (%Args or ($Type ne 'Is' and $Type ne 'In') or not ref $Table) {
-        confess "$0: bad args to New_Prop"
-    }
-
-    if (not $TableInfo{$Type}->{$Name})
-    {
-        $TableInfo{$Type}->{$Name} = $Table;
-        $TableDesc{$Type}->{$Name} = $Desc;
-        if ($Fuzzy) {
-            $FuzzyNames{$Type}->{$Name} = $Name;
+    my $list;
+    if (ref $args_ref eq 'HASH') {
+        foreach my $key (keys %$args_ref) {
+            $args_ref->{$key} = $UNDEF unless defined $args_ref->{$key};
         }
+        $list = join ', ', each %{$args_ref};
     }
+    elsif (ref $args_ref eq 'ARRAY') {
+        foreach my $arg (@$args_ref) {
+            $arg = $UNDEF unless defined $arg;
+        }
+        $list = join ', ', @$args_ref;
+    }
+    else {
+        my_carp_bug("Can't cope with ref "
+                . ref($args_ref)
+                . " . argument to 'carp_extra_args'.  Not checking arguments.");
+        return;
+    }
+
+    my_carp_bug("Unrecognized parameters in options: '$list' to $subroutine.  Skipped.");
+    return;
 }
 
+package main;
+
+{ # Closure
+
+    # This program uses the inside-out method for objects, as recommended in
+    # "Perl Best Practices".  This closure aids in generating those.  There
+    # are two routines.  setup_package() is called once per package to set
+    # things up, and then set_access() is called for each hash representing a
+    # field in the object.  These routines arrange for the object to be
+    # properly destroyed when no longer used, and for standard accessor
+    # functions to be generated.  If you need more complex accessors, just
+    # write your own and leave those accesses out of the call to set_access().
+    # More details below.
+
+    my %constructor_fields; # fields that are to be used in constructors; see
+                            # below
+
+    # The values of this hash will be the package names as keys to other
+    # hashes containing the name of each field in the package as keys, and
+    # references to their respective hashes as values.
+    my %package_fields;
+
+    sub setup_package {
+        # Sets up the package, creating standard DESTROY and dump methods
+        # (unless already defined).  The dump method is used in debugging by
+        # simple_dumper().
+        # The optional parameters are:
+        #   a)  a reference to a hash, that gets populated by later
+        #       set_access() calls with one of the accesses being
+        #       'constructor'.  The caller can then refer to this, but it is
+        #       not otherwise used by these two routines.
+        #   b)  a reference to a callback routine to call during destruction
+        #       of the object, before any fields are actually destroyed
+
+        my %args = @_;
+        my $constructor_ref = delete $args{'Constructor_Fields'};
+        my $destroy_callback = delete $args{'Destroy_Callback'};
+        Carp::carp_extra_args(\@_) if main::DEBUG && %args;
+
+        my %fields;
+        my $package = (caller)[0];
+
+        $package_fields{$package} = \%fields;
+        $constructor_fields{$package} = $constructor_ref;
+
+        unless ($package->can('DESTROY')) {
+            my $destroy_name = "${package}::DESTROY";
+            no strict "refs";
+
+            # Use typeglob to give the anonymous subroutine the name we want
+            *$destroy_name = sub {
+                my $self = shift;
+                my $addr = main::objaddr($self);
+
+                $self->$destroy_callback if $destroy_callback;
+                foreach my $field (keys %{$package_fields{$package}}) {
+                    #print STDERR __LINE__, ": Destroying ", ref $self, " ", sprintf("%04X", $addr), ": ", $field, "\n";
+                    delete $package_fields{$package}{$field}{$addr};
+                }
+                return;
+            }
+        }
+
+        unless ($package->can('dump')) {
+            my $dump_name = "${package}::dump";
+            no strict "refs";
+            *$dump_name = sub {
+                my $self = shift;
+                return dump_inside_out($self, $package_fields{$package}, @_);
+            }
+        }
+        return;
+    }
 
-##
-## Creates a new Table object.
-##
-## Args are key/value pairs:
-##    In => Name         -- Name of "In" property to be associated with
-##    Is => Name         -- Name of "Is" property to be associated with
-##    Fuzzy => Boolean   -- True if name can be accessed "fuzzily"
-##    Desc  => String    -- Description of the property
-##
-## No args are required.
-##
-sub Table::New
-{
-    my $class = shift;
-    my %Args = @_;
+    sub set_access {
+        # Arrange for the input field to be garbage collected when no longer
+        # needed.  Also, creates standard accessor functions for the field
+        # based on the optional parameters-- none if none of these parameters:
+        #   'addable'    creates an 'add_NAME()' accessor function.
+        #   'readable' or 'readable_array'   creates a 'NAME()' accessor
+        #                function.
+        #   'settable'   creates a 'set_NAME()' accessor function.
+        #   'constructor' doesn't create an accessor function, but adds the
+        #                field to the hash that was previously passed to
+        #                setup_package();
+        # Any of the accesses can be abbreviated down, so that 'a', 'ad',
+        # 'add' etc. all mean 'addable'.
+        # The read accessor function will work on both array and scalar
+        # values.  If another accessor in the parameter list is 'a', the read
+        # access assumes an array.  You can also force it to be array access
+        # by specifying 'readable_array' instead of 'readable'
+        #
+        # A sort-of 'protected' access can be set-up by preceding the addable,
+        # readable or settable with some initial portion of 'protected_' (but,
+        # the underscore is required), like 'p_a', 'pro_set', etc.  The
+        # "protection" is only by convention.  All that happens is that the
+        # accessor functions' names begin with an underscore.  So instead of
+        # calling set_foo, the call is _set_foo.  (Real protection could be
+        # accomplished by having a new subroutine, end_package called at the
+        # end of each package, and then storing the __LINE__ ranges and
+        # checking them on every accessor.  But that is way overkill.)
+
+        # We create anonymous subroutines as the accessors and then use
+        # typeglobs to assign them to the proper package and name
+
+        my $name = shift;   # Name of the field
+        my $field = shift;  # Reference to the inside-out hash containing the
+                            # field
+
+        my $package = (caller)[0];
+
+        if (! exists $package_fields{$package}) {
+            croak "$0: Must call 'setup_package' before 'set_access'";
+        }
 
-    my $Table = bless [], $class;
+        # Stash the field so DESTROY can get it.
+        $package_fields{$package}{$name} = $field;
 
-    my $Fuzzy = delete $Args{Fuzzy};
-    my $Desc  = delete $Args{Desc};
+        # Remaining arguments are the accessors.  For each...
+        foreach my $access (@_) {
+            my $access = lc $access;
 
-    for my $Type ('Is', 'In')
-    {
-        if (my $Name = delete $Args{$Type}) {
-            New_Prop($Type => $Name, $Table, Desc => $Desc, Fuzzy => $Fuzzy);
+            my $protected = "";
+
+            # Match the input as far as it goes.
+            if ($access =~ /^(p[^_]*)_/) {
+                $protected = $1;
+                if (substr('protected_', 0, length $protected)
+                    eq $protected)
+                {
+
+                    # Add 1 for the underscore not included in $protected
+                    $access = substr($access, length($protected) + 1);
+                    $protected = '_';
+                }
+                else {
+                    $protected = "";
+                }
+            }
+
+            if (substr('addable', 0, length $access) eq $access) {
+                my $subname = "${package}::${protected}add_$name";
+                no strict "refs";
+
+                # add_ accessor.  Don't add if already there, which we
+                # determine using 'eq' for scalars and '==' otherwise.
+                *$subname = sub {
+                    use strict "refs";
+                    return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
+                    my $self = shift;
+                    my $value = shift;
+                    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+                    if (ref $value) {
+                        return if grep { $value == $_ }
+                                            @{$field->{main::objaddr $self}};
+                    }
+                    else {
+                        return if grep { $value eq $_ }
+                                            @{$field->{main::objaddr $self}};
+                    }
+                    push @{$field->{main::objaddr $self}}, $value;
+                    return;
+                }
+            }
+            elsif (substr('constructor', 0, length $access) eq $access) {
+                if ($protected) {
+                    Carp::my_carp_bug("Can't set-up 'protected' constructors")
+                }
+                else {
+                    $constructor_fields{$package}{$name} = $field;
+                }
+            }
+            elsif (substr('readable_array', 0, length $access) eq $access) {
+
+                # Here has read access.  If one of the other parameters for
+                # access is array, or this one specifies array (by being more
+                # than just 'readable_'), then create a subroutine that
+                # assumes the data is an array.  Otherwise just a scalar
+                my $subname = "${package}::${protected}$name";
+                if (grep { /^a/i } @_
+                    or length($access) > length('readable_'))
+                {
+                    no strict "refs";
+                    *$subname = sub {
+                        use strict "refs";
+                        my $self = shift;
+                        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+                        my $addr = main::objaddr $self;
+                        if (ref $field->{$addr} ne 'ARRAY') {
+                            my $type = ref $field->{$addr};
+                            $type = 'scalar' unless $type;
+                            Carp::my_carp_bug("Trying to read $name as an array when it is a $type.  Big problems.");
+                            return;
+                        }
+                        return scalar @{$field->{$addr}} unless wantarray;
+
+                        # Make a copy; had problems with caller modifying the
+                        # original otherwise
+                        my @return = @{$field->{$addr}};
+                        return @return;
+                    }
+                }
+                else {
+
+                    # Here not an array value, a simpler function.
+                    no strict "refs";
+                    *$subname = sub {
+                        use strict "refs";
+                        my $self = shift;
+                        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+                        return $field->{main::objaddr $self};
+                    }
+                }
+            }
+            elsif (substr('settable', 0, length $access) eq $access) {
+                my $subname = "${package}::${protected}set_$name";
+                no strict "refs";
+                *$subname = sub {
+                    use strict "refs";
+                    return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
+                    my $self = shift;
+                    my $value = shift;
+                    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+                    $field->{main::objaddr $self} = $value;
+                    return;
+                }
+            }
+            else {
+                Carp::my_carp_bug("Unknown accessor type $access.  No accessor set.");
+            }
         }
+        return;
     }
-
-    ## shouldn't have any left over
-    if (%Args) {
-        confess "$0: bad args to Table->New"
-    }
-
-    return $Table;
-}
-
-
-##
-## Returns the maximum code point currently in the table.
-##
-sub Table::Max
-{
-    my $last = $_[0]->[-1];      ## last code point
-    confess "oops" unless $last; ## must have code points to have a max
-    return $last->[RANGE_END];
-}
-
-##
-## Replaces the codepoints in the Table with those in the Table given
-## as an arg. (NOTE: this is not a "deep copy").
-##
-sub Table::Replace($$)
-{
-    my $Table = shift; #self
-    my $New   = shift;
-
-    @$Table = @$New;
-}
-
-##
-## Given a new code point, make the last range of the Table extend to
-## include the new (and all intervening) code points.
-##
-## Takes the time to make sure that the extension is valid.
-##
-sub Table::Extend
-{
-    my $Table = shift; #self
-    my $codepoint = shift;
-
-    my $PrevMax = $Table->Max;
-
-    confess "oops ($codepoint <= $PrevMax)" if $codepoint <= $PrevMax;
-
-    $Table->ExtendNoCheck($codepoint);
-}
-
-
-##
-## Given a new code point, make the last range of the Table extend to
-## include the new (and all intervening) code points.
-##
-## Does NOT check that the extension is valid.  Assumes that the caller
-## has already made this check.
-##
-sub Table::ExtendNoCheck
-{
-    ## Optmized adding: Assumes $Table and $codepoint as parms
-    $_[0]->[-1]->[RANGE_END] = $_[1];
-}
-
-##
-## Given a code point range start and end (and optional name), blindly
-## append them to the list of ranges for the Table.
-##
-## NOTE: Code points must be added in strictly ascending numeric order.
-##
-sub Table::RawAppendRange
-{
-    my $Table = shift; #self
-    my $start = shift;
-    my $end   = shift;
-    my $name  = shift;
-    $name = "" if not defined $name; ## warning: $name can be "0"
-
-    push @$Table, [ $start,    # RANGE_START
-                    $end,      # RANGE_END
-                    $name   ]; # RANGE_NAME
-}
-
-##
-## Given a code point (and optional name), add it to the Table.
-##
-## NOTE: Code points must be added in strictly ascending numeric order.
-##
-sub Table::Append
-{
-    my $Table     = shift; #self
-    my $codepoint = shift;
-    my $name      = shift;
-    $name = "" if not defined $name; ## warning: $name can be "0"
-
-    ##
-    ## If we've already got a range working, and this code point is the next
-    ## one in line, and if the name is the same, just extend the current range.
-    ##
-    my $last = $Table->[-1];
-    if ($last
-        and
-        $last->[RANGE_END] == $codepoint - 1
-        and
-        $last->[RANGE_NAME] eq $name)
-    {
-        $Table->ExtendNoCheck($codepoint);
-    }
-    else
-    {
-        $Table->RawAppendRange($codepoint, $codepoint, $name);
-    }
-}
-
-##
-## Given a code point range starting value and ending value (and name),
-## Add the range to the Table.
-##
-## NOTE: Code points must be added in strictly ascending numeric order.
-##
-sub Table::AppendRange
-{
-    my $Table = shift; #self
-    my $start = shift;
-    my $end   = shift;
-    my $name  = shift;
-    $name = "" if not defined $name; ## warning: $name can be "0"
-
-    $Table->Append($start, $name);
-    $Table->Extend($end) if $end > $start;
-}
-
-##
-## Return a new Table that represents all code points not in the Table.
-##
-sub Table::Invert
-{
-    my $Table = shift; #self
-
-    my $New = Table->New();
-    my $max = -1;
-    for my $range (@$Table)
-    {
-        my $start = $range->[RANGE_START];
-        my $end   = $range->[RANGE_END];
-        if ($start-1 >= $max+1) {
-            $New->AppendRange($max+1, $start-1, "");
-        }
-        $max = $end;
-    }
-    if ($max+1 < $LastUnicodeCodepoint) {
-        $New->AppendRange($max+1, $LastUnicodeCodepoint);
-    }
-    return $New;
-}
-
-##
-## Merges any number of other tables with $self, returning the new table.
-## (existing tables are not modified)
-##
-##
-## Args may be Tables, or individual code points (as integers).
-##
-## Can be called as either a constructor or a method.
-##
-sub Table::Merge
-{
-    shift(@_) if not ref $_[0]; ## if called as a constructor, lose the class
-    my @Tables = @_;
-
-    ## Accumulate all records from all tables
-    my @Records;
-    for my $Arg (@Tables)
-    {
-        if (ref $Arg) {
-            ## arg is a table -- get its ranges
-            push @Records, @$Arg;
-        } else {
-            ## arg is a codepoint, make a range
-            push @Records, [ $Arg, $Arg ]
+}
+
+package Input_file;
+
+# All input files use this object, which stores various attributes about them,
+# and provides for convenient, uniform handling.  The run method wraps the
+# processing.  It handles all the bookkeeping of opening, reading, and closing
+# the file, returning only significant input lines.
+#
+# Each object gets a handler which processes the body of the file, and is
+# called by run().  Most should use the generic, default handler, which has
+# code scrubbed to handle things you might not expect.  A handler should
+# basically be a while(next_line()) {...} loop.
+#
+# You can also set up handlers to
+#   1) call before the first line is read for pre processing
+#   2) call to adjust each line of the input before the main handler gets them
+#   3) call upon EOF before the main handler exits its loop
+#   4) call at the end for post processing
+#
+# $_ is used to store the input line, and is to be filtered by the
+# each_line_handler()s.  So, if the format of the line is not in the desired
+# format for the main handler, these are used to do that adjusting.  They can
+# be stacked (by enclosing them in an [ anonymous array ] in the constructor,
+# so the $_ output of one is used as the input to the next.  None of the other
+# handlers are stackable, but could easily be changed to be so.
+#
+# Most of the handlers can call insert_lines() or insert_adjusted_lines()
+# which insert the parameters as lines to be processed before the next input
+# file line is read.  This allows the EOF handler to flush buffers, for
+# example.  The difference between the two routines is that the lines inserted
+# by insert_lines() are subjected to the each_line_handler()s.  (So if you
+# called it from such a handler, you would get infinite recursion.)  Lines
+# inserted by insert_adjusted_lines() go directly to the main handler without
+# any adjustments.  If the  post-processing handler calls any of these, there
+# will be no effect.  Some error checking for these conditions could be added,
+# but it hasn't been done.
+#
+# carp_bad_line() should be called to warn of bad input lines, which clears $_
+# to prevent further processing of the line.  This routine will output the
+# message as a warning once, and then keep a count of the lines that have the
+# same message, and output that count at the end of the file's processing.
+# This keeps the number of messages down to a manageable amount.
+#
+# get_missings() should be called to retrieve any @missing input lines.
+# Messages will be raised if this isn't done if the options aren't to ignore
+# missings.
+
+sub trace { return main::trace(@_); }
+
+
+{ # Closure
+    # Keep track of fields that are to be put into the constructor.
+    my %constructor_fields;
+
+    main::setup_package(Constructor_Fields => \%constructor_fields);
+
+    my %file; # Input file name, required
+    main::set_access('file', \%file, qw{ c r });
+
+    my %first_released; # Unicode version file was first released in, required
+    main::set_access('first_released', \%first_released, qw{ c r });
+
+    my %handler;    # Subroutine to process the input file, defaults to
+                    # 'process_generic_property_file'
+    main::set_access('handler', \%handler, qw{ c });
+
+    my %property;
+    # name of property this file is for.  defaults to none, meaning not
+    # applicable, or is otherwise determinable, for example, from each line.
+    main::set_access('property', \%property, qw{ c });
+
+    my %optional;
+    # If this is true, the file is optional.  If not present, no warning is
+    # output.  If it is present, the string given by this parameter is
+    # evaluated, and if false the file is not processed.
+    main::set_access('optional', \%optional, 'c', 'r');
+
+    my %non_skip;
+    # This is used for debugging, to skip processing of all but a few input
+    # files.  Add 'non_skip => 1' to the constructor for those files you want
+    # processed when you set the $debug_skip global.
+    main::set_access('non_skip', \%non_skip, 'c');
+
+    my %each_line_handler;
+    # list of subroutines to look at and filter each non-comment line in the
+    # file.  defaults to none.  The subroutines are called in order, each is
+    # to adjust $_ for the next one, and the final one adjusts it for
+    # 'handler'
+    main::set_access('each_line_handler', \%each_line_handler, 'c');
+
+    my %has_missings_defaults;
+    # ? Are there lines in the file giving default values for code points
+    # missing from it?.  Defaults to NO_DEFAULTS.  Otherwise NOT_IGNORED is
+    # the norm, but IGNORED means it has such lines, but the handler doesn't
+    # use them.  Having these three states allows us to catch changes to the
+    # UCD that this program should track
+    main::set_access('has_missings_defaults',
+                                        \%has_missings_defaults, qw{ c r });
+
+    my %pre_handler;
+    # Subroutine to call before doing anything else in the file.  If undef, no
+    # such handler is called.
+    main::set_access('pre_handler', \%pre_handler, qw{ c });
+
+    my %eof_handler;
+    # Subroutine to call upon getting an EOF on the input file, but before
+    # that is returned to the main handler.  This is to allow buffers to be
+    # flushed.  The handler is expected to call insert_lines() or
+    # insert_adjusted() with the buffered material
+    main::set_access('eof_handler', \%eof_handler, qw{ c r });
+
+    my %post_handler;
+    # Subroutine to call after all the lines of the file are read in and
+    # processed.  If undef, no such handler is called.
+    main::set_access('post_handler', \%post_handler, qw{ c });
+
+    my %progress_message;
+    # Message to print to display progress in lieu of the standard one
+    main::set_access('progress_message', \%progress_message, qw{ c });
+
+    my %handle;
+    # cache open file handle, internal.  Is undef if file hasn't been
+    # processed at all, empty if has;
+    main::set_access('handle', \%handle);
+
+    my %added_lines;
+    # cache of lines added virtually to the file, internal
+    main::set_access('added_lines', \%added_lines);
+
+    my %errors;
+    # cache of errors found, internal
+    main::set_access('errors', \%errors);
+
+    my %missings;
+    # storage of '@missing' defaults lines
+    main::set_access('missings', \%missings);
+
+    sub new {
+        my $class = shift;
+
+        my $self = bless \do{ my $anonymous_scalar }, $class;
+        my $addr = main::objaddr($self);
+
+        # Set defaults
+        $handler{$addr} = \&main::process_generic_property_file;
+        $non_skip{$addr} = 0;
+        $has_missings_defaults{$addr} = $NO_DEFAULTS;
+        $handle{$addr} = undef;
+        $added_lines{$addr} = [ ];
+        $each_line_handler{$addr} = [ ];
+        $errors{$addr} = { };
+        $missings{$addr} = [ ];
+
+        # Two positional parameters.
+        $file{$addr} = main::internal_file_to_platform(shift);
+        $first_released{$addr} = shift;
+
+        # The rest of the arguments are key => value pairs
+        # %constructor_fields has been set up earlier to list all possible
+        # ones.  Either set or push, depending on how the default has been set
+        # up just above.
+        my %args = @_;
+        foreach my $key (keys %args) {
+            my $argument = $args{$key};
+
+            # Note that the fields are the lower case of the constructor keys
+            my $hash = $constructor_fields{lc $key};
+            if (! defined $hash) {
+                Carp::my_carp_bug("Unrecognized parameters '$key => $argument' to new() for $self.  Skipped");
+                next;
+            }
+            if (ref $hash->{$addr} eq 'ARRAY') {
+                if (ref $argument eq 'ARRAY') {
+                    foreach my $argument (@{$argument}) {
+                        next if ! defined $argument;
+                        push @{$hash->{$addr}}, $argument;
+                    }
+                }
+                else {
+                    push @{$hash->{$addr}}, $argument if defined $argument;
+                }
+            }
+            else {
+                $hash->{$addr} = $argument;
+            }
+            delete $args{$key};
+        };
+
+        # If the file has a property for it, it means that the property is not
+        # listed in the file's entries.  So add a handler to the list of line
+        # handlers to insert the property name into the lines, to provide a
+        # uniform interface to the final processing subroutine.
+        # the final code doesn't have to worry about that.
+        if ($property{$addr}) {
+            push @{$each_line_handler{$addr}}, \&_insert_property_into_line;
         }
+
+        if ($non_skip{$addr} && ! $debug_skip && $verbosity) {
+            print "Warning: " . __PACKAGE__ . " constructor for $file{$addr} has useless 'non_skip' in it\n";
+        }
+
+        return $self;
     }
 
-    ## sort by range start, with longer ranges coming first.
-    my ($first, @Rest) = sort {
-        ($a->[RANGE_START] <=> $b->[RANGE_START])
-          or
-        ($b->[RANGE_END]   <=> $b->[RANGE_END])
-    } @Records;
 
-    my $New = Table->New();
+    use overload
+        fallback => 0,
+        qw("") => "_operator_stringify",
+        "." => \&main::_operator_dot,
+    ;
 
-    ## Ensuring the first range is there makes the subsequent loop easier
-    $New->AppendRange($first->[RANGE_START],
-                      $first->[RANGE_END]);
+    sub _operator_stringify {
+        my $self = shift;
 
-    ## Fold in records so long as they add new information.
-    for my $set (@Rest)
-    {
-        my $start = $set->[RANGE_START];
-        my $end   = $set->[RANGE_END];
-        if ($start > $New->Max) {
-            $New->AppendRange($start, $end);
-        } elsif ($end > $New->Max) {
-            $New->ExtendNoCheck($end);
+        return __PACKAGE__ . " object for " . $self->file;
+    }
+
+    # flag to make sure extracted files are processed early
+    my $seen_non_extracted_non_age = 0;
+
+    sub run {
+        # Process the input object $self.  This opens and closes the file and
+        # calls all the handlers for it.  Currently,  this can only be called
+        # once per file, as it destroy's the EOF handler
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        my $file = $file{$addr};
+
+        # Don't process if not expecting this file (because released later
+        # than this Unicode version), and isn't there.  This means if someone
+        # copies it into an earlier version's directory, we will go ahead and
+        # process it.
+        return if $first_released{$addr} gt $v_version && ! -e $file;
+
+        # If in debugging mode and this file doesn't have the non-skip
+        # flag set, and isn't one of the critical files, skip it.
+        if ($debug_skip
+            && $first_released{$addr} ne v0
+            && ! $non_skip{$addr})
+        {
+            print "Skipping $file in debugging\n" if $verbosity;
+            return;
+        }
+
+        # File could be optional
+        if ($optional{$addr}){
+            return unless -e $file;
+            my $result = eval $optional{$addr};
+            if (! defined $result) {
+                Carp::my_carp_bug("Got '$@' when tried to eval $optional{$addr}.  $file Skipped.");
+                return;
+            }
+            if (! $result) {
+                if ($verbosity) {
+                    print STDERR "Skipping processing input file '$file' because '$optional{$addr}' is not true\n";
+                }
+                return;
+            }
+        }
+
+        if (! defined $file || ! -e $file) {
+
+            # If the file doesn't exist, see if have internal data for it
+            # (based on first_released being 0).
+            if ($first_released{$addr} eq v0) {
+                $handle{$addr} = 'pretend_is_open';
+            }
+            else {
+                if (! $optional{$addr}  # File could be optional
+                    && $v_version ge $first_released{$addr})
+                {
+                    print STDERR "Skipping processing input file '$file' because not found\n" if $v_version ge $first_released{$addr};
+                }
+                return;
+            }
+        }
+        else {
+
+            # Here, the file exists
+            if ($seen_non_extracted_non_age) {
+                if ($file =~ /$EXTRACTED/) {
+                    Carp::my_carp_bug(join_lines(<<END
+$file should be processed just after the 'Prop..Alias' files, and before
+anything not in the $EXTRACTED_DIR directory.  Proceeding, but the results may
+have subtle problems
+END
+                    ));
+                }
+            }
+            elsif ($EXTRACTED_DIR
+                    && $first_released{$addr} ne v0
+                    && $file !~ /$EXTRACTED/
+                    && $file ne 'DAge.txt')
+            {
+                # We don't set this (by the 'if' above) if we have no
+                # extracted directory, so if running on an early version,
+                # this test won't work.  Not worth worrying about.
+                $seen_non_extracted_non_age = 1;
+            }
+
+            # And mark the file as having being processed, and warn if it
+            # isn't a file we are expecting.  As we process the files,
+            # they are deleted from the hash, so any that remain at the
+            # end of the program are files that we didn't process.
+            Carp::my_carp("Was not expecting '$file'.") if
+                    ! delete $potential_files{File::Spec->rel2abs($file)}
+                    && ! defined $handle{$addr};
+
+            # Open the file, converting the slashes used in this program
+            # into the proper form for the OS
+            my $file_handle;
+            if (not open $file_handle, "<", $file) {
+                Carp::my_carp("Can't open $file.  Skipping: $!");
+                return 0;
+            }
+            $handle{$addr} = $file_handle; # Cache the open file handle
+        }
+
+        if ($verbosity >= $PROGRESS) {
+            if ($progress_message{$addr}) {
+                print "$progress_message{$addr}\n";
+            }
+            else {
+                # If using a virtual file, say so.
+                print "Processing ", (-e $file)
+                                       ? $file
+                                       : "substitute $file",
+                                     "\n";
+            }
+        }
+
+
+        # Call any special handler for before the file.
+        &{$pre_handler{$addr}}($self) if $pre_handler{$addr};
+
+        # Then the main handler
+        &{$handler{$addr}}($self);
+
+        # Then any special post-file handler.
+        &{$post_handler{$addr}}($self) if $post_handler{$addr};
+
+        # If any errors have been accumulated, output the counts (as the first
+        # error message in each class was output when it was encountered).
+        if ($errors{$addr}) {
+            my $total = 0;
+            my $types = 0;
+            foreach my $error (keys %{$errors{$addr}}) {
+                $total += $errors{$addr}->{$error};
+                delete $errors{$addr}->{$error};
+                $types++;
+            }
+            if ($total > 1) {
+                my $message
+                        = "A total of $total lines had errors in $file.  ";
+
+                $message .= ($types == 1)
+                            ? '(Only the first one was displayed.)'
+                            : '(Only the first of each type was displayed.)';
+                Carp::my_carp($message);
+            }
+        }
+
+        if (@{$missings{$addr}}) {
+            Carp::my_carp_bug("Handler for $file didn't look at all the \@missing lines.  Generated tables likely are wrong");
         }
+
+        # If a real file handle, close it.
+        close $handle{$addr} or Carp::my_carp("Can't close $file: $!") if
+                                                        ref $handle{$addr};
+        $handle{$addr} = "";   # Uses empty to indicate that has already seen
+                               # the file, as opposed to undef
+        return;
     }
 
-    return $New;
-}
+    sub next_line {
+        # Sets $_ to be the next logical input line, if any.  Returns non-zero
+        # if such a line exists.  'logical' means that any lines that have
+        # been added via insert_lines() will be returned in $_ before the file
+        # is read again.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        # Here the file is open (or if the handle is not a ref, is an open
+        # 'virtual' file).  Get the next line; any inserted lines get priority
+        # over the file itself.
+        my $adjusted;
+
+        LINE:
+        while (1) { # Loop until find non-comment, non-empty line
+            #local $to_trace = 1 if main::DEBUG;
+            my $inserted_ref = shift @{$added_lines{$addr}};
+            if (defined $inserted_ref) {
+                ($adjusted, $_) = @{$inserted_ref};
+                trace $adjusted, $_ if main::DEBUG && $to_trace;
+                return 1 if $adjusted;
+            }
+            else {
+                last if ! ref $handle{$addr}; # Don't read unless is real file
+                last if ! defined ($_ = readline $handle{$addr});
+            }
+            chomp;
+            trace $_ if main::DEBUG && $to_trace;
+
+            # See if this line is the comment line that defines what property
+            # value that code points that are not listed in the file should
+            # have.  The format or existence of these lines is not guaranteed
+            # by Unicode since they are comments, but the documentation says
+            # that this was added for machine-readability, so probably won't
+            # change.  This works starting in Unicode Version 5.0.  They look
+            # like:
+            #
+            # @missing: 0000..10FFFF; Not_Reordered
+            # @missing: 0000..10FFFF; Decomposition_Mapping; <code point>
+            # @missing: 0000..10FFFF; ; NaN
+            #
+            # Save the line for a later get_missings() call.
+            if (/$missing_defaults_prefix/) {
+                if ($has_missings_defaults{$addr} == $NO_DEFAULTS) {
+                    $self->carp_bad_line("Unexpected \@missing line.  Assuming no missing entries");
+                }
+                elsif ($has_missings_defaults{$addr} == $NOT_IGNORED) {
+                    my @defaults = split /\s* ; \s*/x, $_;
+
+                    # The first field is the @missing, which ends in a
+                    # semi-colon, so can safely shift.
+                    shift @defaults;
+
+                    # Some of these lines may have empty field placeholders
+                    # which get in the way.  An example is:
+                    # @missing: 0000..10FFFF; ; NaN
+                    # Remove them.  Process starting from the top so the
+                    # splice doesn't affect things still to be looked at.
+                    for (my $i = @defaults - 1; $i >= 0; $i--) {
+                        next if $defaults[$i] ne "";
+                        splice @defaults, $i, 1;
+                    }
 
-##
-## Given a filename, write a representation of the Table to a file.
-## May have an optional comment as a 2nd arg.
-## Filename may actually be an arrayref of directories
-##
-sub Table::Write
-{
-    my $Table    = shift; #self
-    my $filename = shift;
-    my $comment  = shift;
+                    # What's left should be just the property (maybe) and the
+                    # default.  Having only one element means it doesn't have
+                    # the property.
+                    my $default;
+                    my $property;
+                    if (@defaults >= 1) {
+                        if (@defaults == 1) {
+                            $default = $defaults[0];
+                        }
+                        else {
+                            $property = $defaults[0];
+                            $default = $defaults[1];
+                        }
+                    }
+
+                    if (@defaults < 1
+                        || @defaults > 2
+                        || ($default =~ /^</
+                            && $default !~ /^<code *point>$/i
+                            && $default !~ /^<none>$/i))
+                    {
+                        $self->carp_bad_line("Unrecognized \@missing line: $_.  Assuming no missing entries");
+                    }
+                    else {
+
+                        # If the property is missing from the line, it should
+                        # be the one for the whole file
+                        $property = $property{$addr} if ! defined $property;
+
+                        # Change <none> to the null string, which is what it
+                        # really means.  If the default is the code point
+                        # itself, set it to <code point>, which is what
+                        # Unicode uses (but sometimes they've forgotten the
+                        # space)
+                        if ($default =~ /^<none>$/i) {
+                            $default = "";
+                        }
+                        elsif ($default =~ /^<code *point>$/i) {
+                            $default = $CODE_POINT;
+                        }
+
+                        # Store them as a sub-arrays with both components.
+                        push @{$missings{$addr}}, [ $default, $property ];
+                    }
+                }
+
+                # There is nothing for the caller to process on this comment
+                # line.
+                next;
+            }
 
-    my @OUT = $HEADER;
+            # Remove comments and trailing space, and skip this line if the
+            # result is empty
+            s/#.*//;
+            s/\s+$//;
+            next if /^$/;
+
+            # Call any handlers for this line, and skip further processing of
+            # the line if the handler sets the line to null.
+            foreach my $sub_ref (@{$each_line_handler{$addr}}) {
+                &{$sub_ref}($self);
+                next LINE if /^$/;
+            }
+
+            # Here the line is ok.  return success.
+            return 1;
+        } # End of looping through lines.
 
-    # files in subdirectories are internal-use-only
-    push @OUT, $INTERNAL_ONLY if ref $filename;
+        # If there is an EOF handler, call it (only once) and if it generates
+        # more lines to process go back in the loop to handle them.
+        if ($eof_handler{$addr}) {
+            &{$eof_handler{$addr}}($self);
+            $eof_handler{$addr} = "";   # Currently only get one shot at it.
+            goto LINE if $added_lines{$addr};
+        }
+
+        # Return failure -- no more lines.
+        return 0;
 
-    if (defined $comment) {
-        $comment =~ s/\s+\Z//;
-        $comment =~ s/^/# /gm;
-        push @OUT, "#\n$comment\n#\n";
     }
-    push @OUT, "return <<'END';\n";
 
-    for my $set (@$Table)
-    {
-        my $start = $set->[RANGE_START];
-        my $end   = $set->[RANGE_END];
-        my $name  = $set->[RANGE_NAME];
+#   Not currently used, not fully tested.
+#    sub peek {
+#        # Non-destructive look-ahead one non-adjusted, non-comment, non-blank
+#        # record.  Not callable from an each_line_handler(), nor does it call
+#        # an each_line_handler() on the line.
+#
+#        my $self = shift;
+#        my $addr = main::objaddr $self;
+#
+#        foreach my $inserted_ref (@{$added_lines{$addr}}) {
+#            my ($adjusted, $line) = @{$inserted_ref};
+#            next if $adjusted;
+#
+#            # Remove comments and trailing space, and return a non-empty
+#            # resulting line
+#            $line =~ s/#.*//;
+#            $line =~ s/\s+$//;
+#            return $line if $line ne "";
+#        }
+#
+#        return if ! ref $handle{$addr}; # Don't read unless is real file
+#        while (1) { # Loop until find non-comment, non-empty line
+#            local $to_trace = 1 if main::DEBUG;
+#            trace $_ if main::DEBUG && $to_trace;
+#            return if ! defined (my $line = readline $handle{$addr});
+#            chomp $line;
+#            push @{$added_lines{$addr}}, [ 0, $line ];
+#
+#            $line =~ s/#.*//;
+#            $line =~ s/\s+$//;
+#            return $line if $line ne "";
+#        }
+#
+#        return;
+#    }
+
+
+    sub insert_lines {
+        # Lines can be inserted so that it looks like they were in the input
+        # file at the place it was when this routine is called.  See also
+        # insert_adjusted_lines().  Lines inserted via this routine go through
+        # any each_line_handler()
+
+        my $self = shift;
+
+        # Each inserted line is an array, with the first element being 0 to
+        # indicate that this line hasn't been adjusted, and needs to be
+        # processed.
+        push @{$added_lines{main::objaddr $self}}, map { [ 0, $_ ] } @_;
+        return;
+    }
+
+    sub insert_adjusted_lines {
+        # Lines can be inserted so that it looks like they were in the input
+        # file at the place it was when this routine is called.  See also
+        # insert_lines().  Lines inserted via this routine are already fully
+        # adjusted, ready to be processed; each_line_handler()s handlers will
+        # not be called.  This means this is not a completely general
+        # facility, as only the last each_line_handler on the stack should
+        # call this.  It could be made more general, by passing to each of the
+        # line_handlers their position on the stack, which they would pass on
+        # to this routine, and that would replace the boolean first element in
+        # the anonymous array pushed here, so that the next_line routine could
+        # use that to call only those handlers whose index is after it on the
+        # stack.  But this is overkill for what is needed now.
+
+        my $self = shift;
+        trace $_[0] if main::DEBUG && $to_trace;
+
+        # Each inserted line is an array, with the first element being 1 to
+        # indicate that this line has been adjusted
+        push @{$added_lines{main::objaddr $self}}, map { [ 1, $_ ] } @_;
+        return;
+    }
 
-        if ($start == $end) {
-            push @OUT, sprintf "%04X\t\t%s\n", $start, $name;
-        } else {
-            push @OUT, sprintf "%04X\t%04X\t%s\n", $start, $end, $name;
+    sub get_missings {
+        # Returns the stored up @missings lines' values, and clears the list.
+        # The values are in an array, consisting of the default in the first
+        # element, and the property in the 2nd.  However, since these lines
+        # can be stacked up, the return is an array of all these arrays.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        # If not accepting a list return, just return the first one.
+        return shift @{$missings{$addr}} unless wantarray;
+
+        my @return = @{$missings{$addr}};
+        undef @{$missings{$addr}};
+        return @return;
+    }
+
+    sub _insert_property_into_line {
+        # Add a property field to $_, if this file requires it.
+
+        my $property = $property{main::objaddr shift};
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        $_ =~ s/(;|$)/; $property$1/;
+        return;
+    }
+
+    sub carp_bad_line {
+        # Output consistent error messages, using either a generic one, or the
+        # one given by the optional parameter.  To avoid gazillions of the
+        # same message in case the syntax of a  file is way off, this routine
+        # only outputs the first instance of each message, incrementing a
+        # count so the totals can be output at the end of the file.
+
+        my $self = shift;
+        my $message = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        $message = 'Unexpected line' unless $message;
+
+        # No trailing punctuation so as to fit with our addenda.
+        $message =~ s/[.:;,]$//;
+
+        # If haven't seen this exact message before, output it now.  Otherwise
+        # increment the count of how many times it has occurred
+        unless ($errors{$addr}->{$message}) {
+            Carp::my_carp("$message in '$_' in "
+                            . $file{main::objaddr $self}
+                            . " at line $..  Skipping this line;");
+            $errors{$addr}->{$message} = 1;
         }
+        else {
+            $errors{$addr}->{$message}++;
+        }
+
+        # Clear the line to prevent any further (meaningful) processing of it.
+        $_ = "";
+
+        return;
     }
+} # End closure
 
-    push @OUT, "END\n";
+package Multi_Default;
 
-    WriteIfChanged($filename, @OUT);
-}
+# Certain properties in early versions of Unicode had more than one possible
+# default for code points missing from the files.  In these cases, one
+# default applies to everything left over after all the others are applied,
+# and for each of the others, there is a description of which class of code
+# points applies to it.  This object helps implement this by storing the
+# defaults, and for all but that final default, an eval string that generates
+# the class that it applies to.
+
+
+{   # Closure
+
+    main::setup_package();
+
+    my %class_defaults;
+    # The defaults structure for the classes
+    main::set_access('class_defaults', \%class_defaults);
+
+    my %other_default;
+    # The default that applies to everything left over.
+    main::set_access('other_default', \%other_default, 'r');
+
+
+    sub new {
+        # The constructor is called with default => eval pairs, terminated by
+        # the left-over default. e.g.
+        # Multi_Default->new(
+        #        'T' => '$gc->table("Mn") + $gc->table("Cf") - 0x200C
+        #               -  0x200D',
+        #        'R' => 'some other expression that evaluates to code points',
+        #        .
+        #        .
+        #        .
+        #        'U'));
+
+        my $class = shift;
+
+        my $self = bless \do{my $anonymous_scalar}, $class;
+        my $addr = main::objaddr($self);
+
+        while (@_ > 1) {
+            my $default = shift;
+            my $eval = shift;
+            $class_defaults{$addr}->{$default} = $eval;
+        }
+
+        $other_default{$addr} = shift;
+
+        return $self;
+    }
+
+    sub get_next_defaults {
+        # Iterates and returns the next class of defaults.
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
 
-## This used only for making the test script.
-## helper function
-sub IsUsable($)
-{
-    my $code = shift;
-    return 0 if $code <= 0x0000;                       ## don't use null
-    return 0 if $code >= $LastUnicodeCodepoint;        ## keep in range
-    return 0 if ($code >= 0xD800 and $code <= 0xDFFF); ## no surrogates
-    return 0 if ($code >= 0xFDD0 and $code <= 0xFDEF); ## utf8.c says no good
-    return 0 if (($code & 0xFFFF) == 0xFFFE);          ## utf8.c says no good
-    return 0 if (($code & 0xFFFF) == 0xFFFF);          ## utf8.c says no good
-    return 1;
+        my $addr = main::objaddr $self;
+
+        return each %{$class_defaults{$addr}};
+    }
 }
 
-## Return a code point that's part of the table.
-## Returns nothing if the table is empty (or covers only surrogates).
-## This used only for making the test script.
-sub Table::ValidCode
-{
-    my $Table = shift; #self
-    for my $set (@$Table) {
-        return $set->[RANGE_END] if IsUsable($set->[RANGE_END]);
+package Alias;
+
+# An alias is one of the names that a table goes by.  This class defines them
+# including some attributes.  Everything is currently setup in the
+# constructor.
+
+
+{   # Closure
+
+    main::setup_package();
+
+    my %name;
+    main::set_access('name', \%name, 'r');
+
+    my %loose_match;
+    # Determined by the constructor code if this name should match loosely or
+    # not.  The constructor parameters can override this, but it isn't fully
+    # implemented, as should have ability to override Unicode one's via
+    # something like a set_loose_match()
+    main::set_access('loose_match', \%loose_match, 'r');
+
+    my %make_pod_entry;
+    # Some aliases should not get their own entries because they are covered
+    # by a wild-card, and some we want to discourage use of.  Binary
+    main::set_access('make_pod_entry', \%make_pod_entry, 'r');
+
+    my %status;
+    # Aliases have a status, like deprecated, or even suppressed (which means
+    # they don't appear in documentation).  Enum
+    main::set_access('status', \%status, 'r');
+
+    my %externally_ok;
+    # Similarly, some aliases should not be considered as usable ones for
+    # external use, such as file names, or we don't want documentation to
+    # recommend them.  Boolean
+    main::set_access('externally_ok', \%externally_ok, 'r');
+
+    sub new {
+        my $class = shift;
+
+        my $self = bless \do { my $anonymous_scalar }, $class;
+        my $addr = main::objaddr($self);
+
+        $name{$addr} = shift;
+        $loose_match{$addr} = shift;
+        $make_pod_entry{$addr} = shift;
+        $externally_ok{$addr} = shift;
+        $status{$addr} = shift;
+
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        # Null names are never ok externally
+        $externally_ok{$addr} = 0 if $name{$addr} eq "";
+
+        return $self;
     }
-    return ();
 }
 
-## Return a code point that's not part of the table
-## Returns nothing if the table covers all code points.
-## This used only for making the test script.
-sub Table::InvalidCode
-{
-    my $Table = shift; #self
+package Range;
 
-    return 0x1234 if not @$Table;
+# A range is the basic unit for storing code points, and is described in the
+# comments at the beginning of the program.  Each range has a starting code
+# point; an ending code point (not less than the starting one); a value
+# that applies to every code point in between the two end-points, inclusive;
+# and an enum type that applies to the value.  The type is for the user's
+# convenience, and has no meaning here, except that a non-zero type is
+# considered to not obey the normal Unicode rules for having standard forms.
+#
+# The same structure is used for both map and match tables, even though in the
+# latter, the value (and hence type) is irrelevant and could be used as a
+# comment.  In map tables, the value is what all the code points in the range
+# map to.  Type 0 values have the standardized version of the value stored as
+# well, so as to not have to recalculate it a lot.
 
-    for my $set (@$Table)
-    {
-        if (IsUsable($set->[RANGE_END] + 1))
-        {
-            return $set->[RANGE_END] + 1;
+sub trace { return main::trace(@_); }
+
+{   # Closure
+
+    main::setup_package();
+
+    my %start;
+    main::set_access('start', \%start, 'r', 's');
+
+    my %end;
+    main::set_access('end', \%end, 'r', 's');
+
+    my %value;
+    main::set_access('value', \%value, 'r');
+
+    my %type;
+    main::set_access('type', \%type, 'r');
+
+    my %standard_form;
+    # The value in internal standard form.  Defined only if the type is 0.
+    main::set_access('standard_form', \%standard_form);
+
+    # Note that if these fields change, the dump() method should as well
+
+    sub new {
+        return Carp::carp_too_few_args(\@_, 3) if main::DEBUG && @_ < 3;
+        my $class = shift;
+
+        my $self = bless \do { my $anonymous_scalar }, $class;
+        my $addr = main::objaddr($self);
+
+        $start{$addr} = shift;
+        $end{$addr} = shift;
+
+        my %args = @_;
+
+        my $value = delete $args{'Value'};  # Can be 0
+        $value = "" unless defined $value;
+        $value{$addr} = $value;
+
+        $type{$addr} = delete $args{'Type'} || 0;
+
+        Carp::carp_extra_args(\%args) if main::DEBUG && %args;
+
+        if (! $type{$addr}) {
+            $standard_form{$addr} = main::standardize($value);
         }
 
-        if (IsUsable($set->[RANGE_START] - 1))
+        return $self;
+    }
+
+    use overload
+        fallback => 0,
+        qw("") => "_operator_stringify",
+        "." => \&main::_operator_dot,
+    ;
+
+    sub _operator_stringify {
+        my $self = shift;
+        my $addr = main::objaddr $self;
+
+        # Output it like '0041..0065 (value)'
+        my $return = sprintf("%04X", $start{$addr})
+                        .  '..'
+                        . sprintf("%04X", $end{$addr});
+        my $value = $value{$addr};
+        my $type = $type{$addr};
+        $return .= ' (';
+        $return .= "$value";
+        $return .= ", Type=$type" if $type != 0;
+        $return .= ')';
+
+        return $return;
+    }
+
+    sub standard_form {
+        # The standard form is the value itself if the standard form is
+        # undefined (that is if the value is special)
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        return $standard_form{$addr} if defined $standard_form{$addr};
+        return $value{$addr};
+    }
+
+    sub dump {
+        # Human, not machine readable.  For machine readable, comment out this
+        # entire routine and let the standard one take effect.
+        my $self = shift;
+        my $indent = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        my $return = $indent
+                    . sprintf("%04X", $start{$addr})
+                    . '..'
+                    . sprintf("%04X", $end{$addr})
+                    . " '$value{$addr}';";
+        if (! defined $standard_form{$addr}) {
+            $return .= "(type=$type{$addr})";
+        }
+        elsif ($standard_form{$addr} ne $value{$addr}) {
+            $return .= "(standard '$standard_form{$addr}')";
+        }
+        return $return;
+    }
+} # End closure
+
+package _Range_List_Base;
+
+# Base class for range lists.  A range list is simply an ordered list of
+# ranges, so that the ranges with the lowest starting numbers are first in it.
+#
+# When a new range is added that is adjacent to an existing range that has the
+# same value and type, it merges with it to form a larger range.
+#
+# Ranges generally do not overlap, except that there can be multiple entries
+# of single code point ranges.  This is because of NameAliases.txt.
+#
+# In this program, there is a standard value such that if two different
+# values, have the same standard value, they are considered equivalent.  This
+# value was chosen so that it gives correct results on Unicode data
+
+# There are a number of methods to manipulate range lists, and some operators
+# are overloaded to handle them.
+
+# Because of the slowness of pure Perl objaddr() on miniperl, and measurements
+# showing this package was using a lot of real time calculating that, the code
+# was changed to only calculate it once per call stack.  This is done by
+# consistently using the package variable $addr in routines, and only calling
+# objaddr() if it isn't defined, and setting that to be local, so that callees
+# will have it already.  It would be a good thing to change this. XXX
+
+sub trace { return main::trace(@_); }
+
+{ # Closure
+
+    our $addr;
+
+    main::setup_package();
+
+    my %ranges;
+    # The list of ranges
+    main::set_access('ranges', \%ranges, 'readable_array');
+
+    my %max;
+    # The highest code point in the list.  This was originally a method, but
+    # actual measurements said it was used a lot.
+    main::set_access('max', \%max, 'r');
+
+    my %each_range_iterator;
+    # Iterator position for each_range()
+    main::set_access('each_range_iterator', \%each_range_iterator);
+
+    my %owner_name_of;
+    # Name of parent this is attached to, if any.  Solely for better error
+    # messages.
+    main::set_access('owner_name_of', \%owner_name_of, 'p_r');
+
+    my %_search_ranges_cache;
+    # A cache of the previous result from _search_ranges(), for better
+    # performance
+    main::set_access('_search_ranges_cache', \%_search_ranges_cache);
+
+    sub new {
+        my $class = shift;
+        my %args = @_;
+
+        # Optional initialization data for the range list.
+        my $initialize = delete $args{'Initialize'};
+
+        my $self;
+
+        # Use _union() to initialize.  _union() returns an object of this
+        # class, which means that it will call this constructor recursively.
+        # But it won't have this $initialize parameter so that it won't
+        # infinitely loop on this.
+        return _union($class, $initialize, %args) if defined $initialize;
+
+        $self = bless \do { my $anonymous_scalar }, $class;
+        local $addr = main::objaddr($self);
+
+        # Optional parent object, only for debug info.
+        $owner_name_of{$addr} = delete $args{'Owner'};
+        $owner_name_of{$addr} = "" if ! defined $owner_name_of{$addr};
+
+        # Stringify, in case it is an object.
+        $owner_name_of{$addr} = "$owner_name_of{$addr}";
+
+        # This is used only for error messages, and so a colon is added
+        $owner_name_of{$addr} .= ": " if $owner_name_of{$addr} ne "";
+
+        Carp::carp_extra_args(\%args) if main::DEBUG && %args;
+
+        # Max is initialized to a negative value that isn't adjacent to 0,
+        # for simpler tests
+        $max{$addr} = -2;
+
+        $_search_ranges_cache{$addr} = 0;
+        $ranges{$addr} = [];
+
+        return $self;
+    }
+
+    use overload
+        fallback => 0,
+        qw("") => "_operator_stringify",
+        "." => \&main::_operator_dot,
+    ;
+
+    sub _operator_stringify {
+        my $self = shift;
+        local $addr = main::objaddr($self) if !defined $addr;
+
+        return "Range_List attached to '$owner_name_of{$addr}'"
+                                                if $owner_name_of{$addr};
+        return "anonymous Range_List " . \$self;
+    }
+
+    sub _union {
+        # Returns the union of the input code points.  It can be called as
+        # either a constructor or a method.  If called as a method, the result
+        # will be a new() instance of the calling object, containing the union
+        # of that object with the other parameter's code points;  if called as
+        # a constructor, the first parameter gives the class the new object
+        # should be, and the second parameter gives the code points to go into
+        # it.
+        # In either case, there are two parameters looked at by this routine;
+        # any additional parameters are passed to the new() constructor.
+        #
+        # The code points can come in the form of some object that contains
+        # ranges, and has a conventionally named method to access them; or
+        # they can be an array of individual code points (as integers); or
+        # just a single code point.
+        #
+        # If they are ranges, this routine doesn't make any effort to preserve
+        # the range values of one input over the other.  Therefore this base
+        # class should not allow _union to be called from other than
+        # initialization code, so as to prevent two tables from being added
+        # together where the range values matter.  The general form of this
+        # routine therefore belongs in a derived class, but it was moved here
+        # to avoid duplication of code.  The failure to overload this in this
+        # class keeps it safe.
+        #
+
+        my $self;
+        my @args;   # Arguments to pass to the constructor
+
+        my $class = shift;
+
+        # If a method call, will start the union with the object itself, and
+        # the class of the new object will be the same as self.
+        if (ref $class) {
+            $self = $class;
+            $class = ref $self;
+            push @args, $self;
+        }
+
+        # Add the other required parameter.
+        push @args, shift;
+        # Rest of parameters are passed on to the constructor
+
+        # Accumulate all records from both lists.
+        my @records;
+        for my $arg (@args) {
+            #local $to_trace = 0 if main::DEBUG;
+            trace "argument = $arg" if main::DEBUG && $to_trace;
+            if (! defined $arg) {
+                my $message = "";
+                if (defined $self) {
+                    $message .= $owner_name_of{main::objaddr $self};
+                }
+                Carp::my_carp_bug($message .= "Undefined argument to _union.  No union done.");
+                return;
+            }
+            $arg = [ $arg ] if ! ref $arg;
+            my $type = ref $arg;
+            if ($type eq 'ARRAY') {
+                foreach my $element (@$arg) {
+                    push @records, Range->new($element, $element);
+                }
+            }
+            elsif ($arg->isa('Range')) {
+                push @records, $arg;
+            }
+            elsif ($arg->can('ranges')) {
+                push @records, $arg->ranges;
+            }
+            else {
+                my $message = "";
+                if (defined $self) {
+                    $message .= $owner_name_of{main::objaddr $self};
+                }
+                Carp::my_carp_bug($message . "Cannot take the union of a $type.  No union done.");
+                return;
+            }
+        }
+
+        # Sort with the range containing the lowest ordinal first, but if
+        # two ranges start at the same code point, sort with the bigger range
+        # of the two first, because it takes fewer cycles.
+        @records = sort { ($a->start <=> $b->start)
+                                      or
+                                    # if b is shorter than a, b->end will be
+                                    # less than a->end, and we want to select
+                                    # a, so want to return -1
+                                    ($b->end <=> $a->end)
+                                   } @records;
+
+        my $new = $class->new(@_);
+
+        # Fold in records so long as they add new information.
+        for my $set (@records) {
+            my $start = $set->start;
+            my $end   = $set->end;
+            my $value   = $set->value;
+            if ($start > $new->max) {
+                $new->_add_delete('+', $start, $end, $value);
+            }
+            elsif ($end > $new->max) {
+                $new->_add_delete('+', $new->max +1, $end, $value);
+            }
+        }
+
+        return $new;
+    }
+
+    sub range_count {        # Return the number of ranges in the range list
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        local $addr = main::objaddr($self) if ! defined $addr;
+
+        return scalar @{$ranges{$addr}};
+    }
+
+    sub min {
+        # Returns the minimum code point currently in the range list, or if
+        # the range list is empty, 2 beyond the max possible.  This is a
+        # method because used so rarely, that not worth saving between calls,
+        # and having to worry about changing it as ranges are added and
+        # deleted.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        local $addr = main::objaddr($self) if ! defined $addr;
+
+        # If the range list is empty, return a large value that isn't adjacent
+        # to any that could be in the range list, for simpler tests
+        return $LAST_UNICODE_CODEPOINT + 2 unless scalar @{$ranges{$addr}};
+        return $ranges{$addr}->[0]->start;
+    }
+
+    sub contains {
+        # Boolean: Is argument in the range list?  If so returns $i such that:
+        #   range[$i]->end < $codepoint <= range[$i+1]->end
+        # which is one beyond what you want; this is so that the 0th range
+        # doesn't return false
+        my $self = shift;
+        my $codepoint = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        local $addr = main::objaddr $self if ! defined $addr;
+
+        my $i = $self->_search_ranges($codepoint);
+        return 0 unless defined $i;
+
+        # The search returns $i, such that
+        #   range[$i-1]->end < $codepoint <= range[$i]->end
+        # So is in the table if and only iff it is at least the start position
+        # of range $i.
+        return 0 if $ranges{$addr}->[$i]->start > $codepoint;
+        return $i + 1;
+    }
+
+    sub value_of {
+        # Returns the value associated with the code point, undef if none
+
+        my $self = shift;
+        my $codepoint = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        local $addr = main::objaddr $self if ! defined $addr;
+
+        my $i = $self->contains($codepoint);
+        return unless $i;
+
+        # contains() returns 1 beyond where we should look
+        return $ranges{$addr}->[$i-1]->value;
+    }
+
+    sub _search_ranges {
+        # Find the range in the list which contains a code point, or where it
+        # should go if were to add it.  That is, it returns $i, such that:
+        #   range[$i-1]->end < $codepoint <= range[$i]->end
+        # Returns undef if no such $i is possible (e.g. at end of table), or
+        # if there is an error.
+
+        my $self = shift;
+        my $code_point = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        local $addr = main::objaddr $self if ! defined $addr;
+
+        return if $code_point > $max{$addr};
+        my $r = $ranges{$addr};                # The current list of ranges
+        my $range_list_size = scalar @$r;
+        my $i;
+
+        use integer;        # want integer division
+
+        # Use the cached result as the starting guess for this one, because,
+        # an experiment on 5.1 showed that 90% of the time the cache was the
+        # same as the result on the next call (and 7% it was one less).
+        $i = $_search_ranges_cache{$addr};
+        $i = 0 if $i >= $range_list_size;   # Reset if no longer valid (prob.
+                                            # from an intervening deletion
+        #local $to_trace = 1 if main::DEBUG;
+        trace "previous \$i is still valid: $i" if main::DEBUG && $to_trace && $code_point <= $r->[$i]->end && ($i == 0 || $r->[$i-1]->end < $code_point);
+        return $i if $code_point <= $r->[$i]->end
+                     && ($i == 0 || $r->[$i-1]->end < $code_point);
+
+        # Here the cache doesn't yield the correct $i.  Try adding 1.
+        if ($i < $range_list_size - 1
+            && $r->[$i]->end < $code_point &&
+            $code_point <= $r->[$i+1]->end)
         {
-            return $set->[RANGE_START] - 1;
+            $i++;
+            trace "next \$i is correct: $i" if main::DEBUG && $to_trace;
+            $_search_ranges_cache{$addr} = $i;
+            return $i;
         }
+
+        # Here, adding 1 also didn't work.  We do a binary search to
+        # find the correct position, starting with current $i
+        my $lower = 0;
+        my $upper = $range_list_size - 1;
+        while (1) {
+            trace "top of loop i=$i:", sprintf("%04X", $r->[$lower]->start), "[$lower] .. ", sprintf("%04X", $r->[$i]->start), "[$i] .. ", sprintf("%04X", $r->[$upper]->start), "[$upper]" if main::DEBUG && $to_trace;
+
+            if ($code_point <= $r->[$i]->end) {
+
+                # Here we have met the upper constraint.  We can quit if we
+                # also meet the lower one.
+                last if $i == 0 || $r->[$i-1]->end < $code_point;
+
+                $upper = $i;        # Still too high.
+
+            }
+            else {
+
+                # Here, $r[$i]->end < $code_point, so look higher up.
+                $lower = $i;
+            }
+
+            # Split search domain in half to try again.
+            my $temp = ($upper + $lower) / 2;
+
+            # No point in continuing unless $i changes for next time
+            # in the loop.
+            if ($temp == $i) {
+
+                # We can't reach the highest element because of the averaging.
+                # So if one below the upper edge, force it there and try one
+                # more time.
+                if ($i == $range_list_size - 2) {
+
+                    trace "Forcing to upper edge" if main::DEBUG && $to_trace;
+                    $i = $range_list_size - 1;
+
+                    # Change $lower as well so if fails next time through,
+                    # taking the average will yield the same $i, and we will
+                    # quit with the error message just below.
+                    $lower = $i;
+                    next;
+                }
+                Carp::my_carp_bug("$owner_name_of{$addr}Can't find where the range ought to go.  No action taken.");
+                return;
+            }
+            $i = $temp;
+        } # End of while loop
+
+        if (main::DEBUG && $to_trace) {
+            trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i;
+            trace "i=  [ $i ]", $r->[$i];
+            trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < $range_list_size - 1;
+        }
+
+        # Here we have found the offset.  Cache it as a starting point for the
+        # next call.
+        $_search_ranges_cache{$addr} = $i;
+        return $i;
     }
-    return ();
-}
 
-###########################################################################
-###########################################################################
-###########################################################################
+    sub _add_delete {
+        # Add, replace or delete ranges to or from a list.  The $type
+        # parameter gives which:
+        #   '+' => insert or replace a range, returning a list of any changed
+        #          ranges.
+        #   '-' => delete a range, returning a list of any deleted ranges.
+        #
+        # The next three parameters give respectively the start, end, and
+        # value associated with the range.  'value' should be null unless the
+        # operation is '+';
+        #
+        # The range list is kept sorted so that the range with the lowest
+        # starting position is first in the list, and generally, adjacent
+        # ranges with the same values are merged into single larger one (see
+        # exceptions below).
+        #
+        # There are more parameters, all are key => value pairs:
+        #   Type    gives the type of the value.  It is only valid for '+'.
+        #           All ranges have types; if this parameter is omitted, 0 is
+        #           assumed.  Ranges with type 0 are assumed to obey the
+        #           Unicode rules for casing, etc; ranges with other types are
+        #           not.  Otherwise, the type is arbitrary, for the caller's
+        #           convenience, and looked at only by this routine to keep
+        #           adjacent ranges of different types from being merged into
+        #           a single larger range, and when Replace =>
+        #           $IF_NOT_EQUIVALENT is specified (see just below).
+        #   Replace  determines what to do if the range list already contains
+        #            ranges which coincide with all or portions of the input
+        #            range.  It is only valid for '+':
+        #       => $NO            means that the new value is not to replace
+        #                         any existing ones, but any empty gaps of the
+        #                         range list coinciding with the input range
+        #                         will be filled in with the new value.
+        #       => $UNCONDITIONALLY  means to replace the existing values with
+        #                         this one unconditionally.  However, if the
+        #                         new and old values are identical, the
+        #                         replacement is skipped to save cycles
+        #       => $IF_NOT_EQUIVALENT means to replace the existing values
+        #                         with this one if they are not equivalent.
+        #                         Ranges are equivalent if their types are the
+        #                         same, and they are the same string, or if
+        #                         both are type 0 ranges, if their Unicode
+        #                         standard forms are identical.  In this last
+        #                         case, the routine chooses the more "modern"
+        #                         one to use.  This is because some of the
+        #                         older files are formatted with values that
+        #                         are, for example, ALL CAPs, whereas the
+        #                         derived files have a more modern style,
+        #                         which looks better.  By looking for this
+        #                         style when the pre-existing and replacement
+        #                         standard forms are the same, we can move to
+        #                         the modern style
+        #       => $MULTIPLE      means that if this range duplicates an
+        #                         existing one, but has a different value,
+        #                         don't replace the existing one, but insert
+        #                         this, one so that the same range can occur
+        #                         multiple times.
+        #       => anything else  is the same as => $IF_NOT_EQUIVALENT
+        #
+        # "same value" means identical for type-0 ranges, and it means having
+        # the same standard forms for non-type-0 ranges.
+
+        return Carp::carp_too_few_args(\@_, 5) if main::DEBUG && @_ < 5;
+
+        my $self = shift;
+        my $operation = shift;   # '+' for add/replace; '-' for delete;
+        my $start = shift;
+        my $end   = shift;
+        my $value = shift;
+
+        my %args = @_;
+
+        $value = "" if not defined $value;        # warning: $value can be "0"
+
+        my $replace = delete $args{'Replace'};
+        $replace = $IF_NOT_EQUIVALENT unless defined $replace;
+
+        my $type = delete $args{'Type'};
+        $type = 0 unless defined $type;
+
+        Carp::carp_extra_args(\%args) if main::DEBUG && %args;
+
+        local $addr = main::objaddr($self) if ! defined $addr;
+
+        if ($operation ne '+' && $operation ne '-') {
+            Carp::my_carp_bug("$owner_name_of{$addr}First parameter to _add_delete must be '+' or '-'.  No action taken.");
+            return;
+        }
+        unless (defined $start && defined $end) {
+            Carp::my_carp_bug("$owner_name_of{$addr}Undefined start and/or end to _add_delete.  No action taken.");
+            return;
+        }
+        unless ($end >= $start) {
+            Carp::my_carp_bug("$owner_name_of{$addr}End of range (" . sprintf("%04X", $end) . ") must not be before start (" . sprintf("%04X", $start) . ").  No action taken.");
+            return;
+        }
+        #local $to_trace = 1 if main::DEBUG;
+
+        if ($operation eq '-') {
+            if ($replace != $IF_NOT_EQUIVALENT) {
+                Carp::my_carp_bug("$owner_name_of{$addr}Replace => \$IF_NOT_EQUIVALENT is required when deleting a range from a range list.  Assuming Replace => \$IF_NOT_EQUIVALENT.");
+                $replace = $IF_NOT_EQUIVALENT;
+            }
+            if ($type) {
+                Carp::my_carp_bug("$owner_name_of{$addr}Type => 0 is required when deleting a range from a range list.  Assuming Type => 0.");
+                $type = 0;
+            }
+            if ($value ne "") {
+                Carp::my_carp_bug("$owner_name_of{$addr}Value => \"\" is required when deleting a range from a range list.  Assuming Value => \"\".");
+                $value = "";
+            }
+        }
+
+        my $r = $ranges{$addr};               # The current list of ranges
+        my $range_list_size = scalar @$r;     # And its size
+        my $max = $max{$addr};                # The current high code point in
+                                              # the list of ranges
+
+        # Do a special case requiring fewer machine cycles when the new range
+        # starts after the current highest point.  The Unicode input data is
+        # structured so this is common.
+        if ($start > $max) {
+
+            trace "$owner_name_of{$addr} $operation", sprintf("%04X", $start) . '..' . sprintf("%04X", $end) . " ($value) type=$type" if main::DEBUG && $to_trace;
+            return if $operation eq '-'; # Deleting a non-existing range is a
+                                         # no-op
+
+            # If the new range doesn't logically extend the current final one
+            # in the range list, create a new range at the end of the range
+            # list.  (max cleverly is initialized to a negative number not
+            # adjacent to 0 if the range list is empty, so even adding a range
+            # to an empty range list starting at 0 will have this 'if'
+            # succeed.)
+            if ($start > $max + 1        # non-adjacent means can't extend.
+                || @{$r}[-1]->value ne $value # values differ, can't extend.
+                || @{$r}[-1]->type != $type # types differ, can't extend.
+            ) {
+                push @$r, Range->new($start, $end,
+                                     Value => $value,
+                                     Type => $type);
+            }
+            else {
+
+                # Here, the new range starts just after the current highest in
+                # the range list, and they have the same type and value.
+                # Extend the current range to incorporate the new one.
+                @{$r}[-1]->set_end($end);
+            }
+
+            # This becomes the new maximum.
+            $max{$addr} = $end;
+
+            return;
+        }
+        #local $to_trace = 0 if main::DEBUG;
+
+        trace "$owner_name_of{$addr} $operation", sprintf("%04X", $start) . '..' . sprintf("%04X", $end) . " ($value) replace=$replace" if main::DEBUG && $to_trace;
+
+        # Here, the input range isn't after the whole rest of the range list.
+        # Most likely 'splice' will be needed.  The rest of the routine finds
+        # the needed splice parameters, and if necessary, does the splice.
+        # First, find the offset parameter needed by the splice function for
+        # the input range.  Note that the input range may span multiple
+        # existing ones, but we'll worry about that later.  For now, just find
+        # the beginning.  If the input range is to be inserted starting in a
+        # position not currently in the range list, it must (obviously) come
+        # just after the range below it, and just before the range above it.
+        # Slightly less obviously, it will occupy the position currently
+        # occupied by the range that is to come after it.  More formally, we
+        # are looking for the position, $i, in the array of ranges, such that:
+        #
+        # r[$i-1]->start <= r[$i-1]->end < $start < r[$i]->start <= r[$i]->end
+        #
+        # (The ordered relationships within existing ranges are also shown in
+        # the equation above).  However, if the start of the input range is
+        # within an existing range, the splice offset should point to that
+        # existing range's position in the list; that is $i satisfies a
+        # somewhat different equation, namely:
+        #
+        #r[$i-1]->start <= r[$i-1]->end < r[$i]->start <= $start <= r[$i]->end
+        #
+        # More briefly, $start can come before or after r[$i]->start, and at
+        # this point, we don't know which it will be.  However, these
+        # two equations share these constraints:
+        #
+        #   r[$i-1]->end < $start <= r[$i]->end
+        #
+        # And that is good enough to find $i.
+
+        my $i = $self->_search_ranges($start);
+        if (! defined $i) {
+            Carp::my_carp_bug("Searching $self for range beginning with $start unexpectedly returned undefined.  Operation '$operation' not performed");
+            return;
+        }
+
+        # The search function returns $i such that:
+        #
+        # r[$i-1]->end < $start <= r[$i]->end
+        #
+        # That means that $i points to the first range in the range list
+        # that could possibly be affected by this operation.  We still don't
+        # know if the start of the input range is within r[$i], or if it
+        # points to empty space between r[$i-1] and r[$i].
+        trace "[$i] is the beginning splice point.  Existing range there is ", $r->[$i] if main::DEBUG && $to_trace;
+
+        # Special case the insertion of data that is not to replace any
+        # existing data.
+        if ($replace == $NO) {  # If $NO, has to be operation '+'
+            #local $to_trace = 1 if main::DEBUG;
+            trace "Doesn't replace" if main::DEBUG && $to_trace;
+
+            # Here, the new range is to take effect only on those code points
+            # that aren't already in an existing range.  This can be done by
+            # looking through the existing range list and finding the gaps in
+            # the ranges that this new range affects, and then calling this
+            # function recursively on each of those gaps, leaving untouched
+            # anything already in the list.  Gather up a list of the changed
+            # gaps first so that changes to the internal state as new ranges
+            # are added won't be a problem.
+            my @gap_list;
+
+            # First, if the starting point of the input range is outside an
+            # existing one, there is a gap from there to the beginning of the
+            # existing range -- add a span to fill the part that this new
+            # range occupies
+            if ($start < $r->[$i]->start) {
+                push @gap_list, Range->new($start,
+                                           main::min($end,
+                                                     $r->[$i]->start - 1),
+                                           Type => $type);
+                trace "gap before $r->[$i] [$i], will add", $gap_list[-1] if main::DEBUG && $to_trace;
+            }
+
+            # Then look through the range list for other gaps until we reach
+            # the highest range affected by the input one.
+            my $j;
+            for ($j = $i+1; $j < $range_list_size; $j++) {
+                trace "j=[$j]", $r->[$j] if main::DEBUG && $to_trace;
+                last if $end < $r->[$j]->start;
+
+                # If there is a gap between when this range starts and the
+                # previous one ends, add a span to fill it.  Note that just
+                # because there are two ranges doesn't mean there is a
+                # non-zero gap between them.  It could be that they have
+                # different values or types
+                if ($r->[$j-1]->end + 1 != $r->[$j]->start) {
+                    push @gap_list,
+                        Range->new($r->[$j-1]->end + 1,
+                                   $r->[$j]->start - 1,
+                                   Type => $type);
+                    trace "gap between $r->[$j-1] and $r->[$j] [$j], will add: $gap_list[-1]" if main::DEBUG && $to_trace;
+                }
+            }
+
+            # Here, we have either found an existing range in the range list,
+            # beyond the area affected by the input one, or we fell off the
+            # end of the loop because the input range affects the whole rest
+            # of the range list.  In either case, $j is 1 higher than the
+            # highest affected range.  If $j == $i, it means that there are no
+            # affected ranges, that the entire insertion is in the gap between
+            # r[$i-1], and r[$i], which we already have taken care of before
+            # the loop.
+            # On the other hand, if there are affected ranges, it might be
+            # that there is a gap that needs filling after the final such
+            # range to the end of the input range
+            if ($r->[$j-1]->end < $end) {
+                    push @gap_list, Range->new(main::max($start,
+                                                         $r->[$j-1]->end + 1),
+                                               $end,
+                                               Type => $type);
+                    trace "gap after $r->[$j-1], will add $gap_list[-1]" if main::DEBUG && $to_trace;
+            }
+
+            # Call recursively to fill in all the gaps.
+            foreach my $gap (@gap_list) {
+                $self->_add_delete($operation,
+                                   $gap->start,
+                                   $gap->end,
+                                   $value,
+                                   Type => $type);
+            }
+
+            return;
+        }
+
+        # Here, we have taken care of the case where $replace is $NO, which
+        # means that whatever action we now take is done unconditionally.  It
+        # still could be that this call will result in a no-op, if duplicates
+        # aren't allowed, and we are inserting a range that merely duplicates
+        # data already in the range list; or also if deleting a non-existent
+        # range.
+        # $i still points to the first potential affected range.  Now find the
+        # highest range affected, which will determine the length parameter to
+        # splice.  (The input range can span multiple existing ones.)  While
+        # we are looking through the range list, see also if this is an
+        # insertion that will change the values of at least one of the
+        # affected ranges.  We don't need to do this check unless this is an
+        # insertion of non-multiples, and also since this is a boolean, we
+        # don't need to do it if have already determined that it will make a
+        # change; just unconditionally change them.  $cdm is created to be 1
+        # if either of these is true. (The 'c' in the name comes from below)
+        my $cdm = ($operation eq '-' || $replace == $MULTIPLE);
+        my $j;        # This will point to the highest affected range
+
+        # For non-zero types, the standard form is the value itself;
+        my $standard_form = ($type) ? $value : main::standardize($value);
+
+        for ($j = $i; $j < $range_list_size; $j++) {
+            trace "Looking for highest affected range; the one at $j is ", $r->[$j] if main::DEBUG && $to_trace;
+
+            # If find a range that it doesn't overlap into, we can stop
+            # searching
+            last if $end < $r->[$j]->start;
+
+            # Here, overlaps the range at $j.  If the value's don't match,
+            # and this is supposedly an insertion, it becomes a change
+            # instead.  This is what the 'c' stands for in $cdm.
+            if (! $cdm) {
+                if ($r->[$j]->standard_form ne $standard_form) {
+                    $cdm = 1;
+                }
+                else {
+
+                    # Here, the two values are essentially the same.  If the
+                    # two are actually identical, replacing wouldn't change
+                    # anything so skip it.
+                    my $pre_existing = $r->[$j]->value;
+                    if ($pre_existing ne $value) {
+
+                        # Here the new and old standardized values are the
+                        # same, but the non-standardized values aren't.  If
+                        # replacing unconditionally, then replace
+                        if( $replace == $UNCONDITIONALLY) {
+                            $cdm = 1;
+                        }
+                        else {
+
+                            # Here, are replacing conditionally.  Decide to
+                            # replace or not based on which appears to look
+                            # the "nicest".  If one is mixed case and the
+                            # other isn't, choose the mixed case one.
+                            my $new_mixed = $value =~ /[A-Z]/
+                                            && $value =~ /[a-z]/;
+                            my $old_mixed = $pre_existing =~ /[A-Z]/
+                                            && $pre_existing =~ /[a-z]/;
+
+                            if ($old_mixed != $new_mixed) {
+                                $cdm = 1 if $new_mixed;
+                                if (main::DEBUG && $to_trace) {
+                                    if ($cdm) {
+                                        trace "Replacing $pre_existing with $value";
+                                    }
+                                    else {
+                                        trace "Retaining $pre_existing over $value";
+                                    }
+                                }
+                            }
+                            else {
+
+                                # Here casing wasn't different between the two.
+                                # If one has hyphens or underscores and the
+                                # other doesn't, choose the one with the
+                                # punctuation.
+                                my $new_punct = $value =~ /[-_]/;
+                                my $old_punct = $pre_existing =~ /[-_]/;
+
+                                if ($old_punct != $new_punct) {
+                                    $cdm = 1 if $new_punct;
+                                    if (main::DEBUG && $to_trace) {
+                                        if ($cdm) {
+                                            trace "Replacing $pre_existing with $value";
+                                        }
+                                        else {
+                                            trace "Retaining $pre_existing over $value";
+                                        }
+                                    }
+                                }   # else existing one is just as "good";
+                                    # retain it to save cycles.
+                            }
+                        }
+                    }
+                }
+            }
+        } # End of loop looking for highest affected range.
+
+        # Here, $j points to one beyond the highest range that this insertion
+        # affects (hence to beyond the range list if that range is the final
+        # one in the range list).
+
+        # The splice length is all the affected ranges.  Get it before
+        # subtracting, for efficiency, so we don't have to later add 1.
+        my $length = $j - $i;
+
+        $j--;        # $j now points to the highest affected range.
+        trace "Final affected range is $j: $r->[$j]" if main::DEBUG && $to_trace;
+
+        # If inserting a multiple record, this is where it goes, after all the
+        # existing ones for this range.  This implies an insertion, and no
+        # change to any existing ranges.  Note that $j can be -1 if this new
+        # range doesn't actually duplicate any existing, and comes at the
+        # beginning of the list, in which case we can handle it like any other
+        # insertion, and is easier to do so.
+        if ($replace == $MULTIPLE && $j >= 0) {
+
+            # This restriction could be remedied with a little extra work, but
+            # it won't hopefully ever be necessary
+            if ($r->[$j]->start != $r->[$j]->end) {
+                Carp::my_carp_bug("$owner_name_of{$addr}Can't cope with adding a multiple when the other range ($r->[$j]) contains more than one code point.  No action taken.");
+                return;
+            }
+
+            # Don't add an exact duplicate, as it isn't really a multiple
+            return if $value eq $r->[$j]->value && $type eq $r->[$j]->type;
+
+            trace "Adding multiple record at $j+1 with $start..$end, $value" if main::DEBUG && $to_trace;
+            my @return = splice @$r,
+                                $j+1,
+                                0,
+                                Range->new($start,
+                                           $end,
+                                           Value => $value,
+                                           Type => $type);
+            if (main::DEBUG && $to_trace) {
+                trace "After splice:";
+                trace 'j-2=[', $j-2, ']', $r->[$j-2] if $j >= 2;
+                trace 'j-1=[', $j-1, ']', $r->[$j-1] if $j >= 1;
+                trace "j  =[", $j, "]", $r->[$j] if $j >= 0;
+                trace 'j+1=[', $j+1, ']', $r->[$j+1] if $j < @$r - 1;
+                trace 'j+2=[', $j+2, ']', $r->[$j+2] if $j < @$r - 2;
+                trace 'j+3=[', $j+3, ']', $r->[$j+3] if $j < @$r - 3;
+            }
+            return @return;
+        }
+
+        # Here, have taken care of $NO and $MULTIPLE replaces.
+        # $j points to the highest affected range.  But it can be < $i or even
+        # -1.  These happen only if the insertion is entirely in the gap
+        # between r[$i-1] and r[$i].  Here's why: j < i means that the j loop
+        # above exited first time through with $end < $r->[$i]->start.  (And
+        # then we subtracted one from j)  This implies also that $start <
+        # $r->[$i]->start, but we know from above that $r->[$i-1]->end <
+        # $start, so the entire input range is in the gap.
+        if ($j < $i) {
+
+            # Here the entire input range is in the gap before $i.
+
+            if (main::DEBUG && $to_trace) {
+                if ($i) {
+                    trace "Entire range is between $r->[$i-1] and $r->[$i]";
+                }
+                else {
+                    trace "Entire range is before $r->[$i]";
+                }
+            }
+            return if $operation ne '+'; # Deletion of a non-existent range is
+                                         # a no-op
+        }
+        else {
+
+            # Here the entire input range is not in the gap before $i.  There
+            # is an affected one, and $j points to the highest such one.
+
+            # At this point, here is the situation:
+            # This is not an insertion of a multiple, nor of tentative ($NO)
+            # data.
+            #   $i  points to the first element in the current range list that
+            #            may be affected by this operation.  In fact, we know
+            #            that the range at $i is affected because we are in
+            #            the else branch of this 'if'
+            #   $j  points to the highest affected range.
+            # In other words,
+            #   r[$i-1]->end < $start <= r[$i]->end
+            # And:
+            #   r[$i-1]->end < $start <= $end <= r[$j]->end
+            #
+            # Also:
+            #   $cdm is a boolean which is set true if and only if this is a
+            #        change or deletion (multiple was handled above).  In
+            #        other words, it could be renamed to be just $cd.
+
+            # We now have enough information to decide if this call is a no-op
+            # or not.  It is a no-op if it is a deletion of a non-existent
+            # range, or an insertion of already existing data.
+
+            if (main::DEBUG && $to_trace && ! $cdm
+                                         && $i == $j
+                                         && $start >= $r->[$i]->start)
+            {
+                    trace "no-op";
+            }
+            return if ! $cdm      # change or delete => not no-op
+                      && $i == $j # more than one affected range => not no-op
+
+                      # Here, r[$i-1]->end < $start <= $end <= r[$i]->end
+                      # Further, $start and/or $end is >= r[$i]->start
+                      # The test below hence guarantees that
+                      #     r[$i]->start < $start <= $end <= r[$i]->end
+                      # This means the input range is contained entirely in
+                      # the one at $i, so is a no-op
+                      && $start >= $r->[$i]->start;
+        }
+
+        # Here, we know that some action will have to be taken.  We have
+        # calculated the offset and length (though adjustments may be needed)
+        # for the splice.  Now start constructing the replacement list.
+        my @replacement;
+        my $splice_start = $i;
+
+        my $extends_below;
+        my $extends_above;
+
+        # See if should extend any adjacent ranges.
+        if ($operation eq '-') { # Don't extend deletions
+            $extends_below = $extends_above = 0;
+        }
+        else {  # Here, should extend any adjacent ranges.  See if there are
+                # any.
+            $extends_below = ($i > 0
+                            # can't extend unless adjacent
+                            && $r->[$i-1]->end == $start -1
+                            # can't extend unless are same standard value
+                            && $r->[$i-1]->standard_form eq $standard_form
+                            # can't extend unless share type
+                            && $r->[$i-1]->type == $type);
+            $extends_above = ($j+1 < $range_list_size
+                            && $r->[$j+1]->start == $end +1
+                            && $r->[$j+1]->standard_form eq $standard_form
+                            && $r->[$j-1]->type == $type);
+        }
+        if ($extends_below && $extends_above) { # Adds to both
+            $splice_start--;     # start replace at element below
+            $length += 2;        # will replace on both sides
+            trace "Extends both below and above ranges" if main::DEBUG && $to_trace;
+
+            # The result will fill in any gap, replacing both sides, and
+            # create one large range.
+            @replacement = Range->new($r->[$i-1]->start,
+                                      $r->[$j+1]->end,
+                                      Value => $value,
+                                      Type => $type);
+        }
+        else {
+
+            # Here we know that the result won't just be the conglomeration of
+            # a new range with both its adjacent neighbors.  But it could
+            # extend one of them.
+
+            if ($extends_below) {
+
+                # Here the new element adds to the one below, but not to the
+                # one above.  If inserting, and only to that one range,  can
+                # just change its ending to include the new one.
+                if ($length == 0 && ! $cdm) {
+                    $r->[$i-1]->set_end($end);
+                    trace "inserted range extends range to below so it is now $r->[$i-1]" if main::DEBUG && $to_trace;
+                    return;
+                }
+                else {
+                    trace "Changing inserted range to start at ", sprintf("%04X",  $r->[$i-1]->start), " instead of ", sprintf("%04X", $start) if main::DEBUG && $to_trace;
+                    $splice_start--;        # start replace at element below
+                    $length++;              # will replace the element below
+                    $start = $r->[$i-1]->start;
+                }
+            }
+            elsif ($extends_above) {
+
+                # Here the new element adds to the one above, but not below.
+                # Mirror the code above
+                if ($length == 0 && ! $cdm) {
+                    $r->[$j+1]->set_start($start);
+                    trace "inserted range extends range to above so it is now $r->[$j+1]" if main::DEBUG && $to_trace;
+                    return;
+                }
+                else {
+                    trace "Changing inserted range to end at ", sprintf("%04X",  $r->[$j+1]->end), " instead of ", sprintf("%04X", $end) if main::DEBUG && $to_trace;
+                    $length++;        # will replace the element above
+                    $end = $r->[$j+1]->end;
+                }
+            }
+
+            trace "Range at $i is $r->[$i]" if main::DEBUG && $to_trace;
+
+            # Finally, here we know there will have to be a splice.
+            # If the change or delete affects only the highest portion of the
+            # first affected range, the range will have to be split.  The
+            # splice will remove the whole range, but will replace it by a new
+            # range containing just the unaffected part.  So, in this case,
+            # add to the replacement list just this unaffected portion.
+            if (! $extends_below
+                && $start > $r->[$i]->start && $start <= $r->[$i]->end)
+            {
+                push @replacement,
+                    Range->new($r->[$i]->start,
+                               $start - 1,
+                               Value => $r->[$i]->value,
+                               Type => $r->[$i]->type);
+            }
+
+            # In the case of an insert or change, but not a delete, we have to
+            # put in the new stuff;  this comes next.
+            if ($operation eq '+') {
+                push @replacement, Range->new($start,
+                                              $end,
+                                              Value => $value,
+                                              Type => $type);
+            }
+
+            trace "Range at $j is $r->[$j]" if main::DEBUG && $to_trace && $j != $i;
+            #trace "$end >=", $r->[$j]->start, " && $end <", $r->[$j]->end if main::DEBUG && $to_trace;
+
+            # And finally, if we're changing or deleting only a portion of the
+            # highest affected range, it must be split, as the lowest one was.
+            if (! $extends_above
+                && $j >= 0  # Remember that j can be -1 if before first
+                            # current element
+                && $end >= $r->[$j]->start
+                && $end < $r->[$j]->end)
+            {
+                push @replacement,
+                    Range->new($end + 1,
+                               $r->[$j]->end,
+                               Value => $r->[$j]->value,
+                               Type => $r->[$j]->type);
+            }
+        }
+
+        # And do the splice, as calculated above
+        if (main::DEBUG && $to_trace) {
+            trace "replacing $length element(s) at $i with ";
+            foreach my $replacement (@replacement) {
+                trace "    $replacement";
+            }
+            trace "Before splice:";
+            trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
+            trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
+            trace "i  =[", $i, "]", $r->[$i];
+            trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
+            trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
+        }
+
+        my @return = splice @$r, $splice_start, $length, @replacement;
+
+        if (main::DEBUG && $to_trace) {
+            trace "After splice:";
+            trace 'i-2=[', $i-2, ']', $r->[$i-2] if $i >= 2;
+            trace 'i-1=[', $i-1, ']', $r->[$i-1] if $i >= 1;
+            trace "i  =[", $i, "]", $r->[$i];
+            trace 'i+1=[', $i+1, ']', $r->[$i+1] if $i < @$r - 1;
+            trace 'i+2=[', $i+2, ']', $r->[$i+2] if $i < @$r - 2;
+            trace "removed @return";
+        }
+
+        # An actual deletion could have changed the maximum in the list.
+        # There was no deletion if the splice didn't return something, but
+        # otherwise recalculate it.  This is done too rarely to worry about
+        # performance.
+        if ($operation eq '-' && @return) {
+            $max{$addr} = $r->[-1]->end;
+        }
+        return @return;
+    }
+
+    sub reset_each_range {  # reset the iterator for each_range();
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        local $addr = main::objaddr $self if ! defined $addr;
+
+        undef $each_range_iterator{$addr};
+        return;
+    }
+
+    sub each_range {
+        # Iterate over each range in a range list.  Results are undefined if
+        # the range list is changed during the iteration.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        local $addr = main::objaddr($self) if ! defined $addr;
+
+        return if $self->is_empty;
 
+        $each_range_iterator{$addr} = -1
+                                if ! defined $each_range_iterator{$addr};
+        $each_range_iterator{$addr}++;
+        return $ranges{$addr}->[$each_range_iterator{$addr}]
+                        if $each_range_iterator{$addr} < @{$ranges{$addr}};
+        undef $each_range_iterator{$addr};
+        return;
+    }
+
+    sub count {        # Returns count of code points in range list
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        local $addr = main::objaddr($self) if ! defined $addr;
+
+        my $count = 0;
+        foreach my $range (@{$ranges{$addr}}) {
+            $count += $range->end - $range->start + 1;
+        }
+        return $count;
+    }
+
+    sub delete_range {    # Delete a range
+        my $self = shift;
+        my $start = shift;
+        my $end = shift;
+
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        return $self->_add_delete('-', $start, $end, "");
+    }
+
+    sub is_empty { # Returns boolean as to if a range list is empty
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        local $addr = main::objaddr($self) if ! defined $addr;
+        return scalar @{$ranges{$addr}} == 0;
+    }
+
+    sub hash {
+        # Quickly returns a scalar suitable for separating tables into
+        # buckets, i.e. it is a hash function of the contents of a table, so
+        # there are relatively few conflicts.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        local $addr = main::objaddr($self) if ! defined $addr;
+
+        # These are quickly computable.  Return looks like 'min..max;count'
+        return $self->min . "..$max{$addr};" . scalar @{$ranges{$addr}};
+    }
+} # End closure for _Range_List_Base
+
+package Range_List;
+use base '_Range_List_Base';
+
+# A Range_List is a range list for match tables; i.e. the range values are
+# not significant.  Thus a number of operations can be safely added to it,
+# such as inversion, intersection.  Note that union is also an unsafe
+# operation when range values are cared about, and that method is in the base
+# class, not here.  But things are set up so that that method is callable only
+# during initialization.  Only in this derived class, is there an operation
+# that combines two tables.  A Range_Map can thus be used to initialize a
+# Range_List, and its mappings will be in the list, but are not significant to
+# this class.
+
+sub trace { return main::trace(@_); }
+
+{ # Closure
+
+    use overload
+        fallback => 0,
+        '+' => sub { my $self = shift;
+                    my $other = shift;
+
+                    return $self->_union($other)
+                },
+        '&' => sub { my $self = shift;
+                    my $other = shift;
+
+                    return $self->_intersect($other, 0);
+                },
+        '~' => "_invert",
+        '-' => "_subtract",
+    ;
+
+    sub _invert {
+        # Returns a new Range_List that gives all code points not in $self.
+
+        my $self = shift;
+
+        my $new = Range_List->new;
+
+        # Go through each range in the table, finding the gaps between them
+        my $max = -1;   # Set so no gap before range beginning at 0
+        for my $range ($self->ranges) {
+            my $start = $range->start;
+            my $end   = $range->end;
+
+            # If there is a gap before this range, the inverse will contain
+            # that gap.
+            if ($start > $max + 1) {
+                $new->add_range($max + 1, $start - 1);
+            }
+            $max = $end;
+        }
+
+        # And finally, add the gap from the end of the table to the max
+        # possible code point
+        if ($max < $LAST_UNICODE_CODEPOINT) {
+            $new->add_range($max + 1, $LAST_UNICODE_CODEPOINT);
+        }
+        return $new;
+    }
+
+    sub _subtract {
+        # Returns a new Range_List with the argument deleted from it.  The
+        # argument can be a single code point, a range, or something that has
+        # a range, with the _range_list() method on it returning them
+
+        my $self = shift;
+        my $other = shift;
+        my $reversed = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        if ($reversed) {
+            Carp::my_carp_bug("Can't cope with a "
+             .  __PACKAGE__
+             . " being the second parameter in a '-'.  Subtraction ignored.");
+            return $self;
+        }
+
+        my $new = Range_List->new(Initialize => $self);
+
+        if (! ref $other) { # Single code point
+            $new->delete_range($other, $other);
+        }
+        elsif ($other->isa('Range')) {
+            $new->delete_range($other->start, $other->end);
+        }
+        elsif ($other->can('_range_list')) {
+            foreach my $range ($other->_range_list->ranges) {
+                $new->delete_range($range->start, $range->end);
+            }
+        }
+        else {
+            Carp::my_carp_bug("Can't cope with a "
+                        . ref($other)
+                        . " argument to '-'.  Subtraction ignored."
+                        );
+            return $self;
+        }
+
+        return $new;
+    }
+
+    sub _intersect {
+        # Returns either a boolean giving whether the two inputs' range lists
+        # intersect (overlap), or a new Range_List containing the intersection
+        # of the two lists.  The optional final parameter being true indicates
+        # to do the check instead of the intersection.
+
+        my $a_object = shift;
+        my $b_object = shift;
+        my $check_if_overlapping = shift;
+        $check_if_overlapping = 0 unless defined $check_if_overlapping;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        if (! defined $b_object) {
+            my $message = "";
+            $message .= $a_object->_owner_name_of if defined $a_object;
+            Carp::my_carp_bug($message .= "Called with undefined value.  Intersection not done.");
+            return;
+        }
+
+        # a & b = !(!a | !b), or in our terminology = ~ ( ~a + -b )
+        # Thus the intersection could be much more simply be written:
+        #   return ~(~$a_object + ~$b_object);
+        # But, this is slower, and when taking the inverse of a large
+        # range_size_1 table, back when such tables were always stored that
+        # way, it became prohibitively slow, hence the code was changed to the
+        # below
+
+        if ($b_object->isa('Range')) {
+            $b_object = Range_List->new(Initialize => $b_object,
+                                        Owner => $a_object->_owner_name_of);
+        }
+        $b_object = $b_object->_range_list if $b_object->can('_range_list');
+
+        my @a_ranges = $a_object->ranges;
+        my @b_ranges = $b_object->ranges;
+
+        #local $to_trace = 1 if main::DEBUG;
+        trace "intersecting $a_object with ", scalar @a_ranges, "ranges and $b_object with", scalar @b_ranges, " ranges" if main::DEBUG && $to_trace;
+
+        # Start with the first range in each list
+        my $a_i = 0;
+        my $range_a = $a_ranges[$a_i];
+        my $b_i = 0;
+        my $range_b = $b_ranges[$b_i];
+
+        my $new = __PACKAGE__->new(Owner => $a_object->_owner_name_of)
+                                                if ! $check_if_overlapping;
+
+        # If either list is empty, there is no intersection and no overlap
+        if (! defined $range_a || ! defined $range_b) {
+            return $check_if_overlapping ? 0 : $new;
+        }
+        trace "range_a[$a_i]=$range_a; range_b[$b_i]=$range_b" if main::DEBUG && $to_trace;
+
+        # Otherwise, must calculate the intersection/overlap.  Start with the
+        # very first code point in each list
+        my $a = $range_a->start;
+        my $b = $range_b->start;
+
+        # Loop through all the ranges of each list; in each iteration, $a and
+        # $b are the current code points in their respective lists
+        while (1) {
+
+            # If $a and $b are the same code point, ...
+            if ($a == $b) {
+
+                # it means the lists overlap.  If just checking for overlap
+                # know the answer now,
+                return 1 if $check_if_overlapping;
+
+                # The intersection includes this code point plus anything else
+                # common to both current ranges.
+                my $start = $a;
+                my $end = main::min($range_a->end, $range_b->end);
+                if (! $check_if_overlapping) {
+                    trace "adding intersection range ", sprintf("%04X", $start) . ".." . sprintf("%04X", $end) if main::DEBUG && $to_trace;
+                    $new->add_range($start, $end);
+                }
+
+                # Skip ahead to the end of the current intersect
+                $a = $b = $end;
+
+                # If the current intersect ends at the end of either range (as
+                # it must for at least one of them), the next possible one
+                # will be the beginning code point in it's list's next range.
+                if ($a == $range_a->end) {
+                    $range_a = $a_ranges[++$a_i];
+                    last unless defined $range_a;
+                    $a = $range_a->start;
+                }
+                if ($b == $range_b->end) {
+                    $range_b = $b_ranges[++$b_i];
+                    last unless defined $range_b;
+                    $b = $range_b->start;
+                }
+
+                trace "range_a[$a_i]=$range_a; range_b[$b_i]=$range_b" if main::DEBUG && $to_trace;
+            }
+            elsif ($a < $b) {
+
+                # Not equal, but if the range containing $a encompasses $b,
+                # change $a to be the middle of the range where it does equal
+                # $b, so the next iteration will get the intersection
+                if ($range_a->end >= $b) {
+                    $a = $b;
+                }
+                else {
+
+                    # Here, the current range containing $a is entirely below
+                    # $b.  Go try to find a range that could contain $b.
+                    $a_i = $a_object->_search_ranges($b);
+
+                    # If no range found, quit.
+                    last unless defined $a_i;
+
+                    # The search returns $a_i, such that
+                    #   range_a[$a_i-1]->end < $b <= range_a[$a_i]->end
+                    # Set $a to the beginning of this new range, and repeat.
+                    $range_a = $a_ranges[$a_i];
+                    $a = $range_a->start;
+                }
+            }
+            else { # Here, $b < $a.
+
+                # Mirror image code to the leg just above
+                if ($range_b->end >= $a) {
+                    $b = $a;
+                }
+                else {
+                    $b_i = $b_object->_search_ranges($a);
+                    last unless defined $b_i;
+                    $range_b = $b_ranges[$b_i];
+                    $b = $range_b->start;
+                }
+            }
+        } # End of looping through ranges.
+
+        # Intersection fully computed, or now know that there is no overlap
+        return $check_if_overlapping ? 0 : $new;
+    }
+
+    sub overlaps {
+        # Returns boolean giving whether the two arguments overlap somewhere
+
+        my $self = shift;
+        my $other = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        return $self->_intersect($other, 1);
+    }
+
+    sub add_range {
+        # Add a range to the list.
 
-##
-## Called like:
-##     New_Alias(Is => 'All', SameAs => 'Any', Fuzzy => 1);
-##
-## The args must be in that order, although the Fuzzy pair may be omitted.
-##
-## This creates 'IsAll' as an alias for 'IsAny'
-##
-sub New_Alias($$$@)
-{
-    my $Type   = shift; ## "Is" or "In"
-    my $Alias  = shift;
-    my $SameAs = shift; # expecting "SameAs" -- just ignored
-    my $Name   = shift;
+        my $self = shift;
+        my $start = shift;
+        my $end = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
 
-    ## remaining args are optional key/val
-    my %Args = @_;
+        return $self->_add_delete('+', $start, $end, "");
+    }
+
+    my $non_ASCII = (ord('A') == 65);   # Assumes test on same platform
+
+    sub is_code_point_usable {
+        # This used only for making the test script.  See if the input
+        # proposed trial code point is one that Perl will handle.  If second
+        # parameter is 0, it won't select some code points for various
+        # reasons, noted below.
+
+        my $code = shift;
+        my $try_hard = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        return 0 if $code < 0;                # Never use a negative
+
+        # For non-ASCII, we shun the characters that don't have Perl encoding-
+        # independent symbols for them.  'A' is such a symbol, so is "\n".
+        # Note, this program hopefully will work on 5.8 Perls, and \v is not
+        # such a symbol in them.
+        return $try_hard if $non_ASCII
+                            && $code <= 0xFF
+                            && ($code >= 0x7F
+                                || ($code >= 0x0E && $code <= 0x1F)
+                                || ($code >= 0x01 && $code <= 0x06)
+                                || $code == 0x0B);  # \v introduced after 5.8
+
+        # shun null.  I'm (khw) not sure why this was done, but NULL would be
+        # the character very frequently used.
+        return $try_hard if $code == 0x0000;
+
+        return 0 if $try_hard;  # XXX Temporary until fix utf8.c
+
+        # shun non-character code points.
+        return $try_hard if $code >= 0xFDD0 && $code <= 0xFDEF;
+        return $try_hard if ($code & 0xFFFE) == 0xFFFE; # includes FFFF
+
+        return $try_hard if $code > $LAST_UNICODE_CODEPOINT;   # keep in range
+        return $try_hard if $code >= 0xD800 && $code <= 0xDFFF; # no surrogate
+
+        return 1;
+    }
+
+    sub get_valid_code_point {
+        # Return a code point that's part of the range list.  Returns nothing
+        # if the table is empty or we can't find a suitable code point.  This
+        # used only for making the test script.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr($self);
+
+        # On first pass, don't choose less desirable code points; if no good
+        # one is found, repeat, allowing a less desirable one to be selected.
+        for my $try_hard (0, 1) {
+
+            # Look through all the ranges for a usable code point.
+            for my $set ($self->ranges) {
+
+                # Try the edge cases first, starting with the end point of the
+                # range.
+                my $end = $set->end;
+                return $end if is_code_point_usable($end, $try_hard);
+
+                # End point didn't, work.  Start at the beginning and try
+                # every one until find one that does work.
+                for my $trial ($set->start .. $end - 1) {
+                    return $trial if is_code_point_usable($trial, $try_hard);
+                }
+            }
+        }
+        return ();  # If none found, give up.
+    }
+
+    sub get_invalid_code_point {
+        # Return a code point that's not part of the table.  Returns nothing
+        # if the table covers all code points or a suitable code point can't
+        # be found.  This used only for making the test script.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        # Just find a valid code point of the inverse, if any.
+        return Range_List->new(Initialize => ~ $self)->get_valid_code_point;
+    }
+} # end closure for Range_List
+
+package Range_Map;
+use base '_Range_List_Base';
+
+# A Range_Map is a range list in which the range values (called maps) are
+# significant, and hence shouldn't be manipulated by our other code, which
+# could be ambiguous or lose things.  For example, in taking the union of two
+# lists, which share code points, but which have differing values, which one
+# has precedence in the union?
+# It turns out that these operations aren't really necessary for map tables,
+# and so this class was created to make sure they aren't accidentally
+# applied to them.
+
+{ # Closure
+
+    sub add_map {
+        # Add a range containing a mapping value to the list
+
+        my $self = shift;
+        # Rest of parameters passed on
+
+        return $self->_add_delete('+', @_);
+    }
+
+    sub add_duplicate {
+        # Adds entry to a range list which can duplicate an existing entry
 
-    my $Fuzzy = delete $Args{Fuzzy};
+        my $self = shift;
+        my $code_point = shift;
+        my $value = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
 
-    ## sanity check a few args
-    if (%Args or ($Type ne 'Is' and $Type ne 'In') or $SameAs ne 'SameAs') {
-        confess "$0: bad args to New_Alias"
+        return $self->add_map($code_point, $code_point,
+                                $value, Replace => $MULTIPLE);
     }
+} # End of closure for package Range_Map
+
+package _Base_Table;
+
+# A table is the basic data structure that gets written out into a file for
+# use by the Perl core.  This is the abstract base class implementing the
+# common elements from the derived ones.  A list of the methods to be
+# furnished by an implementing class is just after the constructor.
+
+sub standardize { return main::standardize($_[0]); }
+sub trace { return main::trace(@_); }
+
+{ # Closure
 
-    $Alias = CanonicalName($Alias) if $Fuzzy;
+    main::setup_package();
 
-    if (not $TableInfo{$Type}->{$Name})
+    my %range_list;
+    # Object containing the ranges of the table.
+    main::set_access('range_list', \%range_list, 'p_r', 'p_s');
+
+    my %full_name;
+    # The full table name.
+    main::set_access('full_name', \%full_name, 'r');
+
+    my %name;
+    # The table name, almost always shorter
+    main::set_access('name', \%name, 'r');
+
+    my %short_name;
+    # The shortest of all the aliases for this table, with underscores removed
+    main::set_access('short_name', \%short_name);
+
+    my %nominal_short_name_length;
+    # The length of short_name before removing underscores
+    main::set_access('nominal_short_name_length',
+                    \%nominal_short_name_length);
+
+    my %property;
+    # Parent property this table is attached to.
+    main::set_access('property', \%property, 'r');
+
+    my %aliases;
+    # Ordered list of aliases of the table's name.  The first ones in the list
+    # are output first in comments
+    main::set_access('aliases', \%aliases, 'readable_array');
+
+    my %comment;
+    # A comment associated with the table for human readers of the files
+    main::set_access('comment', \%comment, 's');
+
+    my %description;
+    # A comment giving a short description of the table's meaning for human
+    # readers of the files.
+    main::set_access('description', \%description, 'readable_array');
+
+    my %note;
+    # A comment giving a short note about the table for human readers of the
+    # files.
+    main::set_access('note', \%note, 'readable_array');
+
+    my %internal_only;
+    # Boolean; if set means any file that contains this table is marked as for
+    # internal-only use.
+    main::set_access('internal_only', \%internal_only);
+
+    my %find_table_from_alias;
+    # The parent property passes this pointer to a hash which this class adds
+    # all its aliases to, so that the parent can quickly take an alias and
+    # find this table.
+    main::set_access('find_table_from_alias', \%find_table_from_alias, 'p_r');
+
+    my %locked;
+    # After this table is made equivalent to another one; we shouldn't go
+    # changing the contents because that could mean it's no longer equivalent
+    main::set_access('locked', \%locked, 'r');
+
+    my %file_path;
+    # This gives the final path to the file containing the table.  Each
+    # directory in the path is an element in the array
+    main::set_access('file_path', \%file_path, 'readable_array');
+
+    my %status;
+    # What is the table's status, normal, $OBSOLETE, etc.  Enum
+    main::set_access('status', \%status, 'r');
+
+    my %status_info;
+    # A comment about its being obsolete, or whatever non normal status it has
+    main::set_access('status_info', \%status_info, 'r');
+
+    my %range_size_1;
+    # Is the table to be output with each range only a single code point?
+    # This is done to avoid breaking existing code that may have come to rely
+    # on this behavior in previous versions of this program.)
+    main::set_access('range_size_1', \%range_size_1, 'r', 's');
+
+    my %perl_extension;
+    # A boolean set iff this table is a Perl extension to the Unicode
+    # standard.
+    main::set_access('perl_extension', \%perl_extension, 'r');
+
+    sub new {
+        # All arguments are key => value pairs, which you can see below, most
+        # of which match fields documented above.  Otherwise: Pod_Entry,
+        # Externally_Ok, and Fuzzy apply to the names of the table, and are
+        # documented in the Alias package
+
+        return Carp::carp_too_few_args(\@_, 2) if main::DEBUG && @_ < 2;
+
+        my $class = shift;
+
+        my $self = bless \do { my $anonymous_scalar }, $class;
+        my $addr = main::objaddr($self);
+
+        my %args = @_;
+
+        $name{$addr} = delete $args{'Name'};
+        $find_table_from_alias{$addr} = delete $args{'_Alias_Hash'};
+        $full_name{$addr} = delete $args{'Full_Name'};
+        $internal_only{$addr} = delete $args{'Internal_Only_Warning'} || 0;
+        $perl_extension{$addr} = delete $args{'Perl_Extension'} || 0;
+        $property{$addr} = delete $args{'_Property'};
+        $range_list{$addr} = delete $args{'_Range_List'};
+        $status{$addr} = delete $args{'Status'} || $NORMAL;
+        $status_info{$addr} = delete $args{'_Status_Info'} || "";
+        $range_size_1{$addr} = delete $args{'Range_Size_1'} || 0;
+
+        my $description = delete $args{'Description'};
+        my $externally_ok = delete $args{'Externally_Ok'};
+        my $loose_match = delete $args{'Fuzzy'};
+        my $note = delete $args{'Note'};
+        my $make_pod_entry = delete $args{'Pod_Entry'};
+
+        # Shouldn't have any left over
+        Carp::carp_extra_args(\%args) if main::DEBUG && %args;
+
+        # Can't use || above because conceivably the name could be 0, and
+        # can't use // operator in case this program gets used in Perl 5.8
+        $full_name{$addr} = $name{$addr} if ! defined $full_name{$addr};
+
+        $aliases{$addr} = [ ];
+        $comment{$addr} = [ ];
+        $description{$addr} = [ ];
+        $note{$addr} = [ ];
+        $file_path{$addr} = [ ];
+        $locked{$addr} = "";
+
+        push @{$description{$addr}}, $description if $description;
+        push @{$note{$addr}}, $note if $note;
+
+        # If hasn't set its status already, see if it is on one of the lists
+        # of properties or tables that have particular statuses; if not, is
+        # normal.  The lists are prioritized so the most serious ones are
+        # checked first
+        my $complete_name = $self->complete_name;
+        if (! $status{$addr}) {
+            if (exists $why_suppressed{$complete_name}) {
+                $status{$addr} = $SUPPRESSED;
+            }
+            elsif (exists $why_deprecated{$complete_name}) {
+                $status{$addr} = $DEPRECATED;
+            }
+            elsif (exists $why_stabilized{$complete_name}) {
+                $status{$addr} = $STABILIZED;
+            }
+            elsif (exists $why_obsolete{$complete_name}) {
+                $status{$addr} = $OBSOLETE;
+            }
+
+            # Existence above doesn't necessarily mean there is a message
+            # associated with it.  Use the most serious message.
+            if ($status{$addr}) {
+                if ($why_suppressed{$complete_name}) {
+                    $status_info{$addr}
+                                = $why_suppressed{$complete_name};
+                }
+                elsif ($why_deprecated{$complete_name}) {
+                    $status_info{$addr}
+                                = $why_deprecated{$complete_name};
+                }
+                elsif ($why_stabilized{$complete_name}) {
+                    $status_info{$addr}
+                                = $why_stabilized{$complete_name};
+                }
+                elsif ($why_obsolete{$complete_name}) {
+                    $status_info{$addr}
+                                = $why_obsolete{$complete_name};
+                }
+            }
+        }
+
+        # By convention what typically gets printed only or first is what's
+        # first in the list, so put the full name there for good output
+        # clarity.  Other routines rely on the full name being first on the
+        # list
+        $self->add_alias($full_name{$addr},
+                            Externally_Ok => $externally_ok,
+                            Fuzzy => $loose_match,
+                            Pod_Entry => $make_pod_entry,
+                            Status => $status{$addr},
+                            );
+
+        # Then comes the other name, if meaningfully different.
+        if (standardize($full_name{$addr}) ne standardize($name{$addr})) {
+            $self->add_alias($name{$addr},
+                            Externally_Ok => $externally_ok,
+                            Fuzzy => $loose_match,
+                            Pod_Entry => $make_pod_entry,
+                            Status => $status{$addr},
+                            );
+        }
+
+        return $self;
+    }
+
+    # Here are the methods that are required to be defined by any derived
+    # class
+    for my $sub qw(
+                    append_to_body
+                    complete_name
+                    pre_body
+                )
+                # append_to_body and pre_body are called in the write() method
+                # to add stuff after the main body of the table, but before
+                # its close; and to prepend stuff before the beginning of the
+                # table.
+                # complete_name returns the complete name of the property and
+                # table, like Script=Latin
     {
-        my $CName = CanonicalName($Name);
-        if ($TableInfo{$Type}->{$CName}) {
-            confess "$0: Use canonical form '$CName' instead of '$Name' for alias.";
-        } else {
-            confess "$0: don't have original $Type => $Name to make alias\n";
-        }
-    }
-    if ($TableInfo{$Alias}) {
-        confess "$0: already have original $Type => $Alias; can't make alias";
-    }
-    $AliasInfo{$Type}->{$Name} = $Alias;
-    if ($Fuzzy) {
-        $FuzzyNames{$Type}->{$Alias} = $Name;
-    }
-
-}
-
-
-## All assigned code points
-my $Assigned = Table->New(Is    => 'Assigned',
-                          Desc  => "All assigned code points",
-                          Fuzzy => 0);
-
-my $Name     = Table->New(); ## all characters, individually by name
-my $General  = Table->New(); ## all characters, grouped by category
-my %General;
-my %Cat;
-
-## Simple Data::Dumper like. Good enough for our needs. We can't use the real
-## thing as we have to run under miniperl
-sub simple_dumper {
-    my @lines;
-    my $item;
-    foreach $item (@_) {
-	if (ref $item) {
-	    if (ref $item eq 'ARRAY') {
-		push @lines, "[\n", simple_dumper (@$item), "],\n";
-	    } elsif (ref $item eq 'HASH') {
-		push @lines, "{\n", simple_dumper (%$item), "},\n";
-	    } else {
-		die "Can't cope with $item";
-	    }
-	} else {
-	    if (defined $item) {
-		my $copy = $item;
-		$copy =~ s/([\'\\])/\\$1/gs;
-		push @lines, "'$copy',\n";
-	    } else {
-		push @lines, "undef,\n";
-	    }
-	}
-    }
-    @lines;
-}
-
-##
-## Process UnicodeData.txt (Categories, etc.)
-##
-# These are the character mappings as defined in the POSIX standard
-# and in the case of PerlSpace and PerlWord as is defined in the test macros
-# for binary strings. IOW, PerlWord is [A-Za-z_] and PerlSpace is [\f\r\n\t ]
-# This differs from Word and the existing SpacePerl (note the prefix/suffix difference)
-# which is basically the Unicode WhiteSpace without the vertical tab included
-#
-my %TRUE_POSIX_PERL_CC= (
-    PosixAlnum => { map { $_ => 1 } ( 0x0030..0x0039, 0x0041..0x005a, 0x0061..0x007a )},
-    PosixAlpha => { map { $_ => 1 } ( 0x0041..0x005a, 0x0061..0x007a )},
-    # Not Needed: Ascii => { map { $_ => 1 } ( 0x0000..0x007f )},
-    PosixBlank => { map { $_ => 1 } ( 0x0009, 0x0020 )},
-    PosixCntrl => { map { $_ => 1 } ( 0x0000..0x001f, 0x007f )},
-    PosixGraph => { map { $_ => 1 } ( 0x0021..0x007e )},
-    PosixLower => { map { $_ => 1 } ( 0x0061..0x007a )},
-    PosixPrint => { map { $_ => 1 } ( 0x0020..0x007e )},
-    PosixPunct => { map { $_ => 1 } ( 0x0021..0x002f, 0x003a..0x0040, 0x005b..0x0060, 0x007b..0x007e )},
-    PosixSpace => { map { $_ => 1 } ( 0x0009..0x000d, 0x0020 )},
-    PosixUpper => { map { $_ => 1 } ( 0x0041..0x005a )},
-    # Not needed:  PosixXdigit => { map { $_ => 1 } ( 0x0030..0x0039, 0x0041..0x0046, 0x0061..0x0066 )},
-    PosixDigit => { map { $_ => 1 } ( 0x0030..0x0039 )},
-    
-    PerlSpace  => { map { $_ => 1 } ( 0x0009..0x000a, 0x000c..0x000d, 0x0020 )},
-    PerlWord   => { map { $_ => 1 } ( 0x0030..0x0039, 0x0041..0x005a, 0x005f, 0x0061..0x007a )},
-);
+        no strict "refs";
+        *$sub = sub {
+            Carp::my_carp_bug( __LINE__
+                              . ": Must create method '$sub()' for "
+                              . ref shift);
+            return;
+        }
+    }
+
+    use overload
+        fallback => 0,
+        "." => \&main::_operator_dot,
+        '!=' => \&main::_operator_not_equal,
+        '==' => \&main::_operator_equal,
+    ;
+
+    sub ranges {
+        # Returns the array of ranges associated with this table.
+
+        return $range_list{main::objaddr shift}->ranges;
+    }
+
+    sub add_alias {
+        # Add a synonym for this table.
+
+        return Carp::carp_too_few_args(\@_, 3) if main::DEBUG && @_ < 3;
+
+        my $self = shift;
+        my $name = shift;       # The name to add.
+        my $pointer = shift;    # What the alias hash should point to.  For
+                                # map tables, this is the parent property;
+                                # for match tables, it is the table itself.
+
+        my %args = @_;
+        my $loose_match = delete $args{'Fuzzy'};
+
+        my $make_pod_entry = delete $args{'Pod_Entry'};
+        $make_pod_entry = $YES unless defined $make_pod_entry;
+
+        my $externally_ok = delete $args{'Externally_Ok'};
+        $externally_ok = 1 unless defined $externally_ok;
+
+        my $status = delete $args{'Status'};
+        $status = $NORMAL unless defined $status;
+
+        Carp::carp_extra_args(\%args) if main::DEBUG && %args;
+
+        # Capitalize the first letter of the alias unless it is one of the CJK
+        # ones which specifically begins with a lower 'k'.  Do this because
+        # Unicode has varied whether they capitalize first letters or not, and
+        # have later changed their minds and capitalized them, but not the
+        # other way around.  So do it always and avoid changes from release to
+        # release
+        $name = ucfirst($name) unless $name =~ /^k[A-Z]/;
+
+        my $addr = main::objaddr $self;
+
+        # Figure out if should be loosely matched if not already specified.
+        if (! defined $loose_match) {
 
-sub UnicodeData_Txt()
-{
-    my $Bidi     = Table->New();
-    my $Deco     = Table->New();
-    my $Comb     = Table->New();
-    my $Number   = Table->New();
-    my $Mirrored = Table->New();#Is    => 'Mirrored',
-                              #Desc  => "Mirrored in bidirectional text",
-                              #Fuzzy => 0);
-
-    my %DC;
-    my %Bidi;
-    my %Number;
-    $DC{Can} = Table->New();
-    $DC{Com} = Table->New();
-
-    ## Initialize Broken Perl-generated categories
-    ## (Categories from UnicodeData.txt are auto-initialized in gencat)
-    $Cat{Alnum}  =
-	Table->New(Is => 'Alnum',  Desc => "[[:Alnum:]]",  Fuzzy => 0);
-    $Cat{Alpha}  =
-	Table->New(Is => 'Alpha',  Desc => "[[:Alpha:]]",  Fuzzy => 0);
-    $Cat{ASCII}  =
-	Table->New(Is => 'ASCII',  Desc => "[[:ASCII:]]",  Fuzzy => 0);
-    $Cat{Blank}  =
-	Table->New(Is => 'Blank',  Desc => "[[:Blank:]]",  Fuzzy => 0);
-    $Cat{Cntrl}  =
-	Table->New(Is => 'Cntrl',  Desc => "[[:Cntrl:]]",  Fuzzy => 0);
-    $Cat{Digit}  =
-	Table->New(Is => 'Digit',  Desc => "[[:Digit:]]",  Fuzzy => 0);
-    $Cat{Graph}  =
-	Table->New(Is => 'Graph',  Desc => "[[:Graph:]]",  Fuzzy => 0);
-    $Cat{Lower}  =
-	Table->New(Is => 'Lower',  Desc => "[[:Lower:]]",  Fuzzy => 0);
-    $Cat{Print}  =
-	Table->New(Is => 'Print',  Desc => "[[:Print:]]",  Fuzzy => 0);
-    $Cat{Punct}  =
-	Table->New(Is => 'Punct',  Desc => "[[:Punct:]]",  Fuzzy => 0);
-    $Cat{Space}  =
-	Table->New(Is => 'Space',  Desc => "[[:Space:]]",  Fuzzy => 0);
-    $Cat{Title}  =
-	Table->New(Is => 'Title',  Desc => "[[:Title:]]",  Fuzzy => 0);
-    $Cat{Upper}  =
-	Table->New(Is => 'Upper',  Desc => "[[:Upper:]]",  Fuzzy => 0);
-    $Cat{XDigit} =
-	Table->New(Is => 'XDigit', Desc => "[[:XDigit:]]", Fuzzy => 0);
-    $Cat{Word}   =
-	Table->New(Is => 'Word',   Desc => "[[:Word:]]",   Fuzzy => 0);
-    $Cat{SpacePerl} =
-	Table->New(Is => 'SpacePerl', Desc => '\s', Fuzzy => 0);
-    $Cat{VertSpace} =
-	Table->New(Is => 'VertSpace', Desc => '\v', Fuzzy => 0);
-    $Cat{HorizSpace} =
-	Table->New(Is => 'HorizSpace', Desc => '\h', Fuzzy => 0);
-    my %To;
-    $To{Upper} = Table->New();
-    $To{Lower} = Table->New();
-    $To{Title} = Table->New();
-    $To{Digit} = Table->New();
-
-    foreach my $cat (keys %TRUE_POSIX_PERL_CC) {
-        $Cat{$cat} = Table->New(Is=>$cat, Fuzzy => 0);
-    }
-
-    sub gencat($$$$)
+            # Is a loose_match if isn't null, and doesn't begin with an
+            # underscore and isn't just a number
+            if ($name ne ""
+                && substr($name, 0, 1) ne '_'
+                && $name !~ qr{^[0-9_.+-/]+$})
+            {
+                $loose_match = 1;
+            }
+            else {
+                $loose_match = 0;
+            }
+        }
+
+        # If this alias has already been defined, do nothing.
+        return if defined $find_table_from_alias{$addr}->{$name};
+
+        # That includes if it is standardly equivalent to an existing alias,
+        # in which case, add this name to the list, so won't have to search
+        # for it again.
+        my $standard_name = main::standardize($name);
+        if (defined $find_table_from_alias{$addr}->{$standard_name}) {
+            $find_table_from_alias{$addr}->{$name}
+                        = $find_table_from_alias{$addr}->{$standard_name};
+            return;
+        }
+
+        # Set the index hash for this alias for future quick reference.
+        $find_table_from_alias{$addr}->{$name} = $pointer;
+        $find_table_from_alias{$addr}->{$standard_name} = $pointer;
+        local $to_trace = 0 if main::DEBUG;
+        trace "adding alias $name to $pointer" if main::DEBUG && $to_trace;
+        trace "adding alias $standard_name to $pointer" if main::DEBUG && $to_trace;
+
+
+        # Put the new alias at the end of the list of aliases unless the final
+        # element begins with an underscore (meaning it is for internal perl
+        # use) or is all numeric, in which case, put the new one before that
+        # one.  This floats any all-numeric or underscore-beginning aliases to
+        # the end.  This is done so that they are listed last in output lists,
+        # to encourage the user to use a better name (either more descriptive
+        # or not an internal-only one) instead.  This ordering is relied on
+        # implicitly elsewhere in this program, like in short_name()
+        my $list = $aliases{$addr};
+        my $insert_position = (@$list == 0
+                                || (substr($list->[-1]->name, 0, 1) ne '_'
+                                    && $list->[-1]->name =~ /\D/))
+                            ? @$list
+                            : @$list - 1;
+        splice @$list,
+                $insert_position,
+                0,
+                Alias->new($name, $loose_match, $make_pod_entry,
+                                                    $externally_ok, $status);
+
+        # This name may be shorter than any existing ones, so clear the cache
+        # of the shortest, so will have to be recalculated.
+        undef $short_name{main::objaddr $self};
+        return;
+    }
+
+    sub short_name {
+        # Returns a name suitable for use as the base part of a file name.
+        # That is, shorter wins.  It can return undef if there is no suitable
+        # name.  The name has all non-essential underscores removed.
+
+        # The optional second parameter is a reference to a scalar in which
+        # this routine will store the length the returned name had before the
+        # underscores were removed, or undef if the return is undef.
+
+        # The shortest name can change if new aliases are added.  So using
+        # this should be deferred until after all these are added.  The code
+        # that does that should clear this one's cache.
+        # Any name with alphabetics is preferred over an all numeric one, even
+        # if longer.
+
+        my $self = shift;
+        my $nominal_length_ptr = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        # For efficiency, don't recalculate, but this means that adding new
+        # aliases could change what the shortest is, so the code that does
+        # that needs to undef this.
+        if (defined $short_name{$addr}) {
+            if ($nominal_length_ptr) {
+                $$nominal_length_ptr = $nominal_short_name_length{$addr};
+            }
+            return $short_name{$addr};
+        }
+
+        # Look at each alias
+        foreach my $alias ($self->aliases()) {
+
+            # Don't use an alias that isn't ok to use for an external name.
+            next if ! $alias->externally_ok;
+
+            my $name = main::Standardize($alias->name);
+            trace $self, $name if main::DEBUG && $to_trace;
+
+            # Take the first one, or a shorter one that isn't numeric.  This
+            # relies on numeric aliases always being last in the array
+            # returned by aliases().  Any alpha one will have precedence.
+            if (! defined $short_name{$addr}
+                || ($name =~ /\D/
+                    && length($name) < length($short_name{$addr})))
+            {
+                # Remove interior underscores.
+                ($short_name{$addr} = $name) =~ s/ (?<= . ) _ (?= . ) //xg;
+
+                $nominal_short_name_length{$addr} = length $name;
+            }
+        }
+
+        # If no suitable external name return undef
+        if (! defined $short_name{$addr}) {
+            $$nominal_length_ptr = undef if $nominal_length_ptr;
+            return;
+        }
+
+        # Don't allow a null external name.
+        if ($short_name{$addr} eq "") {
+            $short_name{$addr} = '_';
+            $nominal_short_name_length{$addr} = 1;
+        }
+
+        trace $self, $short_name{$addr} if main::DEBUG && $to_trace;
+
+        if ($nominal_length_ptr) {
+            $$nominal_length_ptr = $nominal_short_name_length{$addr};
+        }
+        return $short_name{$addr};
+    }
+
+    sub external_name {
+        # Returns the external name that this table should be known by.  This
+        # is usually the short_name, but not if the short_name is undefined.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $short = $self->short_name;
+        return $short if defined $short;
+
+        return '_';
+    }
+
+    sub add_description { # Adds the parameter as a short description.
+
+        my $self = shift;
+        my $description = shift;
+        chomp $description;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        push @{$description{main::objaddr $self}}, $description;
+
+        return;
+    }
+
+    sub add_note { # Adds the parameter as a short note.
+
+        my $self = shift;
+        my $note = shift;
+        chomp $note;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        push @{$note{main::objaddr $self}}, $note;
+
+        return;
+    }
+
+    sub add_comment { # Adds the parameter as a comment.
+
+        my $self = shift;
+        my $comment = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        chomp $comment;
+        push @{$comment{main::objaddr $self}}, $comment;
+
+        return;
+    }
+
+    sub comment {
+        # Return the current comment for this table.  If called in list
+        # context, returns the array of comments.  In scalar, returns a string
+        # of each element joined together with a period ending each.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my @list = @{$comment{main::objaddr $self}};
+        return @list if wantarray;
+        my $return = "";
+        foreach my $sentence (@list) {
+            $return .= '.  ' if $return;
+            $return .= $sentence;
+            $return =~ s/\.$//;
+        }
+        $return .= '.' if $return;
+        return $return;
+    }
+
+    sub initialize {
+        # Initialize the table with the argument which is any valid
+        # initialization for range lists.
+
+        my $self = shift;
+        my $initialization = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        # Replace the current range list with a new one of the same exact
+        # type.
+        my $class = ref $range_list{main::objaddr $self};
+        $range_list{main::objaddr $self} = $class->new(Owner => $self,
+                                        Initialize => $initialization);
+        return;
+
+    }
+
+    sub header {
+        # The header that is output for the table in the file it is written
+        # in.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $return = "";
+        $return .= $DEVELOPMENT_ONLY if $compare_versions;
+        $return .= $HEADER;
+        $return .= $INTERNAL_ONLY if $internal_only{main::objaddr $self};
+        return $return;
+    }
+
+    sub write {
+        # Write a representation of the table to its file.
+
+        my $self = shift;
+        my $tab_stops = shift;       # The number of tab stops over to put any
+                                     # comment.
+        my $suppress_value = shift;  # Optional, if the value associated with
+                                     # a range equals this one, don't write
+                                     # the range
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr($self);
+
+        # Start with the header
+        my @OUT = $self->header;
+
+        # Then the comments
+        push @OUT, "\n", main::simple_fold($comment{$addr}, '# '), "\n"
+                                                        if $comment{$addr};
+
+        # Then any pre-body stuff.
+        my $pre_body = $self->pre_body;
+        push @OUT, $pre_body, "\n" if $pre_body;
+
+        # The main body looks like a 'here' document
+        push @OUT, "return <<'END';\n";
+
+        if ($range_list{$addr}->is_empty) {
+
+            # This is a kludge for empty tables to silence a warning in
+            # utf8.c, which can't really deal with empty tables, but it can
+            # deal with a table that matches nothing, as the inverse of 'Any'
+            # does.
+            push @OUT, "!utf8::IsAny\n";
+        }
+        else {
+            my $range_size_1 = $range_size_1{$addr};
+
+            # Output each range as part of the here document.
+            for my $set ($range_list{$addr}->ranges) {
+                my $start = $set->start;
+                my $end   = $set->end;
+                my $value  = $set->value;
+
+                # Don't output ranges whose value is the one to suppress
+                next if defined $suppress_value && $value eq $suppress_value;
+
+                # If has or wants a single point range output
+                if ($start == $end || $range_size_1) {
+                    for my $i ($start .. $end) {
+                        push @OUT, sprintf "%04X\t\t%s\n", $i, $value;
+                    }
+                }
+                else  {
+                    push @OUT, sprintf "%04X\t%04X\t%s", $start, $end, $value;
+
+                    # Add a comment with the size of the range, if requested.
+                    # Expand Tabs to make sure they all start in the same
+                    # column, and then unexpand to use mostly tabs.
+                    if (! $output_range_counts) {
+                        $OUT[-1] .= "\n";
+                    }
+                    else {
+                        $OUT[-1] = Text::Tabs::expand($OUT[-1]);
+                        my $count = main::clarify_number($end - $start + 1);
+                        use integer;
+
+                        my $width = $tab_stops * 8 - 1;
+                        $OUT[-1] = sprintf("%-*s # [%s]\n",
+                                            $width,
+                                            $OUT[-1],
+                                            $count);
+                        $OUT[-1] = Text::Tabs::unexpand($OUT[-1]);
+                    }
+                }
+            } # End of loop through all the table's ranges
+        }
+
+        # Add anything that goes after the main body, but within the here
+        # document,
+        my $append_to_body = $self->append_to_body;
+        push @OUT, $append_to_body if $append_to_body;
+
+        # And finish the here document.
+        push @OUT, "END\n";
+
+        # All these files have a .pl suffix
+        $file_path{$addr}->[-1] .= '.pl';
+
+        main::write($file_path{$addr}, \@OUT);
+        return;
+    }
+
+    sub set_status {    # Set the table's status
+        my $self = shift;
+        my $status = shift; # The status enum value
+        my $info = shift;   # Any message associated with it.
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr($self);
+
+        $status{$addr} = $status;
+        $status_info{$addr} = $info;
+        return;
+    }
+
+    sub lock {
+        # Don't allow changes to the table from now on.  This stores a stack
+        # trace of where it was called, so that later attempts to modify it
+        # can immediately show where it got locked.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        $locked{$addr} = "";
+
+        my $line = (caller(0))[2];
+        my $i = 1;
+
+        # Accumulate the stack trace
+        while (1) {
+            my ($pkg, $file, $caller_line, $caller) = caller $i++;
+
+            last unless defined $caller;
+
+            $locked{$addr} .= "    called from $caller() at line $line\n";
+            $line = $caller_line;
+        }
+        $locked{$addr} .= "    called from main at line $line\n";
+
+        return;
+    }
+
+    sub carp_if_locked {
+        # Return whether a table is locked or not, and, by the way, complain
+        # if is locked
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        return 0 if ! $locked{$addr};
+        Carp::my_carp_bug("Can't modify a locked table. Stack trace of locking:\n$locked{$addr}\n\n");
+        return 1;
+    }
+
+    sub set_file_path { # Set the final directory path for this table
+        my $self = shift;
+        # Rest of parameters passed on
+
+        @{$file_path{main::objaddr $self}} = @_;
+        return
+    }
+
+    # Accessors for the range list stored in this table.  First for
+    # unconditional
+    for my $sub qw(
+                    contains
+                    count
+                    each_range
+                    hash
+                    is_empty
+                    max
+                    min
+                    range_count
+                    reset_each_range
+                    value_of
+                )
     {
-        my ($name, ## Name ("LATIN CAPITAL LETTER A")
-            $cat,  ## Category ("Lu", "Zp", "Nd", etc.)
-            $code, ## Code point (as an integer)
-            $op) = @_;
-
-        my $MajorCat = substr($cat, 0, 1); ## L, M, Z, S, etc
-
-        $Assigned->$op($code);
-        $Name->$op($code, $name);
-        $General->$op($code, $cat);
-
-        ## add to the sub category (e.g. "Lu", "Nd", "Cf", ..)
-        $Cat{$cat}      ||= Table->New(Is   => $cat,
-                                       Desc => "General Category '$cat'",
-                                       Fuzzy => 0);
-        $Cat{$cat}->$op($code);
-
-        ## add to the major category (e.g. "L", "N", "C", ...)
-        $Cat{$MajorCat} ||= Table->New(Is => $MajorCat,
-                                       Desc => "Major Category '$MajorCat'",
-                                       Fuzzy => 0);
-        $Cat{$MajorCat}->$op($code);
-
-        ($General{$name} ||= Table->New)->$op($code, $name);
-
-        # 005F: SPACING UNDERSCORE
-        $Cat{Word}->$op($code)  if $cat =~ /^[LMN]|Pc/;
-        $Cat{Alnum}->$op($code) if $cat =~ /^[LM]|Nd/;
-        $Cat{Alpha}->$op($code) if $cat =~ /^[LM]/;
-
-	my $isspace = 
-	    ($cat =~ /Zs|Zl|Zp/ &&
-	     $code != 0x200B)	# 200B is ZWSP which is for line break control
-				# and therefore it is not part of "space" even
-				# while it is "Zs" in some versions of Unicode.
-				# In 5.1 it is Cf, so this line is no longer
-				# necessary.
-                                || $code == 0x0009  # 0009: HORIZONTAL TAB
-                                || $code == 0x000A  # 000A: LINE FEED
-                                || $code == 0x000B  # 000B: VERTICAL TAB
-                                || $code == 0x000C  # 000C: FORM FEED
-                                || $code == 0x000D  # 000D: CARRIAGE RETURN
-                                || $code == 0x0085  # 0085: NEL
-
-	    ;
-
-        $Cat{Space}->$op($code) if $isspace;
-
-        $Cat{SpacePerl}->$op($code) if $isspace
-	                               && $code != 0x000B; # Backward compat.
-
-        $Cat{VertSpace}->$op($code) if grep {$code == $_} 
-            ( 0x0A..0x0D,0x85,0x2028,0x2029 );
-
-        $Cat{HorizSpace}->$op($code) if grep {$code == $_} (
-            0x09,   0x20,   0xa0,   0x1680, 0x180e, 0x2000, 0x2001, 0x2002,
-            0x2003, 0x2004, 0x2005, 0x2006, 0x2007, 0x2008, 0x2009, 0x200a,
-            0x202f, 0x205f, 0x3000
-        ); 
-
-        $Cat{Blank}->$op($code) if $isspace
-                                && !($code == 0x000A ||
-				     $code == 0x000B ||
-				     $code == 0x000C ||
-				     $code == 0x000D ||
-				     $code == 0x0085 ||
-				     $cat =~ /^Z[lp]/);
-
-        $Cat{Digit}->$op($code) if $cat eq "Nd";
-        $Cat{Upper}->$op($code) if $cat eq "Lu";
-        $Cat{Lower}->$op($code) if $cat eq "Ll";
-        $Cat{Title}->$op($code) if $cat eq "Lt";
-        $Cat{ASCII}->$op($code) if $code <= 0x007F;
-        $Cat{Cntrl}->$op($code) if $cat =~ /^C/;
-	my $isgraph = !$isspace && $cat !~ /Cc|Cs|Cn/;
-        $Cat{Graph}->$op($code) if $isgraph;
-        $Cat{Print}->$op($code) if $isgraph || $isspace;
-        $Cat{Punct}->$op($code) if $cat =~ /^P/;
-
-        $Cat{XDigit}->$op($code) if ($code >= 0x30 && $code <= 0x39)  ## 0..9
-                                 || ($code >= 0x41 && $code <= 0x46)  ## A..F
-                                 || ($code >= 0x61 && $code <= 0x66); ## a..f
-        if ($code<=0x7F) {
-            foreach my $cat (keys %TRUE_POSIX_PERL_CC) {
-                if ($TRUE_POSIX_PERL_CC{$cat}{$code}) {
-                    $Cat{$cat}->$op($code);
-                }
-            }
-        }
-    }
-
-    ## open and read file.....
-    if (not open IN, "UnicodeData.txt") {
-        die "$0: UnicodeData.txt: $!\n";
-    }
-
-    ##
-    ## For building \p{_CombAbove} and \p{_CanonDCIJ}
-    ##
-    my %_Above_HexCodes; ## Hexcodes for chars with $comb == 230 ("ABOVE")
-
-    my %CodeToDeco;      ## Maps code to decomp. list for chars with first
-                         ## decomp. char an "i" or "j" (for \p{_CanonDCIJ})
-
-    ## This is filled in as we go....
-    my $CombAbove = Table->New(Is   => '_CombAbove',
-                               Desc  => '(for internal casefolding use)',
-                               Fuzzy => 0);
-
-    while (<IN>)
+        no strict "refs";
+        *$sub = sub {
+            use strict "refs";
+            my $self = shift;
+            return $range_list{main::objaddr $self}->$sub(@_);
+        }
+    }
+
+    # Then for ones that should fail if locked
+    for my $sub qw(
+                    delete_range
+                )
     {
-        next unless /^[0-9A-Fa-f]+;/;
-        s/\s+$//;
-
-        my ($hexcode,   ## code point in hex (e.g. "0041")
-            $name,      ## character name (e.g. "LATIN CAPITAL LETTER A")
-            $cat,       ## category (e.g. "Lu")
-            $comb,      ## Canonical combining class (e.g. "230")
-            $bidi,      ## directional category (e.g. "L")
-            $deco,      ## decomposition mapping
-            $decimal,   ## decimal digit value
-            $digit,     ## digit value
-            $number,    ## numeric value
-            $mirrored,  ## mirrored
-            $unicode10, ## name in Unicode 1.0
-            $comment,   ## comment field
-            $upper,     ## uppercase mapping
-            $lower,     ## lowercase mapping
-            $title,     ## titlecase mapping
-              ) = split(/\s*;\s*/);
-
-	# Note that in Unicode 3.2 there will be names like
-	# LINE FEED (LF), which probably means that \N{} needs
-	# to cope also with LINE FEED and LF.
-	$name = $unicode10 if $name eq '<control>' && $unicode10 ne '';
-
-        my $code = hex($hexcode);
-
-        if ($comb and $comb == 230) {
-            $CombAbove->Append($code);
-            $_Above_HexCodes{$hexcode} = 1;
-        }
-
-        ## Used in building \p{_CanonDCIJ}
-        if ($deco and $deco =~ m/^006[9A]\b/) {
-            $CodeToDeco{$code} = $deco;
-        }
-
-        ##
-        ## There are a few pairs of lines like:
-        ##   AC00;<Hangul Syllable, First>;Lo;0;L;;;;;N;;;;;
-        ##   D7A3;<Hangul Syllable, Last>;Lo;0;L;;;;;N;;;;;
-        ## that define ranges.
-        ##
-        if ($name =~ /^<(.+), (First|Last)>$/)
-        {
-            $name = $1;
-            gencat($name, $cat, $code, $2 eq 'First' ? 'Append' : 'Extend');
-            #New_Prop(In => $name, $General{$name}, Fuzzy => 1);
+        no strict "refs";
+        *$sub = sub {
+            use strict "refs";
+            my $self = shift;
+
+            return if $self->carp_if_locked;
+            return $range_list{main::objaddr $self}->$sub(@_);
         }
-        else
-        {
-            ## normal (single-character) lines
-            gencat($name, $cat, $code, 'Append');
+    }
+
+} # End closure
+
+package Map_Table;
+use base '_Base_Table';
+
+# A Map Table is a table that contains the mappings from code points to
+# values.  There are two weird cases:
+# 1) Anomalous entries are ones that aren't maps of ranges of code points, but
+#    are written in the table's file at the end of the table nonetheless.  It
+#    requires specially constructed code to handle these; utf8.c can not read
+#    these in, so they should not go in $map_directory.  As of this writing,
+#    the only case that these happen is for named sequences used in
+#    charnames.pm.   But this code doesn't enforce any syntax on these, so
+#    something else could come along that uses it.
+# 2) Specials are anything that doesn't fit syntactically into the body of the
+#    table.  The ranges for these have a map type of non-zero.  The code below
+#    knows about and handles each possible type.   In most cases, these are
+#    written as part of the header.
+#
+# A map table deliberately can't be manipulated at will unlike match tables.
+# This is because of the ambiguities having to do with what to do with
+# overlapping code points.  And there just isn't a need for those things;
+# what one wants to do is just query, add, replace, or delete mappings, plus
+# write the final result.
+# However, there is a method to get the list of possible ranges that aren't in
+# this table to use for defaulting missing code point mappings.  And,
+# map_add_or_replace_non_nulls() does allow one to add another table to this
+# one, but it is clearly very specialized, and defined that the other's
+# non-null values replace this one's if there is any overlap.
+
+sub trace { return main::trace(@_); }
+
+{ # Closure
+
+    main::setup_package();
+
+    my %default_map;
+    # Many input files omit some entries; this gives what the mapping for the
+    # missing entries should be
+    main::set_access('default_map', \%default_map, 'r');
+
+    my %anomalous_entries;
+    # Things that go in the body of the table which don't fit the normal
+    # scheme of things, like having a range.  Not much can be done with these
+    # once there except to output them.  This was created to handle named
+    # sequences.
+    main::set_access('anomalous_entry', \%anomalous_entries, 'a');
+    main::set_access('anomalous_entries',       # Append singular, read plural
+                    \%anomalous_entries,
+                    'readable_array');
+
+    my %format;
+    # The format of the entries of the table.  This is calculated from the
+    # data in the table (or passed in the constructor).  This is an enum e.g.,
+    # $STRING_FORMAT
+    main::set_access('format', \%format);
+
+    my %core_access;
+    # This is a string, solely for documentation, indicating how one can get
+    # access to this property via the Perl core.
+    main::set_access('core_access', \%core_access, 'r', 's');
+
+    my %has_specials;
+    # Boolean set when non-zero map-type ranges are added to this table,
+    # which happens in only a few tables.  This is purely for performance, to
+    # avoid having to search through every table upon output, so if all the
+    # non-zero maps got deleted before output, this would remain set, and the
+    # only penalty would be performance.  Currently, most map tables that get
+    # output have specials in them, so this doesn't help that much anyway.
+    main::set_access('has_specials', \%has_specials);
+
+    my %to_output_map;
+    # Boolean as to whether or not to write out this map table
+    main::set_access('to_output_map', \%to_output_map, 's');
+
+
+    sub new {
+        my $class = shift;
+        my $name = shift;
+
+        my %args = @_;
+
+        # Optional initialization data for the table.
+        my $initialize = delete $args{'Initialize'};
+
+        my $core_access = delete $args{'Core_Access'};
+        my $default_map = delete $args{'Default_Map'};
+        my $format = delete $args{'Format'};
+        my $property = delete $args{'_Property'};
+        # Rest of parameters passed on
+
+        my $range_list = Range_Map->new(Owner => $property);
+
+        my $self = $class->SUPER::new(
+                                    Name => $name,
+                                    _Property => $property,
+                                    _Range_List => $range_list,
+                                    %args);
+
+        my $addr = main::objaddr $self;
+
+        $anomalous_entries{$addr} = [];
+        $core_access{$addr} = $core_access;
+        $default_map{$addr} = $default_map;
+        $format{$addr} = $format;
+
+        $self->initialize($initialize) if defined $initialize;
+
+        return $self;
+    }
+
+    use overload
+        fallback => 0,
+        qw("") => "_operator_stringify",
+    ;
+
+    sub _operator_stringify {
+        my $self = shift;
+
+        my $name = $self->property->full_name;
+        $name = '""' if $name eq "";
+        return "Map table for Property '$name'";
+    }
+
+    sub complete_name {
+        # The complete name for a map table is just its full name, as that
+        # completely identifies the property it represents
+
+        return shift->full_name;
+    }
+
+    sub add_alias {
+        # Add a synonym for this table (which means the property itself)
+        my $self = shift;
+        my $name = shift;
+        # Rest of parameters passed on.
+
+        $self->SUPER::add_alias($name, $self->property, @_);
+        return;
+    }
+
+    sub add_map {
+        # Add a range of code points to the list of specially-handled code
+        # points.  $MULTI_CP is assumed if the type of special is not passed
+        # in.
+
+        my $self = shift;
+        my $lower = shift;
+        my $upper = shift;
+        my $string = shift;
+        my %args = @_;
+
+        my $type = delete $args{'Type'} || 0;
+        # Rest of parameters passed on
+
+        # Can't change the table if locked.
+        return if $self->carp_if_locked;
+
+        my $addr = main::objaddr $self;
 
-            # No Append() here since since several codes may map into one.
-            $To{Upper}->RawAppendRange($code, $code, $upper) if $upper;
-            $To{Lower}->RawAppendRange($code, $code, $lower) if $lower;
-            $To{Title}->RawAppendRange($code, $code, $title) if $title;
-            $To{Digit}->Append($code, $decimal) if length $decimal;
+        $has_specials{$addr} = 1 if $type;
 
-            $Bidi->Append($code, $bidi);
-            $Comb->Append($code, $comb) if $comb;
-            $Number->Append($code, $number) if length $number;
+        $self->_range_list->add_map($lower, $upper,
+                                    $string,
+                                    @_,
+                                    Type => $type);
+        return;
+    }
+
+    sub append_to_body {
+        # Adds to the written HERE document of the table's body any anomalous
+        # entries in the table..
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        return "" unless @{$anomalous_entries{$addr}};
+        return join("\n", @{$anomalous_entries{$addr}}) . "\n";
+    }
+
+    sub map_add_or_replace_non_nulls {
+        # This adds the mappings in the table $other to $self.  Non-null
+        # mappings from $other override those in $self.  It essentially merges
+        # the two tables, with the second having priority except for null
+        # mappings.
+
+        my $self = shift;
+        my $other = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        return if $self->carp_if_locked;
+
+        if (! $other->isa(__PACKAGE__)) {
+            Carp::my_carp_bug("$other should be a "
+                        . __PACKAGE__
+                        . ".  Not a '"
+                        . ref($other)
+                        . "'.  Not added;");
+            return;
+        }
 
-	    length($decimal) and ($Number{De} ||= Table->New())->Append($code)
-	      or
-	    length($digit)   and ($Number{Di} ||= Table->New())->Append($code)
-	      or
-	    length($number)  and ($Number{Nu} ||= Table->New())->Append($code);
+        my $addr = main::objaddr $self;
+        my $other_addr = main::objaddr $other;
+
+        local $to_trace = 0 if main::DEBUG;
+
+        my $self_range_list = $self->_range_list;
+        my $other_range_list = $other->_range_list;
+        foreach my $range ($other_range_list->ranges) {
+            my $value = $range->value;
+            next if $value eq "";
+            $self_range_list->_add_delete('+',
+                                          $range->start,
+                                          $range->end,
+                                          $value,
+                                          Type => $range->type,
+                                          Replace => $UNCONDITIONALLY);
+        }
 
-            $Mirrored->Append($code) if $mirrored eq "Y";
+        # Copy the specials information from the other table to $self
+        if ($has_specials{$other_addr}) {
+            $has_specials{$addr} = 1;
+        }
 
-            $Bidi{$bidi} ||= Table->New();#Is    => "bt/$bidi",
-                                        #Desc  => "Bi-directional category '$bidi'",
-                                        #Fuzzy => 0);
-            $Bidi{$bidi}->Append($code);
+        return;
+    }
 
-            if ($deco)
+    sub set_default_map {
+        # Define what code points that are missing from the input files should
+        # map to
+
+        my $self = shift;
+        my $map = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        # Convert the input to the standard equivalent, if any (won't have any
+        # for $STRING properties)
+        my $standard = $self->_find_table_from_alias->{$map};
+        $map = $standard->name if defined $standard;
+
+        # Warn if there already is a non-equivalent default map for this
+        # property.  Note that a default map can be a ref, which means that
+        # what it actually means is delayed until later in the program, and it
+        # IS permissible to override it here without a message.
+        my $default_map = $default_map{$addr};
+        if (defined $default_map
+            && ! ref($default_map)
+            && $default_map ne $map
+            && main::Standardize($map) ne $default_map)
+        {
+            my $property = $self->property;
+            my $map_table = $property->table($map);
+            my $default_table = $property->table($default_map);
+            if (defined $map_table
+                && defined $default_table
+                && $map_table != $default_table)
             {
-                $Deco->Append($code, $deco);
-                if ($deco =~/^<(\w+)>/)
-                {
-		    my $dshort = $PVA_reverse{dt}{ucfirst lc $1};
-		    $DC{Com}->Append($code);
-		    $dshort = $PVA_reverse{dt}{lc $1} unless $dshort ne "";
-		    die "No reverse for $1'" unless $dshort ne "";
-		    #$dshort = lc $dshort;   # use lower case only
-		    $DC{$dshort} ||= Table->New();
-		    $DC{$dshort}->Append($code);
-                }
-                else
-                {
-                    $DC{Can}->Append($code);
+                Carp::my_carp("Changing the default mapping for "
+                            . $property
+                            . " from $default_map to $map'");
+            }
+        }
+
+        $default_map{$addr} = $map;
+
+        # Don't also create any missing table for this map at this point,
+        # because if we did, it could get done before the main table add is
+        # done for PropValueAliases.txt; instead the caller will have to make
+        # sure it exists, if desired.
+        return;
+    }
+
+    sub to_output_map {
+        # Returns boolean: should we write this map table?
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        # If overridden, use that
+        return $to_output_map{$addr} if defined $to_output_map{$addr};
+
+        my $full_name = $self->full_name;
+
+        # If table says to output, do so; if says to suppress it, do do.
+        return 1 if grep { $_ eq $full_name } @output_mapped_properties;
+        return 0 if $self->status eq $SUPPRESSED;
+
+        my $type = $self->property->type;
+
+        # Don't want to output binary map tables even for debugging.
+        return 0 if $type == $BINARY;
+
+        # But do want to output string ones.
+        return 1 if $type == $STRING;
+
+        # Otherwise is an $ENUM, don't output it
+        return 0;
+    }
+
+    sub inverse_list {
+        # Returns a Range_List that is gaps of the current table.  That is,
+        # the inversion
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $current = Range_List->new(Initialize => $self->_range_list,
+                                Owner => $self->property);
+        return ~ $current;
+    }
+
+    sub set_final_comment {
+        # Just before output, create the comment that heads the file
+        # containing this table.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        # No sense generating a comment if aren't going to write it out.
+        return if ! $self->to_output_map;
+
+        my $addr = main::objaddr $self;
+
+        my $property = $self->property;
+
+        # Get all the possible names for this property.  Don't use any that
+        # aren't ok for use in a file name, etc.  This is perhaps causing that
+        # flag to do double duty, and may have to be changed in the future to
+        # have our own flag for just this purpose; but it works now to exclude
+        # Perl generated synonyms from the lists for properties, where the
+        # name is always the proper Unicode one.
+        my @property_aliases = grep { $_->externally_ok } $self->aliases;
+
+        my $count = $self->count;
+        my $default_map = $default_map{$addr};
+
+        # The ranges that map to the default aren't output, so subtract that
+        # to get those actually output.  A property with matching tables
+        # already has the information calculated.
+        if ($property->type != $STRING) {
+            $count -= $property->table($default_map)->count;
+        }
+        elsif (defined $default_map) {
+
+            # But for $STRING properties, must calculate now.  Subtract the
+            # count from each range that maps to the default.
+            foreach my $range ($self->_range_list->ranges) {
+        local $to_trace = 1 if main::DEBUG;
+        trace $self, $range;
+                if ($range->value eq $default_map) {
+                    $count -= $range->end +1 - $range->start;
+                }
+            }
+
+        }
+
+        # Get a  string version of $count with underscores in large numbers,
+        # for clarity.
+        my $string_count = main::clarify_number($count);
+
+        my $code_points = ($count == 1)
+                        ? 'single code point'
+                        : "$string_count code points";
+
+        my $mapping;
+        my $these_mappings;
+        my $are;
+        if (@property_aliases <= 1) {
+            $mapping = 'mapping';
+            $these_mappings = 'this mapping';
+            $are = 'is'
+        }
+        else {
+            $mapping = 'synonymous mappings';
+            $these_mappings = 'these mappings';
+            $are = 'are'
+        }
+        my $cp;
+        if ($count >= $MAX_UNICODE_CODEPOINTS) {
+            $cp = "any code point in Unicode Version $string_version";
+        }
+        else {
+            my $map_to;
+            if ($default_map eq "") {
+                $map_to = 'the null string';
+            }
+            elsif ($default_map eq $CODE_POINT) {
+                $map_to = "itself";
+            }
+            else {
+                $map_to = "'$default_map'";
+            }
+            if ($count == 1) {
+                $cp = "the single code point";
+            }
+            else {
+                $cp = "one of the $code_points";
+            }
+            $cp .= " in Unicode Version $string_version for which the mapping is not to $map_to";
+        }
+
+        my $comment = "";
+
+        my $status = $self->status;
+        if ($status) {
+            my $warn = uc $status_past_participles{$status};
+            $comment .= <<END;
+
+!!!!!!!   $warn !!!!!!!!!!!!!!!!!!!
+ All property or property=value combinations contained in this file are $warn.
+ See $unicode_reference_url for what this means.
+
+END
+        }
+        $comment .= "This file returns the $mapping:\n";
+
+        for my $i (0 .. @property_aliases - 1) {
+            $comment .= sprintf("%-8s%s\n",
+                                " ",
+                                $property_aliases[$i]->name . '(cp)'
+                                );
+        }
+        $comment .=
+                "\nwhere 'cp' is $cp.  Note that $these_mappings $are ";
+
+        my $access = $core_access{$addr};
+        if ($access) {
+            $comment .= "accessible through the Perl core via $access.";
+        }
+        else {
+            $comment .= "not accessible through the Perl core directly.";
+        }
+
+        # And append any commentary already set from the actual property.
+        $comment .= "\n\n" . $self->comment if $self->comment;
+        if ($self->description) {
+            $comment .= "\n\n" . join " ", $self->description;
+        }
+        if ($self->note) {
+            $comment .= "\n\n" . join " ", $self->note;
+        }
+        $comment .= "\n";
+
+        if (! $self->perl_extension) {
+            $comment .= <<END;
+
+For information about what this property really means, see:
+$unicode_reference_url
+END
+        }
+
+        if ($count) {        # Format differs for empty table
+                $comment.= "\nThe format of the ";
+            if ($self->range_size_1) {
+                $comment.= <<END;
+main body of lines of this file is: CODE_POINT\\t\\tMAPPING where CODE_POINT
+is in hex; MAPPING is what CODE_POINT maps to.
+END
+            }
+            else {
+
+                # There are tables which end up only having one element per
+                # range, but it is not worth keeping track of for making just
+                # this comment a little better.
+                $comment.= <<END;
+non-comment portions of the main body of lines of this file is:
+START\\tSTOP\\tMAPPING where START is the starting code point of the
+range, in hex; STOP is the ending point, or if omitted, the range has just one
+code point; MAPPING is what each code point between START and STOP maps to.
+END
+                if ($output_range_counts) {
+                    $comment .= <<END;
+Numbers in comments in [brackets] indicate how many code points are in the
+range (omitted when the range is a single code point or if the mapping is to
+the null string).
+END
+                }
+            }
+        }
+        $self->set_comment(main::join_lines($comment));
+        return;
+    }
+
+    my %swash_keys; # Makes sure don't duplicate swash names.
+
+    sub pre_body {
+        # Returns the string that should be output in the file before the main
+        # body of this table.  This includes some hash entries identifying the
+        # format of the body, and what the single value should be for all
+        # ranges missing from it.  It also includes any code points which have
+        # map_types that don't go in the main table.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        my $name = $self->property->swash_name;
+
+        if (defined $swash_keys{$name}) {
+            Carp::my_carp(join_lines(<<END
+Already created a swash name '$name' for $swash_keys{$name}.  This means that
+the same name desired for $self shouldn't be used.  Bad News.  This must be
+fixed before production use, but proceeding anyway
+END
+            ));
+        }
+        $swash_keys{$name} = "$self";
+
+        my $default_map = $default_map{$addr};
+
+        my $pre_body = "";
+        if ($has_specials{$addr}) {
+
+            # Here, some maps with non-zero type have been added to the table.
+            # Go through the table and handle each of them.  None will appear
+            # in the body of the table, so delete each one as we go.  The
+            # code point count has already been calculated, so ok to delete
+            # now.
+
+            my @multi_code_point_maps;
+            my $has_hangul_syllables = 0;
+
+            # The key is the base name of the code point, and the value is an
+            # array giving all the ranges that use this base name.  Each range
+            # is actually a hash giving the 'low' and 'high' values of it.
+            my %names_ending_in_code_point;
+
+            # Inverse mapping.  The list of ranges that have these kinds of
+            # names.  Each element contains the low, high, and base names in a
+            # hash.
+            my @code_points_ending_in_code_point;
+
+            my $range_map = $self->_range_list;
+            foreach my $range ($range_map->ranges) {
+                next unless $range->type != 0;
+                my $low = $range->start;
+                my $high = $range->end;
+                my $map = $range->value;
+                my $type = $range->type;
+
+                # No need to output the range if it maps to the default.  And
+                # the write method won't output it either, so no need to
+                # delete it to keep it from being output, and is faster to
+                # skip than to delete anyway.
+                next if $map eq $default_map;
+
+                # Delete the range to keep write() from trying to output it
+                $range_map->delete_range($low, $high);
+
+                # Switch based on the map type...
+                if ($type == $HANGUL_SYLLABLE) {
+
+                    # These are entirely algorithmically determinable based on
+                    # some constants furnished by Unicode; for now, just set a
+                    # flag to indicate that have them.  Below we will output
+                    # the code that does the algorithm.
+                    $has_hangul_syllables = 1;
                 }
+                elsif ($type == $CP_IN_NAME) {
+
+                    # If the name ends in the code point it represents, are
+                    # also algorithmically determinable, but need information
+                    # about the map to do so.  Both the map and its inverse
+                    # are stored in data structures output in the file.
+                    push @{$names_ending_in_code_point{$map}->{'low'}}, $low;
+                    push @{$names_ending_in_code_point{$map}->{'high'}}, $high;
+
+                    push @code_points_ending_in_code_point, { low => $low,
+                                                              high => $high,
+                                                              name => $map
+                                                            };
+                }
+                elsif ($range->type == $MULTI_CP || $range->type == $NULL) {
+
+                    # Multi-code point maps and null string maps have an entry
+                    # for each code point in the range.  They use the same
+                    # output format.
+                    for my $code_point ($low .. $high) {
+
+                        # The pack() below can't cope with surrogates.
+                        if ($code_point >= 0xD800 && $code_point <= 0xDFFF) {
+                            Carp::my_carp("Surrogage code point '$code_point' in mapping to '$map' in $self.  No map created");
+                            next;
+                        }
+
+                        # Generate the hash entries for these in the form that
+                        # utf8.c understands.
+                        my $tostr = "";
+                        foreach my $to (split " ", $map) {
+                            if ($to !~ /^$code_point_re$/) {
+                                Carp::my_carp("Illegal code point '$to' in mapping '$map' from $code_point in $self.  No map created");
+                                next;
+                            }
+                            $tostr .= sprintf "\\x{%s}", $to;
+                        }
+
+                        # I (khw) have never waded through this line to
+                        # understand it well enough to comment it.
+                        my $utf8 = sprintf(qq["%s" => "$tostr",],
+                                join("", map { sprintf "\\x%02X", $_ }
+                                    unpack("U0C*", pack("U", $code_point))));
+
+                        # Add a comment so that a human reader can more easily
+                        # see what's going on.
+                        push @multi_code_point_maps,
+                                sprintf("%-45s # U+%04X => %s", $utf8,
+                                                                $code_point,
+                                                                $map);
+                    }
+                }
+                else {
+                    Carp::my_carp("Unrecognized map type '$range->type' in '$range' in $self.  Using type 0 instead");
+                    $range_map->add_map($low, $high, $map, Replace => $UNCONDITIONALLY, Type => 0);
+                }
+            } # End of loop through all ranges
+
+            # Here have gone through the whole file.  If actually generated
+            # anything for each map type, add its respective header and
+            # trailer
+            if (@multi_code_point_maps) {
+                $pre_body .= <<END;
+
+# Some code points require special handling because their mappings are each to
+# multiple code points.  These do not appear in the main body, but are defined
+# in the hash below.
+
+# The key: UTF-8 _bytes_, the value: UTF-8 (speed hack)
+%utf8::ToSpec$name = (
+END
+                $pre_body .= join("\n", @multi_code_point_maps) . "\n);\n";
             }
+
+            if ($has_hangul_syllables || @code_points_ending_in_code_point) {
+
+                # Convert these structures to output format.
+                my $code_points_ending_in_code_point =
+                    main::simple_dumper(\@code_points_ending_in_code_point,
+                                        ' ' x 8);
+                my $names = main::simple_dumper(\%names_ending_in_code_point,
+                                                ' ' x 8);
+
+                # Do the same with the Hangul names,
+                my $jamo;
+                my $jamo_l;
+                my $jamo_v;
+                my $jamo_t;
+                my $jamo_re;
+                if ($has_hangul_syllables) {
+
+                    # Construct a regular expression of all the possible
+                    # combinations of the Hangul syllables.
+                    my @L_re;   # Leading consonants
+                    for my $i ($LBase .. $LBase + $LCount - 1) {
+                        push @L_re, $Jamo{$i}
+                    }
+                    my @V_re;   # Middle vowels
+                    for my $i ($VBase .. $VBase + $VCount - 1) {
+                        push @V_re, $Jamo{$i}
+                    }
+                    my @T_re;   # Trailing consonants
+                    for my $i ($TBase + 1 .. $TBase + $TCount - 1) {
+                        push @T_re, $Jamo{$i}
+                    }
+
+                    # The whole re is made up of the L V T combination.
+                    $jamo_re = '('
+                               . join ('|', sort @L_re)
+                               . ')('
+                               . join ('|', sort @V_re)
+                               . ')('
+                               . join ('|', sort @T_re)
+                               . ')?';
+
+                    # These hashes needed by the algorithm were generated
+                    # during reading of the Jamo.txt file
+                    $jamo = main::simple_dumper(\%Jamo, ' ' x 8);
+                    $jamo_l = main::simple_dumper(\%Jamo_L, ' ' x 8);
+                    $jamo_v = main::simple_dumper(\%Jamo_V, ' ' x 8);
+                    $jamo_t = main::simple_dumper(\%Jamo_T, ' ' x 8);
+                }
+
+                $pre_body .= <<END;
+
+# To achieve significant memory savings when this file is read in,
+# algorithmically derivable code points are omitted from the main body below.
+# Instead, the following routines can be used to translate between name and
+# code point and vice versa
+
+{ # Closure
+
+    # Matches legal code point.  4-6 hex numbers, If there are 6, the
+    # first two must be '10'; if there are 5, the first must not be a '0'.
+    my \$code_point_re = qr/$code_point_re/;
+
+    # In the following hash, the keys are the bases of names which includes
+    # the code point in the name, like CJK UNIFIED IDEOGRAPH-4E01.  The values
+    # of each key is another hash which is used to get the low and high ends
+    # for each range of code points that apply to the name
+    my %names_ending_in_code_point = (
+$names
+    );
+
+    # And the following array gives the inverse mapping from code points to
+    # names.  Lowest code points are first
+    my \@code_points_ending_in_code_point = (
+$code_points_ending_in_code_point
+    );
+END
+                # Earlier releases didn't have Jamos.  No sense outputting
+                # them unless will be used.
+                if ($has_hangul_syllables) {
+                    $pre_body .= <<END;
+
+    # Convert from code point to Jamo short name for use in composing Hangul
+    # syllable names
+    my %Jamo = (
+$jamo
+    );
+
+    # Leading consonant (can be null)
+    my %Jamo_L = (
+$jamo_l
+    );
+
+    # Vowel
+    my %Jamo_V = (
+$jamo_v
+    );
+
+    # Optional trailing consonant
+    my %Jamo_T = (
+$jamo_t
+    );
+
+    # Computed re that splits up a Hangul name into LVT or LV syllables
+    my \$syllable_re = qr/$jamo_re/;
+
+    my \$HANGUL_SYLLABLE = "HANGUL SYLLABLE ";
+    my \$HANGUL_SYLLABLE_LENGTH = length \$HANGUL_SYLLABLE;
+
+    # These constants names and values were taken from the Unicode standard,
+    # version 5.1, section 3.12.  They are used in conjunction with Hangul
+    # syllables
+    my \$SBase = 0xAC00;
+    my \$LBase = 0x1100;
+    my \$VBase = 0x1161;
+    my \$TBase = 0x11A7;
+    my \$SCount = 11172;
+    my \$LCount = 19;
+    my \$VCount = 21;
+    my \$TCount = 28;
+    my \$NCount = \$VCount * \$TCount;
+END
+                } # End of has Jamos
+
+                $pre_body .= << 'END';
+
+    sub name_to_code_point_special {
+        my $name = shift;
+
+        # Returns undef if not one of the specially handled names; otherwise
+        # returns the code point equivalent to the input name
+END
+                if ($has_hangul_syllables) {
+                    $pre_body .= << 'END';
+
+        if (substr($name, 0, $HANGUL_SYLLABLE_LENGTH) eq $HANGUL_SYLLABLE) {
+            $name = substr($name, $HANGUL_SYLLABLE_LENGTH);
+            return if $name !~ qr/^$syllable_re$/;
+            my $L = $Jamo_L{$1};
+            my $V = $Jamo_V{$2};
+            my $T = (defined $3) ? $Jamo_T{$3} : 0;
+            return ($L * $VCount + $V) * $TCount + $T + $SBase;
+        }
+END
+                }
+                $pre_body .= << 'END';
+
+        # Name must end in '-code_point' for this to handle.
+        if ($name !~ /^ (.*) - ($code_point_re) $/x) {
+            return;
         }
+
+        my $base = $1;
+        my $code_point = CORE::hex $2;
+
+        # Name must be one of the ones which has the code point in it.
+        return if ! $names_ending_in_code_point{$base};
+
+        # Look through the list of ranges that apply to this name to see if
+        # the code point is in one of them.
+        for (my $i = 0; $i < scalar @{$names_ending_in_code_point{$base}{'low'}}; $i++) {
+            return if $names_ending_in_code_point{$base}{'low'}->[$i] > $code_point;
+            next if $names_ending_in_code_point{$base}{'high'}->[$i] < $code_point;
+
+            # Here, the code point is in the range.
+            return $code_point;
+        }
+
+        # Here, looked like the name had a code point number in it, but
+        # did not match one of the valid ones.
+        return;
     }
-    close IN;
 
-    ## Read in the NameAliases.txt.  It contains other normative names of code
-    ## points not listed in UnicodeData.txt.  This happens when there is an
-    ## error in the name found after the data base was published, but instead of
-    ## changing it, to avoid breaking any code that came to rely on the
-    ## erroneous version, the correct name is added as an alias.
-    
-    my $NameAliases = Table->New();
+    sub code_point_to_name_special {
+        my $code_point = shift;
+
+        # Returns the name of a code point if algorithmically determinable;
+        # undef if not
+END
+                if ($has_hangul_syllables) {
+                    $pre_body .= << 'END';
+
+        # If in the Hangul range, calculate the name based on Unicode's
+        # algorithm
+        if ($code_point >= $SBase && $code_point <= $SBase + $SCount -1) {
+            use integer;
+            my $SIndex = $code_point - $SBase;
+            my $L = $LBase + $SIndex / $NCount;
+            my $V = $VBase + ($SIndex % $NCount) / $TCount;
+            my $T = $TBase + $SIndex % $TCount;
+            $name = "$HANGUL_SYLLABLE $Jamo{$L}$Jamo{$V}";
+            $name .= $Jamo{$T} if $T != $TBase;
+            return $name;
+        }
+END
+                }
+                $pre_body .= << 'END';
 
-    if (not open IN, "NameAliases.txt") {
-        die "$0: NameAliases.txt: $!\n";
+        # Look through list of these code points for one in range.
+        foreach my $hash (@code_points_ending_in_code_point) {
+            return if $code_point < $hash->{'low'};
+            if ($code_point <= $hash->{'high'}) {
+                return sprintf("%s-%04X", $hash->{'name'}, $code_point);
+            }
+        }
+        return;            # None found
     }
+} # End closure
+
+END
+            } # End of has hangul or code point in name maps.
+        } # End of has specials
+
+        # Calculate the format of the table if not already done.
+        my $format = $format{$addr};
+        my $property = $self->property;
+        my $type = $property->type;
+        if (! defined $format) {
+            if ($type == $BINARY) {
+
+                # Don't bother checking the values, because we elsewhere
+                # verify that a binary table has only 2 values.
+                $format = $BINARY_FORMAT;
+            }
+            else {
+                my @ranges = $self->_range_list->ranges;
+
+                # default an empty table based on its type and default map
+                if (! @ranges) {
 
-    while (<IN>)
+                    # But it turns out that the only one we can say is a
+                    # non-string (besides binary, handled above) is when the
+                    # table is a string and the default map is to a code point
+                    if ($type == $STRING && $default_map eq $CODE_POINT) {
+                        $format = $HEX_FORMAT;
+                    }
+                    else {
+                        $format = $STRING_FORMAT;
+                    }
+                }
+                else {
+
+                    # Start with the most restrictive format, and as we find
+                    # something that doesn't fit with that, change to the next
+                    # most restrictive, and so on.
+                    $format = $DECIMAL_FORMAT;
+                    foreach my $range (@ranges) {
+                        my $map = $range->value;
+                        if ($map ne $default_map) {
+                            last if $format eq $STRING_FORMAT;  # already at
+                                                                # least
+                                                                # restrictive
+                            $format = $INTEGER_FORMAT
+                                                if $format eq $DECIMAL_FORMAT
+                                                    && $map !~ / ^ [0-9] $ /x;
+                            $format = $FLOAT_FORMAT
+                                            if $format eq $INTEGER_FORMAT
+                                                && $map !~ / ^ -? [0-9]+ $ /x;
+                            $format = $RATIONAL_FORMAT
+                                if $format eq $FLOAT_FORMAT
+                                    && $map !~ / ^ -? [0-9]+ \. [0-9]* $ /x;
+                            $format = $HEX_FORMAT
+                            if $format eq $RATIONAL_FORMAT
+                                && $map !~ / ^ -? [0-9]+ ( \/ [0-9]+ )? $ /x;
+                            $format = $STRING_FORMAT if $format eq $HEX_FORMAT
+                                                       && $map =~ /[^0-9A-F]/;
+                        }
+                    }
+                }
+            }
+        } # end of calculating format
+
+        my $return = <<END;
+# The name this swash is to be known by, with the format of the mappings in
+# the main body of the table, and what all code points missing from this file
+# map to.
+\$utf8::SwashInfo{'To$name'}{'format'} = '$format'; # $map_table_formats{$format}
+END
+        my $missing = $default_map;
+        if ($missing eq $CODE_POINT
+            && $format ne $HEX_FORMAT
+            && ! defined $format{$addr})    # Is expected if was manually set
+        {
+            Carp::my_carp_bug("Expecting hex format for mapping table for $self, instead got '$format'")
+        }
+        $format{$addr} = $format;
+        $return .= "\$utf8::SwashInfo{'To$name'}{'missing'} = '$missing';";
+        if ($missing eq $CODE_POINT) {
+            $return .= ' # code point maps to itself';
+        }
+        elsif ($missing eq "") {
+            $return .= ' # code point maps to the null string';
+        }
+        $return .= "\n";
+
+        $return .= $pre_body;
+
+        return $return;
+    }
+
+    sub write {
+        # Write the table to the file.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        return $self->SUPER::write(
+            ($self->property == $block)
+                ? 7     # block file needs more tab stops
+                : 3,
+            $default_map{$addr});   # don't write defaulteds
+    }
+
+    # Accessors for the underlying list that should fail if locked.
+    for my $sub qw(
+                    add_duplicate
+                )
     {
-        next unless /^[0-9A-Fa-f]+;/;
-        s/\s+$//;
-
-        my ($hexcode,   ## code point in hex (e.g. "0041")
-            $name,      ## character name (e.g. "LATIN CAPITAL LETTER A")
-              ) = split(/\s*;\s*/);
-
-        my $code = hex($hexcode);
-
-	## One is supposed to enter elements into tables in strictly increasing
-	## order, but this in fact works to append duplicate code points at
-	## the end of the table.  The table is intended to be indexed by name
-	## anyway.
-
-        $Name->RawAppendRange($code, $code, $name);
-    }
-    close IN;
-
-
-    ##
-    ## Tidy up a few special cases....
-    ##
-
-    $Cat{Cn} = $Assigned->Invert; ## Cn is everything that doesn't exist
-    New_Prop(Is => 'Cn',
-             $Cat{Cn},
-             Desc => "General Category 'Cn' [not functional in Perl]",
-             Fuzzy => 0);
-
-    ## Unassigned is the same as 'Cn'
-    New_Alias(Is => 'Unassigned', SameAs => 'Cn', Fuzzy => 0);
-
-    $Cat{C}->Replace($Cat{C}->Merge($Cat{Cn}));  ## Now merge in Cn into C
-
-
-    # LC is Ll, Lu, and Lt.
-    # (used to be L& or L_, but PropValueAliases.txt defines it as LC)
-    New_Prop(Is => 'LC',
-             Table->Merge(@Cat{qw[Ll Lu Lt]}),
-             Desc  => '[\p{Ll}\p{Lu}\p{Lt}]',
-             Fuzzy => 0);
-
-    ## Any and All are all code points.
-    my $Any = Table->New(Is    => 'Any',
-                         Desc  => sprintf("[\\x{0000}-\\x{%X}]",
-                                          $LastUnicodeCodepoint),
-                         Fuzzy => 0);
-    $Any->RawAppendRange(0, $LastUnicodeCodepoint);
-
-    New_Alias(Is => 'All', SameAs => 'Any', Fuzzy => 0);
-
-    ##
-    ## Build special properties for Perl's internal case-folding needs:
-    ##    \p{_CaseIgnorable}
-    ##    \p{_CanonDCIJ}
-    ##    \p{_CombAbove}
-    ## _CombAbove was built above. Others are built here....
-    ##
-
-    ## \p{_CaseIgnorable} is [\p{Mn}\0x00AD\x2010]
-    New_Prop(Is => '_CaseIgnorable',
-             Table->Merge($Cat{Mn},
-                          0x00AD,    #SOFT HYPHEN
-                          0x2010),   #HYPHEN
-             Desc  => '(for internal casefolding use)',
-             Fuzzy => 0);
-
-
-    ## \p{_CanonDCIJ} is fairly complex...
-    my $CanonCDIJ = Table->New(Is    => '_CanonDCIJ',
-                               Desc  => '(for internal casefolding use)',
-                               Fuzzy => 0);
-    ## It contains the ASCII 'i' and 'j'....
-    $CanonCDIJ->Append(0x0069); # ASCII ord("i")
-    $CanonCDIJ->Append(0x006A); # ASCII ord("j")
-    ## ...and any character with a decomposition that starts with either of
-    ## those code points, but only if the decomposition does not have any
-    ## combining character with the "ABOVE" canonical combining class.
-    for my $code (sort { $a <=> $b} keys %CodeToDeco)
+        no strict "refs";
+        *$sub = sub {
+            use strict "refs";
+            my $self = shift;
+
+            return if $self->carp_if_locked;
+            return $self->_range_list->$sub(@_);
+        }
+    }
+} # End closure for Map_Table
+
+package Match_Table;
+use base '_Base_Table';
+
+# A Match table is one which is a list of all the code points that have
+# the same property and property value, for use in \p{property=value}
+# constructs in regular expressions.  It adds very little data to the base
+# structure, but many methods, as these lists can be combined in many ways to
+# form new ones.
+# There are only a few concepts added:
+# 1) Equivalents and Relatedness.
+#    Two tables can match the identical code points, but have different names.
+#    This always happens when there is a perl single form extension
+#    \p{IsProperty} for the Unicode compound form \P{Property=True}.  The two
+#    tables are set to be related, with the Perl extension being a child, and
+#    the Unicode property being the parent.
+#
+#    It may be that two tables match the identical code points and we don't
+#    know if they are related or not.  This happens most frequently when the
+#    Block and Script properties have the exact range.  But note that a
+#    revision to Unicode could add new code points to the script, which would
+#    now have to be in a different block (as the block was filled, or there
+#    would have been 'Unknown' script code points in it and they wouldn't have
+#    been identical).  So we can't rely on any two properties from Unicode
+#    always matching the same code points from release to release, and thus
+#    these tables are considered coincidentally equivalent--not related.  When
+#    two tables are unrelated but equivalent, one is arbitrarily chosen as the
+#    'leader', and the others are 'equivalents'.  This concept is useful
+#    to minimize the number of tables written out.  Only one file is used for
+#    any identical set of code points, with entries in Heavy.pl mapping all
+#    the involved tables to it.
+#
+#    Related tables will always be identical; we set them up to be so.  Thus
+#    if the Unicode one is deprecated, the Perl one will be too.  Not so for
+#    unrelated tables.  Relatedness makes generating the documentation easier.
+#
+# 2) Conflicting.  It may be that there will eventually be name clashes, with
+#    the same name meaning different things.  For a while, there actually were
+#    conflicts, but they have so far been resolved by changing Perl's or
+#    Unicode's definitions to match the other, but when this code was written,
+#    it wasn't clear that that was what was going to happen.  (Unicode changed
+#    because of protests during their beta period.)  Name clashes are warned
+#    about during compilation, and the documentation.  The generated tables
+#    are sane, free of name clashes, because the code suppresses the Perl
+#    version.  But manual intervention to decide what the actual behavior
+#    should be may be required should this happen.  The introductory comments
+#    have more to say about this.
+
+sub standardize { return main::standardize($_[0]); }
+sub trace { return main::trace(@_); }
+
+
+{ # Closure
+
+    main::setup_package();
+
+    my %leader;
+    # The leader table of this one; initially $self.
+    main::set_access('leader', \%leader, 'r');
+
+    my %equivalents;
+    # An array of any tables that have this one as their leader
+    main::set_access('equivalents', \%equivalents, 'readable_array');
+
+    my %parent;
+    # The parent table to this one, initially $self.  This allows us to
+    # distinguish between equivalent tables that are related, and those which
+    # may not be, but share the same output file because they match the exact
+    # same set of code points in the current Unicode release.
+    main::set_access('parent', \%parent, 'r');
+
+    my %children;
+    # An array of any tables that have this one as their parent
+    main::set_access('children', \%children, 'readable_array');
+
+    my %conflicting;
+    # Array of any tables that would have the same name as this one with
+    # a different meaning.  This is used for the generated documentation.
+    main::set_access('conflicting', \%conflicting, 'readable_array');
+
+    my %matches_all;
+    # Set in the constructor for tables that are expected to match all code
+    # points.
+    main::set_access('matches_all', \%matches_all, 'r');
+
+    sub new {
+        my $class = shift;
+
+        my %args = @_;
+
+        # The property for which this table is a listing of property values.
+        my $property = delete $args{'_Property'};
+
+        # Optional
+        my $initialize = delete $args{'Initialize'};
+        my $matches_all = delete $args{'Matches_All'} || 0;
+        # Rest of parameters passed on.
+
+        my $range_list = Range_List->new(Initialize => $initialize,
+                                         Owner => $property);
+
+        my $self = $class->SUPER::new(%args,
+                                      _Property => $property,
+                                      _Range_List => $range_list,
+                                      );
+        my $addr = main::objaddr $self;
+
+        $conflicting{$addr} = [ ];
+        $equivalents{$addr} = [ ];
+        $children{$addr} = [ ];
+        $matches_all{$addr} = $matches_all;
+        $leader{$addr} = $self;
+        $parent{$addr} = $self;
+
+        return $self;
+    }
+
+    # See this program's beginning comment block about overloading these.
+    use overload
+        fallback => 0,
+        qw("") => "_operator_stringify",
+        '=' => sub {
+                    my $self = shift;
+
+                    return if $self->carp_if_locked;
+                    return $self;
+                },
+
+        '+' => sub {
+                        my $self = shift;
+                        my $other = shift;
+
+                        return $self->_range_list + $other;
+                    },
+        '&' => sub {
+                        my $self = shift;
+                        my $other = shift;
+
+                        return $self->_range_list & $other;
+                    },
+        '+=' => sub {
+                        my $self = shift;
+                        my $other = shift;
+
+                        return if $self->carp_if_locked;
+
+                        my $addr = main::objaddr $self;
+
+                        if (ref $other) {
+
+                            # Change the range list of this table to be the
+                            # union of the two.
+                            $self->_set_range_list($self->_range_list
+                                                    + $other);
+                        }
+                        else {    # $other is just a simple value
+                            $self->add_range($other, $other);
+                        }
+                        return $self;
+                    },
+        '-' => sub { my $self = shift;
+                    my $other = shift;
+                    my $reversed = shift;
+
+                    if ($reversed) {
+                        Carp::my_carp_bug("Can't cope with a "
+                            .  __PACKAGE__
+                            . " being the first parameter in a '-'.  Subtraction ignored.");
+                        return;
+                    }
+
+                    return $self->_range_list - $other;
+                },
+        '~' => sub { my $self = shift;
+                    return ~ $self->_range_list;
+                },
+    ;
+
+    sub _operator_stringify {
+        my $self = shift;
+
+        my $name= $self->complete_name;
+        return "Table '$name'";
+    }
+
+    sub add_alias {
+        # Add a synonym for this table.  See the comments in the base class
+
+        my $self = shift;
+        my $name = shift;
+        # Rest of parameters passed on.
+
+        $self->SUPER::add_alias($name, $self, @_);
+        return;
+    }
+
+    sub add_conflicting {
+        # Add the name of some other object to the list of ones that name
+        # clash with this match table.
+
+        my $self = shift;
+        my $conflicting_name = shift;   # The name of the conflicting object
+        my $p = shift || 'p';           # Optional, is this a \p{} or \P{} ?
+        my $conflicting_object = shift; # Optional, the conflicting object
+                                        # itself.  This is used to
+                                        # disambiguate the text if the input
+                                        # name is identical to any of the
+                                        # aliases $self is known by.
+                                        # Sometimes the conflicting object is
+                                        # merely hypothetical, so this has to
+                                        # be an optional parameter.
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        # Check if the conflicting name is exactly the same as any existing
+        # alias in this table (as long as there is a real object there to
+        # disambiguate with).
+        if (defined $conflicting_object) {
+            foreach my $alias ($self->aliases) {
+                if ($alias->name eq $conflicting_name) {
+
+                    # Here, there is an exact match.  This results in
+                    # ambiguous comments, so disambiguate by changing the
+                    # conflicting name to its object's complete equivalent.
+                    $conflicting_name = $conflicting_object->complete_name;
+                    last;
+                }
+            }
+        }
+
+        # Convert to the \p{...} final name
+        $conflicting_name = "\\$p" . "{$conflicting_name}";
+
+        # Only add once
+        return if grep { $conflicting_name eq $_ } @{$conflicting{$addr}};
+
+        push @{$conflicting{$addr}}, $conflicting_name;
+
+        return;
+    }
+
+    sub is_equivalent_to {
+        # Return boolean of whether or not the other object is a table of this
+        # type and has been marked equivalent to this one.
+
+        my $self = shift;
+        my $other = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        return 0 if ! defined $other; # Can happen for incomplete early
+                                      # releases
+        unless ($other->isa(__PACKAGE__)) {
+            my $ref_other = ref $other;
+            my $ref_self = ref $self;
+            Carp::my_carp_bug("Argument to 'is_equivalent_to' must be another $ref_self, not a '$ref_other'.  $other not set equivalent to $self.");
+            return 0;
+        }
+
+        # Two tables are equivalent if they have the same leader.
+        return $leader{main::objaddr $self}
+                == $leader{main::objaddr $other};
+        return;
+    }
+
+    sub matches_identically_to {
+        # Return a boolean as to whether or not two tables match identical
+        # sets of code points.
+
+        my $self = shift;
+        my $other = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        unless ($other->isa(__PACKAGE__)) {
+            my $ref_other = ref $other;
+            my $ref_self = ref $self;
+            Carp::my_carp_bug("Argument to 'matches_identically_to' must be another $ref_self, not a '$ref_other'.  $other not set equivalent to $self.");
+            return 0;
+        }
+
+        # These are ordered in increasing real time to figure out (at least
+        # until a patch changes that and doesn't change this)
+        return 0 if $self->max != $other->max;
+        return 0 if $self->min != $other->min;
+        return 0 if $self->range_count != $other->range_count;
+        return 0 if $self->count != $other->count;
+
+        # Here they could be identical because all the tests above passed.
+        # The loop below is somewhat simpler since we know they have the same
+        # number of elements.  Compare range by range, until reach the end or
+        # find something that differs.
+        my @a_ranges = $self->_range_list->ranges;
+        my @b_ranges = $other->_range_list->ranges;
+        for my $i (0 .. @a_ranges - 1) {
+            my $a = $a_ranges[$i];
+            my $b = $b_ranges[$i];
+            trace "self $a; other $b" if main::DEBUG && $to_trace;
+            return 0 if $a->start != $b->start || $a->end != $b->end;
+        }
+        return 1;
+    }
+
+    sub set_equivalent_to {
+        # Set $self equivalent to the parameter table.
+        # The required Related => 'x' parameter is a boolean indicating
+        # whether these tables are related or not.  If related, $other becomes
+        # the 'parent' of $self; if unrelated it becomes the 'leader'
+        #
+        # Related tables share all characteristics except names; equivalents
+        # not quite so many.
+        # If they are related, one must be a perl extension.  This is because
+        # we can't guarantee that Unicode won't change one or the other in a
+        # later release even if they are idential now.
+
+        my $self = shift;
+        my $other = shift;
+
+        my %args = @_;
+        my $related = delete $args{'Related'};
+
+        Carp::carp_extra_args(\%args) if main::DEBUG && %args;
+
+        return if ! defined $other;     # Keep on going; happens in some early
+                                        # Unicode releases.
+
+        if (! defined $related) {
+            Carp::my_carp_bug("set_equivalent_to must have 'Related => [01] parameter.  Assuming $self is not related to $other");
+            $related = 0;
+        }
+
+        # If already are equivalent, no need to re-do it;  if subroutine
+        # returns null, it found an error, also do nothing
+        my $are_equivalent = $self->is_equivalent_to($other);
+        return if ! defined $are_equivalent || $are_equivalent;
+
+        my $current_leader = ($related)
+                             ? $parent{main::objaddr $self}
+                             : $leader{main::objaddr $self};
+
+        if ($related &&
+            ! $other->perl_extension
+            && ! $current_leader->perl_extension)
+        {
+            Carp::my_carp_bug("set_equivalent_to should have 'Related => 0 for equivalencing two Unicode properties.  Assuming $self is not related to $other");
+            $related = 0;
+        }
+
+        my $leader = main::objaddr $current_leader;
+        my $other_addr = main::objaddr $other;
+
+        # Any tables that are equivalent to or children of this table must now
+        # instead be equivalent to or (children) to the new leader (parent),
+        # still equivalent.  The equivalency includes their matches_all info,
+        # and for related tables, their status
+        # All related tables are of necessity equivalent, but the converse
+        # isn't necessarily true
+        my $status = $other->status;
+        my $status_info = $other->status_info;
+        my $matches_all = $matches_all{other_addr};
+        foreach my $table ($current_leader, @{$equivalents{$leader}}) {
+            next if $table == $other;
+            trace "setting $other to be the leader of $table, status=$status" if main::DEBUG && $to_trace;
+
+            my $table_addr = main::objaddr $table;
+            $leader{$table_addr} = $other;
+            $matches_all{$table_addr} = $matches_all;
+            $self->_set_range_list($other->_range_list);
+            push @{$equivalents{$other_addr}}, $table;
+            if ($related) {
+                $parent{$table_addr} = $other;
+                push @{$children{$other_addr}}, $table;
+                $table->set_status($status, $status_info);
+            }
+        }
+
+        # Now that we've declared these to be equivalent, any changes to one
+        # of the tables would invalidate that equivalency.
+        $self->lock;
+        $other->lock;
+        return;
+    }
+
+    sub add_range { # Add a range to the list for this table.
+        my $self = shift;
+        # Rest of parameters passed on
+
+        return if $self->carp_if_locked;
+        return $self->_range_list->add_range(@_);
+    }
+
+    sub complete_name {
+        # The complete name for a match table includes it's property in a
+        # compound form 'property=table', except if the property is the
+        # pseudo-property, perl, in which case it is just the single form,
+        # 'table'
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $name = $self->full_name;
+        my $property = $self->property;
+        $name = '""' if $name eq "";  # A null name shouldn't happen, but this
+                                      # helps debug if it does
+        return $name if $property == $perl;
+
+        # (If change the '=' must also change the ':' in set_final_comment(),
+        # and the references to colon in its text)
+        return $property->full_name . '=' . $name;
+    }
+
+    sub pre_body {  # Does nothing for match tables.
+        return
+    }
+
+    sub append_to_body {  # Does nothing for match tables.
+        return
+    }
+
+    sub write {
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        return $self->SUPER::write(2); # 2 tab stops
+    }
+
+    sub set_final_comment {
+        # This creates a comment for the file that is to hold the match table
+        # $self.  It is somewhat convoluted to make the English read nicely,
+        # but, heh, it's just a comment.
+        # This should be called only with the leader match table of all the
+        # ones that share the same file.  It lists all such tables, ordered so
+        # that related ones are together.
+
+        my $leader = shift;   # Should only be called on the leader table of
+                              # an equivalent group
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $leader;
+
+        if ($leader{$addr} != $leader) {
+            Carp::my_carp_bug(<<END
+set_final_comment() must be called on a leader table, which $leader is not.
+It is equivalent to $leader{$addr}.  No comment created
+END
+            );
+            return;
+        }
+
+        # Get the number of code points matched by each of the tables in this
+        # file, and add underscores for clarity.
+        my $count = $leader->count;
+        my $string_count = main::clarify_number($count);
+
+        my $loose_count = 0;        # how many aliases loosely matched
+        my $compound_name = "";     # ? Are any names compound?, and if so, an
+                                    # example
+        my $properties_with_compound_names = 0;    # count of these
+
+
+        my %flags;              # The status flags used in the file
+        my $total_entries = 0;  # number of entries written in the comment
+        my $matches_comment = ""; # The portion of the comment about the
+                                  # \p{}'s
+        my @global_comments;    # List of all the tables' comments that are
+                                # there before this routine was called.
+
+        # Get list of all the parent tables that are equivalent to this one
+        # (including itself).
+        my @parents = grep { $parent{main::objaddr $_} == $_ }
+                            main::uniques($leader, @{$equivalents{$addr}});
+        my $has_unrelated = (@parents >= 2);  # boolean, ? are there unrelated
+                                              # tables
+
+        for my $parent (@parents) {
+
+            my $property = $parent->property;
+
+            # Special case 'N' tables in properties with two match tables when
+            # the other is a 'Y' one.  These are likely to be binary tables,
+            # but not necessarily.  In either case, \P{} will match the
+            # complement of \p{}, and so if something is a synonym of \p, the
+            # complement of that something will be the synonym of \P.  This
+            # would be true of any property with just two match tables, not
+            # just those whose values are Y and N; but that would require a
+            # little extra work, and there are none such so far in Unicode.
+            my $perl_p = 'p';        # which is it?  \p{} or \P{}
+            my @yes_perl_synonyms;   # list of any synonyms for the 'Y' table
+
+            if (scalar $property->tables == 2
+                && $parent == $property->table('N')
+                && defined (my $yes = $property->table('Y')))
+            {
+                my $yes_addr = main::objaddr $yes;
+                @yes_perl_synonyms
+                    = grep { $_->property == $perl }
+                                    main::uniques($yes,
+                                                $parent{$yes_addr},
+                                                $parent{$yes_addr}->children);
+
+                # But these synonyms are \P{} ,not \p{}
+                $perl_p = 'P';
+            }
+
+            my @description;        # Will hold the table description
+            my @note;               # Will hold the table notes.
+            my @conflicting;        # Will hold the table conflicts.
+
+            # Look at the parent, any yes synonyms, and all the children
+            for my $table ($parent,
+                           @yes_perl_synonyms,
+                           @{$children{main::objaddr $parent}})
+            {
+                my $table_addr = main::objaddr $table;
+                my $table_property = $table->property;
+
+                # Tables are separated by a blank line to create a grouping.
+                $matches_comment .= "\n" if $matches_comment;
+
+                # The table is named based on the property and value
+                # combination it is for, like script=greek.  But there may be
+                # a number of synonyms for each side, like 'sc' for 'script',
+                # and 'grek' for 'greek'.  Any combination of these is a valid
+                # name for this table.  In this case, there are three more,
+                # 'sc=grek', 'sc=greek', and 'script='grek'.  Rather than
+                # listing all possible combinations in the comment, we make
+                # sure that each synonym occurs at least once, and add
+                # commentary that the other combinations are possible.
+                my @property_aliases = $table_property->aliases;
+                my @table_aliases = $table->aliases;
+
+                Carp::my_carp_bug("$table doesn't have any names.  Proceeding anyway.") unless @table_aliases;
+
+                # The alias lists above are already ordered in the order we
+                # want to output them.  To ensure that each synonym is listed,
+                # we must use the max of the two numbers.
+                my $listed_combos = main::max(scalar @table_aliases,
+                                                scalar @property_aliases);
+                trace "$listed_combos, tables=", scalar @table_aliases, "; names=", scalar @property_aliases if main::DEBUG;
+
+                my $property_had_compound_name = 0;
+
+                for my $i (0 .. $listed_combos - 1) {
+                    $total_entries++;
+
+                    # The current alias for the property is the next one on
+                    # the list, or if beyond the end, start over.  Similarly
+                    # for the table (\p{prop=table})
+                    my $property_alias = $property_aliases
+                                            [$i % @property_aliases]->name;
+                    my $table_alias_object = $table_aliases
+                                                        [$i % @table_aliases];
+                    my $table_alias = $table_alias_object->name;
+                    my $loose_match = $table_alias_object->loose_match;
+
+                    if ($table_alias !~ /\D/) { # Clarify large numbers.
+                        $table_alias = main::clarify_number($table_alias)
+                    }
+
+                    # Add a comment for this alias combination
+                    my $current_match_comment;
+                    if ($table_property == $perl) {
+                        $current_match_comment = "\\$perl_p"
+                                                    . "{$table_alias}";
+                    }
+                    else {
+                        $current_match_comment
+                                        = "\\p{$property_alias=$table_alias}";
+                        $property_had_compound_name = 1;
+                    }
+
+                    # Flag any abnormal status for this table.
+                    my $flag = $property->status
+                                || $table->status
+                                || $table_alias_object->status;
+                    $flags{$flag} = $status_past_participles{$flag} if $flag;
+
+                    $loose_count++;
+
+                    # Pretty up the comment.  Note the \b; it says don't make
+                    # this line a continuation.
+                    $matches_comment .= sprintf("\b%-1s%-s%s\n",
+                                        $flag,
+                                        " " x 7,
+                                        $current_match_comment);
+                } # End of generating the entries for this table.
+
+                # Save these for output after this group of related tables.
+                push @description, $table->description;
+                push @note, $table->note;
+                push @conflicting, $table->conflicting;
+
+                # Compute an alternate compound name using the final property
+                # synonym and the first table synonym with a colon instead of
+                # the equal sign used elsewhere.
+                if ($property_had_compound_name) {
+                    $properties_with_compound_names ++;
+                    if (! $compound_name || @property_aliases > 1) {
+                        $compound_name = $property_aliases[-1]->name
+                                        . ': '
+                                        . $table_aliases[0]->name;
+                    }
+                }
+            } # End of looping through all children of this table
+
+            # Here have assembled in $matches_comment all the related tables
+            # to the current parent (preceded by the same info for all the
+            # previous parents).  Put out information that applies to all of
+            # the current family.
+            if (@conflicting) {
+
+                # But output the conflicting information now, as it applies to
+                # just this table.
+                my $conflicting = join ", ", @conflicting;
+                if ($conflicting) {
+                    $matches_comment .= <<END;
+
+    Note that contrary to what you might expect, the above is NOT the same as
+END
+                    $matches_comment .= "any of: " if @conflicting > 1;
+                    $matches_comment .= "$conflicting\n";
+                }
+            }
+            if (@description) {
+                $matches_comment .= "\n    Meaning: "
+                                    . join('; ', @description)
+                                    . "\n";
+            }
+            if (@note) {
+                $matches_comment .= "\n    Note: "
+                                    . join("\n    ", @note)
+                                    . "\n";
+            }
+        } # End of looping through all tables
+
+
+        my $code_points;
+        my $match;
+        my $any_of_these;
+        if ($count == 1) {
+            $match = 'matches';
+            $code_points = 'single code point';
+        }
+        else {
+            $match = 'match';
+            $code_points = "$string_count code points";
+        }
+
+        my $synonyms;
+        my $entries;
+        if ($total_entries <= 1) {
+            $synonyms = "";
+            $entries = 'entry';
+            $any_of_these = 'this'
+        }
+        else {
+            $synonyms = " any of the following regular expression constructs";
+            $entries = 'entries';
+            $any_of_these = 'any of these'
+        }
+
+        my $comment = "";
+        if ($has_unrelated) {
+            $comment .= <<END;
+This file is for tables that are not necessarily related:  To conserve
+resources, every table that matches the identical set of code points in this
+version of Unicode uses this file.  Each one is listed in a separate group
+below.  It could be that the tables will match the same set of code points in
+other Unicode releases, or it could be purely coincidence that they happen to
+be the same in Unicode $string_version, and hence may not in other versions.
+
+END
+        }
+
+        if (%flags) {
+            foreach my $flag (sort keys %flags) {
+                $comment .= <<END;
+'$flag' below means that this form is $flags{$flag}.  Consult $pod_file.pod
+END
+            }
+            $comment .= "\n";
+        }
+
+        $comment .= <<END;
+This file returns the $code_points in Unicode Version $string_version that
+$match$synonyms:
+
+$matches_comment
+$pod_file.pod should be consulted for the rules on using $any_of_these,
+including if adding or subtracting white space, underscore, and hyphen
+characters matters or doesn't matter, and other permissible syntactic
+variants.  Upper/lower case distinctions never matter.
+END
+
+        if ($compound_name) {
+            $comment .= <<END;
+
+A colon can be substituted for the equals sign, and
+END
+            if ($properties_with_compound_names > 1) {
+                $comment .= <<END;
+within each group above,
+END
+            }
+            $compound_name = sprintf("%-8s\\p{%s}", " ", $compound_name);
+
+            # Note the \b below, it says don't make that line a continuation.
+            $comment .= <<END;
+anything to the left of the equals (or colon) can be combined with anything to
+the right.  Thus, for example,
+$compound_name
+\bis also valid.
+END
+        }
+
+        # And append any comment(s) from the actual tables.  They are all
+        # gathered here, so may not read all that well.
+        $comment .= "\n" . join "\n\n", @global_comments if @global_comments;
+
+        if ($count) {   # The format differs if no code points, and needs no
+                        # explanation in that case
+                $comment.= <<END;
+
+The format of the lines of this file is:
+END
+            $comment.= <<END;
+START\\tSTOP\\twhere START is the starting code point of the range, in hex;
+STOP is the ending point, or if omitted, the range has just one code point.
+END
+            if ($output_range_counts) {
+                $comment .= <<END;
+Numbers in comments in [brackets] indicate how many code points are in the
+range.
+END
+            }
+        }
+
+        $leader->set_comment(main::join_lines($comment));
+        return;
+    }
+
+    # Accessors for the underlying list
+    for my $sub qw(
+                    get_valid_code_point
+                    get_invalid_code_point
+                )
     {
-        ## Need to ensure that all decomposition characters do not have
-        ## a %HexCodeToComb in %AboveCombClasses.
-        my $want = 1;
-        for my $deco_hexcode (split / /, $CodeToDeco{$code})
+        no strict "refs";
+        *$sub = sub {
+            use strict "refs";
+            my $self = shift;
+
+            return $self->_range_list->$sub(@_);
+        }
+    }
+} # End closure for Match_Table
+
+package Property;
+
+# The Property class represents a Unicode property, or the $perl
+# pseudo-property.  It contains a map table initialized empty at construction
+# time, and for properties accessible through regular expressions, various
+# match tables, created through the add_match_table() method, and referenced
+# by the table('NAME') or tables() methods, the latter returning a list of all
+# of the match tables.  Otherwise table operations implicitly are for the map
+# table.
+#
+# Most of the data in the property is actually about its map table, so it
+# mostly just uses that table's accessors for most methods.  The two could
+# have been combined into one object, but for clarity because of their
+# differing semantics, they have been kept separate.  It could be argued that
+# the 'file' and 'directory' fields should be kept with the map table.
+#
+# Each property has a type.  This can be set in the constructor, or in the
+# set_type accessor, but mostly it is figured out by the data.  Every property
+# starts with unknown type, overridden by a parameter to the constructor, or
+# as match tables are added, or ranges added to the map table, the data is
+# inspected, and the type changed.  After the table is mostly or entirely
+# filled, compute_type() should be called to finalize they analysis.
+#
+# There are very few operations defined.  One can safely remove a range from
+# the map table, and property_add_or_replace_non_nulls() adds the maps from another
+# table to this one, replacing any in the intersection of the two.
+
+sub standardize { return main::standardize($_[0]); }
+sub trace { return main::trace(@_) if main::DEBUG && $to_trace }
+
+{   # Closure
+
+    # This hash will contain as keys, all the aliases of all properties, and
+    # as values, pointers to their respective property objects.  This allows
+    # quick look-up of a property from any of its names.
+    my %alias_to_property_of;
+
+    sub dump_alias_to_property_of {
+        # For debugging
+
+        print "\n", main::simple_dumper (\%alias_to_property_of), "\n";
+        return;
+    }
+
+    sub property_ref {
+        # This is a package subroutine, not called as a method.
+        # If the single parameter is a literal '*' it returns a list of all
+        # defined properties.
+        # Otherwise, the single parameter is a name, and it returns a pointer
+        # to the corresponding property object, or undef if none.
+        #
+        # Properties can have several different names.  The 'standard' form of
+        # each of them is stored in %alias_to_property_of as they are defined.
+        # But it's possible that this subroutine will be called with some
+        # variant, so if the initial lookup fails, it is repeated with the
+        # standarized form of the input name.  If found, besides returning the
+        # result, the input name is added to the list so future calls won't
+        # have to do the conversion again.
+
+        my $name = shift;
+
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        if (! defined $name) {
+            Carp::my_carp_bug("Undefined input property.  No action taken.");
+            return;
+        }
+
+        return main::uniques(values %alias_to_property_of) if $name eq '*';
+
+        # Return cached result if have it.
+        my $result = $alias_to_property_of{$name};
+        return $result if defined $result;
+
+        # Convert the input to standard form.
+        my $standard_name = standardize($name);
+
+        $result = $alias_to_property_of{$standard_name};
+        return unless defined $result;        # Don't cache undefs
+
+        # Cache the result before returning it.
+        $alias_to_property_of{$name} = $result;
+        return $result;
+    }
+
+
+    main::setup_package();
+
+    my %map;
+    # A pointer to the map table object for this property
+    main::set_access('map', \%map);
+
+    my %full_name;
+    # The property's full name.  This is a duplicate of the copy kept in the
+    # map table, but is needed because stringify needs it during
+    # construction of the map table, and then would have a chicken before egg
+    # problem.
+    main::set_access('full_name', \%full_name, 'r');
+
+    my %table_ref;
+    # This hash will contain as keys, all the aliases of any match tables
+    # attached to this property, and as values, the pointers to their
+    # respective tables.  This allows quick look-up of a table from any of its
+    # names.
+    main::set_access('table_ref', \%table_ref);
+
+    my %type;
+    # The type of the property, $ENUM, $BINARY, etc
+    main::set_access('type', \%type, 'r');
+
+    my %file;
+    # The filename where the map table will go (if actually written).
+    # Normally defaulted, but can be overridden.
+    main::set_access('file', \%file, 'r', 's');
+
+    my %directory;
+    # The directory where the map table will go (if actually written).
+    # Normally defaulted, but can be overridden.
+    main::set_access('directory', \%directory, 's');
+
+    my %pseudo_map_type;
+    # This is used to affect the calculation of the map types for all the
+    # ranges in the table.  It should be set to one of the values that signify
+    # to alter the calculation.
+    main::set_access('pseudo_map_type', \%pseudo_map_type, 'r');
+
+    my %has_only_code_point_maps;
+    # A boolean used to help in computing the type of data in the map table.
+    main::set_access('has_only_code_point_maps', \%has_only_code_point_maps);
+
+    my %unique_maps;
+    # A list of the first few distinct mappings this property has.  This is
+    # used to disambiguate between binary and enum property types, so don't
+    # have to keep more than three.
+    main::set_access('unique_maps', \%unique_maps);
+
+    sub new {
+        # The only required parameter is the positionally first, name.  All
+        # other parameters are key => value pairs.  See the documentation just
+        # above for the meanings of the ones not passed directly on to the map
+        # table constructor.
+
+        my $class = shift;
+        my $name = shift || "";
+
+        my $self = property_ref($name);
+        if (defined $self) {
+            my $options_string = join ", ", @_;
+            $options_string = ".  Ignoring options $options_string" if $options_string;
+            Carp::my_carp("$self is already in use.  Using existing one$options_string;");
+            return $self;
+        }
+
+        my %args = @_;
+
+        $self = bless \do { my $anonymous_scalar }, $class;
+        my $addr = main::objaddr $self;
+
+        $directory{$addr} = delete $args{'Directory'};
+        $file{$addr} = delete $args{'File'};
+        $full_name{$addr} = delete $args{'Full_Name'} || $name;
+        $type{$addr} = delete $args{'Type'} || $UNKNOWN;
+        $pseudo_map_type{$addr} = delete $args{'Map_Type'};
+        # Rest of parameters passed on.
+
+        $has_only_code_point_maps{$addr} = 1;
+        $table_ref{$addr} = { };
+        $unique_maps{$addr} = { };
+
+        $map{$addr} = Map_Table->new($name,
+                                    Full_Name => $full_name{$addr},
+                                    _Alias_Hash => \%alias_to_property_of,
+                                    _Property => $self,
+                                    %args);
+        return $self;
+    }
+
+    # See this program's beginning comment block about overloading the copy
+    # constructor.  Few operations are defined on properties, but a couple are
+    # useful.  It is safe to take the inverse of a property, and to remove a
+    # single code point from it.
+    use overload
+        fallback => 0,
+        qw("") => "_operator_stringify",
+        "." => \&main::_operator_dot,
+        '==' => \&main::_operator_equal,
+        '!=' => \&main::_operator_not_equal,
+        '=' => sub { return shift },
+        '-=' => "_minus_and_equal",
+    ;
+
+    sub _operator_stringify {
+        return "Property '" .  shift->full_name . "'";
+    }
+
+    sub _minus_and_equal {
+        # Remove a single code point from the map table of a property.
+
+        my $self = shift;
+        my $other = shift;
+        my $reversed = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        if (ref $other) {
+            Carp::my_carp_bug("Can't cope with a "
+                        . ref($other)
+                        . " argument to '-='.  Subtraction ignored.");
+            return $self;
+        }
+        elsif ($reversed) {   # Shouldnt happen in a -=, but just in case
+            Carp::my_carp_bug("Can't cope with a "
+            .  __PACKAGE__
+            . " being the first parameter in a '-='.  Subtraction ignored.");
+            return $self;
+        }
+        else {
+            $map{main::objaddr $self}->delete_range($other, $other);
+        }
+        return $self;
+    }
+
+    sub add_match_table {
+        # Add a new match table for this property, with name given by the
+        # parameter.  It returns a pointer to the table.
+
+        my $self = shift;
+        my $name = shift;
+        my %args = @_;
+
+        my $addr = main::objaddr $self;
+
+        my $table = $table_ref{$addr}{$name};
+        my $standard_name = main::standardize($name);
+        if (defined $table
+            || (defined ($table = $table_ref{$addr}{$standard_name})))
         {
-            if (exists $_Above_HexCodes{$deco_hexcode}) {
-                ## one of the decmposition chars has an ABOVE combination
-                ## class, so we're not interested in this one
-                $want = 0;
-                last;
+            Carp::my_carp("Table '$name' in $self is already in use.  Using existing one");
+            $table_ref{$addr}{$name} = $table;
+            return $table;
+        }
+        else {
+
+            # See if this is a perl extension, if not passed in.
+            my $perl_extension = delete $args{'Perl_Extension'};
+            $perl_extension
+                        = $self->perl_extension if ! defined $perl_extension;
+
+            $table = Match_Table->new(
+                                Name => $name,
+                                Perl_Extension => $perl_extension,
+                                _Alias_Hash => $table_ref{$addr},
+                                _Property => $self,
+
+                                # gets property's status by default
+                                Status => $self->status,
+                                _Status_Info => $self->status_info,
+                                %args,
+                                Internal_Only_Warning => 1); # Override any
+                                                             # input param
+            return unless defined $table;
+        }
+
+        # Save the names for quick look up
+        $table_ref{$addr}{$standard_name} = $table;
+        $table_ref{$addr}{$name} = $table;
+
+        # Perhaps we can figure out the type of this property based on the
+        # fact of adding this match table.  First, string properties don't
+        # have match tables; second, a binary property can't have 3 match
+        # tables
+        if ($type{$addr} == $UNKNOWN) {
+            $type{$addr} = $NON_STRING;
+        }
+        elsif ($type{$addr} == $STRING) {
+            Carp::my_carp("$self Added a match table '$name' to a string property '$self'.  Changed it to a non-string property.  Bad News.");
+            $type{$addr} = $NON_STRING;
+        }
+        elsif ($type{$addr} != $ENUM) {
+            if (scalar main::uniques(values %{$table_ref{$addr}}) > 2
+                && $type{$addr} == $BINARY)
+            {
+                Carp::my_carp("$self now has more than 2 tables (with the addition of '$name'), and so is no longer binary.  Changing its type to 'enum'.  Bad News.");
+                $type{$addr} = $ENUM;
             }
         }
-        if ($want) {
-            $CanonCDIJ->Append($code);
+
+        return $table;
+    }
+
+    sub table {
+        # Return a pointer to the match table (with name given by the
+        # parameter) associated with this property; undef if none.
+
+        my $self = shift;
+        my $name = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        return $table_ref{$addr}{$name} if defined $table_ref{$addr}{$name};
+
+        # If quick look-up failed, try again using the standard form of the
+        # input name.  If that succeeds, cache the result before returning so
+        # won't have to standardize this input name again.
+        my $standard_name = main::standardize($name);
+        return unless defined $table_ref{$addr}{$standard_name};
+
+        $table_ref{$addr}{$name} = $table_ref{$addr}{$standard_name};
+        return $table_ref{$addr}{$name};
+    }
+
+    sub tables {
+        # Return a list of pointers to all the match tables attached to this
+        # property
+
+        return main::uniques(values %{$table_ref{main::objaddr shift}});
+    }
+
+    sub directory {
+        # Returns the directory the map table for this property should be
+        # output in.  If a specific directory has been specified, that has
+        # priority;  'undef' is returned if the type isn't defined;
+        # or $map_directory for everything else.
+
+        my $addr = main::objaddr shift;
+
+        return $directory{$addr} if defined $directory{$addr};
+        return undef if $type{$addr} == $UNKNOWN;
+        return $map_directory;
+    }
+
+    sub swash_name {
+        # Return the name that is used to both:
+        #   1)  Name the file that the map table is written to.
+        #   2)  The name of swash related stuff inside that file.
+        # The reason for this is that the Perl core historically has used
+        # certain names that aren't the same as the Unicode property names.
+        # To continue using these, $file is hard-coded in this file for those,
+        # but otherwise the standard name is used.  This is different from the
+        # external_name, so that the rest of the files, like in lib can use
+        # the standard name always, without regard to historical precedent.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr $self;
+
+        return $file{$addr} if defined $file{$addr};
+        return $map{$addr}->external_name;
+    }
+
+    sub to_create_match_tables {
+        # Returns a boolean as to whether or not match tables should be
+        # created for this property.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        # The whole point of this pseudo property is match tables.
+        return 1 if $self == $perl;
+
+        my $addr = main::objaddr $self;
+
+        # Don't generate tables of code points that match the property values
+        # of a string property.  Such a list would most likely have many
+        # property values, each with just one or very few code points mapping
+        # to it.
+        return 0 if $type{$addr} == $STRING;
+
+        # Don't generate anything for unimplemented properties.
+        return 0 if grep { $self->complete_name eq $_ }
+                                                    @unimplemented_properties;
+        # Otherwise, do.
+        return 1;
+    }
+
+    sub property_add_or_replace_non_nulls {
+        # This adds the mappings in the property $other to $self.  Non-null
+        # mappings from $other override those in $self.  It essentially merges
+        # the two properties, with the second having priority except for null
+        # mappings.
+
+        my $self = shift;
+        my $other = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        if (! $other->isa(__PACKAGE__)) {
+            Carp::my_carp_bug("$other should be a "
+                            . __PACKAGE__
+                            . ".  Not a '"
+                            . ref($other)
+                            . "'.  Not added;");
+            return;
+        }
+
+        return $map{main::objaddr $self}->
+                map_add_or_replace_non_nulls($map{main::objaddr $other});
+    }
+
+    sub set_type {
+        # Set the type of the property.  Mostly this is figured out by the
+        # data in the table.  But this is used to set it explicitly.  The
+        # reason it is not a standard accessor is that when setting a binary
+        # property, we need to make sure that all the true/false aliases are
+        # present, as they were omitted in early Unicode releases.
+
+        my $self = shift;
+        my $type = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        if ($type != $ENUM && $type != $BINARY && $type != $STRING) {
+            Carp::my_carp("Unrecognized type '$type'.  Type not set");
+            return;
+        }
+
+        $type{main::objaddr $self} = $type;
+        return if $type != $BINARY;
+
+        my $yes = $self->table('Y');
+        $yes = $self->table('Yes') if ! defined $yes;
+        $yes = $self->add_match_table('Y') if ! defined $yes;
+        $yes->add_alias('Yes');
+        $yes->add_alias('T');
+        $yes->add_alias('True');
+
+        my $no = $self->table('N');
+        $no = $self->table('No') if ! defined $no;
+        $no = $self->add_match_table('N') if ! defined $no;
+        $no->add_alias('No');
+        $no->add_alias('F');
+        $no->add_alias('False');
+        return;
+    }
+
+    sub add_map {
+        # Add a map to the property's map table.  This also keeps
+        # track of the maps so that the property type can be determined from
+        # its data.
+
+        my $self = shift;
+        my $start = shift;  # First code point in range
+        my $end = shift;    # Final code point in range
+        my $map = shift;    # What the range maps to.
+        # Rest of parameters passed on.
+
+        my $addr = main::objaddr $self;
+
+        # If haven't the type of the property, gather information to figure it
+        # out.
+        if ($type{$addr} == $UNKNOWN) {
+
+            # If the map contains an interior blank or dash, or most other
+            # nonword characters, it will be a string property.  This
+            # heuristic may actually miss some string properties.  If so, they
+            # may need to have explicit set_types called for them.  This
+            # happens in the Unihan properties.
+            if ($map =~ / (?<= . ) [ -] (?= . ) /x
+                || $map =~ / [^\w.\/\ -]  /x)
+            {
+                $self->set_type($STRING);
+
+                # $unique_maps is used for disambiguating between ENUM and
+                # BINARY later; since we know the property is not going to be
+                # one of those, no point in keeping the data around
+                undef $unique_maps{$addr};
+            }
+            else {
+
+                # Not necessarily a string.  The final decision has to be
+                # deferred until all the data are in.  We keep track of if all
+                # the values are code points for that eventual decision.
+                $has_only_code_point_maps{$addr} &=
+                                            $map =~ / ^ $code_point_re $/x;
+
+                # For the purposes of disambiguating between binary and other
+                # enumerations at the end, we keep track of the first three
+                # distinct property values.  Once we get to three, we know
+                # it's not going to be binary, so no need to track more.
+                if (scalar keys %{$unique_maps{$addr}} < 3) {
+                    $unique_maps{$addr}{main::standardize($map)} = 1;
+                }
+            }
         }
+
+        # Add the mapping by calling our map table's method
+        return $map{$addr}->add_map($start, $end, $map, @_);
     }
 
+    sub compute_type {
+        # Compute the type of the property: $ENUM, $STRING, or $BINARY.  This
+        # should be called after the property is mostly filled with its maps.
+        # We have been keeping track of what the property values have been,
+        # and now have the necessary information to figure out the type.
+
+        my $self = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $addr = main::objaddr($self);
 
+        my $type = $type{$addr};
 
-    ##
-    ## Now dump the files.
-    ##
-    $Name->Write("Name.pl");
+        # If already have figured these out, no need to do so again, but we do
+        # a double check on ENUMS to make sure that a string property hasn't
+        # improperly been classified as an ENUM, so continue on with those.
+        return if $type == $STRING || $type == $BINARY;
 
+        # If every map is to a code point, is a string property.
+        if ($type == $UNKNOWN
+            && ($has_only_code_point_maps{$addr}
+                || (defined $map{$addr}->default_map
+                    && $map{$addr}->default_map eq "")))
+        {
+            $self->set_type($STRING);
+        }
+        else {
+
+            # Otherwise, it is to some sort of enumeration.  (The case where
+            # it is a Unicode miscellaneous property, and treated like a
+            # string in this program is handled in add_map()).  Distinguish
+            # between binary and some other enumeration type.  Of course, if
+            # there are more than two values, it's not binary.  But more
+            # subtle is the test that the default mapping is defined means it
+            # isn't binary.  This in fact may change in the future if Unicode
+            # changes the way its data is structured.  But so far, no binary
+            # properties ever have @missing lines for them, so the default map
+            # isn't defined for them.  The few properties that are two-valued
+            # and aren't considered binary have the default map defined
+            # starting in Unicode 5.0, when the @missing lines appeared; and
+            # this program has special code to put in a default map for them
+            # for earlier than 5.0 releases.
+            if ($type == $ENUM
+                || scalar keys %{$unique_maps{$addr}} > 2
+                || defined $self->default_map)
+            {
+                my $tables = $self->tables;
+                my $count = $self->count;
+                if ($verbosity && $count > 500 && $tables/$count > .1) {
+                    Carp::my_carp_bug("It appears that $self should be a \$STRING property, not an \$ENUM because it has too many match tables: $count\n");
+                }
+                $self->set_type($ENUM);
+            }
+            else {
+                $self->set_type($BINARY);
+            }
+        }
+        undef $unique_maps{$addr};  # Garbage collect
+        return;
+    }
+
+    # Most of the accessors for a property actually apply to its map table.
+    # Setup up accessor functions for those, referring to %map
+    for my $sub qw(
+                    add_alias
+                    add_anomalous_entry
+                    add_comment
+                    add_conflicting
+                    add_description
+                    add_duplicate
+                    add_note
+                    aliases
+                    comment
+                    complete_name
+                    core_access
+                    count
+                    default_map
+                    delete_range
+                    description
+                    each_range
+                    external_name
+                    file_path
+                    format
+                    initialize
+                    inverse_list
+                    is_empty
+                    name
+                    note
+                    perl_extension
+                    property
+                    range_count
+                    ranges
+                    range_size_1
+                    reset_each_range
+                    set_comment
+                    set_core_access
+                    set_default_map
+                    set_file_path
+                    set_final_comment
+                    set_range_size_1
+                    set_status
+                    set_to_output_map
+                    short_name
+                    status
+                    status_info
+                    to_output_map
+                    value_of
+                    write
+                )
+                    # 'property' above is for symmetry, so that one can take
+                    # the property of a property and get itself, and so don't
+                    # have to distinguish between properties and tables in
+                    # calling code
     {
-	my @PVA = $HEADER;
-	foreach my $name (qw (PropertyAlias PA_reverse PropValueAlias
-			      PVA_reverse PVA_abbr_map)) {
-	    # Should I really jump through typeglob hoops just to avoid a
-	    # symbolic reference? (%{"utf8::$name})
-	    push @PVA, "\n", "\%utf8::$name = (\n",
-		simple_dumper (%{$utf8::{$name}}), ");\n";
-	}
-	push @PVA, "1;\n";
-	WriteIfChanged("PVA.pl", @PVA);
+        no strict "refs";
+        *$sub = sub {
+            use strict "refs";
+            my $self = shift;
+            return $map{main::objaddr $self}->$sub(@_);
+        }
     }
 
-    # $Bidi->Write("Bidirectional.pl");
-    for (keys %Bidi) {
-	$Bidi{$_}->Write(
-	    ["lib","bc","$_.pl"],
-	    "BidiClass category '$PropValueAlias{bc}{$_}'"
-	);
+
+} # End closure
+
+package main;
+
+sub join_lines($) {
+    # Returns lines of the input joined together, so that they can be folded
+    # properly.
+    # This causes continuation lines to be joined together into one long line
+    # for folding.  A continuation line is any line that doesn't begin with a
+    # space or "\b" (the latter is stripped from the output).  This is so
+    # lines can be be in a HERE document so as to fit nicely in the terminal
+    # width, but be joined together in one long line, and then folded with
+    # indents, '#' prefixes, etc, properly handled.
+    # A blank separates the joined lines except if there is a break; an extra
+    # blank is inserted after a period ending a line.
+
+    # Intialize the return with the first line.
+    my ($return, @lines) = split "\n", shift;
+
+    # If the first line is null, it was an empty line, add the \n back in
+    $return = "\n" if $return eq "";
+
+    # Now join the remainder of the physical lines.
+    for my $line (@lines) {
+
+        # An empty line means wanted a blank line, so add two \n's to get that
+        # effect, and go to the next line.
+        if (length $line == 0) {
+            $return .= "\n\n";
+            next;
+        }
+
+        # Look at the last character of what we have so far.
+        my $previous_char = substr($return, -1, 1);
+
+        # And at the next char to be output.
+        my $next_char = substr($line, 0, 1);
+
+        if ($previous_char ne "\n") {
+
+            # Here didn't end wth a nl.  If the next char a blank or \b, it
+            # means that here there is a break anyway.  So add a nl to the
+            # output.
+            if ($next_char eq " " || $next_char eq "\b") {
+                $previous_char = "\n";
+                $return .= $previous_char;
+            }
+
+            # Add an extra space after periods.
+            $return .= " " if $previous_char eq '.';
+        }
+
+        # Here $previous_char is still the latest character to be output.  If
+        # it isn't a nl, it means that the next line is to be a continuation
+        # line, with a blank inserted between them.
+        $return .= " " if $previous_char ne "\n";
+
+        # Get rid of any \b
+        substr($line, 0, 1) = "" if $next_char eq "\b";
+
+        # And append this next line.
+        $return .= $line;
     }
 
-    $Comb->Write("CombiningClass.pl");
-    for (keys %{ $PropValueAlias{ccc} }) {
-	my ($code, $name) = @{ $PropValueAlias{ccc}{$_} };
-	(my $c = Table->New())->Append($code);
-	$c->Write(
-	    ["lib","ccc","$_.pl"],
-	    "CombiningClass category '$name'"
-	);
+    return $return;
+}
+
+sub simple_fold($;$$$) {
+    # Returns a string of the input (string or an array of strings) folded
+    # into multiple-lines each of no more than $MAX_LINE_WIDTH characters plus
+    # a \n
+    # This is tailored for the kind of text written by this program,
+    # especially the pod file, which can have very long names with
+    # underscores in the middle, or words like AbcDefgHij....  We allow
+    # breaking in the middle of such constructs if the line won't fit
+    # otherwise.  The break in such cases will come either just after an
+    # underscore, or just before one of the Capital letters.
+
+    local $to_trace = 0 if main::DEBUG;
+
+    my $line = shift;
+    my $prefix = shift;     # Optional string to prepend to each output
+                            # line
+    $prefix = "" unless defined $prefix;
+
+    my $hanging_indent = shift; # Optional number of spaces to indent
+                                # continuation lines
+    $hanging_indent = 0 unless $hanging_indent;
+
+    my $right_margin = shift;   # Optional number of spaces to narrow the
+                                # total width by.
+    $right_margin = 0 unless defined $right_margin;
+
+    # Call carp with the 'nofold' option to avoid it from trying to call us
+    # recursively
+    Carp::carp_extra_args(\@_, 'nofold') if main::DEBUG && @_;
+
+    # The space available doesn't include what's automatically prepended
+    # to each line, or what's reserved on the right.
+    my $max = $MAX_LINE_WIDTH - length($prefix) - $right_margin;
+    # XXX Instead of using the 'nofold' perhaps better to look up the stack
+
+    if (DEBUG && $hanging_indent >= $max) {
+        Carp::my_carp("Too large a hanging indent ($hanging_indent); must be < $max.  Using 0", 'nofold');
+        $hanging_indent = 0;
     }
 
-    $Deco->Write("Decomposition.pl");
-    for (keys %DC) {
-	$DC{$_}->Write(
-	    ["lib","dt","$_.pl"],
-	    "DecompositionType category '$PropValueAlias{dt}{$_}'"
-	);
+    # First, split into the current physical lines.
+    my @line;
+    if (ref $line) {        # Better be an array, because not bothering to
+                            # test
+        foreach my $line (@{$line}) {
+            push @line, split /\n/, $line;
+        }
     }
+    else {
+        @line = split /\n/, $line;
+    }
+
+    #local $to_trace = 1 if main::DEBUG;
+    trace "", join(" ", @line), "\n" if main::DEBUG && $to_trace;
+
+    # Look at each current physical line.
+    for (my $i = 0; $i < @line; $i++) {
+        Carp::my_carp("Tabs don't work well.", 'nofold') if $line[$i] =~ /\t/;
+        #local $to_trace = 1 if main::DEBUG;
+        trace "i=$i: $line[$i]\n" if main::DEBUG && $to_trace;
+
+        # Remove prefix, because will be added back anyway, don't want
+        # doubled prefix
+        $line[$i] =~ s/^$prefix//;
+
+        # Remove trailing space
+        $line[$i] =~ s/\s+\Z//;
+
+        # If the line is too long, fold it.
+        if (length $line[$i] > $max) {
+            my $remainder;
+
+            # Here needs to fold.  Save the leading space in the line for
+            # later.
+            $line[$i] =~ /^ ( \s* )/x;
+            my $leading_space = $1;
+            trace "line length", length $line[$i], "; lead length", length($leading_space) if main::DEBUG && $to_trace;
+
+            # If character at final permissible position is white space,
+            # fold there, which will delete that white space
+            if (substr($line[$i], $max - 1, 1) =~ /\s/) {
+                $remainder = substr($line[$i], $max);
+                $line[$i] = substr($line[$i], 0, $max - 1);
+            }
+            else {
+
+                # Otherwise fold at an acceptable break char closest to
+                # the max length.  Look at just the maximal initial
+                # segment of the line
+                my $segment = substr($line[$i], 0, $max - 1);
+                if ($segment =~
+                    /^ ( .{$hanging_indent}   # Don't look before the
+                                              #  indent.
+                        \ *                   # Don't look in leading
+                                              #  blanks past the indent
+                            [^ ] .*           # Find the right-most
+                        (?:                   #  acceptable break:
+                            [ \s = ]          # space or equal
+                            | - (?! [.0-9] )  # or non-unary minus.
+                        )                     # $1 includes the character
+                    )/x)
+                {
+                    # Split into the initial part that fits, and remaining
+                    # part of the input
+                    $remainder = substr($line[$i], length $1);
+                    $line[$i] = $1;
+                    trace $line[$i] if DEBUG && $to_trace;
+                    trace $remainder if DEBUG && $to_trace;
+                }
+
+                # If didn't find a good breaking spot, see if there is a
+                # not-so-good breaking spot.  These are just after
+                # underscores or where the case changes from lower to
+                # upper.  Use \a as a soft hyphen, but give up
+                # and don't break the line if there is actually a \a
+                # already in the input.  We use an ascii character for the
+                # soft-hyphen to avoid any attempt by miniperl to try to
+                # access the files that this program is creating.
+                elsif ($segment !~ /\a/
+                       && ($segment =~ s/_/_\a/g
+                       || $segment =~ s/ ( [a-z] ) (?= [A-Z] )/$1\a/xg))
+                {
+                    # Here were able to find at least one place to insert
+                    # our substitute soft hyphen.  Find the right-most one
+                    # and replace it by a real hyphen.
+                    trace $segment if DEBUG && $to_trace;
+                    substr($segment,
+                            rindex($segment, "\a"),
+                            1) = '-';
+
+                    # Then remove the soft hyphen substitutes.
+                    $segment =~ s/\a//g;
+                    trace $segment if DEBUG && $to_trace;
+
+                    # And split into the initial part that fits, and
+                    # remainder of the line
+                    my $pos = rindex($segment, '-');
+                    $remainder = substr($line[$i], $pos);
+                    trace $remainder if DEBUG && $to_trace;
+                    $line[$i] = substr($segment, 0, $pos + 1);
+                }
+            }
+
+            # Here we know if we can fold or not.  If we can, $remainder
+            # is what remains to be processed in the next iteration.
+            if (defined $remainder) {
+                trace "folded='$line[$i]'" if main::DEBUG && $to_trace;
+
+                # Insert the folded remainder of the line as a new element
+                # of the array.  (It may still be too long, but we will
+                # deal with that next time through the loop.)  Omit any
+                # leading space in the remainder.
+                $remainder =~ s/^\s+//;
+                trace "remainder='$remainder'" if main::DEBUG && $to_trace;
+
+                # But then indent by whichever is larger of:
+                # 1) the leading space on the input line;
+                # 2) the hanging indent.
+                # This preserves indentation in the original line.
+                my $lead = ($leading_space)
+                            ? length $leading_space
+                            : $hanging_indent;
+                $lead = max($lead, $hanging_indent);
+                splice @line, $i+1, 0, (" " x $lead) . $remainder;
+            }
+        }
+
+        # Ready to output the line. Get rid of any trailing space
+        # And prefix by the required $prefix passed in.
+        $line[$i] =~ s/\s+$//;
+        $line[$i] = "$prefix$line[$i]\n";
+    } # End of looping through all the lines.
+
+    return join "", @line;
+}
+
+sub property_ref {  # Returns a reference to a property object.
+    return Property::property_ref(@_);
+}
 
-    # $Number->Write("Number.pl");
-    for (keys %Number) {
-	$Number{$_}->Write(
-	    ["lib","nt","$_.pl"],
-	    "NumericType category '$PropValueAlias{nt}{$_}'"
-	);
+sub force_unlink ($) {
+    my $filename = shift;
+    return unless file_exists($filename);
+    return if CORE::unlink($filename);
+
+    # We might need write permission
+    chmod 0777, $filename;
+    CORE::unlink($filename) or Carp::my_carp("Couldn't unlink $filename.  Proceeding anyway: $!");
+    return;
+}
+
+sub write ($\@) {
+    # Given a filename and a reference to an array of lines, write the lines
+    # to the file
+    # Filename can be given as an arrayref of directory names
+
+    my $file  = shift;
+    my $lines_ref = shift;
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+    if (! defined $lines_ref) {
+        Carp::my_carp("Missing lines to write parameter for $file.  Writing skipped;");
+        return;
     }
 
-    # $General->Write("Category.pl");
+    # Get into a single string if an array, and get rid of, in Unix terms, any
+    # leading '.'
+    $file= File::Spec->join(@$file) if ref $file eq 'ARRAY';
+    $file = File::Spec->canonpath($file);
+
+    # If has directories, make sure that they all exist
+    (undef, my $directories, undef) = File::Spec->splitpath($file);
+    File::Path::mkpath($directories) if $directories && ! -d $directories;
 
-    for my $to (sort keys %To) {
-        $To{$to}->Write(["To","$to.pl"]);
+    push @files_actually_output, $file;
+
+    my $text;
+    if (@$lines_ref) {
+        $text = join "", @$lines_ref;
+    }
+    else {
+        $text = "";
+        Carp::my_carp("Output file '$file' is empty; writing it anyway;");
     }
 
-    for (keys %{ $PropValueAlias{gc} }) {
-	New_Alias(Is => $PropValueAlias{gc}{$_}, SameAs => $_, Fuzzy => 1);
+    force_unlink ($file);
+
+    my $OUT;
+    if (not open $OUT, ">", $file) {
+        Carp::my_carp("can't open $file for output.  Skipping this file: $!");
+        return;
     }
+    print "$file written.\n" if $verbosity >= $VERBOSE;
+
+    print $OUT $text;
+    close $OUT;
+    return;
 }
 
-##
-## Process LineBreak.txt
-##
-sub LineBreak_Txt()
-{
-    if (not open IN, "LineBreak.txt") {
-        die "$0: LineBreak.txt: $!\n";
+
+sub Standardize($) {
+    # This converts the input name string into a standardized equivalent to
+    # use internally.
+
+    my $name = shift;
+    unless (defined $name) {
+      Carp::my_carp_bug("Standardize() called with undef.  Returning undef.");
+      return;
     }
 
-    my $Lbrk = Table->New();
-    my %Lbrk;
+    # Remove any leading or trailing white space
+    $name =~ s/^\s+//g;
+    $name =~ s/\s+$//g;
 
-    while (<IN>)
-    {
-        next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(\w+)/;
+    # Convert interior white space and hypens into underscores.
+    $name =~ s/ (?<= .) [ -]+ (.) /_$1/xg;
+
+    # Capitalize the letter following an underscore, and convert a sequence of
+    # multiple underscores to a single one
+    $name =~ s/ (?<= .) _+ (.) /_\u$1/xg;
+
+    # And capitalize the first letter, but not for the special cjk ones.
+    $name = ucfirst($name) unless $name =~ /^k[A-Z]/;
+    return $name;
+}
+
+sub standardize ($) {
+    # Returns a lower-cased standardized name, without underscores.  This form
+    # is chosen so that it can distinguish between any real versus superficial
+    # Unicode name differences.  It relies on the fact that Unicode doesn't
+    # have interior underscores, white space, nor dashes in any
+    # stricter-matched name.  It should not be used on Unicode code point
+    # names (the Name property), as they mostly, but not always follow these
+    # rules.
+
+    my $name = Standardize(shift);
+    return if !defined $name;
+
+    $name =~ s/ (?<= .) _ (?= . ) //xg;
+    return lc $name;
+}
+
+{   # Closure
+
+    my $indent_increment = " " x 2;
+    my %already_output;
+
+    $main::simple_dumper_nesting = 0;
+
+    sub simple_dumper {
+        # Like Simple Data::Dumper. Good enough for our needs. We can't use
+        # the real thing as we have to run under miniperl.
+
+        # It is designed so that on input it is at the beginning of a line,
+        # and the final thing output in any call is a trailing ",\n".
+
+        my $item = shift;
+        my $indent = shift;
+        $indent = "" if ! defined $indent;
+
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
 
-	my ($first, $last, $lbrk) = (hex($1), hex($2||""), $3);
+        # nesting level is localized, so that as the call stack pops, it goes
+        # back to the prior value.
+        local $main::simple_dumper_nesting = $main::simple_dumper_nesting;
+        undef %already_output if $main::simple_dumper_nesting == 0;
+        $main::simple_dumper_nesting++;
+        #print STDERR __LINE__, ": $main::simple_dumper_nesting: $indent$item\n";
 
-	$Lbrk->Append($first, $lbrk);
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
 
-        $Lbrk{$lbrk} ||= Table->New();
-        $Lbrk{$lbrk}->Append($first);
+        # Determine the indent for recursive calls.
+        my $next_indent = $indent . $indent_increment;
 
-	if ($last) {
-	    $Lbrk->Extend($last);
-	    $Lbrk{$lbrk}->Extend($last);
-	}
+        my $output;
+        if (! ref $item) {
+
+            # Dump of scalar: just output it in quotes if not a number.  To do
+            # so we must escape certain characters, and therefore need to
+            # operate on a copy to avoid changing the original
+            my $copy = $item;
+            $copy = $UNDEF unless defined $copy;
+
+            # Quote non-numbers (numbers also have optional leading '-' and
+            # fractions)
+            if ($copy eq "" || $copy !~ /^ -? \d+ ( \. \d+ )? $/x) {
+
+                # Escape apostrophe and backslash
+                $copy =~ s/ ( ['\\] ) /\\$1/xg;
+                $copy = "'$copy'";
+            }
+            $output = "$indent$copy,\n";
+        }
+        else {
+
+            # Keep track of cycles in the input, and refuse to infinitely loop
+            if (defined $already_output{main::objaddr $item}) {
+                return "${indent}ALREADY OUTPUT: $item\n";
+            }
+            $already_output{main::objaddr $item} = $item;
+
+            if (ref $item eq 'ARRAY') {
+                my $using_brackets;
+                $output = $indent;
+                if ($main::simple_dumper_nesting > 1) {
+                    $output .= '[';
+                    $using_brackets = 1;
+                }
+                else {
+                    $using_brackets = 0;
+                }
+
+                # If the array is empty, put the closing bracket on the same
+                # line.  Otherwise, recursively add each array element
+                if (@$item == 0) {
+                    $output .= " ";
+                }
+                else {
+                    $output .= "\n";
+                    for (my $i = 0; $i < @$item; $i++) {
+
+                        # Indent array elements one level
+                        $output .= &simple_dumper($item->[$i], $next_indent);
+                        $output =~ s/\n$//;      # Remove trailing nl so as to
+                        $output .= " # [$i]\n";  # add a comment giving the
+                                                 # array index
+                    }
+                    $output .= $indent;     # Indent closing ']' to orig level
+                }
+                $output .= ']' if $using_brackets;
+                $output .= ",\n";
+            }
+            elsif (ref $item eq 'HASH') {
+                my $is_first_line;
+                my $using_braces;
+                my $body_indent;
+
+                # No surrounding braces at top level
+                $output .= $indent;
+                if ($main::simple_dumper_nesting > 1) {
+                    $output .= "{\n";
+                    $is_first_line = 0;
+                    $body_indent = $next_indent;
+                    $next_indent .= $indent_increment;
+                    $using_braces = 1;
+                }
+                else {
+                    $is_first_line = 1;
+                    $body_indent = $indent;
+                    $using_braces = 0;
+                }
+
+                # Output hashes sorted alphabetically instead of apparently
+                # random.  Use caseless alphabetic sort
+                foreach my $key (sort { lc $a cmp lc $b } keys %$item)
+                {
+                    if ($is_first_line) {
+                        $is_first_line = 0;
+                    }
+                    else {
+                        $output .= "$body_indent";
+                    }
+
+                    # The key must be a scalar, but this recursive call quotes
+                    # it
+                    $output .= &simple_dumper($key);
+
+                    # And change the trailing comma and nl to the hash fat
+                    # comma for clarity, and so the value can be on the same
+                    # line
+                    $output =~ s/,\n$/ => /;
+
+                    # Recursively call to get the value's dump.
+                    my $next = &simple_dumper($item->{$key}, $next_indent);
+
+                    # If the value is all on one line, remove its indent, so
+                    # will follow the => immediately.  If it takes more than
+                    # one line, start it on a new line.
+                    if ($next !~ /\n.*\n/) {
+                        $next =~ s/^ *//;
+                    }
+                    else {
+                        $output .= "\n";
+                    }
+                    $output .= $next;
+                }
+
+                $output .= "$indent},\n" if $using_braces;
+            }
+            elsif (ref $item eq 'CODE' || ref $item eq 'GLOB') {
+                $output = $indent . ref($item) . "\n";
+                # XXX see if blessed
+            }
+            elsif ($item->can('dump')) {
+
+                # By convention in this program, objects furnish a 'dump'
+                # method.  Since not doing any output at this level, just pass
+                # on the input indent
+                $output = $item->dump($indent);
+            }
+            else {
+                Carp::my_carp("Can't cope with dumping a " . ref($item) . ".  Skipping.");
+            }
+        }
+        return $output;
     }
-    close IN;
+}
 
-    # $Lbrk->Write("Lbrk.pl");
+sub dump_inside_out {
+    # Dump inside-out hashes in an object's state by converting them to a
+    # regular hash and then calling simple_dumper on that.
 
+    my $object = shift;
+    my $fields_ref = shift;
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
 
-    for (keys %Lbrk) {
-	$Lbrk{$_}->Write(
-	    ["lib","lb","$_.pl"],
-	    "Linebreak category '$PropValueAlias{lb}{$_}'"
-	);
+    my $addr = main::objaddr $object;
+
+    my %hash;
+    foreach my $key (keys %$fields_ref) {
+        $hash{$key} = $fields_ref->{$key}{$addr};
     }
+
+    return simple_dumper(\%hash, @_);
+}
+
+sub _operator_dot {
+    # Overloaded '.' method that is common to all packages.  It uses the
+    # package's stringify method.
+
+    my $self = shift;
+    my $other = shift;
+    my $reversed = shift;
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+    $other = "" unless defined $other;
+
+    foreach my $which (\$self, \$other) {
+        next unless ref $$which;
+        if ($$which->can('_operator_stringify')) {
+            $$which = $$which->_operator_stringify;
+        }
+        else {
+            my $ref = ref $$which;
+            my $addr = main::objaddr $$which;
+            $$which = "$ref ($addr)";
+        }
+    }
+    return ($reversed)
+            ? "$other$self"
+            : "$self$other";
+}
+
+sub _operator_equal {
+    # Generic overloaded '==' routine.  To be equal, they must be the exact
+    # same object
+
+    my $self = shift;
+    my $other = shift;
+
+    return 0 unless defined $other;
+    return 0 unless ref $other;
+    return main::objaddr $self == main::objaddr $other;
+}
+
+sub _operator_not_equal {
+    my $self = shift;
+    my $other = shift;
+
+    return ! _operator_equal($self, $other);
 }
 
-##
-## Process ArabicShaping.txt.
-##
-sub ArabicShaping_txt()
-{
-    if (not open IN, "ArabicShaping.txt") {
-        die "$0: ArabicShaping.txt: $!\n";
+sub process_PropertyAliases($) {
+    # This reads in the PropertyAliases.txt file, which contains almost all
+    # the character properties in Unicode and their equivalent aliases:
+    # scf       ; Simple_Case_Folding         ; sfc
+    #
+    # Field 0 is the preferred short name for the property.
+    # Field 1 is the full name.
+    # Any succeeding ones are other accepted names.
+
+    my $file= shift;
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+    # This whole file was non-existent in early releases, so use our own
+    # internal one.
+    $file->insert_lines(get_old_property_aliases())
+                                                if ! -e 'PropertyAliases.txt';
+
+    # Add any cjk properties that may have been defined.
+    $file->insert_lines(@cjk_properties);
+
+    while ($file->next_line) {
+
+        my @data = split /\s*;\s*/;
+
+        my $full = $data[1];
+
+        my $this = Property->new($data[0], Full_Name => $full);
+
+        # Start looking for more aliases after these two.
+        for my $i (2 .. @data - 1) {
+            $this->add_alias($data[$i]);
+        }
+
     }
+    return;
+}
+
+sub finish_property_setup {
+    # Finishes setting up after PropertyAliases.
 
-    my $ArabLink      = Table->New();
-    my $ArabLinkGroup = Table->New();
+    my $file = shift;
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
 
-    my %JoinType;
+    # This entry was missing from this file in earlier Unicode versions
+    if (-e 'Jamo.txt') {
+        my $jsn = property_ref('JSN');
+        if (! defined $jsn) {
+            $jsn = Property->new('JSN', Full_Name => 'Jamo_Short_Name');
+        }
+    }
 
-    while (<IN>)
+    # This entry is still missing as of 5.2, perhaps because no short name for
+    # it.
+    if (-e 'NameAliases.txt') {
+        my $aliases = property_ref('Name_Alias');
+        if (! defined $aliases) {
+            $aliases = Property->new('Name_Alias');
+        }
+    }
+
+    # These are used so much, that we set globals for them.
+    $gc = property_ref('General_Category');
+    $block = property_ref('Block');
+
+    # Perl adds this alias.
+    $gc->add_alias('Category');
+
+    # For backwards compatibility, these property files have particular names.
+    my $upper = property_ref('Uppercase_Mapping');
+    $upper->set_core_access('uc()');
+    $upper->set_file('Upper'); # This is what utf8.c calls it
+
+    my $lower = property_ref('Lowercase_Mapping');
+    $lower->set_core_access('lc()');
+    $lower->set_file('Lower');
+
+    my $title = property_ref('Titlecase_Mapping');
+    $title->set_core_access('ucfirst()');
+    $title->set_file('Title');
+
+    my $fold = property_ref('Case_Folding');
+    $fold->set_file('Fold') if defined $fold;
+
+    # utf8.c can't currently cope with non range-size-1 for these, and even if
+    # it were changed to do so, someone else may be using them, expecting the
+    # old style
+    foreach my $property (qw {
+                                Case_Folding
+                                Lowercase_Mapping
+                                Titlecase_Mapping
+                                Uppercase_Mapping
+                            })
     {
-	next unless /^[0-9A-Fa-f]+;/;
-	s/\s+$//;
+        property_ref($property)->set_range_size_1(1);
+    }
 
-	my ($hexcode, $name, $link, $linkgroup) = split(/\s*;\s*/);
-        my $code = hex($hexcode);
-	$ArabLink->Append($code, $link);
-	$ArabLinkGroup->Append($code, $linkgroup);
+    # These two properties aren't actually used in the core, but unfortunately
+    # the names just above that are in the core interfere with these, so
+    # choose different names.  These aren't a problem unless the map tables
+    # for these files get written out.
+    my $lowercase = property_ref('Lowercase');
+    $lowercase->set_file('IsLower') if defined $lowercase;
+    my $uppercase = property_ref('Uppercase');
+    $uppercase->set_file('IsUpper') if defined $uppercase;
+
+    # Set up the hard-coded default mappings, but only on properties defined
+    # for this release
+    foreach my $property (keys %default_mapping) {
+        my $property_object = property_ref($property);
+        next if ! defined $property_object;
+        my $default_map = $default_mapping{$property};
+        $property_object->set_default_map($default_map);
+
+        # A map of <code point> implies the property is string.
+        if ($property_object->type == $UNKNOWN
+            && $default_map eq $CODE_POINT)
+        {
+            $property_object->set_type($STRING);
+        }
+    }
 
-        $JoinType{$link} ||= Table->New(Is => "JoinType$link");
-        $JoinType{$link}->Append($code);
+    # The following use the Multi_Default class to create objects for
+    # defaults.
+
+    # Bidi class has a complicated default, but the derived file takes care of
+    # the complications, leaving just 'L'.
+    if (file_exists("${EXTRACTED}DBidiClass.txt")) {
+        property_ref('Bidi_Class')->set_default_map('L');
+    }
+    else {
+        my $default;
+
+        # The derived file was introduced in 3.1.1.  The values below are
+        # taken from table 3-8, TUS 3.0
+        my $default_R =
+            'my $default = Range_List->new;
+             $default->add_range(0x0590, 0x05FF);
+             $default->add_range(0xFB1D, 0xFB4F);'
+        ;
+
+        # The defaults apply only to unassigned characters
+        $default_R .= '$gc->table("Cn") & $default;';
+
+        if ($v_version lt v3.0.0) {
+            $default = Multi_Default->new(R => $default_R, 'L');
+        }
+        else {
+
+            # AL apparently not introduced until 3.0:  TUS 2.x references are
+            # not on-line to check it out
+            my $default_AL =
+                'my $default = Range_List->new;
+                 $default->add_range(0x0600, 0x07BF);
+                 $default->add_range(0xFB50, 0xFDFF);
+                 $default->add_range(0xFE70, 0xFEFF);'
+            ;
+
+            # Non-character code points introduced in this release; aren't AL
+            if ($v_version ge 3.1.0) {
+                $default_AL .= '$default->delete_range(0xFDD0, 0xFDEF);';
+            }
+            $default_AL .= '$gc->table("Cn") & $default';
+            $default = Multi_Default->new(AL => $default_AL,
+                                          R => $default_R,
+                                          'L');
+        }
+        property_ref('Bidi_Class')->set_default_map($default);
     }
-    close IN;
 
-    # $ArabLink->Write("ArabLink.pl");
-    # $ArabLinkGroup->Write("ArabLnkGrp.pl");
+    # Joining type has a complicated default, but the derived file takes care
+    # of the complications, leaving just 'U' (or Non_Joining), except the file
+    # is bad in 3.1.0
+    if (file_exists("${EXTRACTED}DJoinType.txt") || -e 'ArabicShaping.txt') {
+        if (file_exists("${EXTRACTED}DJoinType.txt") && $v_version ne 3.1.0) {
+            property_ref('Joining_Type')->set_default_map('Non_Joining');
+        }
+        else {
+
+            # Otherwise, there are not one, but two possibilities for the
+            # missing defaults: T and U.
+            # The missing defaults that evaluate to T are given by:
+            # T = Mn + Cf - ZWNJ - ZWJ
+            # where Mn and Cf are the general category values. In other words,
+            # any non-spacing mark or any format control character, except
+            # U+200C ZERO WIDTH NON-JOINER (joining type U) and U+200D ZERO
+            # WIDTH JOINER (joining type C).
+            my $default = Multi_Default->new(
+               'T' => '$gc->table("Mn") + $gc->table("Cf") - 0x200C - 0x200D',
+               'Non_Joining');
+            property_ref('Joining_Type')->set_default_map($default);
+        }
+    }
 
+    # Line break has a complicated default in early releases. It is 'Unknown'
+    # for non-assigned code points; 'AL' for assigned.
+    if (file_exists("${EXTRACTED}DLineBreak.txt") || -e 'LineBreak.txt') {
+        my $lb = property_ref('Line_Break');
+        if ($v_version gt 3.2.0) {
+            $lb->set_default_map('Unknown');
+        }
+        else {
+            my $default = Multi_Default->new( 'Unknown' => '$gc->table("Cn")',
+                                              'AL');
+            $lb->set_default_map($default);
+        }
 
-    for (keys %JoinType) {
-	$JoinType{$_}->Write(
-	    ["lib","jt","$_.pl"],
-	    "JoiningType category '$PropValueAlias{jt}{$_}'"
-	);
+        # If has the URS property, make sure that the standard aliases are in
+        # it, since not in the input tables in some versions.
+        my $urs = property_ref('Unicode_Radical_Stroke');
+        if (defined $urs) {
+            $urs->add_alias('cjkRSUnicode');
+            $urs->add_alias('kRSUnicode');
+        }
     }
+    return;
 }
 
-##
-## Process EastAsianWidth.txt.
-##
-sub EastAsianWidth_txt()
-{
-    if (not open IN, "EastAsianWidth.txt") {
-        die "$0: EastAsianWidth.txt: $!\n";
+sub get_old_property_aliases() {
+    # Returns what would be in PropertyAliases.txt if it existed in very old
+    # versions of Unicode.  It was derived from the one in 3.2, and pared
+    # down based on the data that was actually in the older releases.
+    # An attempt was made to use the existence of files to mean inclusion or
+    # not of various aliases, but if this was not sufficient, using version
+    # numbers was resorted to.
+
+    my @return;
+
+    # These are to be used in all versions (though some are constructed by
+    # this program if missing)
+    push @return, split /\n/, <<'END';
+bc        ; Bidi_Class
+Bidi_M    ; Bidi_Mirrored
+cf        ; Case_Folding
+ccc       ; Canonical_Combining_Class
+dm        ; Decomposition_Mapping
+dt        ; Decomposition_Type
+gc        ; General_Category
+isc       ; ISO_Comment
+lc        ; Lowercase_Mapping
+na        ; Name
+na1       ; Unicode_1_Name
+nt        ; Numeric_Type
+nv        ; Numeric_Value
+sfc       ; Simple_Case_Folding
+slc       ; Simple_Lowercase_Mapping
+stc       ; Simple_Titlecase_Mapping
+suc       ; Simple_Uppercase_Mapping
+tc        ; Titlecase_Mapping
+uc        ; Uppercase_Mapping
+END
+
+    if (-e 'Blocks.txt') {
+        push @return, "blk       ; Block\n";
+    }
+    if (-e 'ArabicShaping.txt') {
+        push @return, split /\n/, <<'END';
+jg        ; Joining_Group
+jt        ; Joining_Type
+END
+    }
+    if (-e 'PropList.txt') {
+
+        # This first set is in the original old-style proplist.
+        push @return, split /\n/, <<'END';
+Alpha     ; Alphabetic
+Bidi_C    ; Bidi_Control
+Dash      ; Dash
+Dia       ; Diacritic
+Ext       ; Extender
+Hex       ; Hex_Digit
+Hyphen    ; Hyphen
+IDC       ; ID_Continue
+Ideo      ; Ideographic
+Join_C    ; Join_Control
+Math      ; Math
+QMark     ; Quotation_Mark
+Term      ; Terminal_Punctuation
+WSpace    ; White_Space
+END
+        # The next sets were added later
+        if ($v_version ge v3.0.0) {
+            push @return, split /\n/, <<'END';
+Upper     ; Uppercase
+Lower     ; Lowercase
+END
+        }
+        if ($v_version ge v3.0.1) {
+            push @return, split /\n/, <<'END';
+NChar     ; Noncharacter_Code_Point
+END
+        }
+        # The next sets were added in the new-style
+        if ($v_version ge v3.1.0) {
+            push @return, split /\n/, <<'END';
+OAlpha    ; Other_Alphabetic
+OLower    ; Other_Lowercase
+OMath     ; Other_Math
+OUpper    ; Other_Uppercase
+END
+        }
+        if ($v_version ge v3.1.1) {
+            push @return, "AHex      ; ASCII_Hex_Digit\n";
+        }
+    }
+    if (-e 'EastAsianWidth.txt') {
+        push @return, "ea        ; East_Asian_Width\n";
+    }
+    if (-e 'CompositionExclusions.txt') {
+        push @return, "CE        ; Composition_Exclusion\n";
+    }
+    if (-e 'LineBreak.txt') {
+        push @return, "lb        ; Line_Break\n";
+    }
+    if (-e 'BidiMirroring.txt') {
+        push @return, "bmg       ; Bidi_Mirroring_Glyph\n";
+    }
+    if (-e 'Scripts.txt') {
+        push @return, "sc        ; Script\n";
+    }
+    if (-e 'DNormalizationProps.txt') {
+        push @return, split /\n/, <<'END';
+Comp_Ex   ; Full_Composition_Exclusion
+FC_NFKC   ; FC_NFKC_Closure
+NFC_QC    ; NFC_Quick_Check
+NFD_QC    ; NFD_Quick_Check
+NFKC_QC   ; NFKC_Quick_Check
+NFKD_QC   ; NFKD_Quick_Check
+XO_NFC    ; Expands_On_NFC
+XO_NFD    ; Expands_On_NFD
+XO_NFKC   ; Expands_On_NFKC
+XO_NFKD   ; Expands_On_NFKD
+END
+    }
+    if (-e 'DCoreProperties.txt') {
+        push @return, split /\n/, <<'END';
+IDS       ; ID_Start
+XIDC      ; XID_Continue
+XIDS      ; XID_Start
+END
+        # These can also appear in some versions of PropList.txt
+        push @return, "Lower     ; Lowercase\n"
+                                    unless grep { $_ =~ /^Lower\b/} @return;
+        push @return, "Upper     ; Uppercase\n"
+                                    unless grep { $_ =~ /^Upper\b/} @return;
     }
 
-    my %EAW;
+    # This flag requires the DAge.txt file to be copied into the directory.
+    if (DEBUG && $compare_versions) {
+        push @return, 'age       ; Age';
+    }
 
-    while (<IN>)
-    {
-	next unless /^[0-9A-Fa-f]+(\.\.[0-9A-Fa-f]+)?;/;
-	s/#.*//;
-	s/\s+$//;
+    return @return;
+}
 
-	my ($hexcodes, $pv) = split(/\s*;\s*/);
-        $EAW{$pv} ||= Table->New(Is => "EastAsianWidth$pv");
-      my ($start, $end) = split(/\.\./, $hexcodes);
-      if (defined $end) {
-        $EAW{$pv}->AppendRange(hex($start), hex($end));
-      } else {
-        $EAW{$pv}->Append(hex($start));
-      }
+sub process_PropValueAliases {
+    # This file contains values that properties look like:
+    # bc ; AL        ; Arabic_Letter
+    # blk; n/a       ; Greek_And_Coptic                 ; Greek
+    #
+    # Field 0 is the property.
+    # Field 1 is the short name of a property value or 'n/a' if no
+    #                short name exists;
+    # Field 2 is the full property value name;
+    # Any other fields are more synonyms for the property value.
+    # Purely numeric property values are omitted from the file; as are some
+    # others, fewer and fewer in later releases
+
+    # Entries for the ccc property have an extra field before the
+    # abbreviation:
+    # ccc;   0; NR   ; Not_Reordered
+    # It is the numeric value that the names are synonyms for.
+
+    # There are comment entries for values missing from this file:
+    # # @missing: 0000..10FFFF; ISO_Comment; <none>
+    # # @missing: 0000..10FFFF; Lowercase_Mapping; <code point>
+
+    my $file= shift;
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+    # This whole file was non-existent in early releases, so use our own
+    # internal one if necessary.
+    if (! -e 'PropValueAliases.txt') {
+        $file->insert_lines(get_old_property_value_aliases());
     }
-    close IN;
 
+    # Add any explicit cjk values
+    $file->insert_lines(@cjk_property_values);
+
+    # This line is used only for testing the code that checks for name
+    # conflicts.  There is a script Inherited, and when this line is executed
+    # it causes there to be a name conflict with the 'Inherited' that this
+    # program generates for this block property value
+    #$file->insert_lines('blk; n/a; Herited');
 
-    for (keys %EAW) {
-	$EAW{$_}->Write(
-	    ["lib","ea","$_.pl"],
-	    "EastAsianWidth category '$PropValueAlias{ea}{$_}'"
-	);
+
+    # Process each line of the file ...
+    while ($file->next_line) {
+
+        my ($property, @data) = split /\s*;\s*/;
+
+        # The full name for the ccc property value is in field 2 of the
+        # remaining ones; field 1 for all other properties.  Swap ccc fields 1
+        # and 2.  (Rightmost splice removes field 2, returning it; left splice
+        # inserts that into field 1, thus shifting former field 1 to field 2.)
+        splice (@data, 1, 0, splice(@data, 2, 1)) if $property eq 'ccc';
+
+        # If there is no short name, use the full one in element 1
+        $data[0] = $data[1] if $data[0] eq "n/a";
+
+        # Earlier releases had the pseudo property 'qc' that should expand to
+        # the ones that replace it below.
+        if ($property eq 'qc') {
+            if (lc $data[0] eq 'y') {
+                $file->insert_lines('NFC_QC; Y      ; Yes',
+                                    'NFD_QC; Y      ; Yes',
+                                    'NFKC_QC; Y     ; Yes',
+                                    'NFKD_QC; Y     ; Yes',
+                                    );
+            }
+            elsif (lc $data[0] eq 'n') {
+                $file->insert_lines('NFC_QC; N      ; No',
+                                    'NFD_QC; N      ; No',
+                                    'NFKC_QC; N     ; No',
+                                    'NFKD_QC; N     ; No',
+                                    );
+            }
+            elsif (lc $data[0] eq 'm') {
+                $file->insert_lines('NFC_QC; M      ; Maybe',
+                                    'NFKC_QC; M     ; Maybe',
+                                    );
+            }
+            else {
+                $file->carp_bad_line("qc followed by unexpected '$data[0]");
+            }
+            next;
+        }
+
+        # The first field is the short name, 2nd is the full one.
+        my $property_object = property_ref($property);
+        my $table = $property_object->add_match_table($data[0],
+                                                Full_Name => $data[1]);
+
+        # Start looking for more aliases after these two.
+        for my $i (2 .. @data - 1) {
+            $table->add_alias($data[$i]);
+        }
+    } # End of looping through the file
+
+    # As noted in the comments early in the program, it generates tables for
+    # the default values for all releases, even those for which the concept
+    # didn't exist at the time.  Here we add those if missing.
+    my $age = property_ref('age');
+    if (defined $age && ! defined $age->table('Unassigned')) {
+        $age->add_match_table('Unassigned');
+    }
+    $block->add_match_table('No_Block') if -e 'Blocks.txt'
+                                    && ! defined $block->table('No_Block');
+
+
+    # Now set the default mappings of the properties from the file.  This is
+    # done after the loop because a number of properties have only @missings
+    # entries in the file, and may not show up until the end.
+    my @defaults = $file->get_missings;
+    foreach my $default_ref (@defaults) {
+        my $default = $default_ref->[0];
+        my $property = property_ref($default_ref->[1]);
+        $property->set_default_map($default);
     }
+    return;
 }
 
-##
-## Process HangulSyllableType.txt.
-##
-sub HangulSyllableType_txt()
-{
-    if (not open IN, "HangulSyllableType.txt") {
-        die "$0: HangulSyllableType.txt: $!\n";
+sub get_old_property_value_aliases () {
+    # Returns what would be in PropValueAliases.txt if it existed in very old
+    # versions of Unicode.  It was derived from the one in 3.2, and pared
+    # down.  An attempt was made to use the existence of files to mean
+    # inclusion or not of various aliases, but if this was not sufficient,
+    # using version numbers was resorted to.
+
+    my @return = split /\n/, <<'END';
+bc ; AN        ; Arabic_Number
+bc ; B         ; Paragraph_Separator
+bc ; CS        ; Common_Separator
+bc ; EN        ; European_Number
+bc ; ES        ; European_Separator
+bc ; ET        ; European_Terminator
+bc ; L         ; Left_To_Right
+bc ; ON        ; Other_Neutral
+bc ; R         ; Right_To_Left
+bc ; WS        ; White_Space
+
+# The standard combining classes are very much different in v1, so only use
+# ones that look right (not checked thoroughly)
+ccc;   0; NR   ; Not_Reordered
+ccc;   1; OV   ; Overlay
+ccc;   7; NK   ; Nukta
+ccc;   8; KV   ; Kana_Voicing
+ccc;   9; VR   ; Virama
+ccc; 202; ATBL ; Attached_Below_Left
+ccc; 216; ATAR ; Attached_Above_Right
+ccc; 218; BL   ; Below_Left
+ccc; 220; B    ; Below
+ccc; 222; BR   ; Below_Right
+ccc; 224; L    ; Left
+ccc; 228; AL   ; Above_Left
+ccc; 230; A    ; Above
+ccc; 232; AR   ; Above_Right
+ccc; 234; DA   ; Double_Above
+
+dt ; can       ; canonical
+dt ; enc       ; circle
+dt ; fin       ; final
+dt ; font      ; font
+dt ; fra       ; fraction
+dt ; init      ; initial
+dt ; iso       ; isolated
+dt ; med       ; medial
+dt ; n/a       ; none
+dt ; nb        ; noBreak
+dt ; sqr       ; square
+dt ; sub       ; sub
+dt ; sup       ; super
+
+gc ; C         ; Other                            # Cc | Cf | Cn | Co | Cs
+gc ; Cc        ; Control
+gc ; Cn        ; Unassigned
+gc ; Co        ; Private_Use
+gc ; L         ; Letter                           # Ll | Lm | Lo | Lt | Lu
+gc ; LC        ; Cased_Letter                     # Ll | Lt | Lu
+gc ; Ll        ; Lowercase_Letter
+gc ; Lm        ; Modifier_Letter
+gc ; Lo        ; Other_Letter
+gc ; Lu        ; Uppercase_Letter
+gc ; M         ; Mark                             # Mc | Me | Mn
+gc ; Mc        ; Spacing_Mark
+gc ; Mn        ; Nonspacing_Mark
+gc ; N         ; Number                           # Nd | Nl | No
+gc ; Nd        ; Decimal_Number
+gc ; No        ; Other_Number
+gc ; P         ; Punctuation                      # Pc | Pd | Pe | Pf | Pi | Po | Ps
+gc ; Pd        ; Dash_Punctuation
+gc ; Pe        ; Close_Punctuation
+gc ; Po        ; Other_Punctuation
+gc ; Ps        ; Open_Punctuation
+gc ; S         ; Symbol                           # Sc | Sk | Sm | So
+gc ; Sc        ; Currency_Symbol
+gc ; Sm        ; Math_Symbol
+gc ; So        ; Other_Symbol
+gc ; Z         ; Separator                        # Zl | Zp | Zs
+gc ; Zl        ; Line_Separator
+gc ; Zp        ; Paragraph_Separator
+gc ; Zs        ; Space_Separator
+
+nt ; de        ; Decimal
+nt ; di        ; Digit
+nt ; n/a       ; None
+nt ; nu        ; Numeric
+END
+
+    if (-e 'ArabicShaping.txt') {
+        push @return, split /\n/, <<'END';
+jg ; n/a       ; AIN
+jg ; n/a       ; ALEF
+jg ; n/a       ; DAL
+jg ; n/a       ; GAF
+jg ; n/a       ; LAM
+jg ; n/a       ; MEEM
+jg ; n/a       ; NO_JOINING_GROUP
+jg ; n/a       ; NOON
+jg ; n/a       ; QAF
+jg ; n/a       ; SAD
+jg ; n/a       ; SEEN
+jg ; n/a       ; TAH
+jg ; n/a       ; WAW
+
+jt ; C         ; Join_Causing
+jt ; D         ; Dual_Joining
+jt ; L         ; Left_Joining
+jt ; R         ; Right_Joining
+jt ; U         ; Non_Joining
+jt ; T         ; Transparent
+END
+        if ($v_version ge v3.0.0) {
+            push @return, split /\n/, <<'END';
+jg ; n/a       ; ALAPH
+jg ; n/a       ; BEH
+jg ; n/a       ; BETH
+jg ; n/a       ; DALATH_RISH
+jg ; n/a       ; E
+jg ; n/a       ; FEH
+jg ; n/a       ; FINAL_SEMKATH
+jg ; n/a       ; GAMAL
+jg ; n/a       ; HAH
+jg ; n/a       ; HAMZA_ON_HEH_GOAL
+jg ; n/a       ; HE
+jg ; n/a       ; HEH
+jg ; n/a       ; HEH_GOAL
+jg ; n/a       ; HETH
+jg ; n/a       ; KAF
+jg ; n/a       ; KAPH
+jg ; n/a       ; KNOTTED_HEH
+jg ; n/a       ; LAMADH
+jg ; n/a       ; MIM
+jg ; n/a       ; NUN
+jg ; n/a       ; PE
+jg ; n/a       ; QAPH
+jg ; n/a       ; REH
+jg ; n/a       ; REVERSED_PE
+jg ; n/a       ; SADHE
+jg ; n/a       ; SEMKATH
+jg ; n/a       ; SHIN
+jg ; n/a       ; SWASH_KAF
+jg ; n/a       ; TAW
+jg ; n/a       ; TEH_MARBUTA
+jg ; n/a       ; TETH
+jg ; n/a       ; YEH
+jg ; n/a       ; YEH_BARREE
+jg ; n/a       ; YEH_WITH_TAIL
+jg ; n/a       ; YUDH
+jg ; n/a       ; YUDH_HE
+jg ; n/a       ; ZAIN
+END
+        }
     }
 
-    my %HST;
 
-    while (<IN>)
-    {
-        next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(\w+)/;
-	my ($first, $last, $pv) = (hex($1), hex($2||""), $3);
+    if (-e 'EastAsianWidth.txt') {
+        push @return, split /\n/, <<'END';
+ea ; A         ; Ambiguous
+ea ; F         ; Fullwidth
+ea ; H         ; Halfwidth
+ea ; N         ; Neutral
+ea ; Na        ; Narrow
+ea ; W         ; Wide
+END
+    }
 
-        $HST{$pv} ||= Table->New(Is => "HangulSyllableType$pv");
-        $HST{$pv}->Append($first);
+    if (-e 'LineBreak.txt') {
+        push @return, split /\n/, <<'END';
+lb ; AI        ; Ambiguous
+lb ; AL        ; Alphabetic
+lb ; B2        ; Break_Both
+lb ; BA        ; Break_After
+lb ; BB        ; Break_Before
+lb ; BK        ; Mandatory_Break
+lb ; CB        ; Contingent_Break
+lb ; CL        ; Close_Punctuation
+lb ; CM        ; Combining_Mark
+lb ; CR        ; Carriage_Return
+lb ; EX        ; Exclamation
+lb ; GL        ; Glue
+lb ; HY        ; Hyphen
+lb ; ID        ; Ideographic
+lb ; IN        ; Inseperable
+lb ; IS        ; Infix_Numeric
+lb ; LF        ; Line_Feed
+lb ; NS        ; Nonstarter
+lb ; NU        ; Numeric
+lb ; OP        ; Open_Punctuation
+lb ; PO        ; Postfix_Numeric
+lb ; PR        ; Prefix_Numeric
+lb ; QU        ; Quotation
+lb ; SA        ; Complex_Context
+lb ; SG        ; Surrogate
+lb ; SP        ; Space
+lb ; SY        ; Break_Symbols
+lb ; XX        ; Unknown
+lb ; ZW        ; ZWSpace
+END
+    }
 
-	if ($last) { $HST{$pv}->Extend($last) }
+    if (-e 'DNormalizationProps.txt') {
+        push @return, split /\n/, <<'END';
+qc ; M         ; Maybe
+qc ; N         ; No
+qc ; Y         ; Yes
+END
     }
-    close IN;
 
-    for (keys %HST) {
-	$HST{$_}->Write(
-	    ["lib","hst","$_.pl"],
-	    "HangulSyllableType category '$PropValueAlias{hst}{$_}'"
-	);
+    if (-e 'Scripts.txt') {
+        push @return, split /\n/, <<'END';
+sc ; Arab      ; Arabic
+sc ; Armn      ; Armenian
+sc ; Beng      ; Bengali
+sc ; Bopo      ; Bopomofo
+sc ; Cans      ; Canadian_Aboriginal
+sc ; Cher      ; Cherokee
+sc ; Cyrl      ; Cyrillic
+sc ; Deva      ; Devanagari
+sc ; Dsrt      ; Deseret
+sc ; Ethi      ; Ethiopic
+sc ; Geor      ; Georgian
+sc ; Goth      ; Gothic
+sc ; Grek      ; Greek
+sc ; Gujr      ; Gujarati
+sc ; Guru      ; Gurmukhi
+sc ; Hang      ; Hangul
+sc ; Hani      ; Han
+sc ; Hebr      ; Hebrew
+sc ; Hira      ; Hiragana
+sc ; Ital      ; Old_Italic
+sc ; Kana      ; Katakana
+sc ; Khmr      ; Khmer
+sc ; Knda      ; Kannada
+sc ; Laoo      ; Lao
+sc ; Latn      ; Latin
+sc ; Mlym      ; Malayalam
+sc ; Mong      ; Mongolian
+sc ; Mymr      ; Myanmar
+sc ; Ogam      ; Ogham
+sc ; Orya      ; Oriya
+sc ; Qaai      ; Inherited
+sc ; Runr      ; Runic
+sc ; Sinh      ; Sinhala
+sc ; Syrc      ; Syriac
+sc ; Taml      ; Tamil
+sc ; Telu      ; Telugu
+sc ; Thaa      ; Thaana
+sc ; Thai      ; Thai
+sc ; Tibt      ; Tibetan
+sc ; Yiii      ; Yi
+sc ; Zyyy      ; Common
+END
     }
-}
 
-##
-## Process Jamo.txt.
-##
-sub Jamo_txt()
-{
-    if (not open IN, "Jamo.txt") {
-        die "$0: Jamo.txt: $!\n";
+    if ($v_version ge v2.0.0) {
+        push @return, split /\n/, <<'END';
+dt ; com       ; compat
+dt ; nar       ; narrow
+dt ; sml       ; small
+dt ; vert      ; vertical
+dt ; wide      ; wide
+
+gc ; Cf        ; Format
+gc ; Cs        ; Surrogate
+gc ; Lt        ; Titlecase_Letter
+gc ; Me        ; Enclosing_Mark
+gc ; Nl        ; Letter_Number
+gc ; Pc        ; Connector_Punctuation
+gc ; Sk        ; Modifier_Symbol
+END
+    }
+    if ($v_version ge v2.1.2) {
+        push @return, "bc ; S         ; Segment_Separator\n";
+    }
+    if ($v_version ge v2.1.5) {
+        push @return, split /\n/, <<'END';
+gc ; Pf        ; Final_Punctuation
+gc ; Pi        ; Initial_Punctuation
+END
+    }
+    if ($v_version ge v2.1.8) {
+        push @return, "ccc; 240; IS   ; Iota_Subscript\n";
     }
-    my $Short = Table->New();
 
-    while (<IN>)
-    {
-	next unless /^([0-9A-Fa-f]+)\s*;\s*(\w*)/;
-	my ($code, $short) = (hex($1), $2);
+    if ($v_version ge v3.0.0) {
+        push @return, split /\n/, <<'END';
+bc ; AL        ; Arabic_Letter
+bc ; BN        ; Boundary_Neutral
+bc ; LRE       ; Left_To_Right_Embedding
+bc ; LRO       ; Left_To_Right_Override
+bc ; NSM       ; Nonspacing_Mark
+bc ; PDF       ; Pop_Directional_Format
+bc ; RLE       ; Right_To_Left_Embedding
+bc ; RLO       ; Right_To_Left_Override
+
+ccc; 233; DB   ; Double_Below
+END
+    }
+
+    if ($v_version ge v3.1.0) {
+        push @return, "ccc; 226; R    ; Right\n";
+    }
+
+    return @return;
+}
+
+{ # Closure
+    # This is used to store the range list of all the code points usable when
+    # the little used $compare_versions feature is enabled.
+    my $compare_versions_range_list;
+
+    sub process_generic_property_file {
+        # This processes a file containing property mappings and puts them
+        # into internal map tables.  It should be used to handle any property
+        # files that have mappings from a code point or range thereof to
+        # something else.  This means almost all the UCD .txt files.
+        # each_line_handlers() should be set to adjust the lines of these
+        # files, if necessary, to what this routine understands:
+        #
+        # 0374          ; NFD_QC; N
+        # 003C..003E    ; Math
+        #
+        # the fields are: "codepoint range ; property; map"
+        #
+        # meaning the codepoints in the range all have the value 'map' under
+        # 'property'.
+        # Beginning and trailing white space in each field are not signficant.
+        # Note there is not a trailing semi-colon in the above.  A trailing
+        # semi-colon means the map is a null-string.  An omitted map, as
+        # opposed to a null-string, is assumed to be 'Y', based on Unicode
+        # table syntax.  (This could have been hidden from this routine by
+        # doing it in the $file object, but that would require parsing of the
+        # line there, so would have to parse it twice, or change the interface
+        # to pass this an array.  So not done.)
+        #
+        # The map field may begin with a sequence of commands that apply to
+        # this range.  Each such command begins and ends with $CMD_DELIM.
+        # These are used to indicate, for example, that the mapping for a
+        # range has a non-default type.
+        #
+        # This loops through the file, calling it's next_line() method, and
+        # then taking the map and adding it to the property's table.
+        # Complications arise because any number of properties can be in the
+        # file, in any order, interspersed in any way.  The first time a
+        # property is seen, it gets information about that property and
+        # cache's it for quick retrieval later.  It also normalizes the maps
+        # so that only one of many synonym is stored.  The Unicode input files
+        # do use some multiple synonyms.
+
+        my $file = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my %property_info;               # To keep track of what properties
+                                         # have already had entries in the
+                                         # current file, and info about each,
+                                         # so don't have to recompute.
+        my $property_name;               # property currently being worked on
+        my $property_type;               # and its type
+        my $previous_property_name = ""; # name from last time through loop
+        my $property_object;             # pointer to the current property's
+                                         # object
+        my $property_addr;               # the address of that object
+        my $default_map;                 # the string that code points missing
+                                         # from the file map to
+        my $default_table;               # For non-string properties, a
+                                         # reference to the match table that
+                                         # will contain the list of code
+                                         # points that map to $default_map.
+
+        # Get the next real non-comment line
+        LINE:
+        while ($file->next_line) {
+
+            # Default replacement type; means that if parts of the range have
+            # already been stored in our tables, the new map overrides them if
+            # they differ more than cosmetically
+            my $replace = $IF_NOT_EQUIVALENT;
+            my $map_type;            # Default type for the map of this range
+
+            #local $to_trace = 1 if main::DEBUG;
+            trace $_ if main::DEBUG && $to_trace;
+
+            # Split the line into components
+            my ($range, $property_name, $map, @remainder)
+                = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
+
+            # If more or less on the line than we are expecting, warn and skip
+            # the line
+            if (@remainder) {
+                $file->carp_bad_line('Extra fields');
+                next LINE;
+            }
+            elsif ( ! defined $property_name) {
+                $file->carp_bad_line('Missing property');
+                next LINE;
+            }
+
+            # Examine the range.
+            if ($range !~ /^ ($code_point_re) (?:\.\. ($code_point_re) )? $/x)
+            {
+                $file->carp_bad_line("Range '$range' not of the form 'CP1' or 'CP1..CP2' (where CP1,2 are code points in hex)");
+                next LINE;
+            }
+            my $low = hex $1;
+            my $high = (defined $2) ? hex $2 : $low;
+
+            # For the very specialized case of comparing two Unicode
+            # versions...
+            if (DEBUG && $compare_versions) {
+                if ($property_name eq 'Age') {
+
+                    # Only allow code points at least as old as the version
+                    # specified.
+                    my $age = pack "C*", split(/\./, $map);        # v string
+                    next LINE if $age gt $compare_versions;
+                }
+                else {
+
+                    # Again, we throw out code points younger than those of
+                    # the specified version.  By now, the Age property is
+                    # populated.  We use the intersection of each input range
+                    # with this property to find what code points in it are
+                    # valid.   To do the intersection, we have to convert the
+                    # Age property map to a Range_list.  We only have to do
+                    # this once.
+                    if (! defined $compare_versions_range_list) {
+                        my $age = property_ref('Age');
+                        if (! -e 'DAge.txt') {
+                            croak "Need to have 'DAge.txt' file to do version comparison";
+                        }
+                        elsif ($age->count == 0) {
+                            croak "The 'Age' table is empty, but its file exists";
+                        }
+                        $compare_versions_range_list
+                                        = Range_List->new(Initialize => $age);
+                    }
+
+                    # An undefined map is always 'Y'
+                    $map = 'Y' if ! defined $map;
+
+                    # Calculate the intersection of the input range with the
+                    # code points that are known in the specified version
+                    my @ranges = ($compare_versions_range_list
+                                  & Range->new($low, $high))->ranges;
+
+                    # If the intersection is empty, throw away this range
+                    next LINE unless @ranges;
+
+                    # Only examine the first range this time through the loop.
+                    my $this_range = shift @ranges;
+
+                    # Put any remaining ranges in the queue to be processed
+                    # later.  Note that there is unnecessary work here, as we
+                    # will do the intersection again for each of these ranges
+                    # during some future iteration of the LINE loop, but this
+                    # code is not used in production.  The later intersections
+                    # are guaranteed to not splinter, so this will not become
+                    # an infinite loop.
+                    my $line = join ';', $property_name, $map;
+                    foreach my $range (@ranges) {
+                        $file->insert_adjusted_lines(sprintf("%04X..%04X; %s",
+                                                            $range->start,
+                                                            $range->end,
+                                                            $line));
+                    }
+
+                    # And process the first range, like any other.
+                    $low = $this_range->start;
+                    $high = $this_range->end;
+                }
+            } # End of $compare_versions
+
+            # If changing to a new property, get the things constant per
+            # property
+            if ($previous_property_name ne $property_name) {
+
+                $property_object = property_ref($property_name);
+                if (! defined $property_object) {
+                    $file->carp_bad_line("Unexpected property '$property_name'.  Skipped");
+                    next LINE;
+                }
+                $property_addr = main::objaddr($property_object);
+
+                # Defer changing names until have a line that is acceptable
+                # (the 'next' statement above means is unacceptable)
+                $previous_property_name = $property_name;
+
+                # If not the first time for this property, retrieve info about
+                # it from the cache
+                if (defined ($property_info{$property_addr}{'type'})) {
+                    $property_type = $property_info{$property_addr}{'type'};
+                    $default_map = $property_info{$property_addr}{'default'};
+                    $map_type
+                        = $property_info{$property_addr}{'pseudo_map_type'};
+                    $default_table
+                            = $property_info{$property_addr}{'default_table'};
+                }
+                else {
+
+                    # Here, is the first time for this property.  Set up the
+                    # cache.
+                    $property_type = $property_info{$property_addr}{'type'}
+                                   = $property_object->type;
+                    $map_type
+                        = $property_info{$property_addr}{'pseudo_map_type'}
+                        = $property_object->pseudo_map_type;
+
+                    # The Unicode files are set up so that if the map is not
+                    # defined, it is a binary property
+                    if (! defined $map && $property_type != $BINARY) {
+                        if ($property_type != $UNKNOWN
+                            && $property_type != $NON_STRING)
+                        {
+                            $file->carp_bad_line("No mapping defined on a non-binary property.  Using 'Y' for the map");
+                        }
+                        else {
+                            $property_object->set_type($BINARY);
+                            $property_type
+                                = $property_info{$property_addr}{'type'}
+                                = $BINARY;
+                        }
+                    }
+
+                    # Get any @missings default for this property.  This
+                    # should precede the first entry for the property in the
+                    # input file, and is located in a comment that has been
+                    # stored by the Input_file class until we access it here.
+                    # It's possible that there is more than one such line
+                    # waiting for us; collect them all, and parse
+                    my @missings_list = $file->get_missings
+                                            if $file->has_missings_defaults;
+                    foreach my $default_ref (@missings_list) {
+                        my $default = $default_ref->[0];
+                        my $addr = objaddr property_ref($default_ref->[1]);
+
+                        # For string properties, the default is just what the
+                        # file says, but non-string properties should already
+                        # have set up a table for the default property value;
+                        # use the table for these, so can resolve synonyms
+                        # later to a single standard one.
+                        if ($property_type == $STRING
+                            || $property_type == $UNKNOWN)
+                        {
+                            $property_info{$addr}{'missings'} = $default;
+                        }
+                        else {
+                            $property_info{$addr}{'missings'}
+                                        = $property_object->table($default);
+                        }
+                    }
+
+                    # Finished storing all the @missings defaults in the input
+                    # file so far.  Get the one for the current property.
+                    my $missings = $property_info{$property_addr}{'missings'};
+
+                    # But we likely have separately stored what the default
+                    # should be.  (This is to accommodate versions of the
+                    # standard where the @missings lines are absent or
+                    # incomplete.)  Hopefully the two will match.  But check
+                    # it out.
+                    $default_map = $property_object->default_map;
+
+                    # If the map is a ref, it means that the default won't be
+                    # processed until later, so undef it, so next few lines
+                    # will redefine it to something that nothing will match
+                    undef $default_map if ref $default_map;
+
+                    # Create a $default_map if don't have one; maybe a dummy
+                    # that won't match anything.
+                    if (! defined $default_map) {
+
+                        # Use any @missings line in the file.
+                        if (defined $missings) {
+                            if (ref $missings) {
+                                $default_map = $missings->full_name;
+                                $default_table = $missings;
+                            }
+                            else {
+                                $default_map = $missings;
+                            }
+                        
+                            # And store it with the property for outside use.
+                            $property_object->set_default_map($default_map);
+                        }
+                        else {
+
+                            # Neither an @missings nor a default map.  Create
+                            # a dummy one, so won't have to test definedness
+                            # in the main loop.
+                            $default_map = '_Perl This will never be in a file
+                                            from Unicode';
+                        }
+                    }
+
+                    # Here, we have $default_map defined, possibly in terms of
+                    # $missings, but maybe not, and possibly is a dummy one.
+                    if (defined $missings) {
+
+                        # Make sure there is no conflict between the two.
+                        # $missings has priority.
+                        if (ref $missings) {
+                            $default_table 
+                                = $property_object->table($default_map);
+                            if (! defined $default_table
+                                || $default_table != $missings)
+                            {
+                                if (! defined $default_table) {
+                                    $default_table = $UNDEF;
+                                }
+                                $file->carp_bad_line(<<END
+The \@missings line for $property_name in $file says that missings default to
+$missings, but we expect it to be $default_table.  $missings used.
+END
+                                );
+                                $default_table = $missings;
+                                $default_map = $missings->full_name;
+                            }
+                            $property_info{$property_addr}{'default_table'}
+                                                        = $default_table;
+                        }
+                        elsif ($default_map ne $missings) {
+                            $file->carp_bad_line(<<END
+The \@missings line for $property_name in $file says that missings default to
+$missings, but we expect it to be $default_map.  $missings used.
+END
+                            );
+                            $default_map = $missings;
+                        }
+                    }
+
+                    $property_info{$property_addr}{'default'}
+                                                    = $default_map;
+
+                    # If haven't done so already, find the table corresponding
+                    # to this map for non-string properties.
+                    if (! defined $default_table
+                        && $property_type != $STRING
+                        && $property_type != $UNKNOWN)
+                    {
+                        $default_table = $property_info{$property_addr}
+                                                        {'default_table'}
+                                    = $property_object->table($default_map);
+                    }
+                } # End of is first time for this property
+            } # End of switching properties.
+
+            # Ready to process the line.
+            # The Unicode files are set up so that if the map is not defined,
+            # it is a binary property with value 'Y'
+            if (! defined $map) {
+                $map = 'Y';
+            }
+            else {
+
+                # If the map begins with a special command to us (enclosed in
+                # delimiters), extract the command(s).
+                if (substr($map, 0, 1) eq $CMD_DELIM) {
+                    while ($map =~ s/ ^ $CMD_DELIM (.*?) $CMD_DELIM //x) {
+                        my $command = $1;
+                        if ($command =~  / ^ $REPLACE_CMD= (.*) /x) {
+                            $replace = $1;
+                        }
+                        elsif ($command =~  / ^ $MAP_TYPE_CMD= (.*) /x) {
+                            $map_type = $1;
+                        }
+                        else {
+                           $file->carp_bad_line("Unknown command line: '$1'");
+                           next LINE;
+                        }
+                    }
+                }
+            }
+
+            if ($default_map eq $CODE_POINT && $map =~ / ^ $code_point_re $/x)
+            {
+
+                # Here, we have a map to a particular code point, and the
+                # default map is to a code point itself.  If the range
+                # includes the particular code point, change that portion of
+                # the range to the default.  This makes sure that in the final
+                # table only the non-defaults are listed.
+                my $decimal_map = hex $map;
+                if ($low <= $decimal_map && $decimal_map <= $high) {
+
+                    # If the range includes stuff before or after the map
+                    # we're changing, split it and process the split-off parts
+                    # later.
+                    if ($low < $decimal_map) {
+                        $file->insert_adjusted_lines(
+                                            sprintf("%04X..%04X; %s; %s",
+                                                    $low,
+                                                    $decimal_map - 1,
+                                                    $property_name,
+                                                    $map));
+                    }
+                    if ($high > $decimal_map) {
+                        $file->insert_adjusted_lines(
+                                            sprintf("%04X..%04X; %s; %s",
+                                                    $decimal_map + 1,
+                                                    $high,
+                                                    $property_name,
+                                                    $map));
+                    }
+                    $low = $high = $decimal_map;
+                    $map = $CODE_POINT;
+                }
+            }
+
+            # If we can tell that this is a synonym for the default map, use
+            # the default one instead.
+            if ($property_type != $STRING
+                && $property_type != $UNKNOWN)
+            {
+                my $table = $property_object->table($map);
+                if (defined $table && $table == $default_table) {
+                    $map = $default_map;
+                }
+            }
 
-	$Short->Append($code, $short);
+            # And figure out the map type if not known.
+            if (! defined $map_type || $map_type == $COMPUTE_NO_MULTI_CP) {
+                if ($map eq "") {   # Nulls are always $NULL map type
+                    $map_type = $NULL;
+                } # Otherwise, non-strings, and those that don't allow
+                  # $MULTI_CP, and those that aren't multiple code points are
+                  # 0
+                elsif
+                   (($property_type != $STRING && $property_type != $UNKNOWN)
+                   || (defined $map_type && $map_type == $COMPUTE_NO_MULTI_CP)
+                   || $map !~ /^ $code_point_re ( \  $code_point_re )+ $ /x)
+                {
+                    $map_type = 0;
+                }
+                else {
+                    $map_type = $MULTI_CP;
+                }
+            }
+
+            $property_object->add_map($low, $high,
+                                        $map,
+                                        Type => $map_type,
+                                        Replace => $replace);
+        } # End of loop through file's lines
+
+        return;
     }
-    close IN;
-    # $Short->Write("JamoShort.pl");
 }
 
-##
-## Process Scripts.txt.
-##
-sub Scripts_txt()
-{
-    my @ScriptInfo;
+# Unused until revise charnames;
+#sub check_and_handle_compound_name {
+#    This looks at Name properties for parenthesized components and splits
+#    them off.  Thus it finds FF as an equivalent to Form Feed.
+#    my $code_point = shift;
+#    my $name = shift;
+#    if ($name =~ /^ ( .*? ) ( \s* ) \( ( [^)]* ) \) (.*) $/x) {
+#        #local $to_trace = 1 if main::DEBUG;
+#        trace $1, $2, $3, $4 if main::DEBUG && $to_trace;
+#        push @more_Names, "$code_point; $1";
+#        push @more_Names, "$code_point; $3";
+#        Carp::my_carp_bug("Expecting blank space before left parenthesis in '$_'.  Proceeding and assuming it was there;") if $2 ne " ";
+#        Carp::my_carp_bug("Not expecting anything after the right parenthesis in '$_'.  Proceeding and ignoring that;") if $4 ne "";
+#    }
+#    return;
+#}
+
+{ # Closure for UnicodeData.txt handling
+
+    # This file was the first one in the UCD; its design leads to some
+    # awkwardness in processing.  Here is a sample line:
+    # 0041;LATIN CAPITAL LETTER A;Lu;0;L;;;;;N;;;;0061;
+    # The fields in order are:
+    my $i = 0;            # The code point is in field 0, and is shifted off.
+    my $NAME = $i++;      # character name (e.g. "LATIN CAPITAL LETTER A")
+    my $CATEGORY = $i++;  # category (e.g. "Lu")
+    my $CCC = $i++;       # Canonical combining class (e.g. "230")
+    my $BIDI = $i++;      # directional class (e.g. "L")
+    my $PERL_DECOMPOSITION = $i++;  # decomposition mapping
+    my $PERL_DECIMAL_DIGIT = $i++;   # decimal digit value
+    my $NUMERIC_TYPE_OTHER_DIGIT = $i++; # digit value, like a superscript
+                                         # Dual-use in this program; see below
+    my $NUMERIC = $i++;   # numeric value
+    my $MIRRORED = $i++;  # ? mirrored
+    my $UNICODE_1_NAME = $i++; # name in Unicode 1.0
+    my $COMMENT = $i++;   # iso comment
+    my $UPPER = $i++;     # simple uppercase mapping
+    my $LOWER = $i++;     # simple lowercase mapping
+    my $TITLE = $i++;     # simple titlecase mapping
+    my $input_field_count = $i;
+
+    # This routine in addition outputs these extra fields:
+    my $DECOMP_TYPE = $i++; # Decomposition type
+    my $DECOMP_MAP = $i++;  # Must be last; another decomposition mapping
+    my $last_field = $i - 1;
+
+    # All these are read into an array for each line, with the indices defined
+    # above.  The empty fields in the example line above indicate that the
+    # value is defaulted.  The handler called for each line of the input
+    # changes these to their defaults.
+
+    # Here are the official names of the properties, in a parallel array:
+    my @field_names;
+    $field_names[$BIDI] = 'Bidi_Class';
+    $field_names[$CATEGORY] = 'General_Category';
+    $field_names[$CCC] = 'Canonical_Combining_Class';
+    $field_names[$COMMENT] = 'ISO_Comment';
+    $field_names[$DECOMP_MAP] = 'Decomposition_Mapping';
+    $field_names[$DECOMP_TYPE] = 'Decomposition_Type';
+    $field_names[$LOWER] = 'Simple_Lowercase_Mapping';
+    $field_names[$MIRRORED] = 'Bidi_Mirrored';
+    $field_names[$NAME] = 'Name';
+    $field_names[$NUMERIC] = 'Numeric_Value';
+    $field_names[$NUMERIC_TYPE_OTHER_DIGIT] = 'Numeric_Type';
+    $field_names[$PERL_DECIMAL_DIGIT] = 'Perl_Decimal_Digit';
+    $field_names[$PERL_DECOMPOSITION] = 'Perl_Decomposition_Mapping';
+    $field_names[$TITLE] = 'Simple_Titlecase_Mapping';
+    $field_names[$UNICODE_1_NAME] = 'Unicode_1_Name';
+    $field_names[$UPPER] = 'Simple_Uppercase_Mapping';
+
+    # Some of these need a little more explanation.  The $PERL_DECIMAL_DIGIT
+    # field does not lead to an official Unicode property, but is used in
+    # calculating the Numeric_Type.  Perl however, creates a file from this
+    # field, so a Perl property is created from it.  Similarly, the Other
+    # Digit field is used only for calculating the Numeric_Type, and so it can
+    # be safely re-used as the place to store the value for Numeric_Type;
+    # hence it is referred to as $NUMERIC_TYPE_OTHER_DIGIT.  The input field
+    # named $PERL_DECOMPOSITION is a combination of both the decomposition
+    # mapping and its type.  Perl creates a file containing exactly this
+    # field, so it is used for that.  The two properties are separated into
+    # two extra output fields, $DECOMP_MAP and $DECOMP_TYPE.
+
+    # This file is processed like most in this program.  Control is passed to
+    # process_generic_property_file() which calls filter_UnicodeData_line()
+    # for each input line.  This filter converts the input into line(s) that
+    # process_generic_property_file() understands.  There is also a setup
+    # routine called before any of the file is processed, and a handler for
+    # EOF processing, all in this closure.
+
+    # A huge speed-up occurred at the cost of some added complexity when these
+    # routines were altered to buffer the outputs into ranges.  Almost all the
+    # lines of the input file apply to just one code point, and for most
+    # properties, the map for the next code point up is the same as the
+    # current one.  So instead of creating a line for each property for each
+    # input line, filter_UnicodeData_line() remembers what the previous map
+    # of a property was, and doesn't generate a line to pass on until it has
+    # to, as when the map changes; and that passed-on line encompasses the
+    # whole contiguous range of code points that have the same map for that
+    # property.  This means a slight amount of extra setup, and having to
+    # flush these buffers on EOF, testing if the maps have changed, plus
+    # remembering state information in the closure.  But it means a lot less
+    # real time in not having to change the data base for each property on
+    # each line.
+
+    # Another complication is that there are already a few ranges designated
+    # in the input.  There are two lines for each, with the same maps except
+    # the code point and name on each line.  This was actually the hardest
+    # thing to design around.  The code points in those ranges may actually
+    # have real maps not given by these two lines.  These maps will either
+    # be algorthimically determinable, or in the extracted files furnished
+    # with the UCD.  In the event of conflicts between these extracted files,
+    # and this one, Unicode says that this one prevails.  But it shouldn't
+    # prevail for conflicts that occur in these ranges.  The data from the
+    # extracted files prevails in those cases.  So, this program is structured
+    # so that those files are processed first, storing maps.  Then the other
+    # files are processed, generally overwriting what the extracted files
+    # stored.  But just the range lines in this input file are processed
+    # without overwriting.  This is accomplished by adding a special string to
+    # the lines output to tell process_generic_property_file() to turn off the
+    # overwriting for just this one line.
+    # A similar mechanism is used to tell it that the map is of a non-default
+    # type.
+
+    sub setup_UnicodeData { # Called before any lines of the input are read
+        my $file = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $Perl_decomp = Property->new('Perl_Decomposition_Mapping',
+                                        Directory => '.',
+                                        File => 'Decomposition',
+                                        Format => $STRING_FORMAT,
+                                        Internal_Only_Warning => 1,
+                                        Perl_Extension => 1,
+                                        Default_Map => $CODE_POINT,
+
+                                        # This is a specially formatted table
+                                        # explicitly for normalize.pm, which
+                                        # is expecting a particular format,
+                                        # which means that mappings containing
+                                        # multiple code points are in the main
+                                        # body of the table
+                                        Map_Type => $COMPUTE_NO_MULTI_CP,
+                                        Type => $STRING,
+                                        );
+        $Perl_decomp->add_comment(join_lines(<<END
+This mapping is a combination of the Unicode 'Decomposition_Type' and
+'Decomposition_Mapping' properties, formatted for use by normalize.pm.  It is
+identical to the official Unicode 'Decomposition_Mapping'  property except for
+two things:
+ 1) It omits the algorithmically determinable Hangul syllable decompositions,
+which normalize.pm handles algorithmically.
+ 2) It contains the decomposition type as well.  Non-canonical decompositions
+begin with a word in angle brackets, like <super>, which denotes the
+compatible decomposition type.  If the map does not begin with the <angle
+brackets>, the decomposition is canonical.
+END
+        ));
+
+        my $Decimal_Digit = Property->new("Perl_Decimal_Digit",
+                                        Default_Map => "",
+                                        Perl_Extension => 1,
+                                        File => 'Digit',    # Trad. location
+                                        Directory => $map_directory,
+                                        Type => $STRING,
+                                        Range_Size_1 => 1,
+                                        );
+        $Decimal_Digit->add_comment(join_lines(<<END
+This file gives the mapping of all code points which represent a single
+decimal digit [0-9] to their respective digits.  For example, the code point
+U+0031 (an ASCII '1') is mapped to a numeric 1.  These code points are those
+that have Numeric_Type=Decimal; not special things, like subscripts nor Roman
+numerals.
+END
+        ));
+
+        # This property is not used for generating anything else, and is
+        # usually not output.  By making it last in the list, we can just
+        # change the high end of the loop downwards to avoid the work of
+        # generating a table that is just going to get thrown away.
+        if (! property_ref('Decomposition_Mapping')->to_output_map) {
+            $last_field--;
+        }
+        return;
+    }
+
+    my $first_time = 1;                 # ? Is this the first line of the file
+    my $in_range = 0;                   # ? Are we in one of the file's ranges
+    my $previous_cp;                    # hex code point of previous line
+    my $decimal_previous_cp = -1;       # And its decimal equivalent
+    my @start;                          # For each field, the current starting
+                                        # code point in hex for the range
+                                        # being accumulated.
+    my @fields;                         # The input fields;
+    my @previous_fields;                # And those from the previous call
+
+    sub filter_UnicodeData_line {
+        # Handle a single input line from UnicodeData.txt; see comments above
+        # Conceptually this takes a single line from the file containing N
+        # properties, and converts it into N lines with one property per line,
+        # which is what the final handler expects.  But there are
+        # complications due to the quirkiness of the input file, and to save
+        # time, it accumulates ranges where the property values don't change
+        # and only emits lines when necessary.  This is about an order of
+        # magnitude fewer lines emitted.
+
+        my $file = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        # $_ contains the input line.
+        # -1 in split means retain trailing null fields
+        (my $cp, @fields) = split /\s*;\s*/, $_, -1;
+
+        #local $to_trace = 1 if main::DEBUG;
+        trace $cp, @fields , $input_field_count if main::DEBUG && $to_trace;
+        if (@fields > $input_field_count) {
+            $file->carp_bad_line('Extra fields');
+            $_ = "";
+            return;
+        }
+
+        my $decimal_cp = hex $cp;
+
+        # We have to output all the buffered ranges when the next code point
+        # is not exactly one after the previous one, which means there is a
+        # gap in the ranges.
+        my $force_output = ($decimal_cp != $decimal_previous_cp + 1);
+
+        # The decomposition mapping field requires special handling.  It looks
+        # like either:
+        #
+        # <compat> 0032 0020
+        # 0041 0300
+        #
+        # The decomposition type is enclosed in <brackets>; if missing, it
+        # means the type is canonical.  There are two decomposition mapping
+        # tables: the one for use by Perl's normalize.pm has a special format
+        # which is this field intact; the other, for general use is of
+        # standard format.  In either case we have to find the decomposition
+        # type.  Empty fields have None as their type, and map to the code
+        # point itself
+        if ($fields[$PERL_DECOMPOSITION] eq "") {
+            $fields[$DECOMP_TYPE] = 'None';
+            $fields[$DECOMP_MAP] = $fields[$PERL_DECOMPOSITION] = $CODE_POINT;
+        }
+        else {
+            ($fields[$DECOMP_TYPE], my $map) = $fields[$PERL_DECOMPOSITION]
+                                            =~ / < ( .+? ) > \s* ( .+ ) /x;
+            if (! defined $fields[$DECOMP_TYPE]) {
+                $fields[$DECOMP_TYPE] = 'Canonical';
+                $fields[$DECOMP_MAP] = $fields[$PERL_DECOMPOSITION];
+            }
+            else {
+                $fields[$DECOMP_MAP] = $map;
+            }
+        }
+
+        # The 3 numeric fields also require special handling.  The 2 digit
+        # fields must be either empty or match the number field.  This means
+        # that if it is empty, they must be as well, and the numeric type is
+        # None, and the numeric value is 'Nan'.
+        # The decimal digit field must be empty or match the other digit
+        # field.  If the decimal digit field is non-empty, the code point is
+        # a decimal digit, and the other two fields will have the same value.
+        # If it is empty, but the other digit field is non-empty, the code
+        # point is an 'other digit', and the number field will have the same
+        # value as the other digit field.  If the other digit field is empty,
+        # but the number field is non-empty, the code point is a generic
+        # numeric type.
+        if ($fields[$NUMERIC] eq "") {
+            if ($fields[$PERL_DECIMAL_DIGIT] ne ""
+                || $fields[$NUMERIC_TYPE_OTHER_DIGIT] ne ""
+            ) {
+                $file->carp_bad_line("Numeric values inconsistent.  Trying to process anyway");
+            }
+            $fields[$NUMERIC_TYPE_OTHER_DIGIT] = 'None';
+            $fields[$NUMERIC] = 'NaN';
+        }
+        else {
+            $file->carp_bad_line("'$fields[$NUMERIC]' should be a whole or rational number.  Processing as if it were") if $fields[$NUMERIC] !~ qr{ ^ -? \d+ ( / \d+ )? $ }x;
+            if ($fields[$PERL_DECIMAL_DIGIT] ne "") {
+                $file->carp_bad_line("$fields[$PERL_DECIMAL_DIGIT] should equal $fields[$NUMERIC].  Processing anyway") if $fields[$PERL_DECIMAL_DIGIT] != $fields[$NUMERIC];
+                $fields[$NUMERIC_TYPE_OTHER_DIGIT] = 'Decimal';
+            }
+            elsif ($fields[$NUMERIC_TYPE_OTHER_DIGIT] ne "") {
+                $file->carp_bad_line("$fields[$NUMERIC_TYPE_OTHER_DIGIT] should equal $fields[$NUMERIC].  Processing anyway") if $fields[$NUMERIC_TYPE_OTHER_DIGIT] != $fields[$NUMERIC];
+                $fields[$NUMERIC_TYPE_OTHER_DIGIT] = 'Digit';
+            }
+            else {
+                $fields[$NUMERIC_TYPE_OTHER_DIGIT] = 'Numeric';
+
+                # Rationals require extra effort.
+                register_fraction($fields[$NUMERIC])
+                                                if $fields[$NUMERIC] =~ qr{/};
+            }
+        }
+
+        # For the properties that have empty fields in the file, and which
+        # mean something different from empty, change them to that default.
+        # Certain fields just haven't been empty so far in any Unicode
+        # version, so don't look at those, namely $MIRRORED, $BIDI, $CCC,
+        # $CATEGORY.  This leaves just the two fields, and so we hard-code in
+        # the defaults; which are verly unlikely to ever change.
+        $fields[$UPPER] = $CODE_POINT if $fields[$UPPER] eq "";
+        $fields[$LOWER] = $CODE_POINT if $fields[$LOWER] eq "";
+
+        # UAX44 says that if title is empty, it is the same as whatever upper
+        # is,
+        $fields[$TITLE] = $fields[$UPPER] if $fields[$TITLE] eq "";
+
+        # There are a few pairs of lines like:
+        #   AC00;<Hangul Syllable, First>;Lo;0;L;;;;;N;;;;;
+        #   D7A3;<Hangul Syllable, Last>;Lo;0;L;;;;;N;;;;;
+        # that define ranges.  These should be processed after the fields are
+        # adjusted above, as they may override some of them; but mostly what
+        # is left is to possibly adjust the $NAME field.  The names of all the
+        # paired lines start with a '<', but this is also true of '<control>,
+        # which isn't one of these special ones.
+        if ($fields[$NAME] eq '<control>') {
+
+            # Some code points in this file have the pseudo-name
+            # '<control>', but the official name for such ones is the null
+            # string.
+            $fields[$NAME] = "";
+
+            # We had better not be in between range lines.
+            if ($in_range) {
+                $file->carp_bad_line("Expecting a closing range line, not a $fields[$NAME]'.  Trying anyway");
+                $in_range = 0;
+            }
+        }
+        elsif (substr($fields[$NAME], 0, 1) ne '<') {
+
+            # Here is a non-range line.  We had better not be in between range
+            # lines.
+            if ($in_range) {
+                $file->carp_bad_line("Expecting a closing range line, not a $fields[$NAME]'.  Trying anyway");
+                $in_range = 0;
+            }
+            # XXX until charnames catches up.
+#            if ($fields[$NAME] =~ s/- $cp $//x) {
+#
+#                # These are code points whose names end in their code points,
+#                # which means the names are algorithmically derivable from the
+#                # code points.  To shorten the output Name file, the algorithm
+#                # for deriving these is placed in the file instead of each
+#                # code point, so they have map type $CP_IN_NAME
+#                $fields[$NAME] = $CMD_DELIM
+#                                 . $MAP_TYPE_CMD
+#                                 . '='
+#                                 . $CP_IN_NAME
+#                                 . $CMD_DELIM
+#                                 . $fields[$NAME];
+#            }
+
+            # Some official names are really two alternate names with one in
+            # parentheses.  What we do here is use the full official one for
+            # the standard property (stored just above), but for the charnames
+            # table, we add two more entries, one for each of the alternate
+            # ones.
+            # elsif name ne ""
+            #check_and_handle_compound_name($cp, $fields[$NAME]);
+            #check_and_handle_compound_name($cp, $unicode_1_name);
+            # XXX until charnames catches up.
+        }
+        elsif ($fields[$NAME] =~ /^<(.+), First>$/) {
+            $fields[$NAME] = $1;
+
+            # Here we are at the beginning of a range pair.
+            if ($in_range) {
+                $file->carp_bad_line("Expecting a closing range line, not a beginning one, $fields[$NAME]'.  Trying anyway");
+            }
+            $in_range = 1;
+
+            # Because the properties in the range do not overwrite any already
+            # in the db, we must flush the buffers of what's already there, so
+            # they get handled in the normal scheme.
+            $force_output = 1;
+
+        }
+        elsif ($fields[$NAME] !~ s/^<(.+), Last>$/$1/) {
+            $file->carp_bad_line("Unexpected name starting with '<' $fields[$NAME].  Ignoring this line.");
+            $_ = "";
+            return;
+        }
+        else { # Here, we are at the last line of a range pair.
+
+            if (! $in_range) {
+                $file->carp_bad_line("Unexpected end of range $fields[$NAME] when not in one.  Ignoring this line.");
+                $_ = "";
+                return;
+            }
+            $in_range = 0;
+
+            # Check that the input is valid: that the closing of the range is
+            # the same as the beginning.
+            foreach my $i (0 .. $last_field) {
+                next if $fields[$i] eq $previous_fields[$i];
+                $file->carp_bad_line("Expecting '$fields[$i]' to be the same as '$previous_fields[$i]'.  Bad News.  Trying anyway");
+            }
+
+            # The processing differs depending on the type of range,
+            # determined by its $NAME
+            if ($fields[$NAME] =~ /^Hangul Syllable/) {
+
+                # Check that the data looks right.
+                if ($decimal_previous_cp != $SBase) {
+                    $file->carp_bad_line("Unexpected Hangul syllable start = $previous_cp.  Bad News.  Results will be wrong");
+                }
+                if ($decimal_cp != $SBase + $SCount - 1) {
+                    $file->carp_bad_line("Unexpected Hangul syllable end = $cp.  Bad News.  Results will be wrong");
+                }
+
+                # The Hangul syllable range has a somewhat complicated name
+                # generation algorithm.  Each code point in it has a canonical
+                # decomposition also computable by an algorithm.  The
+                # perl decomposition map table built from these is used only
+                # by normalize.pm, which has the algorithm built in it, so the
+                # decomposition maps are not needed, and are large, so are
+                # omitted from it.  If the full decomposition map table is to
+                # be output, the decompositions are generated for it, in the
+                # EOF handling code for this input file.
+
+                $previous_fields[$DECOMP_TYPE] = 'Canonical';
+
+                # This range is stored in our internal structure with its
+                # own map type, different from all others.
+                $previous_fields[$NAME] = $CMD_DELIM
+                                          . $MAP_TYPE_CMD
+                                          . '='
+                                          . $HANGUL_SYLLABLE
+                                          . $CMD_DELIM
+                                          . $fields[$NAME];
+            }
+            elsif ($fields[$NAME] =~ /^CJK/) {
+
+                # The name for these contains the code point itself, and all
+                # are defined to have the same base name, regardless of what
+                # is in the file.  They are stored in our internal structure
+                # with a map type of $CP_IN_NAME
+                $previous_fields[$NAME] = $CMD_DELIM
+                                           . $MAP_TYPE_CMD
+                                           . '='
+                                           . $CP_IN_NAME
+                                           . $CMD_DELIM
+                                           . 'CJK UNIFIED IDEOGRAPH';
+
+            }
+            elsif ($fields[$CATEGORY] eq 'Co'
+                     || $fields[$CATEGORY] eq 'Cs')
+            {
+                # The names of all the code points in these ranges are set to
+                # null, as there are no names for the private use and
+                # surrogate code points.
+
+                $previous_fields[$NAME] = "";
+            }
+            else {
+                $file->carp_bad_line("Unexpected code point range $fields[$NAME] because category is $fields[$CATEGORY].  Attempting to process it.");
+            }
 
-    if (not open(IN, "Scripts.txt")) {
-        die "$0: Scripts.txt: $!\n";
+            # The first line of the range caused everything else to be output,
+            # and then its values were stored as the beginning values for the
+            # next set of ranges, which this one ends.  Now, for each value,
+            # add a command to tell the handler that these values should not
+            # replace any existing ones in our database.
+            foreach my $i (0 .. $last_field) {
+                $previous_fields[$i] = $CMD_DELIM
+                                        . $REPLACE_CMD
+                                        . '='
+                                        . $NO
+                                        . $CMD_DELIM
+                                        . $previous_fields[$i];
+            }
+
+            # And change things so it looks like the entire range has been
+            # gone through with this being the final part of it.  Adding the
+            # command above to each field will cause this range to be flushed
+            # during the next iteration, as it guaranteed that the stored
+            # field won't match whatever value the next one has.
+            $previous_cp = $cp;
+            $decimal_previous_cp = $decimal_cp;
+
+            # We are now set up for the next iteration; so skip the remaining
+            # code in this subroutine that does the same thing, but doesn't
+            # know about these ranges.
+            $_ = "";
+            return;
+        }
+
+        # On the very first line, we fake it so the code below thinks there is
+        # nothing to output, and initialize so that when it does get output it
+        # uses the first line's values for the lowest part of the range.
+        # (One could avoid this by using peek(), but then one would need to
+        # know the adjustments done above and do the same ones in the setup
+        # routine; not worth it)
+        if ($first_time) {
+            $first_time = 0;
+            @previous_fields = @fields;
+            @start = ($cp) x scalar @fields;
+            $decimal_previous_cp = $decimal_cp - 1;
+        }
+
+        # For each field, output the stored up ranges that this code point
+        # doesn't fit in.  Earlier we figured out if all ranges should be
+        # terminated because of changing the replace or map type styles, or if
+        # there is a gap between this new code point and the previous one, and
+        # that is stored in $force_output.  But even if those aren't true, we
+        # need to output the range if this new code point's value for the
+        # given property doesn't match the stored range's.
+        #local $to_trace = 1 if main::DEBUG;
+        foreach my $i (0 .. $last_field) {
+            my $field = $fields[$i];
+            if ($force_output || $field ne $previous_fields[$i]) {
+
+                # Flush the buffer of stored values.
+                $file->insert_adjusted_lines("$start[$i]..$previous_cp; $field_names[$i]; $previous_fields[$i]");
+
+                # Start a new range with this code point and its value
+                $start[$i] = $cp;
+                $previous_fields[$i] = $field;
+            }
+        }
+
+        # Set the values for the next time.
+        $previous_cp = $cp;
+        $decimal_previous_cp = $decimal_cp;
+
+        # The input line has generated whatever adjusted lines are needed, and
+        # should not be looked at further.
+        $_ = "";
+        return;
     }
-    while (<IN>) {
-        next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/;
 
-        # Wait until all the scripts have been read since
-        # they are not listed in numeric order.
-        push @ScriptInfo, [ hex($1), hex($2||""), $3 ];
+    sub EOF_UnicodeData {
+        # Called upon EOF to flush the buffers, and create the Hangul
+        # decomposition mappings if needed.
+
+        my $file = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        # Flush the buffers.
+        foreach my $i (1 .. $last_field) {
+            $file->insert_adjusted_lines("$start[$i]..$previous_cp; $field_names[$i]; $previous_fields[$i]");
+        }
+
+        if (-e 'Jamo.txt') {
+
+            # The algorithm is published by Unicode, based on values in
+            # Jamo.txt, (which should have been processed before this
+            # subroutine), and the results left in %Jamo
+            unless (%Jamo) {
+                Carp::my_carp_bug("Jamo.txt should be processed before Unicode.txt.  Hangul syllables not generated.");
+                return;
+            }
+
+            # If the full decomposition map table is being output, insert
+            # into it the Hangul syllable mappings.  This is to avoid having
+            # to publish a subroutine in it to compute them.  (which would
+            # essentially be this code.)  This uses the algorithm published by
+            # Unicode.
+            if (property_ref('Decomposition_Mapping')->to_output_map) {
+                for (my $S = $SBase; $S < $SBase + $SCount; $S++) {
+                    use integer;
+                    my $SIndex = $S - $SBase;
+                    my $L = $LBase + $SIndex / $NCount;
+                    my $V = $VBase + ($SIndex % $NCount) / $TCount;
+                    my $T = $TBase + $SIndex % $TCount;
+
+                    trace "L=$L, V=$V, T=$T" if main::DEBUG && $to_trace;
+                    my $decomposition = sprintf("%04X %04X", $L, $V);
+                    $decomposition .= sprintf(" %04X", $T) if $T != $TBase;
+                    $file->insert_adjusted_lines(
+                                sprintf("%04X; Decomposition_Mapping; %s",
+                                        $S,
+                                        $decomposition));
+                }
+            }
+        }
+
+        return;
     }
-    close IN;
 
-    # Now append the scripts properties in their code point order.
+    sub filter_v1_ucd {
+        # Fix UCD lines in version 1.  This is probably overkill, but this
+        # fixes some glaring errors in Version 1 UnicodeData.txt.  That file:
+        # 1)    had many Hangul (U+3400 - U+4DFF) code points that were later
+        #       removed.  This program retains them
+        # 2)    didn't include ranges, which it should have, and which are now
+        #       added in @corrected_lines below.  It was hand populated by
+        #       taking the data from Version 2, verified by analyzing
+        #       DAge.txt.
+        # 3)    There is a syntax error in the entry for U+09F8 which could
+        #       cause problems for utf8_heavy, and so is changed.  It's
+        #       numeric value was simply a minus sign, without any number.
+        #       (Eventually Unicode changed the code point to non-numeric.)
+        # 4)    The decomposition types often don't match later versions
+        #       exactly, and the whole syntax of that field is different; so
+        #       the syntax is changed as well as the types to their later
+        #       terminology.  Otherwise normalize.pm would be very unhappy
+        # 5)    Many ccc classes are different.  These are left intact.
+        # 6)    U+FF10 - U+FF19 are missing their numeric values in all three
+        #       fields.  These are unchanged because it doesn't really cause
+        #       problems for Perl.
+        # 7)    A number of code points, such as controls, don't have their
+        #       Unicode Version 1 Names in this file.  These are unchanged.
+
+        my @corrected_lines = split /\n/, <<'END';
+4E00;<CJK Ideograph, First>;Lo;0;L;;;;;N;;;;;
+9FA5;<CJK Ideograph, Last>;Lo;0;L;;;;;N;;;;;
+E000;<Private Use, First>;Co;0;L;;;;;N;;;;;
+F8FF;<Private Use, Last>;Co;0;L;;;;;N;;;;;
+F900;<CJK Compatibility Ideograph, First>;Lo;0;L;;;;;N;;;;;
+FA2D;<CJK Compatibility Ideograph, Last>;Lo;0;L;;;;;N;;;;;
+END
+
+        my $file = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        #local $to_trace = 1 if main::DEBUG;
+        trace $_ if main::DEBUG && $to_trace;
+
+        # -1 => retain trailing null fields
+        my ($code_point, @fields) = split /\s*;\s*/, $_, -1;
+
+        # At the first place that is wrong in the input, insert all the
+        # corrections, replacing the wrong line.
+        if ($code_point eq '4E00') {
+            my @copy = @corrected_lines;
+            $_ = shift @copy;
+            ($code_point, @fields) = split /\s*;\s*/, $_, -1;
+
+            $file->insert_lines(@copy);
+        }
 
-    my %Script;
-    my $Scripts = Table->New();
 
-    for my $script (sort { $a->[0] <=> $b->[0] } @ScriptInfo)
-    {
-        my ($first, $last, $name) = @$script;
-        $Scripts->Append($first, $name);
+        if ($fields[$NUMERIC] eq '-') {
+            $fields[$NUMERIC] = '-1';  # This is what 2.0 made it.
+        }
+
+        if  ($fields[$PERL_DECOMPOSITION] ne "") {
+
+            # Several entries have this change to superscript 2 or 3 in the
+            # middle.  Convert these to the modern version, which is to use
+            # the actual U+00B2 and U+00B3 (the superscript forms) instead.
+            # So 'HHHH HHHH <+sup> 0033 <-sup> HHHH' becomes
+            # 'HHHH HHHH 00B3 HHHH'.
+            # It turns out that all of these that don't have another
+            # decomposition defined at the beginning of the line have the
+            # <square> decomposition in later releases.
+            if ($code_point ne '00B2' && $code_point ne '00B3') {
+                if  ($fields[$PERL_DECOMPOSITION]
+                                    =~ s/<\+sup> 003([23]) <-sup>/00B$1/)
+                {
+                    if (substr($fields[$PERL_DECOMPOSITION], 0, 1) ne '<') {
+                        $fields[$PERL_DECOMPOSITION] = '<square> '
+                        . $fields[$PERL_DECOMPOSITION];
+                    }
+                }
+            }
+
+            # If is like '<+circled> 0052 <-circled>', convert to
+            # '<circled> 0052'
+            $fields[$PERL_DECOMPOSITION] =~
+                            s/ < \+ ( .*? ) > \s* (.*?) \s* <-\1> /<$1> $2/x;
+
+            # Convert '<join> HHHH HHHH <join>' to '<medial> HHHH HHHH', etc.
+            $fields[$PERL_DECOMPOSITION] =~
+                            s/ <join> \s* (.*?) \s* <no-join> /<final> $1/x
+            or $fields[$PERL_DECOMPOSITION] =~
+                            s/ <join> \s* (.*?) \s* <join> /<medial> $1/x
+            or $fields[$PERL_DECOMPOSITION] =~
+                            s/ <no-join> \s* (.*?) \s* <join> /<initial> $1/x
+            or $fields[$PERL_DECOMPOSITION] =~
+                        s/ <no-join> \s* (.*?) \s* <no-join> /<isolated> $1/x;
+
+            # Convert '<break> HHHH HHHH <break>' to '<break> HHHH', etc.
+            $fields[$PERL_DECOMPOSITION] =~
+                    s/ <(break|no-break)> \s* (.*?) \s* <\1> /<$1> $2/x;
+
+            # Change names to modern form.
+            $fields[$PERL_DECOMPOSITION] =~ s/<font variant>/<font>/g;
+            $fields[$PERL_DECOMPOSITION] =~ s/<no-break>/<noBreak>/g;
+            $fields[$PERL_DECOMPOSITION] =~ s/<circled>/<circle>/g;
+            $fields[$PERL_DECOMPOSITION] =~ s/<break>/<fraction>/g;
+
+            # One entry has weird braces
+            $fields[$PERL_DECOMPOSITION] =~ s/[{}]//g;
+        }
+
+        $_ = join ';', $code_point, @fields;
+        trace $_ if main::DEBUG && $to_trace;
+        return;
+    }
+
+    sub filter_v2_1_5_ucd {
+        # A dozen entries in this 2.1.5 file had the mirrored and numeric
+        # columns swapped;  These all had mirrored be 'N'.  So if the numeric
+        # column appears to be N, swap it back.
 
-        $Script{$name} ||= Table->New(Is    => $name,
-                                      Desc  => "Script '$name'",
-                                      Fuzzy => 1);
-        $Script{$name}->Append($first, $name);
+        my ($code_point, @fields) = split /\s*;\s*/, $_, -1;
+        if ($fields[$NUMERIC] eq 'N') {
+            $fields[$NUMERIC] = $fields[$MIRRORED];
+            $fields[$MIRRORED] = 'N';
+            $_ = join ';', $code_point, @fields;
+        }
+        return;
+    }
+} # End closure for UnicodeData
+
+sub process_NamedSequences {
+    # NamedSequences.txt entries are just added to an array.  Because these
+    # don't look like the other tables, they have their own handler.
+    # An example:
+    # LATIN CAPITAL LETTER A WITH MACRON AND GRAVE;0100 0300
+    #
+    # This just adds the sequence to an array for later handling
+
+    return; # XXX Until charnames catches up
+    my $file = shift;
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
 
-        if ($last) {
-            $Scripts->Extend($last);
-            $Script{$name}->Extend($last);
+    while ($file->next_line) {
+        my ($name, $sequence, @remainder) = split /\s*;\s*/, $_, -1;
+        if (@remainder) {
+            $file->carp_bad_line(
+                "Doesn't look like 'KHMER VOWEL SIGN OM;17BB 17C6'");
+            next;
         }
+        push @named_sequences, "$sequence\t\t$name";
     }
+    return;
+}
 
-    # $Scripts->Write("Scripts.pl");
+{ # Closure
+
+    my $first_range;
+
+    sub  filter_early_ea_lb {
+        # Fixes early EastAsianWidth.txt and LineBreak.txt files.  These had a
+        # third field be the name of the code point, which can be ignored in
+        # most cases.  But it can be meaningful if it marks a range:
+        # 33FE;W;IDEOGRAPHIC TELEGRAPH SYMBOL FOR DAY THIRTY-ONE
+        # 3400;W;<CJK Ideograph Extension A, First>
+        #
+        # We need to see the First in the example above to know it's a range.
+        # They did not use the later range syntaxes.  This routine changes it
+        # to use the modern syntax.
+        # $1 is the Input_file object.
+
+        my @fields = split /\s*;\s*/;
+        if ($fields[2] =~ /^<.*, First>/) {
+            $first_range = $fields[0];
+            $_ = "";
+        }
+        elsif ($fields[2] =~ /^<.*, Last>/) {
+            $_ = $_ = "$first_range..$fields[0]; $fields[1]";
+        }
+        else {
+            undef $first_range;
+            $_ = "$fields[0]; $fields[1]";
+        }
 
-    ## Common is everything not explicitly assigned to a Script
-    ##
-    ##    ***shouldn't this be intersected with \p{Assigned}? ******
-    ##
-    New_Prop(Is => 'Common',
-             $Scripts->Invert,
-             Desc  => 'Pseudo-Script of codepoints not in other Unicode scripts',
-             Fuzzy => 1);
+        return;
+    }
 }
 
-##
-## Given a name like "Close Punctuation", return a regex (that when applied
-## with /i) matches any valid form of that name (e.g. "ClosePunctuation",
-## "Close-Punctuation", etc.)
-##
-## Accept any space, dash, or underbar where in the official name there is
-## space or a dash (or underbar, but there never is).
-##
-##
-sub NameToRegex($)
-{
-    my $Name = shift;
-    $Name =~ s/[- _]/(?:[-_]|\\s+)?/g;
-    return $Name;
+sub filter_old_style_arabic_shaping {
+    # Early versions used a different term for the later one.
+
+    my @fields = split /\s*;\s*/;
+    $fields[3] =~ s/<no shaping>/No_Joining_Group/;
+    $fields[3] =~ s/\s+/_/g;                # Change spaces to underscores
+    $_ = join ';', @fields;
+    return;
 }
 
-##
-## Process Blocks.txt.
-##
-sub Blocks_txt()
-{
-    my $Blocks = Table->New();
-    my %Blocks;
+sub filter_arabic_shaping_line {
+    # ArabicShaping.txt has entries that look like:
+    # 062A; TEH; D; BEH
+    # The field containing 'TEH' is not used.  The next field is Joining_Type
+    # and the last is Joining_Group
+    # This generates two lines to pass on, one for each property on the input
+    # line.
+
+    my $file = shift;
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
 
-    if (not open IN, "Blocks.txt") {
-        die "$0: Blocks.txt: $!\n";
+    my @fields = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
+
+    if (@fields > 4) {
+        $file->carp_bad_line('Extra fields');
+        $_ = "";
+        return;
     }
 
-    while (<IN>)
-    {
-        #next if not /Private Use$/;
-	next if not /^([0-9A-Fa-f]+)\.\.([0-9A-Fa-f]+)\s*;\s*(.+?)\s*$/;
+    $file->insert_adjusted_lines("$fields[0]; Joining_Group; $fields[3]");
+    $_ = "$fields[0]; Joining_Type; $fields[2]";
+
+    return;
+}
 
-	my ($first, $last, $name) = (hex($1), hex($2), $3);
+sub setup_special_casing {
+    # SpecialCasing.txt contains the non-simple case change mappings.  The
+    # simple ones are in UnicodeData.txt, and should already have been read
+    # in.
+    # This routine initializes the full mappings to the simple, then as each
+    # line is processed, it overrides the simple ones.
 
-	$Blocks->Append($first, $name);
+    my $file= shift;
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
 
-        $Blocks{$name} ||= Table->New(In    => $name,
-                                      Desc  => "Block '$name'",
-                                      Fuzzy => 1);
-        $Blocks{$name}->Append($first, $name);
+    # For each of the case change mappings...
+    foreach my $case ('lc', 'tc', 'uc') {
 
-	if ($last and $last != $first) {
-	    $Blocks->Extend($last);
-	    $Blocks{$name}->Extend($last);
-	}
+        # The simple version's name in each mapping merely has an 's' in front
+        # of the full one's
+        my $simple = property_ref('s' . $case);
+        unless (defined $simple && ! $simple->is_empty) {
+            Carp::my_carp_bug("Need to process UnicodeData before SpecialCasing.  Only special casing will be generated.");
+        }
+
+        # Initialize the full case mappings with the simple ones.
+        property_ref($case)->initialize($simple);
     }
-    close IN;
 
-    # $Blocks->Write("Blocks.pl");
+    return;
 }
 
-##
-## Read in the PropList.txt.  It contains extended properties not
-## listed in the UnicodeData.txt, such as 'Other_Alphabetic':
-## alphabetic but not of the general category L; many modifiers
-## belong to this extended property category: while they are not
-## alphabets, they are alphabetic in nature.
-##
-sub PropList_txt()
-{
-    my @PropInfo;
+sub filter_special_casing_line {
+    # Change the format of $_ from SpecialCasing.txt into something that the
+    # generic handler understands.  Each input line contains three case
+    # mappings.  This will generate three lines to pass to the generic handler
+    # for each of those.
+
+    # The input syntax (after stripping comments and trailing white space is
+    # like one of the following (with the final two being entries that we
+    # ignore):
+    # 00DF; 00DF; 0053 0073; 0053 0053; # LATIN SMALL LETTER SHARP S
+    # 03A3; 03C2; 03A3; 03A3; Final_Sigma;
+    # 0307; ; 0307; 0307; tr After_I; # COMBINING DOT ABOVE
+    # Note the trailing semi-colon, unlike many of the input files.  That
+    # means that there will be an extra null field generated by the split
+
+    my $file = shift;
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+    my @fields = split /\s*;\s*/, $_, -1; # -1 => retain trailing null fields
+
+    # field #4 is when this mapping is conditional.  If any of these get
+    # implemented, it would be by hard-coding in the casing functions in the
+    # Perl core, not through tables.  But if there is a new condition we don't
+    # know about, output a warning.  We know about all the conditions through
+    # 5.2
+    if ($fields[4] ne "") {
+        my @conditions = split ' ', $fields[4];
+        if ($conditions[0] ne 'tr'  # We know that these languages have
+                                    # conditions, and some are multiple
+            && $conditions[0] ne 'az'
+            && $conditions[0] ne 'lt'
+
+            # And, we know about a single condition Final_Sigma, but
+            # nothing else.
+            && ($v_version gt v5.2.0
+                && (@conditions > 1 || $conditions[0] ne 'Final_Sigma')))
+        {
+            $file->carp_bad_line("Unknown condition '$fields[4]'.  You should inspect it and either add code to handle it, or add to list of those that are to ignore");
+        }
+        elsif ($conditions[0] ne 'Final_Sigma') {
+
+                # Don't print out a message for Final_Sigma, because we have
+                # hard-coded handling for it.  (But the standard could change
+                # what the rule should be, but it wouldn't show up here
+                # anyway.
 
-    if (not open IN, "PropList.txt") {
-        die "$0: PropList.txt: $!\n";
+                print "# SKIPPING Special Casing: $_\n"
+                                                    if $verbosity >= $VERBOSE;
+        }
+        $_ = "";
+        return;
+    }
+    elsif (@fields > 6 || (@fields == 6 && $fields[5] ne "" )) {
+        $file->carp_bad_line('Extra fields');
+        $_ = "";
+        return;
     }
 
-    while (<IN>)
-    {
-	next unless /^([0-9A-Fa-f]+)(?:\.\.([0-9A-Fa-f]+))?\s*;\s*(.+?)\s*\#/;
+    $_ = "$fields[0]; lc; $fields[1]";
+    $file->insert_adjusted_lines("$fields[0]; tc; $fields[2]");
+    $file->insert_adjusted_lines("$fields[0]; uc; $fields[3]");
+
+    return;
+}
+
+sub filter_old_style_case_folding {
+    # This transforms $_ containing the case folding style of 3.0.1, to 3.1
+    # and later style, then calls the handler for the later style.  Different
+    # letters were used.
+
+    my $file = shift;
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
 
-	# Wait until all the extended properties have been read since
-	# they are not listed in numeric order.
-	push @PropInfo, [ hex($1), hex($2||""), $3 ];
+    my @fields = split /\s*;\s*/;
+    if ($fields[0] =~ /^ 013 [01] $/x) { # The two turkish fields
+        $fields[1] = 'I';
     }
-    close IN;
+    elsif ($fields[1] eq 'L') {
+        $fields[1] = 'C';             # L => C always
+    }
+    elsif ($fields[1] eq 'E') {
+        if ($fields[2] =~ / /) {      # E => C if one code point; F otherwise
+            $fields[1] = 'F'
+        }
+        else {
+            $fields[1] = 'C'
+        }
+    }
+    else {
+        $file->carp_bad_line("Expecting L or E in second field");
+        $_ = "";
+        return;
+    }
+    $_ = join("; ", @fields) . ';';
+    return;
+}
 
-    # Now append the extended properties in their code point order.
-    my $Props = Table->New();
-    my %Prop;
+{ # Closure for case folding
 
-    for my $prop (sort { $a->[0] <=> $b->[0] } @PropInfo)
-    {
-        my ($first, $last, $name) = @$prop;
-        $Props->Append($first, $name);
-
-        $Prop{$name} ||= Table->New(Is    => $name,
-                                    Desc  => "Extended property '$name'",
-                                    Fuzzy => 1);
-        $Prop{$name}->Append($first, $name);
-
-        if ($last) {
-            $Props->Extend($last);
-            $Prop{$name}->Extend($last);
-        }
-    }
-
-    for (keys %Prop) {
-	(my $file = $PA_reverse{$_}) =~ tr/_//d;
-	# XXX I'm assuming that the names from %Prop don't suffer 8.3 clashes.
-	$BaseNames{lc $file}++;
-	$Prop{$_}->Write(
-	    ["lib","gc_sc","$file.pl"],
-	    "Binary property '$_'"
-	);
-    }
-
-    # Alphabetic is L, Nl, and Other_Alphabetic.
-    New_Prop(Is    => 'Alphabetic',
-             Table->Merge($Cat{L}, $Cat{Nl}, $Prop{Other_Alphabetic}),
-             Desc  => '[\p{L}\p{Nl}\p{OtherAlphabetic}]', # canonical names
-             Fuzzy => 1);
-
-    # Lowercase is Ll and Other_Lowercase.
-    New_Prop(Is    => 'Lowercase',
-             Table->Merge($Cat{Ll}, $Prop{Other_Lowercase}),
-             Desc  => '[\p{Ll}\p{OtherLowercase}]', # canonical names
-             Fuzzy => 1);
-
-    # Uppercase is Lu and Other_Uppercase.
-    New_Prop(Is => 'Uppercase',
-             Table->Merge($Cat{Lu}, $Prop{Other_Uppercase}),
-             Desc  => '[\p{Lu}\p{OtherUppercase}]', # canonical names
-             Fuzzy => 1);
-
-    # Math is Sm and Other_Math.
-    New_Prop(Is => 'Math',
-             Table->Merge($Cat{Sm}, $Prop{Other_Math}),
-             Desc  => '[\p{Sm}\p{OtherMath}]', # canonical names
-             Fuzzy => 1);
-
-    # ID_Start is Ll, Lu, Lt, Lm, Lo, Nl, and Other_ID_Start.
-    New_Prop(Is => 'ID_Start',
-             Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl]}, $Prop{Other_ID_Start}),
-             Desc  => '[\p{Ll}\p{Lu}\p{Lt}\p{Lm}\p{Lo}\p{Nl}\p{OtherIDStart}]',
-             Fuzzy => 1);
-
-    # ID_Continue is ID_Start, Mn, Mc, Nd, Pc, and Other_ID_Continue.
-    New_Prop(Is => 'ID_Continue',
-             Table->Merge(@Cat{qw[Ll Lu Lt Lm Lo Nl Mn Mc Nd Pc ]},
-                          @Prop{qw[Other_ID_Start Other_ID_Continue]}),
-             Desc  => '[\p{ID_Start}\p{Mn}\p{Mc}\p{Nd}\p{Pc}\p{OtherIDContinue}]',
-             Fuzzy => 1);
-
-    # Default_Ignorable_Code_Point = Other_Default_Ignorable_Code_Point
-    #                     + Cf + Cc + Cs + Noncharacter + Variation_Selector
-    #                     - WhiteSpace - FFF9..FFFB (Annotation Characters)
-
-    my $Annotation = Table->New();
-    $Annotation->RawAppendRange(0xFFF9, 0xFFFB);
-
-    New_Prop(Is => 'Default_Ignorable_Code_Point',
-             Table->Merge(@Cat{qw[Cf Cc Cs]},
-                          $Prop{Noncharacter_Code_Point},
-                          $Prop{Variation_Selector},
-                          $Prop{Other_Default_Ignorable_Code_Point})
-                  ->Invert
-                  ->Merge($Prop{White_Space}, $Annotation)
-                  ->Invert,
-             Desc  => '(?![\p{WhiteSpace}\x{FFF9}-\x{FFFB}])[\p{Cf}\p{Cc}'.
-                      '\p{Cs}\p{NoncharacterCodePoint}\p{VariationSelector}'.
-                      '\p{OtherDefaultIgnorableCodePoint}]',
-             Fuzzy => 1);
-
-}
-
-
-##
-## These are used in:
-##   MakePropTestScript()
-##   WriteAllMappings()
-## for making the test script.
-##
-my %FuzzyNameToTest;
-my %ExactNameToTest;
-
-
-## This used only for making the test script
-sub GenTests($$$$)
-{
-    my $FH = shift;
-    my $Prop = shift;
-    my $MatchCode = shift;
-    my $FailCode = shift;
-
-    if (defined $MatchCode) {
-        printf $FH qq/Expect(1, "\\x{%04X}", '\\p{$Prop}' );\n/, $MatchCode;
-        printf $FH qq/Expect(0, "\\x{%04X}", '\\p{^$Prop}');\n/, $MatchCode;
-        printf $FH qq/Expect(0, "\\x{%04X}", '\\P{$Prop}' );\n/, $MatchCode;
-        printf $FH qq/Expect(1, "\\x{%04X}", '\\P{^$Prop}');\n/, $MatchCode;
-    }
-    if (defined $FailCode) {
-        printf $FH qq/Expect(0, "\\x{%04X}", '\\p{$Prop}' );\n/, $FailCode;
-        printf $FH qq/Expect(1, "\\x{%04X}", '\\p{^$Prop}');\n/, $FailCode;
-        printf $FH qq/Expect(1, "\\x{%04X}", '\\P{$Prop}' );\n/, $FailCode;
-        printf $FH qq/Expect(0, "\\x{%04X}", '\\P{^$Prop}');\n/, $FailCode;
-    }
-}
-
-## This used only for making the test script
-sub ExpectError($$)
-{
-    my $FH = shift;
-    my $prop = shift;
-
-    print $FH qq/Error('\\p{$prop}');\n/;
-    print $FH qq/Error('\\P{$prop}');\n/;
-}
-
-## This used only for making the test script
-my @GoodSeps = (
-                " ",
-                "-",
-                " \t ",
-                "",
-                "",
-                "_",
-               );
-my @BadSeps = (
-               "--",
-               "__",
-               " _",
-               "/"
-              );
-
-## This used only for making the test script
-sub RandomlyFuzzifyName($;$)
-{
-    my $Name = shift;
-    my $WantError = shift;  ## if true, make an error
+    # Create the map for simple only if are going to output it, for otherwise
+    # it takes no part in anything we do.
+    my $to_output_simple;
 
-    my @parts;
-    for my $part (split /[-\s_]+/, $Name)
-    {
-        if (@parts) {
-            if ($WantError and rand() < 0.3) {
-                push @parts, $BadSeps[rand(@BadSeps)];
-                $WantError = 0;
-            } else {
-                push @parts, $GoodSeps[rand(@GoodSeps)];
+    # These are experimental, perhaps will need these to pass to regcomp.c to
+    # handle the cases where for example the Kelvin sign character folds to k,
+    # and in regcomp, we need to know which of the characters can have a
+    # non-latin1 char fold to it, so it doesn't do the optimizations it might
+    # otherwise.
+    my @latin1_singly_folded;
+    my @latin1_folded;
+
+    sub setup_case_folding($) {
+        # Read in the case foldings in CaseFolding.txt.  This handles both
+        # simple and full case folding.
+
+        $to_output_simple
+                        = property_ref('Simple_Case_Folding')->to_output_map;
+
+        return;
+    }
+
+    sub filter_case_folding_line {
+        # Called for each line in CaseFolding.txt
+        # Input lines look like:
+        # 0041; C; 0061; # LATIN CAPITAL LETTER A
+        # 00DF; F; 0073 0073; # LATIN SMALL LETTER SHARP S
+        # 1E9E; S; 00DF; # LATIN CAPITAL LETTER SHARP S
+        #
+        # 'C' means that folding is the same for both simple and full
+        # 'F' that it is only for full folding
+        # 'S' that it is only for simple folding
+        # 'T' is locale-dependent, and ignored
+        # 'I' is a type of 'F' used in some early releases.
+        # Note the trailing semi-colon, unlike many of the input files.  That
+        # means that there will be an extra null field generated by the split
+        # below, which we ignore and hence is not an error.
+
+        my $file = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my ($range, $type, $map, @remainder) = split /\s*;\s*/, $_, -1;
+        if (@remainder > 1 || (@remainder == 1 && $remainder[0] ne "" )) {
+            $file->carp_bad_line('Extra fields');
+            $_ = "";
+            return;
+        }
+
+        if ($type eq 'T') {   # Skip Turkic case folding, is locale dependent
+            $_ = "";
+            return;
+        }
+
+        # C: complete, F: full, or I: dotted uppercase I -> dotless lowercase
+        # I are all full foldings
+        if ($type eq 'C' || $type eq 'F' || $type eq 'I') {
+            $_ = "$range; Case_Folding; $map";
+        }
+        else {
+            $_ = "";
+            if ($type ne 'S') {
+               $file->carp_bad_line('Expecting C F I S or T in second field');
+               return;
             }
         }
-        my $switch = int rand(4);
-        if ($switch == 0) {
-            push @parts, uc $part;
-        } elsif ($switch == 1) {
-            push @parts, lc $part;
-        } elsif ($switch == 2) {
-            push @parts, ucfirst $part;
-        } else {
-            push @parts, $part;
+
+        # C and S are simple foldings, but simple case folding is not needed
+        # unless we explicitly want its map table output.
+        if ($to_output_simple && $type eq 'C' || $type eq 'S') {
+            $file->insert_adjusted_lines("$range; Simple_Case_Folding; $map");
         }
+
+        # Experimental, see comment above
+        if ($type ne 'S' && hex($range) >= 256) {   # assumes range is 1 point
+            my @folded = split ' ', $map;
+            if (hex $folded[0] < 256 && @folded == 1) {
+                push @latin1_singly_folded, hex $folded[0];
+            }
+            foreach my $folded (@folded) {
+                push @latin1_folded, hex $folded if hex $folded < 256;
+            }
+        }
+
+        return;
     }
-    my $new = join('', @parts);
 
-    if ($WantError) {
-        if (rand() >= 0.5) {
-            $new .= $BadSeps[rand(@BadSeps)];
-        } else {
-            $new = $BadSeps[rand(@BadSeps)] . $new;
+    sub post_fold {
+        # Experimental, see comment above
+        return;
+
+        #local $to_trace = 1 if main::DEBUG;
+        @latin1_singly_folded = uniques(@latin1_singly_folded);
+        @latin1_folded = uniques(@latin1_folded);
+        trace "latin1 single folded:", map { chr $_ } sort { $a <=> $b } @latin1_singly_folded if main::DEBUG && $to_trace;
+        trace "latin1 folded:", map { chr $_ } sort { $a <=> $b } @latin1_folded if main::DEBUG && $to_trace;
+        return;
+    }
+} # End case fold closure
+
+sub filter_jamo_line {
+    # Filter Jamo.txt lines.  This routine mainly is used to populate hashes
+    # from this file that is used in generating the Name property for Jamo
+    # code points.  But, it also is used to convert early versions' syntax
+    # into the modern form.  Here are two examples:
+    # 1100; G   # HANGUL CHOSEONG KIYEOK            # Modern syntax
+    # U+1100; G; HANGUL CHOSEONG KIYEOK             # 2.0 syntax
+    #
+    # The input is $_, the output is $_ filtered.
+
+    my @fields = split /\s*;\s*/, $_, -1;  # -1 => retain trailing null fields
+
+    # Let the caller handle unexpected input.  In earlier versions, there was
+    # a third field which is supposed to be a comment, but did not have a '#'
+    # before it.
+    return if @fields > (($v_version gt v3.0.0) ? 2 : 3);
+
+    $fields[0] =~ s/^U\+//;     # Also, early versions had this extraneous
+                                # beginning.
+
+    # Some 2.1 versions had this wrong.  Causes havoc with the algorithm.
+    $fields[1] = 'R' if $fields[0] eq '1105';
+
+    # Add to structure so can generate Names from it.
+    my $cp = hex $fields[0];
+    my $short_name = $fields[1];
+    $Jamo{$cp} = $short_name;
+    if ($cp <= $LBase + $LCount) {
+        $Jamo_L{$short_name} = $cp - $LBase;
+    }
+    elsif ($cp <= $VBase + $VCount) {
+        $Jamo_V{$short_name} = $cp - $VBase;
+    }
+    elsif ($cp <= $TBase + $TCount) {
+        $Jamo_T{$short_name} = $cp - $TBase;
+    }
+    else {
+        Carp::my_carp_bug("Unexpected Jamo code point in $_");
+    }
+
+
+    # Reassemble using just the first two fields to look like a typical
+    # property file line
+    $_ = "$fields[0]; $fields[1]";
+
+    return;
+}
+
+sub filter_numeric_value_comment_missing_line {
+    # Filters out the extra column in DNumValues.txt of this line
+
+    s/^($missing_defaults_prefix)\s*;\s*/$1/;
+    return;
+}
+
+sub register_fraction($) {
+    # This registers the input rational number so that it can be passed on to
+    # utf8_heavy.pl, both in rational and floating forms.
+
+    my $rational = shift;
+
+    my $float = eval $rational;
+    $nv_floating_to_rational{$float} = $rational;
+    return;
+}
+
+sub filter_numeric_value_line {
+    # DNumValues contains lines of a different syntax than the typical
+    # property file:
+    # 0F33          ; -0.5 ; ; -1/2 # No       TIBETAN DIGIT HALF ZERO
+    #
+    # This routine transforms $_ containing the anomalous syntax to the
+    # typical, by filtering out the extra columns, and convert early version
+    # decimal numbers to strings that look like rational numbers.
+
+    my $file = shift;
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+    # Starting in 5.1, there is a rational field.  Just use that, omitting the
+    # extra columns.  Otherwise convert the decimal number in the second field
+    # to a rational, and omit extraneous columns.
+    my @fields = split /\s*;\s*/, $_, -1;
+    my $rational;
+
+    if ($v_version ge v5.1.0) {
+        if (@fields != 4) {
+            $file->carp_bad_line('Not 4 semi-colon separated fields');
+            $_ = "";
+            return;
         }
+        $rational = $fields[3];
+        $_ = join '; ', @fields[ 0, 3 ];
     }
-    return $new;
+    else {
+
+        # Here, is an older Unicode file, which has decimal numbers instead of
+        # rationals in it.  Use the fraction to calculate the denominator and
+        # convert to rational.
+
+        if (@fields != 2 && @fields != 3) {
+            $file->carp_bad_line('Not 2 or 3 semi-colon separated fields');
+            $_ = "";
+            return;
+        }
+
+        my $codepoints = $fields[0];
+        my $decimal = $fields[1];
+        if ($decimal =~ s/\.0+$//) {
+
+            # Anything ending with a decimal followed by nothing but 0's is an
+            # integer
+            $_ = "$codepoints; $decimal";
+            $rational = $decimal;
+        }
+        else {
+
+            my $denominator;
+            if ($decimal =~ /\.50*$/) {
+                $denominator = 2;
+            }
+
+            # Here have the hardcoded repeating decimals in the fraction, and
+            # the denominator they imply.  There were only a few denominators
+            # in the older Unicode versions of this file which this code
+            # handles, so it is easy to convert them.
+
+            # The 4 is because of a round-off error in the Unicode 3.2 files
+            elsif ($decimal =~ /\.33*[34]$/ || $decimal =~ /\.6+7$/) {
+                $denominator = 3;
+            }
+            elsif ($decimal =~ /\.[27]50*$/) {
+                $denominator = 4;
+            }
+            elsif ($decimal =~ /\.[2468]0*$/) {
+                $denominator = 5;
+            }
+            elsif ($decimal =~ /\.16+7$/ || $decimal =~ /\.83+$/) {
+                $denominator = 6;
+            }
+            elsif ($decimal =~ /\.(12|37|62|87)50*$/) {
+                $denominator = 8;
+            }
+            if ($denominator) {
+                my $sign = ($decimal < 0) ? "-" : "";
+                my $numerator = int((abs($decimal) * $denominator) + .5);
+                $rational = "$sign$numerator/$denominator";
+                $_ = "$codepoints; $rational";
+            }
+            else {
+                $file->carp_bad_line("Can't cope with number '$decimal'.");
+                $_ = "";
+                return;
+            }
+        }
+    }
+
+    register_fraction($rational) if $rational =~ qr{/};
+    return;
 }
 
-## This used only for making the test script
-sub MakePropTestScript()
-{
-    ## this written directly -- it's huge.
-    force_unlink ("TestProp.pl");
-    if (not open OUT, ">TestProp.pl") {
-        die "$0: TestProp.pl: $!\n";
+{ # Closure
+    my %unihan_properties;
+    my $iicore;
+
+
+    sub setup_unihan {
+        # Do any special setup for Unihan properties.
+
+        # This property gives the wrong computed type, so override.
+        my $usource = property_ref('kIRG_USource');
+        $usource->set_type($STRING) if defined $usource;
+
+        # This property is to be considered binary, so change all the values
+        # to Y.
+        $iicore = property_ref('kIICore');
+        if (defined $iicore) {
+            $iicore->add_match_table('Y') if ! defined $iicore->table('Y');
+
+            # We have to change the default map, because the @missing line is
+            # misleading, given that we are treating it as binary.
+            $iicore->set_default_map('N');
+            $iicore->set_type($BINARY);
+        }
+
+        return;
     }
-    print OUT <DATA>;
 
-    while (my ($Name, $Table) = each %ExactNameToTest)
-    {
-        GenTests(*OUT, $Name, $Table->ValidCode, $Table->InvalidCode);
-        ExpectError(*OUT, uc $Name) if uc $Name ne $Name;
-        ExpectError(*OUT, lc $Name) if lc $Name ne $Name;
+    sub filter_unihan_line {
+        # Change unihan db lines to look like the others in the db.  Here is
+        # an input sample:
+        #   U+341C        kCangjie        IEKN
+
+        # Tabs are used instead of semi-colons to separate fields; therefore
+        # they may have semi-colons embedded in them.  Change these to periods
+        # so won't screw up the rest of the code.
+        s/;/./g;
+
+        # Remove lines that don't look like ones we accept.
+        if ($_ !~ /^ [^\t]* \t ( [^\t]* ) /x) {
+            $_ = "";
+            return;
+        }
+
+        # Extract the property, and save a reference to its object.
+        my $property = $1;
+        if (! exists $unihan_properties{$property}) {
+            $unihan_properties{$property} = property_ref($property);
+        }
+
+        # Don't do anything unless the property is one we're handling, which
+        # we determine by seeing if there is an object defined for it or not
+        if (! defined $unihan_properties{$property}) {
+            $_ = "";
+            return;
+        }
+
+        # The iicore property is supposed to be a boolean, so convert to our
+        # standard boolean form.
+        if (defined $iicore && $unihan_properties{$property} == $iicore) {
+            $_ =~ s/$property.*/$property\tY/
+        }
+
+        # Convert the tab separators to our standard semi-colons, and convert
+        # the U+HHHH notation to the rest of the standard's HHHH
+        s/\t/;/g;
+        s/\b U \+ (?= $code_point_re )//xg;
+
+        #local $to_trace = 1 if main::DEBUG;
+        trace $_ if main::DEBUG && $to_trace;
+
+        return;
     }
+}
 
+sub filter_blocks_lines {
+    # In the Blocks.txt file, the names of the blocks don't quite match the
+    # names given in PropertyValueAliases.txt, so this changes them so they
+    # do match:  Blanks and hyphens are changed into underscores.  Also makes
+    # early release versions look like later ones
+    #
+    # $_ is transformed to the correct value.
 
-    while (my ($Name, $Table) = each %FuzzyNameToTest)
-    {
-        my $Orig  = $CanonicalToOrig{$Name};
-        my %Names = (
-                     $Name => 1,
-                     $Orig => 1,
-                     RandomlyFuzzifyName($Orig) => 1
-                    );
+    my $file = shift;
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
 
-        for my $N (keys %Names) {
-            GenTests(*OUT, $N, $Table->ValidCode, $Table->InvalidCode);
+    if ($v_version lt v3.2.0) {
+        if (/FEFF.*Specials/) { # Bug in old versions: line wrongly inserted
+            $_ = "";
+            return;
         }
 
-        ExpectError(*OUT, RandomlyFuzzifyName($Orig, 'ERROR'));
+        # Old versions used a different syntax to mark the range.
+        $_ =~ s/;\s+/../ if $v_version lt v3.1.0;
+    }
+
+    my @fields = split /\s*;\s*/, $_, -1;
+    if (@fields != 2) {
+        $file->carp_bad_line("Expecting exactly two fields");
+        $_ = "";
+        return;
     }
 
-    print OUT "Finished();\n";
-    close OUT;
+    # Change hyphens and blanks in the block name field only
+    $fields[1] =~ s/[ -]/_/g;
+    $fields[1] =~ s/_ ( [a-z] ) /_\u$1/g;   # Capitalize first letter of word
+
+    $_ = join("; ", @fields);
+    return;
 }
 
+{ # Closure
+    my $current_property;
+
+    sub filter_old_style_proplist {
+        # PropList.txt has been in Unicode since version 2.0.  Until 3.1, it
+        # was in a completely different syntax.  Ken Whistler of Unicode says
+        # that it was something he used as an aid for his own purposes, but
+        # was never an official part of the standard.  However, comments in
+        # DAge.txt indicate that non-character code points were available in
+        # the UCD as of 3.1.  It is unclear to me (khw) how they could be
+        # there except through this file (but on the other hand, they first
+        # appeared there in 3.0.1), so maybe it was part of the UCD, and maybe
+        # not.  But the claim is that it was published as an aid to others who
+        # might want some more information than was given in the official UCD
+        # of the time.  Many of the properties in it were incorporated into
+        # the later PropList.txt, but some were not.  This program uses this
+        # early file to generate property tables that are otherwise not
+        # accessible in the early UCD's, and most were probably not really
+        # official at that time, so one could argue that it should be ignored,
+        # and you can easily modify things to skip this.  And there are bugs
+        # in this file in various versions.  (For example, the 2.1.9 version
+        # removes from Alphabetic the CJK range starting at 4E00, and they
+        # weren't added back in until 3.1.0.)  Many of this file's properties
+        # were later sanctioned, so this code generates tables for those
+        # properties that aren't otherwise in the UCD of the time but
+        # eventually did become official, and throws away the rest.  Here is a
+        # list of all the ones that are thrown away:
+        #   Bidi=*                       duplicates UnicodeData.txt
+        #   Combining                    never made into official property;
+        #                                is \P{ccc=0}
+        #   Composite                    never made into official property.
+        #   Currency Symbol              duplicates UnicodeData.txt: gc=sc
+        #   Decimal Digit                duplicates UnicodeData.txt: gc=nd
+        #   Delimiter                    never made into official property;
+        #                                removed in 3.0.1
+        #   Format Control               never made into official property;
+        #                                similar to gc=cf
+        #   High Surrogate               duplicates Blocks.txt
+        #   Ignorable Control            never made into official property;
+        #                                similar to di=y
+        #   ISO Control                  duplicates UnicodeData.txt: gc=cc
+        #   Left of Pair                 never made into official property;
+        #   Line Separator               duplicates UnicodeData.txt: gc=zl
+        #   Low Surrogate                duplicates Blocks.txt
+        #   Non-break                    was actually listed as a property
+        #                                in 3.2, but without any code
+        #                                points.  Unicode denies that this
+        #                                was ever an official property
+        #   Non-spacing                  duplicate UnicodeData.txt: gc=mn
+        #   Numeric                      duplicates UnicodeData.txt: gc=cc
+        #   Paired Punctuation           never made into official property;
+        #                                appears to be gc=ps + gc=pe
+        #   Paragraph Separator          duplicates UnicodeData.txt: gc=cc
+        #   Private Use                  duplicates UnicodeData.txt: gc=co
+        #   Private Use High Surrogate   duplicates Blocks.txt
+        #   Punctuation                  duplicates UnicodeData.txt: gc=p
+        #   Space                        different definition than eventual
+        #                                one.
+        #   Titlecase                    duplicates UnicodeData.txt: gc=lt
+        #   Unassigned Code Value        duplicates UnicodeData.txt: gc=cc
+        #   Zero-width                   never made into offical property;
+        #                                subset of gc=cf
+        # Most of the properties have the same names in this file as in later
+        # versions, but a couple do not.
+        #
+        # This subroutine filters $_, converting it from the old style into
+        # the new style.  Here's a sample of the old-style
+        #
+        #   *******************************************
+        #
+        #   Property dump for: 0x100000A0 (Join Control)
+        #
+        #   200C..200D  (2 chars)
+        #
+        # In the example, the property is "Join Control".  It is kept in this
+        # closure between calls to the subroutine.  The numbers beginning with
+        # 0x were internal to Ken's program that generated this file.
+
+        # If this line contains the property name, extract it.
+        if (/^Property dump for: [^(]*\((.*)\)/) {
+            $_ = $1;
+
+            # Convert white space to underscores.
+            s/ /_/g;
+
+            # Convert the few properties that don't have the same name as
+            # their modern counterparts
+            s/Identifier_Part/ID_Continue/
+            or s/Not_a_Character/NChar/;
+
+            # If the name matches an existing property, use it.
+            if (defined property_ref($_)) {
+                trace "new property=", $_ if main::DEBUG && $to_trace;
+                $current_property = $_;
+            }
+            else {        # Otherwise discard it
+                trace "rejected property=", $_ if main::DEBUG && $to_trace;
+                undef $current_property;
+            }
+            $_ = "";    # The property is saved for the next lines of the
+                        # file, but this defining line is of no further use,
+                        # so clear it so that the caller won't process it
+                        # further.
+        }
+        elsif (! defined $current_property || $_ !~ /^$code_point_re/) {
 
-##
-## These are used only in:
-##   RegisterFileForName()
-##   WriteAllMappings()
-##
-my %Exact;      ## will become %utf8::Exact;
-my %Canonical;  ## will become %utf8::Canonical;
-my %CaComment;  ## Comment for %Canonical entry of same key
+            # Here, the input line isn't a header defining a property for the
+            # following section, and either we aren't in such a section, or
+            # the line doesn't look like one that defines the code points in
+            # such a section.  Ignore this line.
+            $_ = "";
+        }
+        else {
 
-##
-## Given info about a name and a datafile that it should be associated with,
-## register that assocation in %Exact and %Canonical.
-sub RegisterFileForName($$$$)
-{
-    my $Type     = shift;
-    my $Name     = shift;
-    my $IsFuzzy  = shift;
-    my $filename = shift;
+            # Here, we have a line defining the code points for the current
+            # stashed property.  Anything starting with the first blank is
+            # extraneous.  Otherwise, it should look like a normal range to
+            # the caller.  Append the property name so that it looks just like
+            # a modern PropList entry.
 
-    ##
-    ## Now in details for the mapping. $Type eq 'Is' has the
-    ## Is removed, as it will be removed in utf8_heavy when this
-    ## data is being checked. In keeps its "In", but a second
-    ## sans-In record is written if it doesn't conflict with
-    ## anything already there.
-    ##
-    if (not $IsFuzzy)
-    {
-        if ($Type eq 'Is') {
-            die "oops[$Name]" if $Exact{$Name};
-            $Exact{$Name} = $filename;
-        } else {
-            die "oops[$Type$Name]" if $Exact{"$Type$Name"};
-            $Exact{"$Type$Name"} = $filename;
-            $Exact{$Name} = $filename if not $Exact{$Name};
+            $_ =~ s/\s.*//;
+            $_ .= "; $current_property";
         }
+        trace $_ if main::DEBUG && $to_trace;
+        return;
     }
-    else
-    {
-        my $CName = lc $Name;
-        if ($Type eq 'Is') {
-            die "oops[$CName]" if $Canonical{$CName};
-            $Canonical{$CName} = $filename;
-            $CaComment{$CName} = $Name if $Name =~ tr/A-Z// >= 2;
-        } else {
-            die "oops[$Type$CName]" if $Canonical{lc "$Type$CName"};
-            $Canonical{lc "$Type$CName"} = $filename;
-            $CaComment{lc "$Type$CName"} = "$Type$Name";
-            if (not $Canonical{$CName}) {
-                $Canonical{$CName} = $filename;
-                $CaComment{$CName} = "$Type$Name";
-            }
-        }
-    }
-}
-
-##
-## Writes the info accumulated in
-##
-##       %TableInfo;
-##       %FuzzyNames;
-##       %AliasInfo;
-##
-##
-sub WriteAllMappings()
-{
-    my @MAP;
-
-    ## 'Is' *MUST* come first, so its names have precidence over 'In's
-    for my $Type ('Is', 'In')
+} # End closure for old style proplist
+
+sub filter_old_style_normalization_lines {
+    # For early releases of Unicode, the lines were like:
+    #        74..2A76    ; NFKD_NO
+    # For later releases this became:
+    #        74..2A76    ; NFKD_QC; N
+    # Filter $_ to look like those in later releases.
+    # Similarly for MAYBEs
+
+    s/ _NO \b /_QC; N/x || s/ _MAYBE \b /_QC; M/x;
+
+    # Also, the property FC_NFKC was abbreviated to FNC
+    s/FNC/FC_NFKC/;
+    return;
+}
+
+sub finish_Unicode() {
+    # This routine should be called after all the Unicode files have been read
+    # in.  It:
+    # 1) Adds the mappings for code points missing from the files which have
+    #    defaults specified for them.
+    # 2) At this this point all mappings are known, so it computes the type of
+    #    each property whose type hasn't been determined yet.
+    # 3) Calculates all the regular expression match tables based on the
+    #    mappings.
+    # 3) Calculates and adds the tables which are defined by Unicode, but
+    #    which aren't derived by them
+
+    # For each property, fill in any missing mappings, and calculate the re
+    # match tables.  If a property has more than one missing mapping, the
+    # default is a reference to a data structure, and requires data from other
+    # properties to resolve.  The sort is used to cause these to be processed
+    # last, after all the other properties have been calculated.
+    # (Fortunately, the missing properties so far don't depend on each other.)
+    foreach my $property
+        (sort { (defined $a->default_map && ref $a->default_map) ? 1 : -1 }
+        property_ref('*'))
     {
-        my %RawNameToFile; ## a per-$Type cache
+        # $perl has been defined, but isn't one of the Unicode properties that
+        # need to be finished up.
+        next if $property == $perl;
+
+        # Handle the properties that have more than one possible default
+        if (ref $property->default_map) {
+            my $default_map = $property->default_map;
+
+            # These properties have stored in the default_map:
+            # One or more of:
+            #   1)  A default map which applies to all code points in a
+            #       certain class
+            #   2)  an expression which will evaluate to the list of code
+            #       points in that class
+            # And
+            #   3) the default map which applies to every other missing code
+            #      point.
+            #
+            # Go through each list.
+            while (my ($default, $eval) = $default_map->get_next_defaults) {
+
+                # Get the class list, and intersect it with all the so-far
+                # unspecified code points yielding all the code points
+                # in the class that haven't been specified.
+                my $list = eval $eval;
+                if ($@) {
+                    Carp::my_carp("Can't set some defaults for missing code points for $property because eval '$eval' failed with '$@'");
+                    last;
+                }
 
-        for my $Name (sort {length $a <=> length $b} keys %{$TableInfo{$Type}})
-        {
-            ## Note: $Name is already canonical
-            my $Table   = $TableInfo{$Type}->{$Name};
-            my $IsFuzzy = $FuzzyNames{$Type}->{$Name};
+                # Narrow down the list to just those code points we don't have
+                # maps for yet.
+                $list = $list & $property->inverse_list;
 
-            ## Need an 8.3 safe filename (which means "an 8 safe" $filename)
-            my $filename;
-            {
-                ## 'Is' items lose 'Is' from the basename.
-                $filename = $Type eq 'Is' ?
-		    ($PVA_reverse{sc}{$Name} || $Name) :
-		    "$Type$Name";
-
-                $filename =~ s/[^\w_]+/_/g; # "L&" -> "L_"
-                substr($filename, 8) = '' if length($filename) > 8;
-
-                ##
-                ## Make sure the basename doesn't conflict with something we
-                ## might have already written. If we have, say,
-                ##     InGreekExtended1
-                ##     InGreekExtended2
-                ## they become
-                ##     InGreekE
-                ##     InGreek2
-                ##
-                while (my $num = $BaseNames{lc $filename}++)
-                {
-                    $num++; ## so basenames with numbers start with '2', which
-                            ## just looks more natural.
-                    ## Want to append $num, but if it'll make the basename longer
-                    ## than 8 characters, pre-truncate $filename so that the result
-                    ## is acceptable.
-                    my $delta = length($filename) + length($num) - 8;
-                    if ($delta > 0) {
-                        substr($filename, -$delta) = $num;
-                    } else {
-                        $filename .= $num;
-                    }
+                # Add mappings to the property for each code point in the list
+                foreach my $range ($list->ranges) {
+                    $property->add_map($range->start, $range->end, $default);
                 }
-            };
+            }
 
-            ##
-            ## Construct a nice comment to add to the file, and build data
-            ## for the "./Properties" file along the way.
-            ##
-            my $Comment;
-            {
-                my $Desc = $TableDesc{$Type}->{$Name} || "";
-                ## get list of names this table is reference by
-                my @Supported = $Name;
-                while (my ($Orig, $Alias) = each %{ $AliasInfo{$Type} })
+            # All remaining code points have the other mapping.  Set that up
+            # so the normal single-default mapping code will work on them
+            $property->set_default_map($default_map->other_default);
+
+            # And fall through to do that
+        }
+
+        # We should have enough data now to compute the type of the property.
+        $property->compute_type;
+        my $property_type = $property->type;
+
+        next if ! $property->to_create_match_tables;
+
+        # Here want to create match tables for this property
+
+        # The Unicode db always (so far, and they claim into the future) have
+        # the default for missing entries in binary properties be 'N' (unless
+        # there is a '@missing' line that specifies otherwise)
+        if ($property_type == $BINARY && ! defined $property->default_map) {
+            $property->set_default_map('N');
+        }
+
+        # Add any remaining code points to the mapping, using the default for
+        # missing code points
+        if (defined (my $default_map = $property->default_map)) {
+            foreach my $range ($property->inverse_list->ranges) {
+                $property->add_map($range->start, $range->end, $default_map);
+            }
+
+            # Make sure there is a match table for the default
+            if (! defined $property->table($default_map)) {
+                $property->add_match_table($default_map);
+            }
+        }
+
+        # Have all we need to populate the match tables.
+        my $property_name = $property->name;
+        foreach my $range ($property->ranges) {
+            my $map = $range->value;
+            my $table = property_ref($property_name)->table($map);
+            if (! defined $table) {
+
+                # Integral and rational property values are not necessarily
+                # defined in PropValueAliases, but all other ones should be,
+                # starting in 5.1
+                if ($v_version ge v5.1.0
+                    && $map !~ /^ -? \d+ ( \/ \d+ )? $/x)
                 {
-                    if ($Orig eq $Name) {
-                        push @Supported, $Alias;
+                    Carp::my_carp("Table '$property_name=$map' should have been defined.  Defining it now.")
+                }
+                $table = property_ref($property_name)->add_match_table($map);
+            }
+
+            $table->add_range($range->start, $range->end);
+        }
+
+        # And add the Is_ prefix synonyms for Perl 5.6 compatibility, in which
+        # all properties have this optional prefix.  These do not get a
+        # separate entry in the pod file, because are covered by a wild-card
+        # entry
+        foreach my $alias ($property->aliases) {
+            my $Is_name = 'Is_' . $alias->name;
+            if (! defined (my $pre_existing = property_ref($Is_name))) {
+                $property->add_alias($Is_name,
+                                     Pod_Entry => 0,
+                                     Status => $alias->status,
+                                     Externally_Ok => 0);
+            }
+            else {
+
+                # It seemed too much work to add in these warnings when it
+                # appears that Unicode has made a decision never to begin a
+                # property name with 'Is_', so this shouldn't happen, but just
+                # in case, it is a warning.
+                Carp::my_carp(<<END
+There is already an alias named $Is_name (from " . $pre_existing . "), so not
+creating this alias for $property.  The generated table and pod files do not
+warn users of this conflict.
+END
+                );
+                $has_Is_conflicts++;
+            }
+        } # End of loop through aliases for this property
+    } # End of loop through all Unicode properties.
+
+    # Fill in the mappings that Unicode doesn't completely furnish.  First the
+    # single letter major general categories.  If Unicode were to start
+    # delivering the values, this would be redundant, but better that than to
+    # try to figure out if should skip and not get it right.  Ths could happen
+    # if a new major category were to be introduced, and the hard-coded test
+    # wouldn't know about it.
+    # This routine depends on the standard names for the general categories
+    # being what it thinks they are, like 'Cn'.  The major categories are the
+    # union of all the general category tables which have the same first
+    # letters. eg. L = Lu + Lt + Ll + Lo + Lm
+    foreach my $minor_table ($gc->tables) {
+        my $minor_name = $minor_table->name;
+        next if length $minor_name == 1;
+        if (length $minor_name != 2) {
+            Carp::my_carp_bug("Unexpected general category '$minor_name'.  Skipped.");
+            next;
+        }
+
+        my $major_name = uc(substr($minor_name, 0, 1));
+        my $major_table = $gc->table($major_name);
+        $major_table += $minor_table;
+    }
+
+    # LC is Ll, Lu, and Lt.  (used to be L& or L_, but PropValueAliases.txt
+    # defines it as LC)
+    my $LC = $gc->table('LC');
+    $LC->add_alias('L_', Status => $DISCOURAGED);   # For backwards...
+    $LC->add_alias('L&', Status => $DISCOURAGED);   # compatibility.
+
+
+    if ($LC->is_empty) { # Assume if not empty that Unicode has started to
+                         # deliver the correct values in it
+        $LC->initialize($gc->table('Ll') + $gc->table('Lu'));
+
+        # Lt not in release 1.
+        $LC += $gc->table('Lt') if defined $gc->table('Lt');
+    }
+    $LC->add_description('[\p{Ll}\p{Lu}\p{Lt}]');
+
+    my $Cs = $gc->table('Cs');
+    if (defined $Cs) {
+        $Cs->add_note('Mostly not usable in Perl.');
+        $Cs->add_comment(join_lines(<<END
+Surrogates are used exclusively for I/O in UTF-16, and should not appear in
+Unicode text, and hence their use will generate (usually fatal) messages
+END
+        ));
+    }
+
+
+    # Folding information was introduced later into Unicode data.  To get
+    # Perl's case ignore (/i) to work at all in releases that don't have
+    # folding, use the best available alternative, which is lower casing.
+    my $fold = property_ref('Simple_Case_Folding');
+    if ($fold->is_empty) {
+        $fold->initialize(property_ref('Simple_Lowercase_Mapping'));
+        $fold->add_note(join_lines(<<END
+WARNING: This table uses lower case as a substitute for missing fold
+information
+END
+        ));
+    }
+
+    # Multiple-character mapping was introduced later into Unicode data.  If
+    # missing, use the single-characters maps as best available alternative
+    foreach my $map (qw {   Uppercase_Mapping
+                            Lowercase_Mapping
+                            Titlecase_Mapping
+                            Case_Folding
+                        } ) {
+        my $full = property_ref($map);
+        if ($full->is_empty) {
+            my $simple = property_ref('Simple_' . $map);
+            $full->initialize($simple);
+            $full->add_comment($simple->comment) if ($simple->comment);
+            $full->add_note(join_lines(<<END
+WARNING: This table uses simple mapping (single-character only) as a
+substitute for missing multiple-character information
+END
+            ));
+        }
+    }
+    return
+}
+
+sub compile_perl() {
+    # Create perl-defined tables.  Almost all are part of the pseudo-property
+    # named 'perl' internally to this program.  Many of these are recommended
+    # in UTS#18 "Unicode Regular Expressions", and their derivations are based
+    # on those found there.
+    # Almost all of these are equivalent to some Unicode property.
+    # A number of these properties have equivalents restricted to the ASCII
+    # range, with their names prefaced by 'Posix', to signify that these match
+    # what the Posix standard says they should match.  A couple are
+    # effectively this, but the name doesn't have 'Posix' in it because there
+    # just isn't any Posix equivalent.
+
+    # 'Any' is all code points.  As an error check, instead of just setting it
+    # to be that, construct it to be the union of all the major categories
+    my $Any = $perl->add_match_table('Any',
+            Description  => "[\\x{0000}-\\x{$LAST_UNICODE_CODEPOINT_STRING}]",
+            Matches_All => 1);
+
+    foreach my $major_table ($gc->tables) {
+
+        # Major categories are the ones with single letter names.
+        next if length($major_table->name) != 1;
+
+        $Any += $major_table;
+    }
+
+    if ($Any->max != $LAST_UNICODE_CODEPOINT) {
+        Carp::my_carp_bug("Generated highest code point ("
+           . sprintf("%X", $Any->max)
+           . ") doesn't match expected value $LAST_UNICODE_CODEPOINT_STRING.")
+    }
+    if ($Any->range_count != 1 || $Any->min != 0) {
+     Carp::my_carp_bug("Generated table 'Any' doesn't match all code points.")
+    }
+
+    $Any->add_alias('All');
+
+    # Assigned is the opposite of gc=unassigned
+    my $Assigned = $perl->add_match_table('Assigned',
+                                Description  => "All assigned code points",
+                                Initialize => ~ $gc->table('Unassigned'),
+                                );
+
+    # Our internal-only property should be treated as more than just a
+    # synonym.
+    $perl->add_match_table('_CombAbove')
+            ->set_equivalent_to(property_ref('ccc')->table('Above'),
+                                                                Related => 1);
+
+    my $ASCII = $perl->add_match_table('ASCII', Description => '[[:ASCII:]]');
+    if (defined $block) {   # This is equivalent to the block if have it.
+        my $Unicode_ASCII = $block->table('Basic_Latin');
+        if (defined $Unicode_ASCII && ! $Unicode_ASCII->is_empty) {
+            $ASCII->set_equivalent_to($Unicode_ASCII, Related => 1);
+        }
+    }
+
+    # Very early releases didn't have blocks, so initialize ASCII ourselves if
+    # necessary
+    if ($ASCII->is_empty) {
+        $ASCII->initialize([ 0..127 ]);
+    }
+
+    # A number of the Perl synonyms have a restricted-range synonym whose name
+    # begins with Posix.  This hash gets filled in with them, so that they can
+    # be populated in a small loop.
+    my %posix_equivalent;
+
+    # Get the best available case definitions.  Early Unicode versions didn't
+    # have Uppercase and Lowercase defined, so use the general category
+    # instead for them.
+    my $Lower = $perl->add_match_table('Lower');
+    my $Unicode_Lower = property_ref('Lowercase');
+    if (defined $Unicode_Lower && ! $Unicode_Lower->is_empty) {
+        $Lower->set_equivalent_to($Unicode_Lower->table('Y'), Related => 1);
+    }
+    else {
+        $Lower->set_equivalent_to($gc->table('Lowercase_Letter'),
+                                                                Related => 1);
+    }
+    $posix_equivalent{'Lower'} = $Lower;
+
+    my $Upper = $perl->add_match_table('Upper');
+    my $Unicode_Upper = property_ref('Uppercase');
+    if (defined $Unicode_Upper && ! $Unicode_Upper->is_empty) {
+        $Upper->set_equivalent_to($Unicode_Upper->table('Y'), Related => 1);
+    }
+    else {
+        $Upper->set_equivalent_to($gc->table('Uppercase_Letter'),
+                                                                Related => 1);
+    }
+    $posix_equivalent{'Upper'} = $Upper;
+
+    # Earliest releases didn't have title case.  Initialize it to empty if not
+    # otherwise present
+    my $Title = $perl->add_match_table('Title');
+    my $lt = $gc->table('Lt');
+    if (defined $lt) {
+        $Title->set_equivalent_to($lt, Related => 1);
+    }
+
+    # If this Unicode version doesn't have Cased, set up our own.  From
+    # Unicode 5.1: Definition D120: A character C is defined to be cased if
+    # and only if C has the Lowercase or Uppercase property or has a
+    # General_Category value of Titlecase_Letter.
+    unless (defined property_ref('Cased')) {
+        my $cased = $perl->add_match_table('Cased',
+                        Initialize => $Lower + $Upper + $Title,
+                        Description => 'Uppercase or Lowercase or Titlecase',
+                        );
+    }
+
+    # Similarly, set up our own Case_Ignorable property if this Unicode
+    # version doesn't have it.  From Unicode 5.1: Definition D121: A character
+    # C is defined to be case-ignorable if C has the value MidLetter or the
+    # value MidNumLet for the Word_Break property or its General_Category is
+    # one of Nonspacing_Mark (Mn), Enclosing_Mark (Me), Format (Cf),
+    # Modifier_Letter (Lm), or Modifier_Symbol (Sk).
+
+    # Perl has long had an internal-only alias for this property.
+    my $perl_case_ignorable = $perl->add_match_table('_Case_Ignorable');
+    my $case_ignorable = property_ref('Case_Ignorable');
+    if (defined $case_ignorable && ! $case_ignorable->is_empty) {
+        $perl_case_ignorable->set_equivalent_to($case_ignorable->table('Y'),
+                                                                Related => 1);
+    }
+    else {
+
+        $perl_case_ignorable->initialize($gc->table('Mn') + $gc->table('Lm'));
+
+        # The following three properties are not in early releases
+        $perl_case_ignorable += $gc->table('Me') if defined $gc->table('Me');
+        $perl_case_ignorable += $gc->table('Cf') if defined $gc->table('Cf');
+        $perl_case_ignorable += $gc->table('Sk') if defined $gc->table('Sk');
+
+        # For versions 4.1 - 5.0, there is no MidNumLet property, and
+        # correspondingly the case-ignorable definition lacks that one.  For
+        # 4.0, it appears that it was meant to be the same definition, but was
+        # inadvertently omitted from the standard's text, so add it if the
+        # property actually is there
+        my $wb = property_ref('Word_Break');
+        if (defined $wb) {
+            my $midlet = $wb->table('MidLetter');
+            $perl_case_ignorable += $midlet if defined $midlet;
+            my $midnumlet = $wb->table('MidNumLet');
+            $perl_case_ignorable += $midnumlet if defined $midnumlet;
+        }
+        else {
+
+            # In earlier versions of the standard, instead of the above two
+            # properties , just the following characters were used:
+            $perl_case_ignorable +=  0x0027  # APOSTROPHE
+                                +   0x00AD  # SOFT HYPHEN (SHY)
+                                +   0x2019; # RIGHT SINGLE QUOTATION MARK
+        }
+    }
+
+    # The remaining perl defined tables are mostly based on Unicode TR 18,
+    # "Annex C: Compatibility Properties".  All of these have two versions,
+    # one whose name generally begins with Posix that is posix-compliant, and
+    # one that matches Unicode characters beyond the Posix, ASCII range
+
+    my $Alpha = $perl->add_match_table('Alpha',
+                        Description => '[[:Alpha:]] extended beyond ASCII');
+
+    # Alphabetic was not present in early releases
+    my $Alphabetic = property_ref('Alphabetic');
+    if (defined $Alphabetic && ! $Alphabetic->is_empty) {
+        $Alpha->set_equivalent_to($Alphabetic->table('Y'), Related => 1);
+    }
+    else {
+
+        # For early releases, we don't get it exactly right.  The below
+        # includes more than it should, which in 5.2 terms is: L + Nl +
+        # Other_Alphabetic.  Other_Alphabetic contains many characters from
+        # Mn and Mc.  It's better to match more than we should, than less than
+        # we should.
+        $Alpha->initialize($gc->table('Letter')
+                            + $gc->table('Mn')
+                            + $gc->table('Mc'));
+        $Alpha += $gc->table('Nl') if defined $gc->table('Nl');
+    }
+    $posix_equivalent{'Alpha'} = $Alpha;
+
+    my $Alnum = $perl->add_match_table('Alnum',
+                        Description => "[[:Alnum:]] extended beyond ASCII",
+                        Initialize => $Alpha + $gc->table('Decimal_Number'),
+                        );
+    $posix_equivalent{'Alnum'} = $Alnum;
+
+    my $Word = $perl->add_match_table('Word',
+                                Description => '\w, including beyond ASCII',
+                                Initialize => $Alnum + $gc->table('Mark'),
+                                );
+    my $Pc = $gc->table('Connector_Punctuation'); # 'Pc' Not in release 1
+    $Word += $Pc if defined $Pc;
+
+    # There is no [[:Word:]], so the name doesn't begin with Posix.
+    $perl->add_match_table('PerlWord',
+                    Description => '\w, restricted to ASCII = [A-Za-z0-9_]',
+                    Initialize => $Word & $ASCII,
+                    );
+
+    my $Blank = $perl->add_match_table('Blank',
+                                Description => '\h, Horizontal white space',
+
+                                # 200B is Zero Width Space which is for line
+                                # break control, and was listed as
+                                # Space_Separator in early releases
+                                Initialize => $gc->table('Space_Separator')
+                                            +   0x0009  # TAB
+                                            -   0x200B, # ZWSP
+                                );
+    $Blank->add_alias('HorizSpace');        # Another name for it.
+    $posix_equivalent{'Blank'} = $Blank;
+
+    my $VertSpace = $perl->add_match_table('VertSpace',
+                            Description => '\v',
+                            Initialize => $gc->table('Line_Separator')
+                                        + $gc->table('Paragraph_Separator')
+                                        + 0x000A  # LINE FEED
+                                        + 0x000B  # VERTICAL TAB
+                                        + 0x000C  # FORM FEED
+                                        + 0x000D  # CARRIAGE RETURN
+                                        + 0x0085, # NEL
+                            );
+    # No Posix equivalent for vertical space
+
+    my $Space = $perl->add_match_table('Space',
+        Description => '\s including beyond ASCII plus vertical tab = [[:Space:]]',
+        Initialize => $Blank + $VertSpace,
+    );
+    $posix_equivalent{'Space'} = $Space;
+
+    # Perl's traditional space doesn't include Vertical Tab
+    my $SpacePerl = $perl->add_match_table('SpacePerl',
+                                  Description => '\s, including beyond ASCII',
+                                  Initialize => $Space - 0x000B,
+                                );
+    $perl->add_match_table('PerlSpace',
+                            Description => '\s, restricted to ASCII',
+                            Initialize => $SpacePerl & $ASCII,
+                            );
+
+    my $Cntrl = $perl->add_match_table('Cntrl',
+                        Description => "[[:Cntrl:]] extended beyond ASCII");
+    $Cntrl->set_equivalent_to($gc->table('Cc'), Related => 1);
+    $posix_equivalent{'Cntrl'} = $Cntrl;
+
+    # $controls is a temporary used to construct Graph.
+    my $controls = Range_List->new(Initialize => $gc->table('Unassigned')
+                                                + $gc->table('Control'));
+    # Cs not in release 1
+    $controls += $gc->table('Surrogate') if defined $gc->table('Surrogate');
+
+    # Graph is  ~space &  ~(Cc|Cs|Cn) = ~(space + $controls)
+    my $Graph = $perl->add_match_table('Graph',
+                        Description => "[[:Graph:]] extended beyond ASCII",
+                        Initialize => ~ ($Space + $controls),
+                        );
+    $posix_equivalent{'Graph'} = $Graph;
+
+    my $Print = $perl->add_match_table('Print',
+                        Description => "[[:Print:]] extended beyond ASCII",
+                        Initialize => $Space + $Graph - $gc->table('Control'),
+                        );
+    $posix_equivalent{'Print'} = $Print;
+
+    my $Punct = $perl->add_match_table('Punct');
+    $Punct->set_equivalent_to($gc->table('Punctuation'), Related => 1);
+
+    # \p{punct} doesn't include the symbols, which posix does
+    $perl->add_match_table('PosixPunct',
+                            Description => "[[:Punct:]]",
+                            Initialize => $ASCII & ($gc->table('Punctuation')
+                                                    + $gc->table('Symbol')),
+                            );
+
+    my $Digit = $perl->add_match_table('Digit',
+                            Description => '\d, extended beyond just [0-9]');
+    $Digit->set_equivalent_to($gc->table('Decimal_Number'), Related => 1);
+    $posix_equivalent{'Digit'} = $Digit;
+
+    # AHex was not present in early releases
+    my $Xdigit = $perl->add_match_table('XDigit',
+                                        Description => '[0-9A-Fa-f]');
+    my $AHex = property_ref('ASCII_Hex_Digit');
+    if (defined $AHex && ! $AHex->is_empty) {
+        $Xdigit->set_equivalent_to($AHex->table('Y'), Related => 1);
+    }
+    else {
+        # (Have to use hex because could be running on an non-ASCII machine,
+        # and we want the Unicode (ASCII) values)
+        $Xdigit->initialize([ 0x30..0x39, 0x41..0x46, 0x61..0x66 ]);
+    }
+
+    # Now, add the ASCII-restricted tables that get uniform treatment
+    while (my ($name, $table) = each %posix_equivalent) {
+        $perl->add_match_table("Posix$name",
+                                Description => "[[:$name:]]",
+                                Initialize => $table & $ASCII,
+                                );
+    }
+    $perl->table('PosixDigit')->add_description('\d, restricted to ASCII');
+    $perl->table('PosixDigit')->add_description('[0-9]');
+
+
+    my $dt = property_ref('Decomposition_Type');
+    $dt->add_match_table('Non_Canon', Full_Name => 'Non_Canonical',
+        Initialize => ~ ($dt->table('None') + $dt->table('Canonical')),
+        Perl_Extension => 1,
+        Note => 'Perl extension consisting of the union of all non-canonical decompositions',
+        );
+
+    # _CanonDCIJ is equivalent to Soft_Dotted, but if on a release earlier
+    # than SD appeared, construct it ourselves, based on the first release SD
+    # was in.
+    my $CanonDCIJ = $perl->add_match_table('_CanonDCIJ');
+    my $soft_dotted = property_ref('Soft_Dotted');
+    if (defined $soft_dotted && ! $soft_dotted->is_empty) {
+        $CanonDCIJ->set_equivalent_to($soft_dotted->table('Y'), Related => 1);
+    }
+    else {
+
+        # This list came from 3.2 Soft_Dotted.
+        $CanonDCIJ->initialize([ 0x0069,
+                                 0x006A,
+                                 0x012F,
+                                 0x0268,
+                                 0x0456,
+                                 0x0458,
+                                 0x1E2D,
+                                 0x1ECB,
+                               ]);
+        $CanonDCIJ = $CanonDCIJ & $Assigned;
+    }
+
+    # This is used in Unicode's definition of \X
+    my $gcb = property_ref('Grapheme_Cluster_Break');
+    if (defined $gcb) {
+        my $extend = $perl->add_match_table('_GCB_Extend',
+                                        Initialize => $gcb->table('Extend'));
+        $extend += $gcb->table('SpacingMark')
+                                        if defined $gcb->table('SpacingMark');
+    }
+    else {    # Old definition, used on early releases.
+        $perl->add_match_table('_X_Extend', Initialize => $gc->table('Mark')
+                                                            + 0x200C    # ZWNJ
+                                                            + 0x200D    # ZWJ
+        );
+    }
+
+    # Create a new property specially located that is a combination of the
+    # various Name properties: Name, Unicode_1_Name, Named Sequences, and
+    # Name_Alias properties.  (The final duplicates elements of the first.)  A
+    # comment for it is constructed based on the actual properties present and
+    # used
+    my $perl_charname = Property->new('Perl_Charnames',
+                                Core_Access => '\N{...} and charnames.pm',
+                                Default_Map => "",
+                                Directory => '.',
+                                File => 'Name',
+                                Internal_Only_Warning => 1,
+                                Perl_Extension => 1,
+                                Range_Size_1 => 1,
+                                Type => $STRING,
+                                Initialize => property_ref('Unicode_1_Name'),
+                                );
+    # Name overrides Unicode_1_Name
+    $perl_charname->property_add_or_replace_non_nulls(property_ref('Name'));
+    my @composition = ('Name', 'Unicode_1_Name');
+
+    if (@named_sequences) {
+        push @composition, 'Named_Sequence';
+        foreach my $sequence (@named_sequences) {
+            $perl_charname->add_anomalous_entry($sequence);
+        }
+    }
+
+    my $alias_sentence = "";
+    my $alias = property_ref('Name_Alias');
+    if (defined $alias) {
+        push @composition, 'Name_Alias';
+        $alias->reset_each_range;
+        while (my ($range) = $alias->each_range) {
+            next if $range->value eq "";
+            if ($range->start != $range->end) {
+                Carp::my_carp("Expecting only one code point in the range $range.  Just to keep going, using just the first code point;");
+            }
+            $perl_charname->add_duplicate($range->start, $range->value);
+        }
+        $alias_sentence = <<END;
+The Name_Alias property adds duplicate code point entries with a corrected
+name.  The original (less correct, but still valid) name will be physically
+first.
+END
+    }
+    my $comment;
+    if (@composition <= 2) { # Always at least 2
+        $comment = join " and ", @composition;
+    }
+    else {
+        $comment = join ", ", @composition[0 .. scalar @composition - 2];
+        $comment .= ", and $composition[-1]";
+    }
+
+    # Wait for charnames to catch up
+#    foreach my $entry (@more_Names,
+#                        split "\n", <<"END"
+#000A; LF
+#000C; FF
+#000D; CR
+#0085; NEL
+#200C; ZWNJ
+#200D; ZWJ
+#FEFF; BOM
+#FEFF; BYTE ORDER MARK
+#END
+#    ) {
+#        #local $to_trace = 1 if main::DEBUG;
+#        trace $entry if main::DEBUG && $to_trace;
+#        my ($code_point, $name) = split /\s*;\s*/, $entry;
+#        $code_point = hex $code_point;
+#        trace $code_point, $name if main::DEBUG && $to_trace;
+#        $perl_charname->add_duplicate($code_point, $name);
+#    }
+#    #$perl_charname->add_comment("This file is for charnames.pm.  It is the union of the $comment properties, plus certain commonly used but unofficial names, such as 'FF' and 'ZWNJ'.  Unicode_1_Name entries are used only for otherwise nameless code points.$alias_sentence");
+    $perl_charname->add_comment(join_lines( <<END
+This file is for charnames.pm.  It is the union of the $comment properties.
+Unicode_1_Name entries are used only for otherwise nameless code
+points.
+$alias_sentence
+END
+    ));
+
+    # The combining class property used by Perl's normalize.pm is not located
+    # in the normal mapping directory; create a copy for it.
+    my $ccc = property_ref('Canonical_Combining_Class');
+    my $perl_ccc = Property->new('Perl_ccc',
+                            Default_Map => $ccc->default_map,
+                            Full_Name => 'Perl_Canonical_Combining_Class',
+                            Internal_Only_Warning => 1,
+                            Perl_Extension => 1,
+                            Pod_Entry =>0,
+                            Type => $ENUM,
+                            Initialize => $ccc,
+                            File => 'CombiningClass',
+                            Directory => '.',
+                            );
+    $perl_ccc->set_to_output_map(1);
+    $perl_ccc->add_comment(join_lines(<<END
+This mapping is for normalize.pm.  It is currently identical to the Unicode
+Canonical_Combining_Class property.
+END
+    ));
+
+    # This one match table for it is needed for calculations on output
+    my $default = $perl_ccc->add_match_table($ccc->default_map,
+                        Initialize => $ccc->table($ccc->default_map),
+                        Status => $SUPPRESSED);
+
+    # Construct the Present_In property from the Age property.
+    if (-e 'DAge.txt' && defined (my $age = property_ref('Age'))) {
+        my $default_map = $age->default_map;
+        my $in = Property->new('In',
+                                Default_Map => $default_map,
+                                Full_Name => "Present_In",
+                                Internal_Only_Warning => 1,
+                                Perl_Extension => 1,
+                                Type => $ENUM,
+                                Initialize => $age,
+                                );
+        $in->add_comment(join_lines(<<END
+This file should not be used for any purpose.  The values in this file are the
+same as for $age, and not for what $in really means.  This is because anything
+defined in a given release should have multiple values: that release and all
+higher ones.  But only one value per code point can be represented in a table
+like this.
+END
+        ));
+
+        # The Age tables are named like 1.5, 2.0, 2.1, ....  Sort so that the
+        # lowest numbered (earliest) come first, with the non-numeric one
+        # last.
+        my ($first_age, @rest_ages) = sort { ($a->name !~ /^[\d.]*$/)
+                                            ? 1
+                                            : ($b->name !~ /^[\d.]*$/)
+                                                ? -1
+                                                : $a->name <=> $b->name
+                                            } $age->tables;
+
+        # The Present_In property is the cumulative age properties.  The first
+        # one hence is identical to the first age one.
+        my $previous_in = $in->add_match_table($first_age->name);
+        $previous_in->set_equivalent_to($first_age, Related => 1);
+
+        my $description_start = "Code point's usage introduced in version ";
+        $first_age->add_description($description_start . $first_age->name);
+
+        # To construct the accumlated values, for each of the age tables
+        # starting with the 2nd earliest, merge the earliest with it, to get
+        # all those code points existing in the 2nd earliest.  Repeat merging
+        # the new 2nd earliest with the 3rd earliest to get all those existing
+        # in the 3rd earliest, and so on.
+        foreach my $current_age (@rest_ages) {
+            next if $current_age->name !~ /^[\d.]*$/;   # Skip the non-numeric
+
+            my $current_in = $in->add_match_table(
+                                    $current_age->name,
+                                    Initialize => $current_age + $previous_in,
+                                    Description => $description_start
+                                                    . $current_age->name
+                                                    . ' or earlier',
+                                    );
+            $previous_in = $current_in;
+
+            # Add clarifying material for the corresponding age file.  This is
+            # in part because of the confusing and contradictory information
+            # given in the Standard's documentation itself, as of 5.2.
+            $current_age->add_description(
+                            "Code point's usage was introduced in version "
+                            . $current_age->name);
+            $current_age->add_note("See also $in");
+
+        }
+
+        # And finally the code points whose usages have yet to be decided are
+        # the same in both properties.  Note that permanently unassigned code
+        # points actually have their usage assigned (as being permanently
+        # unassigned), so that these tables are not the same as gc=cn.
+        my $unassigned = $in->add_match_table($default_map);
+        my $age_default = $age->table($default_map);
+        $age_default->add_description(<<END
+Code point's usage has not been assigned in any Unicode release thus far.
+END
+        );
+        $unassigned->set_equivalent_to($age_default, Related => 1);
+    }
+
+
+    # Finished creating all the perl properties.  All non-internal non-string
+    # ones have a synonym of 'Is_' prefixed.  (Internal properties begin with
+    # an underscore.)  These do not get a separate entry in the pod file
+    foreach my $table ($perl->tables) {
+        foreach my $alias ($table->aliases) {
+            next if $alias->name =~ /^_/;
+            $table->add_alias('Is_' . $alias->name,
+                               Pod_Entry => 0,
+                               Status => $alias->status,
+                               Externally_Ok => 0);
+        }
+    }
+
+    return;
+}
+
+sub add_perl_synonyms() {
+    # A number of Unicode tables have Perl synonyms that are expressed in
+    # the single-form, \p{name}.  These are:
+    #   All the binary property Y tables, so that \p{Name=Y} gets \p{Name} and
+    #       \p{Is_Name} as synonyms
+    #   \p{Script=Value} gets \p{Value}, \p{Is_Value} as synonyms
+    #   \p{General_Category=Value} gets \p{Value}, \p{Is_Value} as synonyms
+    #   \p{Block=Value} gets \p{In_Value} as a synonym, and, if there is no
+    #       conflict, \p{Value} and \p{Is_Value} as well
+    #
+    # This routine generates these synonyms, warning of any unexpected
+    # conflicts.
+
+    # Construct the list of tables to get synonyms for.  Start with all the
+    # binary and the General_Category ones.
+    my @tables = grep { $_->type == $BINARY } property_ref('*');
+    push @tables, $gc->tables;
+
+    # If the version of Unicode includes the Script property, add its tables
+    if (defined property_ref('Script')) {
+        push @tables, property_ref('Script')->tables;
+    }
+
+    # The Block tables are kept separate because they are treated differently.
+    # And the earliest versions of Unicode didn't include them, so add only if
+    # there are some.
+    my @blocks;
+    push @blocks, $block->tables if defined $block;
+
+    # Here, have the lists of tables constructed.  Process blocks last so that
+    # if there are name collisions with them, blocks have lowest priority.
+    # Should there ever be other collisions, manual intervention would be
+    # required.  See the comments at the beginning of the program for a
+    # possible way to handle those semi-automatically.
+    foreach my $table (@tables,  @blocks) {
+
+        # For non-binary properties, the synonym is just the name of the
+        # table, like Greek, but for binary properties the synonym is the name
+        # of the property, and means the code points in its 'Y' table.
+        my $nominal = $table;
+        my $nominal_property = $nominal->property;
+        my $actual;
+        if (! $nominal->isa('Property')) {
+            $actual = $table;
+        }
+        else {
+
+            # Here is a binary property.  Use the 'Y' table.  Verify that is
+            # there
+            my $yes = $nominal->table('Y');
+            unless (defined $yes) {  # Must be defined, but is permissible to
+                                     # be empty.
+                Carp::my_carp_bug("Undefined $nominal, 'Y'.  Skipping.");
+                next;
+            }
+            $actual = $yes;
+        }
+
+        foreach my $alias ($nominal->aliases) {
+
+            # Attempt to create a table in the perl directory for the
+            # candidate table, using whatever aliases in it that don't
+            # conflict.  Also add non-conflicting aliases for all these
+            # prefixed by 'Is_' (and/or 'In_' for Block property tables)
+            PREFIX:
+            foreach my $prefix ("", 'Is_', 'In_') {
+
+                # Only Block properties can have added 'In_' aliases.
+                next if $prefix eq 'In_' and $nominal_property != $block;
+
+                my $proposed_name = $prefix . $alias->name;
+
+                # No Is_Is, In_In, nor combinations thereof
+                trace "$proposed_name is a no-no" if main::DEBUG && $to_trace && $proposed_name =~ /^ I [ns] _I [ns] _/x;
+                next if $proposed_name =~ /^ I [ns] _I [ns] _/x;
+
+                trace "Seeing if can add alias or table: 'perl=$proposed_name' based on $nominal" if main::DEBUG && $to_trace;
+
+                # Get a reference to any existing table in the perl
+                # directory with the desired name.
+                my $pre_existing = $perl->table($proposed_name);
+
+                if (! defined $pre_existing) {
+
+                    # No name collision, so ok to add the perl synonym.
+
+                    my $make_pod_entry;
+                    my $externally_ok;
+                    my $status = $actual->status;
+                    if ($nominal_property == $block) {
+
+                        # For block properties, the 'In' form is preferred for
+                        # external use; the pod file contains wild cards for
+                        # this and the 'Is' form so no entries for those; and
+                        # we don't want people using the name without the
+                        # 'In', so discourage that.
+                        if ($prefix eq "") {
+                            $make_pod_entry = 1;
+                            $status = $status || $DISCOURAGED;
+                            $externally_ok = 0;
+                        }
+                        elsif ($prefix eq 'In_') {
+                            $make_pod_entry = 0;
+                            $status = $status || $NORMAL;
+                            $externally_ok = 1;
+                        }
+                        else {
+                            $make_pod_entry = 0;
+                            $status = $status || $DISCOURAGED;
+                            $externally_ok = 0;
+                        }
+                    }
+                    elsif ($prefix ne "") {
+
+                        # The 'Is' prefix is handled in the pod by a wild
+                        # card, and we won't use it for an external name
+                        $make_pod_entry = 0;
+                        $status = $status || $NORMAL;
+                        $externally_ok = 0;
                     }
+                    else {
+
+                        # Here, is an empty prefix, non block.  This gets its
+                        # own pod entry and can be used for an external name.
+                        $make_pod_entry = 1;
+                        $status = $status || $NORMAL;
+                        $externally_ok = 1;
+                    }
+
+                    # Here, there isn't a perl pre-existing table with the
+                    # name.  Look through the list of equivalents of this
+                    # table to see if one is a perl table.
+                    foreach my $equivalent ($actual->leader->equivalents) {
+                        next if $equivalent->property != $perl;
+
+                        # Here, have found a table for $perl.  Add this alias
+                        # to it, and are done with this prefix.
+                        $equivalent->add_alias($proposed_name,
+                                        Pod_Entry => $make_pod_entry,
+                                        Status => $status,
+                                        Externally_Ok => $externally_ok);
+                        trace "adding alias perl=$proposed_name to $equivalent" if main::DEBUG && $to_trace;
+                        next PREFIX;
+                    }
+
+                    # Here, $perl doesn't already have a table that is a
+                    # synonym for this property, add one.
+                    my $added_table = $perl->add_match_table($proposed_name,
+                                            Pod_Entry => $make_pod_entry,
+                                            Status => $status,
+                                            Externally_Ok => $externally_ok);
+                    # And it will be related to the actual table, since it is
+                    # based on it.
+                    $added_table->set_equivalent_to($actual, Related => 1);
+                    trace "added ", $perl->table($proposed_name) if main::DEBUG && $to_trace;
+                    next;
+                } # End of no pre-existing.
+
+                # Here, there is a pre-existing table that has the proposed
+                # name.  We could be in trouble, but not if this is just a
+                # synonym for another table that we have already made a child
+                # of the pre-existing one.
+                if ($pre_existing->is_equivalent_to($actual)) {
+                    trace "$pre_existing is already equivalent to $actual; adding alias perl=$proposed_name to it" if main::DEBUG && $to_trace;
+                    $pre_existing->add_alias($proposed_name);
+                    next;
                 }
 
-                my $TypeToShow = $Type eq 'Is' ? "" : $Type;
-                my $OrigProp;
+                # Here, there is a name collision, but it still could be ok if
+                # the tables match the identical set of code points, in which
+                # case, we can combine the names.  Compare each table's code
+                # point list to see if they are identical.
+                trace "Potential name conflict with $pre_existing having ", $pre_existing->count, " code points" if main::DEBUG && $to_trace;
+                if ($pre_existing->matches_identically_to($actual)) {
+
+                    # Here, they do match identically.  Not a real conflict.
+                    # Make the perl version a child of the Unicode one, except
+                    # in the non-obvious case of where the perl name is
+                    # already a synonym of another Unicode property.  (This is
+                    # excluded by the test for it being its own parent.)  The
+                    # reason for this exclusion is that then the two Unicode
+                    # properties become related; and we don't really know if
+                    # they are or not.  We generate documentation based on
+                    # relatedness, and this would be misleading.  Code
+                    # later executed in the process will cause the tables to
+                    # be represented by a single file anyway, without making
+                    # it look in the pod like they are necessarily related.
+                    if ($pre_existing->parent == $pre_existing
+                        && ($pre_existing->property == $perl
+                            || $actual->property == $perl))
+                    {
+                        trace "Setting $pre_existing equivalent to $actual since one is \$perl, and match identical sets" if main::DEBUG && $to_trace;
+                        $pre_existing->set_equivalent_to($actual, Related => 1);
+                    }
+                    elsif (main::DEBUG && $to_trace) {
+                        trace "$pre_existing is equivalent to $actual since match identical sets, but not setting them equivalent, to preserve the separateness of the perl aliases";
+                        trace $pre_existing->parent;
+                    }
+                    next PREFIX;
+                }
 
-                $Comment = "This file supports:\n";
-                for my $N (@Supported)
+                # Here they didn't match identically, there is a real conflict
+                # between our new name and a pre-existing property.
+                $actual->add_conflicting($proposed_name, 'p', $pre_existing);
+                $pre_existing->add_conflicting($nominal->full_name,
+                                               'p',
+                                               $actual);
+
+                # Don't output a warning for aliases for the block
+                # properties (unless they start with 'In_') as it is
+                # expected that there will be conflicts and the block
+                # form loses.
+                if ($verbosity >= $NORMAL_VERBOSITY
+                    && ($actual->property != $block || $prefix eq 'In_'))
                 {
-                    my $IsFuzzy = $FuzzyNames{$Type}->{$N};
-                    my $Prop    = "\\p{$TypeToShow$Name}";
-                    $OrigProp = $Prop if not $OrigProp; #cache for aliases
-                    if ($IsFuzzy) {
-                        $Comment .= "\t$Prop (and fuzzy permutations)\n";
-                    } else {
-                        $Comment .= "\t$Prop\n";
-                    }
-                    my $MyDesc = ($N eq $Name) ? $Desc : "Alias for $OrigProp ($Desc)";
+                    print simple_fold(join_lines(<<END
+There is already an alias named $proposed_name (from " . $pre_existing . "),
+so not creating this alias for " . $actual
+END
+                    ), "", 4);
+                }
+
+                # Keep track for documentation purposes.
+                $has_In_conflicts++ if $prefix eq 'In_';
+                $has_Is_conflicts++ if $prefix eq 'Is_';
+            }
+        }
+    }
 
-                    push @MAP, sprintf("%s %-42s %s\n",
-                                       $IsFuzzy ? '*' : ' ', $Prop, $MyDesc);
+    # There are some properties which have No and Yes (and N and Y) as
+    # property values, but aren't binary, and could possibly be confused with
+    # binary ones.  So create caveats for them.  There are tables that are
+    # named 'No', and tables that are named 'N', but confusion is not likely
+    # unless they are the same table.  For example, N meaning Number or
+    # Neutral is not likely to cause confusion, so don't add caveats to things
+    # like them.
+    foreach my $property (grep { $_->type != $BINARY } property_ref('*')) {
+        my $yes = $property->table('Yes');
+        if (defined $yes) {
+            my $y = $property->table('Y');
+            if (defined $y && $yes == $y) {
+                foreach my $alias ($property->aliases) {
+                    $yes->add_conflicting($alias->name);
                 }
-                if ($Desc) {
-                    $Comment .= "\nMeaning: $Desc\n";
+            }
+        }
+        my $no = $property->table('No');
+        if (defined $no) {
+            my $n = $property->table('N');
+            if (defined $n && $no == $n) {
+                foreach my $alias ($property->aliases) {
+                    $no->add_conflicting($alias->name, 'P');
                 }
+            }
+        }
+    }
+
+    return;
+}
+
+sub register_file_for_name($$$) {
+    # Given info about a table and a datafile that it should be associated
+    # with, register that assocation
+
+    my $table = shift;
+    my $directory_ref = shift;   # Array of the directory path for the file
+    my $file = shift;            # The file name in the final directory, [-1].
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+    trace "table=$table, file=$file, directory=@$directory_ref" if main::DEBUG && $to_trace;
+
+    if ($table->isa('Property')) {
+        $table->set_file_path(@$directory_ref, $file);
+        push @map_properties, $table
+                                    if $directory_ref->[0] eq $map_directory;
+        return;
+    }
 
+    # Do all of the work for all equivalent tables when called with the leader
+    # table, so skip if isn't the leader.
+    return if $table->leader != $table;
+
+    # Join all the file path components together, using slashes.
+    my $full_filename = join('/', @$directory_ref, $file);
+
+    # All go in the same subdirectory of unicore
+    if ($directory_ref->[0] ne $matches_directory) {
+        Carp::my_carp("Unexpected directory in "
+                .  join('/', @{$directory_ref}, $file));
+    }
+
+    # For this table and all its equivalents ...
+    foreach my $table ($table, $table->equivalents) {
+
+        # Associate it with its file internally.  Don't include the
+        # $matches_directory first component
+        $table->set_file_path(@$directory_ref, $file);
+        my $sub_filename = join('/', $directory_ref->[1, -1], $file);
+
+        my $property = $table->property;
+        $property = ($property == $perl)
+                    ? ""                # 'perl' is never explicitly stated
+                    : standardize($property->name) . '=';
+
+        my $deprecated = ($table->status eq $DEPRECATED)
+                         ? $table->status_info
+                         : "";
+
+        # And for each of the table's aliases...  This inner loop eventually
+        # goes through all aliases in the UCD that we generate regex match
+        # files for
+        foreach my $alias ($table->aliases) {
+            my $name = $alias->name;
+
+            # Generate an entry in either the loose or strict hashes, which
+            # will translate the property and alias names combination into the
+            # file where the table for them is stored.
+            my $standard;
+            if ($alias->loose_match) {
+                $standard = $property . standardize($alias->name);
+                if (exists $loose_to_file_of{$standard}) {
+                    Carp::my_carp("Can't change file registered to $loose_to_file_of{$standard} to '$sub_filename'.");
+                }
+                else {
+                    $loose_to_file_of{$standard} = $sub_filename;
+                }
+            }
+            else {
+                $standard = lc ($property . $name);
+                if (exists $stricter_to_file_of{$standard}) {
+                    Carp::my_carp("Can't change file registered to $stricter_to_file_of{$standard} to '$sub_filename'.");
+                }
+                else {
+                    $stricter_to_file_of{$standard} = $sub_filename;
+
+                    # Tightly coupled with how utf8_heavy.pl works, for a
+                    # floating point number that is a whole number, get rid of
+                    # the trailing decimal point and 0's, so that utf8_heavy
+                    # will work.  Also note that this assumes that such a
+                    # number is matched strictly; so if that were to change,
+                    # this would be wrong.
+                    if ((my $integer_name = $name)
+                            =~ s/^ ( -? \d+ ) \.0+ $ /$1/x)
+                    {
+                        $stricter_to_file_of{$property . $integer_name}
+                            = $sub_filename;
+                    }
+                }
             }
-            ##
-            ## Okay, write the file...
-            ##
-            $Table->Write(["lib","gc_sc","$filename.pl"], $Comment);
 
-            ## and register it
-            $RawNameToFile{$Name} = $filename;
-            RegisterFileForName($Type => $Name, $IsFuzzy, $filename);
+            # Keep a list of the deprecated properties and their filenames
+            if ($deprecated) {
+                $utf8::why_deprecated{$sub_filename} = $deprecated;
+            }
+        }
+    }
 
-            if ($IsFuzzy)
-            {
-                my $CName = CanonicalName($Type . '_'. $Name);
-                $FuzzyNameToTest{$Name}  = $Table if !$FuzzyNameToTest{$Name};
-                $FuzzyNameToTest{$CName} = $Table if !$FuzzyNameToTest{$CName};
-            } else {
-                $ExactNameToTest{$Name} = $Table;
+    return;
+}
+
+{   # Closure
+    my %base_names;  # Names already used for avoiding DOS 8.3 filesystem
+                     # conflicts
+    my %full_dir_name_of;   # Full length names of directories used.
+
+    sub construct_filename($$$) {
+        # Return a file name for a table, based on the table name, but perhaps
+        # changed to get rid of non-portable characters in it, and to make
+        # sure that it is unique on a file system that allows the names before
+        # any period to be at most 8 characters (DOS).  While we're at it
+        # check and complain if there are any directory conflicts.
+
+        my $name = shift;       # The name to start with
+        my $mutable = shift;    # Boolean: can it be changed?  If no, but
+                                # yet it must be to work properly, a warning
+                                # is given
+        my $directories_ref = shift;  # A reference to an array containing the
+                                # path to the file, with each element one path
+                                # component.  This is used because the same
+                                # name can be used in different directories.
+        Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+        my $warn = ! defined wantarray;  # If true, then if the name is
+                                # changed, a warning is issued as well.
+
+        if (! defined $name) {
+            Carp::my_carp("Undefined name in directory "
+                          . File::Spec->join(@$directories_ref)
+                          . ". '_' used");
+            return '_';
+        }
+
+        # Make sure that no directory names conflict with each other.  Look at
+        # each directory in the input file's path.  If it is already in use,
+        # assume it is correct, and is merely being re-used, but if we
+        # truncate it to 8 characters, and find that there are two directories
+        # that are the same for the first 8 characters, but differ after that,
+        # then that is a problem.
+        foreach my $directory (@$directories_ref) {
+            my $short_dir = substr($directory, 0, 8);
+            if (defined $full_dir_name_of{$short_dir}) {
+                next if $full_dir_name_of{$short_dir} eq $directory;
+                Carp::my_carp("$directory conflicts with $full_dir_name_of{$short_dir}.  Bad News.  Continuing anyway");
+            }
+            else {
+                $full_dir_name_of{$short_dir} = $directory;
             }
+        }
 
+        my $path = join '/', @$directories_ref;
+        $path .= '/' if $path;
+
+        # Remove interior underscores.
+        (my $filename = $name) =~ s/ (?<=.) _ (?=.) //xg;
+
+        # Change any non-word character into an underscore, and truncate to 8.
+        $filename =~ s/\W+/_/g;   # eg., "L&" -> "L_"
+        substr($filename, 8) = "" if length($filename) > 8;
+
+        # Make sure the basename doesn't conflict with something we
+        # might have already written. If we have, say,
+        #     InGreekExtended1
+        #     InGreekExtended2
+        # they become
+        #     InGreekE
+        #     InGreek2
+        my $warned = 0;
+        while (my $num = $base_names{$path}{lc $filename}++) {
+            $num++; # so basenames with numbers start with '2', which
+                    # just looks more natural.
+
+            # Want to append $num, but if it'll make the basename longer
+            # than 8 characters, pre-truncate $filename so that the result
+            # is acceptable.
+            my $delta = length($filename) + length($num) - 8;
+            if ($delta > 0) {
+                substr($filename, -$delta) = $num;
+            }
+            else {
+                $filename .= $num;
+            }
+            if ($warn && ! $warned) {
+                $warned = 1;
+                Carp::my_carp("'$path$name' conflicts with another name on a filesystem with 8 significant characters (like DOS).  Proceeding anyway.");
+            }
         }
 
-        ## Register aliase info
-        for my $Name (sort {length $a <=> length $b} keys %{$AliasInfo{$Type}})
-        {
-            my $Alias    = $AliasInfo{$Type}->{$Name};
-            my $IsFuzzy  = $FuzzyNames{$Type}->{$Alias};
-            my $filename = $RawNameToFile{$Name};
-            die "oops [$Alias]->[$Name]" if not $filename;
-            RegisterFileForName($Type => $Alias, $IsFuzzy, $filename);
-
-            my $Table = $TableInfo{$Type}->{$Name};
-            die "oops" if not $Table;
-            if ($IsFuzzy)
-            {
-                my $CName = CanonicalName($Type .'_'. $Alias);
-                $FuzzyNameToTest{$Alias} = $Table if !$FuzzyNameToTest{$Alias};
-                $FuzzyNameToTest{$CName} = $Table if !$FuzzyNameToTest{$CName};
-            } else {
-                $ExactNameToTest{$Alias} = $Table;
+        return $filename if $mutable;
+
+        # If not changeable, must return the input name, but warn if needed to
+        # change it beyond shortening it.
+        if ($name ne $filename
+            && substr($name, 0, length($filename)) ne $filename) {
+            Carp::my_carp("'$path$name' had to be changed into '$filename'.  Bad News.  Proceeding anyway.");
+        }
+        return $name;
+    }
+}
+
+# The pod file contains a very large table.  Many of the lines in that table
+# would exceed a typical output window's size, and so need to be wrapped with
+# a hanging indent to make them look good.  The pod language is really
+# insufficient here.  There is no general construct to do that in pod, so it
+# is done here by beginning each such line with a space to cause the result to
+# be output without formatting, and doing all the formatting here.  This leads
+# to the result that if the eventual display window is too narrow it won't
+# look good, and if the window is too wide, no advantage is taken of that
+# extra width.  A further complication is that the output may be indented by
+# the formatter so that there is less space than expected.  What I (khw) have
+# done is to assume that that indent is a particular number of spaces based on
+# what it is in my Linux system;  people can always resize their windows if
+# necessary, but this is obviously less than desirable, but the best that can
+# be expected.
+my $automatic_pod_indent = 8;
+
+# Try to format so that uses fewest lines, but few long left column entries
+# slide into the right column.  An experiment on 5.1 data yielded the
+# following percentages that didn't cut into the other side along with the
+# associated first-column widths
+# 69% = 24
+# 80% not too bad except for a few blocks
+# 90% = 33; # , cuts 353/3053 lines from 37 = 12%
+# 95% = 37;
+my $indent_info_column = 27;    # 75% of lines didn't have overlap
+
+my $FILLER = 3;     # Length of initial boiler-plate columns in a pod line
+                    # The 3 is because of:
+                    #   1   for the leading space to tell the pod formatter to
+                    #       output as-is
+                    #   1   for the flag
+                    #   1   for the space between the flag and the main data
+
+sub format_pod_line ($$$;$$) {
+    # Take a pod line and return it, formatted properly
+
+    my $first_column_width = shift;
+    my $entry = shift;  # Contents of left column
+    my $info = shift;   # Contents of right column
+
+    my $status = shift || "";   # Any flag
+
+    my $loose_match = shift;    # Boolean.
+    $loose_match = 1 unless defined $loose_match;
+
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+    my $flags = "";
+    $flags .= $STRICTER if ! $loose_match;
+
+    $flags .= $status if $status;
+
+    # There is a blank in the left column to cause the pod formatter to
+    # output the line as-is.
+    return sprintf " %-*s%-*s %s\n",
+                    # The first * in the format is replaced by this, the -1 is
+                    # to account for the leading blank.  There isn't a
+                    # hard-coded blank after this to separate the flags from
+                    # the rest of the line, so that in the unlikely event that
+                    # multiple flags are shown on the same line, they both
+                    # will get displayed at the expense of that separation,
+                    # but since they are left justified, a blank will be
+                    # inserted in the normal case.
+                    $FILLER - 1,
+                    $flags,
+
+                    # The other * in the format is replaced by this number to
+                    # cause the first main column to right fill with blanks.
+                    # The -1 is for the guaranteed blank following it.
+                    $first_column_width - $FILLER - 1,
+                    $entry,
+                    $info;
+}
+
+my @zero_match_tables;  # List of tables that have no matches in this release
+
+sub make_table_pod_entries($) {
+    # This generates the entries for the pod file for a given table.
+    # Also done at this time are any children tables.  The output looks like:
+    # \p{Common}              \p{Script=Common} (Short: \p{Zyyy}) (5178)
+
+    my $input_table = shift;        # Table the entry is for
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+    # Generate parent and all its children at the same time.
+    return if $input_table->parent != $input_table;
+
+    my $property = $input_table->property;
+    my $type = $property->type;
+    my $full_name = $property->full_name;
+
+    my $count = $input_table->count;
+    my $string_count = clarify_number($count);
+    my $status = $input_table->status;
+    my $status_info = $input_table->status_info;
+
+    my $entry_for_first_table; # The entry for the first table output.
+                           # Almost certainly, it is the parent.
+
+    # For each related table (including itself), we will generate a pod entry
+    # for each name each table goes by
+    foreach my $table ($input_table, $input_table->children) {
+
+        # utf8_heavy.pl cannot deal with null string property values, so don't
+        # output any.
+        next if $table->name eq "";
+
+        # First, gather all the info that applies to this table as a whole.
+
+        push @zero_match_tables, $table if $count == 0;
+
+        my $table_property = $table->property;
+
+        # The short name has all the underscores removed, while the full name
+        # retains them.  Later, we decide whether to output a short synonym
+        # for the full one, we need to compare apples to apples, so we use the
+        # short name's length including underscores.
+        my $table_property_short_name_length;
+        my $table_property_short_name
+            = $table_property->short_name(\$table_property_short_name_length);
+        my $table_property_full_name = $table_property->full_name;
+
+        # Get how much savings there is in the short name over the full one
+        # (delta will always be <= 0)
+        my $table_property_short_delta = $table_property_short_name_length
+                                         - length($table_property_full_name);
+        my @table_description = $table->description;
+        my @table_note = $table->note;
+
+        # Generate an entry for each alias in this table.
+        my $entry_for_first_alias;  # saves the first one encountered.
+        foreach my $alias ($table->aliases) {
+
+            # Skip if not to go in pod.
+            next unless $alias->make_pod_entry;
+
+            # Start gathering all the components for the entry
+            my $name = $alias->name;
+
+            my $entry;      # Holds the left column, may include extras
+            my $entry_ref;  # To refer to the left column's contents from
+                            # another entry; has no extras
+
+            # First the left column of the pod entry.  Tables for the $perl
+            # property always use the single form.
+            if ($table_property == $perl) {
+                $entry = "\\p{$name}";
+                $entry_ref = "\\p{$name}";
+            }
+            else {    # Compound form.
+
+                # Only generate one entry for all the aliases that mean true
+                # or false in binary properties.  Append a '*' to indicate
+                # some are missing.  (The heading comment notes this.)
+                my $wild_card_mark;
+                if ($type == $BINARY) {
+                    next if $name ne 'N' && $name ne 'Y';
+                    $wild_card_mark = '*';
+                }
+                else {
+                    $wild_card_mark = "";
+                }
+
+                # Colon-space is used to give a little more space to be easier
+                # to read;
+                $entry = "\\p{"
+                        . $table_property_full_name
+                        . ": $name$wild_card_mark}";
+
+                # But for the reference to this entry, which will go in the
+                # right column, where space is at a premium, use equals
+                # without a space
+                $entry_ref = "\\p{" . $table_property_full_name . "=$name}";
+            }
+
+            # Then the right (info) column.  This is stored as components of
+            # an array for the moment, then joined into a string later.  For
+            # non-internal only properties, begin the info with the entry for
+            # the first table we encountered (if any), as things are ordered
+            # so that that one is the most descriptive.  This leads to the
+            # info column of an entry being a more descriptive version of the
+            # name column
+            my @info;
+            if ($name =~ /^_/) {
+                push @info,
+                        '(For internal use by Perl, not necessarily stable)';
+            }
+            elsif ($entry_for_first_alias) {
+                push @info, $entry_for_first_alias;
+            }
+
+            # If this entry is equivalent to another, add that to the info,
+            # using the first such table we encountered
+            if ($entry_for_first_table) {
+                if (@info) {
+                    push @info, "(= $entry_for_first_table)";
+                }
+                else {
+                    push @info, $entry_for_first_table;
+                }
             }
+
+            # If the name is a large integer, add an equivalent with an
+            # exponent for better readability
+            if ($name =~ /^[+-]?[\d]+$/ && $name >= 10_000) {
+                push @info, sprintf "(= %.1e)", $name
+            }
+
+            my $parenthesized = "";
+            if (! $entry_for_first_alias) {
+
+                # This is the first alias for the current table.  The alias
+                # array is ordered so that this is the fullest, most
+                # descriptive alias, so it gets the fullest info.  The other
+                # aliases are mostly merely pointers to this one, using the
+                # information already added above.
+
+                # Display any status message, but only on the parent table
+                if ($status && ! $entry_for_first_table) {
+                    push @info, $status_info;
+                }
+
+                # Put out any descriptive info
+                if (@table_description || @table_note) {
+                    push @info, join "; ", @table_description, @table_note;
+                }
+
+                # Look to see if there is a shorter name we can point people
+                # at
+                my $standard_name = standardize($name);
+                my $short_name;
+                my $proposed_short = $table->short_name;
+                if (defined $proposed_short) {
+                    my $standard_short = standardize($proposed_short);
+
+                    # If the short name is shorter than the standard one, or
+                    # even it it's not, but the combination of it and its
+                    # short property name (as in \p{prop=short} ($perl doesn't
+                    # have this form)) saves at least two characters, then,
+                    # cause it to be listed as a shorter synonym.
+                    if (length $standard_short < length $standard_name
+                        || ($table_property != $perl
+                            && (length($standard_short)
+                                - length($standard_name)
+                                + $table_property_short_delta)  # (<= 0)
+                                < -2))
+                    {
+                        $short_name = $proposed_short;
+                        if ($table_property != $perl) {
+                            $short_name = $table_property_short_name
+                                          . "=$short_name";
+                        }
+                        $short_name = "\\p{$short_name}";
+                    }
+                }
+
+                # And if this is a compound form name, see if there is a
+                # single form equivalent
+                my $single_form;
+                if ($table_property != $perl) {
+
+                    # Special case the binary N tables, so that will print
+                    # \P{single}, but use the Y table values to populate
+                    # 'single', as we haven't populated the N table.
+                    my $test_table;
+                    my $p;
+                    if ($type == $BINARY
+                        && $input_table == $property->table('No'))
+                    {
+                        $test_table = $property->table('Yes');
+                        $p = 'P';
+                    }
+                    else {
+                        $test_table = $input_table;
+                        $p = 'p';
+                    }
+
+                    # Look for a single form amongst all the children.
+                    foreach my $table ($test_table->children) {
+                        next if $table->property != $perl;
+                        my $proposed_name = $table->short_name;
+                        next if ! defined $proposed_name;
+
+                        # Don't mention internal-only properties as a possible
+                        # single form synonym
+                        next if substr($proposed_name, 0, 1) eq '_';
+
+                        $proposed_name = "\\$p\{$proposed_name}";
+                        if (! defined $single_form
+                            || length($proposed_name) < length $single_form)
+                        {
+                            $single_form = $proposed_name;
+
+                            # The goal here is to find a single form; not the
+                            # shortest possible one.  We've already found a
+                            # short name.  So, stop at the first single form
+                            # found, which is likely to be closer to the
+                            # original.
+                            last;
+                        }
+                    }
+                }
+
+                # Ouput both short and single in the same parenthesized
+                # expression, but with only one of 'Single', 'Short' if there
+                # are both items.
+                if ($short_name || $single_form || $table->conflicting) {
+                    $parenthesized .= '(';
+                    $parenthesized .= "Short: $short_name" if $short_name;
+                    if ($short_name && $single_form) {
+                        $parenthesized .= ', ';
+                    }
+                    elsif ($single_form) {
+                        $parenthesized .= 'Single: ';
+                    }
+                    $parenthesized .= $single_form if $single_form;
+                }
+            }
+
+
+            # Warn if this property isn't the same as one that a
+            # semi-casual user might expect.  The other components of this
+            # parenthesized structure are calculated only for the first entry
+            # for this table, but the conflicting is deemed important enough
+            # to go on every entry.
+            my $conflicting = join " NOR ", $table->conflicting;
+            if ($conflicting) {
+                $parenthesized .= '(' if ! $parenthesized;
+                $parenthesized .=  '; ' if $parenthesized ne '(';
+                $parenthesized .= "NOT $conflicting";
+            }
+            $parenthesized .= ')' if $parenthesized;
+
+            push @info, $parenthesized if $parenthesized;
+            push @info, "($string_count)" if $output_range_counts;
+
+            # Now, we have both the entry and info so add them to the
+            # list of all the properties.
+            push @match_properties,
+                format_pod_line($indent_info_column,
+                                $entry,
+                                join( " ", @info),
+                                $alias->status,
+                                $alias->loose_match);
+
+            $entry_for_first_alias = $entry_ref unless $entry_for_first_alias;
+        } # End of looping through the aliases for this table.
+
+        if (! $entry_for_first_table) {
+            $entry_for_first_table = $entry_for_first_alias;
         }
+    } # End of looping through all the related tables
+    return;
+}
+
+sub pod_alphanumeric_sort {
+    # Sort pod entries alphanumerically.
+
+    # The first few character columns are filler; and get rid of all the
+    # trailing stuff, starting with the trailing '}', so as to sort on just
+    # '\p{Name=Value'
+    my $a = lc substr($a, $FILLER);
+    $a =~ s/}.*//;
+    my $b = lc substr($b, $FILLER);
+    $b =~ s/}.*//;
+
+    # Determine if the two operands are numeric property values or not.
+    # A numeric property will look like \p{xyz: 3}.  But the number
+    # can begin with an optional minus sign, and may have a
+    # fraction or rational component, like \p{xyz: 3/2}.  If either
+    # isn't numeric, use alphabetic sort.
+    my ($a_initial, $a_number) =
+        ($a =~ /^\\p{ ( [^:=]+ [:=] \s* ) (-? \d+ (?: [.\/] \d+)? )/ix);
+    return $a cmp $b unless defined $a_number;
+    my ($b_initial, $b_number) =
+        ($b =~ /^\\p{ ( [^:=]+ [:=] \s* ) (-? \d+ (?: [.\/] \d+)? )/ix);
+    return $a cmp $b unless defined $b_number;
+
+    # Here they are both numeric, but use alphabetic sort if the
+    # initial parts don't match
+    return $a cmp $b if $a_initial ne $b_initial;
+
+    # Convert rationals to floating for the comparison.
+    $a_number = eval $a_number if $a_number =~ qr{/};
+    $b_number = eval $b_number if $b_number =~ qr{/};
+
+    return $a_number <=> $b_number;
+}
+
+sub make_pod () {
+    # Create the .pod file.  This generates the various subsections and then
+    # combines them in one big HERE document.
+
+    return unless defined $pod_directory;
+    print "Making pod file\n" if $verbosity >= $PROGRESS;
+
+    my $exception_message =
+    '(Any exceptions are individually noted beginning with the word NOT.)';
+    my @block_warning;
+    if (-e 'Blocks.txt') {
+
+        # Add the line: '\p{In_*}    \p{Block: *}', with the warning message
+        # if the global $has_In_conflicts indicates we have them.
+        push @match_properties, format_pod_line($indent_info_column,
+                                                '\p{In_*}',
+                                                '\p{Block: *}'
+                                                    . (($has_In_conflicts)
+                                                      ? " $exception_message"
+                                                      : ""));
+        @block_warning = << "END";
+
+Matches in the Block property have shortcuts that begin with 'In_'.  For
+example, \\p{Block=Latin1} can be written as \\p{In_Latin1}.  For backward
+compatibility, if there is no conflict with another shortcut, these may also
+be written as \\p{Latin1} or \\p{Is_Latin1}.  But, N.B., there are numerous
+such conflicting shortcuts.  Use of these forms for Block is discouraged, and
+are flagged as such, not only because of the potential confusion as to what is
+meant, but also because a later release of Unicode may preempt the shortcut,
+and your program would no longer be correct.  Use the 'In_' form instead to
+avoid this, or even more clearly, use the compound form, e.g.,
+\\p{blk:latin1}.  See L<perlunicode/"Blocks"> for more information about this.
+END
+    }
+    my $text = "If an entry has flag(s) at its beginning, like '$DEPRECATED', the 'Is_' form has the same flag(s)";
+    $text = "$exception_message $text" if $has_Is_conflicts;
+
+    # And the 'Is_ line';
+    push @match_properties, format_pod_line($indent_info_column,
+                                            '\p{Is_*}',
+                                            "\\p{*} $text");
+
+    # Sort the properties array for output.  It is sorted alphabetically
+    # except numerically for numeric properties, and only output unique lines.
+    @match_properties = sort pod_alphanumeric_sort uniques @match_properties;
+
+    my $formatted_properties = simple_fold(\@match_properties,
+                                        "",
+                                        # indent succeeding lines by two extra
+                                        # which looks better
+                                        $indent_info_column + 2,
+
+                                        # shorten the line length by how much
+                                        # the formatter indents, so the folded
+                                        # line will fit in the space
+                                        # presumably available
+                                        $automatic_pod_indent);
+    # Add column headings, indented to be a little more centered, but not
+    # exactly
+    $formatted_properties =  format_pod_line($indent_info_column,
+                                                    '    NAME',
+                                                    '           INFO')
+                                    . "\n"
+                                    . $formatted_properties;
+
+    # Generate pod documentation lines for the tables that match nothing
+    my $zero_matches;
+    if (@zero_match_tables) {
+        @zero_match_tables = uniques(@zero_match_tables);
+        $zero_matches = join "\n\n",
+                        map { $_ = '=item \p{' . $_->complete_name . "}" }
+                            sort { $a->complete_name cmp $b->complete_name }
+                            uniques(@zero_match_tables);
+
+        $zero_matches = <<END;
+
+=head2 Legal \\p{} and \\P{} constructs that match no characters
+
+Unicode has some property-value pairs that currently don't match anything.
+This happens generally either because they are obsolete, or for symmetry with
+other forms, but no language has yet been encoded that uses them.  In this
+version of Unicode, the following match zero code points:
+
+=over 4
+
+$zero_matches
+
+=back
+
+END
     }
 
-    ##
-    ## Write out the property list
-    ##
-    {
-        my @OUT = (
-                   "##\n",
-                   "## This file created by $0\n",
-                   "## List of built-in \\p{...}/\\P{...} properties.\n",
-                   "##\n",
-                   "## '*' means name may be 'fuzzy'\n",
-                   "##\n\n",
-                   sort { substr($a,2) cmp substr($b, 2) } @MAP,
-                  );
-        WriteIfChanged('Properties', @OUT);
-    }
-
-    ## Write Exact.pl
+    # Generate list of properties that we don't accept, grouped by the reasons
+    # why.  This is so only put out the 'why' once, and then list all the
+    # properties that have that reason under it.
+
+    my %why_list;   # The keys are the reasons; the values are lists of
+                    # properties that have the key as their reason
+
+    # For each property, add it to the list that are suppressed for its reason
+    # The sort will cause the alphabetically first properties to be added to
+    # each list first, so each list will be sorted.
+    foreach my $property (sort keys %why_suppressed) {
+        push @{$why_list{$why_suppressed{$property}}}, $property;
+    }
+
+    # For each reason (sorted by the first property that has that reason)...
+    my @bad_re_properties;
+    foreach my $why (sort { $why_list{$a}->[0] cmp $why_list{$b}->[0] }
+                     keys %why_list)
     {
-        my @OUT = (
-                   $HEADER,
-                   "##\n",
-                   "## Data in this file used by ../utf8_heavy.pl\n",
-                   "##\n\n",
-                   "## Mapping from name to filename in ./lib/gc_sc\n",
-                   "%utf8::Exact = (\n",
-                  );
-
-	$Exact{InGreek} = 'InGreekA';  # this is evil kludge
-        for my $Name (sort keys %Exact)
+        # Add to the output, all the properties that have that reason.  Start
+        # with an empty line.
+        push @bad_re_properties, "\n\n";
+
+        my $has_item = 0;   # Flag if actually output anything.
+        foreach my $name (@{$why_list{$why}}) {
+
+            # Split compound names into $property and $table components
+            my $property = $name;
+            my $table;
+            if ($property =~ / (.*) = (.*) /x) {
+                $property = $1;
+                $table = $2;
+            }
+
+            # This release of Unicode may not have a property that is
+            # suppressed, so don't reference a non-existent one.
+            $property = property_ref($property);
+            next if ! defined $property;
+
+            # And since this list is only for match tables, don't list the
+            # ones that don't have match tables.
+            next if ! $property->to_create_match_tables;
+
+            # Find any abbreviation, and turn it into a compound name if this
+            # is a property=value pair.
+            my $short_name = $property->name;
+            $short_name .= '=' . $property->table($table)->name if $table;
+
+            # And add the property as an item for the reason.
+            push @bad_re_properties, "\n=item I<$name> ($short_name)\n";
+            $has_item = 1;
+        }
+
+        # And add the reason under the list of properties, if such a list
+        # actually got generated.  Note that the header got added
+        # unconditionally before.  But pod ignores extra blank lines, so no
+        # harm.
+        push @bad_re_properties, "\n$why\n" if $has_item;
+
+    } # End of looping through each reason.
+
+    # Generate a list of the properties whose map table we output, from the
+    # global @map_properties.
+    my @map_tables_actually_output;
+    my $info_indent = 20;       # Left column is narrower than \p{} table.
+    foreach my $property (@map_properties) {
+
+        # Get the path to the file; don't output any not in the standard
+        # directory.
+        my @path = $property->file_path;
+        next if $path[0] ne $map_directory;
+        shift @path;    # Remove the standard name
+
+        my $file = join '/', @path; # In case is in sub directory
+        my $info = $property->full_name;
+        my $short_name = $property->name;
+        if ($info ne $short_name) {
+            $info .= " ($short_name)";
+        }
+        foreach my $more_info ($property->description,
+                               $property->note,
+                               $property->status_info)
         {
-            my $File = $Exact{$Name};
-            $Name = $Name =~ m/\W/ ? qq/'$Name'/ : " $Name ";
-            my $Text = sprintf("%-15s => %s,\n", $Name, qq/'$File'/);
-            push @OUT, Text::Tabs::unexpand($Text);
+            next unless $more_info;
+            $info =~ s/\.\Z//;
+            $info .= ".  $more_info";
         }
-        push @OUT, ");\n1;\n";
+        push @map_tables_actually_output, format_pod_line($info_indent,
+                                                          $file,
+                                                          $info,
+                                                          $property->status);
+    }
 
-        WriteIfChanged('Exact.pl', @OUT);
+    # Sort alphabetically, and fold for output
+    @map_tables_actually_output = sort
+                            pod_alphanumeric_sort @map_tables_actually_output;
+    @map_tables_actually_output
+                        = simple_fold(\@map_tables_actually_output,
+                                        ' ',
+                                        $info_indent,
+                                        $automatic_pod_indent);
+
+    # Generate a list of the formats that can appear in the map tables.
+    my @map_table_formats;
+    foreach my $format (sort keys %map_table_formats) {
+        push @map_table_formats, " $format    $map_table_formats{$format}\n";
     }
 
-    ## Write Canonical.pl
-    {
-        my @OUT = (
-                   $HEADER,
-                   "##\n",
-                   "## Data in this file used by ../utf8_heavy.pl\n",
-                   "##\n\n",
-                   "## Mapping from lc(canonical name) to filename in ./lib\n",
-                   "%utf8::Canonical = (\n",
-                  );
-        my $Trail = ""; ## used just to keep the spacing pretty
-        for my $Name (sort keys %Canonical)
+    # Everything is ready to assemble.
+    my @OUT = << "END";
+=begin comment
+
+$HEADER
+
+To change this file, edit $0 instead.
+
+=end comment
+
+=head1 NAME
+
+$pod_file - Complete index of Unicode Version $string_version properties in
+the Perl core.
+
+=head1 DESCRIPTION
+
+There are many properties in Unicode, and Perl provides access to almost all of
+them, as well as some additional extensions and short-cut synonyms.
+
+And just about all of the few that aren't accessible through the Perl
+core are accessible through the modules: Unicode::Normalize and
+Unicode::UCD, and for Unihan properties, via the CPAN module Unicode::Unihan.
+
+This document merely lists all available properties and does not attempt to
+explain what each property really means.  There is a brief description of each
+Perl extension.  There is some detail about Blocks, Scripts, General_Category,
+and Bidi_Class in L<perlunicode>, but to find out about the intricacies of the
+Unicode properties, refer to the Unicode standard.  A good starting place is
+L<$unicode_reference_url>.  More information on the Perl extensions is in
+L<perlrecharclass>.
+
+Note that you can define your own properties; see
+L<perlunicode/"User-Defined Character Properties">.
+
+=head1 Properties accessible through \\p{} and \\P{}
+
+The Perl regular expression \\p{} and \\P{} constructs give access to most of
+the Unicode character properties.  The table below shows all these constructs,
+both single and compound forms.
+
+B<Compound forms> consist of two components, separated by an equals sign or a
+colon.  The first component is the property name, and the second component is
+the particular value of the property to match against, for example,
+'\\p{Script: Greek}' or '\\p{Script=Greek}' both mean to match characters
+whose Script property is Greek.
+
+B<Single forms>, like '\\p{Greek}', are mostly Perl-defined shortcuts for
+their equivalent compound forms.  The table shows these equivalences.  (In our
+example, '\\p{Greek}' is a just a shortcut for '\\p{Script=Greek}'.)
+There are also a few Perl-defined single forms that are not shortcuts for a
+compound form.  One such is \\p{Word}.  These are also listed in the table.
+
+In parsing these constructs, Perl always ignores Upper/lower case differences
+everywhere within the {braces}.  Thus '\\p{Greek}' means the same thing as
+'\\p{greek}'.  But note that changing the case of the 'p' or 'P' before the
+left brace completely changes the meaning of the construct, from "match" (for
+'\\p{}') to "doesn't match" (for '\\P{}').  Casing in this document is for
+improved legibility.
+
+Also, white space, hyphens, and underscores are also normally ignored
+everywhere between the {braces}, and hence can be freely added or removed
+even if the C</x> modifier hasn't been specified on the regular expression.
+But $a_bold_stricter at the beginning of an entry in the table below
+means that tighter (stricter) rules are used for that entry:
+
+=over 4
+
+=item Single form (\\p{name}) tighter rules:
+
+White space, hyphens, and underscores ARE significant
+except for:
+
+=over 4
+
+=item * white space adjacent to a non-word character
+
+=item * underscores separating digits in numbers
+
+=back
+
+That means, for example, that you can freely add or remove white space
+adjacent to (but within) the braces without affecting the meaning.
+
+=item Compound form (\\p{name=value} or \\p{name:value}) tighter rules:
+
+The tighter rules given above for the single form apply to everything to the
+right of the colon or equals; the looser rules still apply to everything to
+the left.
+
+That means, for example, that you can freely add or remove white space
+adjacent to (but within) the braces and the colon or equal sign.
+
+=back
+
+Some properties are considered obsolete, but still available.  There are
+several varieties of obsolesence:
+
+=over 4
+
+=item Obsolete
+
+Properties marked with $a_bold_obsolete in the table are considered
+obsolete.  At the time of this writing (Unicode version 5.2) there is no
+information in the Unicode standard about the implications of a property being
+obsolete.
+
+=item Stabilized
+
+Obsolete properties may be stabilized.  This means that they are not actively
+maintained by Unicode, and will not be extended as new characters are added to
+the standard.  Such properties are marked with $a_bold_stabilized in the
+table.  At the time of this writing (Unicode version 5.2) there is no further
+information in the Unicode standard about the implications of a property being
+stabilized.
+
+=item Deprecated
+
+Obsolete properties may be deprecated.  This means that their use is strongly
+discouraged, so much so that a warning will be issued if used, unless the
+regular expression is in the scope of a C<S<no warnings 'deprecated'>>
+statement.  $A_bold_deprecated flags each such entry in the table, and
+the entry there for the longest, most descriptive version of the property will
+give the reason it is deprecated, and perhaps advice.  Perl may issue such a
+warning, even for properties that aren't officially deprecated by Unicode,
+when there used to be characters or code points that were matched by them, but
+no longer.  This is to warn you that your program may not work like it did on
+earlier Unicode releases.
+
+A deprecated property may be made unavailable in a future Perl version, so it
+is best to move away from them.
+
+=back
+
+Some Perl extensions are present for backwards compatibility and are
+discouraged from being used, but not obsolete.  $A_bold_discouraged
+flags each such entry in the table.
+
+@block_warning
+
+The table below has two columns.  The left column contains the \\p{}
+constructs to look up, possibly preceeded by the flags mentioned above; and
+the right column contains information about them, like a description, or
+synonyms.  It shows both the single and compound forms for each property that
+has them.  If the left column is a short name for a property, the right column
+will give its longer, more descriptive name; and if the left column is the
+longest name, the right column will show any equivalent shortest name, in both
+single and compound forms if applicable.
+
+The right column will also caution you if a property means something different
+than what might normally be expected.
+
+Numbers in (parentheses) indicate the total number of code points matched by
+the property.  For emphasis, those properties that match no code points at all
+are listed as well in a separate section following the table.
+
+There is no description given for most non-Perl defined properties (See
+$unicode_reference_url for that).
+
+For compactness, 'B<*>' is used as a wildcard instead of showing all possible
+combinations.  For example, entries like:
+
+ \\p{Gc: *}                                  \\p{General_Category: *}
+
+mean that 'Gc' is a synonym for 'General_Category', and anything that is valid
+for the latter is also valid for the former.  Similarly,
+
+ \\p{Is_*}                                   \\p{*}
+
+means that if and only if, for example, \\p{Foo} exists, then \\p{Is_Foo} and
+\\p{IsFoo} are also valid and all mean the same thing.  And similarly,
+\\p{Foo=Bar} means the same as \\p{Is_Foo=Bar} and \\p{IsFoo=Bar}.  '*' here
+is restricted to something not beginning with an underscore.
+
+Also, in binary properties, 'Yes', 'T', and 'True' are all synonyms for 'Y'.
+And 'No', 'F', and 'False' are all synonyms for 'N'.  The table shows 'Y*' and
+'N*' to indicate this, and doesn't have separate entries for the other
+possibilities.  Note that not all properties which have values 'Yes' and 'No'
+are binary, and they have all their values spelled out without using this wild
+card, and a C<NOT> clause in their description that highlights their not being
+binary.  These also require the compound form to match them, whereas true
+binary properties have both single and compound forms available.
+
+Note that all non-essential underscores are removed in the display of the
+short names below.
+
+B<Summary legend:>
+
+=over 4
+
+=item B<*> is a wild-card
+
+=item B<(\\d+)> in the info column gives the number of code points matched by
+this property.
+
+=item B<$DEPRECATED> means this is deprecated.
+
+=item B<$OBSOLETE> means this is obsolete.
+
+=item B<$STABILIZED> means this is stabilized.
+
+=item B<$STRICTER> means tighter (stricter) name matching applies.
+
+=item B<$DISCOURAGED> means use of this form is discouraged.
+
+=back
+
+$formatted_properties
+
+$zero_matches
+
+=head1 Properties not accessible through \\p{} and \\P{}
+
+A few properties are accessible in Perl via various function calls only.
+These are:
+ Lowercase_Mapping          lc() and lcfirst()
+ Titlecase_Mapping          ucfirst()
+ Uppercase_Mapping          uc()
+
+Case_Folding is accessible through the /i modifier in regular expressions.
+
+The Name property is accessible through the \\N{} interpolation in
+double-quoted strings and regular expressions, but both usages require a C<use
+charnames;> to be specified, which also contains related functions viacode()
+and vianame().
+
+=head1 Unicode regular expression properties that are NOT accepted by Perl
+
+Perl will generate an error for a few character properties in Unicode when
+used in a regular expression.  The non-Unihan ones are listed below, with the
+reasons they are not accepted, perhaps with work-arounds.  The short names for
+the properties are listed enclosed in (parentheses).
+
+=over 4
+
+@bad_re_properties
+
+=back
+
+An installation can choose to allow any of these to be matched by changing the
+controlling lists contained in the program C<\$Config{privlib}>/F<unicore/$0>
+and then re-running F<$0>.  (C<\%Config> is available from the Config module).
+
+=head1 Files in the I<To> directory (for serious hackers only)
+
+All Unicode properties are really mappings (in the mathematical sense) from
+code points to their respective values.  As part of its build process,
+Perl constructs tables containing these mappings for all properties that it
+deals with.  But only a few of these are written out into files.
+Those written out are in the directory C<\$Config{privlib}>/F<unicore/To/>
+(%Config is available from the Config module).
+
+Those ones written are ones needed by Perl internally during execution, or for
+which there is some demand, and those for which there is no access through the
+Perl core.  Generally, properties that can be used in regular expression
+matching do not have their map tables written, like Script.  Nor are the
+simplistic properties that have a better, more complete version, such as
+Simple_Uppercase_Mapping  (Uppercase_Mapping is written instead).
+
+None of the properties in the I<To> directory are currently directly
+accessible through the Perl core, although some may be accessed indirectly.
+For example, the uc() function implements the Uppercase_Mapping property and
+uses the F<Upper.pl> file found in this directory.
+
+The available files with their properties (short names in parentheses),
+and any flags or comments about them, are:
+
+@map_tables_actually_output
+
+An installation can choose to change which files are generated by changing the
+controlling lists contained in the program C<\$Config{privlib}>/F<unicore/$0>
+and then re-running F<$0>.
+
+Each of these files defines two hash entries to help reading programs decipher
+it.  One of them looks like this:
+
+    \$utf8::SwashInfo{'ToNAME'}{'format'} = 's';
+
+where 'NAME' is a name to indicate the property.  For backwards compatibility,
+this is not necessarily the property's official Unicode name.  (The 'To' is
+also for backwards compatibility.)  The hash entry gives the format of the
+mapping fields of the table, currently one of the following:
+
+ @map_table_formats
+
+This format applies only to the entries in the main body of the table.
+Entries defined in hashes or ones that are missing from the list can have a
+different format.
+
+The value that the missing entries have is given by the other SwashInfo hash
+entry line; it looks like this:
+
+    \$utf8::SwashInfo{'ToNAME'}{'missing'} = 'NaN';
+
+This example line says that any Unicode code points not explicitly listed in
+the file have the value 'NaN' under the property indicated by NAME.  If the
+value is the special string C<< <code point> >>, it means that the value for
+any missing code point is the code point itself.  This happens, for example,
+in the file for Uppercase_Mapping (To/Upper.pl), in which code points like the
+character 'A', are missing because the uppercase of 'A' is itself.
+
+=head1 SEE ALSO
+
+L<$unicode_reference_url>
+
+L<perlrecharclass>
+
+L<perlunicode>
+
+END
+
+    # And write it.
+    main::write([ $pod_directory, "$pod_file.pod" ], @OUT);
+    return;
+}
+
+sub make_Heavy () {
+    # Create and write Heavy.pl, which passes info about the tables to
+    # utf8_heavy.pl
+
+    my @heavy = <<END;
+$HEADER
+$INTERNAL_ONLY
+
+# This file is for the use of utf8_heavy.pl
+
+# Maps property names in loose standard form to its standard name
+\%utf8::loose_property_name_of = (
+END
+
+    push @heavy, simple_dumper (\%loose_property_name_of, ' ' x 4);
+    push @heavy, <<END;
+);
+
+# Maps property, table to file for those using stricter matching
+\%utf8::stricter_to_file_of = (
+END
+    push @heavy, simple_dumper (\%stricter_to_file_of, ' ' x 4);
+    push @heavy, <<END;
+);
+
+# Maps property, table to file for those using loose matching
+\%utf8::loose_to_file_of = (
+END
+    push @heavy, simple_dumper (\%loose_to_file_of, ' ' x 4);
+    push @heavy, <<END;
+);
+
+# Maps floating point to fractional form
+\%utf8::nv_floating_to_rational = (
+END
+    push @heavy, simple_dumper (\%nv_floating_to_rational, ' ' x 4);
+    push @heavy, <<END;
+);
+
+# If a floating point number doesn't have enough digits in it to get this
+# close to a fraction, it isn't considered to be that fraction even if all the
+# digits it does have match.
+\$utf8::max_floating_slop = $MAX_FLOATING_SLOP;
+
+# Deprecated tables to generate a warning for.  The key is the file containing
+# the table, so as to avoid duplication, as many property names can map to the
+# file, but we only need one entry for all of them.
+\%utf8::why_deprecated = (
+END
+
+    push @heavy, simple_dumper (\%utf8::why_deprecated, ' ' x 4);
+    push @heavy, <<END;
+);
+
+1;
+END
+
+    main::write("Heavy.pl", @heavy);
+    return;
+}
+
+sub write_all_tables() {
+    # Write out all the tables generated by this program to files, as well as
+    # the supporting data structures, pod file, and .t file.
+
+    my @writables;              # List of tables that actually get written
+    my %match_tables_to_write;  # Used to collapse identical match tables
+                                # into one file.  Each key is a hash function
+                                # result to partition tables into buckets.
+                                # Each value is an array of the tables that
+                                # fit in the bucket.
+
+    # For each property ...
+    # (sort so that if there is an immutable file name, it has precedence, so
+    # some other property can't come in and take over its file name.  If b's
+    # file name is defined, will return 1, meaning to take it first; don't
+    # care if both defined, as they had better be different anyway)
+    PROPERTY:
+    foreach my $property (sort { defined $b->file } property_ref('*')) {
+        my $type = $property->type;
+
+        # And for each table for that property, starting with the mapping
+        # table for it ...
+        TABLE:
+        foreach my $table($property,
+
+                        # and all the match tables for it (if any), sorted so
+                        # the ones with the shortest associated file name come
+                        # first.  The length sorting prevents problems of a
+                        # longer file taking a name that might have to be used
+                        # by a shorter one.  The alphabetic sorting prevents
+                        # differences between releases
+                        sort {  my $ext_a = $a->external_name;
+                                return 1 if ! defined $ext_a;
+                                my $ext_b = $b->external_name;
+                                return -1 if ! defined $ext_b;
+                                my $cmp = length $ext_a <=> length $ext_b;
+
+                                # Return result if lengths not equal
+                                return $cmp if $cmp;
+
+                                # Alphabetic if lengths equal
+                                return $ext_a cmp $ext_b
+                        } $property->tables
+                    )
         {
-            my $File = $Canonical{$Name};
-            if ($CaComment{$Name}) {
-                push @OUT, "\n" if not $Trail;
-                push @OUT, " # $CaComment{$Name}\n";
-                $Trail = "\n";
-            } else {
-                $Trail = "";
+
+            # Here we have a table associated with a property.  It could be
+            # the map table (done first for each property), or one of the
+            # other tables.  Determine which type.
+            my $is_property = $table->isa('Property');
+
+            my $name = $table->name;
+            my $complete_name = $table->complete_name;
+
+            # See if should suppress the table if is empty, but warn if it
+            # contains something.
+            my $suppress_if_empty_warn_if_not = grep { $complete_name eq $_ }
+                                    keys %why_suppress_if_empty_warn_if_not;
+
+            # Calculate if this table should have any code points associated
+            # with it or not.
+            my $expected_empty =
+
+                # $perl should be empty, as well as properties that we just
+                # don't do anything with
+                ($is_property
+                    && ($table == $perl
+                        || grep { $complete_name eq $_ }
+                                                    @unimplemented_properties
+                    )
+                )
+
+                # Match tables in properties we skipped populating should be
+                # empty
+                || (! $is_property && ! $property->to_create_match_tables)
+
+                # Tables and properties that are expected to have no code
+                # points should be empty
+                || $suppress_if_empty_warn_if_not
+            ;
+
+            # Set a boolean if this table is the complement of an empty binary
+            # table
+            my $is_complement_of_empty_binary =
+                $type == $BINARY &&
+                (($table == $property->table('Y')
+                    && $property->table('N')->is_empty)
+                || ($table == $property->table('N')
+                    && $property->table('Y')->is_empty));
+
+
+            # Some tables should match everything
+            my $expected_full =
+                ($is_property)
+                ? # All these types of map tables will be full because
+                  # they will have been populated with defaults
+                  ($type == $ENUM || $type == $BINARY)
+
+                : # A match table should match everything if its method
+                  # shows it should
+                  ($table->matches_all
+
+                  # The complement of an empty binary table will match
+                  # everything
+                  || $is_complement_of_empty_binary
+                  )
+            ;
+
+            if ($table->is_empty) {
+
+
+                if ($suppress_if_empty_warn_if_not) {
+                    $table->set_status($SUPPRESSED,
+                        $why_suppress_if_empty_warn_if_not{$complete_name});
+                }
+
+                # Suppress expected empty tables.
+                next TABLE if $expected_empty;
+
+                # And setup to later output a warning for those that aren't
+                # known to be allowed to be empty.  Don't do the warning if
+                # this table is a child of another one to avoid duplicating
+                # the warning that should come from the parent one.
+                if (($table == $property || $table->parent == $table)
+                    && $table->status ne $SUPPRESSED
+                    && ! grep { $complete_name =~ /^$_$/ }
+                                                    @tables_that_may_be_empty)
+                {
+                    push @unhandled_properties, "$table";
+                }
+            }
+            elsif ($expected_empty) {
+                my $because = "";
+                if ($suppress_if_empty_warn_if_not) {
+                    $because = " because $why_suppress_if_empty_warn_if_not{$complete_name}";
+                }
+
+                Carp::my_carp("Not expecting property $table$because.  Generating file for it anyway.");
+            }
+
+            my $count = $table->count;
+            if ($expected_full) {
+                if ($count != $MAX_UNICODE_CODEPOINTS) {
+                    Carp::my_carp("$table matches only "
+                    . clarify_number($count)
+                    . " Unicode code points but should match "
+                    . clarify_number($MAX_UNICODE_CODEPOINTS)
+                    . " (off by "
+                    .  clarify_number(abs($MAX_UNICODE_CODEPOINTS - $count))
+                    . ").  Proceeding anyway.");
+                }
+
+                # Here is expected to be full.  If it is because it is the
+                # complement of an (empty) binary table that is to be
+                # suppressed, then suppress this one as well.
+                if ($is_complement_of_empty_binary) {
+                    my $opposing_name = ($name eq 'Y') ? 'N' : 'Y';
+                    my $opposing = $property->table($opposing_name);
+                    my $opposing_status = $opposing->status;
+                    if ($opposing_status) {
+                        $table->set_status($opposing_status,
+                                           $opposing->status_info);
+                    }
+                }
+            }
+            elsif ($count == $MAX_UNICODE_CODEPOINTS) {
+                if ($table == $property || $table->leader == $table) {
+                    Carp::my_carp("$table unexpectedly matches all Unicode code points.  Proceeding anyway.");
+                }
+            }
+
+            if ($table->status eq $SUPPRESSED) {
+                if (! $is_property) {
+                    my @children = $table->children;
+                    foreach my $child (@children) {
+                        if ($child->status ne $SUPPRESSED) {
+                            Carp::my_carp_bug("'$table' is suppressed and has a child '$child' which isn't");
+                        }
+                    }
+                }
+                next TABLE;
+
+            }
+            if (! $is_property) {
+
+                # Several things need to be done just once for each related
+                # group of match tables.  Do them on the parent.
+                if ($table->parent == $table) {
+
+                    # Add an entry in the pod file for the table; it also does
+                    # the children.
+                    make_table_pod_entries($table);
+
+                    # See if the the table matches identical code points with
+                    # something that has already been output.  In that case,
+                    # no need to have two files with the same code points in
+                    # them.  We use the table's hash() method to store these
+                    # in buckets, so that it is quite likely that if two
+                    # tables are in the same bucket they will be identical, so
+                    # don't have to compare tables frequently.  The tables
+                    # have to have the same status to share a file, so add
+                    # this to the bucket hash.  (The reason for this latter is
+                    # that Heavy.pl associates a status with a file.)
+                    my $hash = $table->hash . ';' . $table->status;
+
+                    # Look at each table that is in the same bucket as this
+                    # one would be.
+                    foreach my $comparison (@{$match_tables_to_write{$hash}})
+                    {
+                        if ($table->matches_identically_to($comparison)) {
+                            $table->set_equivalent_to($comparison,
+                                                                Related => 0);
+                            next TABLE;
+                        }
+                    }
+
+                    # Here, not equivalent, add this table to the bucket.
+                    push @{$match_tables_to_write{$hash}}, $table;
+                }
             }
-            $Name = $Name =~ m/\W/ ? qq/'$Name'/ : " $Name ";
-            my $Text = sprintf("  %-41s => %s,\n$Trail", $Name, qq/'$File'/);
-            push @OUT, Text::Tabs::unexpand($Text);
+            else {
+
+                # Here is the property itself.
+                # Don't write out or make references to the $perl property
+                next if $table == $perl;
+
+                if ($type != $STRING) {
+
+                    # There is a mapping stored of the various synonyms to the
+                    # standardized name of the property for utf8_heavy.pl.
+                    # Also, the pod file contains entries of the form:
+                    # \p{alias: *}         \p{full: *}
+                    # rather than show every possible combination of things.
+
+                    my @property_aliases = $property->aliases;
+
+                    # The full name of this property is stored by convention
+                    # first in the alias array
+                    my $full_property_name =
+                                '\p{' . $property_aliases[0]->name . ': *}';
+                    my $standard_property_name = standardize($table->name);
+
+                    # For each synonym ...
+                    for my $i (0 .. @property_aliases - 1)  {
+                        my $alias = $property_aliases[$i];
+                        my $alias_name = $alias->name;
+                        my $alias_standard = standardize($alias_name);
+
+                        # Set the mapping for utf8_heavy of the alias to the
+                        # property
+                        if (exists ($loose_property_name_of{$alias_standard}))
+                        {
+                            Carp::my_carp("There already is a property with the same standard name as $alias_name: $loose_property_name_of{$alias_standard}.  Old name is retained");
+                        }
+                        else {
+                            $loose_property_name_of{$alias_standard}
+                                                = $standard_property_name;
+                        }
+
+                        # Now for the pod entry for this alias.  Skip
+                        # the first one, which is the full name so won't have
+                        # an entry like: '\p{full: *}   \p{full: *}', and skip
+                        # if don't want an entry for this one.
+                        next if $i == 0 || ! $alias->make_pod_entry;
+
+                        push @match_properties,
+                            format_pod_line($indent_info_column,
+                                        '\p{' . $alias->name . ': *}',
+                                        $full_property_name,
+                                        $alias->status);
+                    }
+                } # End of non-string-like property code
+
+
+                # Don't output a mapping file if not desired.
+                next if ! $property->to_output_map;
+            }
+
+            # Here, we know we want to write out the table, but don't do it
+            # yet because there may be other tables that come along and will
+            # want to share the file, and the file's comments will change to
+            # mention them.  So save for later.
+            push @writables, $table;
+
+        } # End of looping through the property and all its tables.
+    } # End of looping through all properties.
+
+    # Now have all the tables that will have files written for them.  Do it.
+    foreach my $table (@writables) {
+        my @directory;
+        my $filename;
+        my $property = $table->property;
+        my $is_property = ($table == $property);
+        if (! $is_property) {
+
+            # Match tables for the property go in lib/$subdirectory, which is
+            # the property's name.  Don't use the standard file name for this,
+            # as may get an unfamiliar alias
+            @directory = ($matches_directory, $property->external_name);
         }
-        push @OUT, ");\n1\n";
-        WriteIfChanged('Canonical.pl', @OUT);
+        else {
+
+            @directory = $table->directory;
+            $filename = $table->file;
+        }
+
+        # Use specified filename if avaliable, or default to property's
+        # shortest name.  We need an 8.3 safe filename (which means "an 8
+        # safe" filename, since after the dot is only 'pl', which is < 3)
+        # The 2nd parameter is if the filename shouldn't be changed, and
+        # it shouldn't iff there is a hard-coded name for this table.
+        $filename = construct_filename(
+                                $filename || $table->external_name,
+                                ! $filename,    # mutable if no filename
+                                \@directory);
+
+        register_file_for_name($table, \@directory, $filename);
+
+        # Only need to write one file when shared by more than one
+        # property
+        next if ! $is_property && $table->leader != $table;
+
+        # Construct a nice comment to add to the file
+        $table->set_final_comment;
+
+        $table->write;
     }
 
-    MakePropTestScript() if $MakeTestScript;
+
+    # Write out the pod file
+    make_pod;
+
+    # And Heavy.pl
+    make_Heavy;
+
+    make_property_test_script() if $make_test_script;
+    return;
 }
 
+my @white_space_separators = ( # This used only for making the test script.
+                            "",
+                            ' ',
+                            "\t",
+                            '   '
+                        );
 
-sub SpecialCasing_txt()
-{
-    #
-    # Read in the special cases.
-    #
+sub generate_separator($) {
+    # This used only for making the test script.  It generates the colon or
+    # equal separator between the property and property value, with random
+    # white space surrounding the separator
+
+    my $lhs = shift;
+
+    return "" if $lhs eq "";  # No separator if there's only one (the r) side
 
-    my %CaseInfo;
+    # Choose space before and after randomly
+    my $spaces_before =$white_space_separators[rand(@white_space_separators)];
+    my $spaces_after = $white_space_separators[rand(@white_space_separators)];
 
-    if (not open IN, "SpecialCasing.txt") {
-        die "$0: SpecialCasing.txt: $!\n";
+    # And return the whole complex, half the time using a colon, half the
+    # equals
+    return $spaces_before
+            . (rand() < 0.5) ? '=' : ':'
+            . $spaces_after;
+}
+
+sub generate_tests($$$$$$) {
+    # This used only for making the test script.  It generates test cases that
+    # are expected to compile successfully in perl.  Note that the lhs and
+    # rhs are assumed to already be as randomized as the caller wants.
+
+    my $file_handle = shift;   # Where to output the tests
+    my $lhs = shift;           # The property: what's to the left of the colon
+                               #  or equals separator
+    my $rhs = shift;           # The property value; what's to the right
+    my $valid_code = shift;    # A code point that's known to be in the
+                               # table given by lhs=rhs; undef if table is
+                               # empty
+    my $invalid_code = shift;  # A code point known to not be in the table;
+                               # undef if the table is all code points
+    my $warning = shift;
+
+    # Get the colon or equal
+    my $separator = generate_separator($lhs);
+
+    # The whole 'property=value'
+    my $name = "$lhs$separator$rhs";
+
+    # Create a complete set of tests, with complements.
+    if (defined $valid_code) {
+        printf $file_handle
+                    qq/Expect(1, $valid_code, '\\p{$name}', $warning);\n/;
+        printf $file_handle
+                    qq/Expect(0, $valid_code, '\\p{^$name}', $warning);\n/;
+        printf $file_handle
+                    qq/Expect(0, $valid_code, '\\P{$name}', $warning);\n/;
+        printf $file_handle
+                    qq/Expect(1, $valid_code, '\\P{^$name}', $warning);\n/;
     }
-    while (<IN>) {
-        next unless /^[0-9A-Fa-f]+;/;
-        s/\#.*//;
-        s/\s+$//;
+    if (defined $invalid_code) {
+        printf $file_handle
+                    qq/Expect(0, $invalid_code, '\\p{$name}', $warning);\n/;
+        printf $file_handle
+                    qq/Expect(1, $invalid_code, '\\p{^$name}', $warning);\n/;
+        printf $file_handle
+                    qq/Expect(1, $invalid_code, '\\P{$name}', $warning);\n/;
+        printf $file_handle
+                    qq/Expect(0, $invalid_code, '\\P{^$name}', $warning);\n/;
+    }
+    return;
+}
 
-        my ($code, $lower, $title, $upper, $condition) = split(/\s*;\s*/);
+sub generate_error($$$$) {
+    # This used only for making the test script.  It generates test cases that
+    # are expected to not only not match, but to be syntax or similar errors
+
+    my $file_handle = shift;        # Where to output to.
+    my $lhs = shift;                # The property: what's to the left of the
+                                    # colon or equals separator
+    my $rhs = shift;                # The property value; what's to the right
+    my $already_in_error = shift;   # Boolean; if true it's known that the
+                                # unmodified lhs and rhs will cause an error.
+                                # This routine should not force another one
+    # Get the colon or equal
+    my $separator = generate_separator($lhs);
+
+    # Since this is an error only, don't bother to randomly decide whether to
+    # put the error on the left or right side; and assume that the rhs is
+    # loosely matched, again for convenience rather than rigor.
+    $rhs = randomize_loose_name($rhs, 'ERROR') unless $already_in_error;
+
+    my $property = $lhs . $separator . $rhs;
+
+    print $file_handle qq/Error('\\p{$property}');\n/;
+    print $file_handle qq/Error('\\P{$property}');\n/;
+    return;
+}
 
-        if ($condition) { # not implemented yet
-            print "# SKIPPING $_\n" if $Verbose;
-            next;
+# These are used only for making the test script
+# XXX Maybe should also have a bad strict seps, which includes underscore.
+
+my @good_loose_seps = (
+            " ",
+            "-",
+            "\t",
+            "",
+            "_",
+           );
+my @bad_loose_seps = (
+           "/a/",
+           ':=',
+          );
+
+sub randomize_stricter_name {
+    # This used only for making the test script.  Take the input name and
+    # return a randomized, but valid version of it under the stricter matching
+    # rules.
+
+    my $name = shift;
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+    # If the name looks like a number (integer, floating, or rational), do
+    # some extra work
+    if ($name =~ qr{ ^ ( -? ) (\d+ ( ( [./] ) \d+ )? ) $ }x) {
+        my $sign = $1;
+        my $number = $2;
+        my $separator = $3;
+
+        # If there isn't a sign, part of the time add a plus
+        # Note: Not testing having any denominator having a minus sign
+        if (! $sign) {
+            $sign = '+' if rand() <= .3;
+        }
+
+        # And add 0 or more leading zeros.
+        $name = $sign . ('0' x int rand(10)) . $number;
+
+        if (defined $separator) {
+            my $extra_zeros = '0' x int rand(10);
+
+            if ($separator eq '.') {
+
+                # Similarly, add 0 or more trailing zeros after a decimal
+                # point
+                $name .= $extra_zeros;
+            }
+            else {
+
+                # Or, leading zeros before the denominator
+                $name =~ s,/,/$extra_zeros,;
+            }
         }
+    }
 
-        # Wait until all the special cases have been read since
-        # they are not listed in numeric order.
-        my $ix = hex($code);
-        push @{$CaseInfo{Lower}}, [ $ix, $code, $lower ]
-	    unless $code eq $lower;
-        push @{$CaseInfo{Title}}, [ $ix, $code, $title ]
-	    unless $code eq $title;
-        push @{$CaseInfo{Upper}}, [ $ix, $code, $upper ]
-	    unless $code eq $upper;
+    # For legibility of the test, only change the case of whole sections at a
+    # time.  To do this, first split into sections.  The split returns the
+    # delimiters
+    my @sections;
+    for my $section (split / ( [ - + \s _ . ]+ ) /x, $name) {
+        trace $section if main::DEBUG && $to_trace;
+
+        if (length $section > 1 && $section !~ /\D/) {
+
+            # If the section is a sequence of digits, about half the time
+            # randomly add underscores between some of them.
+            if (rand() > .5) {
+
+                # Figure out how many underscores to add.  max is 1 less than
+                # the number of digits.  (But add 1 at the end to make sure
+                # result isn't 0, and compensate earlier by subtracting 2
+                # instead of 1)
+                my $num_underscores = int rand(length($section) - 2) + 1;
+
+                # And add them evenly throughout, for convenience, not rigor
+                use integer;
+                my $spacing = (length($section) - 1)/ $num_underscores;
+                my $temp = $section;
+                $section = "";
+                for my $i (1 .. $num_underscores) {
+                    $section .= substr($temp, 0, $spacing, "") . '_';
+                }
+                $section .= $temp;
+            }
+            push @sections, $section;
+        }
+        else {
+
+            # Here not a sequence of digits.  Change the case of the section
+            # randomly
+            my $switch = int rand(4);
+            if ($switch == 0) {
+                push @sections, uc $section;
+            }
+            elsif ($switch == 1) {
+                push @sections, lc $section;
+            }
+            elsif ($switch == 2) {
+                push @sections, ucfirst $section;
+            }
+            else {
+                push @sections, $section;
+            }
+        }
     }
-    close IN;
+    trace "returning", join "", @sections if main::DEBUG && $to_trace;
+    return join "", @sections;
+}
+
+sub randomize_loose_name($;$) {
+    # This used only for making the test script
 
-    # Now write out the special cases properties in their code point order.
-    # Prepend them to the To/{Upper,Lower,Title}.pl.
+    my $name = shift;
+    my $want_error = shift;  # if true, make an error
+    Carp::carp_extra_args(\@_) if main::DEBUG && @_;
+
+    $name = randomize_stricter_name($name);
+
+    my @parts;
+    push @parts, $good_loose_seps[rand(@good_loose_seps)];
+    for my $part (split /[-\s_]+/, $name) {
+        if (@parts) {
+            if ($want_error and rand() < 0.3) {
+                push @parts, $bad_loose_seps[rand(@bad_loose_seps)];
+                $want_error = 0;
+            }
+            else {
+                push @parts, $good_loose_seps[rand(@good_loose_seps)];
+            }
+        }
+        push @parts, $part;
+    }
+    my $new = join("", @parts);
+    trace "$name => $new" if main::DEBUG && $to_trace;
+
+    if ($want_error) {
+        if (rand() >= 0.5) {
+            $new .= $bad_loose_seps[rand(@bad_loose_seps)];
+        }
+        else {
+            $new = $bad_loose_seps[rand(@bad_loose_seps)] . $new;
+        }
+    }
+    return $new;
+}
+
+# Used to make sure don't generate duplicate test cases.
+my %test_generated;
+
+sub make_property_test_script() {
+    # This used only for making the test script
+    # this written directly -- it's huge.
+
+    print "Making test script\n" if $verbosity >= $PROGRESS;
+
+    # This uses randomness to test different possibilities without testing all
+    # possibilities.  To ensure repeatability, set the seed to 0.  But if
+    # tests are added, it will perturb all later ones in the .t file
+    srand 0;
+
+    force_unlink ($t_path);
+    push @files_actually_output, $t_path;
+    my $OUT;
+    if (not open $OUT, "> $t_path") {
+        Carp::my_carp("Can't open $t_path.  Skipping: $!");
+        return;
+    }
 
-    for my $case (qw(Lower Title Upper))
+    # Keep going down an order of magnitude
+    # until find that adding this quantity to
+    # 1 remains 1; but put an upper limit on
+    # this so in case this algorithm doesn't
+    # work properly on some platform, that we
+    # won't loop forever.
+    my $digits = 0;
+    my $min_floating_slop = 1;
+    while (1+ $min_floating_slop != 1
+            && $digits++ < 50)
     {
-        my $NormalCase = do "To/$case.pl" || die "$0: $@\n";
-
-        my @OUT =
-	    (
-	     $HEADER, $INTERNAL_ONLY, "\n",
-	     "# The key: UTF-8 _bytes_, the value: UTF-8 (speed hack)\n",
-	     "%utf8::ToSpec$case =\n(\n",
-	    );
-
-        for my $prop (sort { $a->[0] <=> $b->[0] } @{$CaseInfo{$case}}) {
-            my ($ix, $code, $to) = @$prop;
-            my $tostr =
-              join "", map { sprintf "\\x{%s}", $_ } split ' ', $to;
-            push @OUT, sprintf qq["%s" => "$tostr",\n], join("", map { sprintf "\\x%02X", $_ } unpack("U0C*", pack("U", $ix)));
-	    # Remove any single-character mappings for
-	    # the same character since we are going for
-	    # the special casing rules.
-	    $NormalCase =~ s/^$code\t\t\w+\n//m;
-        }
-        push @OUT, (
-                    ");\n\n",
-                    "return <<'END';\n",
-                    $NormalCase,
-                    "END\n"
-                    );
-        WriteIfChanged(["To","$case.pl"], @OUT);
+        my $next = $min_floating_slop / 10;
+        last if $next == 0; # If underflows,
+                            # use previous one
+        $min_floating_slop = $next;
     }
+    print $OUT $HEADER, <DATA>;
+
+    foreach my $property (property_ref('*')) {
+        foreach my $table ($property->tables) {
+
+            # Find code points that match, and don't match this table.
+            my $valid = $table->get_valid_code_point;
+            my $invalid = $table->get_invalid_code_point;
+            my $warning = ($table->status eq $DEPRECATED)
+                            ? "'deprecated'"
+                            : '""';
+
+            # Test each possible combination of the property's aliases with
+            # the table's.  If this gets to be too many, could do what is done
+            # in the set_final_comment() for Tables
+            my @table_aliases = $table->aliases;
+            my @property_aliases = $table->property->aliases;
+            my $max = max(scalar @table_aliases, scalar @property_aliases);
+            for my $j (0 .. $max - 1) {
+
+                # The current alias for property is the next one on the list,
+                # or if beyond the end, start over.  Similarly for table
+                my $property_name
+                            = $property_aliases[$j % @property_aliases]->name;
+
+                $property_name = "" if $table->property == $perl;
+                my $table_alias = $table_aliases[$j % @table_aliases];
+                my $table_name = $table_alias->name;
+                my $loose_match = $table_alias->loose_match;
+
+                # If the table doesn't have a file, any test for it is
+                # already guaranteed to be in error
+                my $already_error = ! $table->file_path;
+
+                # Generate error cases for this alias.
+                generate_error($OUT,
+                                $property_name,
+                                $table_name,
+                                $already_error);
+
+                # If the table is guaranteed to always generate an error,
+                # quit now without generating success cases.
+                next if $already_error;
+
+                # Now for the success cases.
+                my $random;
+                if ($loose_match) {
+
+                    # For loose matching, create an extra test case for the
+                    # standard name.
+                    my $standard = standardize($table_name);
+
+                    # $test_name should be a unique combination for each test
+                    # case; used just to avoid duplicate tests
+                    my $test_name = "$property_name=$standard";
+
+                    # Don't output duplicate test cases.
+                    if (! exists $test_generated{$test_name}) {
+                        $test_generated{$test_name} = 1;
+                        generate_tests($OUT,
+                                        $property_name,
+                                        $standard,
+                                        $valid,
+                                        $invalid,
+                                        $warning,
+                                    );
+                    }
+                    $random = randomize_loose_name($table_name)
+                }
+                else { # Stricter match
+                    $random = randomize_stricter_name($table_name);
+                }
+
+                # Now for the main test case for this alias.
+                my $test_name = "$property_name=$random";
+                if (! exists $test_generated{$test_name}) {
+                    $test_generated{$test_name} = 1;
+                    generate_tests($OUT,
+                                    $property_name,
+                                    $random,
+                                    $valid,
+                                    $invalid,
+                                    $warning,
+                                );
+
+                    # If the name is a rational number, add tests for the
+                    # floating point equivalent.
+                    if ($table_name =~ qr{/}) {
+
+                        # Calculate the float, and find just the fraction.
+                        my $float = eval $table_name;
+                        my ($whole, $fraction)
+                                            = $float =~ / (.*) \. (.*) /x;
+
+                        # Starting with one digit after the decimal point,
+                        # create a test for each possible precision (number of
+                        # digits past the decimal point) until well beyond the
+                        # native number found on this machine.  (If we started
+                        # with 0 digits, it would be an integer, which could
+                        # well match an unrelated table)
+                        PLACE:
+                        for my $i (1 .. $min_floating_slop + 3) {
+                            my $table_name = sprintf("%.*f", $i, $float);
+                            if ($i < $MIN_FRACTION_LENGTH) {
+
+                                # If the test case has fewer digits than the
+                                # minimum acceptable precision, it shouldn't
+                                # succeed, so we expect an error for it.
+                                # E.g., 2/3 = .7 at one decimal point, and we
+                                # shouldn't say it matches .7.  We should make
+                                # it be .667 at least before agreeing that the
+                                # intent was to match 2/3.  But at the
+                                # less-than- acceptable level of precision, it
+                                # might actually match an unrelated number.
+                                # So don't generate a test case if this
+                                # conflating is possible.  In our example, we
+                                # don't want 2/3 matching 7/10, if there is
+                                # a 7/10 code point.
+                                for my $existing
+                                        (keys %nv_floating_to_rational)
+                                {
+                                    next PLACE
+                                        if abs($table_name - $existing)
+                                                < $MAX_FLOATING_SLOP;
+                                }
+                                generate_error($OUT,
+                                            $property_name,
+                                            $table_name,
+                                            1   # 1 => already an error
+                                );
+                            }
+                            else {
+
+                                # Here the number of digits exceeds the
+                                # minimum we think is needed.  So generate a
+                                # success test case for it.
+                                generate_tests($OUT,
+                                                $property_name,
+                                                $table_name,
+                                                $valid,
+                                                $invalid,
+                                                $warning,
+                                );
+                            }
+                        }
+                    }
+                }
+            }
+        }
+    }
+    print $OUT "Finished();\n";
+    close $OUT;
+    return;
 }
 
+# This is a list of the input files and how to handle them.  The files are
+# processed in their order in this list.  Some reordering is possible if
+# desired, but the v0 files should be first, and the extracted before the
+# others except DAge.txt (as data in an extracted file can be over-ridden by
+# the non-extracted.  Some other files depend on data derived from an earlier
+# file, like UnicodeData requires data from Jamo, and the case changing and
+# folding requires data from Unicode.  Mostly, it safest to order by first
+# version releases in (except the Jamo).  DAge.txt is read before the
+# extracted ones because of the rarely used feature $compare_versions.  In the
+# unlikely event that there were ever an extracted file that contained the Age
+# property information, it would have to go in front of DAge.
 #
-# Read in the case foldings.
-#
-# We will do full case folding, C + F + I (see CaseFolding.txt).  Note that
-# there are no I entries starting with Unicode 3.2, but leaving it in allows
-# for backward compatibility.
-#
-sub CaseFolding_txt()
-{
-    if (not open IN, "CaseFolding.txt") {
-	die "$0: CaseFolding.txt: $!\n";
+# The version strings allow the program to know whether to expect a file or
+# not, but if a file exists in the directory, it will be processed, even if it
+# is in a version earlier than expected, so you can copy files from a later
+# release into an earlier release's directory.
+my @input_file_objects = (
+    Input_file->new('PropertyAliases.txt', v0,
+                    Handler => \&process_PropertyAliases,
+                    ),
+    Input_file->new(undef, v0,  # No file associated with this
+                    Progress_Message => 'Finishing Property Setup',
+                    Handler => \&finish_property_setup,
+                    ),
+    Input_file->new('PropValueAliases.txt', v0,
+                     Handler => \&process_PropValueAliases,
+                     Has_Missings_Defaults => $NOT_IGNORED,
+                     ),
+    Input_file->new('DAge.txt', v3.2.0,
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    Property => 'Age'
+                    ),
+    Input_file->new("${EXTRACTED}DGeneralCategory.txt", v3.1.0,
+                    Property => 'General_Category',
+                    ),
+    Input_file->new("${EXTRACTED}DCombiningClass.txt", v3.1.0,
+                    Property => 'Canonical_Combining_Class',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+    Input_file->new("${EXTRACTED}DNumType.txt", v3.1.0,
+                    Property => 'Numeric_Type',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+    Input_file->new("${EXTRACTED}DEastAsianWidth.txt", v3.1.0,
+                    Property => 'East_Asian_Width',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+    Input_file->new("${EXTRACTED}DLineBreak.txt", v3.1.0,
+                    Property => 'Line_Break',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+    Input_file->new("${EXTRACTED}DBidiClass.txt", v3.1.1,
+                    Property => 'Bidi_Class',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+    Input_file->new("${EXTRACTED}DDecompositionType.txt", v3.1.0,
+                    Property => 'Decomposition_Type',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+    Input_file->new("${EXTRACTED}DBinaryProperties.txt", v3.1.0),
+    Input_file->new("${EXTRACTED}DNumValues.txt", v3.1.0,
+                    Property => 'Numeric_Value',
+                    Each_Line_Handler => \&filter_numeric_value_line,
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+    Input_file->new("${EXTRACTED}DJoinGroup.txt", v3.1.0,
+                    Property => 'Joining_Group',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+
+    Input_file->new("${EXTRACTED}DJoinType.txt", v3.1.0,
+                    Property => 'Joining_Type',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+    Input_file->new('Jamo.txt', v2.0.0,
+                    Property => 'Jamo_Short_Name',
+                    Each_Line_Handler => \&filter_jamo_line,
+                    ),
+    Input_file->new('UnicodeData.txt', v1.1.5,
+                    Pre_Handler => \&setup_UnicodeData,
+
+                    # We clean up this file for some early versions.
+                    Each_Line_Handler => [ (($v_version lt v2.0.0 )
+                                            ? \&filter_v1_ucd
+                                            : ($v_version eq v2.1.5)
+                                                ? \&filter_v2_1_5_ucd
+                                                : undef),
+
+                                            # And the main filter
+                                            \&filter_UnicodeData_line,
+                                         ],
+                    EOF_Handler => \&EOF_UnicodeData,
+                    ),
+    Input_file->new('ArabicShaping.txt', v2.0.0,
+                    Each_Line_Handler =>
+                        [ ($v_version lt 4.1.0)
+                                    ? \&filter_old_style_arabic_shaping
+                                    : undef,
+                        \&filter_arabic_shaping_line,
+                        ],
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+    Input_file->new('Blocks.txt', v2.0.0,
+                    Property => 'Block',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    Each_Line_Handler => \&filter_blocks_lines
+                    ),
+    Input_file->new('PropList.txt', v2.0.0,
+                    Each_Line_Handler => (($v_version lt v3.1.0)
+                                            ? \&filter_old_style_proplist
+                                            : undef),
+                    ),
+    Input_file->new('Unihan.txt', v2.0.0,
+                    Pre_Handler => \&setup_unihan,
+                    Optional => 1,
+                    Each_Line_Handler => \&filter_unihan_line,
+                        ),
+    Input_file->new('SpecialCasing.txt', v2.1.8,
+                    Each_Line_Handler => \&filter_special_casing_line,
+                    Pre_Handler => \&setup_special_casing,
+                    ),
+    Input_file->new(
+                    'LineBreak.txt', v3.0.0,
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    Property => 'Line_Break',
+                    # Early versions had problematic syntax
+                    Each_Line_Handler => (($v_version lt v3.1.0)
+                                        ? \&filter_early_ea_lb
+                                        : undef),
+                    ),
+    Input_file->new('EastAsianWidth.txt', v3.0.0,
+                    Property => 'East_Asian_Width',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    # Early versions had problematic syntax
+                    Each_Line_Handler => (($v_version lt v3.1.0)
+                                        ? \&filter_early_ea_lb
+                                        : undef),
+                    ),
+    Input_file->new('CompositionExclusions.txt', v3.0.0,
+                    Property => 'Composition_Exclusion',
+                    ),
+    Input_file->new('BidiMirroring.txt', v3.0.1,
+                    Property => 'Bidi_Mirroring_Glyph',
+                    ),
+    Input_file->new('CaseFolding.txt', v3.0.1,
+                    Pre_Handler => \&setup_case_folding,
+                    Each_Line_Handler =>
+                        [ ($v_version lt v3.1.0)
+                                 ? \&filter_old_style_case_folding
+                                 : undef,
+                           \&filter_case_folding_line
+                        ],
+                    Post_Handler => \&post_fold,
+                    ),
+    Input_file->new('DCoreProperties.txt', v3.1.0,
+                    # 5.2 changed this file
+                    Has_Missings_Defaults => (($v_version ge v5.2.0)
+                                            ? $NOT_IGNORED
+                                            : $NO_DEFAULTS),
+                    ),
+    Input_file->new('Scripts.txt', v3.1.0,
+                    Property => 'Script',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+    Input_file->new('DNormalizationProps.txt', v3.1.0,
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    Each_Line_Handler => (($v_version lt v4.0.1)
+                                      ? \&filter_old_style_normalization_lines
+                                      : undef),
+                    ),
+    Input_file->new('HangulSyllableType.txt', v4.0.0,
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    Property => 'Hangul_Syllable_Type'),
+    Input_file->new("$AUXILIARY/WordBreakProperty.txt", v4.1.0,
+                    Property => 'Word_Break',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+    Input_file->new("$AUXILIARY/GraphemeBreakProperty.txt", v4.1.0,
+                    Property => 'Grapheme_Cluster_Break',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+    Input_file->new("$AUXILIARY/SentenceBreakProperty.txt", v4.1.0,
+                    Property => 'Sentence_Break',
+                    Has_Missings_Defaults => $NOT_IGNORED,
+                    ),
+    Input_file->new('NamedSequences.txt', v4.1.0,
+                    Handler => \&process_NamedSequences
+                    ),
+    Input_file->new('NameAliases.txt', v5.0.0,
+                    Property => 'Name_Alias',
+                    ),
+    Input_file->new('UnihanIndicesDictionary.txt', v5.2.0,
+                    Optional => 1,
+                    Each_Line_Handler => \&filter_unihan_line,
+                    ),
+    Input_file->new('UnihanDataDictionaryLike.txt', v5.2.0,
+                    Optional => 1,
+                    Each_Line_Handler => \&filter_unihan_line,
+                    ),
+    Input_file->new('UnihanIRGSources.txt', v5.2.0,
+                    Optional => 1,
+                    Pre_Handler => \&setup_unihan,
+                    Each_Line_Handler => \&filter_unihan_line,
+                    ),
+    Input_file->new('UnihanNumericValues.txt', v5.2.0,
+                    Optional => 1,
+                    Each_Line_Handler => \&filter_unihan_line,
+                    ),
+    Input_file->new('UnihanOtherMappings.txt', v5.2.0,
+                    Optional => 1,
+                    Each_Line_Handler => \&filter_unihan_line,
+                    ),
+    Input_file->new('UnihanRadicalStrokeCounts.txt', v5.2.0,
+                    Optional => 1,
+                    Each_Line_Handler => \&filter_unihan_line,
+                    ),
+    Input_file->new('UnihanReadings.txt', v5.2.0,
+                    Optional => 1,
+                    Each_Line_Handler => \&filter_unihan_line,
+                    ),
+    Input_file->new('UnihanVariants.txt', v5.2.0,
+                    Optional => 1,
+                    Each_Line_Handler => \&filter_unihan_line,
+                    ),
+);
+
+# End of all the preliminaries.
+# Do it...
+
+if ($compare_versions) {
+    Carp::my_carp(<<END
+Warning.  \$compare_versions is set.  Output is not suitable for production
+END
+    );
+}
+
+# Put into %potential_files a list of all the files in the directory structure
+# that could be inputs to this program, excluding those that we should ignore.
+# Also don't consider test files.  Use absolute file names because it makes it
+# easier across machine types.
+my @ignored_files_full_names = map { File::Spec->rel2abs(
+                                     internal_file_to_platform($_))
+                                } keys %ignored_files;
+File::Find::find({
+    wanted=>sub {
+        return unless /\.txt$/i;
+        return if /Test\.txt$/i;
+        my $full = File::Spec->rel2abs($_);
+        $potential_files{$full} = 1
+                        if ! grep { $full eq $_ } @ignored_files_full_names;
+        return;
     }
+}, File::Spec->curdir());
 
-    my $Fold = Table->New();
-    my %Fold;
+my @mktables_list_output_files;
 
-    while (<IN>) {
-	# Skip status 'S', simple case folding
-	next unless /^([0-9A-Fa-f]+)\s*;\s*([CFI])\s*;\s*([0-9A-Fa-f]+(?: [0-9A-Fa-f]+)*)\s*;/;
+if ($write_unchanged_files) {
+    print "Not checking file list '$file_list'.\n" if $verbosity >= $VERBOSE;
+}
+else {
+    print "Reading file list '$file_list'\n" if $verbosity >= $VERBOSE;
+    my $file_handle;
+    if (! open $file_handle,"<",$file_list) {
+        Carp::my_carp("Failed to open '$file_list', turning on -globlist option instead: $!");
+        $glob_list = 1;
+    }
+    else {
+        my @input;
+
+        # Read and parse mktables.lst, placing the results from the first part
+        # into @input, and the second part into @mktables_list_output_files
+        for my $list ( \@input, \@mktables_list_output_files ) {
+            while (<$file_handle>) {
+                s/^ \s+ | \s+ $//xg;
+                next if /^ \s* (?: \# .* )? $/x;
+                last if /^ =+ $/x;
+                my ( $file ) = split /\t/;
+                push @$list, $file;
+            }
+            @$list = uniques(@$list);
+            next;
+        }
 
-	my ($code, $status, $fold) = (hex($1), $2, $3);
+        # Look through all the input files
+        foreach my $input (@input) {
+            next if $input eq 'version'; # Already have checked this.
 
-	if ($status eq 'C') { # Common: one-to-one folding
-	    # No append() since several codes may fold into one.
-	    $Fold->RawAppendRange($code, $code, $fold);
-	} else { # F: full, or I: dotted uppercase I -> dotless lowercase I
-	    $Fold{$code} = $fold;
-	}
+            # Ignore if doesn't exist.  The checking about whether we care or
+            # not is done via the Input_file object.
+            next if ! file_exists($input);
+
+            # The paths are stored with relative names, and with '/' as the
+            # delimiter; convert to absolute on this machine
+            my $full = File::Spec->rel2abs(internal_file_to_platform($input));
+            $potential_files{$full} = 1
+                        if ! grep { $full eq $_ } @ignored_files_full_names;
+        }
     }
-    close IN;
 
-    $Fold->Write("To/Fold.pl");
+    close $file_handle;
+}
 
-    #
-    # Prepend the special foldings to the common foldings.
-    #
-    my $CommonFold = do "To/Fold.pl" || die "$0: To/Fold.pl: $!\n";
-
-    my @OUT =
-	(
-	 $HEADER, $INTERNAL_ONLY, "\n",
-	 "#  The key: UTF-8 _bytes_, the value: UTF-8 (speed hack)\n",
-	 "%utf8::ToSpecFold =\n(\n",
-	);
-    for my $code (sort { $a <=> $b } keys %Fold) {
-        my $foldstr =
-          join "", map { sprintf "\\x{%s}", $_ } split ' ', $Fold{$code};
-        push @OUT, sprintf qq["%s" => "$foldstr",\n], join("", map { sprintf "\\x%02X", $_ } unpack("U0C*", pack("U", $code)));
-    }
-    push @OUT, (
-                ");\n\n",
-                "return <<'END';\n",
-                $CommonFold,
-                "END\n",
-               );
-
-    WriteIfChanged(["To","Fold.pl"], @OUT);
-}
-
-## Do it....
-
-Build_Aliases();
-UnicodeData_Txt();
-PropList_txt();
-
-Scripts_txt();
-Blocks_txt();
-
-WriteAllMappings();
-
-LineBreak_Txt();
-ArabicShaping_txt();
-EastAsianWidth_txt();
-HangulSyllableType_txt();
-Jamo_txt();
-SpecialCasing_txt();
-CaseFolding_txt();
-
-if ( $FileList and $MakeList ) {
-    
-    print "Updating '$FileList'\n"
-        if ($Verbose);
-        
-    open my $ofh,">",$FileList 
-        or die "Can't write to '$FileList':$!";
-    print $ofh <<"EOFHEADER";
-#
-# mktables.lst -- File list for mktables.
+if ($glob_list) {
+
+    # Here wants to process all .txt files in the directory structure.
+    # Convert them to full path names.  They are stored in the platform's
+    # relative style
+    my @known_files = map { File::Spec->rel2abs($_->file) }
+                                                        @input_file_objects;
+
+    my @unknown_input_files;
+    foreach my $file (keys %potential_files) {
+        next if grep { $file eq $_ } @known_files;
+
+        # Here, the file is unknown to us.  Get relative path name
+        $file = File::Spec->abs2rel($file);
+        push @unknown_input_files, $file;
+
+        # What will happen is we create a data structure for it, and add it to
+        # the list of input files to process.  First get the subdirectories
+        # into an array
+        my (undef, $directories, undef) = File::Spec->splitpath($file);
+        $directories =~ s;/$;;;     # Can have extraneous trailing '/'
+        my @directories = File::Spec->splitdir($directories);
+
+        # If the file isn't extracted (meaning none of the directories is the
+        # extracted one), just add it to the end of the list of inputs.
+        if (! grep { $EXTRACTED_DIR eq $_ } @directories) {
+            push @input_file_objects, Input_file->new($file);
+        }
+        else {
+
+            # Here, the file is extracted.  It needs to go ahead of most other
+            # processing.  Search for the first input file that isn't a
+            # special required property (that is, find one whose first_release
+            # is non-0), and isn't extracted.  Also, the Age property file is
+            # processed before the extracted ones, just in case
+            # $compare_versions is set.
+            for (my $i = 0; $i < @input_file_objects; $i++) {
+                if ($input_file_objects[$i]->first_released ne v0
+                    && $input_file_objects[$i]->file ne 'DAge.txt'
+                    && $input_file_objects[$i]->file !~ /$EXTRACTED_DIR/)
+                {
+                    splice @input_file_objects, $i, 0, Input_file->new($file);
+                    last;
+                }
+            }
+
+        }
+    }
+    if (@unknown_input_files) {
+        print STDERR simple_fold(join_line(<<END
+
+The following files are unknown as to how to handle.  Assuming they are
+typical property files.  You'll know by later error messages if it worked or
+not:
+END
+        ) . join(", ", @unknown_input_files) . "\n\n");
+    }
+} # End of looking through directory structure for more .txt files.
+
+if ( $make_list ) {
+    foreach my $file (@mktables_list_output_files) {
+        unlink internal_file_to_platform($file);
+    }
+}
+
+# Create the list of input files from the objects we have defined, plus
+# version
+my @input_files = 'version';
+foreach my $object (@input_file_objects) {
+    my $file = $object->file;
+    next if ! defined $file;    # Not all objects have files
+    next if $object->optional && ! -e $file;
+    push @input_files,  $file;
+}
+
+if ( $verbosity >= $VERBOSE ) {
+    print "Expecting ".scalar( @input_files )." input files. ",
+         "Checking ".scalar( @mktables_list_output_files )." output files.\n";
+}
+
+# We set $youngest to be the most recently changed input file, including this
+# program itself (done much earlier in this file)
+foreach my $in (@input_files) {
+    my $age = -M $in;
+    next unless defined $age;        # Keep going even if missing a file
+    $youngest = $age if $age < $youngest;
+
+    # See that the input files have distinct names, to warn someone if they
+    # are adding a new one
+    if ($make_list) {
+        my ($volume, $directories, $file ) = File::Spec->splitpath($in);
+        $directories =~ s;/$;;;     # Can have extraneous trailing '/'
+        my @directories = File::Spec->splitdir($directories);
+        my $base = $file =~ s/\.txt$//;
+        construct_filename($file, 'mutable', \@directories);
+    }
+}
+
+my $ok = ! $write_unchanged_files
+        && scalar @mktables_list_output_files;        # If none known, rebuild
+
+# Now we check to see if any output files are older than youngest, if
+# they are, we need to continue on, otherwise we can presumably bail.
+if ($ok) {
+    foreach my $out (@mktables_list_output_files) {
+        if ( ! file_exists($out)) {
+            print "'$out' is missing.\n" if $verbosity >= $VERBOSE;
+            $ok = 0;
+            last;
+         }
+        #local $to_trace = 1 if main::DEBUG;
+        trace $youngest, -M $out if main::DEBUG && $to_trace;
+        if ( -M $out > $youngest ) {
+            #trace "$out: age: ", -M $out, ", youngest: $youngest\n" if main::DEBUG && $to_trace;
+            print "'$out' is too old.\n" if $verbosity >= $VERBOSE;
+            $ok = 0;
+            last;
+        }
+    }
+}
+if ($ok) {
+    print "Files seem to be ok, not bothering to rebuild.\n";
+    exit(0);
+}
+print "Must rebuild tables.\n" if $verbosity >= $VERBOSE;
+
+# Ready to do the major processing.  First create the perl pseudo-property.
+$perl = Property->new('perl', Type => $NON_STRING, Perl_Extension => 1);
+
+# Process each input file
+foreach my $file (@input_file_objects) {
+    $file->run;
+}
+
+# Finish the table generation.
+
+print "Finishing processing Unicode properties\n" if $verbosity >= $PROGRESS;
+finish_Unicode();
+
+print "Compiling Perl properties\n" if $verbosity >= $PROGRESS;
+compile_perl();
+
+print "Creating Perl synonyms\n" if $verbosity >= $PROGRESS;
+add_perl_synonyms();
+
+print "Writing tables\n" if $verbosity >= $PROGRESS;
+write_all_tables();
+
+# Write mktables.lst
+if ( $file_list and $make_list ) {
+
+    print "Updating '$file_list'\n" if $verbosity >= $PROGRESS;
+    foreach my $file (@input_files, @files_actually_output) {
+        my (undef, $directories, $file) = File::Spec->splitpath($file);
+        my @directories = File::Spec->splitdir($directories);
+        $file = join '/', @directories, $file;
+    }
+
+    my $ofh;
+    if (! open $ofh,">",$file_list) {
+        Carp::my_carp("Can't write to '$file_list'.  Skipping: $!");
+        return
+    }
+    else {
+        print $ofh <<"END";
+#
+# $file_list -- File list for $0.
 #
 #   Autogenerated on @{[scalar localtime]}
 #
 # - First section is input files
-#   (mktables itself is automatically included)
+#   ($0 itself is not listed but is automatically considered an input)
 # - Section seperator is /^=+\$/
 # - Second section is a list of output files.
 # - Lines matching /^\\s*#/ are treated as comments
@@ -2235,91 +13700,172 @@ if ( $FileList and $MakeList ) {
 
 # Input files:
 
-EOFHEADER
-    my @input=("version",glob('*.txt'));
-    print $ofh "$_\n" for 
-        sort(@input),
-        "\n=================================\n",
-        "# Output files:\n",
-        # special files
-        "Properties";
-        
-    
-    require File::Find;
-    my @output_files;
-    File::Find::find({
-        no_chdir=>1,
-        wanted=>sub {
-          if (/\.pl$/) {
-            s!^\./!!;
-            push @output_files, "$_\n";
-          }
-        },
-    },"."); 
-    
-    print $ofh sort @output_files;
-    print $ofh "\n# ",scalar(@input)," input files\n",
-               "# ",scalar(@output_files)+1," output files\n\n",
-               "# End list\n";  
-    close $ofh 
-        or warn "Failed to close $ofh: $!";
-    
-    print "Filelist has ",scalar(@input)," input files and ",
-          scalar(@output_files)+1," output files\n"
-        if $Verbose;
-}
-print "All done\n" if $Verbose;
+END
+        print $ofh "$_\n" for sort(@input_files);
+        print $ofh "\n=================================\n# Output files:\n\n";
+        print $ofh "$_\n" for sort @files_actually_output;
+        print $ofh "\n# ",scalar(@input_files)," input files\n",
+                "# ",scalar(@files_actually_output)+1," output files\n\n",
+                "# End list\n";
+        close $ofh
+            or Carp::my_carp("Failed to close $ofh: $!");
+
+        print "Filelist has ",scalar(@input_files)," input files and ",
+            scalar(@files_actually_output)+1," output files\n"
+            if $verbosity >= $VERBOSE;
+    }
+}
+
+# Output these warnings unless -q explicitly specified.
+if ($verbosity >= $NORMAL_VERBOSITY) {
+    if (@unhandled_properties) {
+        print "\nProperties and tables that unexpectedly have no code points\n";
+        foreach my $property (sort @unhandled_properties) {
+            print $property, "\n";
+        }
+    }
+
+    if (%potential_files) {
+        print "\nInput files that are not considered:\n";
+        foreach my $file (sort keys %potential_files) {
+            print File::Spec->abs2rel($file), "\n";
+        }
+    }
+    print "\nAll done\n" if $verbosity >= $VERBOSE;
+}
 exit(0);
 
-## TRAILING CODE IS USED BY MakePropTestScript()
+# TRAILING CODE IS USED BY make_property_test_script()
 __DATA__
+
 use strict;
 use warnings;
 
+# Test the \p{} regular expression constructs.  This file is constructed by
+# mktables from the tables it generates, so if mktables is buggy, this won't
+# necessarily catch those bugs.  Tests are generated for all feasible
+# properties; a few aren't currently feasible; see is_code_point_usable()
+# in mktables for details.
+
+# Standard test packages are not used because this manipulates SIG_WARN.  It
+# exits 0 if every non-skipped test succeeded; -1 if any failed.
+
 my $Tests = 0;
 my $Fails = 0;
+my $Skips = 0;
+
+my $non_ASCII = (ord('A') == 65);
+
+# The first 127 ASCII characters in ordinal order, with the ones that don't
+# have Perl names (as of 5.8) replaced by dots.  The 127th is used as the
+# string delimiter
+my $ascii_to_ebcdic = "\0......\a\b\t\n.\f\r.................. !\"#\$\%&'()*+,-./0123456789:;<=>?\@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~";
+#for my $i (0..126) {
+#    print $i, ": ", substr($ascii_to_ebcdic, $i, 1), "\n";
+#}
+
+sub Expect($$$$) {
+    my $expected = shift;
+    my $ord = shift;
+    my $regex  = shift;
+    my $warning_type = shift;   # Type of warning message, like 'deprecated'
+                                # or empty if none
+    my $line   = (caller)[2];
+
+    # Convert the code point to hex form
+    my $string = sprintf "\"\\x{%04X}\"", $ord; 
+
+    # Convert the non-ASCII code points expressible as characters in Perl 5.8
+    # to their ASCII equivalents, and skip the others.
+    if ($non_ASCII && $ord < 255) {
+
+        # Dots are used as place holders in the conversion string for the
+        # non-convertible ones, so check for it first.
+        if ($ord == 0x2E) {
+            $ord = ord('.');
+        }
+        elsif ($ord < 0x7F
+                  # Any dots returned are non-convertible.
+                 && ((my $char = substr($ascii_to_ebcdic, $ord, 1)) ne '.'))
+        {
+            #print STDERR "$ord, $char, \n";
+            $ord = ord($char);
+        }
+        else {
+            $Tests++;
+            $Skips++;
+            print "ok $Tests - $string =~ $regex # Skipped: non-ASCII\n";
+            return;
+        }
+    }
 
-sub Expect($$$)
-{
-    my $Expect = shift;
-    my $String = shift;
-    my $Regex  = shift;
-    my $Line   = (caller)[2];
+    # The first time through, use all warnings.
+    my @tests = "";
 
-    $Tests++;
-    my $RegObj;
-    my $result = eval {
-        $RegObj = qr/$Regex/;
-        $String =~ $RegObj ? 1 : 0
-    };
-    
-    if (not defined $result) {
-        print "couldn't compile /$Regex/ on $0 line $Line: $@\n";
-        $Fails++;
-    } elsif ($result ^ $Expect) {
-        print "bad result (expected $Expect) on $0 line $Line: $@\n";
-        $Fails++;
+    # If the input should generate a warning, add another time through with
+    # them turned off
+    push @tests, "no warnings '$warning_type';" if $warning_type;
+
+    foreach my $no_warnings (@tests) {
+
+        # Store any warning messages instead of outputting them
+        local $SIG{__WARN__} = $SIG{__WARN__};
+        my $warning_message;
+        $SIG{__WARN__} = sub { $warning_message = $_[0] };
+
+        $Tests++;
+
+        # A string eval is needed because of the 'no warnings'.
+        # Assumes no parens in the regular expression
+        my $result = eval "$no_warnings
+                            my \$RegObj = qr($regex);
+                            $string =~ \$RegObj ? 1 : 0";
+        if (not defined $result) {
+            print "not ok $Tests - couldn't compile /$regex/; line $line: $@\n";
+            $Fails++;
+        }
+        elsif ($result ^ $expected) {
+            print "not ok $Tests - expected $expected but got $result for $string =~ qr/$regex/; line $line\n";
+            $Fails++;
+        }
+        elsif ($warning_message) {
+            if (! $warning_type || ($warning_type && $no_warnings)) {
+                print "not ok $Tests - for qr/$regex/ did not expect warning message '$warning_message'; line $line\n";
+                $Fails++;
+            }
+            else {
+                print "ok $Tests - expected and got a warning message for qr/$regex/; line $line\n";
+            }
+        }
+        elsif ($warning_type && ! $no_warnings) {
+            print "not ok $Tests - for qr/$regex/ expected a $warning_type warning message, but got none; line $line\n";
+            $Fails++;
+        }
+        else {
+            print "ok $Tests - got $result for $string =~ qr/$regex/; line $line\n";
+        }
     }
+    return;
 }
 
-sub Error($)
-{
-    my $Regex  = shift;
+sub Error($) {
+    my $regex  = shift;
     $Tests++;
-    if (eval { 'x' =~ qr/$Regex/; 1 }) {
+    if (eval { 'x' =~ qr/$regex/; 1 }) {
         $Fails++;
-        my $Line = (caller)[2];
-        print "expected error for /$Regex/ on $0 line $Line: $@\n";
+        my $line = (caller)[2];
+        print "not ok $Tests - re compiled ok, but expected error for qr/$regex/; line $line: $@\n";
     }
+    else {
+        my $line = (caller)[2];
+        print "ok $Tests - got and expected error for qr/$regex/; line $line\n";
+    }
+    return;
 }
 
-sub Finished()
-{
-   if ($Fails == 0) {
-      print "All $Tests tests passed.\n";
-      exit(0);
-   } else {
-      print "$Tests tests, $Fails failed!\n";
-      exit(-1);
-   }
+sub Finished() {
+    print "1..$Tests.\n";
+    exit($Fails ? -1 : 0);
 }
+
+Error('\p{Script=InGreek}');    # Bug #69018
diff --git a/lib/unicore/mktables.lst b/lib/unicore/mktables.lst
index f4c55bb64e..db6b0b7a4f 100644
--- a/lib/unicore/mktables.lst
+++ b/lib/unicore/mktables.lst
@@ -1,10 +1,10 @@
 #
-# mktables.lst -- File list for mktables.
+# mktables.lst -- File list for lib/unicore/mktables.
 #
-#   Autogenerated on Mon Jan 26 17:57:26 2009
+#   Autogenerated on Sat Nov 21 22:17:15 2009
 #
 # - First section is input files
-#   (mktables itself is automatically included)
+#   (lib/unicore/mktables itself is not listed but is automatically considered an input)
 # - Section seperator is /^=+$/
 # - Second section is a list of output files.
 # - Lines matching /^\s*#/ are treated as comments
@@ -18,546 +18,812 @@ BidiMirroring.txt
 Blocks.txt
 CaseFolding.txt
 CompositionExclusions.txt
+DAge.txt
+DCoreProperties.txt
+DNormalizationProps.txt
 EastAsianWidth.txt
 HangulSyllableType.txt
-Index.txt
 Jamo.txt
 LineBreak.txt
 NameAliases.txt
 NamedSequences.txt
-NamedSqProv.txt
-NamesList.txt
-NormalizationCorrections.txt
 PropList.txt
 PropValueAliases.txt
 PropertyAliases.txt
-ReadMe.txt
 Scripts.txt
 SpecialCasing.txt
-StandardizedVariants.txt
 UnicodeData.txt
+auxiliary/GraphemeBreakProperty.txt
+auxiliary/SentenceBreakProperty.txt
+auxiliary/WordBreakProperty.txt
+extracted/DBidiClass.txt
+extracted/DBinaryProperties.txt
+extracted/DCombiningClass.txt
+extracted/DDecompositionType.txt
+extracted/DEastAsianWidth.txt
+extracted/DGeneralCategory.txt
+extracted/DJoinGroup.txt
+extracted/DJoinType.txt
+extracted/DLineBreak.txt
+extracted/DNumType.txt
+extracted/DNumValues.txt
 version
 
 =================================
-
 # Output files:
 
-Properties
-Canonical.pl
+../../pod/perluniprops.pod
+../../t/re/uniprops.t
 CombiningClass.pl
 Decomposition.pl
-Exact.pl
+Heavy.pl
 Name.pl
-PVA.pl
+To/Bmg.pl
 To/Digit.pl
 To/Fold.pl
 To/Lower.pl
 To/Title.pl
 To/Upper.pl
-lib/bc/AL.pl
-lib/bc/AN.pl
-lib/bc/B.pl
-lib/bc/BN.pl
-lib/bc/CS.pl
-lib/bc/EN.pl
-lib/bc/ES.pl
-lib/bc/ET.pl
-lib/bc/L.pl
-lib/bc/LRE.pl
-lib/bc/LRO.pl
-lib/bc/NSM.pl
-lib/bc/ON.pl
-lib/bc/PDF.pl
-lib/bc/R.pl
-lib/bc/RLE.pl
-lib/bc/RLO.pl
-lib/bc/S.pl
-lib/bc/WS.pl
-lib/ccc/A.pl
-lib/ccc/AL.pl
-lib/ccc/AR.pl
-lib/ccc/ATAR.pl
-lib/ccc/ATB.pl
-lib/ccc/ATBL.pl
-lib/ccc/B.pl
-lib/ccc/BL.pl
-lib/ccc/BR.pl
-lib/ccc/DA.pl
-lib/ccc/DB.pl
-lib/ccc/IS.pl
-lib/ccc/KV.pl
-lib/ccc/L.pl
-lib/ccc/NK.pl
-lib/ccc/NR.pl
-lib/ccc/OV.pl
-lib/ccc/R.pl
-lib/ccc/VR.pl
-lib/dt/Can.pl
-lib/dt/Com.pl
-lib/dt/Enc.pl
-lib/dt/Fin.pl
-lib/dt/Font.pl
-lib/dt/Fra.pl
-lib/dt/Init.pl
-lib/dt/Iso.pl
-lib/dt/Med.pl
-lib/dt/Nar.pl
-lib/dt/Nb.pl
-lib/dt/Sml.pl
-lib/dt/Sqr.pl
-lib/dt/Sub.pl
-lib/dt/Sup.pl
-lib/dt/Vert.pl
-lib/dt/Wide.pl
-lib/ea/A.pl
-lib/ea/F.pl
-lib/ea/H.pl
-lib/ea/N.pl
-lib/ea/Na.pl
-lib/ea/W.pl
-lib/gc_sc/AHex.pl
-lib/gc_sc/ASCII.pl
-lib/gc_sc/Alnum.pl
-lib/gc_sc/Alpha.pl
-lib/gc_sc/Alphabet.pl
-lib/gc_sc/Any.pl
-lib/gc_sc/Arab.pl
-lib/gc_sc/Armn.pl
-lib/gc_sc/AsciiHex.pl
-lib/gc_sc/Assigned.pl
-lib/gc_sc/Bali.pl
-lib/gc_sc/Beng.pl
-lib/gc_sc/BidiC.pl
-lib/gc_sc/BidiCont.pl
-lib/gc_sc/Blank.pl
-lib/gc_sc/Bopo.pl
-lib/gc_sc/Brai.pl
-lib/gc_sc/Bugi.pl
-lib/gc_sc/Buhd.pl
-lib/gc_sc/C.pl
-lib/gc_sc/Canadian.pl
-lib/gc_sc/Cari.pl
-lib/gc_sc/Cc.pl
-lib/gc_sc/Cf.pl
-lib/gc_sc/Cham.pl
-lib/gc_sc/Cher.pl
-lib/gc_sc/Cn.pl
-lib/gc_sc/Cntrl.pl
-lib/gc_sc/Co.pl
-lib/gc_sc/Copt.pl
-lib/gc_sc/Cprt.pl
-lib/gc_sc/Cs.pl
-lib/gc_sc/Cyrl.pl
-lib/gc_sc/Dash.pl
-lib/gc_sc/Dash2.pl
-lib/gc_sc/DefaultI.pl
-lib/gc_sc/Dep.pl
-lib/gc_sc/Deprecat.pl
-lib/gc_sc/Deva.pl
-lib/gc_sc/Dia.pl
-lib/gc_sc/Diacriti.pl
-lib/gc_sc/Digit.pl
-lib/gc_sc/Dsrt.pl
-lib/gc_sc/Ethi.pl
-lib/gc_sc/Ext.pl
-lib/gc_sc/Extender.pl
-lib/gc_sc/Geor.pl
-lib/gc_sc/Glag.pl
-lib/gc_sc/Goth.pl
-lib/gc_sc/Graph.pl
-lib/gc_sc/Grek.pl
-lib/gc_sc/Gujr.pl
-lib/gc_sc/Guru.pl
-lib/gc_sc/Hang.pl
-lib/gc_sc/Hani.pl
-lib/gc_sc/Hano.pl
-lib/gc_sc/Hebr.pl
-lib/gc_sc/Hex.pl
-lib/gc_sc/HexDigit.pl
-lib/gc_sc/Hira.pl
-lib/gc_sc/HorizSpa.pl
-lib/gc_sc/Hyphen.pl
-lib/gc_sc/Hyphen2.pl
-lib/gc_sc/IDSB.pl
-lib/gc_sc/IDST.pl
-lib/gc_sc/IdContin.pl
-lib/gc_sc/IdStart.pl
-lib/gc_sc/Ideo.pl
-lib/gc_sc/Ideograp.pl
-lib/gc_sc/IdsBinar.pl
-lib/gc_sc/IdsTrina.pl
-lib/gc_sc/InAegean.pl
-lib/gc_sc/InAlphab.pl
-lib/gc_sc/InAncie2.pl
-lib/gc_sc/InAncie3.pl
-lib/gc_sc/InAncien.pl
-lib/gc_sc/InArabi2.pl
-lib/gc_sc/InArabi3.pl
-lib/gc_sc/InArabi4.pl
-lib/gc_sc/InArabic.pl
-lib/gc_sc/InArmeni.pl
-lib/gc_sc/InArrows.pl
-lib/gc_sc/InBaline.pl
-lib/gc_sc/InBasicL.pl
-lib/gc_sc/InBengal.pl
-lib/gc_sc/InBlockE.pl
-lib/gc_sc/InBopom2.pl
-lib/gc_sc/InBopomo.pl
-lib/gc_sc/InBoxDra.pl
-lib/gc_sc/InBraill.pl
-lib/gc_sc/InBugine.pl
-lib/gc_sc/InBuhid.pl
-lib/gc_sc/InByzant.pl
-lib/gc_sc/InCarian.pl
-lib/gc_sc/InCham.pl
-lib/gc_sc/InCherok.pl
-lib/gc_sc/InCjkCo2.pl
-lib/gc_sc/InCjkCo3.pl
-lib/gc_sc/InCjkCo4.pl
-lib/gc_sc/InCjkCom.pl
-lib/gc_sc/InCjkRad.pl
-lib/gc_sc/InCjkStr.pl
-lib/gc_sc/InCjkSym.pl
-lib/gc_sc/InCjkUn2.pl
-lib/gc_sc/InCjkUn3.pl
-lib/gc_sc/InCjkUni.pl
-lib/gc_sc/InCombi2.pl
-lib/gc_sc/InCombi3.pl
-lib/gc_sc/InCombi4.pl
-lib/gc_sc/InCombin.pl
-lib/gc_sc/InContro.pl
-lib/gc_sc/InCoptic.pl
-lib/gc_sc/InCounti.pl
-lib/gc_sc/InCunei2.pl
-lib/gc_sc/InCuneif.pl
-lib/gc_sc/InCurren.pl
-lib/gc_sc/InCyprio.pl
-lib/gc_sc/InCyril2.pl
-lib/gc_sc/InCyril3.pl
-lib/gc_sc/InCyril4.pl
-lib/gc_sc/InCyrill.pl
-lib/gc_sc/InDesere.pl
-lib/gc_sc/InDevana.pl
-lib/gc_sc/InDingba.pl
-lib/gc_sc/InDomino.pl
-lib/gc_sc/InEnclo2.pl
-lib/gc_sc/InEnclos.pl
-lib/gc_sc/InEthio2.pl
-lib/gc_sc/InEthio3.pl
-lib/gc_sc/InEthiop.pl
-lib/gc_sc/InGenera.pl
-lib/gc_sc/InGeomet.pl
-lib/gc_sc/InGeorg2.pl
-lib/gc_sc/InGeorgi.pl
-lib/gc_sc/InGlagol.pl
-lib/gc_sc/InGothic.pl
-lib/gc_sc/InGreekA.pl
-lib/gc_sc/InGreekE.pl
-lib/gc_sc/InGujara.pl
-lib/gc_sc/InGurmuk.pl
-lib/gc_sc/InHalfwi.pl
-lib/gc_sc/InHangu2.pl
-lib/gc_sc/InHangu3.pl
-lib/gc_sc/InHangul.pl
-lib/gc_sc/InHanuno.pl
-lib/gc_sc/InHebrew.pl
-lib/gc_sc/InHighPr.pl
-lib/gc_sc/InHighSu.pl
-lib/gc_sc/InHiraga.pl
-lib/gc_sc/InIdeogr.pl
-lib/gc_sc/InIpaExt.pl
-lib/gc_sc/InKanbun.pl
-lib/gc_sc/InKangxi.pl
-lib/gc_sc/InKannad.pl
-lib/gc_sc/InKatak2.pl
-lib/gc_sc/InKataka.pl
-lib/gc_sc/InKayahL.pl
-lib/gc_sc/InKharos.pl
-lib/gc_sc/InKhmer.pl
-lib/gc_sc/InKhmerS.pl
-lib/gc_sc/InLao.pl
-lib/gc_sc/InLatin1.pl
-lib/gc_sc/InLatin2.pl
-lib/gc_sc/InLatin3.pl
-lib/gc_sc/InLatin4.pl
-lib/gc_sc/InLatin5.pl
-lib/gc_sc/InLatinE.pl
-lib/gc_sc/InLepcha.pl
-lib/gc_sc/InLetter.pl
-lib/gc_sc/InLimbu.pl
-lib/gc_sc/InLinea2.pl
-lib/gc_sc/InLinear.pl
-lib/gc_sc/InLowSur.pl
-lib/gc_sc/InLycian.pl
-lib/gc_sc/InLydian.pl
-lib/gc_sc/InMahjon.pl
-lib/gc_sc/InMalaya.pl
-lib/gc_sc/InMathe2.pl
-lib/gc_sc/InMathem.pl
-lib/gc_sc/InMisce2.pl
-lib/gc_sc/InMisce3.pl
-lib/gc_sc/InMisce4.pl
-lib/gc_sc/InMisce5.pl
-lib/gc_sc/InMiscel.pl
-lib/gc_sc/InModifi.pl
-lib/gc_sc/InMongol.pl
-lib/gc_sc/InMusica.pl
-lib/gc_sc/InMyanma.pl
-lib/gc_sc/InNewTai.pl
-lib/gc_sc/InNko.pl
-lib/gc_sc/InNumber.pl
-lib/gc_sc/InOgham.pl
-lib/gc_sc/InOlChik.pl
-lib/gc_sc/InOldIta.pl
-lib/gc_sc/InOldPer.pl
-lib/gc_sc/InOptica.pl
-lib/gc_sc/InOriya.pl
-lib/gc_sc/InOsmany.pl
-lib/gc_sc/InPhagsP.pl
-lib/gc_sc/InPhaist.pl
-lib/gc_sc/InPhoeni.pl
-lib/gc_sc/InPhone2.pl
-lib/gc_sc/InPhonet.pl
-lib/gc_sc/InPrivat.pl
-lib/gc_sc/InRejang.pl
-lib/gc_sc/InRunic.pl
-lib/gc_sc/InSauras.pl
-lib/gc_sc/InShavia.pl
-lib/gc_sc/InSinhal.pl
-lib/gc_sc/InSmallF.pl
-lib/gc_sc/InSpacin.pl
-lib/gc_sc/InSpecia.pl
-lib/gc_sc/InSundan.pl
-lib/gc_sc/InSupers.pl
-lib/gc_sc/InSuppl2.pl
-lib/gc_sc/InSuppl3.pl
-lib/gc_sc/InSuppl4.pl
-lib/gc_sc/InSuppl5.pl
-lib/gc_sc/InSuppl6.pl
-lib/gc_sc/InSupple.pl
-lib/gc_sc/InSyloti.pl
-lib/gc_sc/InSyriac.pl
-lib/gc_sc/InTagalo.pl
-lib/gc_sc/InTagban.pl
-lib/gc_sc/InTags.pl
-lib/gc_sc/InTaiLe.pl
-lib/gc_sc/InTaiXua.pl
-lib/gc_sc/InTamil.pl
-lib/gc_sc/InTelugu.pl
-lib/gc_sc/InThaana.pl
-lib/gc_sc/InThai.pl
-lib/gc_sc/InTibeta.pl
-lib/gc_sc/InTifina.pl
-lib/gc_sc/InUgarit.pl
-lib/gc_sc/InUnifie.pl
-lib/gc_sc/InVai.pl
-lib/gc_sc/InVaria2.pl
-lib/gc_sc/InVariat.pl
-lib/gc_sc/InVertic.pl
-lib/gc_sc/InYiRadi.pl
-lib/gc_sc/InYiSyll.pl
-lib/gc_sc/InYijing.pl
-lib/gc_sc/JoinC.pl
-lib/gc_sc/JoinCont.pl
-lib/gc_sc/Kana.pl
-lib/gc_sc/KayahLi.pl
-lib/gc_sc/Khar.pl
-lib/gc_sc/Khmr.pl
-lib/gc_sc/Knda.pl
-lib/gc_sc/L.pl
-lib/gc_sc/LC.pl
-lib/gc_sc/LOE.pl
-lib/gc_sc/Laoo.pl
-lib/gc_sc/Latn.pl
-lib/gc_sc/Lepc.pl
-lib/gc_sc/Limb.pl
-lib/gc_sc/LinearB.pl
-lib/gc_sc/Ll.pl
-lib/gc_sc/Lm.pl
-lib/gc_sc/Lo.pl
-lib/gc_sc/LogicalO.pl
-lib/gc_sc/Lower.pl
-lib/gc_sc/Lowercas.pl
-lib/gc_sc/Lt.pl
-lib/gc_sc/Lu.pl
-lib/gc_sc/Lyci.pl
-lib/gc_sc/Lydi.pl
-lib/gc_sc/M.pl
-lib/gc_sc/Math.pl
-lib/gc_sc/Mc.pl
-lib/gc_sc/Me.pl
-lib/gc_sc/Mlym.pl
-lib/gc_sc/Mn.pl
-lib/gc_sc/Mong.pl
-lib/gc_sc/Mymr.pl
-lib/gc_sc/N.pl
-lib/gc_sc/NChar.pl
-lib/gc_sc/Nd.pl
-lib/gc_sc/NewTaiLu.pl
-lib/gc_sc/Nkoo.pl
-lib/gc_sc/Nl.pl
-lib/gc_sc/No.pl
-lib/gc_sc/Nonchara.pl
-lib/gc_sc/OAlpha.pl
-lib/gc_sc/ODI.pl
-lib/gc_sc/OGrExt.pl
-lib/gc_sc/OIDC.pl
-lib/gc_sc/OIDS.pl
-lib/gc_sc/OLower.pl
-lib/gc_sc/OMath.pl
-lib/gc_sc/OUpper.pl
-lib/gc_sc/Ogam.pl
-lib/gc_sc/OlChiki.pl
-lib/gc_sc/OldItali.pl
-lib/gc_sc/OldPersi.pl
-lib/gc_sc/Orya.pl
-lib/gc_sc/Osma.pl
-lib/gc_sc/OtherAlp.pl
-lib/gc_sc/OtherDef.pl
-lib/gc_sc/OtherGra.pl
-lib/gc_sc/OtherIdC.pl
-lib/gc_sc/OtherIdS.pl
-lib/gc_sc/OtherLow.pl
-lib/gc_sc/OtherMat.pl
-lib/gc_sc/OtherUpp.pl
-lib/gc_sc/P.pl
-lib/gc_sc/PatSyn.pl
-lib/gc_sc/PatWS.pl
-lib/gc_sc/PatternS.pl
-lib/gc_sc/PatternW.pl
-lib/gc_sc/Pc.pl
-lib/gc_sc/Pd.pl
-lib/gc_sc/Pe.pl
-lib/gc_sc/PerlSpac.pl
-lib/gc_sc/PerlWord.pl
-lib/gc_sc/Pf.pl
-lib/gc_sc/PhagsPa.pl
-lib/gc_sc/Phnx.pl
-lib/gc_sc/Pi.pl
-lib/gc_sc/Po.pl
-lib/gc_sc/PosixAln.pl
-lib/gc_sc/PosixAlp.pl
-lib/gc_sc/PosixBla.pl
-lib/gc_sc/PosixCnt.pl
-lib/gc_sc/PosixDig.pl
-lib/gc_sc/PosixGra.pl
-lib/gc_sc/PosixLow.pl
-lib/gc_sc/PosixPri.pl
-lib/gc_sc/PosixPun.pl
-lib/gc_sc/PosixSpa.pl
-lib/gc_sc/PosixUpp.pl
-lib/gc_sc/Print.pl
-lib/gc_sc/Ps.pl
-lib/gc_sc/Punct.pl
-lib/gc_sc/QMark.pl
-lib/gc_sc/Qaai.pl
-lib/gc_sc/Quotatio.pl
-lib/gc_sc/Radical.pl
-lib/gc_sc/Radical2.pl
-lib/gc_sc/Rjng.pl
-lib/gc_sc/Runr.pl
-lib/gc_sc/S.pl
-lib/gc_sc/SD.pl
-lib/gc_sc/STerm.pl
-lib/gc_sc/Saur.pl
-lib/gc_sc/Sc.pl
-lib/gc_sc/Shaw.pl
-lib/gc_sc/Sinh.pl
-lib/gc_sc/Sk.pl
-lib/gc_sc/Sm.pl
-lib/gc_sc/So.pl
-lib/gc_sc/SoftDott.pl
-lib/gc_sc/Space.pl
-lib/gc_sc/SpacePer.pl
-lib/gc_sc/Sterm2.pl
-lib/gc_sc/Sund.pl
-lib/gc_sc/SylotiNa.pl
-lib/gc_sc/Syrc.pl
-lib/gc_sc/Tagb.pl
-lib/gc_sc/TaiLe.pl
-lib/gc_sc/Taml.pl
-lib/gc_sc/Telu.pl
-lib/gc_sc/Term.pl
-lib/gc_sc/Terminal.pl
-lib/gc_sc/Tfng.pl
-lib/gc_sc/Tglg.pl
-lib/gc_sc/Thaa.pl
-lib/gc_sc/Thai.pl
-lib/gc_sc/Tibt.pl
-lib/gc_sc/Title.pl
-lib/gc_sc/UIdeo.pl
-lib/gc_sc/Ugar.pl
-lib/gc_sc/UnifiedI.pl
-lib/gc_sc/Upper.pl
-lib/gc_sc/Uppercas.pl
-lib/gc_sc/VS.pl
-lib/gc_sc/Vaii.pl
-lib/gc_sc/Variatio.pl
-lib/gc_sc/VertSpac.pl
-lib/gc_sc/WSpace.pl
-lib/gc_sc/WhiteSpa.pl
-lib/gc_sc/Word.pl
-lib/gc_sc/XDigit.pl
-lib/gc_sc/Xsux.pl
-lib/gc_sc/Yiii.pl
-lib/gc_sc/Z.pl
-lib/gc_sc/Zl.pl
-lib/gc_sc/Zp.pl
-lib/gc_sc/Zs.pl
-lib/gc_sc/Zyyy.pl
-lib/gc_sc/_CanonDC.pl
-lib/gc_sc/_CaseIgn.pl
-lib/gc_sc/_CombAbo.pl
-lib/hst/L.pl
-lib/hst/LV.pl
-lib/hst/LVT.pl
-lib/hst/T.pl
-lib/hst/V.pl
-lib/jt/C.pl
-lib/jt/D.pl
-lib/jt/R.pl
-lib/jt/U.pl
-lib/lb/AI.pl
-lib/lb/AL.pl
-lib/lb/B2.pl
-lib/lb/BA.pl
-lib/lb/BB.pl
-lib/lb/BK.pl
-lib/lb/CB.pl
-lib/lb/CL.pl
-lib/lb/CM.pl
-lib/lb/CR.pl
-lib/lb/EX.pl
-lib/lb/GL.pl
-lib/lb/H2.pl
-lib/lb/H3.pl
-lib/lb/HY.pl
-lib/lb/ID.pl
-lib/lb/IN.pl
-lib/lb/IS.pl
-lib/lb/JL.pl
-lib/lb/JT.pl
-lib/lb/JV.pl
-lib/lb/LF.pl
-lib/lb/NL.pl
-lib/lb/NS.pl
-lib/lb/NU.pl
-lib/lb/OP.pl
-lib/lb/PO.pl
-lib/lb/PR.pl
-lib/lb/QU.pl
-lib/lb/SA.pl
-lib/lb/SG.pl
-lib/lb/SP.pl
-lib/lb/SY.pl
-lib/lb/WJ.pl
-lib/lb/XX.pl
-lib/lb/ZW.pl
-lib/nt/De.pl
-lib/nt/Di.pl
-lib/nt/Nu.pl
+lib/AHex/N.pl
+lib/AHex/Y.pl
+lib/Age/1_1.pl
+lib/Age/2_0.pl
+lib/Age/2_1.pl
+lib/Age/3_0.pl
+lib/Age/3_1.pl
+lib/Age/3_2.pl
+lib/Age/4_0.pl
+lib/Age/4_1.pl
+lib/Age/5_0.pl
+lib/Age/5_1.pl
+lib/Age/Unassign.pl
+lib/Alpha/N.pl
+lib/Alpha/Y.pl
+lib/Bc/AL.pl
+lib/Bc/AN.pl
+lib/Bc/B.pl
+lib/Bc/BN.pl
+lib/Bc/CS.pl
+lib/Bc/EN.pl
+lib/Bc/ES.pl
+lib/Bc/ET.pl
+lib/Bc/L.pl
+lib/Bc/LRE.pl
+lib/Bc/LRO.pl
+lib/Bc/NSM.pl
+lib/Bc/ON.pl
+lib/Bc/PDF.pl
+lib/Bc/R.pl
+lib/Bc/RLE.pl
+lib/Bc/RLO.pl
+lib/Bc/S.pl
+lib/Bc/WS.pl
+lib/BidiC/N.pl
+lib/BidiC/Y.pl
+lib/BidiM/N.pl
+lib/BidiM/Y.pl
+lib/Blk/ASCII.pl
+lib/Blk/AegeanNu.pl
+lib/Blk/Alphabet.pl
+lib/Blk/Ancient2.pl
+lib/Blk/AncientG.pl
+lib/Blk/AncientS.pl
+lib/Blk/Arabic.pl
+lib/Blk/ArabicP2.pl
+lib/Blk/ArabicPr.pl
+lib/Blk/ArabicSu.pl
+lib/Blk/Armenian.pl
+lib/Blk/Arrows.pl
+lib/Blk/Balinese.pl
+lib/Blk/Bengali.pl
+lib/Blk/BlockEle.pl
+lib/Blk/Bopomof2.pl
+lib/Blk/Bopomofo.pl
+lib/Blk/BoxDrawi.pl
+lib/Blk/BrailleP.pl
+lib/Blk/Buginese.pl
+lib/Blk/Buhid.pl
+lib/Blk/Byzantin.pl
+lib/Blk/CJKComp2.pl
+lib/Blk/CJKComp3.pl
+lib/Blk/CJKComp4.pl
+lib/Blk/CJKCompa.pl
+lib/Blk/CJKRadic.pl
+lib/Blk/CJKStrok.pl
+lib/Blk/CJKSymbo.pl
+lib/Blk/CJKUnif2.pl
+lib/Blk/CJKUnif3.pl
+lib/Blk/CJKUnifi.pl
+lib/Blk/Canadian.pl
+lib/Blk/Carian.pl
+lib/Blk/Cham.pl
+lib/Blk/Cherokee.pl
+lib/Blk/Combini2.pl
+lib/Blk/Combini3.pl
+lib/Blk/Combini4.pl
+lib/Blk/Combinin.pl
+lib/Blk/ControlP.pl
+lib/Blk/Coptic.pl
+lib/Blk/Counting.pl
+lib/Blk/Cuneifo2.pl
+lib/Blk/Cuneifor.pl
+lib/Blk/Currency.pl
+lib/Blk/CypriotS.pl
+lib/Blk/Cyrilli2.pl
+lib/Blk/Cyrilli3.pl
+lib/Blk/Cyrilli4.pl
+lib/Blk/Cyrillic.pl
+lib/Blk/Deseret.pl
+lib/Blk/Devanaga.pl
+lib/Blk/Dingbats.pl
+lib/Blk/DominoTi.pl
+lib/Blk/Enclose2.pl
+lib/Blk/Enclosed.pl
+lib/Blk/Ethiopi2.pl
+lib/Blk/Ethiopi3.pl
+lib/Blk/Ethiopic.pl
+lib/Blk/GeneralP.pl
+lib/Blk/Geometri.pl
+lib/Blk/Georgia2.pl
+lib/Blk/Georgian.pl
+lib/Blk/Glagolit.pl
+lib/Blk/Gothic.pl
+lib/Blk/Greek.pl
+lib/Blk/GreekExt.pl
+lib/Blk/Gujarati.pl
+lib/Blk/Gurmukhi.pl
+lib/Blk/Halfwidt.pl
+lib/Blk/HangulCo.pl
+lib/Blk/HangulJa.pl
+lib/Blk/HangulSy.pl
+lib/Blk/Hanunoo.pl
+lib/Blk/Hebrew.pl
+lib/Blk/HighPriv.pl
+lib/Blk/HighSurr.pl
+lib/Blk/Hiragana.pl
+lib/Blk/IPAExten.pl
+lib/Blk/Ideograp.pl
+lib/Blk/Kanbun.pl
+lib/Blk/KangxiRa.pl
+lib/Blk/Kannada.pl
+lib/Blk/Katakan2.pl
+lib/Blk/Katakana.pl
+lib/Blk/KayahLi.pl
+lib/Blk/Kharosht.pl
+lib/Blk/Khmer.pl
+lib/Blk/KhmerSym.pl
+lib/Blk/Lao.pl
+lib/Blk/Latin1.pl
+lib/Blk/LatinEx2.pl
+lib/Blk/LatinEx3.pl
+lib/Blk/LatinEx4.pl
+lib/Blk/LatinEx5.pl
+lib/Blk/LatinExt.pl
+lib/Blk/Lepcha.pl
+lib/Blk/Letterli.pl
+lib/Blk/Limbu.pl
+lib/Blk/LinearBI.pl
+lib/Blk/LinearBS.pl
+lib/Blk/LowSurro.pl
+lib/Blk/Lycian.pl
+lib/Blk/Lydian.pl
+lib/Blk/MahjongT.pl
+lib/Blk/Malayala.pl
+lib/Blk/Mathema2.pl
+lib/Blk/Mathemat.pl
+lib/Blk/Miscell2.pl
+lib/Blk/Miscell3.pl
+lib/Blk/Miscell4.pl
+lib/Blk/Miscell5.pl
+lib/Blk/Miscella.pl
+lib/Blk/Modifier.pl
+lib/Blk/Mongolia.pl
+lib/Blk/MusicalS.pl
+lib/Blk/Myanmar.pl
+lib/Blk/NKo.pl
+lib/Blk/NewTaiLu.pl
+lib/Blk/NoBlock.pl
+lib/Blk/NumberFo.pl
+lib/Blk/Ogham.pl
+lib/Blk/OlChiki.pl
+lib/Blk/OldItali.pl
+lib/Blk/OldPersi.pl
+lib/Blk/OpticalC.pl
+lib/Blk/Oriya.pl
+lib/Blk/Osmanya.pl
+lib/Blk/PhagsPa.pl
+lib/Blk/Phaistos.pl
+lib/Blk/Phoenici.pl
+lib/Blk/Phoneti2.pl
+lib/Blk/Phonetic.pl
+lib/Blk/PrivateU.pl
+lib/Blk/Rejang.pl
+lib/Blk/Runic.pl
+lib/Blk/Saurasht.pl
+lib/Blk/Shavian.pl
+lib/Blk/Sinhala.pl
+lib/Blk/SmallFor.pl
+lib/Blk/SpacingM.pl
+lib/Blk/Specials.pl
+lib/Blk/Sundanes.pl
+lib/Blk/Superscr.pl
+lib/Blk/Supplem2.pl
+lib/Blk/Supplem3.pl
+lib/Blk/Supplem4.pl
+lib/Blk/Supplem5.pl
+lib/Blk/Supplem6.pl
+lib/Blk/Suppleme.pl
+lib/Blk/SylotiNa.pl
+lib/Blk/Syriac.pl
+lib/Blk/Tagalog.pl
+lib/Blk/Tagbanwa.pl
+lib/Blk/Tags.pl
+lib/Blk/TaiLe.pl
+lib/Blk/TaiXuanJ.pl
+lib/Blk/Tamil.pl
+lib/Blk/Telugu.pl
+lib/Blk/Thaana.pl
+lib/Blk/Thai.pl
+lib/Blk/Tibetan.pl
+lib/Blk/Tifinagh.pl
+lib/Blk/Ugaritic.pl
+lib/Blk/Vai.pl
+lib/Blk/Variati2.pl
+lib/Blk/Variatio.pl
+lib/Blk/Vertical.pl
+lib/Blk/YiRadica.pl
+lib/Blk/YiSyllab.pl
+lib/Blk/YijingHe.pl
+lib/CE/N.pl
+lib/CE/Y.pl
+lib/Ccc/10.pl
+lib/Ccc/103.pl
+lib/Ccc/107.pl
+lib/Ccc/11.pl
+lib/Ccc/118.pl
+lib/Ccc/12.pl
+lib/Ccc/122.pl
+lib/Ccc/129.pl
+lib/Ccc/13.pl
+lib/Ccc/130.pl
+lib/Ccc/132.pl
+lib/Ccc/14.pl
+lib/Ccc/15.pl
+lib/Ccc/16.pl
+lib/Ccc/17.pl
+lib/Ccc/18.pl
+lib/Ccc/19.pl
+lib/Ccc/20.pl
+lib/Ccc/21.pl
+lib/Ccc/214.pl
+lib/Ccc/22.pl
+lib/Ccc/23.pl
+lib/Ccc/24.pl
+lib/Ccc/25.pl
+lib/Ccc/26.pl
+lib/Ccc/27.pl
+lib/Ccc/28.pl
+lib/Ccc/29.pl
+lib/Ccc/30.pl
+lib/Ccc/31.pl
+lib/Ccc/32.pl
+lib/Ccc/33.pl
+lib/Ccc/34.pl
+lib/Ccc/35.pl
+lib/Ccc/36.pl
+lib/Ccc/84.pl
+lib/Ccc/91.pl
+lib/Ccc/A.pl
+lib/Ccc/AL.pl
+lib/Ccc/AR.pl
+lib/Ccc/ATAR.pl
+lib/Ccc/ATB.pl
+lib/Ccc/ATBL.pl
+lib/Ccc/B.pl
+lib/Ccc/BL.pl
+lib/Ccc/BR.pl
+lib/Ccc/DA.pl
+lib/Ccc/DB.pl
+lib/Ccc/IS.pl
+lib/Ccc/KV.pl
+lib/Ccc/L.pl
+lib/Ccc/NK.pl
+lib/Ccc/NR.pl
+lib/Ccc/OV.pl
+lib/Ccc/R.pl
+lib/Ccc/VR.pl
+lib/CompEx/N.pl
+lib/CompEx/Y.pl
+lib/DI/N.pl
+lib/DI/Y.pl
+lib/Dash/N.pl
+lib/Dash/Y.pl
+lib/Dep/N.pl
+lib/Dep/Y.pl
+lib/Dia/N.pl
+lib/Dia/Y.pl
+lib/Dt/Com.pl
+lib/Dt/Enc.pl
+lib/Dt/Fin.pl
+lib/Dt/Font.pl
+lib/Dt/Fra.pl
+lib/Dt/Init.pl
+lib/Dt/Iso.pl
+lib/Dt/Med.pl
+lib/Dt/Nar.pl
+lib/Dt/Nb.pl
+lib/Dt/NonCanon.pl
+lib/Dt/None.pl
+lib/Dt/Sml.pl
+lib/Dt/Sqr.pl
+lib/Dt/Sub.pl
+lib/Dt/Sup.pl
+lib/Dt/Vert.pl
+lib/Ea/A.pl
+lib/Ea/F.pl
+lib/Ea/H.pl
+lib/Ea/N.pl
+lib/Ea/Na.pl
+lib/Ea/W.pl
+lib/Ext/N.pl
+lib/Ext/Y.pl
+lib/GCB/CN.pl
+lib/GCB/CR.pl
+lib/GCB/EX.pl
+lib/GCB/L.pl
+lib/GCB/LF.pl
+lib/GCB/LV.pl
+lib/GCB/LVT.pl
+lib/GCB/PP.pl
+lib/GCB/SM.pl
+lib/GCB/T.pl
+lib/GCB/V.pl
+lib/GCB/XX.pl
+lib/Gc/C.pl
+lib/Gc/Cc.pl
+lib/Gc/Cf.pl
+lib/Gc/Cn.pl
+lib/Gc/Co.pl
+lib/Gc/Cs.pl
+lib/Gc/L.pl
+lib/Gc/LC.pl
+lib/Gc/Ll.pl
+lib/Gc/Lm.pl
+lib/Gc/Lo.pl
+lib/Gc/Lt.pl
+lib/Gc/Lu.pl
+lib/Gc/M.pl
+lib/Gc/Mc.pl
+lib/Gc/Me.pl
+lib/Gc/Mn.pl
+lib/Gc/N.pl
+lib/Gc/Nl.pl
+lib/Gc/No.pl
+lib/Gc/P.pl
+lib/Gc/Pd.pl
+lib/Gc/Pe.pl
+lib/Gc/Pf.pl
+lib/Gc/Pi.pl
+lib/Gc/Po.pl
+lib/Gc/Ps.pl
+lib/Gc/S.pl
+lib/Gc/Sc.pl
+lib/Gc/Sk.pl
+lib/Gc/Sm.pl
+lib/Gc/So.pl
+lib/Gc/Z.pl
+lib/Gc/Zl.pl
+lib/Gc/Zp.pl
+lib/Gc/Zs.pl
+lib/GrBase/N.pl
+lib/GrBase/Y.pl
+lib/GrExt/N.pl
+lib/GrExt/Y.pl
+lib/Hex/N.pl
+lib/Hex/Y.pl
+lib/Hst/NA.pl
+lib/Hyphen/N.pl
+lib/Hyphen/Y.pl
+lib/IDC/N.pl
+lib/IDC/Y.pl
+lib/IDS/N.pl
+lib/IDS/Y.pl
+lib/IDSB/N.pl
+lib/IDSB/Y.pl
+lib/IDST/N.pl
+lib/IDST/Y.pl
+lib/Ideo/N.pl
+lib/Ideo/Y.pl
+lib/In/2_0.pl
+lib/In/2_1.pl
+lib/In/3_0.pl
+lib/In/3_1.pl
+lib/In/3_2.pl
+lib/In/4_0.pl
+lib/In/4_1.pl
+lib/In/5_0.pl
+lib/In/5_1.pl
+lib/Jg/Ain.pl
+lib/Jg/Alaph.pl
+lib/Jg/Alef.pl
+lib/Jg/Beh.pl
+lib/Jg/Beth.pl
+lib/Jg/Burushas.pl
+lib/Jg/Dal.pl
+lib/Jg/DalathRi.pl
+lib/Jg/E.pl
+lib/Jg/Fe.pl
+lib/Jg/Feh.pl
+lib/Jg/FinalSem.pl
+lib/Jg/Gaf.pl
+lib/Jg/Gamal.pl
+lib/Jg/Hah.pl
+lib/Jg/HamzaOnH.pl
+lib/Jg/He.pl
+lib/Jg/Heh.pl
+lib/Jg/HehGoal.pl
+lib/Jg/Heth.pl
+lib/Jg/Kaf.pl
+lib/Jg/Kaph.pl
+lib/Jg/Khaph.pl
+lib/Jg/KnottedH.pl
+lib/Jg/Lam.pl
+lib/Jg/Lamadh.pl
+lib/Jg/Meem.pl
+lib/Jg/Mim.pl
+lib/Jg/NoJoinin.pl
+lib/Jg/Noon.pl
+lib/Jg/Nun.pl
+lib/Jg/Pe.pl
+lib/Jg/Qaf.pl
+lib/Jg/Qaph.pl
+lib/Jg/Reh.pl
+lib/Jg/Reversed.pl
+lib/Jg/Sad.pl
+lib/Jg/Sadhe.pl
+lib/Jg/Seen.pl
+lib/Jg/Semkath.pl
+lib/Jg/Shin.pl
+lib/Jg/SwashKaf.pl
+lib/Jg/SyriacWa.pl
+lib/Jg/Tah.pl
+lib/Jg/Taw.pl
+lib/Jg/TehMarbu.pl
+lib/Jg/Teth.pl
+lib/Jg/Waw.pl
+lib/Jg/Yeh.pl
+lib/Jg/YehBarre.pl
+lib/Jg/YehWithT.pl
+lib/Jg/Yudh.pl
+lib/Jg/YudhHe.pl
+lib/Jg/Zain.pl
+lib/Jg/Zhain.pl
+lib/JoinC/N.pl
+lib/JoinC/Y.pl
+lib/Jt/C.pl
+lib/Jt/D.pl
+lib/Jt/R.pl
+lib/Jt/T.pl
+lib/Jt/U.pl
+lib/LOE/N.pl
+lib/Lb/AI.pl
+lib/Lb/AL.pl
+lib/Lb/B2.pl
+lib/Lb/BA.pl
+lib/Lb/BB.pl
+lib/Lb/BK.pl
+lib/Lb/CB.pl
+lib/Lb/CL.pl
+lib/Lb/CM.pl
+lib/Lb/EX.pl
+lib/Lb/GL.pl
+lib/Lb/HY.pl
+lib/Lb/ID.pl
+lib/Lb/IN.pl
+lib/Lb/IS.pl
+lib/Lb/NL.pl
+lib/Lb/NS.pl
+lib/Lb/NU.pl
+lib/Lb/OP.pl
+lib/Lb/PO.pl
+lib/Lb/PR.pl
+lib/Lb/QU.pl
+lib/Lb/SA.pl
+lib/Lb/SG.pl
+lib/Lb/SP.pl
+lib/Lb/SY.pl
+lib/Lb/WJ.pl
+lib/Lb/XX.pl
+lib/Lb/ZW.pl
+lib/Lower/N.pl
+lib/Lower/Y.pl
+lib/Math/N.pl
+lib/Math/Y.pl
+lib/NChar/N.pl
+lib/NChar/Y.pl
+lib/NFCQC/Y.pl
+lib/NFDQC/N.pl
+lib/NFDQC/Y.pl
+lib/NFKCQC/M.pl
+lib/NFKCQC/N.pl
+lib/NFKCQC/Y.pl
+lib/NFKDQC/N.pl
+lib/Nt/De.pl
+lib/Nt/Di.pl
+lib/Nt/Nu.pl
+lib/Nv/0.pl
+lib/Nv/1.pl
+lib/Nv/10.pl
+lib/Nv/100.pl
+lib/Nv/1000.pl
+lib/Nv/10000.pl
+lib/Nv/100000.pl
+lib/Nv/10000000.pl
+lib/Nv/10000002.pl
+lib/Nv/11.pl
+lib/Nv/11_2.pl
+lib/Nv/12.pl
+lib/Nv/13.pl
+lib/Nv/13_2.pl
+lib/Nv/14.pl
+lib/Nv/15.pl
+lib/Nv/15_2.pl
+lib/Nv/16.pl
+lib/Nv/17.pl
+lib/Nv/17_2.pl
+lib/Nv/18.pl
+lib/Nv/19.pl
+lib/Nv/1_2.pl
+lib/Nv/1_3.pl
+lib/Nv/1_4.pl
+lib/Nv/1_5.pl
+lib/Nv/1_6.pl
+lib/Nv/1_8.pl
+lib/Nv/2.pl
+lib/Nv/20.pl
+lib/Nv/200.pl
+lib/Nv/2000.pl
+lib/Nv/20000.pl
+lib/Nv/21.pl
+lib/Nv/22.pl
+lib/Nv/23.pl
+lib/Nv/24.pl
+lib/Nv/25.pl
+lib/Nv/26.pl
+lib/Nv/27.pl
+lib/Nv/28.pl
+lib/Nv/29.pl
+lib/Nv/2_3.pl
+lib/Nv/2_5.pl
+lib/Nv/3.pl
+lib/Nv/30.pl
+lib/Nv/300.pl
+lib/Nv/3000.pl
+lib/Nv/30000.pl
+lib/Nv/31.pl
+lib/Nv/32.pl
+lib/Nv/33.pl
+lib/Nv/34.pl
+lib/Nv/35.pl
+lib/Nv/36.pl
+lib/Nv/37.pl
+lib/Nv/38.pl
+lib/Nv/39.pl
+lib/Nv/3_2.pl
+lib/Nv/3_4.pl
+lib/Nv/3_5.pl
+lib/Nv/3_8.pl
+lib/Nv/4.pl
+lib/Nv/40.pl
+lib/Nv/400.pl
+lib/Nv/4000.pl
+lib/Nv/40000.pl
+lib/Nv/41.pl
+lib/Nv/42.pl
+lib/Nv/43.pl
+lib/Nv/44.pl
+lib/Nv/45.pl
+lib/Nv/46.pl
+lib/Nv/47.pl
+lib/Nv/48.pl
+lib/Nv/49.pl
+lib/Nv/4_5.pl
+lib/Nv/5.pl
+lib/Nv/50.pl
+lib/Nv/500.pl
+lib/Nv/5000.pl
+lib/Nv/50000.pl
+lib/Nv/5_2.pl
+lib/Nv/5_6.pl
+lib/Nv/5_8.pl
+lib/Nv/6.pl
+lib/Nv/60.pl
+lib/Nv/600.pl
+lib/Nv/6000.pl
+lib/Nv/60000.pl
+lib/Nv/7.pl
+lib/Nv/70.pl
+lib/Nv/700.pl
+lib/Nv/7000.pl
+lib/Nv/70000.pl
+lib/Nv/7_2.pl
+lib/Nv/7_8.pl
+lib/Nv/8.pl
+lib/Nv/80.pl
+lib/Nv/800.pl
+lib/Nv/8000.pl
+lib/Nv/80000.pl
+lib/Nv/9.pl
+lib/Nv/90.pl
+lib/Nv/900.pl
+lib/Nv/9000.pl
+lib/Nv/90000.pl
+lib/Nv/9_2.pl
+lib/Nv/NaN.pl
+lib/Nv/_1_2.pl
+lib/OAlpha/N.pl
+lib/OAlpha/Y.pl
+lib/ODI/N.pl
+lib/ODI/Y.pl
+lib/OGrExt/N.pl
+lib/OGrExt/Y.pl
+lib/OIDC/N.pl
+lib/OIDC/Y.pl
+lib/OIDS/N.pl
+lib/OIDS/Y.pl
+lib/OLower/N.pl
+lib/OLower/Y.pl
+lib/OMath/N.pl
+lib/OMath/Y.pl
+lib/OUpper/N.pl
+lib/OUpper/Y.pl
+lib/PatSyn/N.pl
+lib/PatSyn/Y.pl
+lib/PatWS/N.pl
+lib/PatWS/Y.pl
+lib/Perl/Alnum.pl
+lib/Perl/Any.pl
+lib/Perl/Assigned.pl
+lib/Perl/Blank.pl
+lib/Perl/Cased.pl
+lib/Perl/Graph.pl
+lib/Perl/PerlSpac.pl
+lib/Perl/PerlWord.pl
+lib/Perl/PosixAln.pl
+lib/Perl/PosixAlp.pl
+lib/Perl/PosixBla.pl
+lib/Perl/PosixCnt.pl
+lib/Perl/PosixDig.pl
+lib/Perl/PosixGra.pl
+lib/Perl/PosixLow.pl
+lib/Perl/PosixPri.pl
+lib/Perl/PosixPun.pl
+lib/Perl/PosixSpa.pl
+lib/Perl/PosixUpp.pl
+lib/Perl/Print.pl
+lib/Perl/SpacePer.pl
+lib/Perl/VertSpac.pl
+lib/Perl/Word.pl
+lib/Perl/_CaseIgn.pl
+lib/Perl/_GCBExte.pl
+lib/QMark/N.pl
+lib/QMark/Y.pl
+lib/Radical/N.pl
+lib/Radical/Y.pl
+lib/SB/AT.pl
+lib/SB/CL.pl
+lib/SB/EX.pl
+lib/SB/FO.pl
+lib/SB/LE.pl
+lib/SB/LO.pl
+lib/SB/SC.pl
+lib/SB/SE.pl
+lib/SB/ST.pl
+lib/SB/Sp.pl
+lib/SB/UP.pl
+lib/SB/XX.pl
+lib/SD/N.pl
+lib/SD/Y.pl
+lib/STerm/N.pl
+lib/STerm/Y.pl
+lib/Sc/Arab.pl
+lib/Sc/Armn.pl
+lib/Sc/Bali.pl
+lib/Sc/Beng.pl
+lib/Sc/Bopo.pl
+lib/Sc/Bugi.pl
+lib/Sc/Buhd.pl
+lib/Sc/Cans.pl
+lib/Sc/Cari.pl
+lib/Sc/Cham.pl
+lib/Sc/Cher.pl
+lib/Sc/Copt.pl
+lib/Sc/Cprt.pl
+lib/Sc/Cyrl.pl
+lib/Sc/Deva.pl
+lib/Sc/Ethi.pl
+lib/Sc/Geor.pl
+lib/Sc/Glag.pl
+lib/Sc/Goth.pl
+lib/Sc/Grek.pl
+lib/Sc/Gujr.pl
+lib/Sc/Guru.pl
+lib/Sc/Han.pl
+lib/Sc/Hang.pl
+lib/Sc/Hano.pl
+lib/Sc/Hebr.pl
+lib/Sc/Hira.pl
+lib/Sc/Ital.pl
+lib/Sc/Kana.pl
+lib/Sc/Khar.pl
+lib/Sc/Khmr.pl
+lib/Sc/Knda.pl
+lib/Sc/Lao.pl
+lib/Sc/Latn.pl
+lib/Sc/Lepc.pl
+lib/Sc/Limb.pl
+lib/Sc/Linb.pl
+lib/Sc/Lyci.pl
+lib/Sc/Lydi.pl
+lib/Sc/Mlym.pl
+lib/Sc/Mong.pl
+lib/Sc/Mymr.pl
+lib/Sc/Nko.pl
+lib/Sc/Ogam.pl
+lib/Sc/Orya.pl
+lib/Sc/Osma.pl
+lib/Sc/Phag.pl
+lib/Sc/Phnx.pl
+lib/Sc/Qaai.pl
+lib/Sc/Rjng.pl
+lib/Sc/Runr.pl
+lib/Sc/Saur.pl
+lib/Sc/Sinh.pl
+lib/Sc/Sund.pl
+lib/Sc/Sylo.pl
+lib/Sc/Syrc.pl
+lib/Sc/Tagb.pl
+lib/Sc/Tale.pl
+lib/Sc/Talu.pl
+lib/Sc/Taml.pl
+lib/Sc/Telu.pl
+lib/Sc/Tfng.pl
+lib/Sc/Tglg.pl
+lib/Sc/Thaa.pl
+lib/Sc/Thai.pl
+lib/Sc/Tibt.pl
+lib/Sc/Ugar.pl
+lib/Sc/Vai.pl
+lib/Sc/Xpeo.pl
+lib/Sc/Xsux.pl
+lib/Sc/Yi.pl
+lib/Sc/Zyyy.pl
+lib/Sc/Zzzz.pl
+lib/Space/N.pl
+lib/Space/Y.pl
+lib/Term/N.pl
+lib/Term/Y.pl
+lib/UIdeo/N.pl
+lib/UIdeo/Y.pl
+lib/Upper/N.pl
+lib/Upper/Y.pl
+lib/VS/N.pl
+lib/VS/Y.pl
+lib/WB/EX.pl
+lib/WB/KA.pl
+lib/WB/LE.pl
+lib/WB/MB.pl
+lib/WB/ML.pl
+lib/WB/MN.pl
+lib/WB/NL.pl
+lib/WB/NU.pl
+lib/WB/XX.pl
+lib/XIDC/N.pl
+lib/XIDC/Y.pl
+lib/XIDS/N.pl
+lib/XIDS/Y.pl
 
-# 24 input files
-# 514 output files
+# 35 input files
+# 771 output files
 
 # End list
diff --git a/lib/utf8_heavy.pl b/lib/utf8_heavy.pl
index 41a0662fd8..250eb69f05 100644
--- a/lib/utf8_heavy.pl
+++ b/lib/utf8_heavy.pl
@@ -8,8 +8,6 @@ sub DESTROY {}
 
 my %Cache;
 
-our (%PropertyAlias, %PA_reverse, %PropValueAlias, %PVA_reverse, %PVA_abbr_map);
-
 sub croak { require Carp; Carp::croak(@_) }
 
 ##
@@ -17,272 +15,450 @@ sub croak { require Carp; Carp::croak(@_) }
 ## It's a data structure that encodes a set of Unicode characters.
 ##
 
-sub SWASHNEW {
-    my ($class, $type, $list, $minbits, $none) = @_;
-    local $^D = 0 if $^D;
-
-    print STDERR "SWASHNEW @_\n" if DEBUG;
-
-    ##
-    ## Get the list of codepoints for the type.
-    ## Called from swash_init (see utf8.c) or SWASHNEW itself.
-    ##
-    ## Callers of swash_init:
-    ##     op.c:pmtrans             -- for tr/// and y///
-    ##     regexec.c:regclass_swash -- for /[]/, \p, and \P
-    ##     utf8.c:is_utf8_common    -- for common Unicode properties
-    ##     utf8.c:to_utf8_case      -- for lc, uc, ucfirst, etc. and //i
-    ##
-    ## Given a $type, our goal is to fill $list with the set of codepoint
-    ## ranges. If $type is false, $list passed is used.
-    ##
-    ## $minbits:
-    ##     For binary properties, $minbits must be 1.
-    ##     For character mappings (case and transliteration), $minbits must
-    ##     be a number except 1.
-    ##
-    ## $list (or that filled according to $type):
-    ##     Refer to perlunicode.pod, "User-Defined Character Properties."
-    ##     
-    ##     For binary properties, only characters with the property value
-    ##     of True should be listed. The 3rd column, if any, will be ignored.
-    ##
-    ## To make the parsing of $type clear, this code takes the a rather
-    ## unorthodox approach of last'ing out of the block once we have the
-    ## info we need. Were this to be a subroutine, the 'last' would just
-    ## be a 'return'.
-    ##
-    my $file; ## file to load data from, and also part of the %Cache key.
-    my $ListSorted = 0;
-
-    if ($type)
-    {
-        $type =~ s/^\s+//;
-        $type =~ s/\s+$//;
-
-        print STDERR "type = $type\n" if DEBUG;
-
-      GETFILE:
+{
+    # If a floating point number is within this distance from the value of a
+    # fraction, it is considered to be that fraction, even if many more digits
+    # are specified that don't exactly match.
+    my $min_floating_slop;
+
+    sub SWASHNEW {
+        my ($class, $type, $list, $minbits, $none) = @_;
+        local $^D = 0 if $^D;
+
+        print STDERR __LINE__, ": ", join(", ", @_), "\n" if DEBUG;
+
+        ##
+        ## Get the list of codepoints for the type.
+        ## Called from swash_init (see utf8.c) or SWASHNEW itself.
+        ##
+        ## Callers of swash_init:
+        ##     op.c:pmtrans             -- for tr/// and y///
+        ##     regexec.c:regclass_swash -- for /[]/, \p, and \P
+        ##     utf8.c:is_utf8_common    -- for common Unicode properties
+        ##     utf8.c:to_utf8_case      -- for lc, uc, ucfirst, etc. and //i
+        ##
+        ## Given a $type, our goal is to fill $list with the set of codepoint
+        ## ranges. If $type is false, $list passed is used.
+        ##
+        ## $minbits:
+        ##     For binary properties, $minbits must be 1.
+        ##     For character mappings (case and transliteration), $minbits must
+        ##     be a number except 1.
+        ##
+        ## $list (or that filled according to $type):
+        ##     Refer to perlunicode.pod, "User-Defined Character Properties."
+        ##     
+        ##     For binary properties, only characters with the property value
+        ##     of True should be listed. The 3rd column, if any, will be ignored
+        ##
+        ## $none is undocumented, so I'm (khw) trying to do some documentation
+        ## of it now.  It appears to be if there is a mapping in an input file
+        ## that maps to 'XXXX', then that is replaced by $none+1, expressed in
+        ## hexadecimal.  The only place I found it possibly used was in
+        ## S_pmtrans in op.c.
+        ##
+        ## To make the parsing of $type clear, this code takes the a rather
+        ## unorthodox approach of last'ing out of the block once we have the
+        ## info we need. Were this to be a subroutine, the 'last' would just
+        ## be a 'return'.
+        ##
+        my $file; ## file to load data from, and also part of the %Cache key.
+        my $ListSorted = 0;
+
+        # Change this to get a different set of Unicode tables
+        my $unicore_dir = 'unicore';
+
+        if ($type)
         {
-	    ##
-	    ## It could be a user-defined property.
-	    ##
-
-	    my $caller1 = $type =~ s/(.+)::// ? $1 : caller(1);
-
-	    if (defined $caller1 && $type =~ /^(?:\w+)$/) {
-		my $prop = "${caller1}::$type";
-		if (exists &{$prop}) {
-		    no strict 'refs';
-		    
-		    $list = &{$prop};
-		    last GETFILE;
-		}
-	    }
-
-            my $wasIs;
-
-            ($wasIs = $type =~ s/^Is(?:\s+|[-_])?//i)
-              or
-            $type =~ s/^(?:(?:General(?:\s+|_)?)?Category|gc)\s*[:=]\s*//i
-              or
-            $type =~ s/^(?:Script|sc)\s*[:=]\s*//i
-              or
-            $type =~ s/^Block\s*[:=]\s*/In/i;
-
-
-	    ##
-	    ## See if it's in some enumeration.
-	    ##
-	    require "unicore/PVA.pl";
-	    if ($type =~ /^([\w\s]+)[:=]\s*(.*)/) {
-		my ($enum, $val) = (lc $1, lc $2);
-		$enum =~ tr/ _-//d;
-		$val =~ tr/ _-//d;
-
-		my $pa = $PropertyAlias{$enum} ? $enum : $PA_reverse{$enum};
-		my $f = $PropValueAlias{$pa}{$val} ? $val : $PVA_reverse{$pa}{lc $val};
-
-		if ($pa and $f) {
-		    $pa = "gc_sc" if $pa eq "gc" or $pa eq "sc";
-		    $file = "unicore/lib/$pa/$PVA_abbr_map{$pa}{lc $f}.pl";
-		    last GETFILE;
-		}
-	    }
-	    else {
-		my $t = lc $type;
-		$t =~ tr/ _-//d;
-
-		if ($PropValueAlias{gc}{$t} or $PropValueAlias{sc}{$t}) {
-		    $file = "unicore/lib/gc_sc/$PVA_abbr_map{gc_sc}{$t}.pl";
-		    last GETFILE;
-		}
-	    }
-
-            ##
-            ## See if it's in the direct mapping table.
-            ##
-            require "unicore/Exact.pl";
-            if (my $base = $utf8::Exact{$type}) {
-                $file = "unicore/lib/gc_sc/$base.pl";
-                last GETFILE;
+            $type =~ s/^\s+//;
+            $type =~ s/\s+$//;
+
+            print STDERR __LINE__, ": type = $type\n" if DEBUG;
+
+        GETFILE:
+            {
+                ##
+                ## It could be a user-defined property.
+                ##
+
+                my $caller1 = $type =~ s/(.+)::// ? $1 : caller(1);
+
+                if (defined $caller1 && $type =~ /^(?:\w+)$/) {
+                    my $prop = "${caller1}::$type";
+                    if (exists &{$prop}) {
+                        no strict 'refs';
+                        
+                        $list = &{$prop};
+                        last GETFILE;
+                    }
+                }
+
+                require "$unicore_dir/Heavy.pl";
+
+                # Everything is caseless matching
+                my $property_and_table = lc $type;
+                print STDERR __LINE__, ": $property_and_table\n" if DEBUG;
+
+                # See if is of the compound form 'property=value', where the
+                # value indicates the table we should use.
+                my ($property, $table, @remainder) =
+                                    split /\s*[:=]\s*/, $property_and_table, -1;
+                return $type if @remainder;
+
+                my $prefix;
+                if (! defined $table) {
+                        
+                    # Here, is the single form.  The property becomes empty, and
+                    # the whole value is the table.
+                    $table = $property;
+                    $prefix = $property = "";
+                } else {
+                    print STDERR __LINE__, ": $property\n" if DEBUG;
+
+                    # Here it is the compound property=table form.  The property
+                    # name is always loosely matched, which means remove any of
+                    # these:
+                    $property =~ s/[_\s-]//g;
+
+                    # And convert to canonical form.  Quit if not valid.
+                    $property = $utf8::loose_property_name_of{$property};
+                    return $type unless defined $property;
+
+                    $prefix = "$property=";
+
+                    # If the rhs looks like it is a number...
+                    print STDERR __LINE__, ": table=$table\n" if DEBUG;
+                    if ($table =~ qr{ ^ [ \s 0-9 _  + / . -]+ $ }x) {
+                        print STDERR __LINE__, ": table=$table\n" if DEBUG;
+
+                        # Don't allow leading nor trailing slashes 
+                        return $type if $table =~ / ^ \/ | \/ $ /x;
+
+                        # Split on slash, in case it is a rational, like \p{1/5}
+                        my @parts = split qr{ \s* / \s* }x, $table, -1;
+                        print __LINE__, ": $type\n" if @parts > 2 && DEBUG;
+
+                        # Can have maximum of one slash
+                        return $type if @parts > 2;
+
+                        foreach my $part (@parts) {
+                            print __LINE__, ": part=$part\n" if DEBUG;
+
+                            $part =~ s/^\+\s*//;    # Remove leading plus
+                            $part =~ s/^-\s*/-/;    # Remove blanks after unary
+                                                    # minus
+
+                            # Remove underscores between digits.
+                            $part =~ s/( ?<= [0-9] ) _ (?= [0-9] ) //xg;
+
+                            # No leading zeros (but don't make a single '0'
+                            # into a null string)
+                            $part =~ s/ ^ ( -? ) 0+ /$1/x;
+                            $part .= '0' if $part eq '-' || $part eq "";
+
+                            # No trailing zeros after a decimal point
+                            $part =~ s/ ( \. .*? ) 0+ $ /$1/x;
+
+                            # Begin with a 0 if a leading decimal point
+                            $part =~ s/ ^ ( -? ) \. /${1}0./x;
+
+                            # Ensure not a trailing decimal point: turn into an
+                            # integer
+                            $part =~ s/ \. $ //x;
+
+                            print STDERR __LINE__, ": part=$part\n" if DEBUG;
+                            #return $type if $part eq "";
+                            
+                            # Result better look like a number.  (This test is
+                            # needed because, for example could have a plus in
+                            # the middle.)
+                            return $type if $part
+                                            !~ / ^ -? [0-9]+ ( \. [0-9]+)? $ /x;
+                        }
+
+                        #  If a rational...
+                        if (@parts == 2) {
+
+                            # If denominator is negative, get rid of it, and ...
+                            if ($parts[1] =~ s/^-//) {
+
+                                # If numerator is also negative, convert the
+                                # whole thing to positive, or move the minus to
+                                # the numerator
+                                if ($parts[0] !~ s/^-//) {
+                                    $parts[0] = '-' . $parts[0];
+                                }
+                            }
+                            $table = join '/', @parts;
+                        }
+                        elsif ($property ne 'nv' || $parts[0] !~ /\./) {
+
+                            # Here is not numeric value, or doesn't have a
+                            # decimal point.  No further manipulation is
+                            # necessary.  (Note the hard-coded property name.
+                            # This could fail if other properties eventually
+                            # had fractions as well; perhaps the cjk ones
+                            # could evolve to do that.  This hard-coding could
+                            # be fixed by mktables generating a list of
+                            # properties that could have fractions.)
+                            $table = $parts[0];
+                        } else {
+
+                            # Here is a floating point numeric_value.  Try to
+                            # convert to rational.  First see if is in the list
+                            # of known ones.
+                            if (exists $utf8::nv_floating_to_rational{$parts[0]}) {
+                                $table = $utf8::nv_floating_to_rational{$parts[0]};
+                            } else {
+
+                                # Here not in the list.  See if is close
+                                # enough to something in the list.  First
+                                # determine what 'close enough' means.  It has
+                                # to be as tight as what mktables says is the
+                                # maximum slop, and as tight as how many
+                                # digits we were passed.  That is, if the user
+                                # said .667, .6667, .66667, etc.  we match as
+                                # many digits as they passed until get to
+                                # where it doesn't matter any more due to the
+                                # machine's precision.  If they said .6666668,
+                                # we fail.
+                                (my $fraction = $parts[0]) =~ s/^.*\.//;
+                                my $epsilon = 10 ** - (length($fraction));
+                                if ($epsilon > $utf8::max_floating_slop) {
+                                    $epsilon = $utf8::max_floating_slop;
+                                }
+
+                                # But it can't be tighter than the minimum
+                                # precision for this machine.  If haven't
+                                # already calculated that minimum, do so now.
+                                if (! defined $min_floating_slop) {
+
+                                    # Keep going down an order of magnitude
+                                    # until find that adding this quantity to
+                                    # 1 remains 1; but put an upper limit on
+                                    # this so in case this algorithm doesn't
+                                    # work properly on some platform, that we
+                                    # won't loop forever.
+                                    my $count = 0;
+                                    $min_floating_slop = 1;
+                                    while (1+ $min_floating_slop != 1
+                                           && $count++ < 50)
+                                    {
+                                        my $next = $min_floating_slop / 10;
+                                        last if $next == 0; # If underflows,
+                                                            # use previous one
+                                        $min_floating_slop = $next;
+                                        print STDERR __LINE__, ": min_float_slop=$min_floating_slop\n" if DEBUG;
+                                    }
+
+                                    # Back off a couple orders of magnitude,
+                                    # just to be safe.
+                                    $min_floating_slop *= 100;
+                                }
+                                    
+                                if ($epsilon < $min_floating_slop) {
+                                    $epsilon = $min_floating_slop;
+                                }
+                                print STDERR __LINE__, ": fraction=.$fraction; epsilon=$epsilon\n" if DEBUG;
+
+                                undef $table;
+
+                                # And for each possible rational in the table,
+                                # see if it is within epsilon of the input.
+                                foreach my $official
+                                        (keys %utf8::nv_floating_to_rational)
+                                {
+                                    print STDERR __LINE__, ": epsilon=$epsilon, official=$official, diff=", abs($parts[0] - $official), "\n" if DEBUG;
+                                    if (abs($parts[0] - $official) < $epsilon) {
+                                      $table =
+                                      $utf8::nv_floating_to_rational{$official};
+                                        last;
+                                    }
+                                }
+
+                                # Quit if didn't find one.
+                                return $type unless defined $table;
+                            }
+                        }
+                        print STDERR __LINE__, ": $property=$table\n" if DEBUG;
+                    }
+                }
+
+                # Combine lhs (if any) and rhs to get something that matches
+                # the syntax of the lookups.
+                $property_and_table = "$prefix$table";
+                print STDERR __LINE__, ": $property_and_table\n" if DEBUG;
+
+                # First try stricter matching.
+                $file = $utf8::stricter_to_file_of{$property_and_table};
+
+                # If didn't find it, try again with looser matching by editing
+                # out the applicable characters on the rhs and looking up
+                # again.
+                if (! defined $file) {
+                    $table =~ s/ [_\s-] //xg;
+                    $property_and_table = "$prefix$table";
+                    print STDERR __LINE__, ": $property_and_table\n" if DEBUG;
+                    $file = $utf8::loose_to_file_of{$property_and_table};
+                }
+
+                # Add the constant and go fetch it in.
+                if (defined $file) {
+                    if ($utf8::why_deprecated{$file}) {
+                        warnings::warnif('deprecated', "Use of '$type' in \\p{} or \\P{} is deprecated because: $utf8::why_deprecated{$file};");
+                    }
+                    $file= "$unicore_dir/lib/$file.pl";
+                    last GETFILE;
+                }
+                print STDERR __LINE__, ": didn't find $property_and_table\n" if DEBUG;
+
+                ##
+                ## See if it's a user-level "To".
+                ##
+
+                my $caller0 = caller(0);
+
+                if (defined $caller0 && $type =~ /^To(?:\w+)$/) {
+                    my $map = $caller0 . "::" . $type;
+
+                    if (exists &{$map}) {
+                        no strict 'refs';
+                        
+                        $list = &{$map};
+                        last GETFILE;
+                    }
+                }
+
+                ##
+                ## Last attempt -- see if it's a standard "To" name
+                ## (e.g. "ToLower")  ToTitle is used by ucfirst().
+                ## The user-level way to access ToDigit() and ToFold()
+                ## is to use Unicode::UCD.
+                ##
+                if ($type =~ /^To(Digit|Fold|Lower|Title|Upper)$/) {
+                    $file = "$unicore_dir/To/$1.pl";
+                    ## would like to test to see if $file actually exists....
+                    last GETFILE;
+                }
+
+                ##
+                ## If we reach this line, it's because we couldn't figure
+                ## out what to do with $type. Ouch.
+                ##
+
+                return $type;
             }
 
-            ##
-            ## If not there exactly, try the canonical form. The canonical
-            ## form is lowercased, with any separators (\s+|[-_]) removed.
-            ##
-            my $canonical = lc $type;
-            $canonical =~ s/(?<=[a-z\d])(?:\s+|[-_])(?=[a-z\d])//g;
-            print STDERR "canonical = $canonical\n" if DEBUG;
-
-            require "unicore/Canonical.pl";
-	    { no warnings "uninitialized";
-            if (my $base = ($utf8::Canonical{$canonical} || $utf8::Canonical{ lc $utf8::PropertyAlias{$canonical} })) {
-                $file = "unicore/lib/gc_sc/$base.pl";
-                last GETFILE;
-            }
+            if (defined $file) {
+                print STDERR __LINE__, ": found it (file='$file')\n" if DEBUG;
+
+                ##
+                ## If we reach here, it was due to a 'last GETFILE' above
+                ## (exception: user-defined properties and mappings), so we
+                ## have a filename, so now we load it if we haven't already.
+                ## If we have, return the cached results. The cache key is the
+                ## class and file to load.
+                ##
+                my $found = $Cache{$class, $file};
+                if ($found and ref($found) eq $class) {
+                    print STDERR __LINE__, ": Returning cached '$file' for \\p{$type}\n" if DEBUG;
+                    return $found;
+                }
+
+                local $@;
+                local $!;
+                $list = do $file; die $@ if $@;
             }
 
-	    ##
-	    ## See if it's a user-level "To".
-	    ##
-
-	    my $caller0 = caller(0);
-
-	    if (defined $caller0 && $type =~ /^To(?:\w+)$/) {
-		my $map = $caller0 . "::" . $type;
-
-		if (exists &{$map}) {
-		    no strict 'refs';
-		    
-		    $list = &{$map};
-		    last GETFILE;
-		}
-	    }
-
-            ##
-            ## Last attempt -- see if it's a standard "To" name
-	    ## (e.g. "ToLower")  ToTitle is used by ucfirst().
-	    ## The user-level way to access ToDigit() and ToFold()
-	    ## is to use Unicode::UCD.
-            ##
-            if ($type =~ /^To(Digit|Fold|Lower|Title|Upper)$/) {
-                $file = "unicore/To/$1.pl";
-                ## would like to test to see if $file actually exists....
-                last GETFILE;
-            }
-
-            ##
-            ## If we reach this line, it's because we couldn't figure
-            ## out what to do with $type. Ouch.
-            ##
-
-            return $type;
+            $ListSorted = 1; ## we know that these lists are sorted
         }
 
-	if (defined $file) {
-	    print STDERR "found it (file='$file')\n" if DEBUG;
-
-	    ##
-	    ## If we reach here, it was due to a 'last GETFILE' above
-	    ## (exception: user-defined properties and mappings), so we
-	    ## have a filename, so now we load it if we haven't already.
-	    ## If we have, return the cached results. The cache key is the
-	    ## class and file to load.
-	    ##
-	    my $found = $Cache{$class, $file};
-	    if ($found and ref($found) eq $class) {
-		print STDERR "Returning cached '$file' for \\p{$type}\n" if DEBUG;
-		return $found;
-	    }
-
-	    local $@;
-	    local $!;
-	    $list = do $file; die $@ if $@;
-	}
-
-        $ListSorted = 1; ## we know that these lists are sorted
-    }
-
-    my $extras;
-    my $bits = $minbits;
-
-    my $ORIG = $list;
-    if ($list) {
-	my @tmp = split(/^/m, $list);
-	my %seen;
-	no warnings;
-	$extras = join '', grep /^[^0-9a-fA-F]/, @tmp;
-	$list = join '',
-	    map  { $_->[1] }
-	    sort { $a->[0] <=> $b->[0] }
-	    map  { /^([0-9a-fA-F]+)/; [ CORE::hex($1), $_ ] }
-	    grep { /^([0-9a-fA-F]+)/ and not $seen{$1}++ } @tmp; # XXX doesn't do ranges right
-    }
+        my $extras;
+        my $bits = $minbits;
+
+        my $ORIG = $list;
+        if ($list) {
+            my @tmp = split(/^/m, $list);
+            my %seen;
+            no warnings;
+            $extras = join '', grep /^[^0-9a-fA-F]/, @tmp;
+            $list = join '',
+                map  { $_->[1] }
+                sort { $a->[0] <=> $b->[0] }
+                map  { /^([0-9a-fA-F]+)/; [ CORE::hex($1), $_ ] }
+                grep { /^([0-9a-fA-F]+)/ and not $seen{$1}++ } @tmp; # XXX doesn't do ranges right
+        }
 
-    if ($none) {
-	my $hextra = sprintf "%04x", $none + 1;
-	$list =~ s/\tXXXX$/\t$hextra/mg;
-    }
+        if ($none) {
+            my $hextra = sprintf "%04x", $none + 1;
+            $list =~ s/\tXXXX$/\t$hextra/mg;
+        }
 
-    if ($minbits != 1 && $minbits < 32) { # not binary property
-	my $top = 0;
-	while ($list =~ /^([0-9a-fA-F]+)(?:[\t]([0-9a-fA-F]+)?)(?:[ \t]([0-9a-fA-F]+))?/mg) {
-	    my $min = CORE::hex $1;
-	    my $max = defined $2 ? CORE::hex $2 : $min;
-	    my $val = defined $3 ? CORE::hex $3 : 0;
-	    $val += $max - $min if defined $3;
-	    $top = $val if $val > $top;
-	}
-	my $topbits =
-	    $top > 0xffff ? 32 :
-	    $top > 0xff ? 16 : 8;
-	$bits = $topbits if $bits < $topbits;
-    }
+        if ($minbits != 1 && $minbits < 32) { # not binary property
+            my $top = 0;
+            while ($list =~ /^([0-9a-fA-F]+)(?:[\t]([0-9a-fA-F]+)?)(?:[ \t]([0-9a-fA-F]+))?/mg) {
+                my $min = CORE::hex $1;
+                my $max = defined $2 ? CORE::hex $2 : $min;
+                my $val = defined $3 ? CORE::hex $3 : 0;
+                $val += $max - $min if defined $3;
+                $top = $val if $val > $top;
+            }
+            my $topbits =
+                $top > 0xffff ? 32 :
+                $top > 0xff ? 16 : 8;
+            $bits = $topbits if $bits < $topbits;
+        }
 
-    my @extras;
-    for my $x ($extras) {
-	pos $x = 0;
-	while ($x =~ /^([^0-9a-fA-F\n])(.*)/mg) {
-	    my $char = $1;
-	    my $name = $2;
-	    print STDERR "$1 => $2\n" if DEBUG;
-	    if ($char =~ /[-+!&]/) {
-		my ($c,$t) = split(/::/, $name, 2);	# bogus use of ::, really
-		my $subobj;
-		if ($c eq 'utf8') {
-		    $subobj = utf8->SWASHNEW($t, "", $minbits, 0);
-		}
-		elsif (exists &$name) {
-		    $subobj = utf8->SWASHNEW($name, "", $minbits, 0);
-		}
-		elsif ($c =~ /^([0-9a-fA-F]+)/) {
-		    $subobj = utf8->SWASHNEW("", $c, $minbits, 0);
-		}
-		return $subobj unless ref $subobj;
-		push @extras, $name => $subobj;
-		$bits = $subobj->{BITS} if $bits < $subobj->{BITS};
-	    }
-	}
-    }
+        my @extras;
+        if ($extras) {
+            for my $x ($extras) {
+                pos $x = 0;
+                while ($x =~ /^([^0-9a-fA-F\n])(.*)/mg) {
+                    my $char = $1;
+                    my $name = $2;
+                    print STDERR __LINE__, ": $1 => $2\n" if DEBUG;
+                    if ($char =~ /[-+!&]/) {
+                        my ($c,$t) = split(/::/, $name, 2);	# bogus use of ::, really
+                        my $subobj;
+                        if ($c eq 'utf8') {
+                            $subobj = utf8->SWASHNEW($t, "", $minbits, 0);
+                        }
+                        elsif (exists &$name) {
+                            $subobj = utf8->SWASHNEW($name, "", $minbits, 0);
+                        }
+                        elsif ($c =~ /^([0-9a-fA-F]+)/) {
+                            $subobj = utf8->SWASHNEW("", $c, $minbits, 0);
+                        }
+                        return $subobj unless ref $subobj;
+                        push @extras, $name => $subobj;
+                        $bits = $subobj->{BITS} if $bits < $subobj->{BITS};
+                    }
+                }
+            }
+        }
 
-    print STDERR "CLASS = $class, TYPE => $type, BITS => $bits, NONE => $none\nEXTRAS =>\n$extras\nLIST =>\n$list\n" if DEBUG;
+        if (DEBUG) {
+            print STDERR __LINE__, ": CLASS = $class, TYPE => $type, BITS => $bits, NONE => $none";
+            print STDERR "\nLIST =>\n$list" if defined $list;
+            print STDERR "\nEXTRAS =>\n$extras" if defined $extras;
+            print STDERR "\n";
+        }
 
-    my $SWASH = bless {
-	TYPE => $type,
-	BITS => $bits,
-	EXTRAS => $extras,
-	LIST => $list,
-	NONE => $none,
-	@extras,
-    } => $class;
+        my $SWASH = bless {
+            TYPE => $type,
+            BITS => $bits,
+            EXTRAS => $extras,
+            LIST => $list,
+            NONE => $none,
+            @extras,
+        } => $class;
+
+        if ($file) {
+            $Cache{$class, $file} = $SWASH;
+        }
 
-    if ($file) {
-        $Cache{$class, $file} = $SWASH;
+        return $SWASH;
     }
-
-    return $SWASH;
 }
 
 # Now SWASHGET is recasted into a C function S_swash_get (see utf8.c).