path: root/doc/groff.texi

\input texinfo

@c
@c Please convert this manual with `texi2dvi -e groff.texi' due to
@c problems in texinfo regarding expansion of user-defined macros.
@c
@c You need texinfo 5.0 or newer to format this document!
@c

@c %**start of header (This is for running Texinfo on a region.)
@setfilename groff.info
@settitle The GNU Troff Manual
@setchapternewpage odd
@footnotestyle separate
@c %**end of header (This is for running Texinfo on a region.)

@documentlanguage en
@documentencoding ISO-8859-1


@smallbook

@finalout


@copying
This manual documents GNU @code{troff} version 1.23.0.

Copyright @copyright{} 1994--2023 Free Software Foundation, Inc.

@quotation
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.3 or
any later version published by the Free Software Foundation; with no
Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts. A
copy of the license is included in the section entitled ``GNU Free
Documentation License''.
@end quotation
@end copying


@c We use the following indices:
@c
@c   cindex: concepts
@c   rqindex: requests
@c   esindex: escape sequences
@c   vindex: registers
@c   kindex: commands in font files
@c   pindex: programs and files
@c   tindex: environment variables
@c   maindex: macros
@c   stindex: strings
@c   opindex: operators
@c
@c tindex and cindex are merged.

@defcodeindex rq
@defcodeindex es
@defcodeindex ma
@defcodeindex st
@defcodeindex op
@syncodeindex tp cp


@c To avoid uppercasing in @deffn while converting to info, we define
@c our special @Var{}.

@macro Var{arg}
@r{@slanted{\arg\}}
@end macro


@c To assure correct HTML translation, some ugly hacks are necessary.
@c While processing a @def... request, the HTML translator looks at the
@c next line to decide whether to start indentation, and if the line
@c starts with @def... (e.g. @deffnx), indentation is started.  We must
@c therefore ensure that a @def... is seen, during macro expansion.
@c
@c The following macros have to be used:
@c
@c One item:
@c
@c   @Def...
@c
@c Two items:
@c
@c   @Def...List
@c   @Def...ListEnd
@c
@c More than two:
@c
@c   @Def...List
@c   @Def...Item
@c   @Def...Item
@c   ...
@c   @Def...ListEnd
@c
@c The definition block must end with
@c
@c   @endDef...
@c
@c The above is valid for texinfo 4.0f and above.
@c
@c By default, only the first item generates an index entry.  To
@c override this, use a variant with a trailing `x' (like
@c `@DefmacItemx').


@c a dummy macro to assure the `@def...'

@macro defdummy
@c
@end macro


@c definition of requests

@macro Defreq{name, arg}
@deffn Request @t{.\name\} \arg\
@rqindex \name\
@c
@end macro

@macro DefreqList{name, arg}
@deffn Request @t{.\name\} \arg\
@defdummy
@rqindex \name\
@c
@end macro

@macro DefreqItem{name, arg}
@deffnx Request @t{.\name\} \arg\
@defdummy
@c
@end macro

@macro DefreqItemx{name, arg}
@deffnx Request @t{.\name\} \arg\
@defdummy
@rqindex \name\
@c
@end macro

@macro DefreqListEnd{name, arg}
@deffnx Request @t{.\name\} \arg\
@c
@end macro

@macro DefreqListEndx{name, arg}
@deffnx Request @t{.\name\} \arg\
@rqindex \name\
@c
@end macro

@macro endDefreq
@end deffn
@end macro


@c definition of escape sequences

@macro Defesc{name, delimI, arg, delimII}
@deffn Escape@tie{}sequence @t{\name\\delimI\}@Var{\arg\}@t{\delimII\}
@esindex \name\
@c
@end macro

@macro DefescList{name, delimI, arg, delimII}
@deffn Escape@tie{}sequence @t{\name\\delimI\}@Var{\arg\}@t{\delimII\}
@defdummy
@esindex \name\
@c
@end macro

@macro DefescItem{name, delimI, arg, delimII}
@deffnx Escape@tie{}sequence @t{\name\\delimI\}@Var{\arg\}@t{\delimII\}
@defdummy
@c
@end macro

@macro DefescItemx{name, delimI, arg, delimII}
@deffnx Escape@tie{}sequence @t{\name\\delimI\}@Var{\arg\}@t{\delimII\}
@defdummy
@esindex \name\
@c
@end macro

@macro DefescListEnd{name, delimI, arg, delimII}
@deffnx Escape@tie{}sequence @t{\name\\delimI\}@Var{\arg\}@t{\delimII\}
@c
@end macro

@macro DefescListEndx{name, delimI, arg, delimII}
@deffnx Escape@tie{}sequence @t{\name\\delimI\}@Var{\arg\}@t{\delimII\}
@esindex \name\
@c
@end macro

@macro endDefesc
@end deffn
@end macro


@c definition of registers (built in to GNU troff)

@macro Defreg{name}
@deffn Register @t{\\n[\name\]}
@vindex \name\
@c
@end macro

@macro DefregList{name}
@deffn Register @t{\\n[\name\]}
@defdummy
@vindex \name\
@c
@end macro

@macro DefregItem{name}
@deffnx Register @t{\\n[\name\]}
@defdummy
@c
@end macro

@macro DefregItemx{name}
@deffnx Register @t{\\n[\name\]}
@defdummy
@vindex \name\
@c
@end macro

@macro DefregListEnd{name}
@deffnx Register @t{\\n[\name\]}
@c
@end macro

@macro DefregListEndx{name}
@deffnx Register @t{\\n[\name\]}
@vindex \name\
@c
@end macro

@macro endDefreg
@end deffn
@end macro


@c string definitions (built in to GNU troff)

@macro Defstr{name}
@deffn String @t{\\*[\name\]}
@stindex \name\
@c
@end macro

@macro DefstrList{name}
@deffn String @t{\\*[\name\]}
@defdummy
@stindex \name\
@c
@end macro

@macro DefstrItem{name}
@deffnx String @t{\\*[\name\]}
@defdummy
@c
@end macro

@macro DefstrItemx{name}
@deffnx String @t{\\*[\name\]}
@defdummy
@stindex \name\
@c
@end macro

@macro DefstrListEnd{name}
@deffnx String @t{\\*[\name\]}
@c
@end macro

@macro DefstrListEndx{name}
@deffnx String @t{\\*[\name\]}
@stindex \name\
@c
@end macro

@macro endDefstr
@end deffn
@end macro


@c register definitions specific to macro packages, preprocessors, ...

@macro Defmpreg{name, package}
@deffn Register @t{\\n[\name\]}
@vindex \name\ @r{[}\package\@r{]}
@c
@end macro

@macro DefmpregList{name, package}
@deffn Register @t{\\n[\name\]}
@defdummy
@vindex \name\ @r{[}\package\@r{]}
@c
@end macro

@macro DefmpregItem{name, package}
@deffnx Register @t{\\n[\name\]}
@defdummy
@c
@end macro

@macro DefmpregItemx{name, package}
@deffnx Register @t{\\n[\name\]}
@defdummy
@vindex \name\ @r{[}\package\@r{]}
@c
@end macro

@macro DefmpregListEnd{name, package}
@deffnx Register @t{\\n[\name\]}
@c
@end macro

@macro DefmpregListEndx{name, package}
@deffnx Register @t{\\n[\name\]}
@vindex \name\ @r{[}\package\@r{]}
@c
@end macro

@macro endDefmpreg
@end deffn
@end macro


@c definition of macros

@macro Defmac{name, arg, package}
@defmac @t{.\name\} \arg\
@maindex \name\ @r{[}\package\@r{]}
@c
@end macro

@macro DefmacList{name, arg, package}
@defmac @t{.\name\} \arg\
@defdummy
@maindex \name\ @r{[}\package\@r{]}
@c
@end macro

@macro DefmacItem{name, arg, package}
@defmacx @t{.\name\} \arg\
@defdummy
@c
@end macro

@macro DefmacItemx{name, arg, package}
@defmacx @t{.\name\} \arg\
@defdummy
@maindex \name\ @r{[}\package\@r{]}
@c
@end macro

@macro DefmacListEnd{name, arg, package}
@defmacx @t{.\name\} \arg\
@c
@end macro

@macro DefmacListEndx{name, arg, package}
@defmacx @t{.\name\} \arg\
@maindex \name\ @r{[}\package\@r{]}
@c
@end macro

@macro endDefmac
@end defmac
@end macro


@c string definitions specific to macro packages, preprocessors, ...

@macro Defmpstr{name, package}
@deffn String @t{\\*[\name\]}
@stindex \name\ @r{[}\package\@r{]}
@c
@end macro

@macro DefmpstrList{name, package}
@deffn String @t{\\*[\name\]}
@defdummy
@stindex \name\ @r{[}\package\@r{]}
@c
@end macro

@macro DefmpstrItem{name, package}
@deffnx String @t{\\*[\name\]}
@defdummy
@c
@end macro

@macro DefmpstrItemx{name, package}
@deffnx String @t{\\*[\name\]}
@defdummy
@stindex \name\ @r{[}\package\@r{]}
@c
@end macro

@macro DefmpstrListEnd{name, package}
@deffnx String @t{\\*[\name\]}
@c
@end macro

@macro DefmpstrListEndx{name, package}
@deffnx String @t{\\*[\name\]}
@stindex \name\ @r{[}\package\@r{]}
@c
@end macro

@macro endDefmpstr
@end deffn
@end macro


@c our example macros

@macro Example
@example
@group
@end macro

@macro endExample
@end group
@end example
@end macro

@macro CartoucheExample
@cartouche
@example
@end macro

@macro endCartoucheExample
@end example
@end cartouche
@end macro


@c Render text with angle brackets around it, as in <text>.

@macro angles{text}
@guilsinglleft{}@r{\text\}@guilsinglright{}
@end macro


@c Note: We say `Roman numerals' but `roman font'.


@dircategory Typesetting
@direntry
* Groff: (groff).               The GNU roff document formatting system.
@end direntry


@titlepage
@title groff
@subtitle The GNU implementation of @code{troff}
@subtitle Edition 1.23.0
@subtitle May 2023
@author Trent@tie{}A.@: Fisher
@author Werner Lemberg
@author G.@tie{}Branden Robinson

@page
@vskip 0pt plus 1filll
@insertcopying
@end titlepage

@contents

@ifnottex
@node Top, Introduction, (dir), (dir)
@top GNU @code{troff}
@end ifnottex

@menu
* Introduction::
* Invoking groff::
* Tutorial for Macro Users::
* Major Macro Packages::
* GNU troff Reference::
* File Formats::
* Copying This Manual::
* Request Index::
* Escape Sequence Index::
* Operator Index::
* Register Index::
* Macro Index::
* String Index::
* File Keyword Index::
* Program and File Index::
* Concept Index::
@end menu

@ifnottex
@insertcopying
@end ifnottex



@c =====================================================================
@c =====================================================================

@codequotebacktick on
@codequoteundirected on

@node Introduction, Invoking groff, Top, Top
@chapter Introduction
@cindex introduction

GNU @code{roff} (or @code{groff}) is a programming system for
typesetting documents.  It is highly flexible and has been used
extensively for over thirty years.

@menu
* Background::
* What Is @code{groff}?::
* @code{groff} Capabilities::
* Macro Package Intro::
* Preprocessor Intro::
* Output Device Intro::
* Conventions Used in This Manual::
* Installation::
* Credits::
@end menu


@c =====================================================================

@node Background, What Is @code{groff}?, Introduction, Introduction
@section Background
@cindex background

M.@: Douglas McIlroy, formerly of AT&T Bell Laboratories and present at
the creation of the Unix operating system, offers an authoritative
historical summary.

@quotation
The prime reason for Unix was the desire of Ken [Thompson], Dennis
[Ritchie], and Joe Ossanna to have a pleasant environment for software
development.  The fig leaf that got the nod from @dots{}
management was that an early use would be to develop a ``stand-alone''
word-processing system for use in typing pools and secretarial offices.
Perhaps they had in mind ``dedicated'', as distinct from
``stand-alone''; that's what eventuated in various cases, most notably
in the legal/patent department and in the AT&T CEO's office.

Both those systems were targets of opportunity, not foreseen from the
start.  When Unix was up and running on the PDP-11, Joe got wind of
the legal department having installed a commercial word processor.
He went to pitch Unix as an alternative and clinched a trial by
promising to make @code{roff} able to number lines by tomorrow in order
to fulfill a patent-office requirement that the commercial system did
not support.

Modems were installed so legal-department secretaries could try the
Research machine.  They liked it and Joe's superb customer service.
Soon the legal department got a system of their own.  Joe went on to
create @code{nroff} and @code{troff}.  Document preparation became a
widespread use of Unix, but no stand-alone word-processing system was
ever undertaken.
@end quotation
@c https://minnie.tuhs.org/pipermail/tuhs/2022-March/025535.html

A history relating @code{groff} to its predecessors @code{roff},
@code{nroff}, and @code{troff} is available in the @cite{roff@r{(7)}}
man page.


@c =====================================================================

@node What Is @code{groff}?, @code{groff} Capabilities, Introduction, Introduction
@section What Is @code{groff}?
@cindex what is @code{groff}?
@cindex @code{groff}---what is it?

@c BEGIN Keep parallel with groff(1), section "Description" (after the
@c first sentence).
@c This language is slightly expanded from that in the "ANNOUNCE" file
@c and on the groff home page.
@code{groff} (GNU @code{roff}) is a typesetting system that reads plain
text input files that include formatting commands to produce output in
PostScript, PDF, HTML, DVI, or other formats, or for display to a
terminal.  Formatting commands can be low-level typesetting primitives,
macros from a supplied package, or user-defined macros.  All three
approaches can be combined.

A reimplementation and extension of the typesetter from @acronym{AT&T}
Unix, @code{groff} is present on most @acronym{POSIX} systems owing to
its long association with Unix manuals (including man pages).  It and
its predecessor are notable for their production of several best-selling
software engineering texts.  @code{groff} is capable of producing
typographically sophisticated documents while consuming minimal system
resources.
@c END Keep parallel with groff(1), section "Description" (after the
@c first sentence).


@c =====================================================================

@node @code{groff} Capabilities, Macro Package Intro, What Is @code{groff}?, Introduction
@section @code{groff} Capabilities
@cindex @code{groff} capabilities
@cindex capabilities of @code{groff}

So what exactly is @code{groff} capable of doing?  @code{groff} provides
a wide range of low-level text formatting operations.  Using these, it
is possible to perform a wide range of formatting tasks, such as
footnotes, table of contents, multiple columns, etc.  Here's a list of
the most important operations supported by @code{groff}:

@itemize @bullet
@item
text filling, adjustment, and centering

@item
hyphenation

@item
page control

@item
font and glyph size control

@item
vertical spacing (e.g., double-spacing)

@item
line length and indenting

@item
macros, strings, diversions, and traps

@item
registers

@item
tabs, leaders, and fields

@item
input and output conventions and character translation

@item
overstrike, bracket, line drawing, and zero-width functions

@item
local horizontal and vertical motions and the width function

@item
three-part titles

@item
output line numbering

@item
conditional acceptance of input

@item
environment switching

@item
insertions from the standard input

@item
input/output file switching

@item
output and error messages
@end itemize


@c =====================================================================

@node Macro Package Intro, Preprocessor Intro, @code{groff} Capabilities, Introduction
@section Macro Packages
@cindex macro package, introduction
@cindex package, macro, introduction

Since @code{groff} provides such low-level facilities, it can be quite
difficult to use by itself.  However, @code{groff} provides a
@dfn{macro} facility to specify how certain routine operations, such as
starting paragraphs, or printing headers and footers, should be done.
These macros can be collected together into a @dfn{macro package}.
There are a number of macro packages available; the most common (and the
ones described in this manual) are @file{man}, @file{mdoc}, @file{me},
@file{ms}, and @file{mm}.


@c =====================================================================

@node Preprocessor Intro, Output Device Intro, Macro Package Intro, Introduction
@section Preprocessors
@cindex preprocessors

Although @code{groff} provides most functions needed to format a
document, some operations would be unwieldy (e.g., to draw pictures).
Therefore, programs called @dfn{preprocessors} were written that
understand their own language and produce the necessary @code{groff}
operations.  These preprocessors are able to differentiate their own
input from the rest of the document via markers.

To use a preprocessor, Unix pipes are used to feed the output from the
preprocessor into @code{groff}.  Any number of preprocessors may be used
on a given document; in this case, the preprocessors are linked together
into one pipeline.  However, with @code{groff}, the user does not need
to construct the pipe, but only tell @code{groff} what preprocessors to
use.

@code{groff} currently has preprocessors for producing tables
(@code{tbl}), typesetting equations (@code{eqn}), drawing pictures
(@code{pic} and @code{grn}), processing bibliographies
(@code{refer}), and drawing chemical structures (@code{chem}).  An
associated program that is useful when dealing with preprocessors is
@code{soelim}.

A free implementation of @code{grap}, a preprocessor for drawing graphs,
can be obtained as an extra package; @code{groff} can use @code{grap}
also.

Unique to @code{groff} is the @code{preconv} preprocessor that enables
@code{groff} to handle documents in various input encodings.

Other preprocessors exist, but, unfortunately, no free implementations
are available.  Among them is a preprocessor for drawing mathematical
pictures (@code{ideal}).


@c =====================================================================

@node Output Device Intro, Installation, Preprocessor Intro, Introduction
@section Output Devices
@cindex postprocessors
@cindex output devices
@cindex devices for output

@code{groff} produces device-independent code that may be fed into a
postprocessor to produce output for a particular device.  Currently,
@code{groff} has postprocessors for PostScript devices, character
terminals, X11 (for previewing), DVI, HP LaserJet@tie{}4 and Canon LBP
printers (which use @acronym{CaPSL}), @acronym{HTML}, @acronym{XHTML},
and @acronym{PDF}.


@c =====================================================================

@node Installation, Conventions Used in This Manual, Output Device Intro, Introduction
@section Installation
@cindex installation

Installation procedures are documented by the files @file{INSTALL},
@file{INSTALL.extra}, and @file{INSTALL.REPO} in the @code{groff} source
distribution.


@c =====================================================================

@node Conventions Used in This Manual, Credits, Installation, Introduction
@section Conventions Used in This Manual

We apply the term ``groff'' to the language documented here, the GNU
implementation of the overall system, the project that develops that
system, and the command of that name.  In the first sense, @code{groff}
is an extended dialect of the @code{roff} language, for which many
similar implementations exist.

The @code{roff} language features several major categories for which
many items are predefined.  Presentations of these items feature the
form in which the item is most commonly used on the left, and, aligned
to the right margin, the name of the category in brackets.

@deffn Register \n[example]
The register @samp{example} is one that that @code{groff} @emph{doesn't}
predefine.  You can create it yourself, though; see @ref{Setting
Registers}.
@end deffn

To make this document useful as a reference and not merely amiable
bedtime reading, we tend to present these syntax items in exhaustive
detail when they arise.  References to topics discussed later in the
text are frequent; skip material you don't understand yet.

We use Texinfo's ``result'' (@result{}) and @error{} notations to
present output written to the standard output and standard error
streams, respectively.  Diagnostic messages from the GNU @code{troff}
formatter and other programs are examples of the latter, but the
formatter can also be directed to write user-specified messages to the
standard error stream.  The notation then serves to identify the
output stream and does not necessarily mean that an error has
occurred.@footnote{Unix and related operating systems distinguish
standard output and standard error streams @emph{because} of
@code{troff}:@:
@uref{https://minnie.tuhs.org/pipermail/tuhs/2013-December/006113.html}.}

@Example
$ echo "Twelve o'clock and" | groff -Tascii | sed '/^$/d'
    @result{} Twelve o'clock and
$ echo '.tm all is well.' | groff > /dev/null
    @error{} all is well.
@endExample

Sometimes we use @result{} somewhat abstractly to represent formatted
text that you will need to use a PostScript or PDF viewer program (or a
printer) to observe.  While arguably an abuse of notation, we think this
preferable to requiring the reader to understand the syntax of these
page description languages.

We also present diagnostic messages in an abbreviated form, often
omitting the name of the program issuing them, the input file name, and
line number or other positional information when such data do not serve
to illuminate the topic under discussion.

Most examples are of @code{roff} language input that would be placed in
a text file.  Occasionally, we start an example with a @samp{$}
character to indicate a shell prompt, as seen above.

You are encouraged to try the examples yourself, and to alter them to
better learn @code{groff}'s behavior.  Our examples frequently need to
direct the formatter to set a line length (with @samp{.ll}) that will
fit within the page margins of this manual.  We mention this so that you
know why it is there before we discuss the @code{ll} request
formally.@footnote{@xref{Line Layout}.}


@c =====================================================================

@node Credits,  , Conventions Used in This Manual, Introduction
@section Credits
@cindex credits

Large portions of this manual were taken from existing documents---most
notably, the manual pages for the @code{groff} package by James Clark,
and Eric Allman's papers on the @file{me} macro package.  Larry Kollar
contributed much of the material on the @file{ms} macro package.


@c =====================================================================
@c =====================================================================

@codequotebacktick on
@codequoteundirected on

@node Invoking groff, Tutorial for Macro Users, Introduction, Top
@chapter Invoking @code{groff}
@cindex invoking @code{groff}
@cindex @code{groff} invocation

This chapter focuses on how to invoke the @code{groff} front end.  This
front end takes care of the details of constructing the pipeline among
the preprocessors, @code{gtroff} and the postprocessor.

It has become a tradition that GNU programs get the prefix @samp{g} to
distinguish them from their original counterparts provided by the host
(@pxref{Environment}).  Thus, for example, @code{geqn} is GNU
@code{eqn}.  On operating systems like GNU/Linux or the Hurd, which
don't contain proprietary versions of @code{troff}, and on
MS-DOS/MS-Windows, where @code{troff} and associated programs are not
available at all, this prefix is omitted since GNU @code{troff} is the
only incarnation of @code{troff} used.  Exception: @samp{groff} is never
replaced by @samp{roff}.

In this document, we consequently say @samp{gtroff} when talking about
the GNU @code{troff} program.  @c XXX: Not for much longer... -- GBR
All other implementations of @code{troff} are called @acronym{AT&T}
@code{troff}, which is the common origin of almost all @code{troff}
implementations@footnote{Besides @code{groff}, @code{neatroff} is an
exception.} (with more or less compatible changes).  Similarly, we say
@samp{gpic}, @samp{geqn}, and so on.

@menu
* Groff Options::
* Environment::
* Macro Directories::
* Font Directories::
* Paper Format::
* Invocation Examples::
@end menu


@c =====================================================================

@node Groff Options, Environment, Invoking groff, Invoking groff
@section Options
@cindex options

@pindex groff
@pindex gtroff
@pindex gpic
@pindex geqn
@pindex ggrn
@pindex grap
@pindex gtbl
@pindex gchem
@pindex grefer
@pindex gsoelim
@pindex preconv
@code{groff} normally runs the @code{gtroff} program and a
postprocessor appropriate for the selected device.  The default device
is @samp{ps} (but it can be changed when @code{groff} is configured and
built).  It can optionally preprocess with any of @code{gpic},
@code{geqn}, @code{gtbl}, @code{ggrn}, @code{grap}, @code{gchem},
@code{grefer}, @code{gsoelim}, or @code{preconv}.

This section only documents options to the @code{groff} front end.  Many
of the arguments to @code{groff} are passed on to @code{gtroff},
therefore those are also included.  Arguments to preprocessors and
output drivers can be found in the man pages @cite{gpic@r{(1)}},
@cite{geqn@r{(1)}}, @cite{gtbl@r{(1)}}, @cite{ggrn@r{(1)}},
@cite{grefer@r{(1)}}, @cite{gchem@r{(1)}}, @cite{gsoelim@r{(1)}},
@cite{preconv@r{(1)}}, @cite{grotty@r{(1)}}, @cite{grops@r{(1)}},
@cite{gropdf@r{(1)}}, @cite{grohtml@r{(1)}}, @cite{grodvi@r{(1)}},
@cite{grolj4@r{(1)}}, @cite{grolbp@r{(1)}}, and @cite{gxditview@r{(1)}}.

The command-line format for @code{groff} is:

@Example
groff [ -abceghijklpstvzCEGNRSUVXZ ] [ -d@var{cs} ] [ -D@var{arg} ]
      [ -f@var{fam} ] [ -F@var{dir} ] [ -I@var{dir} ] [ -K@var{arg} ]
      [ -L@var{arg} ] [ -m@var{name} ] [ -M@var{dir} ] [ -n@var{num} ]
      [ -o@var{list} ] [ -P@var{arg} ] [ -r@var{cn} ] [ -T@var{dev} ]
      [ -w@var{name} ] [ -W@var{name} ] [ @var{files}@dots{} ]
@endExample

The command-line format for @code{gtroff} is as follows.

@Example
gtroff [ -abcivzCERU ] [ -d@var{cs} ] [ -f@var{fam} ] [ -F@var{dir} ]
       [ -m@var{name} ] [ -M@var{dir} ] [ -n@var{num} ] [ -o@var{list} ]
       [ -r@var{cn} ] [ -T@var{name} ] [ -w@var{name} ] [ -W@var{name} ]
       [ @var{files}@dots{} ]
@endExample

@noindent
Obviously, many of the options to @code{groff} are actually passed on to
@code{gtroff}.

Options without an argument can be grouped behind a
single@tie{}@option{-}.  A filename of@tie{}@file{-} denotes the
standard input.  Whitespace is permitted between an option and its
argument.

The @code{grog} command can be used to guess the correct @code{groff}
command to format a file.  See its man page @cite{grog@r{(1)}}; type
@samp{man grog} at the command line to view it.

@command{groff}'s command-line options are as follows.

@cindex command-line options
@table @samp
@item -a
@cindex plain text approximation output register (@code{.A})
Generate a plain text approximation of the typeset output.  The
read-only register @code{.A} is set to@tie{}1.  @xref{Built-in
Registers}.  This option produces a sort of abstract preview of the
formatted output.

@itemize @bullet
@item
Page breaks are marked by a phrase in angle brackets; for example,
@samp{<beginning of page>}.

@item
Lines are broken where they would be in the formatted output.

@item
A horizontal motion of any size is represented as one space.  Adjacent
horizontal motions are not combined.  Inter-sentence space nodes (those
arising from the second argument to the @code{ss} request) are not
represented.

@item
Vertical motions are not represented.

@item
Special characters are rendered in angle brackets; for example, the
default soft hyphen character appears as @samp{<hy>}.
@end itemize

The above description should not be considered a specification; the
details of @option{-a} output are subject to change.

@item -b
Write a backtrace reporting the state of @command{gtroff}'s input parser
to the standard error stream with each diagnostic message.  The line
numbers given in the backtrace might not always be correct, because
@command{gtroff}'s idea of line numbers can be confused by requests that
append to
@c XXX: strings or (??? strings never contain newlines)
macros.

@item -c
Start with color output disabled.

@item -C
Enable AT&T @command{troff} compatibility mode; implies @option{-c}.
@xref{Implementation Differences}, for the list of incompatibilities
between @command{groff} and @acronym{AT&T} @command{troff}.

@item -d@var{c}@var{text}
@itemx -d@var{string}=@var{text}
Define @code{roff} string @var{c} or @var{string} as@tie{}@var{t} or
@var{text}.  @var{c}@tie{}must be one character; @var{string} can be
of arbitrary length.  Such string assignments happen before any macro
file is loaded, including the startup file.  Due to @code{getopt_long}
limitations, @var{c}@tie{}cannot be, and @var{string} cannot contain, an
equals sign, even though that is a valid character in a @code{roff}
identifier.

@item -D@var{enc}
Set fallback input encoding used by @command{preconv} to @var{enc};
implies @option{-k}.

@item -e
Run @command{geqn} preprocessor.

@item -E
Inhibit @command{gtroff} error messages.  This option does @emph{not}
suppress messages sent to the standard error stream by documents or
macro packages using @code{tm} or related requests.

@item -f@var{fam}
Use @var{fam} as the default font family.  @xref{Font Families}.

@item -F@var{dir}
Search in directory @file{@var{dir}} for the selected output device's
directory of device and font description files.  See the description of
@env{GROFF_FONT_PATH} in @ref{Environment} below for the default search
locations and ordering.

@item -g
Run @command{ggrn} preprocessor.

@item -G
Run @command{grap} preprocessor; implies @option{-p}.

@item -h
Display a usage message and exit.

@item -i
Read the standard input after all the named input files have been
processed.

@item -I@var{dir}
Search the directory @var{dir} for files named in several contexts;
implies @option{-g} and @option{-s}.

@itemize
@item
@command{gsoelim} replaces @code{so} requests with the contents of their
file name arguments.

@item
@command{gtroff} searches for files named as operands in its command
line and as arguments to @code{psbb}, @code{so}, and @code{soquiet}
requests.

@item
Output drivers may search for files; for instance, @command{grops} looks
for files named in @samp{\X'ps: import @r{@dots{}}'}, @samp{\X'ps: file
@r{@dots{}}'}, and @samp{\X'pdf: pdfpic @r{@dots{}}'} device control
escape sequences.
@end itemize

This option may be specified more than once; the directories are
searched in the order specified.  If you want to search the current
directory before others, add @samp{-I .} at the desired place.  The
current working directory is otherwise searched last.  @option{-I} works
similarly to, and is named for, the ``include'' option of Unix C
compilers.

@option{-I} options are passed to @command{gsoelim}, @command{gtroff},
and output drivers; with the flag letter changed to @option{-M}, they
are also passed to @command{ggrn}.

@item -j
Run @command{gchem} preprocessor.  Implies @option{-p}.

@item -k
Run @command{preconv} preprocessor.  Refer to its man page for its
behavior if neither of @command{groff}'s @option{-K} or @option{-D}
options is also specified.

@item -K@var{enc}
Set input encoding used by @command{preconv} to @var{enc}; implies
@option{-k}.

@item -l
Send the output to a spooler for printing.  The @code{print} directive
in the device description file specifies the default command to be used;
see @ref{Device and Font Description Files}.
@c XXX: This document is not parameterized in configuration variables.
@c If no such directive is present for the output device,
@c .ie '@PSPRINT@'' \{\
@c this option is ignored.
@c .\}
@c .el \{\
@c output is piped to
@c .MR @PSPRINT@ 1 .
@c .\}
See options @option{-L} and @option{-X}.

@item -L@var{arg}
Pass @var{arg} to the print spooler program.  If multiple @var{arg}s are
required, pass each with a separate @option{-L} option.  @command{groff}
does not prefix an option dash to @var{arg} before passing it to the
spooler program.

@item -m@var{name}
Process the file @file{@var{name}.tmac} prior to any input files.
If not found, @file{tmac.@var{name}} is attempted.  @var{name}
(in both arrangements) is presumed to be a macro file; see the
description of @env{GROFF_TMAC_PATH} in @ref{Environment} below for the
default search locations and ordering.  This option and its argument are
also passed to @command{geqn}, @command{grap}, and @command{ggrn}.

@item -M@var{dir}
Search directory @file{@var{dir}} for macro files;  see the description
of @env{GROFF_TMAC_PATH} in @ref{Environment} below for the default
search locations and ordering.  This option and its argument are also
passed to @command{geqn}, @command{grap}, and @command{ggrn}.

@item -n@var{num}
Number the first page @var{num}.

@item -N
Prohibit newlines between @code{eqn} delimiters:@: pass @option{-N} to
@command{geqn}.

@item -o@var{list}
@cindex print current page register (@code{.P})
Output only pages in @var{list}, which is a comma-separated list of page
ranges; @samp{@var{n}} means page@tie{}@var{n}, @samp{@var{m}-@var{n}}
means every page between @var{m} and@tie{}@var{n}, @samp{-@var{n}} means
every page up to@tie{}@var{n}, @samp{@var{n}-} means every page from
@var{n}@tie{}on.  @command{gtroff} stops processing and exits after
formatting the last page enumerated in @var{list}.

@item -p
Run @command{gpic} preprocessor.

@item -P@var{arg}
Pass @var{arg} to the postprocessor.  If multiple @var{arg}s are
required, pass each with a separate @option{-P} option.  @command{groff}
does not prefix an option dash to @var{arg} before passing it to the
postprocessor.

@item -r@var{c}@var{numeric-expression}
@itemx -r@var{register}=@var{expr}
Set @code{roff} register@tie{}@var{c} or @var{register} to the value
@var{numeric-expression} (@pxref{Numeric Expressions}).
@var{c}@tie{}must be one character; @var{register} can be of arbitrary
length.  Such register assignments happen before any macro file is
loaded, including the startup file.  Due to @code{getopt_long}
limitations, @var{c}@tie{}cannot be, and @var{register} cannot contain,
an equals sign, even though that is a valid character in a @code{roff}
identifier.

@item -R
Run @command{grefer} preprocessor.  No mechanism is provided for passing
arguments to @command{grefer} because most @command{grefer} options have
equivalent language elements that can be specified within the document.

@pindex troffrc
@pindex troffrc-end
@command{gtroff} also accepts a @option{-R} option, which is not
accessible via @command{groff}.  This option prevents the loading of the
@file{troffrc} and @file{troffrc-end} files.

@item -s
Run @command{gsoelim} preprocessor.

@item -S
@cindex @code{open} request, and safer mode
@cindex @code{opena} request, and safer mode
@cindex @code{pso} request, and safer mode
@cindex @code{sy} request, and safer mode
@cindex @code{pi} request, and safer mode
@cindex safer mode
@cindex mode, safer
Operate in ``safer'' mode; see @option{-U} below for its opposite.  For
security reasons, safer mode is enabled by default.

@item -t
Run @command{gtbl} preprocessor.

@item -T@var{dev}
Direct @command{gtroff} to format the input for the output device
@var{dev}.  @command{groff} then calls an output driver to convert
@command{gtroff}'s output to a form appropriate for @var{dev}.  The
following output devices are available.

@table @code
@item ps
For PostScript printers and previewers.

@item pdf
For @acronym{PDF} viewers or printers.

@item dvi
For @TeX{} DVI format.

@item X75
For a 75@dmn{dpi} X11 previewer.

@item X75-12
For a 75@dmn{dpi} X11 previewer with a 12-point base font in the
document.

@item X100
For a 100@dmn{dpi} X11 previewer.

@item X100-12
For a 100@dmn{dpi} X11 previewer with a 12-point base font in the
document.

@item ascii
@cindex encoding, output, @acronym{ASCII}
@cindex encoding, output, ISO@tie{}646
@cindex @acronym{ASCII}, output encoding
@cindex ISO@tie{}646, output encoding
@cindex output encoding, @acronym{ASCII}
@cindex output encoding, ISO@tie{}646
For typewriter-like devices using the (7-bit) @acronym{ASCII}
(ISO@tie{}646) character set.

@item latin1
@cindex encoding, output, @w{Latin-1} (ISO @w{8859-1})
@cindex @w{Latin-1} (ISO @w{8859-1}), output encoding
@cindex ISO @w{8859-1} (@w{Latin-1}), output encoding
@cindex output encoding, @w{Latin-1} (ISO @w{8859-1})
For typewriter-like devices that support the @w{Latin-1}
(ISO@tie{}@w{8859-1}) character set.

@item utf8
@cindex encoding, output, @w{UTF-8}
@cindex @w{UTF-8}, output encoding
@cindex output encoding, @w{UTF-8}
For typewriter-like devices that use the Unicode (ISO@tie{}10646)
character set with @w{UTF-8} encoding.

@item cp1047
@cindex encoding, output, @acronym{EBCDIC}
@cindex @acronym{EBCDIC}, output encoding
@cindex output encoding, @acronym{EBCDIC}
@cindex encoding, output, code page 1047
@cindex code page 1047, output encoding
@cindex output encoding, code page 1047
@cindex IBM code page 1047 output encoding
For typewriter-like devices that use the @acronym{EBCDIC} encoding IBM
code page 1047.

@item lj4
For HP LaserJet4-compatible (or other PCL5-compatible) printers.

@item lbp
For Canon @acronym{CaPSL} printers (@w{LBP-4} and @w{LBP-8} series laser
printers).

@pindex pre-grohtml
@pindex post-grohtml
@cindex @code{grohtml}, the program
@item html
@itemx xhtml
To produce @acronym{HTML} and @acronym{XHTML} output, respectively.
This driver consists of two parts, a preprocessor
(@command{pre-grohtml}) and a postprocessor (@command{post-grohtml}).
@end table

@cindex output device name string (@code{.T})
@cindex output device usage register (@code{.T})
The predefined GNU @code{troff} string @code{.T} contains the name of
the output device; the read-only register @code{.T} is set to@tie{}1 if
this option is used (which is always true if @command{groff} is used to
call GNU @command{troff}).  @xref{Built-in Registers}.

The postprocessor to be used for a device is specified by the
@code{postpro} command in the device description file.  (@xref{Device
and Font Description Files}.)  This can be overridden with the
@option{-X} option.

@item -U
@cindex mode, unsafe
@cindex unsafe mode
Operate in @dfn{unsafe mode}, which enables the @code{open},
@code{opena}, @code{pi}, @code{pso}, and @code{sy} requests.  These
requests are disabled by default because they allow an untrusted input
document to write to arbitrary file names and run arbitrary commands.
This option also adds the current directory to the macro package search
path; see the @option{-m} option above.  @option{-U} is passed to
@command{gpic} and @command{gtroff}.

@item -v
Write version information for @command{groff} and all programs run by it
to the standard output stream; that is, the given command line is
processed in the usual way, passing @option{-v} to the formatter and any
pre- or postprocessors invoked.

@item -V
Output the pipeline that would be run by @command{groff}
(as a wrapper program) to the standard output stream, but do not execute
it.  If given more than once, the pipeline is both written to the
standard error stream and run.

@item -w@var{category}
Enable warnings in @var{category}.  Categories are listed in
@ref{Warnings}.

@item -W@var{category}
Inhibit warnings in @var{category}.  Categories are listed in
@ref{Warnings}.

@item -X
Use @command{gxditview} instead of the usual postprocessor to (pre)view
a document on an X11 display.  Combining this option with
@option{-Tps} uses the font metrics of the PostScript device, whereas
the @option{-TX75} and @option{-TX100} options use the metrics of X11
fonts.

@item -z
Suppress formatted output from @command{gtroff}.

@item -Z
Disable postprocessing.  @command{gtroff} output will appear on the
standard output stream (unless suppressed with @option{-z}; see
@ref{gtroff Output} for a description of this format.
@end table


@c =====================================================================

@node Environment, Macro Directories, Groff Options, Invoking groff
@section Environment
@cindex environment variables
@cindex variables in environment

There are also several environment variables (of the operating system,
not within @code{gtroff}) that can modify the behavior of @code{groff}.

@table @code
@item GROFF_BIN_PATH
@tindex GROFF_BIN_PATH@r{, environment variable}
This search path, followed by @code{PATH}, is used for commands executed
by @code{groff}.

@item GROFF_COMMAND_PREFIX
@tindex GROFF_COMMAND_PREFIX@r{, environment variable}
@cindex command prefix
@cindex prefix, for commands
If this is set to@tie{}@var{X}, then @command{groff} runs
@command{@var{X}troff} instead of @command{gtroff}.  This also applies
to @command{tbl}, @command{pic}, @command{eqn}, @command{grn},
@command{chem}, @command{refer}, and @command{soelim}.  It does not
apply to @command{grops}, @command{grodvi}, @command{grotty},
@command{pre-grohtml}, @command{post-grohtml}, @command{preconv},
@command{grolj4}, @command{gropdf}, and @command{gxditview}.

The default command prefix is determined during the installation
process.  If a non-GNU @code{troff} system is found, prefix @samp{g} is
used, none otherwise.

@item GROFF_ENCODING
@tindex GROFF_ENCODING@r{, environment variable}
The value of this variable is passed to the @code{preconv}
preprocessor's @option{-e} option to select the character encoding of
input files.  This variable's existence implies the @code{groff} option
@option{-k}.  If set but empty, @code{groff} calls @code{preconv}
without an @option{-e} option.  @code{groff}'s @option{-K} option
overrides @env{GROFF_ENCODING}.  See the @cite{preconv@r{(7)}} man page;
type @samp{man preconv} at the command line to view it.

@item GROFF_FONT_PATH
@tindex GROFF_FONT_PATH@r{, environment variable}
A list of directories in which to seek the selected output device's
directory of device and font description files.  GNU @code{troff}
will search directories given as arguments to any specified @option{-F}
options before these, and a built-in list of directories after them.
@xref{Font Directories} and the @cite{troff@r{(1)}} or
@cite{gtroff@r{(1)}} man pages.

@item GROFF_TMAC_PATH
@tindex GROFF_TMAC_PATH@r{, environment variable}
A list of directories in which to seek macro files.  GNU @code{troff}
will search directories given as arguments to any specified @option{-M}
options before these, and a built-in list of directories after them.
@xref{Macro Directories} and the @cite{troff@r{(1)}} or
@cite{gtroff@r{(1)}} man pages.

@item GROFF_TMPDIR
@tindex GROFF_TMPDIR@r{, environment variable}
@tindex TMPDIR@r{, environment variable}
The directory in which @code{groff} creates temporary files.  If this is
not set and @env{TMPDIR} is set, temporary files are created in that
directory.  Otherwise temporary files are created in a system-dependent
default directory (on Unix and GNU/Linux systems, this is usually
@file{/tmp}).  @code{grops}, @code{grefer}, @code{pre-grohtml}, and
@code{post-grohtml} can create temporary files in this directory.

@item GROFF_TYPESETTER
@tindex GROFF_TYPESETTER@r{, environment variable}
Sets the default output device.  If empty or not set, a build-time
default (often @code{ps}) is used.  The @option{-T@var{dev}} option
overrides @env{GROFF_TYPESETTER}.

@item SOURCE_DATE_EPOCH
@tindex SOURCE_DATE_EPOCH@r{, environment variable}
A timestamp (expressed as seconds since the Unix epoch) to use as the
output creation timestamp in place of the current time.  The time is
converted to human-readable form using @cite{localtime@r{(3)}} when the
formatter starts up and stored in registers usable by documents and
macro packages (@pxref{Built-in Registers}).

@item TZ
@tindex TZ@r{, environment variable}
The time zone to use when converting the current time (or value of
@env{SOURCE_DATE_EPOCH}) to human-readable form; see
@cite{tzset@r{(3)}}.
@end table

MS-DOS and MS-Windows ports of @code{groff} use semicolons, rather than
colons, to separate the directories in the lists described above.


@c =====================================================================

@node Macro Directories, Font Directories, Environment, Invoking groff
@section Macro Directories
@cindex macro directories
@cindex directories for macros
@cindex searching macros
@cindex macros, searching

A macro file must have a name in the form @code{@var{name}.tmac} or
@code{tmac.@var{name}} and be placed in a @dfn{tmac directory} to be
found by the @option{-m@var{name}} command-line option.@footnote{The
@code{mso} request does not have these limitations.  @xref{I/O}.}
@cindex tmac, directory
@cindex directory, for tmac files
@cindex tmac, path
@cindex path, for tmac files
@cindex locating macro files
@cindex macro file search path
@cindex file, macro, search path
@cindex locating macro packages
@cindex macro package search path
@cindex package, macro, search path
Together, these directories constitute the @dfn{tmac path}.  Each
directory is searched in the following order until the desired macro
file is found or the list is exhausted.

@itemize @bullet
@item
Directories specified with GNU @code{troff}'s or @code{groff}'s
@option{-M} command-line option.

@item
@tindex GROFF_TMAC_PATH@r{, environment variable}
Directories listed in the @env{GROFF_TMAC_PATH} environment variable.

@item
@cindex safer mode
@cindex mode, safer
@cindex unsafe mode
@cindex mode, unsafe
@cindex current directory
@cindex directory, current
The current working directory (only if in unsafe mode using the
@option{-U} command-line option).

@item
@cindex home directory
@cindex directory, home
The user's home directory, @env{HOME}.

@item
@cindex site-local directory
@cindex directory, site-local
@cindex platform-specific directory
@cindex directory, platform-specific
A platform-dependent directory, a site-local (platform-independent)
directory, and the main @slanted{tmac} directory.  The locations
corresponding to your installation are listed in section ``Environment''
of @cite{gtroff@r{(1)}}.  If not otherwise configured, they are as
follows.

@Example
/usr/local/lib/groff/site-tmac
/usr/local/share/groff/site-tmac
/usr/local/share/groff/1.23.0/tmac
@endExample

@noindent
The foregoing assumes that the version of @code{groff} is 1.23.0, and
that the installation prefix was @file{/usr/local}.  It is possible to
fine-tune these locations during the source configuration process.
@end itemize


@c =====================================================================

@node Font Directories, Paper Format, Macro Directories, Invoking groff
@section Font Directories
@cindex font directories
@cindex directories for fonts
@cindex searching fonts
@cindex fonts, searching

@code{groff} enforces few restrictions on how font description files are
named.  For its family/style mechanism to work (@pxref{Font Families}),
the names of fonts within a family should start with the family name,
followed by the style.  For example, the Times family uses @samp{T} for
the family name and @samp{R}, @samp{B}, @samp{I}, and @samp{BI} to
indicate the styles `roman', `bold', `italic', and `bold italic',
respectively.  Thus the final font names are @samp{TR}, @samp{TB},
@samp{TI}, and @samp{TBI}.

@cindex font path
@cindex path, for font files
Font description files are kept in @dfn{font directories}, which
together constitute the @dfn{font path}.  The search procedure
always appends the directory @code{dev}@var{name}, where @var{name} is
the name of the output device.  Assuming @TeX{} DVI output, and
@file{/foo/bar} as a font directory, the font description files for
@command{grodvi} must be in @file{/foo/bar/devdvi}.
Each directory in the font path is searched in the following order until
the desired font description file is found or the list is exhausted.

@itemize @bullet
@item
Directories specified with GNU @code{troff}'s or @code{groff}'s
@option{-f} command-line option.  All output drivers (and some
preprocessors) support this option as well, because they require
information about the glyphs to be rendered in the document.

@item
@tindex GROFF_FONT_PATH@r{, environment variable}
Directories listed in the @env{GROFF_FONT_PATH} environment variable.

@item
@cindex site-local directory
@cindex directory, site-local
A site-local directory and the main font description directory.
The locations corresponding to your installation are listed in section
``Environment'' of @cite{gtroff@r{(1)}}.  If not otherwise configured,
they are as follows.

@Example
/usr/local/share/groff/site-font
/usr/local/share/groff/1.23.0/font
@endExample

@noindent
The foregoing assumes that the version of @code{groff} is 1.23.0, and
that the installation prefix was @file{/usr/local}.  It is possible to
fine-tune these locations during the source configuration process.
@end itemize


@c =====================================================================

@node Paper Format, Invocation Examples, Font Directories, Invoking groff
@section Paper Format
@cindex paper format
@cindex format, paper
@cindex paper size
@cindex size, paper
@cindex landscape page orientation
@cindex orientation, landscape
@cindex page orientation, landscape

In @code{groff}, the page dimensions for the formatter GNU @code{troff}
and for output devices are handled separately.  @xref{Page Layout}, for
vertical manipulation of the page size, and @xref{Line Layout}, for
horizontal changes.
@pindex papersize.tmac
@pindex troffrc
The @file{papersize} macro package, normally loaded by @file{troffrc} at
startup, provides an interface for configuring page dimensions by
convenient names, like @samp{letter} or @samp{a4}; see
@cite{groff_tmac@r{(5)}}.  The default used by the formatter depends on
its build configuration, but is usually one of the foregoing, as
geographically appropriate.
@c groff(1), being generated, says what the default is.

It is up to each macro package to respect the page dimensions configured
in this way.

For each output device, the size of the output medium can be set in its
@file{DESC} file.  Most output drivers also recognize a command-line
option @option{-p} to override the default dimensions and an option
@option{-l} to use landscape orientation.  @xref{DESC File Format}, for
a description of the @code{papersize} keyword, which takes an argument
of the same form as @option{-p}.  The output driver's man page, such as
@cite{grops@r{(1)}}, may also be helpful.

@code{groff} uses the command-line option @option{-P} to pass options to
postprocessors; for example, use the following for PostScript output on
A4 paper in landscape orientation.

@Example
groff -Tps -dpaper=a4l -P-pa4 -P-l -ms foo.ms > foo.ps
@endExample


@c =====================================================================

@c BEGIN Keep parallel with groff(1), section "Examples".
@node Invocation Examples,  , Paper Format, Invoking groff
@section Invocation Examples
@cindex invocation examples
@cindex examples of invocation

@code{roff} systems are best known for formatting man pages.  Once a
@command{man} librarian program has located a man page, it may execute
a @code{groff} command much like the following.

@Example
groff -t -man -Tutf8 /usr/share/man/man1/groff.1
@endExample

The librarian will also pipe the output through a pager, which might not
interpret the SGR terminal escape sequences @command{groff} emits for
boldface, underlining, or italics; see the @cite{grotty@r{(1)}} man page
for a discussion.

To process a @code{roff} input file using the preprocessors
@command{gtbl} and @command{gpic} and the @file{me} macro package in the
way to which AT&T @code{troff} users were accustomed, one would type (or
script) a pipeline.

@Example
gpic foo.me | gtbl | gtroff -me -Tutf8 | grotty
@endExample

Using @command{groff}, this pipe can be shortened to an equivalent
command.

@Example
groff -p -t -me -T utf8 foo.me
@endExample

An even easier way to do this is to use @command{grog} to guess the
preprocessor and macro options and execute the result by using the
command substitution feature of the shell.

@Example
$(grog -Tutf8 foo.me)
@endExample

Each command-line option to a postprocessor must be specified with any
required leading dashes @samp{-}
@c No GNU roff postprocessor uses long options for anything except
@c --help or --version.
@c or @samp{--}
because @command{groff} passes the arguments as-is to the postprocessor;
this permits arbitrary arguments to be transmitted.  For example, to
pass a title to the @command{gxditview} postprocessor,
the shell commands

@Example
groff -X -P -title -P 'trial run' mydoc.t
@endExample

@noindent
and

@Example
groff -X -Z mydoc.t | gxditview -title 'trial run' -
@endExample

@noindent
are equivalent.
@c END Keep parallel with groff(1), section "Examples".

@codequotebacktick off
@codequoteundirected off



@c =====================================================================
@c =====================================================================

@node Tutorial for Macro Users, Major Macro Packages, Invoking groff, Top
@chapter Tutorial for Macro Users
@cindex tutorial for macro users
@cindex macros, tutorial for users
@cindex user's tutorial for macros
@cindex user's macro tutorial

Most users of the @code{roff} language employ a macro package to format
their documents.  Successful macro packages tend to ease the composition
process; their users need not have mastered the full formatting
language, nor even some of its major features like diversions, traps,
and environments.  A familiarity with some basic concepts and mechanisms
common to macro packages (like ``displays'') remains helpful; this
chapter aims to bring you to this level.  If you prefer a meticulous and
comprehensive presentation, try @ref{GNU troff Reference} instead.

@menu
* Basics::
* Common Features::
@end menu


@c =====================================================================

@node Basics, Common Features, Tutorial for Macro Users, Tutorial for Macro Users
@section Basics
@cindex basics of macros
@cindex macro basics

This section covers some of the basic concepts necessary to understand
how to use a macro package.@footnote{This section is derived from
@cite{Writing Papers with nroff using -me} by Eric@tie{}P.@: Allman.}
References are made throughout to more detailed information, if desired.

GNU @code{troff} reads an input file prepared by the user and outputs a
formatted document suitable for publication or framing.  The input
consists of text, or words to be printed, and embedded commands
(@dfn{requests} and @dfn{escape sequences}), which tell GNU @code{troff}
how to format the output.  For more detail on this, see @ref{Formatter
Instructions}.

The word @dfn{argument} is used in this chapter to mean a word or number
that appears on the same line as a request, and which modifies the
meaning of that request.  For example, the request

@Example
.sp
@endExample

@noindent
spaces one line, but

@Example
.sp 4
@endExample

@noindent
spaces four lines.  The number@tie{}4 is an argument to the @code{sp}
request, which says to space four lines instead of one.  Arguments are
separated from the request and from each other by spaces (@emph{no}
tabs).  @xref{Invoking Requests}.

The primary function of @code{gtroff} is to collect words from input
lines, fill output lines with those words, justify the right-hand margin
by inserting extra spaces in the line, and output the result.  For
example, the input:

@Example
Now is the time
for all good men
to come to the aid
of their party.
Four score and seven
years ago, etc.
@endExample

@noindent
is read, packed onto output lines, and justified to produce:

@quotation
Now is the time for all good men to come to the aid of their party.
Four score and seven years ago, etc.
@end quotation

@cindex break
@cindex line break
Sometimes a new output line should be started even though the current
line is not yet full; for example, at the end of a paragraph.  To do
this it is possible to cause a @dfn{break}, which starts a new output
line.  Some requests cause a break automatically, as normally do blank
input lines and input lines beginning with a space.

Not all input lines are text to be formatted.  Some input lines are
requests that describe how to format the text.  Requests always have a
period (@samp{.}) or an apostrophe (@samp{'}) as the first character of
the input line.

The text formatter also does more complex things, such as automatically
numbering pages, skipping over page boundaries, putting footnotes in the
correct place, and so forth.

Here are a few hints for preparing text for input to @code{gtroff}.

@itemize @bullet
@item
First, keep the input lines short.  Short input lines are easier to
edit, and @code{gtroff} packs words onto longer lines anyhow.

@item
In keeping with this, it is helpful to begin a new line after every
comma or phrase, since common corrections are to add or delete sentences
or phrases.

@item
End each sentence with two spaces---or better, start each sentence on a
new line.  @code{gtroff} recognizes characters that usually end a
sentence, and inserts inter-sentence space accordingly.

@item
Do not hyphenate words at the end of lines---@code{gtroff} is smart
enough to hyphenate words as needed, but is not smart enough to take
hyphens out and join a word back together.  Also, words such as
``mother-in-law'' should not be broken over a line, since then a space
can occur where not wanted, such as ``@w{mother- in}-law''.
@end itemize

@cindex double-spacing (@code{ls})
@cindex spacing
@code{gtroff} double-spaces output text automatically if you use the
request @w{@samp{.ls 2}}.  Reactivate single-spaced mode by typing
@w{@samp{.ls 1}}.@footnote{If you need finer granularity of the vertical
space, use the @code{pvs} request (@pxref{Changing the Type Size}).}

A number of requests allow you to change the way the output is arranged
on the page, sometimes called the @dfn{layout} of the output page.

@cindex new page (@code{bp})
The @code{bp} request starts a new page, causing a line break.

@cindex blank line (@code{sp})
@cindex empty line (@code{sp})
@cindex line, empty (@code{sp})
The request @w{@samp{.sp @var{N}}} leaves @var{N}@tie{}lines of blank
space.  @var{N}@tie{}can be omitted (meaning skip a single line) or can
be of the form @var{N}i (for @var{N}@tie{}inches) or @var{N}c (for
@var{N}@tie{}centimeters).  For example, the input:

@Example
.sp 1.5i
My thoughts on the subject
.sp
@endExample

@noindent
leaves one and a half inches of space, followed by the line ``My
thoughts on the subject'', followed by a single blank line (more
measurement units are available, see @ref{Measurements}).

@cindex centering lines (@code{ce})
@cindex lines, centering (@code{ce})
Text lines can be centered by using the @code{ce} request.  The line
after @code{ce} is centered (horizontally) on the page.  To center more
than one line, use @w{@samp{.ce @var{N}}} (where @var{N} is the number
of lines to center), followed by the @var{N}@tie{}lines.  To center many
lines without counting them, type:

@Example
.ce 1000
lines to center
.ce 0
@endExample

@noindent
The @w{@samp{.ce 0}} request tells @code{groff} to center zero more
lines, in other words, stop centering.

@cindex line break (@code{br})
@cindex break (@code{br})
All of these requests cause a break; that is, they always start a new
line.  To start a new line without performing any other action, use
@code{br}.


@c =====================================================================

@node Common Features,  , Basics, Tutorial for Macro Users
@section Common Features
@cindex common features
@cindex features, common

@code{gtroff} provides very low-level operations for formatting a
document.  There are many common routine operations that are done in
all documents.  These common operations are written into @dfn{macros}
and collected into a @dfn{macro package}.

All macro packages provide certain common capabilities that fall into
the following categories.

@menu
* Paragraphs::
* Sections and Chapters::
* Headers and Footers::
* Page Layout Adjustment::
* Displays::
* Footnotes and Annotations::
* Table of Contents::
* Indices::
* Paper Formats::
* Multiple Columns::
* Font and Size Changes::
* Predefined Strings::
* Preprocessor Support::
* Configuration and Customization::
@end menu

@c ---------------------------------------------------------------------

@node Paragraphs, Sections and Chapters, Common Features, Common Features
@subsection Paragraphs
@cindex paragraphs

One of the most common and most used capability is starting a paragraph.
There are a number of different types of paragraphs, any of which can be
initiated with macros supplied by the macro package.  Normally,
paragraphs start with a blank line and the first line indented, like the
text in this manual.  There are also block style paragraphs, which omit
the indentation:

@Example
Some   men  look   at  constitutions   with  sanctimonious
reverence, and deem them like the ark of the covenant, too
sacred to be touched.
@endExample

@noindent
And there are also indented paragraphs, which begin with a tag or label
at the margin and the remaining text indented.

@Example
one   This is  the first paragraph.  Notice  how the first
      line of  the resulting  paragraph lines up  with the
      other lines in the paragraph.
@endExample

@Example
longlabel
      This  paragraph   had  a  long   label.   The  first
      character of text on the first line does not line up
      with  the  text  on  second  and  subsequent  lines,
      although they line up with each other.
@endExample

A variation of this is a bulleted list.

@Example
.     Bulleted lists start with a bullet.   It is possible
      to use other glyphs instead of the bullet.  In nroff
      mode using the ASCII character set for output, a dot
      is used instead of a real bullet.
@endExample

@c ---------------------------------------------------------------------

@node Sections and Chapters, Headers and Footers, Paragraphs, Common Features
@subsection Sections and Chapters

Most macro packages supply some form of section headers.  The simplest
kind is simply the heading on a line by itself in bold type.  Others
supply automatically numbered section heading or different heading
styles at different levels.  Some, more sophisticated, macro packages
supply macros for starting chapters and appendices.

@c ---------------------------------------------------------------------

@node Headers and Footers, Page Layout Adjustment, Sections and Chapters, Common Features
@subsection Headers and Footers

Every macro package gives some way to manipulate the @dfn{headers} and
@dfn{footers} (also called @dfn{titles}) on each page.  This is text put
at the top and bottom of each page, respectively, which contain data
like the current page number, the current chapter title, and so on.  Its
appearance is not affected by the running text.  Some packages allow for
different ones on the even and odd pages (for material printed in a book
form).

The titles are called @dfn{three-part titles}, that is, there is a
left-justified part, a centered part, and a right-justified part.  An
automatically generated page number may be put in any of these fields
with the @samp{%} character (@pxref{Page Layout}).

@c ---------------------------------------------------------------------

@node Page Layout Adjustment, Displays, Headers and Footers, Common Features
@subsection Page Layout

Most macro packages let the user specify top and bottom margins and
other details about the appearance of the printed pages.

@c ---------------------------------------------------------------------

@node Displays, Footnotes and Annotations, Page Layout Adjustment, Common Features
@subsection Displays
@cindex displays

@dfn{Displays} are sections of text to be set off from the body of the
paper.  Major quotes, tables, and figures are types of displays, as are
all the examples used in this document.

@cindex quotes, major
@cindex major quotes
@dfn{Major quotes} are quotes that are several lines long, and hence
are set in from the rest of the text without quote marks around them.

@cindex list
A @dfn{list} is an indented, single-spaced, unfilled display.  Lists
should be used when the material to be printed should not be filled and
justified like normal text, such as columns of figures or the examples
used in this paper.

@cindex keep
A @dfn{keep} is a display of lines that are kept on a single page if
possible.  An example for a keep might be a diagram.  Keeps differ from
lists in that lists may be broken over a page boundary whereas keeps are
not.

@cindex keep, floating
@cindex floating keep
@dfn{Floating keeps} move relative to the text.  Hence, they are good
for things that are referred to by name, such as ``See figure@tie{}3''.
A floating keep appears at the bottom of the current page if it fits;
otherwise, it appears at the top of the next page.  Meanwhile, the
surrounding text `flows' around the keep, thus leaving no blank areas.

@c ---------------------------------------------------------------------

@node Footnotes and Annotations, Table of Contents, Displays, Common Features
@subsection Footnotes and Annotations
@cindex footnotes
@cindex annotations

There are a number of requests to save text for later printing.

@dfn{Footnotes} are printed at the bottom of the current page.

@cindex delayed text
@dfn{Delayed text} is very similar to a footnote except that it is
printed when called for explicitly.  This allows a list of references to
appear (for example) at the end of each chapter, as is the convention in
some disciplines.

Most macro packages that supply this functionality also supply a means
of automatically numbering either type of annotation.

@c ---------------------------------------------------------------------

@node Table of Contents, Indices, Footnotes and Annotations, Common Features
@subsection Table of Contents
@cindex table of contents
@cindex contents, table of

@dfn{Tables of contents} are a type of delayed text having a tag
(usually the page number) attached to each entry after a row of dots.
The table accumulates throughout the paper until printed, usually after
the paper has ended.  Many macro packages provide the ability to have
several tables of contents (e.g., a standard table of contents, a list
of tables, etc).

@c ---------------------------------------------------------------------

@node Indices, Paper Formats, Table of Contents, Common Features
@subsection Indices
@cindex index, in macro package

While some macro packages use the term @dfn{index}, none actually
provide that functionality.  The facilities they call indices are
actually more appropriate for tables of contents.

@pindex makeindex
To produce a real index in a document, external tools like the
@code{makeindex} program are necessary.

@c ---------------------------------------------------------------------

@node Paper Formats, Multiple Columns, Indices, Common Features
@subsection Paper Formats
@cindex paper formats

Some macro packages provide stock formats for various kinds of
documents.  Many of them provide a common format for the title and
opening pages of a technical paper.  The @file{mm} macros in particular
provide formats for letters and memoranda.

@c ---------------------------------------------------------------------

@node Multiple Columns, Font and Size Changes, Paper Formats, Common Features
@subsection Multiple Columns

Some macro packages (but not @file{man}) provide the ability to have two
or more columns on a page.

@c ---------------------------------------------------------------------

@node Font and Size Changes, Predefined Strings, Multiple Columns, Common Features
@subsection Font and Size Changes

The built-in font and size functions are not always intuitive, so all
macro packages provide macros to make these operations simpler.

@c ---------------------------------------------------------------------

@node Predefined Strings, Preprocessor Support, Font and Size Changes, Common Features
@subsection Predefined Strings

Most macro packages provide various predefined strings for a variety of
uses; examples are sub- and superscripts, printable dates, quotes and
various special characters.

@c ---------------------------------------------------------------------

@node Preprocessor Support, Configuration and Customization, Predefined Strings, Common Features
@subsection Preprocessor Support

All macro packages provide support for various preprocessors and may
extend their functionality.

For example, all macro packages mark tables (which are processed with
@code{gtbl}) by placing them between @code{TS} and @code{TE} macros.
The @file{ms} macro package has an option, @samp{.TS@tie{}H}, that
prints a caption at the top of a new page (when the table is too long to
fit on a single page).

@c ---------------------------------------------------------------------

@node Configuration and Customization,  , Preprocessor Support, Common Features
@subsection Configuration and Customization

Some macro packages provide means of customizing many of the details of
how the package behaves.  This ranges from setting the default type size
to changing the appearance of section headers.



@c =====================================================================
@c =====================================================================

@node Major Macro Packages, GNU troff Reference, Tutorial for Macro Users, Top
@chapter Macro Packages
@cindex major macro package
@cindex package, macro, major
@cindex macro package, major

This chapter surveys the ``major'' macro packages that come with
@code{groff}.  One, @file{ms}, is presented in detail.

@cindex full-service macro package
@cindex package, macro, full-service
@cindex macro package, full-service
Major macro packages are also sometimes described as @dfn{full-service}
due to the breadth of features they provide and because more than one
cannot be used by the same document; for example

@Example
groff -m man foo.man -m ms bar.doc
@endExample

@noindent
doesn't work.  Option arguments are processed before non-option
arguments; the above (failing) sample is thus reordered to

@Example
groff -m man -m ms foo.man bar.doc
@endExample

@cindex minor macro package
@cindex package, macro, minor
@cindex macro package, minor
@cindex auxiliary macro package
@cindex package, macro, auxiliary
@cindex macro package, auxiliary
Many auxiliary, or ``minor'' macro packages are also available.  They
may in general be used with any full-service macro package and handle a
variety of tasks from character encoding selection, to language
localization, to inlining of raster images.  See the
@cite{groff_tmac@r{(5)}} man page for a list.  Type @samp{man
groff_tmac} at the command line to view it.

@menu
* man::
* mdoc::
* me::
* mm::
* mom::
* ms::
@end menu


@c =====================================================================

@node man, mdoc, Major Macro Packages, Major Macro Packages
@section @file{man}
@cindex manual pages
@cindex man pages
@pindex an.tmac
@pindex man.tmac

The @code{man} macro package is the most widely used and probably the
most important ever developed for @code{troff}.  It is easy to use, and
a vast majority of manual pages (``man pages'') are written in it.

@code{groff}'s implementation is documented in the
@cite{groff_man@r{(7)}} man page.  Type @samp{man groff_man} at the
command line to view it.

@menu
* Optional man extensions::
@end menu

@c ---------------------------------------------------------------------

@node Optional man extensions,  ,  , man
@subsection Optional @file{man} extensions

@pindex man.local
Use the file @file{man.local} for local extensions to the @code{man}
macros or for style changes.

@unnumberedsubsubsec Custom headers and footers
@cindex @code{man} macros, custom headers and footers

In @code{groff} versions 1.18.2 and later, you can specify custom
headers and footers by redefining the following macros in
@file{man.local}.

@Defmac {PT, , man}
Control the content of the headers.  Normally, the header prints the
command name and section number on either side, and the optional fifth
argument to @code{TH} in the center.
@endDefmac

@Defmac {BT, , man}
Control the content of the footers.  Normally, the footer prints the
page number and the third and fourth arguments to @code{TH}.

Use the @code{FT} register to specify the footer position.  The default
is @minus{}0.5@dmn{i}.
@endDefmac

@unnumberedsubsubsec Ultrix-specific man macros
@cindex Ultrix-specific @code{man} macros
@cindex @code{man} macros, Ultrix-specific

@pindex man.ultrix
The @code{groff} source distribution includes a file named
@file{man.ultrix}, containing macros compatible with the Ultrix variant
of @code{man}.  Copy this file into @file{man.local} (or use the
@code{mso} request to load it) to enable the following macros.

@Defmac {CT, @Var{key}, man}
Print @samp{<CTRL/@var{key}>}.
@endDefmac

@Defmac {CW, , man}
Print subsequent text using a ``constant-width'' (monospaced) typeface
(Courier roman).
@endDefmac

@Defmac {Ds, , man}
Begin a non-filled display.
@endDefmac

@Defmac {De, , man}
End a non-filled display started with @code{Ds}.
@endDefmac

@Defmac {EX, [@Var{indent}], man}
Begin a non-filled display using a monospaced typeface (Courier roman).
Use the optional @var{indent} argument to indent the display.
@endDefmac

@Defmac {EE, , man}
End a non-filled display started with @code{EX}.
@endDefmac

@Defmac {G, [@Var{text}], man}
Set @var{text} in Helvetica.  If no text is present on the line where
the macro is called, then the text of the next line appears in
Helvetica.
@endDefmac

@Defmac {GL, [@Var{text}], man}
Set @var{text} in Helvetica oblique.  If no text is present on the line
where the macro is called, then the text of the next line appears in
Helvetica Oblique.
@endDefmac

@Defmac {HB, [@Var{text}], man}
Set @var{text} in Helvetica bold.  If no text is present on the line
where the macro is called, then all text up to the next @code{HB}
appears in Helvetica bold.
@endDefmac

@Defmac {TB, [@Var{text}], man}
Identical to @code{HB}.
@endDefmac

@Defmac {MS, @Var{title} @Var{sect} [@Var{punct}], man}
Set a man page reference in Ultrix format.  The @var{title} is in
Courier instead of italic.  Optional punctuation follows the section
number without an intervening space.
@endDefmac

@Defmac {NT, [@code{C}] [@Var{title}], man}
Begin a note.  Print the optional @Var{title}, or the word ``Note'',
centered on the page.  Text following the macro makes up the body of the
note, and is indented on both sides.  If the first argument is @code{C},
the body of the note is printed centered (the second argument replaces
the word ``Note'' if specified).
@endDefmac

@Defmac {NE, , man}
End a note begun with @code{NT}.
@endDefmac

@Defmac {PN, @Var{path} [@Var{punct}], man}
Set the path name in a monospaced typeface (Courier roman), followed by
optional punctuation.
@endDefmac

@Defmac {Pn, [@Var{punct}] @Var{path} [@Var{punct}], man}
If called with two arguments, identical to @code{PN}.  If called with
three arguments, set the second argument in a monospaced typeface
(Courier roman), bracketed by the first and third arguments in the
current font.
@endDefmac

@Defmac {R, , man}
Switch to roman font and turn off any underlining in effect.
@endDefmac

@Defmac {RN, , man}
Print the string @samp{<RETURN>}.
@endDefmac

@Defmac {VS, [@code{4}], man}
Start printing a change bar in the margin if the number@tie{}@code{4} is
specified.  Otherwise, this macro does nothing.
@endDefmac

@Defmac {VE, , man}
End printing the change bar begun by @code{VS}.
@endDefmac

@unnumberedsubsubsec Simple example

The following example @file{man.local} file alters the @code{SH} macro
to add some extra vertical space before printing the heading.  Headings
are printed in Helvetica bold.

@Example
.\" Make the heading fonts Helvetica
.ds HF HB
.
.\" Put more space in front of headings.
.rn SH SH-orig
.de SH
.  if t .sp (u;\\n[PD]*2)
.  SH-orig \\$*
..
@endExample


@c =====================================================================

@node mdoc, me, man, Major Macro Packages
@section @file{mdoc}
@cindex @code{mdoc} macros

@code{groff}'s implementation of the BSD @file{doc} package for man
pages is documented in the @cite{groff_mdoc@r{(7)}} man page.  Type
@samp{man groff_mdoc} at the command line to view it.


@c =====================================================================

@node me, mm, mdoc, Major Macro Packages
@section @file{me}
@cindex @code{me} macro package

@code{groff}'s implementation of the BSD @file{me} macro package is
documented using itself.  A tutorial, @file{meintro.me}, and reference,
@file{meref.me}, are available in @code{groff}'s documentation
directory.  A @cite{groff_me@r{(7)}} man page is also available and
identifies the installation path for these documents.  Type @samp{man
groff_me} at the command line to view it.

A French translation of the tutorial is available as
@file{meintro_fr.me} and installed parallel to the English version.


@c =====================================================================

@node mm, mom, me, Major Macro Packages
@section @file{mm}
@cindex @code{mm} macro package

@code{groff}'s implementation of the @acronym{AT&T} memorandum macro
package is documented in the @cite{groff_mm@r{(7)}} man page.  Type
@samp{man groff_mm} at the command line) to view it.

A Swedish localization of @file{mm} is also available; see
@cite{groff_mmse@r{(7)}}.


@c =====================================================================

@node mom, ms, mm, Major Macro Packages
@section @file{mom}
@cindex @code{mom} macro package

The main documentation files for the @file{mom} macros are in
@acronym{HTML} format.  Additional, useful documentation is in
@acronym{PDF} format.  See the @cite{groff@r{(1)}} man page, section
``Installation Directories'', for their location.

@itemize @bullet
@item
@file{toc.html}
@noindent
Entry point to the full mom manual.

@item
@file{macrolist.html}
@noindent
Hyperlinked index of macros with brief descriptions, arranged by
category.

@item
@file{mom-pdf.pdf}
@noindent
@acronym{PDF} features and usage.
@end itemize

The mom macros are in active development between @code{groff} releases.
The most recent version, along with up-to-date documentation, is
available at @uref{http://www.schaffter.ca/mom/mom-05.html}.

The @cite{groff_mom@r{(7)}} man page (type @samp{man groff_mom} at
the command line) contains a partial list of available macros, however
their usage is best understood by consulting the @acronym{HTML}
documentation.


@c =====================================================================

@codequotebacktick on
@codequoteundirected on

@node ms,  , mom, Major Macro Packages
@section @file{ms}
@cindex @file{ms} macros

The @file{ms} (``manuscript'') package is suitable for the preparation
of letters, memoranda, reports, and books.  These @code{groff}
macros feature cover page and table of contents generation,
automatically numbered headings, several paragraph styles, a variety of
text styling options, footnotes, and multi-column page layouts.
@file{ms} supports the @command{tbl}, @command{eqn}, @command{pic}, and
@command{refer} preprocessors for inclusion of tables, mathematical
equations, diagrams, and standardized bibliographic citations.  This
implementation is mostly compatible with the documented interface and
behavior of AT&T Unix Version@tie{}7 @file{ms}.  Many extensions from
4.2BSD (Berkeley)
@c Few changes were made in 4.3, Tahoe, Reno, or 4.4.
and Tenth Edition Research Unix have been recreated.

@menu
* ms Introduction::
* ms Document Structure::
* ms Document Control Settings::
* ms Document Description Macros::
* ms Body Text::
* ms Page Layout::
* Differences from AT&T ms::
* ms Legacy Features::
* ms Naming Conventions::
@end menu

@c ---------------------------------------------------------------------

@node ms Introduction, ms Document Structure, ms, ms
@subsection Introduction

The @file{ms} macros are the oldest surviving package for @code{roff}
systems.@footnote{While manual @emph{pages} are older, early ones used
macros supplanted by the @file{man} package of Seventh Edition Unix
(1979).  @file{ms} shipped with Sixth Edition (1975) and was documented
by Mike Lesk in a Bell Labs internal memorandum.}  While the @file{man}
package was designed for brief reference documents, the @file{ms} macros
are also suitable for longer works intended for printing and possible
publication.

@menu
* ms basic information::
@end menu

@c ---------------------------------------------------------------------

@node ms basic information, ms Document Structure, ms Introduction, ms Introduction
@subsubsection Basic information

@file{ms} documents are plain text files; prepare them with your
preferred text editor.  If you're in a hurry to start, know that
@file{ms} needs one of its macros called at the beginning of a document
so that it can initialize.  A @dfn{macro} is a formatting instruction to
@file{ms}.  Put a macro call on a line by itself.  Use @samp{.PP} if you
want your paragraph's first line to be indented, or @samp{.LP} if you
don't.

After that, start typing normally.  It is a good practice to start each
sentence on a new line, or to put two spaces after sentence-ending
puncutation, so that the formatter knows where the sentence boundaries
are.  You can separate paragraphs with further paragraphing macros, or
with blank lines, and you can indent with tabs.  When you need one of
the features mentioned earlier (@pxref{ms}), return to this part of the
manual.

Format the document with the @command{groff} command.  @command{nroff}
can be useful for previewing.

@CartoucheExample
$ editor radical.ms
$ nroff -ww -z -ms radical.ms # check for errors
$ nroff -ms radical.ms | less -R
$ groff -T ps -ms radical.ms > radical.ps
$ see radical.ps
@endCartoucheExample

Our @file{radical.ms} document might look like this.

@CartoucheExample
.LP
Radical novelties are so disturbing that they tend to be
suppressed or ignored, to the extent that even the
possibility of their existence in general is more often
denied than admitted.

@arrow{}That's what Dijkstra said, anyway.
@endCartoucheExample

@file{ms} exposes many aspects of document layout to user control via
@code{groff}'s @dfn{registers} and @dfn{strings}, which store numbers
and text, respectively.  Measurements in @code{groff} are expressed with
a suffix called a @dfn{scaling unit}.

@table @code
@item i
inches

@item c
centimeters

@item p
points (1/72 inch)

@item P
picas (1/6 inch)

@item v
vees; current vertical spacing

@item m
ems; width of an ``M'' in the current font

@item n
ens; one-half em
@end table

Set registers with the @code{nr} request and strings with the @code{ds}
request.  @dfn{Requests} are like macro calls; they go on lines by
themselves and start with the @dfn{control character}, a dot (@code{.}).
The difference is that they directly instruct the formatter program,
rather than the macro package.  We'll discuss a few as applicable.  It
is wise to specify a scaling unit when setting any register that
represents a length, size, or distance.

@CartoucheExample
.nr PS 10.5p \" Use 10.5-point type.
.ds FAM P    \" Use Palatino font family.
@endCartoucheExample

@noindent
In the foregoing, we see that @code{\"} begins a comment.  This is an
example of an @dfn{escape sequence}, the other kind of formatting
instruction.  Escape sequences can appear anywhere.  They begin with the
escape character (@code{\}) and are followed by at least one more
character.  @file{ms} documents
@c like this one
tend to use only a few of @code{groff}'s many requests and escape
sequences; see @ref{Request Index} and @ref{Escape Sequence Index} or
the @cite{groff@r{(7)}} man page for complete lists.

@table @code
@item \"
Begin comment; ignore remainder of line.

@item \n[@var{reg}]
Interpolate value of register @var{reg}.

@item \*[@var{str}]
Interpolate contents of string @var{str}.

@item \*@var{s}
abbreviation of @code{\*[@var{s}]}; the name @var{s} must be only one
character

@item \[@var{char}]
Interpolate glyph of special character named @var{char}.

@item \&
dummy character

@item \~
Insert an unbreakable space that is adjustable like a normal space.

@item \|
Move horizontally by one-sixth em (``thin space'').
@end table

Prefix any words that start with a dot @samp{.} or neutral apostrophe
@samp{'} with @code{\&} if they are at the beginning of an input line
(or might become that way in editing) to prevent them from being
interpreted as macro calls or requests.  Suffix @samp{.}, @samp{?}, and
@samp{!} with @code{\&} when needed to cancel end-of-sentence detection.

@CartoucheExample
My exposure was \&.5 to \&.6 Sv of neutrons, said Dr.\&
Wallace after the criticality incident.
@endCartoucheExample

@c ---------------------------------------------------------------------

@node ms Document Structure, ms Document Control Settings, ms Introduction, ms
@subsection Document Structure
@cindex @file{ms} macros, general structure

The @file{ms} macro package expects a certain amount of structure:
a well-formed document contains at least one paragraphing or heading
macro call.  Longer documents have a structure as follows.

@table @strong
@item Document type
Calling the @code{RP} macro at the beginning of your document puts the
document description (see below) on a cover page.  Otherwise, @file{ms}
places the information (if any) on the first page, followed immediately
by the body text.  Some document types found in other @file{ms}
implementations are specific to @acronym{AT&T} or Berkeley, and are not
supported by @code{groff} @file{ms}.

@item Format and layout
By setting registers and strings, you can configure your document's
typeface, margins, spacing, headers and footers, and footnote
arrangement.  @xref{ms Document Control Settings}.

@item Document description
A document description consists of any of: a title, one or more authors'
names and affiliated institutions, an abstract, and a date or other
identifier.  @xref{ms Document Description Macros}.

@item Body text
The main matter of your document follows its description (if any).
@file{ms} supports highly structured text consisting of paragraphs
interspersed with multi-level headings (chapters, sections, subsections,
and so forth) and augmented by lists, footnotes, tables, diagrams, and
similar material.  @xref{ms Body Text}.

@item Tables of contents
Macros enable the collection of entries for a table of contents (or
index) as the material they discuss appears in the document.  You then
call a macro to emit the table of contents at the end of your document.
The table of contents must necessarily follow the rest of the text since
GNU @code{troff} is a single-pass formatter; it thus cannot determine
the page number of a division of the text until it has been set and
output.  Since @file{ms} was designed for the production of hard copy,
the traditional procedure was to manually relocate the pages containing
the table of contents between the cover page and the body text.  Today,
page resequencing can be done in the digital domain with tools like
@cite{pdfjam@r{(1)}}.  An index works similarly, but because it
typically needs to be sorted after collection, its preparation requires
separate processing.
@end table

@c ---------------------------------------------------------------------

@node ms Document Control Settings, ms Document Description Macros, ms Document Structure, ms
@subsection Document Control Settings
@cindex @file{ms} macros, document control settings

@file{ms} exposes many aspects of document layout to user control via
@code{groff} requests.  To use them, you must understand how to define
registers and strings.

@Defreq {nr, reg value}
Set register @var{reg} to @var{value}.  If @var{reg} doesn't exist, GNU
@code{troff} creates it.
@endDefreq

@Defreq {ds, name contents}
Set string @var{name} to @var{contents}.
@endDefreq

A list of document control registers and strings follows.  For any
parameter whose default is unsatisfactory, define its register or string
before calling any @file{ms} macro other than @code{RP}.

@unnumberedsubsubsec Margin settings

@Defmpreg {PO, ms}
Defines the page offset (i.e., the left margin).
@c not in V6

Effective: next page.

Default: Varies by output device and paper format; 1@dmn{i} is used for
typesetters using U.S.@: letter paper, and zero for terminals.
@xref{Paper Format}.
@endDefmpreg

@Defmpreg {LL, ms}
Defines the line length (i.e., the width of the body text).

Effective: next paragraph.

Default: Varies by output device and paper format; 6.5@dmn{i} is used
for typesetters using U.S.@: letter paper (@pxref{Paper Format}) and
65@dmn{n} on terminals.
@endDefmpreg

@Defmpreg {LT, ms}
Defines the title line length (i.e., the header and footer width).  This
is usually the same as @code{LL}, but need not be.

Effective: next paragraph.

Default: Varies by output device and paper format; 6.5@dmn{i} is used
for typesetters using U.S.@: letter paper (@pxref{Paper Format}) and
65@dmn{n} on terminals.
@endDefmpreg

@Defmpreg {HM, ms}
Defines the header margin height at the top of the page.
@c not in V6

Effective: next page.

Default: 1@dmn{i}.
@endDefmpreg

@Defmpreg {FM, ms}
Defines the footer margin height at the bottom of the page.
@c not in V6

Effective: next page.

Default: 1@dmn{i}.
@endDefmpreg

@unnumberedsubsubsec Titles (headers, footers)

@Defmpstr {LH, ms}
Defines the text displayed in the left header position.

Effective: next header.

Default: empty.
@endDefmpstr

@Defmpstr {CH, ms}
Defines the text displayed in the center header position.

Effective: next header.

Default: @samp{-\n[%]-}.
@endDefmpstr

@Defmpstr {RH, ms}
Defines the text displayed in the right header position.

Effective: next header.

Default: empty.
@endDefmpstr

@Defmpstr {LF, ms}
Defines the text displayed in the left footer position.

Effective: next footer.

Default: empty.
@endDefmpstr

@Defmpstr {CF, ms}
Defines the text displayed in the center footer position.

Effective: next footer.

Default: empty.
@endDefmpstr

@Defmpstr {RF, ms}
Defines the text displayed in the right footer position.

Effective: next footer.

Default: empty.
@endDefmpstr

@unnumberedsubsubsec Text settings

@Defmpreg {PS, ms}
Defines the type size of the body text.

Effective: next paragraph.

Default: 10@dmn{p}.
@endDefmpreg

@Defmpreg {VS, ms}
Defines the vertical spacing (type size plus leading).

Effective: next paragraph.

Default: 12@dmn{p}.
@endDefmpreg

@Defmpreg {HY, ms}
Defines the automatic hyphenation mode used with the @code{hy} request.
Setting @code{HY} to@tie{}0 is equivalent to using the @code{nh}
request.  This is a Tenth Edition Research Unix extension.
@c possibly 9th, but definitely not Berkeley

Effective: next paragraph.

Default: 6.
@endDefmpreg

@Defmpstr {FAM, ms}
Defines the font family used to typeset the document.  This is a GNU
extension.

Effective: next paragraph.

Default: defined by the output device; often @samp{T} (@pxref{ms Body
Text})
@endDefmpstr

@unnumberedsubsubsec Paragraph settings

@Defmpreg {PI, ms}
Defines the indentation amount used by the @code{PP}, @code{IP} (unless
overridden by an optional argument), @code{XP}, and @code{RS} macros.
@c not in V6

Effective: next paragraph.

Default: 5@dmn{n}.
@endDefmpreg

@Defmpreg {PD, ms}
Defines the space between paragraphs.
@c not in V6

Effective: next paragraph.

Default: 0.3@dmn{v} (1@dmn{v} on low-resolution devices).
@endDefmpreg

@Defmpreg {QI, ms}
Defines the indentation amount used on both sides of a paragraph set
with the @code{QP} or between the @code{QS} and @code{QE} macros.

Effective: next paragraph.

Default: 5@dmn{n}.
@endDefmpreg

@Defmpreg {PORPHANS, ms}
Defines the minimum number of initial lines of any paragraph that must
be kept together to avoid isolated lines at the bottom of a page.  If a
new paragraph is started close to the bottom of a page, and there is
insufficient space to accommodate @code{PORPHANS} lines before an
automatic page break, then a page break is forced before the start of
the paragraph.  This is a GNU extension.

Effective: next paragraph.

Default: 1.
@endDefmpreg

@unnumberedsubsubsec Heading settings

@Defmpreg {PSINCR, ms}
Defines an increment in type size to be applied to a heading at a
lesser depth than that specified in @code{GROWPS}.  The value of
@code{PSINCR} should be specified in points with the @dmn{p} scaling
unit and may include a fractional component; for example, @w{@samp{.nr
PSINCR 1.5p}} sets a type size increment of 1.5@dmn{p}.  This is a GNU
extension.

Effective: next heading.

Default: 1@dmn{p}.
@endDefmpreg

@Defmpreg {GROWPS, ms}
Defines the heading depth above which the type size increment set by
@code{PSINCR} becomes effective.  For each heading depth less than the
value of @code{GROWPS}, the type size is increased by @code{PSINCR}.
Setting @code{GROWPS} to any value less than@tie{}2 disables the
incremental heading size feature.  This is a GNU extension.

Effective: next heading.

Default: 0.
@endDefmpreg

@Defmpreg {HORPHANS, ms}
Defines the minimum number of lines of an immediately succeeding
paragraph that should be kept together with any heading introduced by
the @code{NH} or @code{SH} macros.  If a heading is placed close to the
bottom of a page, and there is insufficient space to accommodate both
the heading and at least @code{HORPHANS} lines of the following
paragraph, before an automatic page break, then the page break is forced
before the heading.  This is a GNU extension.

Effective: next paragraph.

Default: 1.
@endDefmpreg

@Defmpstr {SN-STYLE, ms}
Defines the style used to print numbered headings.  @xref{Headings in
ms}.  This is a GNU extension.

Effective: next heading.

Default: alias of @code{SN-DOT}
@endDefmpstr

@unnumberedsubsubsec Footnote settings

@Defmpreg {FI, ms}
Defines the footnote indentation.  This is a Berkeley extension.

Effective: next footnote.

Default: 2@dmn{n}.
@endDefmpreg

@Defmpreg {FF, ms}
Defines the format of automatically numbered footnotes,
and those for which the @code{FS} request is given a marker argument, at
the bottom of a column or page.  This is a Berkeley extension.
@table @code
@item 0
Set an automatic number@footnote{defined in @ref{ms Footnotes}} as a
superscript (on typesetter devices) or surrounded by square brackets (on
terminals).  The footnote paragraph is indented as with @code{PP} if
there is an @code{FS} argument or an automatic number, and as with
@code{LP} otherwise.  This is the default.

@item 1
As @code{0}, but set the marker as regular text and follow an
automatic number with a period.

@item 2
As @code{1}, but without indentation (like @code{LP}).

@item 3
As @code{1}, but set the footnote paragraph with the marker hanging
(like @code{IP}).
@end table

Effective: next footnote.

Default: 0.
@endDefmpreg

@Defmpreg {FPS, ms}
Defines the footnote type size.

Effective: next footnote.

Default: @code{\n[PS] - 2p}.
@endDefmpreg

@Defmpreg {FVS, ms}
Defines the footnote vertical spacing.

Effective: next footnote.

Default: @code{\n[FPS] + 2p}.
@endDefmpreg

@Defmpreg {FPD, ms}
Defines the footnote paragraph spacing.  This is a GNU extension.

Effective: next footnote.

Default: @code{\n[PD] / 2}.
@endDefmpreg

@Defmpstr {FR, ms}
Defines the ratio of the footnote line length to the current line
length.  This is a GNU extension.

Effective: next footnote in single-column arrangements, next page
otherwise.

Default: @code{11/12}.
@endDefmpstr

@unnumberedsubsubsec Display settings

@Defmpreg {DD, ms}
Sets the display distance---the vertical spacing before and after a
display, a @code{tbl} table, an @code{eqn} equation, or a @code{pic}
image.  This is a Berkeley extension.

Effective: next display boundary.

Default: 0.5@dmn{v} (1@dmn{v} on low-resolution devices).
@endDefmpreg

@Defmpreg {DI, ms}
Sets the default amount by which to indent a display started with
@code{DS} and @code{ID} without arguments, to @samp{.DS@tie{}I} without
an indentation argument, and to equations set with @samp{.EQ@tie{}I}.
This is a GNU extension.

Effective: next indented display.

Default: 0.5@dmn{i}.
@endDefmpreg

@unnumberedsubsubsec Other settings

@Defmpreg {MINGW, ms}
Defines the default minimum width between columns in a multi-column
document.  This is a GNU extension.

Effective: next page.

Default: 2@dmn{n}.
@endDefmpreg

@Defmpreg {TC-MARGIN, ms}
Defines the width of the field in which page numbers are set in a table
of contents entry; the right margin thus moves inboard by this amount.
This is a GNU extension.

Effective: next @code{PX} call.

Default: @code{\w'000'}
@endDefmpreg

@c XXX: Normally we'd have an entry for TC-LEADER here, but it's a
@c special character and we have no custom Texinfo macros for defining
@c (and indexing) these.  There would be little point in an index for
@c one item, and the plan is to drop this entire @section from this
@c manual once doc/ms.ms is ready.  See Savannah #60061.

@c ---------------------------------------------------------------------

@node ms Document Description Macros, ms Body Text, ms Document Control Settings, ms
@subsection Document Description Macros
@cindex @file{ms} macros, document description
@cindex document description macros, [@file{ms}]

Only the simplest document lacks a title.@footnote{Distinguish a
document title from ``titles'', which are what @code{roff} systems call
headers and footers collectively.}  As its level of sophistication (or
complexity) increases, it tends to acquire a date of revision,
explicitly identified authors, sponsoring institutions for authors, and,
at the rarefied heights, an abstract of its content.  Define these
data by calling the macros below in the order shown; @code{DA} or
@code{ND} can be called to set the document date (or other identifier)
at any time before (a) the abstract, if present, or (b) its information
is required in a header or footer.  Use of these macros is optional,
except that @code{TL} is mandatory if any of @code{RP}, @code{AU},
@code{AI}, or @code{AB} is called, and @code{AE} is mandatory if
@code{AB} is called.

@Defmac {RP, [@code{no-repeat-info}] [@code{no-renumber}], ms}
Use the ``report'' (@acronym{AT&T}: ``released paper'') format for your
document, creating a separate cover page.  The default arrangement is to
place most of the document description (title, author names and
institutions, and abstract, but not the date) at the top of the first
page.  If the optional @code{no-repeat-info} argument is given,
@file{ms} produces a cover page but does not repeat any of its
information subsequently (but see the @code{DA} macro below regarding
the date).  Normally, @code{RP} sets the page number following the cover
page to@tie{}1.  Specifying the optional @code{no-renumber} argument
suppresses this alteration.  Optional arguments can occur in any order.
@code{no} is recognized as a synonym of @code{no-repeat-info} for
@code{AT&T} compatibility.
@endDefmac

@Defmac {TL, , ms}
Specify the document title.  @file{ms} collects text on input lines
following this call into the title until reaching @code{AU}, @code{AB},
or a heading or paragraphing macro call.
@endDefmac

@Defmac {AU, , ms}
Specify an author's name.  @file{ms} collects text on input lines
following this call into the author's name until reaching @code{AI},
@code{AB}, another @code{AU}, or a heading or paragraphing macro call.
Call it repeatedly to specify multiple authors.
@endDefmac

@Defmac {AI, , ms}
Specify the preceding author's institution.  An @code{AU} call is
usefully followed by at most one @code{AI} call; if there are more, the
last @code{AI} call controls.  @file{ms} collects text on input lines
following this call into the author's institution until reaching
@code{AU}, @code{AB}, or a heading or paragraphing macro call.
@endDefmac

@Defmac {DA, [@Var{x} @dots{}], ms}
Typeset the current date, or any arguments @var{x}, in the center
footer, and, if @code{RP} is also called, left-aligned at the end of the
description information on the cover page.
@endDefmac

@Defmac {ND, [@Var{x} @dots{}], ms}
Typeset the current date, or any arguments @var{x}, if @code{RP} is also
called, left-aligned at the end of the document description on the cover
page.  This is @code{groff} @file{ms}'s default.
@endDefmac

@Defmac {AB, [@code{no}], ms}
Begin the abstract.  @file{ms} collects text on input lines following
this call into the abstract until reaching an @code{AE} call.  By
default, @file{ms} places the word ``ABSTRACT'' centered and in italics
above the text of the abstract.  The optional argument @code{no}
suppresses this heading.
@endDefmac

@Defmac {AE, , ms}
End the abstract.
@endDefmac

An example document description, using a cover page, follows.
@cindex cover page in [@file{ms}], example markup
@cindex example markup, cover page in [@file{ms}]

@CartoucheExample
.RP
.TL
The Inevitability of Code Bloat
in Commercial and Free Software
.AU
J.\& Random Luser
.AI
University of West Bumblefuzz
.AB
This report examines the long-term growth of the code
bases in two large,
popular software packages;
the free Emacs and the commercial Microsoft Word.
While differences appear in the type or order of
features added,
due to the different methodologies used,
the results are the same in the end.
.PP
The free software approach is shown to be superior in
that while free software can become as bloated as
commercial offerings,
free software tends to have fewer serious bugs and the
added features are more in line with user demand.
.AE

@r{@dots{}the rest of the paper@dots{}}
@endCartoucheExample

@c ---------------------------------------------------------------------

@node ms Body Text, ms Page Layout, ms Document Description Macros, ms
@subsection Body Text
@cindex @file{ms} macros, body text

A variety of macros, registers, and strings can be used to structure and
style the body of your document.  They organize your text into
paragraphs, headings, footnotes, and inclusions of material such as
tables and figures.

@menu
* Text settings in ms::
* Typographical symbols in ms::
* Paragraphs in ms::
* Headings in ms::
* Typeface and decoration::
* Lists in ms::
* Indented regions in ms::
* ms keeps and displays::
* ms Insertions::
* ms Footnotes::
* ms language and localization::
@end menu

@c ---------------------------------------------------------------------

@node Text settings in ms, Typographical symbols in ms, ms Body Text, ms Body Text
@subsubsection Text settings
@cindex @file{ms} macros, text settings

The @code{FAM} string, a GNU extension, sets the font family for body
text; the default is @samp{T}.  The @code{PS} and @code{VS} registers
set the type size and vertical spacing (distance between text
baselines), respectively.  The font family and type size are ignored on
terminal devices.  Setting these parameters before the first call of a
heading, paragraphing, or (non-date) document description macro also
applies them to headers, footers, and (for @code{FAM}) footnotes.

Which font families are available depends on the output device; as a
convention, @code{T} selects a serif family (``Times''), @code{H} a
sans-serif family (``Helvetica''), and @code{C} a monospaced family
(``Courier'').  The man page for the output driver documents its font
repertoire.  Consult the @cite{groff@r{(1)}} man page for lists of
available output devices and their drivers.

The hyphenation mode (as used by the @code{hy} request) is set from the
@code{HY} register.  Setting @code{HY} to @samp{0} is equivalent to
using the @code{nh} request.  This is a Tenth Edition Research Unix
extension.

@c ---------------------------------------------------------------------

@node Typographical symbols in ms, Paragraphs in ms, Text settings in ms, ms Body Text
@subsubsection Typographical symbols
@cindex @file{ms} macros, obtaining typographical symbols

@file{ms} provides a few strings to obtain typographical symbols not
easily entered with the keyboard.  These and many others are available
as special character escape sequences---see the @cite{groff_char@r{(7)}}
man page.

@Defmpstr {-, ms}
Interpolate an em dash.
@endDefmpstr

@DefmpstrList {Q, ms}
@DefmpstrListEndx {U, ms}
Interpolate typographer's quotation marks where available, and neutral
double quotes otherwise.  @code{\*Q} is the left quote and @code{\*U}
the right.
@endDefmpstr

@c ---------------------------------------------------------------------

@node Paragraphs in ms, Headings in ms, Typographical symbols in ms, ms Body Text
@subsubsection Paragraphs
@cindex @file{ms} macros, paragraph handling

Paragraphing macros @dfn{break}, or terminate, any pending output line
so that a new paragraph can begin.  Several paragraph types are
available, differing in how indentation applies to them: to left, right,
or both margins; to the first output line of the paragraph, all output
lines, or all but the first.  All paragraphing macro calls cause the
insertion of vertical space in the amount stored in the @code{PD}
register, except at page or column breaks.  Alternatively, a blank input
line breaks the output line and vertically spaces by one vee.

@Defmac {LP, , ms}
Set a paragraph without any (additional) indentation.
@endDefmac

@Defmac {PP, , ms}
Set a paragraph with a first-line left indentation in the amount stored
in the @code{PI} register.
@endDefmac

@Defmac {IP, [@Var{marker} [@Var{width}]], ms}
Set a paragraph with a left indentation.  The optional @var{marker} is
not indented and is empty by default.  It has several applications;
see @ref{Lists in ms}.  @var{width} overrides the indentation amount
stored in the @code{PI} register; its default unit is @samp{n}.  Once
specified, @var{width} applies to further @code{IP} calls until
specified again or a heading or different paragraphing macro is called.
@endDefmac

@Defmac {QP, , ms}
Set a paragraph indented from both left and right margins by the amount
stored in the @code{QI} register.
@endDefmac

@DefmacList {QS, , ms}
@DefmacListEndx {QE, , ms}
Begin (@code{QS}) and end (@code{QE}) a region where each paragraph is
indented from both margins by the amount stored in the @code{QI}
register.  The text between @code{QS} and @code{QE} can be structured
further by use of other paragraphing macros.
@endDefmac

@Defmac {XP, , ms}
Set an ``exdented'' paragraph---one with a left indentation in the
amount stored in the @code{PI} register on every line @emph{except} the
first (also known as a hanging indent).  This is a Berkeley extension.
@endDefmac

The following example illustrates the use of paragraphing macros.

@CartoucheExample
.NH 2
Cases used in the 2001 study
.LP
Two software releases were considered for this report.
.PP
The first is commercial software;
the second is free.
.IP \[bu]
Microsoft Word for Windows,
starting with version 1.0 through the current version
(Word 2000).
.IP \[bu]
GNU Emacs,
from its first appearance as a standalone editor through
the current version (v20).
See [Bloggs 2002] for details.
.QP
Franklin's Law applied to software:
software expands to outgrow both RAM and disk space over
time.
.SH
Bibliography
.XP
Bloggs, Joseph R.,
.I "Everyone's a Critic" ,
Underground Press, March 2002.
A definitive work that answers all questions and
criticisms about the quality and usability of free
software.
@endCartoucheExample

@c ---------------------------------------------------------------------

@node Headings in ms, Typeface and decoration, Paragraphs in ms, ms Body Text
@subsubsection Headings
@cindex @file{ms} macros, headings

Use headings to create a sequential or hierarchical structure for your
document.  The @file{ms} macros print headings in @strong{bold} using
the same font family and, by default, type size as the body text.
Headings are available with and without automatic numbering.  Text on
input lines following the macro call becomes the heading's title.  Call
a paragraphing macro to end the heading text and start the section's
content.

@DefmacList {NH, [@Var{depth}], ms}
@DefmacListEnd {NH, @t{S} @Var{heading-depth-index} @dots{}, ms}
Set an automatically numbered heading.

@file{ms} produces a numbered heading the form @var{a.b.c@dots{}}, to
any depth desired, with the numbering of each depth increasing
automatically and being reset to zero when a more significant level is
increased.  ``1''@tie{}is the most significant or coarsest division of
the document.  Only nonzero values are output.  If @var{depth} is
omitted, it is taken to be @samp{1}.

If you specify @var{depth} such that an ascending gap occurs relative to
the previous @code{NH} call---that is, you ``skip a depth'', as by
@samp{.NH 1} and then @samp{.NH 3}---@code{groff} @file{ms} emits a
warning on the standard error stream.

Alternatively, you can give @code{NH} a first argument of@tie{}@code{S},
followed by integers to number the heading depths explicitly.  Further
automatic numbering, if used, resumes using the specified indices as
their predecessors.
@c Although undocumented in Tuthill's 4.2BSD ms.diffs paper...
This feature is a Berkeley extension.
@endDefmac

An example may be illustrative.

@CartoucheExample
.NH 1
Animalia
.NH 2
Arthropoda
.NH 3
Crustacea
.NH 2
Chordata
.NH S 6 6 6
Daimonia
.NH 1
Plantae
@endCartoucheExample

The above results in numbering as follows; the vertical space that
normally precedes each heading is omitted.

@Example
1.  Animalia
1.1.  Arthropoda
1.1.1.  Crustacea
1.2.  Chordata
6.6.6.  Daimonia
7.  Plantae
@endExample

@DefmpstrList {SN-STYLE, ms}
@DefmpstrItemx {SN-DOT, ms}
@DefmpstrItemx {SN-NO-DOT, ms}
@DefmpstrListEndx {SN, ms}
After @code{NH} is called, the assigned number is made available in the
strings @code{SN-DOT} (as it appears in a printed heading with default
formatting, followed by a terminating period) and @code{SN-NO-DOT} (with
the terminating period omitted).  These are GNU extensions.

You can control the style used to print numbered headings by defining an
appropriate alias for the string @code{SN-STYLE}.  By default,
@code{SN-STYLE} is aliased to @code{SN-DOT}.  If you prefer to omit the
terminating period from numbers appearing in numbered headings, you may
define the alias as follows.

@Example
.als SN-STYLE SN-NO-DOT
@endExample

@noindent
Any such change in numbering style becomes effective from the next use
of @code{NH} following redefinition of the alias for @code{SN-STYLE}.
The formatted number of the current heading is available in the
@code{SN} string (a feature first documented by Berkeley), which
facilitates its inclusion in, for example, table captions, equation
labels, and @code{XS}/@code{XA}/@code{XE} table of contents entries.
@endDefmpstr

@Defmac {SH, [@Var{depth}], ms}
Set an unnumbered heading.

The optional @var{depth} argument is a GNU extension indicating the
heading depth corresponding to the @var{depth} argument of @code{NH}.
It matches the type size at which the heading is set to that of a
numbered heading at the same depth when the @code{GROWPS} and
@code{PSINCR} heading size adjustment mechanism is in effect.
@endDefmac

If the @code{GROWPS} register is set to a value greater than the
@var{level} argument to @code{NH} or @code{SH}, the type size of a
heading produced by these macros increases by @code{PSINCR} units over
the size specified by @code{PS} multiplied by the difference of
@code{GROWPS} and @var{level}.  The value stored in @code{PSINCR} is
interpreted in @code{groff} basic units; the @code{p} scaling unit
should be employed when assigning a value specified in points.  For
example, the sequence

@CartoucheExample
.nr PS 10
.nr GROWPS 3
.nr PSINCR 1.5p
.NH 1
Carnivora
.NH 2
Felinae
.NH 3
Felis catus
.SH 2
Machairodontinae
@endCartoucheExample

@noindent
will cause ``1. Carnivora'' to be printed in 13-point text, followed by
``1.1. Felinae'' in 11.5-point text, while ``1.1.1. Felis catus'' and
all more deeply nested heading levels will remain in the 10-point text
specified by the @code{PS} register.  ``Machairodontinae'' is printed at
11.5 points, since it corresponds to heading level@tie{}2.

The @code{HORPHANS} register operates in conjunction with the @code{NH}
and @code{SH} macros to inhibit the printing of isolated headings at the
bottom of a page; it specifies the minimum number of lines of an
immediately subsequent paragraph that must be kept on the same page as
the heading.  If insufficient space remains on the current page to
accommodate the heading and this number of lines of paragraph text, a
page break is forced before the heading is printed.  Any display macro
call or @code{tbl}, @code{pic}, or @code{eqn} region between the heading
and the subsequent paragraph suppresses this grouping.  @xref{ms keeps
and displays} and @ref{ms Insertions}.

@c ---------------------------------------------------------------------

@node Typeface and decoration, Lists in ms, Headings in ms, ms Body Text
@subsubsection Typeface and decoration

The @file{ms} macros provide a variety of ways to style text.
Attend closely to the ordering of arguments labeled @var{pre} and
@var{post}, which is not intuitive.  Support for @var{pre}
arguments is a GNU extension.@footnote{This idiosyncrasy arose through
feature accretion; for example, the @code{B} macro in Version@tie{}6
Unix @file{ms} (1975) accepted only one argument, the text to be set in
boldface.  By Version@tie{}7 (1979) it recognized a second argument; in
1990, @code{groff} @file{ms} added a ``pre'' argument, placing it third
to avoid breaking support for older documents.}

@Defmac {B, [@Var{text} [@Var{post} [@Var{pre}]]], ms}
Style @var{text} in @b{bold}, followed by @var{post} in the previous
font style without intervening space, and preceded by @var{pre}
similarly.  Without arguments, @file{ms} styles subsequent text in bold
until the next paragraphing, heading, or no-argument typeface macro
call.
@endDefmac

@Defmac {R, [@Var{text} [@Var{post} [@Var{pre}]]], ms}
As @code{B}, but use the roman style (upright text of normal weight)
instead of bold.  Argument recognition is a GNU extension.
@endDefmac

@Defmac {I, [@Var{text} [@Var{post} [@Var{pre}]]], ms}
As @code{B}, but use an @i{italic} or oblique style instead of bold.
@endDefmac

@Defmac {BI, [@Var{text} [@Var{post} [@Var{pre}]]], ms}
As @code{B}, but use a bold italic or bold oblique style instead of
upright bold.  This is a Tenth Edition Research Unix extension.
@c possibly 9th, but definitely not Berkeley
@endDefmac

@Defmac {CW, [@Var{text} [@Var{post} [@Var{pre}]]], ms}
As @code{B}, but use a @t{constant-width} (monospaced) roman typeface
instead of bold.  This is a Tenth Edition Research Unix extension.
@c possibly 9th, but definitely not Berkeley
@endDefmac

@Defmac {BX, [@Var{text}], ms}
Typeset @var{text} and draw a box around it.  On terminal devices,
reverse video is used instead.  If you want @var{text} to contain space,
use unbreakable space or horizontal motion escape sequences (@code{\~},
@code{\@key{SP}}, @code{\^}, @code{\|}, @code{\0} or @code{\h}).
@endDefmac

@Defmac {UL, [@Var{text} [@Var{post}]], ms}
Typeset @var{text} with an underline.  @var{post}, if present, is set
after @var{text} with no intervening space.
@endDefmac

@Defmac {LG, , ms}
Set subsequent text in larger type (two points larger than the
current size) until the next type size, paragraphing, or heading macro
call.  You can specify this macro multiple times to enlarge the type
size as needed.
@endDefmac

@Defmac {SM, , ms}
Set subsequent text in smaller type (two points smaller than the current
size) until the next type size, paragraphing, or heading macro call.
You can specify this macro multiple times to reduce the type size as
needed.
@endDefmac

@Defmac {NL, , ms}
Set subsequent text at the normal type size (the amount in the @code{PS}
register).
@endDefmac

@var{pre} and @var{post} arguments are typically used to simplify the
attachment of punctuation to styled words.  When @var{pre} is used,
a hyphenation control escape sequence @code{\%} that would ordinarily
start @var{text} must start @var{pre} instead to have the desired
effect.

@CartoucheExample
The CS course's students found one C language keyword
.CW static ) \%(
most troublesome.
@endCartoucheExample

The foregoing example produces output as follows.

@CartoucheExample
@r{The CS course's students found one C language keyword (@t{static})
most troublesome.}
@endCartoucheExample

You can use the output line continuation escape sequence @code{\c} to
achieve the same result (@pxref{Line Continuation}).  It is also
portable to older @file{ms} implementations.

@CartoucheExample
The CS course's students found one C language keyword
\%(\c
.CW \%static )
most troublesome.
@endCartoucheExample

@code{groff} @file{ms} also offers strings to begin and end super- and
subscripting.  These are GNU extensions.

@DefmpstrList {@lbracechar{}, ms}
@DefmpstrListEndx {@rbracechar{}, ms}
Begin and end superscripting, respectively.
@endDefmpstr

@DefmpstrList {<, ms}
@DefmpstrListEndx {>, ms}
Begin and end subscripting, respectively.
@endDefmpstr

Rather than calling the @code{CW} macro, in @code{groff} @file{ms} you
might prefer to change the font family to Courier by setting the
@code{FAM} string to @samp{C}.  You can then use all four style macros
above, returning to the default family (Times) with @samp{.ds FAM T}.
Because changes to @code{FAM} take effect only at the next paragraph,
@code{CW} remains useful to ``inline'' a change to the font family,
similarly to the practice of this document in noting syntactical
elements of @file{ms} and @code{groff}.

@c ---------------------------------------------------------------------

@node Lists in ms, Indented regions in ms, Typeface and decoration, ms Body Text
@subsubsection Lists
@cindex @file{ms} macros, lists

The @var{marker} argument to the @code{IP} macro can be employed to
present a variety of lists; for instance, you can use a bullet glyph
(@code{\[bu]}) for unordered lists, a number (or auto-incrementing
register) for numbered lists, or a word or phrase for glossary-style or
definition lists.  If you set the paragraph indentation register
@code{PI} before calling @code{IP}, you can later reorder the items in
the list without having to ensure that a @var{width} argument remains
affixed to the first call.

The following is an example of a bulleted list.
@cindex example markup, bulleted list [@file{ms}]
@cindex bulleted list, example markup [@file{ms}]

@CartoucheExample
.nr PI 2n
A bulleted list:
.IP \[bu]
lawyers
.IP \[bu]
guns
.IP \[bu]
money
@endCartoucheExample

@Example
A bulleted list:

@bullet{} lawyers

@bullet{} guns

@bullet{} money
@endExample

The following is an example of a numbered list.
@cindex example markup, numbered list [@file{ms}]
@cindex numbered list, example markup [@file{ms}]

@CartoucheExample
.nr step 0 1
.nr PI 3n
A numbered list:
.IP \n+[step]
lawyers
.IP \n+[step]
guns
.IP \n+[step]
money
@endCartoucheExample

@Example
A numbered list:

1. lawyers

2. guns

3. money
@endExample

Here we have employed the @code{nr} request to create a register of our
own, @samp{step}.  We initialized it to zero and assigned it an
auto-increment of 1.  Each time we use the escape sequence
@samp{\n+[PI]} (note the plus sign), the formatter applies the increment
just before interpolating the register's value.  Preparing the @code{PI}
register as well enables us to rearrange the list without the tedium of
updating macro calls.

The next example illustrates a glossary-style list.
@cindex example markup, glossary-style list [@file{ms}]
@cindex glossary-style list, example markup [@file{ms}]

@CartoucheExample
A glossary-style list:
.IP lawyers 0.4i
Two or more attorneys.
.IP guns
Firearms,
preferably large-caliber.
.IP money
Gotta pay for those
lawyers and guns!
@endCartoucheExample

@Example
A glossary-style list:

lawyers
      Two or more attorneys.

guns  Firearms, preferably large-caliber.

money
      Gotta pay for those lawyers and guns!
@endExample

In the previous example, observe how the @code{IP} macro places the
definition on the same line as the term if it has enough space.  If this
is not what you want, there are a few workarounds we will illustrate by
modifying the example.  First, you can use a @code{br} request to force
a break after printing the term or label.

@CartoucheExample
.IP guns
.br
Firearms,
@endCartoucheExample

Second, you could apply the @code{\p} escape sequence to force a break.
The space following the escape sequence is important; if you omit it,
@code{groff} prints the first word of the paragraph text on the same
line as the term or label (if it fits) @emph{then} breaks the line.

@CartoucheExample
.IP guns
\p Firearms,
@endCartoucheExample

Finally, you may append a horizontal motion to the marker with the
@code{\h} escape sequence; using the same amount as the indentation will
ensure that the marker is too wide for @code{groff} to treat it as
``fitting'' on the same line as the paragraph text.

@CartoucheExample
.IP guns\h'0.4i'
Firearms,
@endCartoucheExample

In each case, the result is the same.

@Example
A glossary-style list:

lawyers
      Two or more attorneys.

guns
      Firearms, preferably large-caliber.

money
      Gotta pay for those lawyers and guns!
@endExample

@c ---------------------------------------------------------------------

@node Indented regions in ms, ms keeps and displays, Lists in ms, ms Body Text
@subsubsection Indented regions

You may need to indent a region of text while otherwise formatting it
normally.  Indented regions can be nested; you can change @code{\n[PI]}
before each call to vary the amount of inset.

@Defmac {RS, , ms}
Begin a region where headings, paragraphs, and displays are indented
(further) by the amount stored in the @code{PI} register.
@endDefmac

@Defmac {RE, , ms}
End the (next) most recent indented region.
@endDefmac

This feature enables you to easily line up text under hanging and
indented paragraphs.
@cindex @file{ms} macros, nested lists
@cindex nested lists [@file{ms}]
For example, you may wish to structure lists hierarchically.

@CartoucheExample
.IP \[bu] 2
Lawyers:
.RS
.IP \[bu]
Dewey,
.IP \[bu]
Cheatham,
and
.IP \[bu]
and Howe.
.RE
.IP \[bu]
Guns
@endCartoucheExample

@Example
@bullet{} Lawyers:

  @bullet{}  Dewey,

  @bullet{}  Cheatham, and

  @bullet{}  Howe.

@bullet{} Guns
@endExample

@c ---------------------------------------------------------------------

@node ms keeps and displays, ms Insertions, Indented regions in ms, ms Body Text
@subsubsection Keeps, boxed keeps, and displays
@cindex @file{ms} macros, displays
@cindex @file{ms} macros, keeps
@cindex keeps [@file{ms}]

On occasion, you may want to @dfn{keep} several lines of text, or a
region of a document, together on a single page, preventing an automatic
page break within certain boundaries.  This can cause a page break to
occur earlier than it normally would.  For example, you may want to keep
two paragraphs together, or a paragraph that refers to a table, list, or
figure adjacent to the item it discusses.  @file{ms} provides the
@code{KS} and @code{KE} macros for this purpose.

You can alternatively specify a @dfn{floating keep}:@: if a keep cannot
fit on the current page, @file{ms} holds its contents and
allows material following the keep (in the source document) to fill the
remainder of the current page.  When the page breaks, whether by
reaching the end or @code{bp} request, @file{ms} puts the floating keep
at the beginning of the next page.  This is useful for placing large
graphics or tables that do not need to appear exactly where they occur
in the source document.

@DefmacList {KS, , ms}
@DefmacItemx {KF, , ms}
@DefmacListEndx {KE, , ms}
@code{KS} begins a keep, @code{KF} a floating keep, and @code{KE} ends a
keep of either kind.
@endDefmac

As an alternative to the keep mechanism, the @code{ne} request forces a
page break if there is not at least the amount of vertical space
specified in its argument remaining on the page (@pxref{Page Control}).
One application of @code{ne} is to reserve space on the page for a
figure or illustration to be included later.

@cindex boxes [@file{ms}]
A @dfn{boxed keep} has a frame drawn around it.

@DefmacList {B1, , ms}
@DefmacListEndx {B2, , ms}
@code{B1} begins a keep with a box drawn around it.  @code{B2} ends a
boxed keep.
@endDefmac

Boxed keep macros cause breaks; if you need to box a word or phrase
within a line, see the @code{BX} macro in @ref{Typeface and decoration}.
Box lines are drawn as close as possible to the text they enclose so
that they are usable within paragraphs.  If you wish to box one or more
paragraphs, you may improve the appearance by calling @code{B1} after
the first paragraphing macro, and by adding a small amount of vertical
space before calling @code{B2}.

@c Wrap example at 58 columns.
@CartoucheExample
.LP
.B1
.I Warning:
Happy Fun Ball may suddenly accelerate to dangerous
speeds.
.sp \n[PD]/2 \" space by half the inter-paragraph distance
.B2
@endCartoucheExample

If you want a boxed keep to float, you will need to enclose the
@code{B1} and @code{B2} calls within a pair of @code{KF} and @code{KE}
calls.

@cindex displays [@file{ms}]
@dfn{Displays} turn off filling; lines of verse or program code are
shown with their lines broken as in the source document without
requiring @code{br} requests between lines.  Displays can be kept on a
single page or allowed to break across pages.  The @code{DS} macro
begins a kept display of the layout specified in its first argument;
non-kept displays are begun with dedicated macros corresponding to their
layout.

@DefmacList {DS, @t{L}, ms}
@DefmacListEndx {LD, , ms}
Begin (@code{DS}:@: kept) left-aligned display.
@endDefmac

@DefmacList {DS, [@t{I} [@Var{indent}]], ms}
@DefmacListEndx {ID, [@Var{indent}], ms}
Begin (@code{DS}:@: kept) display indented by @var{indent} if specified,
and by the amount of the @code{DI} register otherwise.
@endDefmac

@DefmacList {DS, @t{B}, ms}
@DefmacListEndx {BD, , ms}
Begin a (@code{DS}:@: kept) a block display:@: the entire display is
left-aligned, but indented such that the longest line in the display
is centered on the page.
@endDefmac

@DefmacList {DS, @t{C}, ms}
@DefmacListEndx {CD, , ms}
Begin a (@code{DS}:@: kept) centered display:@: each line in the display
is centered.
@endDefmac

@DefmacList {DS, @t{R}, ms}
@DefmacListEndx {RD, , ms}
Begin a (@code{DS}:@: kept) right-aligned display.  This is a GNU
extension.
@endDefmac

@Defmac {DE, , ms}
End any display.
@endDefmac

The distance stored in the @code{DD} register is inserted before and
after each pair of display macros; this is a Berkeley extension.  In
@code{groff} @file{ms}, this distance replaces any adjacent
inter-paragraph distance or subsequent spacing prior to a section
heading.  The @code{DI} register is a GNU extension; its value is an
indentation applied to displays created with @samp{.DS} and @samp{.ID}
without arguments, to @samp{.DS I} without an indentation argument, and
to indented equations set with @samp{.EQ}.  Changes to either register
take effect at the next display boundary.

@c ---------------------------------------------------------------------

@node ms Insertions, ms Footnotes, ms keeps and displays, ms Body Text
@subsubsection Tables, figures, equations, and references
@cindex @file{ms} macros, tables
@cindex @file{ms} macros, figures
@cindex @file{ms} macros, equations
@cindex @file{ms} macros, references
@cindex tables [@file{ms}]
@cindex figures [@file{ms}]
@cindex equations [@file{ms}]
@cindex references [@file{ms}]

The @file{ms} package is often used with the @code{tbl}, @code{pic},
@code{eqn}, and @code{refer} preprocessors.
@pindex tbl
@pindex pic
@pindex eqn
@pindex refer
Mark text meant for preprocessors by enclosing it in pairs of tokens
as follows, with nothing between the dot and the macro name.  The
preprocessors match these tokens only at the start of an input line.

@DefmacList {TS, [@code{H}], ms}
@DefmacListEndx {TE, , ms}
Demarcate a table to be processed by the @code{tbl} preprocessor.  The
optional argument@tie{}@code{H} to @code{TS} instructs @file{ms} to
repeat table rows (often column headings) at the top of each new page
the table spans, if applicable; calling the @code{TH} macro marks the
end of such rows.  The GNU @cite{tbl@r{(1)}} man page provides a
comprehensive reference to the preprocessor and offers examples of its
use.
@endDefmac

@DefmacList {PS, , ms}
@DefmacItemx {PE, , ms}
@DefmacListEndx {PF, , ms}
@code{PS} begins a picture to be processed by the @command{gpic}
preprocessor; either of @code{PE} or @code{PF} ends it, the latter with
``flyback'' to the vertical position at its top.  You can create
@code{pic} input manually or with a program such as @code{xfig}.
@endDefmac

@DefmacList {EQ, [@Var{align} [@Var{label}]], ms}
@DefmacListEndx {EN, , ms}
Demarcate an equation to be processed by the @code{eqn} preprocessor.
The equation is centered by default; @var{align} can be @samp{C},
@samp{L}, or @samp{I} to (explicitly) center, left-align, or indent it
by the amount stored in the @code{DI} register, respectively.  If
specified, @var{label} is set right-aligned.
@endDefmac

@DefmacList {[, , ms}
@DefmacListEndx {], , ms}
Demarcate a bibliographic citation to be processed by the @code{refer}
preprocessor.  The GNU @cite{refer@r{(1)}} man page provides a
comprehensive reference to the preprocessor and the format of its
bibliographic database.  Type @samp{man refer} at the command line to
view it.
@endDefmac

When @code{refer} emits collected references (as might be done on a
``Works Cited'' page), it interpolates the @code{REFERENCES} string as
an unnumbered heading (@code{SH}).

@cindex table, multi-page, example [@file{ms}]
@cindex multi-page table example [@file{ms}]
The following is an example of how to set up a table that may print
across two or more pages.

@CartoucheExample
.TS H
allbox;
Cb | Cb .
Part@arrow{}Description
_
.TH
.T&
GH-1978@arrow{}Fribulating gonkulator
@r{@dots{}the rest of the table follows@dots{}}
.TE
@endCartoucheExample

@noindent
Attempting to place a multi-page table inside a keep can lead to
unpleasant results, particularly if the @code{tbl} @code{allbox} option
is used.

@cindex equation example [@file{ms}]
Mathematics can be typeset using the language of the @code{eqn}
preprocessor.

@CartoucheExample
.EQ C (\*[SN-NO-DOT]a)
p ~ = ~ q sqrt @{ ( 1 + ~ ( x / q sup 2 ) @}
.EN
@endCartoucheExample

@noindent
This input formats a labelled equation.  We used the @code{SN-NO-DOT}
string to base the equation label on the current heading number, giving
us more flexibility to reorganize the document.

Use @command{groff} options to run preprocessors on the input:@:
@option{-e} for @command{geqn}, @option{-p} for @command{gpic},
@option{-R} for @command{grefer}, and @option{-t} for @command{gtbl}.

@c ---------------------------------------------------------------------

@node ms Footnotes, , ms Insertions, ms Body Text
@subsubsection Footnotes
@cindex @file{ms} macros, footnotes
@cindex footnotes [@file{ms}]

@cindex footnote marker [@file{ms}]
@cindex marker, footnote [@file{ms}]
A footnote is typically anchored to a place in the text with a
@dfn{marker}, which is a small integer, a symbol such as a dagger, or
arbitrary user-specified text.

@Defmpstr {*, ms}
Place an @dfn{automatic number}, an automatically generated numeric
footnote marker, in the text.  Each time this string is interpolated,
the number it produces increments by one.  Automatic numbers start at 1.
This is a Berkeley extension.
@endDefesc

Enclose the footnote text in @code{FS} and @code{FE} macro calls to set
it at the nearest available ``foot'', or bottom, of a text column or
page.

@DefmacList {FS, [@Var{marker}], ms}
@DefmacListEndx {FE, , ms}
Begin (@code{FS}) and end (@code{FE}) a footnote.  @code{FS} calls
@code{FS-MARK} with any supplied @var{marker} argument, which is then
also placed at the beginning of the footnote text.  If @var{marker} is
omitted, the next pending automatic footnote number enqueued by
interpolation of the @code{*} string is used, and if none exists,
nothing is prefixed.
@endDefmac

You may not desire automatically numbered footnotes in spite of their
convenience.  You can indicate a footnote with a symbol or other text by
specifying its marker at the appropriate place (for example, by using
@code{\[dg]} for the dagger glyph) @emph{and} as an argument to the
@code{FS} macro.  Such manual marks should be repeated as arguments to
@code{FS} or as part of the footnote text to disambiguate their
correspondence.  You may wish to use @code{\*@{} and @code{\*@}} to
superscript the marker at the anchor point, in the footnote text, or
both.

@code{groff} @file{ms} provides a hook macro, @code{FS-MARK}, for
user-determined operations to be performed when the @code{FS} macro is
called.  It is passed the same arguments as @code{FS} itself.  An
application of @code{FS-MARK} is anchor placement for a hyperlink
reference, so that a footnote can link back to its referential
context.@footnote{``Portable Document Format Publishing with GNU
Troff'', @file{pdfmark.ms} in the @code{groff} distribution, uses this
technique.}  By default, this macro has an empty definition.
@code{FS-MARK} is a GNU extension.

@cindex footnotes, and keeps [@file{ms}]
@cindex keeps, and footnotes [@file{ms}]
@cindex footnotes, and displays [@file{ms}]
@cindex displays, and footnotes [@file{ms}]
Footnotes can be safely used within keeps and displays, but you should
avoid using automatically numbered footnotes within floating keeps.  You
can place a second @code{\**} interpolation between a @code{\**} and its
corresponding @code{FS} call as long as each @code{FS} call occurs
@emph{after} the corresponding @code{\**} and occurrences of @code{FS}
are in the same order as corresponding occurrences of @code{\**}.

Footnote text is formatted as paragraphs are, using analogous
parameters.  The registers @code{FI}, @code{FPD}, @code{FPS}, and
@code{FVS} correspond to @code{PI}, @code{PD}, @code{PS}, and @code{CS},
respectively; @code{FPD}, @code{FPS}, and @code{FVS} are GNU extensions.

The @code{FF} register controls the formatting of automatically numbered
footnote paragraphs and those for which @code{FS} is given a marker
argument.  @xref{ms Document Control Settings}.

The default footnote line length is 11/12ths of the normal line length
for compatibility with the expectations of historical @file{ms}
documents; you may wish to set the @code{FR} string to @samp{1} to align
with contemporary typesetting practices.  In the
past,@footnote{Unix Version@tie{}7 @file{ms}, its descendants, and GNU
@file{ms} prior to @code{groff} version 1.23.0} an @code{FL} register
was used for the line length in footnotes; however, setting this
register at document initialization time had no effect on the footnote
line length in multi-column arrangements.@footnote{You could reset it
after each call to @code{.1C}, @code{.2C}, or @code{.MC}.}

@code{FR} should be used in preference to the old @code{FL} register in
contemporary documents.  The footnote line length is effectively
computed as @samp{@slanted{column-width} * \*[FR]}.  If an absolute
footnote line length is required, recall that arithmetic expressions in
@code{roff} input are evaluated strictly from left to right, with no
operator precedence (parentheses are honored).

@Example
.ds FR 0+3i \" Set footnote line length to 3 inches.
@endExample

@c ---------------------------------------------------------------------

@node ms language and localization, ms Page Layout, ms Footnotes, ms Body Text
@subsubsection Language and localization
@cindex @file{ms} macros, language
@cindex @file{ms} macros, localization
@cindex language [@file{ms}]
@cindex localization [@file{ms}]

@code{groff} @file{ms} provides several strings that you can customize
for your own purposes, or redefine to adapt the macro package to
languages other than English.  It is already localized for
@c cs, de, fr, it, sv
Czech, German, French, Italian, and Swedish.  Load the desired
localization macro package after @file{ms}; see the
@cite{groff_tmac@r{(5)}} man page.

@CartoucheExample
$ groff -ms -mfr bienvenue.ms
@endCartoucheExample

The following strings are available.

@Defmpstr {REFERENCES, ms}
Contains the string printed at the beginning of a references
(bibliography) page produced with GNU @cite{refer@r{(1)}}.  The default
is @samp{References}.
@c XXX: Use of refer(1) with ms is insufficiently documented.
@endDefmpstr

@Defmpstr {ABSTRACT, ms}
Contains the string printed at the beginning of the abstract.  The
default is @samp{\f[I]ABSTRACT\f[]}; it includes font selection escape
sequences to set the word in italics.
@endDefmpstr

@Defmpstr {TOC, ms}
Contains the string printed at the beginning of the table of contents.
The default is @samp{Table of Contents}.
@endDefmpstr

@DefmpstrList {MONTH1, ms}
@DefmpstrItemx {MONTH2, ms}
@DefmpstrItemx {MONTH3, ms}
@DefmpstrItemx {MONTH4, ms}
@DefmpstrItemx {MONTH5, ms}
@DefmpstrItemx {MONTH6, ms}
@DefmpstrItemx {MONTH7, ms}
@DefmpstrItemx {MONTH8, ms}
@DefmpstrItemx {MONTH9, ms}
@DefmpstrItemx {MONTH10, ms}
@DefmpstrItemx {MONTH11, ms}
@DefmpstrListEndx {MONTH12, ms}
Contain the full names of the calendar months.  The defaults are in
English: @samp{January}, @samp{February}, and so on.
@endDefmpstr

@c ---------------------------------------------------------------------

@node ms Page Layout, Differences from AT&T ms, ms Body Text, ms
@subsection Page layout
@cindex @file{ms} macros, page layout
@cindex page layout [@file{ms}]

@file{ms}'s default page layout arranges text in a single column with
the page number between hyphens centered in a header on each page except
the first, and produces no footers.  You can customize this arrangement.

@menu
* ms Headers and Footers::
* Tab Stops in ms::
* ms Margins::
* ms Multiple Columns::
* ms TOC::
@end menu

@c ---------------------------------------------------------------------

@node ms Headers and Footers, Tab Stops in ms, ms Page Layout, ms Page Layout
@subsubsection Headers and footers
@cindex @file{ms} macros, headers
@cindex @file{ms} macros, footers
@cindex headers [@file{ms}]
@cindex footers [@file{ms}]

There are multiple ways to produce headers and footers.  One is to
define the strings @code{LH}, @code{CH}, and @code{RH} to set the left,
center, and right headers, respectively; and @code{LF}, @code{CF}, and
@code{RF} to set the left, center, and right footers.  This approach
suffices for documents that do not distinguish odd- and even-numbered
pages.

Another method is to call macros that set headers or footers for odd- or
even-numbered pages.  Each such macro takes a delimited argument
separating the left, center, and right header or footer texts from each
other.  You can replace the neutral apostrophes (@code{'}) shown below
with any character not appearing in the header or footer text.  These
macros are Berkeley extensions.

@DefmacList {OH, @code{'}@Var{left}@code{'}@Var{center}@code{'}@Var{right}@code{'}, ms}
@DefmacItemx {EH, @code{'}@Var{left}@code{'}@Var{center}@code{'}@Var{right}@code{'}, ms}
@DefmacItemx {OF, @code{'}@Var{left}@code{'}@Var{center}@code{'}@Var{right}@code{'}, ms}
@DefmacListEndx {EF, @code{'}@Var{left}@code{'}@Var{center}@code{'}@Var{right}@code{'}, ms}
The @code{OH} and @code{EH} macros define headers for odd- (recto)
and even-numbered (verso) pages, respectively; the @code{OF} and
@code{EF} macros define footers for them.
@endDefmac

With either method, a percent sign @code{%} in header or footer text is
replaced by the current page number.  By default, @file{ms} places no
header on a page numbered ``1'' (regardless of its number format).

@Defmac {P1, , ms}
Typeset the header even on page@tie{}1.  To be effective, this macro
must be called before the header trap is sprung on any page numbered
``1''; in practice, unless your page numbering is unusual, this means
that you should call it early, before @code{TL} or any heading or
paragraphing macro.  This is a Berkeley extension.
@endDefmac

For even greater flexibility, @file{ms} is designed to permit the
redefinition of the macros that are called when the @code{groff} traps
that ordinarily cause the headers and footers to be output are sprung.
@code{PT} (``page trap'') is called by @file{ms} when the header is to
be written, and @code{BT} (``bottom trap'') when the footer is to be.
The @code{groff} page location trap that @file{ms} sets up to format the
header also calls the (normally undefined) @code{HD} macro after
@code{PT}; you can define @code{HD} if you need additional processing
after setting the header (for example, to draw a line below it).
@c Although undocumented in Tuthill's 4.2BSD ms.diffs paper...
The @code{HD} hook is a Berkeley extension.  Any such macros you
(re)define must implement any desired specialization for odd-, even-, or
first numbered pages.

@c ---------------------------------------------------------------------

@node Tab Stops in ms, ms Margins, ms Headers and Footers, ms Page Layout
@subsubsection Tab stops

Use the @code{ta} request to define tab stops as needed.  @xref{Tabs and
Fields}.

@Defmac {TA, , ms}
Reset the tab stops to the @file{ms} default (every 5 ens).
Redefine this macro to create a different set of default tab stops.
@endDefmac

@c ---------------------------------------------------------------------

@node ms Margins, ms Multiple Columns, Tab Stops in ms, ms Page Layout
@subsubsection Margins
@cindex @file{ms} macros, margins

Control margins using the registers summarized in ``Margin settings'' in
@ref{ms Document Control Settings} above.  There is no setting for the
right margin; the combination of page offset @code{\n[PO]} and line
length @code{\n[LL]} determines it.

@c ---------------------------------------------------------------------

@node ms Multiple Columns, ms TOC, ms Margins, ms Page Layout
@subsubsection Multiple columns
@cindex @file{ms} macros, multiple columns
@cindex multiple columns [@file{ms}]

@file{ms} can set text in as many columns as reasonably fit on the page.
The following macros force a page break if a multi-column layout is
active when they are called.  The @code{MINGW} register stores the
default minimum gutter width; it is a GNU extension.  When multiple
columns are in use, keeps and the @code{HORPHANS} and @code{PORPHANS}
registers work with respect to column breaks instead of page breaks.

@Defmac {1C, , ms}
Arrange page text in a single column (the default).
@endDefmac

@Defmac {2C, , ms}
Arrange page text in two columns.
@endDefmac

@Defmac {MC, [@Var{column-width} [@Var{gutter-width}]], ms}
Arrange page text in multiple columns.  If you specify no arguments, it
is equivalent to the @code{2C} macro.  Otherwise, @var{column-width} is
the width of each column and @var{gutter-width} is the minimum distance
between columns.
@endDefmac

@c ---------------------------------------------------------------------

@node ms TOC, Differences from AT&T ms, ms Multiple Columns, ms Page Layout
@subsubsection Creating a table of contents
@cindex @file{ms} macros, creating table of contents
@cindex table of contents, creating [@file{ms}]

Because @code{roff} formatters process their input in a single pass,
material on page 50, for example, cannot influence what appears on
page@tie{}1---this poses a challenge for a table of contents at its
traditional location in front matter, if you wish to avoid manually
maintaining it.  @file{ms} enables the collection of material to be
presented in the table of contents as it appears, saving its page number
along with it, and then emitting the collected contents on demand toward
the end of the document.  The table of contents can then be resequenced
to its desired location as part of post-processing---with a PDF page
relocation tool, or by physically rearranging the pages of a printed
document, depending on the output format and circumstances.

Define an entry to appear in the table of contents by bracketing its
text between calls to the @code{XS} and @code{XE} macros.  A typical
application is to call them immediately after @code{NH} or @code{SH} and
repeat the heading text within them.  The @code{XA} macro, used within
@samp{.XS}/@samp{.XE} pairs, supplements an entry---for instance, when
it requires multiple output lines, whether because a heading is too long
to fit or because style dictates that page numbers not be repeated.  You
may wish to indent the text thus wrapped to correspond to its heading
depth; this can be done in the entry text by prefixing it with tabs or
horizontal motion escape sequences, or by providing a second argument to
the @code{XA} macro.  @code{XS} and @code{XA} automatically associate
the page number where they are called with the text following them, but
they accept arguments to override this behavior.  At the end of the
document, call @code{TC} or @code{PX} to emit the table of contents;
@code{TC} resets the page number to @samp{i} (Roman numeral one), and
then calls @code{PX}.  All of these macros are Berkeley extensions.

@DefmacList {XS, [@Var{page-number}], ms}
@DefmacItemx {XA, [@Var{page-number} [@Var{indentation}]], ms}
@DefmacListEndx {XE, , ms}
Begin, supplement, and end a table of contents entry.  Each entry is
associated with @var{page-number} (otherwise the current page number); a
@var{page-number} of @samp{no} prevents a leader and page number from
being emitted for that entry.  Use of @code{XA} within
@code{XS}/@code{XE} is optional; it can be repeated.  If
@var{indentation} is present, a supplemental entry is indented by that
amount; ens are assumed if no unit is indicated.  Text on input lines
between @code{XS} and @code{XE} is stored for later recall by @code{PX}.
@endDefmac

@Defmac {PX, [@code{no}], ms}
Switch to single-column layout.  Unless @code{no} is specified, center
and interpolate the @code{TOC} string in bold and two points larger than
the body text.  Emit the table of contents entries.
@endDefmac

@Defmac {TC, [@code{no}], ms}
Set the page number to@tie{}1, the page number format to lowercase Roman
numerals, and call @code{PX} (with a @code{no} argument, if present).
@endDefmac

Here's an example of typical @file{ms} table of contents preparation.
We employ horizontal escape sequences @code{\h} to indent the entries by
sectioning depth.

@CartoucheExample
.NH 1
Introduction
.XS
Introduction
.XE
@r{@dots{}}
.NH 2
Methodology
.XS
\h'2n'Methodology
.XA
\h'4n'Fassbinder's Approach
\h'4n'Kahiu's Approach
.XE
@r{@dots{}}
.NH 1
Findings
.XS
Findings
.XE
@r{@dots{}}
.TC
@endCartoucheExample

The remaining features in this subsubsection are GNU extensions.
@code{groff} @file{ms} obviates the need to repeat heading text after
@code{XS} calls.  Call @code{XN} and @code{XH} after @code{NH} and
@code{SH}, respectively.

@DefmacList {XN, @Var{heading-text}, ms}
@DefmacListEndx {XH, @Var{depth} @Var{heading-text}, ms}
Format @var{heading-text} and create a corresponding table of contents
entry.  @code{XN} computes the indentation from the depth of the
preceding @code{NH} call; @code{XH} requires a @var{depth} argument to
do so.
@endDefmac

@code{groff} @file{ms} encourages customization of table of contents
entry production.

@DefmacList {XN-REPLACEMENT, @Var{heading-text}, ms}
@DefmacListEndx {XH-REPLACEMENT, @Var{depth} @Var{heading-text}, ms}
These hook macros implement @code{XN} and @code{XH}, respectively.
They call @code{XN-INIT} and pass their @var{heading-text} arguments to
@code{XH-UPDATE-TOC}.
@endDefmac

@DefmacList {XN-INIT, , ms}
@DefmacListEndx {XH-UPDATE-TOC, @Var{depth} @Var{heading-text}, ms}
The @code{XN-INIT} hook macro does nothing by default.
@code{XH-UPDATE-TOC} brackets @var{heading-text} with @code{XS} and
@code{XE} calls, indenting it by 2 ens per level of @var{depth} beyond
the first.
@endDefmac

We could therefore produce a table of contents similar to that in the
previous example with fewer macro calls.  (The difference is that this
input follows the ``Approach'' entries with leaders and page numbers.)

@CartoucheExample
.NH 1
.XN Introduction
@r{@dots{}}
.NH 2
.XN Methodology
.XH 3 "Fassbinder's Approach"
.XH 3 "Kahiu's Approach"
@r{@dots{}}
.NH 1
.XN Findings
@r{@dots{}}
@endCartoucheExample

To get the section number of the numbered headings into the table of
contents entries, we might define @code{XN-REPLACEMENT} as follows.
(We obtain the heading depth from @code{groff} @file{ms}'s internal
register @code{nh*hl}.)

@CartoucheExample
.de XN-REPLACEMENT
.XN-INIT
.XH-UPDATE-TOC \\n[nh*hl] \\$@@
\&\\*[SN] \\$*
..
@endCartoucheExample

You can change the style of the leader that bridges each table of
contents entry with its page number; define the @code{TC-LEADER} special
character by using the @code{char} request.  A typical leader combines
the dot glyph @samp{.} with a horizontal motion escape sequence to
spread the dots.  The width of the page number field is stored in the
@code{TC-MARGIN} register.

@c ---------------------------------------------------------------------

@node Differences from AT&T ms, ms Naming Conventions, ms Page Layout, ms
@subsection Differences from @acronym{AT&T} @file{ms}
@cindex @file{ms} macros, @code{groff} differences from @acronym{AT&T}
@cindex @acronym{AT&T} @file{ms}, macro package differences

The @code{groff} @file{ms} macros are an independent reimplementation,
using no @acronym{AT&T} code.  Since they take advantage of the extended
features of @code{groff}, they cannot be used with @acronym{AT&T}
@code{troff}.  @code{groff} @file{ms} supports features described above
as Berkeley and Tenth Edition Research Unix extensions, and adds several
of its own.

@itemize @bullet
@item
The internals of @code{groff} @file{ms} differ from the internals of
@acronym{AT&T} @file{ms}.  Documents that depend upon implementation
details of @acronym{AT&T} @file{ms} may not format properly with
@code{groff} @file{ms}.  Such details include macros whose function was
not documented in the @acronym{AT&T} @file{ms}
manual.@footnote{@cite{Typing Documents on the UNIX System: Using the
-ms Macros with Troff and Nroff}, M.@tie{}E.@: Lesk, Bell Laboratories,
1978}
@c XXX: We support RT anyway; maybe we should stop?

@item
The error-handling policy of @code{groff} @file{ms} is to detect and
report errors, rather than to ignore them silently.

@item
Tenth Edition @c possibly 9th
Research Unix supported @code{P1}/@code{P2} macros to bracket code
examples; @code{groff} @file{ms} does not.

@item
@code{groff} @file{ms} does not work in GNU @code{troff}'s
@acronym{AT&T} compatibility mode.  If loaded when that mode is enabled,
it aborts processing with a diagnostic message.

@item
Multiple line spacing is not supported.  Use a larger vertical spacing
instead.

@item
@code{groff} @file{ms} uses the same header and footer defaults in both
@code{nroff} and @code{troff} modes as @acronym{AT&T} @file{ms} does in
@code{troff} mode; @acronym{AT&T}'s default in @code{nroff} mode is to
put the date, in U.S.@: traditional format (e.g., ``January 1, 2021''),
in the center footer (the @code{CF} string).

@item
Many @code{groff} @file{ms} macros, including those for paragraphs,
headings, and displays, cause a reset of paragraph rendering parameters,
and may change the indentation; they do so not by incrementing or
decrementing it, but by setting it absolutely.  This can cause problems
for documents that define additional macros of their own that try to
manipulate indentation.  Use the @file{ms} @code{RS} and @code{RE}
macros instead of the @code{in} request.

@item
@cindex fractional type sizes in @file{ms} macros
@cindex @file{ms} macros, fractional type sizes in
@acronym{AT&T} @file{ms} interpreted the values of the registers
@code{PS} and @code{VS} in points, and did not support the use of
scaling units with them.  @code{groff} @file{ms} interprets values of
the registers @code{PS}, @code{VS}, @code{FPS}, and @code{FVS} equal to
or larger than@tie{}1,000 (one thousand) as decimal fractions multiplied
by@tie{}1,000.@footnote{Register values are converted to and stored as
basic units.  @xref{Measurements}.}  This threshold makes use of a
scaling unit with these parameters practical for high-resolution
devices while preserving backward compatibility.  It also permits
expression of non-integral type sizes.  For example, @samp{groff
-rPS=10.5p} at the shell prompt is equivalent to placing @samp{.nr PS
10.5p} at the beginning of the document.

@item
@acronym{AT&T} @file{ms}'s @code{AU} macro supported arguments used with
some document types; @code{groff} @file{ms} does not.

@item
Right-aligned displays are available.  The @acronym{AT&T} @file{ms}
manual observes that ``it is tempting to assume that @samp{.DS R} will
right adjust lines, but it doesn't work''.  In @code{groff} @file{ms},
it does.

@item
To make @code{groff} @file{ms} use the default page offset (which also
specifies the left margin), the @code{PO} register must stay undefined
until the first @file{ms} macro is called.

This implies that @samp{\n[PO]} should not be used early in the
document, unless it is changed also: accessing an undefined register
automatically defines it.

@item
@code{groff} @file{ms} supports the @code{PN} register, but it is not
necessary; you can access the page number via the usual @code{%}
register and invoke the @code{af} request to assign a different format
to it if desired.@footnote{If you redefine the @file{ms} @code{PT} macro
@c I wouldn't mention that, but Lesk 1978 encourages doing so.  :-/
and desire special treatment of certain page numbers (like @samp{1}),
you may need to handle a non-Arabic page number format, as @code{groff}
@file{ms}'s @code{PT} does; see the macro package source.  @code{groff}
@file{ms} aliases the @code{PN} register to @code{%}.}

@item
The @acronym{AT&T} @file{ms} manual documents registers @code{CW} and
@code{GW} as setting the default column width and ``intercolumn gap'',
respectively, and which applied when @code{MC} was called with fewer
than two arguments.  @code{groff} @file{ms} instead treats @code{MC}
without arguments as synonymous with @code{2C}; there is thus no
occasion for a default column width register.  Further, the @code{MINGW}
register and the second argument to @code{MC} specify a @emph{minimum}
space between columns, not the fixed gutter width of @acronym{AT&T}
@file{ms}.

@item
The @acronym{AT&T} @file{ms} manual did not document the @code{QI}
register; Berkeley and @code{groff} @file{ms} do.
@end itemize

@Defmpreg {GS, ms}
The register @code{GS} is set to@tie{}1 by the @code{groff} @file{ms}
macros, but is not used by the @acronym{AT&T} @file{ms} package.
Documents that need to determine whether they are being formatted with
@code{groff} @file{ms} or another implementation should test this
register.
@endDefmpreg

@menu
* Missing Unix Version 7 ms Macros::
@end menu

@c ---------------------------------------------------------------------

@node Missing Unix Version 7 ms Macros, , Differences from AT&T ms, Differences from AT&T ms
@subsubsection Unix Version 7 @file{ms} macros not implemented by @code{groff} @file{ms}

Several macros described in the Unix Version@tie{}7 @file{ms}
documentation are unimplemented by @code{groff} @file{ms} because they
are specific to the requirements of documents produced internally by
Bell Laboratories, some of which also require a glyph for the Bell
System logo that @code{groff} does not support.  These macros
implemented several document type formats
(@code{EG}, @c engineer's notes
@code{IM}, @c internal memorandum
@code{MF}, @c memorandum for file
@code{MR}, @c memorandum for record
@code{TM}, @c technical memorandum
@code{TR}), @c technical report
were meaningful only in conjunction with the use of certain document
types
(@code{AT}, @c attachments
@code{CS}, @c cover sheet info for `TM` documents
@code{CT}, @c copies to
@code{OK}, @c "other keywords" for `TM` documents
@code{SG}), @c signatures for `TM` documents
stored the postal addresses of Bell Labs sites
(@code{HO}, @c Holmdel
@code{IH}, @c Naperville
@code{MH}, @c Murray Hill
@code{PY}, @c Piscataway
@code{WH}), @c Whippany
or lacked a stable definition over time
(@code{UX}). @c Unix; on 1st use, add footnote id'ing trademark owner
To compatibly render historical @file{ms} documents using these macros,
we advise your documents to invoke the @code{rm} request to remove any
such macros it uses and then define replacements with an authentically
typeset original at hand.@footnote{The removal beforehand is necessary
because @code{groff} @file{ms} aliases these macros to a diagnostic
macro, and you want to redefine the aliased name, not its target.}  For
informal purposes, a simple definition of @code{UX} should maintain the
readability of the document's substance.

@CartoucheExample
.rm UX
.ds UX Unix\"
@endCartoucheExample

@c ---------------------------------------------------------------------

@node ms Legacy Features, ms Naming Conventions, Differences from AT&T ms, ms
@subsection Legacy Features
@cindex @file{ms} macros, strings
@cindex @file{ms} macros, special characters
@cindex @file{ms} macros, accent marks
@cindex accent marks [@file{ms}]
@cindex special characters [@file{ms}]
@cindex strings [@file{ms}]

@code{groff} @file{ms} retains some legacy features solely to support
formatting of historical documents; contemporary ones should not use
them because they can render poorly.  See the @cite{groff_char@r{(7)}}
man page.

@unnumberedsubsubsec AT&T accent mark strings

AT&T @file{ms} defined accent mark strings as follows.

@Defmpstr {@code{'}, ms}
Apply acute accent to subsequent glyph.
@endDefmpstr

@Defmpstr {@code{`}, ms}
Apply grave accent to subsequent glyph.
@endDefmpstr

@Defmpstr {:, ms}
Apply dieresis (umlaut) to subsequent glyph.
@endDefmpstr

@Defmpstr {^, ms}
Apply circumflex accent to subsequent glyph.
@endDefmpstr

@Defmpstr {~, ms}
Apply tilde accent to subsequent glyph.
@endDefmpstr

@Defmpstr {C, ms}
Apply caron to subsequent glyph.
@endDefmpstr

@Defmpstr {\,, ms}
Apply cedilla to subsequent glyph.
@endDefmpstr

@unnumberedsubsubsec Berkeley accent mark and glyph strings

Berkeley @file{ms} offered an @code{AM} macro; calling it redefined the
AT&T accent mark strings (except for @samp{\*C}), applied them to the
@emph{preceding} glyph, and defined additional strings, some for spacing
glyphs.

@Defmac {AM, , ms}
Enable alternative accent mark and glyph-producing strings.
@endDefmac

@Defmpstr {@code{'}, ms}
Apply acute accent to preceding glyph.
@endDefmpstr

@Defmpstr {@code{`}, ms}
Apply grave accent to preceding glyph.
@endDefmpstr

@Defmpstr {:, ms}
Apply dieresis (umlaut) to preceding glyph.
@endDefmpstr

@Defmpstr {^, ms}
Apply circumflex accent to preceding glyph.
@endDefmpstr

@Defmpstr {~, ms}
Apply tilde accent to preceding glyph.
@endDefmpstr

@Defmpstr {\,, ms}
Apply cedilla to preceding glyph.
@endDefmpstr

@Defmpstr {/, ms}
Apply stroke (slash) to preceding glyph.
@endDefmpstr

@Defmpstr {v, ms}
Apply caron to preceding glyph.
@endDefmpstr

@Defmpstr {_, ms}
Apply macron to preceding glyph.
@endDefmpstr

@Defmpstr {., ms}
Apply underdot to preceding glyph.
@endDefmpstr

@Defmpstr {o, ms}
Apply ring accent to preceding glyph.
@endDefmpstr

@Defmpstr {?, ms}
Interpolate inverted question mark.
@endDefmpstr

@Defmpstr {!, ms}
Interpolate inverted exclamation mark.
@endDefmpstr

@Defmpstr {8, ms}
Interpolate small letter sharp s.
@endDefmpstr

@Defmpstr {q, ms}
Interpolate small letter o with hook accent (ogonek).
@endDefmpstr

@Defmpstr {3, ms}
Interpolate small letter yogh.
@endDefmpstr

@Defmpstr {d-, ms}
Interpolate small letter eth.
@endDefmpstr

@Defmpstr {D-, ms}
Interpolate capital letter eth.
@endDefmpstr

@Defmpstr {th, ms}
Interpolate small letter thorn.
@endDefmpstr

@Defmpstr {Th, ms}
Interpolate capital letter thorn.
@endDefmpstr

@Defmpstr {ae, ms}
Interpolate small æ ligature.
@endDefmpstr

@Defmpstr {Ae, ms}
Interpolate capital Æ ligature.
@endDefmpstr

@Defmpstr {oe, ms}
Interpolate small oe ligature.
@endDefmpstr

@Defmpstr {OE, ms}
Interpolate capital OE ligature.
@endDefmpstr

@c ---------------------------------------------------------------------

@node ms Naming Conventions,  , ms Legacy Features, ms
@subsection Naming Conventions
@cindex @file{ms} macros, naming conventions
@cindex naming conventions, @file{ms} macros

The following conventions are used for names of macros, strings, and
registers.  External names available to documents that use the
@code{groff} @file{ms} macros contain only uppercase letters and digits.

Internally, the macros are divided into modules.  Conventions for
identifier names are as follows.

@itemize @bullet
@item
Names used only within one module are of the form
@var{module}@code{*}@var{name}.

@item
Names used outside the module in which they are defined are of the form
@var{module}@code{@@}@var{name}.

@item
Names associated with a particular environment are of the form
@var{environment}@code{:}@var{name}; these are used only within the
@code{par} module.

@item
@var{name} does not have a module prefix.

@item
Constructed names used to implement arrays are of the form
@var{array}@code{!}@var{index}.
@end itemize

Thus the @code{groff} @file{ms} macros reserve the following names.

@itemize @bullet
@item
Names containing the characters @code{*}, @code{@@}, and@tie{}@code{:}.

@item
Names containing only uppercase letters and digits.
@end itemize

@codequotebacktick off
@codequoteundirected off


@c =====================================================================
@c =====================================================================

@node GNU troff Reference, File Formats, Major Macro Packages, Top
@chapter GNU @code{troff} Reference
@cindex reference, @code{gtroff}
@cindex @code{gtroff}, reference

This chapter covers @emph{all} of the facilities of the GNU
@code{troff} formatting engine.  Users of macro packages may skip it if
not interested in details.


@menu
* Text::
* Page Geometry::
* Measurements::
* Numeric Expressions::
* Identifiers::
* Formatter Instructions::
* Comments::
* Registers::
* Manipulating Filling and Adjustment::
* Manipulating Hyphenation::
* Manipulating Spacing::
* Tabs and Fields::
* Character Translations::
* @code{troff} and @code{nroff} Modes::
* Line Layout::
* Line Continuation::
* Page Layout::
* Page Control::
* Fonts and Symbols::
* Manipulating Type Size and Vertical Spacing::
* Colors::
* Strings::
* Conditionals and Loops::
* Writing Macros::
* Page Motions::
* Drawing Requests::
* Deferring Output::
* Traps::
* Diversions::
* Punning Names::
* Environments::
* Suppressing Output::
* I/O::
* Postprocessor Access::
* Miscellaneous::
* Gtroff Internals::
* Debugging::
* Implementation Differences::
@end menu


@c =====================================================================

@codequotebacktick on
@codequoteundirected on

@c BEGIN Keep roughly parallel with roff(7) section "Concepts".
@node Text, Measurements, GNU troff Reference, GNU troff Reference
@section Text
@cindex text, GNU @code{troff} processing

@acronym{AT&T} @code{troff} was designed to take input as it would be
composed on a typewriter, including the teletypewriters used as early
computer terminals, and relieve the user drafting a document of concern
with details like line length, hyphenation breaking, and the achievement
of straight margins.  Early in its development, the program gained the
ability to prepare output for a phototypesetter; a document could then
be prepared for output to either a teletypewriter, a phototypesetter, or
both.  GNU @code{troff} continues this tradition of permitting an author
to compose a single master version of a document which can then be
rendered for a variety of output formats or devices.

@code{roff} input files contain text interspersed with instructions to
control the formatter.  Even in the absence of such instructions, GNU
@code{troff} still processes its input in several ways, by filling,
hyphenating, breaking, and adjusting it, and supplementing it with
inter-sentence space.

@menu
* Filling::
* Hyphenation::
* Sentences::
* Breaking::
* Adjustment::
* Tabs and Leaders::
* Requests and Macros::
* Macro Packages::
* Input Encodings::
* Input Conventions::
@end menu

@c ---------------------------------------------------------------------

@node Filling, Sentences, Text, Text
@subsection Filling

When GNU @code{troff} starts up, it obtains information about the device
for which it is preparing output.@footnote{@xref{Device and Font
Description Files}.}  An essential property is the length of the output
line, such as ``6.5 inches''.

@cindex word, definition of
@cindex filling
GNU @code{troff} interprets plain text files employing the Unix
line-ending convention.  It reads input a character at a time,
collecting words as it goes, and fits as many words together on an
output line as it can---this is known as @dfn{filling}.  To GNU
@code{troff}, a @dfn{word} is any sequence of one or more characters
that aren't spaces, tabs, or newlines.  The exceptions separate
words.@footnote{There are also @emph{escape sequences} which can
function as word characters, word separators, or neither---the last
simply have no effect on GNU @code{troff}'s idea of whether an input
character is within a word or not.}  To disable filling, see
@ref{Manipulating Filling and Adjustment}.

@Example
It is a truth universally acknowledged
that a single man in possession of a
good fortune must be in want of a wife.
    @result{} It is a truth universally acknowledged that a
    @result{} single man in possession of a good fortune must
    @result{} be in want of a wife.
@endExample

@c ---------------------------------------------------------------------

@node Sentences, Hyphenation, Filling, Text
@subsection Sentences
@cindex sentences

A passionate debate has raged for decades among writers of the English
language over whether more space should appear between adjacent
sentences than between words within a sentence, and if so, how much, and
what other circumstances should influence this spacing.@footnote{A
well-researched jeremiad appreciated by @code{groff} contributors on
both sides of the sentence-spacing debate can be found at
@uref{https://web.archive.org@//web@//20171217060354@//http://www.heracliteanriver.com@//?p=324}.}
GNU @code{troff} follows the example of @acronym{AT&T} @code{troff};
it attempts to detect the boundaries between sentences, and supplies
additional inter-sentence space between them.

@Example
Hello, world!
Welcome to groff.
    @result{} Hello, world!  Welcome to groff.
@endExample

@cindex end-of-sentence characters
@cindex sentence space
@cindex space between sentences
@cindex French spacing
GNU @code{troff} flags certain characters (normally @samp{!}, @samp{?},
and @samp{.}) as potentially ending a sentence.  When GNU @code{troff}
encounters one of these @dfn{end-of-sentence characters} at the end of
an input line, or one of them is followed by two (unescaped) spaces on
the same input line, it appends an inter-word space followed by an
inter-sentence space in the output.

@Example
R. Harper subscribes to a maxim of P. T. Barnum.
    @result{} R. Harper subscribes to a maxim of P. T. Barnum.
@endExample

In the above example, inter-sentence space is not added after @samp{P.}
or @samp{T.} because the periods do not occur at the end of an input
line, nor are they followed by two or more spaces.  Let's imagine that
we've heard something about defamation from Mr.@: Harper's attorney,
recast the sentence, and reflowed it in our text editor.

@Example
I submit that R. Harper subscribes to a maxim of P. T.
Barnum.
    @result{} I submit that R. Harper subscribes to a maxim of
    @result{} P. T.  Barnum.
@endExample

``Barnum'' doesn't begin a sentence!  What to do?  Let us meet our first
@dfn{escape sequence}, a series of input characters that give
instructions to GNU @code{troff} instead of being used to construct
output device glyphs.@footnote{This statement oversimplifes; there are
escape sequences whose purpose is precisely to produce glyphs on the
output device, and input characters that @emph{aren't} part of escape
sequences can undergo a great deal of processing before getting to the
output.} An escape sequence begins with the backslash character @code{\}
by default, an uncommon character in natural language text, and is
@emph{always} followed by at least one other character, hence the term
``sequence''.

@cindex @code{\&}, at end of sentence
The dummy character escape sequence @code{\&} can be used after an
end-of-sentence character to defeat end-of-sentence detection on a
per-instance basis.  We can therefore rewrite our input more
defensively.

@Example
I submit that R.\& Harper subscribes to a maxim of P.\&
T.\& Barnum.
    @result{} I submit that R. Harper subscribes to a maxim of
    @result{} P. T. Barnum.
@endExample

Adding text caused our input to wrap; now, we don't need @code{\&} after
@samp{T.} but we do after @samp{P.}.  Consistent use of the escape
sequence ensures that potential sentence boundaries are robust to
editing activities.  Further advice along these lines will follow in
@ref{Input Conventions}.

@cindex end-of-sentence transparent characters
@cindex characters, end-of-sentence transparent
@cindex @code{dg} glyph, at end of sentence
@cindex @code{dd} glyph, at end of sentence
@cindex @code{rq} glyph, at end of sentence
@cindex @code{cq} glyph, at end of sentence
@cindex @code{"}, at end of sentence
@cindex @code{'}, at end of sentence
@cindex @code{)}, at end of sentence
@cindex @code{]}, at end of sentence
@cindex @code{*}, at end of sentence
@cindex special characters
@cindex characters, special
Normally, the occurrence of a visible non-end-of-sentence character (as
opposed to a space or tab) immediately after an end-of-sentence
character cancels detection of the end of a sentence.  For example, it
would be incorrect for GNU @code{troff} to infer the end of a sentence
after the dot in @samp{3.14159}.  However, several characters are
treated @emph{transparently} after the occurrence of an end-of-sentence
character.  That is, GNU @code{troff} does not cancel end-of-sentence
detection when it processes them.  This is because such characters are
often used as footnote markers or to close quotations and
parentheticals.  The default set is @samp{"}, @samp{'}, @samp{)},
@samp{]}, @samp{*}, @code{\[dg]}, @code{\[dd]}, @code{\[rq]}, and
@code{\[cq]}.  The last four are examples of @dfn{special characters},
escape sequences whose purpose is to obtain glyphs that are not easily
typed at the keyboard, or which have special meaning to GNU @code{troff}
(like @code{\} itself).@footnote{The mnemonics for the special
characters shown here are ``dagger'', ``double dagger'', ``right
(double) quote'', and ``closing (single) quote''.  See the
@cite{groff_char@r{(7)}} man page.}

@Example
\[lq]The idea that the poor should have leisure has always
been shocking to the rich.\[rq]
(Bertrand Russell, 1935)
@c XXX: @iftex puts a blank line on the output.  This seems like a bug.
@c @newline works around it.  But we need a weird inverse indent.
@iftex @
  @result{} @quotedblleft{}The idea that the poor should have
    @result{} leisure has always been shocking to
    @result{} the rich.@quotedblright{}  (Bertrand Russell, 1935)
@end iftex
@ifnottex
    @result{} "The idea that the poor should have
    @result{} leisure has always been shocking to
    @result{} the rich."  (Bertrand Russell, 1935)
@end ifnottex
@endExample

The sets of characters that potentially end sentences or are transparent
to sentence endings are configurable.  See the @code{cflags} request in
@ref{Using Symbols}.  To change the additional inter-sentence space
amount---even to remove it entirely---see @ref{Manipulating Filling and
Adjustment}.

@c ---------------------------------------------------------------------

@node Hyphenation, Breaking, Sentences, Text
@subsection Hyphenation
@cindex hyphenation

When an output line is nearly full, it is uncommon for the next word
collected from the input to exactly fill it---typically, there is room
left over only for part of the next word.  The process of splitting a
word so that it appears partially on one line (with a hyphen to indicate
to the reader that the word has been broken) with its remainder on the
next is @dfn{hyphenation}.  Hyphenation points can be manually
specified; GNU @code{troff} also uses a hyphenation algorithm and
language-specific pattern files (based on those used in @TeX{}) to
decide which words can be hyphenated and where.

Hyphenation does not always occur even when the hyphenation rules for a
word allow it; it can be disabled, and when not disabled there are
several parameters that can prevent it in certain circumstances.
@xref{Manipulating Hyphenation}.

@c ---------------------------------------------------------------------

@node Breaking, Adjustment, Hyphenation, Text
@subsection Breaking
@cindex break
@cindex implicit line break
@cindex line break, output
@cindex output line break

Once an output line has been filled, whether or not hyphenation has
occurred on that line, the next word read from the input will be placed
on a different output line; this is called a @dfn{break}.  In this
manual and in @code{roff} discussions generally, a ``break'' if not
further qualified always refers to the termination of an output line.
When the formatter is filling text, it introduces breaks automatically
to keep output lines from exceeding the configured line length.  After
an automatic break, GNU @code{troff} adjusts the line if applicable (see
below), and then resumes collecting and filling text on the next output
line.

Sometimes, a line cannot be broken automatically.  This usually does
not happen with natural language text unless the output line length has
been manipulated to be extremely short, but it can with specialized
text like program source code.  We can use @code{perl} at the shell
prompt to contrive an example of failure to break the line.  We also
employ the @option{-z} option to suppress normal output.

@Example
$ perl -e 'print "#" x 80, "\n";' | nroff -z
    @error{} warning: cannot break line
@endExample

The remedy for these cases is to tell GNU @code{troff} where the line
may be broken without hyphens.  This is done with the non-printing break
point escape sequence @samp{\:}; see @ref{Manipulating Hyphenation}.

@cindex blank line
@cindex empty line
@cindex line, blank
@cindex blank line macro (@code{blm})
What if the document author wants to stop filling lines temporarily, for
instance to start a new paragraph?  There are several solutions.  A
blank input line not only causes a break, but by default it also outputs
a one-line vertical space (effectively a blank output line).  This
behavior can be modified; see @ref{Blank Line Traps}.  Macro packages
may discourage or disable the blank line method of paragraphing in favor
of their own macros.

@cindex leading spaces
@cindex spaces, leading and trailing
@cindex trailing spaces on text lines
@cindex leading space macro (@code{lsm})
A line that begins with one or more spaces causes a break.  The spaces
are output at the beginning of the next line without being
@emph{adjusted} (see below); however, this behavior can be modified
(@pxref{Leading Space Traps}).  Again, macro packages may provide other
methods of producing indented paragraphs.  Trailing spaces on text lines
are discarded.@footnote{``Text lines'' are defined in @ref{Requests and
Macros}.}

What if the file ends before enough words have been collected to fill an
output line?  Or the output line is exactly full but not yet broken, and
there is no more input?  GNU @code{troff} interprets the end of input as
a break.  Certain requests also cause breaks, implicitly or explicitly.
This is discussed in @ref{Manipulating Filling and Adjustment}.

@c ---------------------------------------------------------------------

@node Adjustment, Tabs and Leaders, Breaking, Text
@subsection Adjustment

@cindex extra spaces between words
After GNU @code{troff} performs an automatic break, it may then
@dfn{adjust} the line, widening inter-word spaces until the text reaches
the right margin.  Extra spaces between words are preserved.  Leading
and trailing spaces are handled as noted above.  Text can be aligned to
the left or right margin only, or centered; see @ref{Manipulating
Filling and Adjustment}.
@c END Keep roughly parallel with roff(7) section "Concepts".

@c ---------------------------------------------------------------------

@node Tabs and Leaders, Input Conventions, Adjustment, Text
@subsection Tabs and Leaders

@cindex horizontal tab character
@cindex tab character
@cindex character, horizontal tab
@cindex leader character
@cindex character, leader
@cindex tab stops
@cindex stops, tab
GNU @code{troff} translates input horizontal tab characters (``tabs'')
and @key{Control+A} characters (``leaders'') into movements to the next
tab stop.  Tabs simply move to the next tab stop; leaders place enough
periods to fill the space.  Tab stops are by default located every half
inch measured from the drawing position corresponding to the beginning
of the input line; see @ref{Page Geometry}.   Tabs and leaders do not
cause breaks and therefore do not interrupt filling.  Below, we use
arrows @arrow{} and bullets @bullet{} to indicate input tabs and
leaders, respectively.

@Example
1
@arrow{} 2 @arrow{} 3 @bullet{} 4
@arrow{} @bullet{} 5
@result{} 1         2       3.......4         ........5
@endExample

Tabs and leaders lend themselves to table construction.@footnote{``Tab''
is short for ``tabulation'', revealing the term's origin as a spacing
mechanism for table arrangement.} The tab and leader glyphs can be
configured, and further facilities for sophisticated table composition
are available; see @ref{Tabs and Fields}.  There are many details to
track when using such low-level features, so most users turn to the
@cite{tbl@r{(1)}} preprocessor for table construction.

@c ---------------------------------------------------------------------

@node Requests and Macros, Macro Packages, Tabs and Leaders, Text
@subsection Requests and Macros

We have now encountered almost all of the syntax there is in the
@code{roff} language, with an exception already noted in passing.
@cindex request
@cindex control character (@code{.})
@cindex character, control (@code{.})
@cindex no-break control character (@code{'})
@cindex character, no-break control (@code{'})
@cindex control character, no-break (@code{'})
A @dfn{request} is an instruction to the formatter that occurs after a
@dfn{control character}, which is recognized at the beginning of an
input line.  The regular control character is a dot (@code{.}).  Its
counterpart, the @dfn{no-break control character}, a neutral apostrophe
(@code{'}), suppresses the break that is implied by some requests.
These characters were chosen because it is uncommon for lines of text in
natural languages to begin with them.
@cindex dummy character (@code{\&}), as control character suppressor
@cindex character, dummy (@code{\&}), as control character suppressor
If you require a formatted period or apostrophe (closing single
quotation mark) where GNU @code{troff} is expecting a control character,
prefix the dot or neutral apostrophe with the dummy character escape
sequence, @samp{\&}.

@cindex control line
An input line beginning with a control character is called a
@dfn{control line}.
@cindex text line
Every line of input that is not a control line is a @dfn{text
line}.@footnote{The @code{\@key{RET}} escape sequence can alter how an
input line is classified; see @ref{Line Continuation}.}

@cindex argument
Requests often take @dfn{arguments}, words (separated from the request
name and each other by spaces) that specify details of the action GNU
@code{troff} is expected to perform.  If a request is meaningless
without arguments, it is typically ignored.

GNU @code{troff}'s requests and escape sequences comprise the control
language of the formatter.  Of key importance are the requests that
define macros.  Macros are invoked like requests, enabling the request
repertoire to be extended or overridden.@footnote{Argument handling in
macros is more flexible but also more complex.  @xref{Calling Macros}.}

@cindex macro
@cindex calling a macro
@cindex interpolation
A @dfn{macro} can be thought of as an abbreviation you can define for a
collection of control and text lines.  When the macro is @dfn{called} by
giving its name after a control character, it is replaced with what it
stands for.  The process of textual replacement is known as
@dfn{interpolation}.@footnote{Some escape sequences undergo
interpolation as well.}  Interpolations are handled as soon as they are
recognized, and once performed, a @code{roff} formatter scans the
replacement for further requests, macro calls, and escape sequences.

In @code{roff} systems, the @code{de} request defines a
macro.@footnote{GNU @code{troff} offers additional ones.  @xref{Writing
Macros}.}

@Example
.de DATE
2020-11-14
..
@endExample

@noindent
The foregoing input produces no output by itself; all we have done is
store some information.  Observe the pair of dots that ends the macro
definition.  This is a default; you can specify your own terminator for
the macro definition as the second argument to the @code{de} request.

@Example
.de NAME ENDNAME
Heywood Jabuzzoff
.ENDNAME
@endExample

In fact, the ending marker is itself the name of a macro to be
called, or a request to be invoked, if it is defined at the time its
contol line is read.

@Example
.de END
Big Rip
..
.de START END
Big Bang
.END
.START
    @result{} Big Rip Big Bang
@endExample

@noindent
In the foregoing example, ``Big Rip'' printed before ``Big Bang''
because its macro was @emph{called} first.  Consider what would happen
if we dropped @code{END} from the @samp{.de START} line and added
@code{..} after @code{.END}.  Would the order change?

Let us consider a more elaborate example.

@Example
.de DATE
2020-10-05
..
.
.de BOSS
D.\& Kruger,
J.\& Peterman
..
.
.de NOTICE
Approved:
.DATE
by
.BOSS
..
.
Insert tedious regulatory compliance paragraph here.

.NOTICE

Insert tedious liability disclaimer paragraph here.

.NOTICE
    @result{} Insert tedious regulatory compliance paragraph here.
    @result{}
    @result{} Approved: 2020-10-05 by D. Kruger, J. Peterman
    @result{}
    @result{} Insert tedious liability disclaimer paragraph here.
    @result{}
    @result{} Approved: 2020-10-05 by D. Kruger, J. Peterman
@endExample

@noindent
The above document started with a series of control lines.  Three macros
were defined, with a @code{de} request declaring each macro's name, and
the ``body'' of the macro starting on the next line and continuing until
a line with two dots @samp{@code{..}} marked its end.  The text proper
began only after the macros were defined; this is a common pattern.
Only the @code{NOTICE} macro was called ``directly'' by the document;
@code{DATE} and @code{BOSS} were called only by @code{NOTICE} itself.
Escape sequences were used in @code{BOSS}, two levels of macro
interpolation deep.

The advantage in typing and maintenance economy may not be obvious from
such a short example, but imagine a much longer document with dozens of
such paragraphs, each requiring a notice of managerial approval.
Consider what must happen if you are in charge of generating a new
version of such a document with a different date, for a different boss.
With well-chosen macros, you only have to change each datum in one
place.

In practice, we would probably use strings (@pxref{Strings}) instead of
macros for such simple interpolations; what is important here is to
glimpse the potential of macros and the power of recursive
interpolation.

We could have defined @code{DATE} and @code{BOSS} in the opposite order;
perhaps less obviously, we could also have defined them @emph{after}
@code{NOTICE}.  ``Forward references'' like this are acceptable because
the body of a macro definition is not (completely) interpreted, but
stored instead (@pxref{Copy Mode}).  While a macro is being defined (or
appended to), requests are not interpreted and macros not interpolated,
whereas some commonly used escape sequences @emph{are} interpreted.
@code{roff} systems also support recursive macro calls, as long as you
have a way to break the recursion (@pxref{Conditionals and Loops}).
Maintainable @code{roff} documents tend to arrange macro definitions to
minimize forward references.

@c ---------------------------------------------------------------------

@node Macro Packages, Input Encodings, Requests and Macros, Text
@subsection Macro Packages
@cindex macro package
@cindex package, macro

@c TODO: Consider parallelizing with groff_tmac(5) "Description".
Macro definitions can be collected into @dfn{macro files}, @code{roff}
input files designed to produce no output themselves but instead ease
the preparation of other @code{roff} documents.  There is no syntactical
difference between a macro file and any other @code{roff} document; only
its purpose distinguishes it.  When a macro file is installed at a
standard location and suitable for use by a general audience, it is
often termed a @dfn{macro package}.@footnote{Macro files and packages
frequently define registers and strings as well.}  Macro packages can be
loaded by supplying the @option{-m} option to GNU @command{troff} or a
@code{groff} front end.  A document wishing to use a macro package can
load it with the @code{mso} (``macro source'') request.

@c ---------------------------------------------------------------------

@c TODO: Move a lot of this node to the "Invoking groff" chapter.  Some
@c of the discussion is better placed in discussion of output drivers
@c (e.g., what character encodings _they_ support for output and their
@c responsibility for converting to them) as well.

@node Input Encodings, Input Conventions, Macro Packages, Text
@subsection Input Encodings

The @command{groff} command's @option{-k} option calls the
@command{preconv} preprocessor to perform input character encoding
conversions.  Input to the GNU @code{troff} formatter itself, on the
other hand, must be in one of two encodings it can recognize.

@table @code
@item cp1047
@cindex encoding, input, @acronym{EBCDIC}
@cindex @acronym{EBCDIC}, input encoding
@cindex input encoding, @acronym{EBCDIC}
@cindex encoding, input, code page 1047
@cindex code page 1047, input encoding
@cindex input encoding, code page 1047
@cindex IBM code page 1047 input encoding
@pindex cp1047.tmac
The code page 1047 input encoding works only on @acronym{EBCDIC}
platforms (and conversely, the other input encodings don't work with
@acronym{EBCDIC}); the file @file{cp1047.tmac} is loaded at startup.

@item latin1
@cindex encoding, input, @w{Latin-1} (ISO @w{8859-1})
@cindex @w{Latin-1} (ISO @w{8859-1}), input encoding
@cindex ISO @w{8859-1} (@w{Latin-1}), input encoding
@cindex input encoding, @w{Latin-1} (ISO @w{8859-1})
@pindex latin1.tmac
ISO @w{Latin-1}, an encoding for Western European languages, is the
default input encoding on non-@acronym{EBCDIC} platforms; the file
@file{latin1.tmac} is loaded at startup.
@end table

@noindent
Any document that is encoded in ISO 646:1991 (a descendant of USAS
@w{X3.4-1968} or ``US-ASCII''), or, equivalently, uses only code points
from the ``C0 Controls'' and ``Basic Latin'' parts of the Unicode
character set is also a valid ISO @w{Latin-1} document; the standards
are interchangeable in their first 128 code points.@footnote{The
@emph{semantics} of certain punctuation code points have gotten stricter
with the successive standards, a cause of some frustration among man
page writers; see the @cite{groff_char@r{(7)}} man page.}

Other encodings are supported by means of macro packages.

@table @code
@item latin2
@cindex encoding, input, @w{Latin-2} (ISO @w{8859-2})
@cindex @w{Latin-2} (ISO @w{8859-2}), input encoding
@cindex ISO @w{8859-2} (@w{Latin-2}), input encoding
@cindex input encoding, @w{Latin-2} (ISO @w{8859-2})
@pindex latin2.tmac
To use ISO @w{Latin-2}, an encoding for Central and Eastern European
languages, invoke @w{@samp{.mso latin2.tmac}} at the beginning of your
document or supply @samp{-mlatin2} as a command-line argument to
@code{groff}.

@item latin5
@cindex encoding, input, @w{Latin-5} (ISO @w{8859-9})
@cindex @w{Latin-5} (ISO @w{8859-9}), input encoding
@cindex ISO @w{8859-9} (@w{Latin-5}), input encoding
@cindex input encoding, @w{Latin-5} (ISO @w{8859-9})
@pindex latin5.tmac
To use ISO @w{Latin-5}, an encoding for the Turkish language, invoke
@w{@samp{.mso latin5.tmac}} at the beginning of your document or
supply @samp{-mlatin5} as a command-line argument to @code{groff}.

@item latin9
@cindex encoding, input, @w{Latin-9} (ISO @w{8859-15})
@cindex @w{Latin-9} (ISO @w{8859-15}), input encoding
@cindex ISO @w{8859-15} (@w{Latin-9}), input encoding
@cindex input encoding, @w{Latin-9} (ISO @w{8859-15})
@pindex latin9.tmac
ISO @w{Latin-9} succeeds @w{Latin-1}; it includes a Euro sign and better
glyph coverage for French.  To use this encoding, invoke @w{@samp{.mso
latin9.tmac}} at the beginning of your document or supply
@samp{-mlatin9} as a command-line argument to @code{groff}.
@end table

Some characters from an input encoding may not be available with a
particular output driver, or their glyphs may not have representation in
the font used.  For terminal devices, fallbacks are defined, like
@samp{EUR} for the Euro sign and @samp{(C)} for the copyright sign.  For
typesetter devices, you may need to ``mount'' fonts that support glyphs
required by the document.  @xref{Font Positions}.

@pindex freeeuro.pfa
@pindex ec.tmac
Because a Euro glyph was not historically defined in PostScript fonts,
@code{groff} comes with a font called @file{freeeuro.pfa} that provides
the Euro in several styles.  Standard PostScript fonts contain the
glyphs from @w{Latin-5} and @w{Latin-9} that @w{Latin-1} lacks, so these
encodings are supported for the @option{ps} and @option{pdf} output
devices as @code{groff} ships, while @w{Latin-2} is not.

Unicode supports characters from all other input encodings; the
@option{utf8} output driver for terminals therefore does as well.  The
DVI output driver supports the @w{Latin-2} and @w{Latin-9} encodings if
the command-line option @option{-mec} is used as well.  @footnote{The
DVI output device defaults to using the Computer Modern (CM) fonts;
@file{ec.tmac} loads the EC fonts instead, which provide Euro
@samp{\[Eu]} and per mille @samp{\[%0]} glyphs.}

@c ---------------------------------------------------------------------

@node Input Conventions, , Input Encodings, Text
@subsection Input Conventions
@cindex input conventions
@cindex conventions for input

Since GNU @code{troff} fills text automatically, it is common practice
in the @code{roff} language to avoid visual composition of text in input
files: the esthetic appeal of the formatted output is what matters.
Therefore, @code{roff} input should be arranged such that it is easy for
authors and maintainers to compose and develop the document, understand
the syntax of @code{roff} requests, macro calls, and preprocessor
languages used, and predict the behavior of the formatter.  Several
traditions have accrued in service of these goals.

@itemize @bullet
@item
Follow sentence endings in the input with newlines to ease their
recognition (@pxref{Sentences}).  It is frequently convenient to end
text lines after colons and semicolons as well, as these typically
precede independent clauses.  Consider doing so after commas; they often
occur in lists that become easy to scan when itemized by line, or
constitute supplements to the sentence that are added, deleted, or
updated to clarify it.  Parenthetical and quoted phrases are also good
candidates for placement on text lines by themselves.

@item
Set your text editor's line length to 72 characters or
fewer.@footnote{Emacs: @code{fill-column: 72}; Vim: @code{textwidth=72}}
This limit, combined with the previous item of advice, makes it less
common that an input line will wrap in your text editor, and thus will
help you perceive excessively long constructions in your text.  Recall
that natural languages originate in speech, not writing, and that
punctuation is correlated with pauses for breathing and changes in
prosody.

@item
Use @code{\&} after @samp{!}, @samp{?}, and @samp{.} if they are
followed by space, tab, or newline characters and don't end a sentence.

@item
In filled text lines, use @code{\&} before @samp{.} and @samp{'} if they
are preceded by space, so that reflowing the input doesn't turn them
into control lines.

@item
Do not use spaces to perform indentation or align columns of a table.
Leading spaces are reliable when text is not being filled.

@item
Comment your document.  It is never too soon to apply comments to
record information of use to future document maintainers (including your
future self).  We thus introduce another escape sequence, @code{\"},
which causes GNU @code{troff} to ignore the remainder of the input line.

@item
Use the empty request---a control character followed immediately by a
newline---to visually manage separation of material in input files.
Many of the @code{groff} project's own documents use an empty request
between sentences, after macro definitions, and where a break is
expected, and two empty requests between paragraphs or other requests or
macro calls that will introduce vertical space into the document.

You can combine the empty request with the comment escape sequence to
include whole-line comments in your document, and even ``comment out''
sections of it.
@end itemize

We conclude this section with an example sufficiently long to illustrate
most of the above suggestions in practice.  For the purpose of fitting
the example between the margins of this manual with the font used for
its typeset version, we have shortened the input line length to 56
columns.  As before, an arrow @arrow{} indicates a tab character.

@c Wrap example at 56 columns (not counting @arrow{}).
@CartoucheExample
.\"   nroff this_file.roff | less
.\"   groff -T ps this_file.roff > this_file.ps
@arrow{}The theory of relativity is intimately connected with
the theory of space and time.
.
I shall therefore begin with a brief investigation of
the origin of our ideas of space and time,
although in doing so I know that I introduce a
controversial subject.  \" remainder of paragraph elided
.
.

@arrow{}The experiences of an individual appear to us arranged
in a series of events;
in this series the single events which we remember
appear to be ordered according to the criterion of
\[lq]earlier\[rq] and \[lq]later\[rq], \" punct swapped
which cannot be analysed further.
.
There exists,
therefore,
for the individual,
an I-time,
or subjective time.
.
This itself is not measurable.
.
I can,
indeed,
associate numbers with the events,
in such a way that the greater number is associated with
the later event than with an earlier one;
but the nature of this association may be quite
arbitrary.
.
This association I can define by means of a clock by
comparing the order of events furnished by the clock
with the order of a given series of events.
.
We understand by a clock something which provides a
series of events which can be counted,
and which has other properties of which we shall speak
later.
.\" Albert Einstein, _The Meaning of Relativity_, 1922
@endCartoucheExample

@node Page Geometry, Measurements, Text, GNU troff Reference
@section Page Geometry
@cindex page, geometry of
@cindex geometry, page

@code{roff} systems format text under certain assumptions about the size
of the output medium, or page.  For the formatter to correctly break a
line it is filling, it must know the line length, which it derives from
the page width (@pxref{Line Layout}).  For it to decide whether to write
an output line to the current page or wait until the next one, it must
know the page length (@pxref{Page Layout}).

@cindex device resolution
@cindex resolution, device
@cindex basic units
@cindex units, basic
@cindex machine units
@cindex units, machine
A device's @dfn{resolution} converts practical units like inches or
centimeters to @dfn{basic units}, a convenient length measure for the
output device or file format.  The formatter and output driver use basic
units to reckon page measurements.  The device description file defines
its resolution and page dimensions (@pxref{DESC File Format}).

@cindex page
A @dfn{page} is a two-dimensional structure upon which a @code{roff}
system imposes a rectangular coordinate system with its upper left
corner as the origin.  Coordinate values are in basic units and increase
down and to the right.  Useful ones are therefore always positive and
within numeric ranges corresponding to the page boundaries.

@cindex drawing position
@cindex position, drawing
While the formatter (and, later, output driver) is processing a page, it
keeps track of its @dfn{drawing position}, which is the location at
which the next glyph will be written, from which the next motion will be
measured, or where a geometric primitive will commence rendering.
@cindex text baseline
@cindex baseline, text
Notionally, glyphs are drawn from the text baseline upward and to the
right.@footnote{@code{groff} does not yet support right-to-left
scripts.}  The @dfn{text baseline} is a (usually invisible) line upon
which the glyphs of a typeface are aligned.  A glyph therefore
``starts'' at its bottom-left corner.  If drawn at the origin, a typical
letter glyph would lie partially or wholly off the page, depending on
whether, like ``g'', it features a decender below the baseline.

@cindex page offset
@cindex offset, page
Such a situation is nearly always undesirable.  It is furthermore
conventional not to write or draw at the extreme edges of the page.
Therefore the initial drawing position of a @code{roff} formatter is not
at the origin, but below and to the right of it.  This rightward shift
from the left edge is known as the @dfn{page
offset}.@footnote{@code{groff}'s terminal output devices have page
offsets of zero.}  The downward shift leaves room for a text output
line.

Text is arranged on a one-dimensional lattice of text baselines from the
top to the bottom of the page.
@cindex vertical spacing
@cindex spacing, vertical
@cindex vee
@dfn{Vertical spacing} is the distance between adjacent text baselines.
Typographic tradition sets this quantity to 120% of the type size.  The
initial drawing position is one unit of vertical spacing below the page
top.  Typographers term this unit a @slanted{vee}.

@cindex page break
@cindex break, page
Vertical spacing has an impact on page-breaking decisions.  Generally,
when a break occurs, the formatter moves the drawing position to the
next text baseline automatically.  If the formatter were already writing
to the last line that would fit on the page, advancing by one vee would
place the next text baseline off the page.  Rather than let that happen,
@code{roff} formatters instruct the output driver to eject the page,
start a new one, and again set the drawing position to one vee below the
page top; this is a @dfn{page break}.

When the last line of input text corresponds to the last output line
that fits on the page, the break caused by the end of input will also
break the page, producing a useless blank one.  Macro packages keep
users from having to confront this difficulty by setting ``traps''
(@pxref{Traps}); moreover, all but the simplest page layouts tend to
have headers and footers, or at least bear vertical margins larger than
one vee.


@c =====================================================================
@c TODO: Add a section here about interpolations and input processing.
@c
@c We need to level up the reader's macro brain from reasoning about
@c interpolation at the scope of input lines to interpolations _within_
@c lines.  It is also a good time to introduce the \n and \* escape
@c sequences to avoid painful, "WTF"-producing forward references later.
@c Some materal from groff_mm(7) might be adaptable to this purpose.
@c
@c Earlier material from @Defesc{\\n}:
@c "This means that the value of the register is expanded in place while
@c GNU @code{troff} is parsing the input line.  Nested assignments (also
@c called indirect assignments) are possible."
@c
@c We can probably drop the term "indirect assignments"; there's nothing
@c special about these--they are a consequence of *roffs' left-to-right
@c parsing and they apply to escape sequences in general.
@c =====================================================================

@c BEGIN Keep (roughly) parallel with section "Measurements" of
@c groff(7).
@node Measurements, Numeric Expressions, Text, GNU troff Reference
@section Measurements
@cindex measurements
@cindex scaling indicator
@cindex indicator, scaling

@cindex units of measurement
@cindex measurement units
The formatter sometimes requires the input of numeric parameters to
specify measurements.  These are specified as integers or decimal
fractions with an optional @dfn{scaling unit} suffixed.  A scaling unit
is a letter that immediately follows the last digit of a number.  Digits
after the decimal point are optional.  Measurement expressions include
@samp{10.5p}, @samp{11i}, and @samp{3.c}.

@cindex basic units, conversion to
@cindex units, basic, conversion to
@cindex conversion to basic units
Measurements are scaled by the scaling unit and stored internally (with
any fractional part discarded) in basic units.
@cindex device resolution, obtaining in the formatter
@cindex resolution, device, obtaining in the formatter
The device resolution can therefore be obtained by storing a value of
@samp{1i} to a register.  The only constraint on the basic unit is that
it is at least as small as any other unit.
@c That's a fib.  A device resolution of around 2^31 would surely also
@c cause problems.  But nobody does that.

@table @code
@cindex basic scaling unit (@code{u})
@cindex @code{u} scaling unit
@cindex unit, scaling, @code{u}
@cindex scaling unit @code{u}
@item u
Basic unit.

@item i
@cindex inch scaling unit (@code{i})
@cindex @code{i} scaling unit
@cindex unit, scaling, @code{i}
@cindex scaling unit @code{i}
Inch; defined as 2.54@tie{}centimeters.

@item c
@cindex centimeter scaling unit (@code{c})
@cindex @code{c} scaling unit
@cindex unit, scaling, @code{c}
@cindex scaling unit @code{c}
Centimeter; a centimeter is about 0.3937@tie{}inches.

@item p
@cindex point scaling unit (@code{p})
@cindex @code{p} scaling unit
@cindex unit, scaling, @code{p}
@cindex scaling unit @code{p}
Point; a typesetter's unit used for measuring type size.
There are 72@tie{}points to an inch.

@item P
@cindex pica scaling unit (@code{P})
@cindex @code{P} scaling unit
@cindex unit, scaling, @code{P}
@cindex scaling unit @code{P}
Pica; another typesetter's unit.  There are 6@tie{}picas to an inch and
12@tie{}points to a pica.

@item s
@itemx z
@xref{Using Fractional Type Sizes}, for a discussion of these units.

@item f
GNU @code{troff} defines this unit to scale decimal fractions in the
interval [0, 1] to 16-bit unsigned integers.  It multiplies a quantity
by 65,536.  @xref{Colors}, for usage.
@end table

The magnitudes of other scaling units depend on the text formatting
parameters in effect.  These are useful when specifying measurements
that need to scale with the typeface or vertical spacing.

@table @code
@item m
@cindex em scaling unit (@code{m})
@cindex @code{m} scaling unit
@cindex unit, scaling, @code{m}
@cindex scaling unit @code{m}
Em; an em is equal to the current type size in points.  It is named thus
because it is approximately the width of the letter@tie{}@samp{M}.

@item n
@cindex en scaling unit (@code{n})
@cindex @code{n} scaling unit
@cindex unit, scaling, @code{n}
@cindex scaling unit @code{n}
En; an en is one-half em.

@item v
@cindex vertical space unit (@code{v})
@cindex space, vertical, unit (@code{v})
@cindex vee scaling unit (@code{v})
@cindex @code{v} scaling unit
@cindex unit, scaling, @code{v}
@cindex scaling unit @code{v}
Vee; recall @ref{Page Geometry}.

@item M
@cindex @code{M} scaling unit
@cindex unit, scaling, @code{M}
@cindex scaling unit @code{M}
Hundredth of an em.
@end table

@menu
* Motion Quanta::
* Default Units::
@end menu
@c END Keep (roughly) parallel with section "Measurements" of groff(7).

@c ---------------------------------------------------------------------

@node Motion Quanta, Default Units, Measurements, Measurements
@subsection Motion Quanta
@cindex motion quanta
@cindex quanta, motion

@c BEGIN Keep (roughly) parallel with subsection "Motion quanta" of
@c groff(7).
An output device's basic unit @code{u} is not necessarily its smallest
addressable length; @code{u} can be smaller to avoid problems with
integer roundoff.  The minimum distances that a device can work with in
the horizontal and vertical directions are termed its @dfn{motion
quanta}.  Measurements are rounded to applicable motion quanta.
Half-quantum fractions round toward zero.

@cindex horizontal motion quantum register (@code{.H})
@cindex motion quantum, horizontal, register (@code{.H})
@cindex horizontal resolution register (@code{.H})
@cindex resolution, horizontal, register (@code{.H})
@DefregList {.H}
@DefregListEndx {.V}
These read-only registers interpolate the horizontal and vertical motion
quanta, respectively, of the output device in basic units.
@endDefreg

For example, we might draw short baseline rules on a terminal device as
follows.  @xref{Drawing Requests}.

@Example
.tm \n[.H]
    @error{} 24
.nf
\l'36u' 36u
\l'37u' 37u
    @result{} _ 36u
    @result{} __ 37u
@endExample
@c END Keep (roughly) parallel with subsection "Motion quanta" of
@c groff(7).

@c ---------------------------------------------------------------------

@node Default Units,  , Motion Quanta, Measurements
@subsection Default Units
@cindex default units
@cindex units, default

@c BEGIN Keep (roughly) parallel with subsection "Default units" of
@c groff(7).
A general-purpose register (one created or updated with the @code{nr}
request; see @pxref{Registers}) is implicitly dimensionless, or reckoned
in basic units if interpreted in a measurement context.  But it is
convenient for many requests and escape sequences to infer a scaling
unit for an argument if none is specified.  An explicit scaling unit
(not after a closing parenthesis) can override an undesirable default.
Effectively, the default unit is suffixed to the expression if a scaling
unit is not already present.  GNU @code{troff}'s use of integer
arithmetic should also be kept in mind (@pxref{Numeric Expressions}).

The @code{ll} request interprets its argument in ems by default.
Consider several attempts to set a line length of 3.5@tie{}inches when
the type size is 10@tie{}points on a terminal device with a resolution
of 240 basic units and horizontal motion quantum of 24.  Some
expressions become zero; the request clamps them to that quantum.

@Example
.ll 3.5i      \" 3.5i (= 840u)
.ll 7/2       \" 7u/2u -> 3u -> 3m -> 0, clamped to 24u
.ll (7 / 2)u  \" 7u/2u -> as above
.ll 7/2i      \" 7u/2i -> 7u/480u -> 0 -> as above
.ll 7i/2      \" 7i/2u -> 1680u/2m -> 1680u/24u -> 35u
.ll 7i/2u     \" 3.5i (= 840u)
@endExample

@noindent
@cindex measurements, specifying safely
The safest way to specify measurements is to attach a scaling unit.  To
multiply or divide by a dimensionless quantity, use @samp{u} as its
scaling unit.
@c END Keep (roughly) parallel with subsection "Default units" of
@c groff(7).


@c =====================================================================

@c BEGIN Keep (roughly) parallel with section "Numeric expressions" of
@c groff(7).
@node Numeric Expressions, Identifiers, Measurements, GNU troff Reference
@section Numeric Expressions
@cindex numeric expressions
@cindex expressions, numeric

A @dfn{numeric expression} evaluates to an integer:@: it can be as
simple as a literal @samp{0} or it can be a complex sequence of register
and string interpolations interleaved with measurements and operators.

GNU @code{troff} provides a set of mathematical and logical operators
familiar to programmers---as well as some unusual ones---but supports
only integer arithmetic.@footnote{Provision is made for interpreting and
reporting decimal fractions in certain cases.}  The internal data type
used for computing results is usually a 32-bit signed integer, which
suffices to represent magnitudes within a range of ±2
billion.@footnote{If that's not enough, see the @cite{groff_tmac@r{(5)}}
man page for the @file{62bit.tmac} macro package.}

@cindex arithmetic operators
@cindex operators, arithmetic
@cindex truncating division
@cindex addition
@cindex subtraction
@cindex multiplication
@cindex division, truncating
@cindex modulus
@opindex +
@opindex -
@opindex *
@opindex /
@opindex %
Arithmetic infix operators perform a function on the numeric expressions
to their left and right; they are @code{+} (addition), @code{-}
(subtraction), @code{*} (multiplication), @code{/} (truncating
division), and @code{%} (modulus).  @dfn{Truncating division} rounds to
the integer nearer to zero, no matter how large the fractional portion.
Overflow and division (or modulus) by zero are errors and abort
evaluation of a numeric expression.
@cindex unary arithmetic operators
@cindex operators, unary arithmetic
@cindex negation
@cindex assertion (arithmetic operator)
@opindex -
@opindex +
@cindex @code{if} request, and the @samp{!} operator
@cindex @code{while} request, and the @samp{!} operator

Arithmetic unary operators operate on the numeric expression to their
right; they are @code{-} (negation) and @code{+} (assertion---for
completeness; it does nothing).  The unary minus must often be used
with parentheses to avoid confusion with the decrementation operator,
discussed below.

Observe the rounding behavior and effect of negative operands on the
modulus and truncating division operators.

@Example
.nr T 199/100
.nr U 5/2
.nr V (-5)/2
.nr W 5/-2
.nr X 5%2
.nr Y (-5)%2
.nr Z 5%-2
T=\n[T] U=\n[U] V=\n[V] W=\n[W] X=\n[X] Y=\n[Y] Z=\n[Z]
    @result{} T=1 U=2 V=-2 W=-2 X=1 Y=-1 Z=1
@endExample

@noindent
The sign of the modulus of operands of mixed signs is determined by the
sign of the first.  Division and modulus operators satisfy the following
property:@: given a dividend@tie{}@var{a} and a divisor@tie{}@var{b}, a
quotient@tie{}@var{q} formed by @samp{(a / b)} and a
remainder@tie{}@var{r} by @samp{(a % b)}, then @math{qb + r = a}.

@cindex scaling operator
@cindex operator, scaling
@opindex ;
GNU @code{troff}'s scaling operator, used with parentheses as
@code{(@var{c};@var{e})}, evaluates a numeric expression@tie{}@var{e}
using@tie{}@var{c} as the default scaling unit.  If @var{c} is omitted,
scaling units are ignored in the evaluation of@tie{}@var{e}.  This
operator can save typing by avoiding the attachment of scaling units to
every operand out of caution.  Your macros can select a sensible default
unit in case the user neglects to supply one.

@Example
.\" Indent by amount given in first argument; assume ens.
.de Indent
.  in (n;\\$1)
..
@endExample

@noindent
Without the scaling operator, the foregoing macro would, if called with
a unitless argument, cause indentation by the @code{in} request's
default scaling unit (ems).  The result would be twice as much
indentation as expected.

@cindex extremum operators (@code{>?}, @code{<?})
@cindex operators, extremum (@code{>?}, @code{<?})
@cindex maximum operator
@cindex minimum operator
@opindex >?
@opindex <?
GNU @code{troff} also provides a pair of operators to compute the
extrema of two operands: @code{>?} (maximum) and @code{<?} (minimum).

@Example
.nr slots 5
.nr candidates 3
.nr salaries (\n[slots] <? \n[candidates])
Looks like we'll end up paying \n[salaries] salaries.
    @result{} Looks like we'll end up paying 3 salaries.
@endExample

@cindex comparison operators
@cindex operators, comparison
@cindex greater than (or equal to) operator
@cindex less than (or equal to) operator
@cindex equality operator
@opindex <
@opindex >
@opindex >=
@opindex <=
@opindex =
@opindex ==
Comparison operators comprise @code{<} (less than), @code{>} (greater
than), @code{<=} (less than or equal), @code{>=} (greater than or
equal), and @code{=} (equal).  @code{==} is a synonym for @code{=}.
When evaluated, a comparison is replaced with @samp{0} if it is false
and @samp{1} if true.  In the @code{roff} language, positive values are
true, others false.

@cindex logical operators
@cindex operators, logical
@cindex logical ``and'' operator
@cindex logical conjunction operator
@cindex logical ``or'' operator
@cindex logical disjunction operator
@opindex &
@ifnotinfo
@opindex :
@end ifnotinfo
@ifinfo
@opindex @r{<colon>}
@end ifinfo
We can operate on truth values with the logical operators @code{&}
(logical conjunction or ``and'') and @code{:} (logical disjunction or
``or'').  They evaluate as comparison operators do.

@opindex !
@cindex complementation, logical
@cindex logical complementation operator
@cindex logical not, limitation in expression
@cindex expression, limitation of logical not in
A logical complementation (``not'') operator, @code{!}, works only
within @code{if}, @code{ie}, and @code{while} requests.
@c This is worded to avoid implying that the operator doesn't apply
@c to conditional expressions in general, albeit without mentioning them
@c because they're out of scope.
Furthermore, @code{!} is recognized only at the beginning of a numeric
expression not contained by another numeric expression.  In other words,
it must be the ``outermost'' operator.  Including it elsewhere in the
expression produces a warning in the @samp{number} category
(@pxref{Warnings}), and its expression evaluates false.  This
unfortunate limitation maintains compatibility with @acronym{AT&T}
@code{troff}.  You can test a numeric expression for falsity by
comparing it to a false value.

@Example
.nr X 1
.nr Y 0
.\" This does not work as expected.
.if (\n[X])&(!\n[Y]) .nop A: X is true, Y is false
.
.\" Use this construct instead.
.if (\n[X])&(\n[Y]<=0) .nop B: X is true, Y is false
    @error{} warning: expected numeric expression, got '!'
    @result{} B: X is true, Y is false
@endExample

@cindex parentheses
@cindex order of evaluation in expressions
@cindex expression, order of evaluation
@opindex (
@opindex )
The @code{roff} language has no operator precedence:@: expressions are
evaluated strictly from left to right, in contrast to schoolhouse
arithmetic.  Use parentheses @code{(} @code{)} to impose a desired
precedence upon subexpressions.

@Example
.nr X 3+5*4
.nr Y (3+5)*4
.nr Z 3+(5*4)
X=\n[X] Y=\n[Y] Z=\n[Z]
    @result{} X=32 Y=32 Z=23
@endExample

@cindex @code{+}, and page motion
@cindex @code{-}, and page motion
@cindex motion operators
@cindex operators, motion
@opindex + @r{(unary)}
@opindex - @r{(unary)}
For many requests and escape sequences that cause motion on the page,
the unary operators @code{+} and @code{-} work differently when leading
a numeric expression.  They then indicate a motion relative to the
drawing position:@: positive is down in vertical contexts, right in
horizontal ones.

@cindex @code{bp} request, using @code{+} and@tie{}@code{-} with
@cindex @code{in} request, using @code{+} and@tie{}@code{-} with
@cindex @code{ll} request, using @code{+} and@tie{}@code{-} with
@cindex @code{lt} request, using @code{+} and@tie{}@code{-} with
@cindex @code{nm} request, using @code{+} and@tie{}@code{-} with
@cindex @code{nr} request, using @code{+} and@tie{}@code{-} with
@cindex @code{pl} request, using @code{+} and@tie{}@code{-} with
@cindex @code{pn} request, using @code{+} and@tie{}@code{-} with
@cindex @code{po} request, using @code{+} and@tie{}@code{-} with
@cindex @code{ps} request, using @code{+} and@tie{}@code{-} with
@cindex @code{pvs} request, using @code{+} and@tie{}@code{-} with
@cindex @code{rt} request, using @code{+} and@tie{}@code{-} with
@cindex @code{ti} request, using @code{+} and@tie{}@code{-} with
@cindex @code{\H}, using @code{+} and@tie{}@code{-} with
@cindex @code{\R}, using @code{+} and@tie{}@code{-} with
@cindex @code{\s}, using @code{+} and@tie{}@code{-} with
@code{+} and @code{-} are also treated differently by the following
requests and escape sequences:@: @code{bp}, @code{in}, @code{ll},
@code{lt}, @code{nm}, @code{nr}, @code{pl}, @code{pn}, @code{po},
@code{ps}, @code{pvs}, @code{rt}, @code{ti}, @code{\H}, @code{\R}, and
@code{\s}.  Here, leading plus and minus signs serve as incrementation
and decrementation operators, respectively.  To negate an expression,
subtract it from zero or include the unary minus in parentheses with its
argument.  @xref{Setting Registers}, for examples.

@opindex |
@cindex @code{|}, and page motion
@cindex absolute @slanted{(sic)} position operator (@code{|})
@cindex position, absolute @slanted{(sic)} operator (@code{|})
@cindex boundary-relative motion operator (@code{|})
@c "motion" and "operators" already indexed above
A leading @code{|} operator indicates a motion relative not to the
drawing position but to a boundary.  For horizontal motions, the
measurement specifies a distance relative to a drawing position
corresponding to the beginning of the @emph{input} line.  By default,
tab stops reckon movements in this way.  Most escape sequences do not;
@c XXX: Which ones do?
@code{|} tells them to do so.

@Example
Mind the \h'1.2i'gap.
.br
Mind the \h'|1.2i'gap.
.br
Mind the
\h'|1.2i'gap.
@c 13 spaces, 4 spaces, 13 spaces
    @result{} Mind the             gap.
    @result{} Mind the    gap.
    @result{} Mind the             gap.
@endExample

One use of this feature is to define macros whose scope is limited to
the output they format.

@Example
.\" underline word $1 with trailing punctuation $2
.de Underline
.  nop \\$1\l'|0\[ul]'\\$2
..
Typographical emphasis is best used
.Underline sparingly .
@endExample

@noindent
In the above example, @samp{|0} specifies a negative motion from the
current position (at the end of the argument just emitted, @code{\$1})
to the beginning of the input line.  Thus, the @code{\l} escape sequence
in this case draws a line from right to left.  A macro call occurs at
the beginning of an input line;@footnote{Control structure syntax
creates an exception to this rule, but is designed to remain useful:
recalling our example, @samp{.if 1 .Underline this} would underline only
``this'', precisely.  @xref{Conditionals and Loops}.} if the @code{|}
operator were omitted, then the underline would be drawn at zero
distance from the current position, producing device-dependent, and
likely undesirable, results.  On the @samp{ps} output device, it
underlines the period.

For vertical movements, the @code{|} operator specifies a distance from
the first text baseline on the page or in the current diversion, using
the current vertical spacing.

@Example
A
.br
B \Z'C'\v'|0'D
    @result{} A D
    @result{} B C
@endExample

In the foregoing example, we've used the @code{\Z} escape sequence
(@pxref{Page Motions}) to restore the drawing position after formatting
@samp{C}, then moved vertically to the first text baseline on the page.

@Defesc {\\B, @code{'}, anything, @code{'}}
@cindex numeric expression, valid
@cindex valid numeric expression
Interpolate@tie{}1 if @var{anything} is a valid numeric expression,
and@tie{}0 otherwise.  The delimiter need not be a neutral apostrophe;
see @ref{Delimiters}.
@endDefesc

You might use @code{\B} along with the @code{if} request to filter out
invalid macro or string arguments.  @xref{Conditionals and Loops}.

@Example
.\" Indent by amount given in first argument; assume ens.
.de Indent
.  if \B'\\$1' .in (n;\\$1)
..
@endExample

A register interpolated as an operand in a numeric expression must have
an Arabic format; luckily, this is the default.  @xref{Assigning
Register Formats}.

@cindex space characters, in expressions
@cindex expressions, and space characters
Because spaces separate arguments to requests, spaces are not allowed in
numeric expressions unless the (sub)expression containing them is
surrounded by parentheses.  @xref{Invoking Requests}, and
@ref{Conditionals and Loops}.

@Example
.nf
.nr a 1+2 + 2+1
\na
    @error{} expected numeric expression, got a space
    @result{} 3
.nr a 1+(2 + 2)+1
\na
    @result{} 6
@endExample

The @code{nr} request (@pxref{Setting Registers}) expects its second and
optional third arguments to be numeric expressions; a bare @code{+} does
not qualify, so our first attempt got a warning.
@c END Keep (roughly) parallel with section "Numeric expressions" of
@c groff(7).


@c =====================================================================

@c BEGIN Keep (roughly) parallel with section "Identifiers" of groff(7).
@node Identifiers, Formatter Instructions, Numeric Expressions, GNU troff Reference
@section Identifiers
@cindex identifiers

An @dfn{identifier} labels a GNU @code{troff} datum such as a register,
name (macro, string, or diversion), typeface, color, special character,
character class, environment, or stream.  Valid identifiers consist of
one or more ordinary characters.
@cindex ordinary character
@cindex character, ordinary
An @slanted{ordinary character} is an input character that is not a
leader, tab, newline, or invalid as GNU @code{troff} input.

@c XXX: We might move this discussion earlier since it is applicable to
@c troff input in general, and include a reference to the `trin`
@c request.
@cindex invalid input characters
@cindex input characters, invalid
@cindex characters, invalid input
@cindex Unicode
Invalid input characters are a subset of control characters (from the
sets ``C0 Controls'' and ``C1 Controls'' as Unicode describes them).
When GNU @code{troff} encounters one in an identifier, it produces a
warning in category @samp{input} (@pxref{Warnings}).  They are removed
during interpretation: an identifier @samp{foo}, followed by an invalid
character and then @samp{bar}, is processed as @samp{foobar}.

On a machine using the ISO 646, 8859, or 10646 character encodings,
invalid input characters are @code{0x00}, @code{0x08}, @code{0x0B},
@code{0x0D}--@code{0x1F}, and @code{0x80}--@code{0x9F}.  On an
@acronym{EBCDIC} host, they are @code{0x00}--@code{0x01}, @code{0x08},
@code{0x09}, @code{0x0B}, @code{0x0D}--@code{0x14},
@code{0x17}--@code{0x1F}, and
@code{0x30}--@code{0x3F}.@footnote{Historically, control characters like
ASCII STX, ETX, and BEL (@key{Control+B}, @key{Control+C}, and
@key{Control+G}) have been observed in @code{roff} documents,
particularly in macro packages employing them as delimiters with the
output comparison operator to try to avoid collisions with the content
of arbitrary user-supplied parameters (@pxref{Operators in
Conditionals}).  We discourage this expedient; in GNU @code{troff} it is
unnecessary (outside of compatibility mode) because delimited arguments
are parsed at a different input level than the surrounding context.
@xref{Implementation Differences}.}  Some of these code points are used
by GNU @code{troff} internally, making it non-trivial to extend the
program to accept UTF-8 or other encodings that use characters from
these ranges.@footnote{Consider what happens when a C1 control
@code{0x80}--@code{0x9F} is necessary as a continuation byte in a UTF-8
sequence.}

Thus, the identifiers @samp{br}, @samp{PP}, @samp{end-list},
@samp{ref*normal-print}, @samp{|}, @samp{@@_}, and @samp{!"#$%'()*+,-./}
are all valid.  Discretion should be exercised to prevent confusion.
Identifiers starting with @samp{(} or @samp{[} require care.

@Example
.nr x 9
.nr y 1
.nr (x 2
.nr [y 3
.nr sum1 (\n(x + \n[y])
    @error{} a space character is not allowed in an escape
    @error{}   sequence parameter
A:2+3=\n[sum1]
.nr sum2 (\n((x + \n[[y])
B:2+3=\n[sum2]
.nr sum3 (\n[(x] + \n([y)
C:2+3=\n[sum3]
    @result{} A:2+3=1 B:2+3=5 C:2+3=5
@endExample

@cindex @code{]}, as part of an identifier
@noindent
An identifier with a closing bracket (@samp{]}) in its name can't be
accessed with bracket-form escape sequences that expect an identifier as
a parameter.  For example, @samp{\[foo]]} accesses the glyph @samp{foo},
followed by @samp{]} in whatever the surrounding context is, whereas
@samp{\C'foo]'} formats a glyph named @samp{foo]}.  Similarly, the
identifier @samp{(} can't be interpolated @emph{except} with bracket
forms.

@cindex @code{refer}, and macro names starting with @code{[} or @code{]}
@cindex @code{[}, macro names starting with, and @code{refer}
@cindex @code{]}, macro names starting with, and @code{refer}
@cindex macro names, starting with @code{[} or @code{]}, and @code{refer}
If you begin a macro, string, or diversion name with either of the
characters @samp{[} or @samp{]}, you foreclose use of the @code{grefer}
preprocessor, which recognizes @samp{.[} and @samp{.]} as bibliographic
reference delimiters.

@Defesc {\\A, @code{'}, anything, @code{'}}
Interpolate@tie{}1 if @var{anything} is a valid identifier, and@tie{}0
otherwise.  The delimiter need not be a neutral apostrophe; see
@ref{Delimiters}.  Because invalid input characters are removed (see
above), invalid identifiers are empty or contain spaces, tabs, or
newlines.

You can employ @code{\A} to validate a macro argument before using it to
construct another escape sequence or identifier.

@Example
.\" usage: .init-coordinate-pair name val1 val2
.\" Create a coordinate pair where name!x=val1 and
.\" name!y=val2.
.de init-coordinate-pair
.  if \A'\\$1' \@{\
.    if \B'\\$2' .nr \\$1!x \\$2
.    if \B'\\$3' .nr \\$1!y \\$3
.  \@}
..
.init-coordinate-pair center 5 10
The center is at (\n[center!x], \n[center!y]).
.init-coordinate-pair "poi@arrow{}nt" trash garbage \" ignored
.init-coordinate-pair point trash garbage \" ignored
    @result{} The center is at (5, 10).
@endExample

@noindent
In this example, we also validated the numeric arguments; the registers
@samp{point!x} and @samp{point!y} remain undefined.  @xref{Numeric
Expressions} for the @code{\B} escape sequence.
@endDefesc

@cindex undefined identifiers
@cindex identifiers, undefined
How GNU @code{troff} handles the interpretation of an undefined
identifier depends on the context.  There is no way to invoke an
undefined request; such syntax is interpreted as a macro call instead.
If the identifier is being interpreted as a string, macro, or diversion,
GNU @code{troff} emits a warning in category @samp{mac}, defines it as
empty, and interpolates nothing.  If the identifier is being interpreted
as a register, GNU @code{troff} emits a warning in category @samp{reg},
initializes it to zero, and interpolates that value.  @xref{Warnings},
@ref{Interpolating Registers}, and @ref{Strings}.  Attempting to use an
undefined typeface, glyph, color, character class, environment, or
stream generally provokes an error diagnostic.

@need 1000
@cindex name space, common, of macros, diversions, and strings
@cindex common name space of macros, diversions, and strings
@cindex macros, shared name space with strings and diversions
@cindex strings, shared name space with macros and diversions
@cindex diversions, shared name space with macros and strings
Identifiers for requests, macros, strings, and diversions share one name
space; special characters and character classes another.  No other
object types do.

@Example
.de xxx
.  nop foo
..
@c . slack line for pagination management
.di xxx
bar
.br
.di
.
.xxx
    @result{} bar
@endExample

@noindent
The foregoing example shows that GNU @code{troff} reuses the identifier
@samp{xxx}, changing it from a macro to a diversion.  No warning is
emitted, and the previous contents of @samp{xxx} are lost.
@c END Keep (roughly) parallel with section "Identifiers" of groff(7).


@c =====================================================================

@node Formatter Instructions, Registers, Identifiers, GNU troff Reference
@section Formatter Instructions
@cindex formatter instructions
@cindex instructing the formatter

To support documents that require more than filling, automatic line
breaking and hyphenation, adjustment, and supplemental inter-sentence
space, the @code{roff} language offers two means of embedding
instructions to the formatter.

@cindex request
One is a @dfn{request}, which begins with a control character and takes
up the remainder of the input line.  Requests often perform relatively
large-scale operations such as setting the page length, breaking the
line, or starting a new page.  They also conduct internal operations
like defining macros.

@cindex escape sequence
@cindex sequence, escape
The other is an @dfn{escape sequence}, which begins with the escape
character and can be embedded anywhere in the input, even in arguments
to requests and other escape sequences.  Escape sequences interpolate
special characters, strings, or registers, and handle comparatively
minor formatting tasks like sub- and superscripting.

Some operations, such as font selection and type size alteration, are
available via both requests and escape sequences.

@menu
* Control Characters::
* Invoking Requests::
* Calling Macros::
* Using Escape Sequences::
* Delimiters::
@end menu

@c ---------------------------------------------------------------------

@node Control Characters, Invoking Requests, Formatter Instructions, Formatter Instructions
@subsection Control Characters
@cindex control characters
@cindex configuring control characters
@cindex changing control characters

The mechanism of using @code{roff}'s control characters to invoke
requests and call macros was introduced in @ref{Requests and Macros}.
Control characters are recognized only at the beginning of an input
line, or at the beginning of the branch of a control structure request;
see @ref{Conditionals and Loops}.

A few requests cause a break implicitly; use the no-break control
character to prevent the break.  Break suppression is its sole
behavioral distinction.  Employing the no-break control character to
invoke requests that don't cause breaks is harmless but poor style.
@xref{Manipulating Filling and Adjustment}.

@cindex control character, changing (@code{cc})
@cindex character, control, changing (@code{cc})
@cindex no-break control character, changing (@code{c2})
@cindex character, no-break control, changing (@code{c2})
@cindex control character, no-break, changing (@code{c2})
The control @samp{.} and no-break control @samp{'} characters can each
be changed to any ordinary character@footnote{Recall @ref{Identifiers}.}
with the @code{cc} and @code{c2} requests, respectively.

@Defreq {cc, [@Var{o}]}
Recognize the ordinary character@tie{}@var{o} as the control character.
If@tie{}@var{o} is absent or invalid, the default control character
@samp{.} is selected.  The identity of the control character is
associated with the environment (@pxref{Environments}).
@endDefreq

@Defreq {c2, [@Var{o}]}
Recognize the ordinary character@tie{}@var{o} as the no-break control
character.  If@tie{}@var{o} is absent or invalid, the default no-break
control character @samp{'} is selected.  The identity of the no-break
control character is associated with the environment
(@pxref{Environments}).
@endDefreq

When writing a macro, you might wish to know which control character was
used to call it.

@Defreg {.br}
This read-only register interpolates@tie{}1 if the currently executing
macro was called using the normal control character and@tie{}0
otherwise.  If a macro is interpolated as a string, the @code{.br}
register's value is inherited from the context of the string
interpolation.  @xref{Strings}.

@cindex intercepting requests
@cindex requests, intercepting
@cindex modifying requests
@cindex requests, modifying
Use this register to reliably intercept requests that imply breaks.

@Example
.als bp*orig bp
.de bp
.  ie \\n[.br] .bp*orig
.  el          'bp*orig
..
@endExample

Testing the @code{.br} register outside of a macro definition makes no
sense.
@endDefreg

@c ---------------------------------------------------------------------

@c BEGIN Keep (roughly) parallel with section "Requests" of groff(7).
@node Invoking Requests, Calling Macros, Control Characters, Formatter Instructions
@subsection Invoking Requests
@cindex invoking requests
@cindex requests, invoking

A control character is optionally followed by tabs and/or spaces and
then an identifier naming a request or macro.  The invocation of an
unrecognized request is interpreted as a macro call.  Defining a macro
with the same name as a request replaces the request.  Deleting a
request name with the @code{rm} request makes it unavailable.  The
@code{als} request can alias requests, permitting them to be wrapped or
non-destructively replaced.  @xref{Strings}.

@cindex request arguments
@cindex arguments to requests
@cindex tabs, and macro arguments
@cindex macro arguments, and tabs
@cindex arguments to macros, and tabs
@cindex tabs, and request arguments
@cindex request arguments, and tabs
@cindex arguments to requests, and tabs
There is no general limit on argument length or quantity.  Most
requests take one or more arguments, and ignore any they do not expect.
A request may be separated from its arguments by tabs or spaces, but
only spaces can separate an argument from its successor.  Only one
between arguments is necessary; any excess is ignored.  GNU @code{troff}
does not allow tabs for argument separation.@footnote{In compatibility
mode, a space is not necessary after a request or macro name of two
characters' length.  Also, Plan@tie{}9 @code{troff} allows tabs to
separate arguments.}

Generally, a space @emph{within} a request argument is not relevant, not
meaningful, or is supported by bespoke provisions, as with the @code{tl}
request's delimiters (@pxref{Page Layout}).  Some requests, like
@code{ds}, interpret the remainder of the control line as a single
argument.  @xref{Strings}.

@need 1000
@cindex structuring source code of documents or macro packages
@cindex documents, structuring the source of
@cindex macro package, structuring the source of
@cindex package, package, structuring the source of
@cindex indentation, of @code{roff} source code
Spaces and tabs immediately after a control character are ignored.
Commonly, authors structure the source of documents or macro files with
them.

@Example
.de center
.  if \\n[.br] \
.    br
.  ce \\$@@
..
.
.
.de right-align
.@arrow{}if \\n[.br] \
.@arrow{}@arrow{}br
.@arrow{}rj \\$@@
..
@endExample

@cindex blank line trap (@code{blm})
@cindex blank line macro (@code{blm})
If you assign an empty blank line trap, you can separate macro
definitions (or any input lines) with blank lines.

@Example
.de do-nothing
..
.blm do-nothing  \" activate blank line trap

.de center
.  if \\n[.br] \
.    br
.  ce \\$@@
..


.de right-align
.@arrow{}if \\n[.br] \
.@arrow{}@arrow{}br
.@arrow{}rj \\$@@
..

.blm             \" deactivate blank line trap
@endExample

@xref{Blank Line Traps}.
@c END Keep (roughly) parallel with section "Requests" of groff(7).

@c ---------------------------------------------------------------------

@need 1000
@node Calling Macros, Using Escape Sequences, Invoking Requests, Formatter Instructions
@subsection Calling Macros
@cindex calling macros
@cindex macro arguments
@cindex arguments to macros

If a macro of the desired name does not exist when called, it is
created, assigned an empty definition, and a warning in category
@samp{mac} is emitted.  Calling an undefined macro @emph{does} end a
macro definition naming it as its end macro (@pxref{Writing Macros}).

@cindex spaces, in a macro argument
To embed spaces @emph{within} a macro argument, enclose the argument in
neutral double quotes @code{"}.  Horizontal motion escape sequences are
sometimes a better choice for arguments to be formatted as text.

Consider calls to a hypothetical section heading macro @samp{uh}.

@Example
.uh The Mouse Problem
.uh "The Mouse Problem"
.uh The\~Mouse\~Problem
.uh The\ Mouse\ Problem
@endExample

@cindex @code{\~}, difference to @code{\@key{SP}}
@cindex @code{\@key{SP}}, difference to @code{\~}
@noindent
The first line calls @code{uh} with three arguments: @samp{The},
@samp{Mouse}, and @samp{Problem}.  The remainder call the @code{uh}
macro with one argument, @samp{The Mouse Problem}.  The last solution,
using escaped spaces, can be found in documents prepared for
@acronym{AT&T} @code{troff}.  It can cause surprise when text is
adjusted, because @code{\@key{SP}} inserts a @emph{fixed-width},
non-breaking space.  GNU @code{troff}'s @code{\~} escape sequence
inserts an adjustable, non-breaking space.@footnote{@code{\~} is fairly
portable; see @ref{Other Differences}.}

@cindex @code{"}, embedding in a macro argument
@cindex double quote, embedding in a macro argument
@cindex @code{\}, embedding in a macro argument
@cindex backslash, embedding in a macro argument
The foregoing raises the question of how to embed neutral double quotes
or backslashes in macro arguments when @emph{those} characters are
desired as literals.  In GNU @code{troff}, the special character escape
sequence @code{\[rs]} produces a backslash and @code{\[dq]} a neutral
double quote.

In GNU @code{troff}'s @acronym{AT&T} compatibility mode, these
characters remain available as @code{\(rs} and @code{\(dq},
respectively.  @acronym{AT&T} @code{troff} did not consistently define
these special characters,
@c It seems that AT&T troff never recognized \(rs, though DWB 3.3
@c defined \(bs as an alias of "\" on its "Latin1" device, in
@c deliberate(?) collision with the Bell System logo identifier.  It
@c also defined \(dq for several devices (pcl, Latin1, nroff, ...) along
@c with \(aq.
but its descendants can be made to support them.  @xref{Device and Font
Description Files}.

If even that is not feasible, options remain.  To obtain a literal
escape character in a macro argument, you can simply type it if you
change or disable the escape character first.  @xref{Using Escape
Sequences}.  Otherwise, you must escape the escape character repeatedly
to a context-dependent extent.  @xref{Copy Mode}.

For the (neutral) double quote, you have recourse to an obscure
syntactical feature of @acronym{AT&T} @code{troff}.  Because a double
quote can begin a macro argument, the formatter keeps track of whether
the current argument was started thus, and doesn't require a space after
the double quote that ends it.@footnote{Strictly, you can neglect to
close the last quoted macro argument, relying on the end of the control
line to do so.   We consider this lethargic practice poor style.}  In
the argument list to a macro, a double quote that @emph{isn't} preceded
by a space @emph{doesn't} start a macro argument.  If not preceded by a
double quote that began an argument, this double quote becomes part of
the argument.  Furthermore, within a quoted argument, a pair of adjacent
double quotes becomes a literal double quote.

@Example
.de eq
.  tm arg1:\\$1 arg2:\\$2 arg3:\\$3
.  tm arg4:\\$4 arg5:\\$5 arg6:\\$6
.. \" 4 backslashes on the next line
.eq a" "b c" "de"f\\\\g" h""i "j""k"
    @error{} arg1:a" arg2:b c arg3:de
    @error{} arg4:f\g" arg5:h""i arg6:j"k
@endExample

Apart from the complexity of the rules, this traditional solution has
the disadvantage that double quotes don't survive repeated argument
expansion in @acronym{AT&T} @code{troff} or GNU @code{troff}'s
compatibility mode.  This can frustrate efforts to pass such arguments
intact through multiple macro calls.

@Example
.cp 1
.de eq
.  tm arg1:\\$1 arg2:\\$2 arg3:\\$3
.  tm arg4:\\$4 arg5:\\$5 arg6:\\$6
..
.de xe
.  eq \\$1 \\$2 \\$3 \\$4 \\$5 \\$6
.. \" 8 backslashes on the next line
.xe a" "b c" "de"f\\\\\\\\g" h""i "j""k"
    @error{} arg1:a" arg2:b arg3:c
    @error{} arg4:de arg5:f\g" arg6:h""i
@endExample

@cindex input level
@cindex level, input
@cindex interpolation depth
@cindex depth, interpolation
Outside of compatibility mode, GNU @code{troff} doesn't exhibit this
problem because it tracks the nesting depth of interpolations.
@xref{Implementation Differences}.

@c ---------------------------------------------------------------------

@c BEGIN Keep (roughly) parallel with section "Using escape sequences"
@c of groff(7).
@node Using Escape Sequences, Delimiters, Calling Macros, Formatter Instructions
@subsection Using Escape Sequences
@cindex using escape sequences
@cindex escape sequences

Whereas requests must occur on control lines, escape sequences can occur
intermixed with text and may appear in arguments to requests, macros,
and other escape sequences.
@esindex \
An escape sequence is introduced by the escape character, a backslash
@code{\} (but see the @code{ec} request below).  The next character
selects the escape's function.

Escape sequences vary in length.  Some take an argument, and of those,
some have different syntactical forms for a one-character,
two-character, or arbitrary-length argument.  Others accept @emph{only}
an arbitrary-length argument.  In the former scheme, a one-character
argument follows the function character immediately, an opening
parenthesis @samp{(} introduces a two-character argument (no closing
parenthesis is used), and an argument of arbitrary length is enclosed in
brackets @samp{[]}.  In the latter scheme, the user selects a delimiter
character.  A few escape sequences are idiosyncratic, and support both
of the foregoing conventions (@code{\s}), designate their own
termination sequence (@code{\?}), consume input until the next newline
(@code{\!}, @code{\"}, @code{\#}), or support an additional modifier
character (@code{\s} again, and @code{\n}).  As with requests, use of
some escape sequences in source documents may interact poorly with a
macro package you use; consult its documentation to learn of ``safe''
sequences or alternative facilities it provides to achieve the desired
result.

If an escape character is followed by a character that does not
identify a defined operation, the escape character is ignored (producing
a diagnostic of the @samp{escape} warning category, which is not enabled
by default) and the following character is processed normally.

@Example
$ groff -Tps -ww
.nr N 12
.ds co white
.ds animal elephant
I have \fI\nN \*(co \*[animal]s,\f[]
said \P.\&\~Pseudo Pachyderm.
    @error{} warning: escape character ignored before 'P'
    @result{} I have @slanted{12 white elephants,} said P. Pseudo Pachyderm.
@endExample

Escape sequence interpolation is of higher precedence than escape
sequence argument interpretation.  This rule affords flexibility in
using escape sequences to construct parameters to other escape
sequences.
@c END Keep (roughly) parallel with section "Escape sequences" of
@c groff(7).

@Example
.ds family C\" Courier
.ds style I\" oblique
Choice a typeface \f(\*[family]\*[style]wisely.
    @result{} Choose a typeface @slanted{wisely.}
@endExample

@noindent
In the above, the syntax form @samp{\f(} accepts only two characters for
an argument; the example works because the subsequent escape sequences
are interpolated before the selection escape sequence argument is
processed, and strings @code{family} and @code{style} interpolate one
character each.@footnote{The omission of spaces before the comment
escape sequences is necessary; see @ref{Strings}.}

@c @need 1000
The escape character is nearly always interpreted when encountered; it
is therefore desirable to have a way to interpolate it.

@cindex formatting the escape character (@code{\e})
@cindex escape character, formatting (@code{\e})
@Defesc {\\e, , , }
Interpolate the escape character.
@endDefesc

@cindex formatting a backslash glyph (@code{\[rs]})
@cindex backslash glyph, formatting (@code{\[rs]})
The @code{\[rs]} special character escape sequence formats a backslash
glyph.  In macro and string definitions, the input sequences @code{\\}
and @code{\E} defer interpretation of escape sequences.  @xref{Copy
Mode}.

@Defreq {eo, }
@cindex disabling @code{\} (@code{eo})
@cindex @code{\}, disabling (@code{eo})
Disable the escape mechanism except in copy mode.  Once this request is
invoked, no input character is recognized as starting an escape
sequence in interpretation mode.
@endDefreq

@Defreq {ec, [@Var{o}]}
@cindex escape character, changing (@code{ec})
@cindex character, escape, changing (@code{ec})
Recognize the ordinary character@tie{}@var{o} as the escape character.
If@tie{}@var{o} is absent or invalid, the default escape character
@samp{\} is selected.
@endDefreq

Switching escape sequence interpretation off to define a macro and back
on afterward can obviate the need to double the escape character within
the definition.  @xref{Writing Macros}.  This technique is not available
if your macro needs to interpolate values at the time it is
@emph{defined}---but many do not.

@Example
.\" simplified `BR` macro from the man(7) macro package
.eo
.de BR
.  ds result \&
.  while (\n[.$] >= 2) \@{\
.    as result \fB\$1\fR\$2\"
.    shift 2
.  \@}
.  if \n[.$] .as result \fB\$1\"
\*[result]
.  rm result
.  ft R
..
.ec
@endExample

@DefreqList {ecs, }
@DefreqListEndx {ecr, }
The @code{ecs} request stores the escape character for recall with
@code{ecr}.  @code{ecr} sets the escape character to @samp{\} if none
has been saved.

Use these requests together to temporarily change the escape character.
@endDefreq

Using a different escape character, or disbling it, when calling macros
not under your control will likely cause errors, since GNU @code{troff}
has no mechanism to ``intern'' macros---that is, to convert a macro
definition into a form independent of its
representation.@footnote{@TeX{} does have such a mechanism.}  When a
macro is called, its contents are interpreted literally.
@c XXX: all that stuff mapped into the C0 and C1 controls seems pretty
@c close to an interning mechanism to me, though... --GBR

@c XXX: Motivation?  Why are we directing the reader to these?
@c @xref{Diversions}, and @ref{Identifiers}.

@c BEGIN Keep (roughly) parallel with subsection "Delimiters" of
@c groff(7).
@node Delimiters, , Using Escape Sequences, Formatter Instructions
@subsection Delimiters
@cindex delimiting escape sequence arguments
@cindex escape sequence argument delimiters
@cindex delimiters, for escape sequence arguments
@cindex arguments, for escape sequences, delimiting

@cindex @code{'}, as delimiter
@cindex @code{"}, as delimiter
Some escape sequences that require parameters use delimiters.  The
neutral apostrophe @code{'} is a popular choice and shown in this
document.  The neutral double quote @code{"} is also commonly seen.
Letters, numerals, and leaders can be used.  Punctuation characters
are likely better choices, except for those defined as infix operators
in numeric expressions; see below.

@Example
\l'1.5i\[bu]' \" draw 1.5 inches of bullet glyphs
@endExample

@cindex @code{\%}, as delimiter
@cindex @code{\@key{SP}}, as delimiter
@cindex @code{\|}, as delimiter
@cindex @code{\^}, as delimiter
@cindex @code{\@{}, as delimiter
@cindex @code{\@}}, as delimiter
@cindex @code{\'}, as delimiter
@cindex @code{\`}, as delimiter
@cindex @code{\-}, as delimiter
@cindex @code{\_}, as delimiter
@cindex @code{\!}, as delimiter
@cindex @code{\?}, as delimiter
@cindex @code{\)}, as delimiter
@cindex @code{\/}, as delimiter
@cindex @code{\,}, as delimiter
@cindex @code{\&}, as delimiter
@cindex @code{\:}, as delimiter
@cindex @code{\~}, as delimiter
@cindex @code{\0}, as delimiter
@cindex @code{\a}, as delimiter
@cindex @code{\c}, as delimiter
@cindex @code{\d}, as delimiter
@cindex @code{\e}, as delimiter
@cindex @code{\E}, as delimiter
@cindex @code{\p}, as delimiter
@cindex @code{\r}, as delimiter
@cindex @code{\t}, as delimiter
@cindex @code{\u}, as delimiter
The following escape sequences don't take arguments and thus are allowed
as delimiters:
@code{\@key{SP}}, @code{\%}, @code{\|}, @code{\^}, @code{\@{},
@code{\@}}, @code{\'}, @code{\`}, @code{\-}, @code{\_}, @code{\!},
@code{\?}, @code{\)}, @code{\/}, @code{\,}, @code{\&}, @code{\:},
@code{\~}, @code{\0}, @code{\a}, @code{\c}, @code{\d}, @code{\e},
@code{\E}, @code{\p}, @code{\r}, @code{\t}, and @code{\u}.  However,
using them this way is discouraged; they can make the input confusing to
read.

@cindex @code{\A}, delimiters allowed by
@cindex @code{\b}, delimiters allowed by
@cindex @code{\o}, delimiters allowed by
@cindex @code{\w}, delimiters allowed by
@cindex @code{\X}, delimiters allowed by
@cindex @code{\Z}, delimiters allowed by
@cindex newline, as delimiter
A few escape sequences,
@code{\A},
@code{\b},
@code{\o},
@code{\w},
@code{\X},
and @code{\Z}, accept a newline as a delimiter.  Newlines that serve
as delimiters continue to be recognized as input line terminators.

@Example
A caf\o
e\(aa
in Paris
    @result{} A café in Paris
@endExample

@noindent
Use of newlines as delimiters in escape sequences is also discouraged.

@cindex @code{\D}, delimiters allowed by
@cindex @code{\h}, delimiters allowed by
@cindex @code{\H}, delimiters allowed by
@cindex @code{\l}, delimiters allowed by
@cindex @code{\L}, delimiters allowed by
@cindex @code{\N}, delimiters allowed by
@cindex @code{\R}, delimiters allowed by
@cindex @code{\s}, delimiters allowed by
@cindex @code{\S}, delimiters allowed by
@cindex @code{\v}, delimiters allowed by
@cindex @code{\x}, delimiters allowed by
Finally, the escape sequences @code{\D}, @code{\h}, @code{\H},
@code{\l}, @code{\L}, @code{\N}, @code{\R}, @code{\s}, @code{\S},
@code{\v}, and @code{\x} prohibit many delimiters.

@itemize @bullet
@item
@cindex numerals, as delimiters
@cindex digits, as delimiters
@cindex @code{.}, as delimiter
@cindex decimal point, as delimiter
@cindex dot, as delimiter
the numerals @code{0}-@code{9} and the decimal point @code{.}

@item
@cindex operators, as delimiters
@cindex @code{+}, as delimiter
@cindex @code{-}, as delimiter
@cindex @code{/}, as delimiter
@cindex @code{*}, as delimiter
@cindex @code{%}, as delimiter
@cindex @code{<}, as delimiter
@cindex @code{>}, as delimiter
@cindex @code{=}, as delimiter
@cindex @code{&}, as delimiter
@ifnotinfo
@cindex @code{:}, as delimiter
@end ifnotinfo
@ifinfo
@cindex <colon>, as delimiter
@end ifinfo
@cindex @code{(}, as delimiter
@cindex @code{)}, as delimiter
the (single-character) operators @samp{+-/*%<>=&:()}

@item
@cindex space character, as delimiter
@cindex tab character, as delimiter
the space and tab characters

@item
@cindex @code{\%}, as delimiter
@cindex @code{\:}, as delimiter
@cindex @code{\@{}, as delimiter
@cindex @code{\@}}, as delimiter
@cindex @code{\'}, as delimiter
@cindex @code{\`}, as delimiter
@cindex @code{\-}, as delimiter
@cindex @code{\_}, as delimiter
@cindex @code{\!}, as delimiter
@cindex @code{\/}, as delimiter
@cindex @code{\c}, as delimiter
@cindex @code{\e}, as delimiter
@cindex @code{\p}, as delimiter
any escape sequences other than @code{\%}, @code{\:}, @code{\@{},
@code{\@}}, @code{\'}, @code{\`}, @code{\-}, @code{\_}, @code{\!},
@code{\/}, @code{\c}, @code{\e}, and @code{\p}
@end itemize

Delimiter syntax is complex and flexible primarily for historical
reasons; the foregoing restrictions need be kept in mind mainly when
using @code{groff} in @acronym{AT&T} compatibility mode.  GNU
@code{troff} keeps track of the nesting depth of escape sequence
interpolations, so the only characters you need to avoid using as
delimiters are those that appear in the arguments you input, not any
that result from interpolation. Typically, @code{'} works fine.
@xref{Implementation Differences}.

@Example
$ groff -Tps
.de Mw
.  nr wd \w'\\$1'
.  tm "\\$1" is \\n(wd units wide.
..
.Mw Wet'suwet'en
.Mw Wet+200i
.cp 1 \" turn on compatibility mode
.Mw Wet'suwet'en
.Mw Wet'
.Mw Wet+200i
    @error{} "Wet'suwet'en" is 54740 units wide.
    @error{} "Wet'+200i" is 42610 units wide.
    @error{} "Wet'suwet'en" is 15860 units wide.
    @error{} "Wet'" is 15860 units wide.
    @error{} "Wet'+200i" is 14415860 units wide.
@endExample

We see here that in compatibility mode, the part of the argument after
the @code{'} delimiter escapes from its context and, if nefariously
crafted, influences the computation of the @var{wd} register's value in
a surprising way.
@c END Keep (roughly) parallel with subsection " Delimiters" of
@c groff(7).

@node Comments, Registers, Formatter Instructions, GNU troff Reference
@section Comments
@cindex comments

One of the most common forms of escape sequence is the
comment.@footnote{This claim may be more aspirational than descriptive.}

@Defesc {\\", , , }
Start a comment.  Everything up to the next newline is ignored.

This may sound simple, but it can be tricky to keep the comments from
interfering with the appearance of the output.
@cindex @code{ds}, @code{ds1} requests, and comments
@cindex @code{as}, @code{as1} requests, and comments
If the escape sequence is to the right of some text or a request, that
portion of the line is ignored, but spaces preceding it are processed
normally by GNU @code{troff}.  This affects only the @code{ds} and
@code{as} requests and their variants.

@cindex tabs, before comments
@cindex comments, lining up with tabs
One possibly irritating idiosyncrasy is that tabs should not be used to
vertically align comments in the source document.  Tab characters are
not treated as separators between a request name and its first argument,
nor between arguments.

@cindex undefined request
@cindex request, undefined
A comment on a line by itself is treated as a blank line, because after
eliminating the comment, that is all that remains.

@Example
Test
\" comment
Test
    @result{} Test
    @result{}
    @result{} Test
@endExample

To avoid this, it is common to combine the empty request with the
comment escape sequence as @samp{.\"}, causing the input line to be
ignored.

@cindex @code{'}, as a comment
Another commenting scheme sometimes seen is three consecutive single
quotes (@code{'''}) at the beginning of a line.  This works, but GNU
@code{troff} emits a warning diagnostic (if enabled) about an undefined
macro (namely @samp{''}).
@endDefesc

@Defesc {\\#, , , }
Start a comment; everything up to and including the next newline is
ignored.  This @code{groff} extension was introduced to avoid the
problems described above.

@Example
Test
\# comment
Test
    @result{} Test Test
@endExample
@endDefesc

@Defreq {ig, [@Var{end}]}
Ignore input until, in the current conditional block (if
any),@footnote{@xref{Conditional Blocks}.} the macro @var{end} is called
at the start of a control line, or the control line @samp{..} is
encountered if @var{end} is not specified.  @code{ig} is parsed as if it
were a macro definition, but its contents are discarded, not
stored.@footnote{Exception: auto-incrementing registers defined outside
the ignored region @emph{will} be modified if interpolated with
@code{\n±} inside it.  @xref{Auto-increment}.}

@c Wrap example at 56 columns.
@Example
hand\c
.de TX
fasting
..
.ig TX
This is part of a large block of input that has been
temporarily(?) commented out.
We can restore it simply by removing the .ig request and
the call of its end macro.
.TX
@endExample
@Example
    @result{} handfasting
@endExample
@endDefreq


@c =====================================================================

@c BEGIN Keep (roughly) parallel with subsection "Registers" of
@c groff(7).
@node Registers, Manipulating Filling and Adjustment, Formatter Instructions, GNU troff Reference
@section Registers
@cindex registers

In the @code{roff} language, numbers can be stored in @dfn{registers}.
Many built-in registers exist, supplying anything from the date to
details of formatting parameters.  You can also define your own.
@xref{Identifiers}, for information on constructing a valid name for a
register.

@menu
* Setting Registers::
* Interpolating Registers::
* Auto-increment::
* Assigning Register Formats::
* Built-in Registers::
@end menu

@c ---------------------------------------------------------------------

@node Setting Registers, Interpolating Registers, Registers, Registers
@subsection Setting Registers
@cindex setting registers (@code{nr}, @code{\R})
@cindex registers, setting (@code{nr}, @code{\R})

Define registers and update their values with the @code{nr} request or
the @code{\R} escape sequence.

@DefreqList {nr, ident value}
@DefescListEndx {\\R, @code{'}, ident value, @code{'}}
Set register @var{ident} to @var{value}.  If @var{ident} doesn't exist,
GNU @code{troff} creates it.  In the @code{\R} escape sequence, the
delimiter need not be a neutral apostrophe; see @ref{Delimiters}.  It
also does not produce an input token in GNU @code{troff}.  @xref{Gtroff
Internals}.

@Example
.nr a (((17 + (3 * 4))) % 4)
\n[a]
.\R'a (((17 + (3 * 4))) % 4)'
\n[a]
    @result{} 1 1
@endExample

(Later, we will discuss additional forms of @code{nr} and @code{\R} that
can change a register's value after it is dereferenced but before it is
interpolated.  @xref{Auto-increment}.)

The complete transparency of @code{\R} can cause surprising effects if
you use registers like @code{.k}, which get evaluated at the time they
are accessed.

@Example
.ll 1.6i
.
aaa bbb ccc ddd eee fff ggg hhh\R':k \n[.k]'
.tm :k == \n[:k]
    @result{} :k == 126950
.
.br
.
aaa bbb ccc ddd eee fff ggg hhh\h'0'\R':k \n[.k]'
.tm :k == \n[:k]
    @result{} :k == 15000
@endExample

If you process this with the PostScript device (@code{-Tps}), there will
be a line break eventually after @code{ggg} in both input lines.
However, after processing the space after @code{ggg}, the partially
collected line is not overfull yet, so GNU @code{troff} continues to
collect input until it sees the space (or in this case, the newline)
after @code{hhh}.  At this point, the line is longer than the line
length, and the line gets broken.

In the first input line, since the @code{\R} escape sequence leaves no
traces, the check for the overfull line hasn't been done yet at the
point where @code{\R} gets handled, and you get a value for the
@code{.k} register that is even greater than the current line length.

In the second input line, the insertion of @code{\h'0'} to emit an
invisible zero-width space forces GNU @code{troff} to check the line
length, which in turn causes the start of a new output line.  Now
@code{.k} returns the expected value.
@endDefreq

@code{nr} and @code{\R} each have two additional special forms to
increment or decrement a register.

@DefreqList {nr, ident @t{+}@Var{value}}
@DefreqItem {nr, ident @t{-}@Var{value}}
@DefescItemx {\\R, @code{'}, ident @t{+}value, @code{'}}
@DefescListEnd {\\R, @code{'}, ident @t{-}value, @code{'}}
Increment (decrement) register @var{ident} by @var{value}.  In the
@code{\R} escape sequence, the delimiter need not be a neutral
apostrophe; see @ref{Delimiters}.

@Example
.nr a 1
.nr a +1
\na
    @result{} 2
@endExample

@cindex negating register values
A leading minus sign in @var{value} is always interpreted as a
decrementation operator, not an algebraic sign.  To assign a register a
negative value or the negated value of another register, you can
force GNU @code{troff} to interpret @samp{-} as a negation or minus,
rather than decrementation, operator: enclose it with its operand in
parentheses or subtract it from zero.

@Example
.nr a 7
.nr b 3
.nr a -\nb
\na
    @result{} 4
.nr a (-\nb)
\na
    @result{} -3
.nr a 0-\nb
\na
    @result{} -3
@endExample

If a register's prior value does not exist (the register was undefined),
an increment or decrement is applied as if to@tie{}0.
@endDefreq

@Defreq {rr, ident}
@cindex removing a register (@code{rr})
@cindex register, removing (@code{rr})
Remove register @var{ident}.  If @var{ident} doesn't exist, the request
is ignored.  Technically, only the name is removed; the register's
contents are still accessible under aliases created with @code{aln}, if
any.
@endDefreq

@Defreq {rnn, ident1 ident2}
@cindex renaming a register (@code{rnn})
@cindex register, renaming (@code{rnn})
Rename register @var{ident1} to @var{ident2}.  If @var{ident1} doesn't
exist, the request is ignored.  Renaming a built-in register does not
otherwise alter its properties.
@endDefreq

@Defreq {aln, new old}
@cindex alias, register, creation (@code{aln})
@cindex creating alias for register (@code{aln})
@cindex register, creating alias for (@code{aln})
Create an alias @var{new} for an existing register @var{old}, causing
the names to refer to the same stored object.  If @var{old} is
undefined, a warning in category @samp{reg} is produced and the request
is ignored.  @xref{Warnings}, for information about the enablement and
suppression of warnings.

@cindex alias, register, removing (@code{aln})
@cindex removing alias for register (@code{aln})
@cindex register, removing alias for (@code{aln})
To remove a register alias, call @code{rr} on its name.  A register's
contents do not become inaccessible until it has no more names.
@endDefreq
@c END Keep (roughly) parallel with subsection "Registers" of groff(7).

@c ---------------------------------------------------------------------

@node Interpolating Registers, Auto-increment, Setting Registers, Registers
@subsection Interpolating Registers
@cindex interpolating registers (@code{\n})
@cindex registers, interpolating (@code{\n})

Register contents are interpolated with the @code{\n} escape sequence.

@DefescList {\\n, , i, }
@DefescItem {\\n, (, id, }
@DefescListEnd {\\n, [, ident, ]}
@cindex nested assignments
@cindex assignments, nested
@cindex indirect assignments
@cindex assignments, indirect
Interpolate register with name @var{ident} (one-character
name@tie{}@var{i}, two-character name @var{id}).  @code{\n} is
interpreted even in copy mode (@pxref{Copy Mode}).  If the register is
undefined, it is created, assigned a value of@tie{}@samp{0}, that value
is interpolated, and a warning in category @samp{reg} is emitted.
@xref{Warnings}, for information about the enablement and suppression
of warnings.

@Example
.nr a 5
.nr as \na+\na
\n(as
    @result{} 10
@endExample

@Example
.nr a1 5
.nr ab 6
.ds str b
.ds num 1
\n[a\n[num]]
    @result{} 5
\n[a\*[str]]
    @result{} 6
@endExample
@endDefesc

@c ---------------------------------------------------------------------

@node Auto-increment, Assigning Register Formats, Interpolating Registers, Registers
@subsection Auto-increment
@cindex auto-incrementation of a register
@cindex incrementation, automatic, of a register
@cindex decrementation, automatic, of a register

Registers can also be incremented or decremented by a configured amount
at the time they are interpolated.  The value of the increment is
specified with a third argument to the @code{nr} request, and a special
interpolation syntax is used to alter and then retrieve the register's
value.  Together, these features are called
@dfn{auto-increment}.@footnote{A negative auto-increment can be
considered an ``auto-decrement''.}

@Defreq {nr, ident value incr}
@cindex @code{\R}, difference to @code{nr}
Set register @var{ident} to @var{value} and its auto-incrementation
amount to to @var{incr}.  The @code{\R} escape sequence doesn't support
an @var{incr} argument.
@endDefreq

Auto-incrementation is not @emph{completely} automatic; the @code{\n}
escape sequence in its basic form never alters the value of a register.
To apply auto-incrementation to a register, interpolate it with
@samp{\n±}.

@DefescList {\\n, +, i, }
@DefescItem {\\n, -, i, }
@DefescItem {\\n, +(, id, }
@DefescItem {\\n, -(, id, }
@DefescItem {\\n, +[, ident, ]}
@DefescListEnd {\\n, -[, ident, ]}
Increment or decrement @var{ident} (one-character
name@tie{}@var{i}, two-character name @var{id}) by the register's
auto-incrementation value and then interpolate the new register value.
If @var{ident} has no auto-incrementation value, interpolate as with
@code{\n}.
@endDefesc

@need 1000
@Example
.nr a 0 1
.nr xx 0 5
.nr foo 0 -2
\n+a, \n+a, \n+a, \n+a, \n+a
.br
\n-(xx, \n-(xx, \n-(xx, \n-(xx, \n-(xx
.br
\n+[foo], \n+[foo], \n+[foo], \n+[foo], \n+[foo]
    @result{} 1, 2, 3, 4, 5
    @result{} -5, -10, -15, -20, -25
    @result{} -2, -4, -6, -8, -10
@endExample

@cindex increment value without changing the register
@cindex value, incrementing without changing the register
To change the increment value without changing the value of a register,
assign the register's value to itself by interpolating it, and specify
the desired increment normally.  Apply an increment of @samp{0} to
disable auto-incrementation of the register.

@c ---------------------------------------------------------------------

@node Assigning Register Formats, Built-in Registers, Auto-increment, Registers
@subsection Assigning Register Formats
@cindex assigning formats (@code{af})
@cindex formats, assigning (@code{af})

A writable register's value can be interpolated in several number
formats.  By default, conventional Arabic numerals are used.
Other formats see use in sectioning and outlining schemes and
alternative page numbering arrangements.

@Defreq {af, reg fmt}
Use number format @var{fmt} when interpolating register @var{reg}.
Valid number formats are as follows.

@table @code
@item 0@r{@dots{}}
Arabic numerals 0, 1, 2, and so on.
Any decimal digit is equivalent to @samp{0}; the formatter merely counts
the digits specified.  Multiple Arabic numerals in @var{fmt} cause
interpolations to be zero-padded on the left if necessary to at least as
many digits as specified (interpolations never truncate a register
value).  A register with format @samp{00} interpolates values 1, 2, 3 as
@samp{01}, @samp{02}, @samp{03}.  The default format for all writable
registers is @samp{0}.

@item I
@cindex Roman numerals
@cindex numerals, Roman
Uppercase Roman numerals: 0, I, II, III, IV,@tie{}@enddots{}

@item i
Lowercase Roman numerals: 0, i, ii, iii, iv,@tie{}@enddots{}

@item A
Uppercase letters: 0, A, B, C, @dots{},@tie{}Z, AA, AB,@tie{}@enddots{}

@item a
Lowercase letters: 0, a, b, c, @dots{},@tie{}z, aa, ab,@tie{}@enddots{}
@end table

Omitting @var{fmt} causes a warning in category @samp{missing}.
@xref{Warnings}, for information about the enablement and suppression of
warnings.  Specifying an unrecognized format is an error.

Zero values are interpolated as @samp{0} in non-Arabic formats.
Negative quantities are prefixed with @samp{-} irrespective of format.
In Arabic formats, the sign supplements the field width.  If @var{reg}
doesn't exist, it is created with a zero value.

@Example
.nr a 10
.af a 0           \" the default format
\na,
.af a I
\na,
.af a 321
.nr a (-\na)
\na,
.af a a
\na
    @result{} 10, X, -010, -j
@endExample

@cindex Roman numerals, extrema (maximum and minimum)
@cindex extreme values representable with Roman numerals
@cindex maximum value representable with Roman numerals
@cindex minimum value representable with Roman numerals
The representable extrema in the @samp{i} and @samp{I} formats
correspond to Arabic ±39,999.  GNU @code{troff} uses @samp{w} and
@samp{z} to represent 5,000 and 10,000 in Roman numerals, respectively,
following the convention of @acronym{AT&T} @code{troff}---currently, the
correct glyphs for Roman numerals five thousand (@code{U+2181}) and ten
thousand (@code{U+2182}) are not used.

@cindex read-only register, changing format
@cindex changing format, and read-only registers
Assigning the format of a read-only register is an error.  Instead, copy
the read-only register's value to, and assign the format of, a writable
register.
@endDefreq

@DefescList {\\g, , r, }
@DefescItem {\\g, (, rg, }
@DefescListEnd {\\g, [, reg, ]}
@cindex format of register (@code{\g})
@cindex register, format (@code{\g})
Interpolate the format of the register @var{reg} (one-character
name@tie{}@var{r}, two-character name @var{rg}).  Zeroes represent
Arabic formats.  If @var{reg} is not defined, @var{reg} is not created
and nothing is interpolated.  @code{\g} is interpreted even in copy mode
(@pxref{Copy Mode}).
@endDefesc

@cindex register format, in expressions
@cindex expressions, and register format
GNU @code{troff} interprets only Arabic numerals.  The Roman numeral or
alphabetic formats cannot be used as operands to arithmetic operators in
expressions (@pxref{Numeric Expressions}).  For instance, it may be
desirable to test the page number independently of its format.

@Example
.af % i \" front matter
.de header-trap
.  \" To test the page number, we need it in Arabic.
.  ds saved-page-number-format \\g%\"
.  af % 0
.  nr page-number-in-decimal \\n%
.  af % \\*[saved-page-number-format]
.  ie \\n[page-number-in-decimal]=1 .do-first-page-stuff
.  el \@{\
.    ie o .do-odd-numbered-page-stuff
.    el   .do-even-numbered-page-stuff
.  \@}
.  rm saved-page-number-format
..
.wh 0 header-trap
@endExample

@c ---------------------------------------------------------------------

@node Built-in Registers,  , Assigning Register Formats, Registers
@subsection Built-in Registers
@cindex built-in registers
@cindex registers, built-in

Predefined registers whose identifiers start with a dot are read-only.
Many are Boolean-valued, interpolating a true or false value testable
with the @code{if}, @code{ie}, or @code{while} requests.  Some read-only
registers are string-valued, meaning that they interpolate text.

@cindex removing a built-in register
@cindex register, built-in, removing
@cindex built-in register, removing
@strong{Caution:@:} Built-in registers are subject to removal like
others; once removed, they can be recreated only as normal writable
registers and will not reflect formatter state.

A register name (without the dot) is often associated with a request of
the same name.  A complete listing of all built-in registers can be
found in @ref{Register Index}.

We present here a few built-in registers that are not described
elsewhere in this manual; they have to do with invariant properties of
GNU @code{troff}, or obtain information about the formatter's
command-line options or the operating environment.

@table @code
@item \n[.A]
@vindex .A
@cindex approximation output register (@code{.A})
@cindex plain text approximation output register (@code{.A})
Approximate output is being formatted (Boolean-valued); see
@command{groff} @option{-a} option (@ref{Groff Options}).

@item \n[.c]
@vindex .c
@itemx \n[c.]
@vindex c.
@cindex input line number register (@code{.c}, @code{c.})
@cindex line number, input, register (@code{.c}, @code{c.})
Input line number.  @samp{c.} is a writable synonym,
@c introduced in AT&T device-independent troff (CSTR #54, 1981-01)
affecting subsequent interpolations of both @samp{.c} and @samp{c.}.

@item \n[.F]
@cindex current input file name register (@code{.F})
@cindex input file name, current, register (@code{.F})
@vindex .F
Name of input file (string-valued).

@item \n[.g]
@vindex .g
@cindex GNU @code{troff}, identification register (@code{.g})
@cindex GNU-specific register (@code{.g})
Always true in GNU @code{troff} (Boolean-valued).  Documents can use
this to ask the formatter if it claims @code{groff} compatibility.

@item \n[.P]
@vindex .P
Output page selection status (Boolean-valued); see @command{groff}
@option{-o} option (@ref{Groff Options}).

@item \n[.R]
@cindex number of registers register (@code{.R})
@cindex registers, number of, register (@code{.R})
@vindex .R
Count of available unused registers; always 10,000 in GNU
@code{troff}.@footnote{GNU @code{troff} dynamically allocates memory for
as many registers as required.}

@item \n[.T]
@vindex .T
Indicator of output device selection (Boolean-valued); see
@command{groff} @option{-T} option (@ref{Groff Options}).

@item \n[.U]
@cindex safer mode
@cindex mode, safer
@cindex unsafe mode
@cindex mode, unsafe
@vindex .U
Unsafe mode enablement status (Boolean-valued); see @command{groff}
@option{-U} option (@ref{Groff Options}).

@item \n[.x]
@vindex .x
@cindex major version number register (@code{.x})
@cindex version number, major, register (@code{.x})
Major version number of the running GNU @code{troff} formatter.  For
example, if the version number is 1.23.0, then @code{.x}
contains@tie{}@samp{1}.

@item \n[.y]
@vindex .y
@cindex minor version number register (@code{.y})
@cindex version number, minor, register (@code{.y})
Minor version number of the running GNU @code{troff} formatter.  For
example, if the version number is 1.23.0, then @code{.y}
contains@tie{}@samp{23}.

@item \n[.Y]
@vindex .Y
@cindex revision number register (@code{.Y})
Revision number of the running GNU @code{troff} formatter.  For example,
if the version number is 1.23.0, then @code{.Y} contains@tie{}@samp{0}.

@item \n[$$]
@vindex $$
@cindex process ID of GNU @code{troff} register (@code{$$})
@cindex PID of GNU @code{troff} register (@code{$$})
@cindex GNU @code{troff}, process ID register (@code{$$})
@cindex GNU @code{troff}, PID register (@code{$$})
Process identifier (PID) of the GNU @code{troff} program in its
operating environment.
@end table

Date- and time-related registers are set per the local time as
determined by @cite{localtime@r{(3)}} when the formatter launches.  This
initialization can be overridden by @env{SOURCE_DATE_EPOCH} and
@env{TZ}; see @ref{Environment}.

@table @code
@item \n[seconds]
@cindex seconds, current time (@code{seconds})
@cindex time, current, seconds (@code{seconds})
@cindex current time, seconds (@code{seconds})
@vindex seconds
Count of seconds elapsed in the minute (0--60). @c not 59; see POSIX

@item \n[minutes]
@cindex minutes, current time (@code{minutes})
@cindex time, current, minutes (@code{minutes})
@cindex current time, minutes (@code{minutes})
@vindex minutes
Count of minutes elapsed in the hour (0--59).

@item \n[hours]
@cindex hours, current time (@code{hours})
@cindex time, current, hours (@code{hours})
@cindex current time, hours (@code{hours})
@vindex hours
Count of hours elapsed since midnight (0--23).

@item \n[dw]
@cindex day of the week register (@code{dw})
@cindex date, day of the week register (@code{dw})
@vindex dw
Day of the week (1--7; 1 is Sunday).

@item \n[dy]
@cindex day of the month register (@code{dy})
@cindex date, day of the month register (@code{dy})
@vindex dy
Day of the month (1--31).

@item \n[mo]
@cindex month of the year register (@code{mo})
@cindex date, month of the year register (@code{mo})
@vindex mo
Month of the year (1--12).

@item \n[year]
@cindex date, year register (@code{year}, @code{yr})
@cindex year, current, register (@code{year}, @code{yr})
@vindex year
Gregorian year.

@cindex CSTR@tie{}#54 errata
@cindex CSTR@tie{}#54 erratum, @code{yr} register
@item \n[yr]
@vindex yr
Gregorian year minus@tie{}1900.  This register is incorrectly documented
in the @acronym{AT&T} @code{troff} manual as storing the last two digits
of the current year.  That claim stopped being true in 2000.  Old
@code{troff} input that looks like:

@Example
'\" The year number is a surprise after 1999.
This document was formatted in 19\n(yr.
@endExample

@noindent
can be corrected to:

@Example
This document was formatted in \n[year].
@endExample

@noindent
or, for portability across many @code{roff} programs, to the following.

@Example
.nr y4 1900+\n(yr
This document was formatted in \n(y4.
@endExample
@end table


@c =====================================================================

@node Manipulating Filling and Adjustment, Manipulating Hyphenation, Registers, GNU troff Reference
@section Manipulating Filling and Adjustment
@cindex manipulating filling and adjustment
@cindex filling and adjustment, manipulating
@cindex adjustment and filling, manipulating
@cindex justifying text
@cindex text, justifying

@cindex break
@cindex line break
@cindex @code{bp} request, causing implicit break
@cindex @code{ce} request, causing implicit break
@cindex @code{cf} request, causing implicit break
@cindex @code{fi} request, causing implicit break
@cindex @code{fl} request, causing implicit break
@cindex @code{in} request, causing implicit break
@cindex @code{nf} request, causing implicit break
@cindex @code{rj} request, causing implicit break
@cindex @code{sp} request, causing implicit break
@cindex @code{ti} request, causing implicit break
@cindex @code{trf} request, causing implicit break
When an output line is pending (see below), a break moves the drawing
position to the beginning of the next text baseline, interrupting
filling.  Various ways of causing breaks were shown in @ref{Breaking}.
The @code{br} request likewise causes a break.  Several other requests
imply breaks:@: @code{bp}, @code{ce}, @code{cf}, @code{fi}, @code{fl},
@code{in}, @code{nf}, @code{rj}, @code{sp}, @code{ti}, and @code{trf}.
If the no-break control character is used with any of these requests,
GNU @code{troff} suppresses the break; instead the requested operation
takes effect at the next break.  @samp{'br} does nothing.

@Example
.ll 55n
This line is normally filled and adjusted.
.br
A line's alignment is decided
'ce \" Center the next input line (no break).
when it is output.
This line returns to normal filling and adjustment.
    @result{} This line is normally filled and adjusted.
    @result{}    A line's alignment is decided when it is output.
    @result{} This line returns to normal filling and adjustment.
@endExample

@noindent
@cindex pending output line
@cindex partially collected line
@cindex output line properties
@cindex properties of output lines
Output line properties like page offset, indentation, adjustment, and
even the location of its text baseline, are not determined until the
line has been broken.  An output line is said to be @dfn{pending} if
some input has been collected but an output line corresponding to it has
not yet been written; such an output line is also termed @dfn{partially
collected}.  If no output line is pending, it is as if a break has
already happened; additional breaks, whether explicit or implicit, have
no effect.  If the vertical drawing position is negative---as it is when
the formatter starts up---a break starts a new page (even if no output
line is pending) unless an end-of-input macro is being interpreted.
@xref{End-of-input Traps}.

@Defreq {br, }
Break the line: emit any pending output line without adjustment.

@Example
foo bar
.br
baz
'br
qux
    @result{} foo bar
    @result{} baz qux
@endExample
@endDefreq

Sometimes you want to prevent a break within a phrase or between a
quantity and its units.

@Defesc {\\~, , , }
@cindex unbreakable space (@code{\~})
@cindex space, unbreakable (@code{\~})
Insert an unbreakable space that is adjustable like an ordinary space.
It is discarded from the end of an output line if a break is forced.

@Example
Set the output speed to\~1.
There are 1,024\~bytes in 1\~KiB.
J.\~F.\~Ossanna wrote the original CSTR\~#54.
@endExample
@endDefesc

By default, GNU @code{troff} fills text and adjusts it to reach the
output line length.  The @code{nf} request disables filling; the
@code{fi} request reënables it.

@DefreqList {fi, }
@DefregListEndx {.u}
@cindex filling of output, enabling (@code{fi})
@cindex output, filling, enablement of (@code{fi})
@cindex fill mode (@code{fi}), enabling
@cindex mode, fill (@code{fi}), enabling
Enable filling of output lines; a pending output line is broken.  The
read-only register @code{.u} is set to@tie{}1.  The filling enablement
status, sometimes called @dfn{fill mode}, is associated with the
environment (@pxref{Environments}).  @xref{Line Continuation}, for
interaction with the @code{\c} escape sequence.
@endDefreq

@Defreq {nf, }
@cindex filling of output, disabling (@code{nf})
@cindex output, filling, disablement of (@code{nf})
@cindex no-fill mode
@cindex mode, no-fill
@cindex fill mode, disabling
@cindex mode, fill, disabling
Disable filling of output lines: the output line length (@pxref{Line
Layout}) is ignored and output lines are broken where the input lines
are.  A pending output line is broken and adjustment is suppressed.  The
read-only register @code{.u} is set to@tie{}0.  The filling enablement
status is associated with the environment (@pxref{Environments}).  See
@ref{Line Continuation}, for interaction with the @code{\c} escape
sequence.
@endDefreq

@DefreqList {ad, [@Var{mode}]}
@DefregListEndx {.j}
Enable output line adjustment in @var{mode}, taking effect when the
pending (or next) output line is broken.  Adjustment is suppressed when
filling is.  @var{mode} can have one of the following values.

@table @code
@item b
@itemx n
Adjust ``normally'':@: if the output line does not consume the distance
between the indentation and the configured output line length, GNU
@code{troff} stretches adjustable spaces within the line until that
length is reached.  When the indentation is zero, this mode spreads the
line to both the left and right margins.  This is the GNU @code{troff}
default.

@item c
@cindex centered text (filled)
Center filled text.  Contrast with the @code{ce} request, which centers
text @emph{without} filling it.

@item l
@cindex ragged-right text
Align text to the left without adjusting it.

@item r
@cindex ragged-left text
Align text to the right without adjusting it.
@end table

@var{mode} can also be a value previously stored in the @code{.j}
register.  Using @code{ad} without an argument is the same as @samp{.ad
\n[.j]}; unless filling is disabled, GNU @code{troff} resumes adjusting
lines in the same way it did before adjustment was disabled by
invocation of the @code{na} request.

@Example
.ll 48n
.de AD
.  br
.  ad \\$1
..
@c . @c XXX: Restore this line when the page has room for it.
.de NA
.  br
.  na
..
@c . @c XXX: Restore this line when the page has room for it.
left
.AD r
.nr ad \n(.j
right
.AD c
center
.NA
left
.AD
center
.AD \n(ad
right
@endExample
@Example
    @result{} left
    @result{}                                            right
    @result{}                      center
    @result{} left
    @result{}                      center
    @result{}                                            right
@endExample

@cindex adjustment mode register (@code{.j})
The adjustment mode and enablement status are encoded in the read-only
register @code{.j}.  These parameters are associated with the
environment (@pxref{Environments}).

The value of @code{.j} for any adjustment mode is an implementation
detail and should not be relied upon as a programmer's interface.  Do
not write logic to interpret or perform arithmetic on it.
@endDefreq

@Defreq {na, }
Disable output line adjustment.  This produces the same output as
left-alignment, but the value of the adjustment mode register @code{.j}
is altered differently.  The adjustment mode and enablement status are
associated with the environment (@pxref{Environments}).
@endDefreq

@DefreqList {brp, }
@DefescListEndx {\\p, , , }
Break, adjusting the line per the current adjustment mode.  @code{\p}
schedules a break with adjustment at the next word boundary.  The escape
sequence is itself neither a break nor a space of any kind; it can thus
be placed in the middle of a word to cause a break at the end of that
word.

Breaking with immediate adjustment can produce ugly results since GNU
@code{troff} doesn't have a sophisticated paragraph-building algorithm,
as @TeX{} has, for example.  Instead, GNU @code{troff} fills and adjusts
a paragraph line by line.

@Example
.ll 4.5i
This is an uninteresting sentence.
This is an uninteresting sentence.\p
This is an uninteresting sentence.
@endExample

@noindent
is formatted as follows.

@Example
This  is  an uninteresting sentence.  This is
an          uninteresting           sentence.
This is an uninteresting sentence.
@endExample
@endDefreq

@cindex productive input line
@cindex input line, productive
@cindex line, productive input
To clearly present the next couple of requests, we must introduce the
concept of ``productive'' input lines.  A @dfn{productive input line} is
one that directly produces formatted output.  Text lines produce
output,@footnote{unless diverted; see @ref{Diversions}} as do control
lines containing requests like @code{tl} or escape sequences like
@code{\D}.  Macro calls are not @emph{directly} productive, and thus not
counted, but their interpolated contents can be.  Empty requests, and
requests and escape sequences that define registers or strings or alter
the formatting environment (as with changes to the size, face, height,
slant, or color of the type) are not productive.  We will also preview
the output line continuation escape sequence, @code{\c}, which
``connects'' two input lines that would otherwise be counted separately.
@footnote{@xref{Line Continuation}.}

@Example
@c .ll 56n
.de hello
Hello, world!
..
.ce \" center output of next productive input line
.
.nr junk-reg 1
.ft I
Chorus: \c
.ft
.hello
Went the day well?
  @result{}                  @slanted{Chorus:} Hello, world!
  @result{} Went the day well?
@endExample

@DefreqList {ce, [@Var{nnn}]}
@DefregListEndx {.ce}
@cindex centered text (unfilled)
@cindex centering lines (@code{ce})
@cindex lines, centering (@code{ce})
Break (unless the no-break control character is used), center the output
of the next @var{nnn} productive input lines with respect to the line
length and indentation, then break again regarding of the invoking
control character.  Temporary indentation is ignored.  The count of
lines remaining to be centered is stored in the read-only register
@code{.ce} and is associated with the environment
(@pxref{Environments}).

@cindex @code{ce} request, difference from @w{@samp{.ad c}}
While the @w{@samp{.ad c}} request also centers text, it fills the text
as well.

@c Wrap example at 56 columns.
@Example
.de FR
This is a small text fragment that shows the differences
between the `.ce' and the `.ad c' requests.
..
.ll 4i
.ce 1000
.FR
.ce 0

.ad c
.FR
    @result{} This is a small text fragment that shows
    @result{}              the differences
    @result{} between the @quoteleft{}.ce@quoteright{} and the @quoteleft{}.ad c@quoteright{} requests.
    @result{}
    @result{} This is a small text fragment that shows
    @result{}  the differences between the @quoteleft{}.ce@quoteright{} and
    @result{}         the @quoteleft{}.ad c@quoteright{} requests.
@endExample

The previous example illustrates a common idiom of turning centering on
for a quantity of lines far in excess of what is required, and off again
after the text to be centered.  This technique relieves humans of
counting lines for requests that take a count of input lines as an
argument.
@endDefreq

@DefreqList {rj, [@Var{nnn}]}
@DefregListEndx {.rj}
@cindex justifying text (@code{rj})
@cindex text, justifying (@code{rj})
@cindex right-justifying (@code{rj})
Break (unless the no-break control character is used), align the output
of the next @var{nnn} productive input lines to the right margin without
filling, then break again (regardless of the control character).  The
count of lines remaining to be right-aligned is stored in the read-only
register @code{.rj} and is associated with the environment
(@pxref{Environments}).  @var{nnn} specifies the number of lines to
right-align.  If the argument is not positive, right-alignment is
disabled.  Omitting the argument implies an @var{nnn} of @samp{1}.
@endDefreq

@need 2000
@DefreqList {ss, word-space-size [@Var{additional-sentence-space-size}]}
@DefregItemx {.ss}
@DefregListEndx {.sss}
@cindex word space size register (@code{.ss})
@cindex size of word space register (@code{.ss})
@cindex space between words register (@code{.ss})
@cindex inter-sentence space size register (@code{.sss})
@cindex sentence space size register (@code{.sss})
@cindex size of sentence space register (@code{.sss})
@cindex space between sentences register (@code{.sss})
Set the sizes of spaces between words and
sentences@footnote{Recall @ref{Filling} and @ref{Sentences} for the
definitions of word and sentence boundaries, respectively.} in twelfths
of current font's space width (usually one-third em for Western
scripts).  Initially, both the @var{word-space-size} and
@var{additional-sentence-space-size} are@tie{}12.  Negative values are
erroneous.
@cindex inter-word spacing, minimal
@cindex minimal inter-word spacing
@cindex space, between words
The first argument, the inter-word space size, is a minimum; if an
output line undergoes adjustment, such spaces may increase in width.
@cindex inter-sentence space, additional
@cindex additional inter-sentence space
@cindex space, between sentences
The optional second argument sets the amount of additional space
separating sentences on the same output line.  If omitted, this amount
is set to @var{word-space-size}.  The request is ignored if there are no
parameters.

@cindex filling, and inter-sentence space
@cindex mode, fill, and inter-sentence space
Additional inter-sentence space is used only if the output line is not
full when the end of a sentence occurs in the input.  If a sentence ends
at the end of an input line, then both an inter-word space and an
inter-sentence space are added to the output; if two spaces follow the
end of a sentence in the middle of an input line, then the second space
becomes an inter-sentence space in the output.  Additional
inter-sentence space is not adjusted, but the inter-word space that
always precedes it may be.  Further input spaces after the second, if
present, are adjusted as normal.

The read-only registers @code{.ss} and @code{.sss} hold the minimal
inter-word space and additional inter-sentence space amounts,
respectively.  These parameters are part of the environment
(@pxref{Environments}), and rounded down to the nearest multiple
of@tie{}12 on terminals.

@cindex discardable horizontal space
@cindex space, discardable, horizontal
@cindex horizontal discardable space
The @code{ss} request can insert discardable horizontal space; that is,
space that is discarded at a break.  For example, some footnote styles
collect the notes into a single paragraph with large gaps between
each note.

@Example
.ll 48n
1.\~J. Fict. Ch. Soc. 6 (2020), 3\[en]14.
.ss 12 48 \" applies to next sentence ending
Reprints no longer available through FCS.
.ss 12 \" go back to normal
2.\~Better known for other work.
    @result{} 1.  J.  Fict. Ch. Soc. 6 (2020), 3-14.  Reprints
    @result{} no longer available through FCS.      2.  Better
    @result{} known for other work.
@endExample

@noindent
If @emph{undiscardable} space is required, use the @code{\h} escape
sequence.
@endDefreq


@c =====================================================================

@node Manipulating Hyphenation, Manipulating Spacing, Manipulating Filling and Adjustment, GNU troff Reference
@section Manipulating Hyphenation
@cindex manipulating hyphenation
@cindex hyphenation, manipulating

@cindex hyphenation, automatic
@cindex automatic hyphenation
When filling, GNU @code{troff} hyphenates words as needed at
user-specified and automatically determined hyphenation points.  The
machine-driven determination of hyphenation points in words requires
algorithms and data, and is susceptible to conventions and preferences.
Before tackling such @dfn{automatic hyphenation}, let us consider how
hyphenation points can be set manually.

@cindex hyphenation, explicit
@cindex explicit hyphenation
@cindex hyphenation, manual
@cindex manual hyphenation
Explicitly hyphenated words such as ``mother-in-law'' are eligible for
breaking after each of their hyphens.  Relatively few words in a
language offer such obvious break points, however, and automatic
detection of syllabic (or phonetic) boundaries for hyphenation is not
perfect,@footnote{Whether a perfect algorithm for this application is
even possible is an unsolved problem in computer science:@:
@url{https://tug.org/docs/liang/liang-thesis.pdf}.} particularly for
unusual words found in technical literature.  We can instruct GNU
@code{troff} how to hyphenate specific words if the need arises.

@cindex hyphenation exceptions
@Defreq {hw, word @dots{}}
Define each @dfn{hyphenation exception} @var{word} with each hyphen `-'
in the word indicating a hyphenation point.  For example, the request

@Example
.hw in-sa-lub-rious alpha
@endExample

@c Serendipitously, in PDF output, the "alpha" below gets hyphenated.
@c Try to preserve this felicity in future edits.
marks potential hyphenation points in ``insalubrious'', and prevents
``alpha'' from being hyphenated at all.

Besides the space character, any character whose hyphenation code is
zero can be used to separate the arguments of @code{hw} (see the
@code{hcode} request below).  In addition, this request can be used more
than once.

@cindex @code{hw} request, and @code{hy} restrictions
Hyphenation points specified with @code{hw} are not subject to the
within-word placement restrictions imposed by the @code{hy} request (see
below).

Hyphenation exceptions specified with the @code{hw} request are
associated with the hyphenation language (see the @code{hla} request
below) and environment (@pxref{Environments}); invoking the @code{hw}
request in the absence of a hyphenation language is an error.

The request is ignored if there are no parameters.
@endDefreq

These are known as hyphenation @slanted{exceptions} in the expectation
that most users will avail themselves of automatic hyphenation; these
exceptions override any rules that would normally apply to a word
matching a hyphenation exception defined with @code{hw}.

Situations also arise when only a specific occurrence of a word needs
its hyphenation altered or suppressed, or when a URL or similar string
needs to be breakable in sensible places without hyphenation.

@DefescList {\\%, , , }
@DefescListEndx {\:, , , }
@cindex hyphenation character (@code{\%})
@cindex character, hyphenation (@code{\%})
@cindex disabling hyphenation (@code{\%})
@cindex hyphenation, disabling (@code{\%})
To tell GNU @code{troff} how to hyphenate words as they occur in input,
use the @code{\%} escape sequence; it is the default @dfn{hyphenation
character}.  Each instance within a word indicates to GNU @code{troff}
that the word may be hyphenated at that point, while prefixing a word
with this escape sequence prevents it from being otherwise hyphenated.
This mechanism affects only that occurrence of the word; to change the
hyphenation of a word for the remainder of input processing, use the
@code{hw} request.

@cindex @code{\X}, followed by @code{\%}
@cindex @code{\Y}, followed by @code{\%}
@cindex @code{\%}, following @code{\X} or @code{\Y}
GNU @code{troff} regards the escape sequences @code{\X} and @code{\Y} as
starting a word; that is, the @code{\%} escape sequence in, say,
@w{@samp{\X'...'\%foobar}} or @w{@samp{\Y'...'\%foobar}} no longer
prevents hyphenation of @samp{foobar} but inserts a hyphenation point
just prior to it; most likely this isn't what you want.
@xref{Postprocessor Access}.

@cindex non-printing break point (@code{\:})
@cindex breaking without hyphens (@code{\:})
@cindex file names, breaking (@code{\:})
@cindex breaking file names (@code{\:})
@cindex URLs, breaking (@code{\:})
@cindex breaking URLs (@code{\:})
@code{\:} inserts a non-printing break point; that is, a word can break
there, but the soft hyphen glyph (see below) is not written to the
output if it does.  This escape sequence is an input word boundary, so
the remainder of the word is subject to hyphenation as normal.

You can combine @code{\:} and @code{\%} to control breaking of a file
name or URL, or to permit hyphenation only after certain explicit
hyphens within a word.

@Example
@c Wrap example at 56 columns.
The \%Lethbridge-Stewart-\:\%Sackville-Baggins divorce
was, in retrospect, inevitable once the contents of
\%/var/log/\:\%httpd/\:\%access_log on the family web
server came to light, revealing visitors from Hogwarts.
@endExample
@endDefesc

@Defreq {hc, [@Var{char}]}
Change the hyphenation character to @var{char}.  This character then
works as the @code{\%} escape sequence normally does, and thus no longer
appears in the output.@footnote{@code{\%} itself stops marking
hyphenation points but still produces no output glyph.}  Without an
argument, @code{hc} resets the hyphenation character to @code{\%} (the
default).  The hyphenation character is associated with the environment
(@pxref{Environments}).
@endDefreq

@Defreq {shc, [@Var{glyph}]}
@cindex soft hyphen character, setting (@code{shc})
@cindex character, soft hyphen, setting (@code{shc})
@cindex glyph, soft hyphen (@code{hy})
@cindex soft hyphen glyph (@code{hy})
@cindex @code{char} request, and soft hyphen character
@cindex @code{tr} request, and soft hyphen character
Set the @dfn{soft hyphen character},@footnote{``Soft hyphen
@emph{character}'' is a misnomer since it is an output glyph.} inserted
when a word is hyphenated automatically or at a hyphenation character,
to@tie{}@var{glyph}.@footnote{``Soft'' because it appears in output only
where a hyphenation break is performed; a ``hard'' hyphen, as in
``long-term'', always appears.}  If the argument is omitted, the soft
hyphen glyph is set to the default, @code{\[hy]}.  If the selected glyph
does not exist in the font in use at a potential hyphenation point, then
the line is not broken at that point.  Neither character definitions
(specified with the @code{char} and similar requests) nor translations
(specified with the @code{tr} request) are considered when assigning the
soft hyphen glyph.
@endDefreq

@cindex hyphenation parameters, automatic
@cindex automatic hyphenation parameters
Several requests influence automatic hyphenation.  Because conventions
vary, a variety of hyphenation modes is available to the @code{hy}
request; these determine whether hyphenation will apply to a
word prior to breaking a line at the end of a page (more or less; see
below for details), and at which positions within that word
automatically determined hyphenation points are permissible.  The places
within a word that are eligible for hyphenation are determined by
language-specific data and lettercase relationships.  Furthermore,
hyphenation of a word might be suppressed due to a limit on
consecutive hyphenated lines (@code{hlm}), a minimum line length
threshold, certain minimum length (@code{hym}), or because the line can
instead be adjusted with additional inter-word space (@code{hys}).

@cindex hyphenation mode register (@code{.hy})
@DefreqList {hy, [@Var{mode}]}
@DefregListEndx {.hy}
Set automatic hyphenation mode to @var{mode}, an integer encoding
conditions for hyphenation; if omitted, @samp{1} is implied.  The
hyphenation mode is available in the read-only register @samp{.hy}; it
is associated with the environment (@pxref{Environments}).  The default
hyphenation mode depends on the localization file loaded when GNU
@code{troff} starts up; see the @code{hpf} request below.

Typesetting practice generally does not avail itself of every
opportunity for hyphenation, but the details differ by language and site
mandates.  The hyphenation modes of @acronym{AT&T} @code{troff} were
implemented with English-language publishing practices of the 1970s in
mind, not a scrupulous enumeration of conceivable parameters.  GNU
@code{troff} extends those modes such that finer-grained control is
possible, favoring compatibility with older implementations over a more
intuitive arrangement.  The means of hyphenation mode control is a set
of numbers that can be added up to encode the behavior
sought.@footnote{The mode is a vector of Booleans encoded as an integer.
To a programmer, this fact is easily deduced from the exclusive use of
powers of two for the configuration parameters; they are computationally
easy to ``mask off'' and compare to zero.  To almost everyone else, the
arrangement seems recondite and unfriendly.}  The entries in the
following table are termed @dfn{values}; the sum of the desired
values is the @dfn{mode}.

@table @code
@item 0
disables hyphenation.

@item 1
enables hyphenation except after the first and before the last character
of a word.
@end table

The remaining values ``imply'' 1; that is, they enable hyphenation
under the same conditions as @samp{.hy 1}, and then apply or lift
restrictions relative to that basis.

@table @code
@item 2
disables hyphenation of the last word on a page,@footnote{Hyphenation is
prevented if the next page location trap is closer to the vertical
drawing position than the next text baseline would be.  GNU @code{troff}
automatically inserts an implicit vertical position trap at the end of
each page to cause a page transition.  Users or macro packages can set
such traps explicitly to prevent hyphenation of the last word in a
column in multi-column page layouts or before floating figures or
tables.  @xref{Page Location Traps}.} even for manually hyphenated
words.

@item 4
disables hyphenation before the last two characters of a word.

@item 8
disables hyphenation after the first two characters of a word.

@item 16
enables hyphenation before the last character of a word.

@item 32
enables hyphenation after the first character of a word.
@end table

Apart from value@tie{}2, restrictions imposed by the hyphenation mode
are @emph{not} respected for words whose hyphenations have been
specified with the hyphenation character (@samp{\%} by default) or the
@code{hw} request.

Nonzero values in the previous table are additive.  For example,
mode@tie{}12 causes GNU @code{troff} to hyphenate neither the last two
nor the first two characters of a word.  Some values cannot be used
together because they contradict; for instance, values 4 and@tie{}16,
and values 8 and@tie{}32.  As noted, it is superfluous to add 1 to any
nonzero even mode.

@cindex hyphenation pattern files
@cindex pattern files, for hyphenation
The automatic placement of hyphens in words is determined by
@dfn{pattern files}, which are derived from @TeX{} and available for
several languages.  The number of characters at the beginning of a word
after which the first hyphenation point should be inserted is determined
by the patterns themselves; it can't be reduced further without
introducing additional, invalid hyphenation points (unfortunately, this
information is not part of a pattern file---you have to know it in
advance).  The same is true for the number of characters at the end of
a word before the last hyphenation point should be inserted.  For
example, you can supply the following input to @samp{echo $(nroff)}.

@Example
.ll 1
.hy 48
splitting
@endExample

@noindent
You will get

@Example
s- plit- t- in- g
@endExample

@noindent
instead of the correct `split- ting'.  English patterns as distributed
with GNU @code{troff} need two characters at the beginning and three
characters at the end; this means that value@tie{}4 of @code{hy} is
mandatory.  Value@tie{}8 is possible as an additional restriction, but
values@tie{}16 and@tie{}32 should be avoided, as should mode@tie{}1.
Modes@tie{}4 and@tie{}6 are typical.

A table of left and right minimum character counts for hyphenation as
needed by the patterns distributed with GNU @code{troff} follows; see
the @cite{groff_tmac@r{(5)}} man page for more information on GNU
@code{troff}'s language macro files.

@multitable {German traditional}    {pattern name}    {left min}    {right min}
@headitem    language           @tab pattern name @tab left min @tab right min
@item        Czech              @tab cs           @tab 2        @tab 2
@item        English            @tab en           @tab 2        @tab 3
@item        French             @tab fr           @tab 2        @tab 3
@item        German traditional @tab det          @tab 2        @tab 2
@item        German reformed    @tab den          @tab 2        @tab 2
@item        Italian            @tab it           @tab 2        @tab 2
@item        Swedish            @tab sv           @tab 1        @tab 2
@end multitable

Hyphenation exceptions within pattern files (i.e., the words within a
@TeX{} @code{\hyphenation} group) obey the hyphenation restrictions
given by @code{hy}.
@endDefreq

@Defreq {nh, }
Disable automatic hyphenation; i.e., set the hyphenation mode to@tie{}0
(see above).  The hyphenation mode of the last call to @code{hy} is not
remembered.
@endDefreq

@DefreqList {hpf, pattern-file}
@DefreqItemx {hpfa, pattern-file}
@DefreqListEndx {hpfcode, a b [c d] @dots{}}
@cindex hyphenation patterns (@code{hpf})
@cindex patterns for hyphenation (@code{hpf})
Read hyphenation patterns from @var{pattern-file}, which is sought
in the same way that macro files are with the @code{mso} request or the
@option{-m@var{name}} command-line option to @code{groff}.  The
@var{pattern-file} should have the same format as (simple) @TeX{}
pattern files.  More specifically, the following scanning rules are
implemented.

@itemize @bullet
@item
A percent sign starts a comment (up to the end of the line) even if
preceded by a backslash.

@item
``Digraphs'' like @code{\$} are not supported.

@item
@code{^^@var{xx}} (where each @var{x} is 0--9 or a--f) and
@code{^^@var{c}} (character @var{c} in the code point range 0--127
decimal) are recognized; other uses of @code{^} cause an error.

@item
No macro expansion is performed.

@item
@code{hpf} checks for the expression @code{\patterns@{@dots{}@}}
(possibly with whitespace before or after the braces).  Everything
between the braces is taken as hyphenation patterns.  Consequently,
@code{@{} and @code{@}} are not allowed in patterns.

@item
Similarly, @code{\hyphenation@{@dots{}@}} gives a list of hyphenation
exceptions.

@item
@code{\endinput} is recognized also.

@item
For backward compatibility, if @code{\patterns} is missing, the whole
file is treated as a list of hyphenation patterns (except that the
@code{%} character is recognized as the start of a comment).
@end itemize

The @code{hpfa} request appends a file of patterns to the current list.

The @code{hpfcode} request defines mapping values for character codes in
pattern files.  It is an older mechanism no longer used by GNU
@code{troff}'s own macro files; for its successor, see @code{hcode}
below.  @code{hpf} or @code{hpfa} apply the mapping after reading the
patterns but before replacing or appending to the active list of
patterns.  Its arguments are pairs of character codes---integers from 0
to@tie{}255.  The request maps character code@tie{}@var{a} to
code@tie{}@var{b}, code@tie{}@var{c} to code@tie{}@var{d}, and so on.
Character codes that would otherwise be invalid in GNU @code{troff} can
be used.  By default, every code maps to itself except those for letters
`A' to `Z', which map to those for `a' to `z'.

@cindex localization
@pindex troffrc
@pindex cs.tmac
@pindex de.tmac
@pindex en.tmac
@pindex fr.tmac
@pindex it.tmac
@pindex ja.tmac
@pindex sv.tmac
@pindex zh.tmac
The set of hyphenation patterns is associated with the language set by
the @code{hla} request (see below).  The @code{hpf} request is usually
invoked by a localization file loaded by the @file{troffrc}
file.@footnote{For more on localization, see the
@cite{groff_tmac@r{(5)}} man page.}

A second call to @code{hpf} (for the same language) replaces the
hyphenation patterns with the new ones.  Invoking @code{hpf} or
@code{hpfa} causes an error if there is no hyphenation language.  If no
@code{hpf} request is specified (either in the document, in a file
loaded at startup, or in a macro package), GNU @code{troff} won't
automatically hyphenate at all.
@endDefreq

@Defreq {hcode, c1 code1 [c2 code2] @dots{}}
@cindex hyphenation code (@code{hcode})
@cindex code, hyphenation (@code{hcode})
Set the hyphenation code of character @var{c1} to @var{code1}, that of
@var{c2} to @var{code2}, and so on.  A hyphenation code must be an
ordinary character (not a special character escape sequence) other than
a digit or a space.  The request is ignored if given no arguments.

For hyphenation to work, hyphenation codes must be set up.  At
startup, GNU @code{troff} assigns hyphenation codes to the letters
@samp{a}--@samp{z} (mapped to themselves), to the letters
@samp{A}--@samp{Z} (mapped to @samp{a}--@samp{z}), and zero to all other
characters.  Normally, hyphenation patterns contain only lowercase
letters which should be applied regardless of case.  In other words,
they assume that the words `FOO' and `Foo' should be hyphenated exactly
as `foo' is.  The @code{hcode} request extends this principle to letters
outside the Unicode basic Latin alphabet; without it, words containing
such letters won't be hyphenated properly even if the corresponding
hyphenation patterns contain them.

For example, the following @code{hcode} requests are necessary to assign
hyphenation codes to the letters @samp{ÄäÖöÜüß}, needed for German.

@Example
.hcode ä ä  Ä ä
.hcode ö ö  Ö ö
.hcode ü ü  Ü ü
.hcode ß ß
@endExample

Without these assignments, GNU @code{troff} treats the German word
@w{`Kindergärten'} (the plural form of `kindergarten') as two words
@w{`kinderg'} and @w{`rten'} because the hyphenation code of the
umlaut@tie{}a is zero by default, just like a space.  There is a German
hyphenation pattern that covers @w{`kinder'}, so GNU @code{troff} finds
the hyphenation `kin-der'.  The other two hyphenation points
(`kin-der-gär-ten') are missed.
@endDefreq

@DefreqList {hla, lang}
@DefregListEndx {.hla}
@cindex @code{hpf} request, and hyphenation language
@cindex @code{hw} request, and hyphenation language
@pindex troffrc
@pindex troffrc-end
Set the hyphenation language to @var{lang}.  Hyphenation exceptions
specified with the @code{hw} request and hyphenation patterns and
exceptions specified with the @code{hpf} and @code{hpfa} requests are
associated with the hyphenation language.  The @code{hla} request is
usually invoked by a localization file, which is turn loaded by the
@file{troffrc} or @file{troffrc-end} file; see the @code{hpf} request
above.

@cindex hyphenation language register (@code{.hla})
The hyphenation language is available in the read-only string-valued
register @samp{.hla}; it is associated with the environment
(@pxref{Environments}).
@endDefreq

@DefreqList {hlm, [@Var{n}]}
@DefregItemx {.hlm}
@DefregListEndx {.hlc}
@cindex explicit hyphen (@code{\%})
@cindex hyphen, explicit (@code{\%})
@cindex consecutive hyphenated lines (@code{hlm})
@cindex lines, consecutive hyphenated (@code{hlm})
@cindex hyphenated lines, consecutive (@code{hlm})
Set the maximum quantity of consecutive hyphenated lines to @var{n}.  If
@var{n} is negative, there is no maximum.  If omitted, @var{n}
is@tie{}@minus{}1.  This value is associated with the environment
(@pxref{Environments}).  Only lines output from a given environment
count toward the maximum associated with that environment.  Hyphens
resulting from @code{\%} are counted; explicit hyphens are not.

@cindex hyphenation consecutive line limit register (@code{.hlm})
@cindex hyphenation consecutive line count register (@code{.hlc})
The @code{.hlm} read-only register stores this maximum.  The count of
immediately preceding consecutive hyphenated lines is available in the
read-only register @code{.hlc}.
@endDefreq

@DefreqList {hym, [@Var{length}]}
@DefregListEndx {.hym}
@cindex hyphenation margin (@code{hym})
@cindex margin for hyphenation (@code{hym})
@cindex @code{ad} request, and hyphenation margin
Set the (right) hyphenation margin to @var{length}.  If the adjustment
mode is not @samp{b} or @samp{n}, the line is not hyphenated if it is
shorter than @var{length}.  Without an argument, the hyphenation margin
is reset to its default value, 0.  The default scaling indicator is
@samp{m}.  The hyphenation margin is associated with the environment
(@pxref{Environments}).

A negative argument resets the hyphenation margin to zero, emitting a
warning in category @samp{range}.

@cindex hyphenation margin register (@code{.hym})
The hyphenation margin is available in the @code{.hym} read-only
register.
@endDefreq

@DefreqList {hys, [@Var{hyphenation-space}]}
@DefregListEndx {.hys}
@cindex hyphenation space (@code{hys})
@cindex hyphenation space adjustment threshold
@cindex @code{ad} request, and hyphenation space
Suppress hyphenation of the line in adjustment modes @samp{b} or
@samp{n} if it can be justified by adding no more than
@var{hyphenation-space} extra space to each inter-word space.  Without
an argument, the hyphenation space adjustment threshold is set to its
default value, 0.  The default scaling indicator is @samp{m}.  The
hyphenation space adjustment threshold is associated with the
environment (@pxref{Environments}).

A negative argument resets the hyphenation space adjustment threshold to
zero, emitting a warning in category @samp{range}.

@cindex hyphenation space adjustment threshold register (@code{.hys})
The hyphenation space adjustment threshold is available in the
@code{.hys} read-only register.
@endDefreq


@c =====================================================================

@node Manipulating Spacing, Tabs and Fields, Manipulating Hyphenation, GNU troff Reference
@section Manipulating Spacing
@cindex manipulating spacing
@cindex spacing, manipulating

@Defreq {sp, [@Var{distance}]}
Break the line and place the next text baseline @var{distance} lower
than it would otherwise be, or until springing a page location
trap.@footnote{@xref{Page Location Traps}.}  If invoked with the
no-break control character, @code{sp} moves the pending output line's
text baseline by @var{distance}.  If @var{distance} would move the
text baseline below the bottom of the page, the page is broken and
any leftover distance discarded.  A negative distance moves the text
baseline up the page, but will not reduce it below zero.  Inside a
diversion, any @var{distance} argument is ignored.  The default scaling
unit is @samp{v}.  If @var{distance} is not specified, @samp{1v} is
assumed.

@Example
.pl 5v \" Set page length to 5 vees.
.de xx
\-\-\-
.  br
..
.wh 0 xx \" Set a trap at the top of the page.
foo on page \n%
.sp 2v
bar on page \n%
.sp 50v \" This will cause a page break.
baz on page \n%
.pl \n(nlu \" Truncate page to current position.
    @result{} ---
    @result{} foo on page 1
    @result{}
    @result{}
    @result{} bar on page 1
    @result{} ---
    @result{} baz on page 2
@endExample

You might use the following macros to set the baseline of the next
output text at a given distance from the top or the bottom of the page.
We subtract one line height (@code{\n[.v]}) because the @code{|}
operator moves to one vee below the page top (recall @ref{Numeric
Expressions}).

@Example
.de y-from-top-down
.  sp |\\$1-\\n[.v]u
..
.
.de y-from-bot-up
.  sp |\\n[.p]u-\\$1-\\n[.v]u
..
@endExample

@noindent
A call to @samp{.y-from-bot-up 10c} means that the next text baseline
will be 10@tie{}cm from the bottom edge of the paper.
@endDefreq

@DefreqList {ls, [@Var{count}]}
@DefregListEndx {.L}
@cindex double-spacing (@code{ls})
Set the line spacing; add @w{@var{count}@minus{}1} blank lines after each
line of text.  With no argument, GNU @code{troff} uses the previous
value before the last @code{ls} call.  The default is @code{1}.

@c This example is fairly obvious, doesn't realistically reflect the
@c fact that formatted text would occur between each of these requests,
@c and doesn't fit well on the (PDF) page as of this writing.
@c @Example
@c .ls 2    \" begin double-spaced output
@c .ls 3    \" begin triple-spaced output
@c .ls      \" return to double-spaced output
@c @endExample

@cindex line spacing register (@code{.L})
The read-only register @code{.L} contains the current line spacing; it
is associated with the environment (@pxref{Environments}).
@endDefreq

The @code{ls} request is a coarse mechanism.  @xref{Changing the Type
Size}, for the requests @code{vs} and @code{pvs} as alternatives to
@code{ls}.

@DefescList {\\x, @code{'}, spacing, @code{'}}
@DefregListEndx {.a}
Sometimes, an output line requires additional vertical spacing, for
instance to allow room for a tall construct like an inline equation with
exponents or subscripts (particularly if they are iterated).  The
@code{\x} escape sequence takes a delimited measurement (like
@samp{\x'3p'}) to increase the vertical spacing of the pending output
line.  The default scaling unit is @samp{v}.  If the measurement is
positive, extra vertical space is inserted below the current line; a
negative measurement adds space above.  If @code{\x} is applied to the
pending output line multiple times, the maxima of the positive and
negative adjustments are separately applied.  The delimiter need not be
a neutral apostrophe; see @ref{Delimiters}.

@cindex extra post-vertical line space register (@code{.a})
The @code{.a} read-only register contains the extra vertical spacing
@emph{after} the text baseline of the most recently emitted output line.
(In other words, it is the largest positive argument to @code{\x}
encountered on that line.)  This quantity is exposed via a register
because if an output line requires this ``extra post-vertical line
spacing'', and the subsequent output line requires ``extra pre-vertical
line spacing'' (a negative argument to @code{\x}), then applying both
can lead to excessive spacing between the output lines.  Text that is
piling high on line @var{n} might not require (as much) extra
pre-vertical line spacing if line @var{n}@minus{}1 carries extra
post-vertical line spacing.

Use of @code{\x} can be necessary in combination with the
bracket-building escape sequence @code{\b},@footnote{@xref{Drawing
Requests}.} as the following example shows.

@Example
.nf
This is a test of \[rs]b (1).
This is a test of \[rs]b (2).
This is a test of \b'xyz'\x'-1m'\x'1m' (3).
This is a test of \[rs]b (4).
This is a test of \[rs]b (5).
    @result{} This is a test of \b (1).
    @result{} This is a test of \b (2).
    @result{}                   x
    @result{} This is a test of y (3).
    @result{}                   z
    @result{} This is a test of \b (4).
    @result{} This is a test of \b (5).
@endExample
@endDefesc

@noindent
Without @code{\x}, the backslashes on the lines marked @samp{(2)} and
@samp{(4)} would be overprinted.

@DefreqList {ns, }
@DefreqItemx {rs, }
@DefregListEndx {.ns}
@cindex @code{sp} request, and no-space mode
@cindex no-space mode (@code{ns})
@cindex mode, no-space (@code{ns})
@cindex blank lines, disabling
@cindex lines, blank, disabling
Enable @dfn{no-space mode}.  Vertical spacing, whether by @code{sp}
requests or blank input lines, is disabled.  The @code{bp} request to
advance to the next page is also disabled, unless it is accompanied by a
page number (@pxref{Page Control}).  No-space mode ends automatically
when text@footnote{or geometric primitives; see @ref{Drawing Requests}}
is formatted for output @footnote{to the top-level diversion; see
@ref{Diversions}} or the @code{rs} request is invoked, which ends
no-space mode.  The read-only register @code{.ns} interpolates a Boolean
value indicating the enablement of no-space mode.

A paragraphing macro might ordinarily insert vertical space to separate
paragraphs.  A section heading macro could invoke @code{ns} to suppress
this spacing for the first paragraph in a section.
@endDefreq


@c =====================================================================

@node Tabs and Fields, Character Translations, Manipulating Spacing, GNU troff Reference
@section Tabs and Fields
@cindex tabs, and fields
@cindex fields, and tabs

@cindex tab character encoding
A tab character (@acronym{ISO} code point@tie{}9, @acronym{EBCDIC}
code point@tie{}5) causes a horizontal movement to the next tab stop, if
any.

@Defesc {\\t, , , }
@cindex tab character, non-interpreted (@code{\t})
@cindex character, tab, non-interpreted (@code{\t})
@cindex @code{\t}, and copy mode
@cindex copy mode, and @code{\t}
@cindex mode, copy, and @code{\t}
Interpolate a tab in copy mode; see @ref{Copy Mode}.
@endDefesc

@DefreqList {ta, [[@Var{n1} @Var{n2} @dots{} @Var{nn} ]@t{T} @Var{r1} @
 @Var{r2} @dots{} @Var{rn}]}
@DefregListEndx {.tabs}
Change tab stop positions.  This request takes a series of tab
specifiers as arguments (optionally divided into two groups with the
letter @samp{T}) that indicate where each tab stop is to be, overriding
any previous settings.  The default scaling unit is @samp{m}.  Invoking
@code{ta} without an argument removes all tab stops.
@cindex default tab stops
@cindex tab stops, default
GNU @code{troff}'s startup value is @w{@samp{T 0.5i}}.

Tab stops can be specified absolutely---as distances from the left
margin.  The following example sets six tab stops, one every inch.

@Example
.ta 1i 2i 3i 4i 5i 6i
@endExample

Tab stops can also be specified using a leading @samp{+}, which means
that the specified tab stop is set relative to the previous tab stop.
For example, the following is equivalent to the previous example.

@Example
.ta 1i +1i +1i +1i +1i +1i
@endExample

GNU @code{troff} supports an extended syntax to specify repeating tab
stops.  These stops appear after a @samp{T} argument.  Their values are
always taken as distances relative to the previous tab stop.  This is
the idiomatic way to specify tab stops at equal intervals in
@code{groff}.  The following is, yet again, the same as the previous
examples.  It does more, in fact, since it defines an infinite number of
tab stops at one-inch intervals.

@Example
.ta T 1i
@endExample

Now we are ready to interpret the full syntax given above.  The
@code{ta} request sets tabs at positions @var{n1}, @var{n2}, @dots{},
@var{nn}, then at @var{nn}+@var{r1}, @var{nn}+@var{r2}, @dots{},
@var{nn}+@var{rn}, then at @var{nn}+@var{rn}+@var{r1},
@var{nn}+@var{rn}+@var{r2}, @dots{}, @var{nn}+@var{rn}+@var{rn}, and so
on.

For example, @samp{4c +6c T 3c 5c 2c} is equivalent to @samp{4c 10c 13c
18c 20c 23c 28c 30c @dots{}}.

Text written to a tab column (i.e., between two tab stops, or between a
tab stop and an output line boundary) may be aligned to the right or
left, or centered in the column.  This alignment is determined by
appending @samp{R}, @samp{L}, or @samp{C} to the tab specifier.  The
default is @samp{L}.

@Example
.ta 1i 2iC 3iR
@endExample

The beginning of an output line is not a tab stop; the text that begins
an output line is placed according to the configured alignment and
indentation; see @ref{Manipulating Filling and Adjustment} and @ref{Line
Layout}.

A tab stop is converted into a non-breakable horizontal movement that
cannot be adjusted.

@Example
.ll 2i
.ds foo a\tb\tc
.ta T 1i
\*[foo]
    @error{} warning: cannot break line
    @result{} a         b         c
@endExample

@noindent
The above creates a single output line that is a bit longer than two
inches (we use a string to show exactly where the tab stops are).
Now consider the following.

@Example
.ll 2i
.ds bar a\tb c\td
.ta T 1i
\*[bar]
    @error{} warning: cannot adjust line
    @result{} a         b
    @result{} c       d
@endExample

@noindent
GNU @code{troff} first converts the line's tab stops into unbreakable
horizontal movements, then breaks after @samp{b}.  This usually isn't
what you want.

Superfluous tab characters---those that do not correspond to a tab
stop---are ignored except for the first, which delimits the characters
belonging to the last tab stop for right-alignment or centering.

@Example
.ds Z   foo\tbar\tbaz
.ds ZZ  foo\tbar\tbazqux
.ds ZZZ foo\tbar\tbaz\tqux
.ta 2i 4iR
\*[Z]
.br
\*[ZZ]
.br
\*[ZZZ]
.br
    @result{} foo                 bar              baz
    @result{} foo                 bar           bazqux
    @result{} foo                 bar              bazqux
@endExample

@noindent
The first line right-aligns ``baz'' within the second tab stop.  The
second line right-aligns ``bazqux'' within it.  The third line
right-aligns only ``baz'' because of the additional tab character, which
marks the end of the text occupying the last tab stop defined.

Tab stops are associated with the environment (@pxref{Environments}).

@cindex tab stop settings register (@code{.tabs})
@cindex @code{.S} register, Plan@tie{}9 alias for @code{.tabs}
@cindex @code{.tabs} register, Plan@tie{}9 alias (@code{.S})
The read-only register @code{.tabs} contains a string
representation of the current tab settings suitable for use as an
argument to the @code{ta} request.@footnote{Plan@tie{}9 @code{troff}
uses the register @code{.S} for this purpose.}

@Example
.ds tab-string \n[.tabs]
\*[tab-string]
    @result{} T120u
@endExample
@endDefreq

@Defreq {tc, [@Var{fill-glyph}]}
@cindex tab repetition character (@code{tc})
@cindex character, tab repetition (@code{tc})
@cindex glyph, tab repetition (@code{tc})
Normally, GNU @code{troff} writes no glyph when moving to a tab stop
(some output devices may explicitly output space characters to achieve
this motion).  A @dfn{tab repetition character} can be specified with
the @code{tc} request, causing GNU @code{troff} to write as many
instances of @var{fill-glyph} as are necessary to occupy the interval
from the current horizontal location to the next tab stop.  With no
argument, GNU @code{troff} reverts to the default behavior.  The tab
repetition character is associated with the environment
(@pxref{Environments}).@footnote{Tab repetition @emph{character} is a
misnomer since it is an output glyph.}  Only a single @var{fill-glyph}
is recognized; any excess is ignored.
@endDefreq

@DefreqList {linetabs, n}
@DefregListEndx {.linetabs}
@cindex tab, line-tabs mode
@cindex line-tabs mode
@cindex mode, line-tabs
If @var{n} is missing or nonzero, activate @dfn{line-tabs}; deactivate
it otherwise (the default).  Active line-tabs cause GNU @code{troff}
to compute tab distances relative to the start of the output line
instead of the input line.

@Example
.de Tabs
.  ds x a\t\c
.  ds y b\t\c
.  ds z c
.  ta 1i 3i
\\*x
\\*y
\\*z
..
.Tabs
.br
.linetabs
.Tabs
    @result{} a         b         c
    @result{} a         b                   c
@endExample

Line-tabs activation is associated with the environment
(@pxref{Environments}).  The read-only register @code{.linetabs}
interpolates@tie{}1 if line-tabs are active, and 0 otherwise.
@endDefreq

@menu
* Leaders::
* Fields::
@end menu

@c ---------------------------------------------------------------------

@node Leaders, Fields, Tabs and Fields, Tabs and Fields
@subsection Leaders
@cindex leaders

Sometimes it is desirable to fill a tab stop with a given glyph,
but also use tab stops normally on the same output line.  An example is
a table of contents entry that uses dots to bridge the entry name with
its page number, which is itself aligned within a tab stop.  The
@code{roff} language provides @dfn{leaders} for this
purpose.@footnote{This is pronounced to rhyme with ``feeder'', and
refers to how the glyphs ``lead'' the eye across the page to the
corresponding page number or other datum.}

@cindex leader character
A leader character (@acronym{ISO} and @acronym{EBCDIC} code
point@tie{}1, also known as @acronym{SOH} or ``start of heading''),
behaves similarly to a tab character:@: it moves to the next tab stop.
The difference is that for this movement, the default fill glyph is a
period @samp{.}.

@Defesc {\\a, , , }
@cindex leader character, non-interpreted (@code{\a})
@cindex character, leader, non-interpreted (@code{\a})
@cindex @code{\a}, and copy mode
@cindex copy mode, and @code{\a}
@cindex mode, copy, and @code{\a}
Interpolate a leader in copy mode; see @ref{Copy Mode}.
@endDefesc

@Defreq {lc, [@Var{fill-glyph}]}
@cindex leader repetition character (@code{lc})
@cindex character, leader repetition (@code{lc})
@cindex glyph, leader repetition (@code{lc})
When writing a leader, GNU @code{troff} fills the space to the next tab
stop with dots @samp{.}.  A different @dfn{leader repetition character}
can be specified with the @code{lc} request, causing GNU @code{troff} to
write as many instances of @var{fill-glyph} as are necessary to occupy
the interval from the current horizontal location to the next tab stop.
With no argument, GNU @code{troff} treats leaders the same as tabs.  The
leader repetition character is associated with the environment
(@pxref{Environments}).@footnote{Leader repetition @emph{character} is a
misnomer since it is an output glyph.}  Only a single @var{fill-glyph}
is recognized; any excess is ignored.
@endDefreq

@cindex table of contents
@cindex contents, table of
A table of contents, for example, may define tab stops after a section
number, a title, and a gap to be filled with leader dots.  The page
number follows the leader, after a right-aligned final tab stop wide
enough to house the largest page number occurring in the document.

@Example
.ds entry1 19.\tThe Prophet\a\t98
.ds entry2 20.\tAll Astir\a\t101
.ta .5i 4.5i +.5iR
.nf
\*[entry1]
\*[entry2]
    @result{} 19.  The Prophet.............................   98
    @result{} 20.  All Astir...............................  101
@endExample

@c ---------------------------------------------------------------------

@node Fields,  , Leaders, Tabs and Fields
@subsection Fields
@cindex fields

@cindex field delimiting character (@code{fc})
@cindex delimiting character, for fields (@code{fc})
@cindex character, field delimiting (@code{fc})
@cindex field padding character (@code{fc})
@cindex padding character, for fields (@code{fc})
@cindex character, field padding (@code{fc})
@dfn{Fields} are a more general way of laying out tabular data.  A field
is defined as the data between a pair of @dfn{delimiting characters}.
It contains substrings that are separated by @dfn{padding characters}.
The width of a field is the distance on the @emph{input} line from the
position where the field starts to the next tab stop.  A padding
character inserts an adjustable space similar to @TeX{}'s @code{\hss}
command (thus it can even be negative) to make the sum of all substring
lengths plus the adjustable space equal to the field width.  If more
than one padding character is inserted, the available space is evenly
distributed among them.

@Defreq {fc, [@Var{delim-char} [@Var{padding-char}]]}
Define a delimiting and a padding character for fields.  If the latter
is missing, the padding character defaults to a space character.  If
there is no argument at all, the field mechanism is disabled (which is
the default).  In contrast to, e.g., the tab repetition character,
delimiting and padding characters are @emph{not} associated with the
environment (@pxref{Environments}).

@Example
.fc # ^
.ta T 3i
#foo^bar^smurf#
.br
#foo^^bar^smurf#
    @result{} foo         bar          smurf
    @result{} foo            bar       smurf
@endExample
@endDefreq


@c =====================================================================

@node Character Translations, @code{troff} and @code{nroff} Modes, Tabs and Fields, GNU troff Reference
@section Character Translations
@cindex character translations
@cindex translations of characters

A @dfn{translation} is a mapping of an input character to an output
glyph.  The mapping occurs at output time, i.e., the input character
gets assigned the metric information of the mapped output character
right before input tokens are converted to nodes (@pxref{Gtroff
Internals}, for more on this process).

@DefreqList {tr, @Var{a}@Var{b}@Var{c}@Var{d}@dots{}}
@DefreqListEndx {trin, @Var{a}@Var{b}@Var{c}@Var{d}@dots{}}
Translate character @var{a} to glyph@tie{}@var{b}, character @var{c} to
glyph@tie{}@var{d}, and so on.  If there is an odd number of characters
in the argument, the last one is translated to a fixed-width space (the
same one obtained by the @code{\@key{SP}} escape sequence).

The @code{trin} request is identical to @code{tr}, but when you unformat
a diversion with @code{asciify} it ignores the translation.
@xref{Diversions}, for details about the @code{asciify} request.

Some notes:

@itemize @bullet
@item
@cindex @code{\(}, and translations
@cindex @code{\[}, and translations
@cindex @code{\'}, and translations
@cindex @code{\`}, and translations
@cindex @code{\-}, and translations
@cindex @code{\_}, and translations
@cindex @code{\C}, and translations
@cindex @code{\N}, and translations
@cindex @code{char} request, and translations
@cindex special characters
@cindex character, special
@cindex numbered glyph (@code{\N})
@cindex glyph, numbered (@code{\N})
Special characters (@code{\(@var{xx}}, @code{\[@var{xxx}]},
@code{\C'@var{xxx}'}, @code{\'}, @code{\`}, @code{\-}, @code{\_}),
glyphs defined with the @code{char} request, and numbered glyphs
(@code{\N'@var{xxx}'}) can be translated also.

@item
@cindex @code{\e}, and translations
The @code{\e} escape can be translated also.

@item
@cindex @code{\%}, and translations
@cindex @code{\~}, and translations
Characters can be mapped onto the @code{\%} and @code{\~} escape
sequences (but @code{\%} and @code{\~} can't be mapped onto another
glyph).

@item
@cindex backspace character, and translations
@cindex character, backspace, and translations
@cindex leader character, and translations
@cindex character, leader, and translations
@cindex newline character, and translations
@cindex character, newline, and translations
@cindex tab character, and translations
@cindex character, tab, and translations
@cindex @code{\a}, and translations
@cindex @code{\t}, and translations
The following characters can't be translated: space (with one exception,
see below), backspace, newline, leader (and @code{\a}), tab (and
@code{\t}).

@item
@cindex @code{shc} request, and translations
Translations are not considered for finding the soft hyphen character
set with the @code{shc} request.

@item
@cindex @code{\&}, and translations
The pair @samp{@var{c}\&} (an arbitrary character@tie{}@var{c} followed
by the dummy character) maps this character to nothing.

@Example
.tr a\&
foo bar
    @result{} foo br
@endExample

@noindent
Even the space character can be mapped to the dummy character.

@Example
.tr aa \&
foo bar
    @result{} foobar
@endExample

@noindent
As shown in the example, the space character can't be the first
character/glyph pair as an argument of @code{tr}.  Additionally, it is
not possible to map the space character to any other glyph; requests
like @w{@samp{.tr aa x}} undo @w{@samp{.tr aa \&}} instead.

If justification is active, lines are justified in spite of the `empty'
space character (but there is no minimal distance, i.e., the space
character, between words).

@item
After an output glyph has been constructed (this happens at the moment
immediately before the glyph is appended to an output glyph list, either
by direct output, in a macro, diversion, or string), it is no longer
affected by @code{tr}.

@item
Translating character to glyphs where one of them or both are undefined
is possible also; @code{tr} does not check whether the elements of its
argument exist.

@xref{Gtroff Internals}.

@item
Without an argument, the @code{tr} request is ignored.
@end itemize
@endDefreq

@Defreq {trnt, @Var{a}@Var{b}@Var{c}@Var{d}@dots{}}
@cindex @code{\!}, and @code{trnt}
@code{trnt} is the same as the @code{tr} request except that the
translations do not apply to text that is transparently throughput into
a diversion with @code{\!}.  @xref{Diversions}.

For example,

@Example
.tr ab
.di x
\!.tm a
.di
.x
@endExample

@noindent
prints @samp{b} to the standard error stream; if @code{trnt} is used
instead of @code{tr} it prints @samp{a}.
@endDefreq


@c =====================================================================

@node @code{troff} and @code{nroff} Modes, Line Layout, Character Translations, GNU troff Reference
@section @code{troff} and @code{nroff} Modes
@cindex @code{troff} mode
@cindex mode, @code{troff}
@cindex @code{nroff} mode
@cindex mode, @code{nroff}

Historically, @code{nroff} and @code{troff} were two separate programs;
the former for terminal output, the latter for typesetters.  GNU
@code{troff} merges both functions into one executable@footnote{A
GNU @command{nroff} program is available for convenience; it calls GNU
@code{troff} to perform the formatting.} that sends its output to a
device driver (@code{grotty} for terminal devices, @code{grops} for
PostScript, and so on) which interprets this intermediate output format.
When discussing @acronym{AT&T} @code{troff}, it makes sense to talk
about @dfn{@code{nroff} mode} and @dfn{@code{troff} mode} since the
differences are hard-coded.  GNU @code{troff} takes information from
device and font description files without handling requests specially if
a terminal output device is used, so such a strong distinction is
unnecessary.

Usually, a macro package can be used with all output devices.
Nevertheless, it is sometimes necessary to make a distinction between
terminal and non-terminal devices: GNU @code{troff} provides two
built-in conditions @samp{n} and @samp{t} for the @code{if}, @code{ie},
and @code{while} requests to decide whether GNU @code{troff} shall
behave like @code{nroff} or like @code{troff}.

@Defreq {troff, }
@pindex troffrc
@pindex troffrc-end
Make the @samp{t} built-in condition true (and the @samp{n} built-in
condition false) for @code{if}, @code{ie}, and @code{while} conditional
requests.  This is the default if GNU @code{troff} (@emph{not}
@code{groff}) is started with the @option{-R} switch to avoid loading of
the startup files @file{troffrc} and @file{troffrc-end}.  Without
@option{-R}, GNU @code{troff} stays in @code{troff} mode if the output
device is not a terminal (e.g., `ps').
@endDefreq

@Defreq {nroff, }
@pindex tty.tmac
Make the @samp{n} built-in condition true (and the @samp{t} built-in
condition false) for @code{if}, @code{ie}, and @code{while} conditional
requests.  This is the default if GNU @code{troff} uses a terminal
output device; the code for switching to @code{nroff} mode is in the
file @file{tty.tmac}, which is loaded by the startup file
@code{troffrc}.
@endDefreq

@xref{Conditionals and Loops}, for more details on built-in conditions.

@codequotebacktick off
@codequoteundirected off


@c =====================================================================

@node Line Layout, Line Continuation, @code{troff} and @code{nroff} Modes, GNU troff Reference
@section Line Layout
@cindex line layout
@cindex layout, line

@cindex dimensions, line
@cindex line dimensions
The following drawing shows the dimensions that @code{gtroff} uses for
placing a line of output onto the page.  They are labeled with the
request that manipulates each dimension.

@Example
     -->| in |<--
        |<-----------ll------------>|
   +----+----+----------------------+----+
   |    :    :                      :    |
   +----+----+----------------------+----+
-->| po |<--
   |<--------paper width---------------->|
@endExample

@noindent
These dimensions are:

@ftable @code
@item po
@cindex left margin (@code{po})
@cindex margin, left (@code{po})
@cindex page offset (@code{po})
@cindex offset, page (@code{po})
@dfn{Page offset}---this is the leftmost position of text on the final
output, defining the @dfn{left margin}.

@item in
@cindex indentation (@code{in})
@cindex line indentation (@code{in})
@dfn{Indentation}---this is the distance from the left margin where
text is printed.

@item ll
@cindex line length (@code{ll})
@cindex length of line (@code{ll})
@dfn{Line length}---this is the distance from the left margin to right
margin.
@end ftable

@cindex margin, right
@cindex right margin
The right margin is not explicitly configured; the combination of page
offset and line length provides the information necessary to derive it.

A simple demonstration:

@Example
.ll 3i
This is text without indentation.
The line length has been set to 3\~inches.
.in +.5i
.ll -.5i
Now the left and right margins are both increased.
.in
.ll
Calling .in and .ll without parameters restores
the previous values.
@endExample

Result:

@Example
This  is text without indenta-
tion.   The  line  length  has
been set to 3 inches.
     Now   the  left  and
     right  margins   are
     both increased.
Calling  .in  and  .ll without
parameters restores the previ-
ous values.
@endExample

@DefreqList {po, [@Var{offset}]}
@DefreqItem {po, @t{+}@Var{offset}}
@DefreqItem {po, @t{-}@Var{offset}}
@DefregListEndx {.o}
@pindex tty.tmac
Set page offset to @var{offset} (or increment or decrement its current
value by @var{offset}).  If invoked without an argument, the page offset
is restored to the value before the previous @code{po} request.
This request does not cause a break; the page offset in effect when an
output line is broken prevails (@pxref{Manipulating Filling and
Adjustment}).  The initial value is 1@dmn{i} and the default scaling
unit is @samp{m}.  On terminal devices, the page offset is set to zero
by a driver-specific macro file, @file{tty.tmac}.  The current page
offset can be found in the read-only register @samp{.o}.
@cindex CSTR@tie{}#54 errata
@cindex CSTR@tie{}#54 erratum, @code{po} request
This request is incorrectly documented in the @acronym{AT&T}
@code{troff} manual as using a default scaling unit of @samp{v}.

@Example
.po 3i
\n[.o]
    @result{} 720
.po -1i
\n[.o]
    @result{} 480
.po
\n[.o]
    @result{} 720
@endExample
@endDefreq

@DefreqList {in, [@Var{indent}]}
@DefreqItem {in, @t{+}@Var{indent}}
@DefreqItem {in, @t{-}@Var{indent}}
@DefregListEndx {.i}
Set indentation to @var{indent} (or increment or decrement the current
value by @var{indent}).  This request causes a break.  Initially, there
is no indentation.

If @code{in} is called without an argument, the indentation is reset to
the previous value before the last call to @code{in}.  The default
scaling indicator is @samp{m}.

If a negative indentation value is specified (which is not allowed),
@code{gtroff} emits a warning in category @samp{range} and sets the
indentation to zero.

The effect of @code{in} is delayed until a partially collected line (if
it exists) is output.  A temporary indentation value is reset to zero
also.

The current indentation (as set by @code{in}) can be found in the
read-only register @samp{.i}.  The indentation is associated with the
environment (@pxref{Environments}).
@endDefreq

@DefreqList {ti, offset}
@DefreqItem {ti, @t{+}@Var{offset}}
@DefreqItem {ti, @t{-}@Var{offset}}
@DefregListEndx {.in}
Temporarily indent the next output line by @var{offset}.  If an
increment or decrement value is specified, adjust the temporary
indentation relative to the value set by the @code{in} request.

This request causes a break; its value is associated with the
environment (@pxref{Environments}).  The default scaling indicator is
@samp{m}.  A call of @code{ti} without an argument is ignored.

If the total indentation value is negative (which is not allowed),
@code{gtroff} emits a warning in category @samp{range} and sets the
temporary indentation to zero.  `Total indentation' is either
@var{offset} if specified as an absolute value, or the temporary plus
normal indentation, if @var{offset} is given as a relative value.

The effect of @code{ti} is delayed until a partially collected line (if
it exists) is output.

The read-only register @code{.in} is the indentation that applies to the
current output line.

The difference between @code{.i} and @code{.in} is that the latter takes
into account whether a partially collected line still uses the old
indentation value or a temporary indentation value is active.
@endDefreq

@DefreqList {ll, [@Var{length}]}
@DefreqItem {ll, @t{+}@Var{length}}
@DefreqItem {ll, @t{-}@Var{length}}
@DefregItemx {.l}
@DefregListEndx {.ll}
Set the line length to @var{length} (or increment or decrement the
current value by @var{length}).  Initially, the line length is set to
6.5@dmn{i}.  The effect of @code{ll} is delayed until a partially
collected line (if it exists) is output.  The default scaling indicator
is @samp{m}.

If @code{ll} is called without an argument, the line length is reset to
the previous value before the last call to @code{ll}.  If a negative
line length is specified (which is not allowed), @code{gtroff} emits a
warning in category @samp{range} and sets the line length to zero.  The
line length is associated with the environment (@pxref{Environments}).

@cindex line length register (@code{.l})
The current line length (as set by @code{ll}) can be found in the
read-only register @samp{.l}.  The read-only register @code{.ll} is the
line length that applies to the current output line.

Similar to @code{.i} and @code{.in}, the difference between @code{.l}
and @code{.ll} is that the latter takes into account whether a partially
collected line still uses the old line length value.
@endDefreq


@c =====================================================================

@node Line Continuation, Page Layout, Line Layout, GNU troff Reference
@section Line Continuation
@cindex line control
@cindex control, line

When filling is enabled, input and output line breaks generally do not
correspond.  The @code{roff} language therefore distinguishes input and
output line continuation.

@Defesc {\\@key{RET}, , ,}
@cindex input line continuation (@code{\@key{RET}})
@cindex line, input, continuation (@code{\@key{RET}})
@cindex continuation, input line (@code{\@key{RET}})
@c We use the following notation in our man pages; Texinfo is bound to
@c the GNU Emacs dialect.
@esindex \@slanted{newline}
@code{\@key{RET}} (a backslash immediately followed by a newline)
suppresses the effects of that newline in the input.  The next input
line thus retains the classification of its predecessor as a control or
text line.  @code{\@key{RET}} is useful for managing line lengths in the
input during document maintenance; you can break an input line in the
middle of a request invocation, macro call, or escape sequence.  Input
line continuation is invisible to the formatter, with two exceptions:
the @code{|} operator recognizes the new input line
(@pxref{Numeric Expressions}), and the input line counter register
@code{.c} is incremented.

@c Wrap example at 56 columns (on the _output_).  We use 50n in the
@c groff input to avoid line adjustment.
@Example
.ll 50n
.de I
.  ft I
.  nop \\$*
.  ft
..
Our film class watched
.I The Effect of Gamma Rays on Man-in-the-Moon
Marigolds. \" whoops, the input line wrapped
.br
.I My own opus begins on line \n[.c] \
and ends on line \n[.c].
@endExample
@Example
    @result{} Our film class watched @i{The Effect of Gamma Rays on}
    @result{} @i{Man-in-the-Moon} Marigolds.
    @result{} @i{My own opus begins on line 11 and ends on line 12.}
@endExample
@endDefesc

@DefescList {\\c, , ,}
@DefregListEndx {.int}
@cindex output line, continuation (@code{\c})
@cindex line, output, continuation (@code{\c})
@cindex continuation, output line (@code{\c})
@cindex interrupted line
@cindex line, interrupted
@cindex @code{\R}, after @code{\c}
@code{\c} continues an output line.  Nothing on the input line after it
is formatted.  In contrast to @code{\@key{RET}}, a line after @code{\c}
is treated as a new input line, so a control character is recognized at
its beginning.  The visual results depend on whether filling is enabled;
see @ref{Manipulating Filling and Adjustment}.

@itemize @bullet
@item
@cindex @code{\c}, when filling enabled
@cindex fill mode, and @code{\c}
@cindex mode, fill, and @code{\c}
If filling is enabled, a word interrupted with @code{\c} is continued
with the text on the next input text line, without an intervening space.

@Example
This is a te\c
st.
    @result{} This is a test.
@endExample

@item
@cindex @code{\c}, when filling disabled
@cindex no-fill mode, and @code{\c}
@cindex mode, no-fill, and @code{\c}
If filling is disabled, the next input text line after @code{\c} is
handled as a continuation of the same input text line.

@Example
.nf
This is a \c
test.
    @result{} This is a test.
@endExample
@end itemize

An intervening control line that causes a break overrides @code{\c},
flushing out the pending output line in the usual way.

@cindex interrupted line register (@code{.int})
@cindex continued output line register (@code{.int})
The @code{.int} register contains a positive value if the last output
line was continued with @code{\c}; this datum is associated with the
environment (@pxref{Environments}).@footnote{Historically, the @code{\c}
escape sequence has proven challenging to characterize.  Some sources
say it ``connects the next input text'' (to the input line on which it
appears); others describe it as ``interrupting'' text, on the grounds
that a text line is interrupted without breaking, perhaps to inject a
request invocation or macro call.}
@endDefesc


@c =====================================================================

@codequotebacktick on
@codequoteundirected on

@node Page Layout, Page Control, Line Continuation, GNU troff Reference
@section Page Layout
@cindex page layout
@cindex layout, page

GNU @code{troff} provides some primitive operations for controlling page
layout.

@DefreqList {pl, [@Var{length}]}
@DefreqItem {pl, @t{+}@Var{length}}
@DefreqItem {pl, @t{-}@Var{length}}
@DefregListEndx {.p}
@cindex page length (@code{pl})
@cindex length of page (@code{pl})
Set the @dfn{page length} to @var{length} (or increment or decrement the
current value by @var{length}).  This is the length of the physical
output page.  The default scaling indicator is @samp{v}.

@cindex page length register (@code{.p})
The current setting can be found in the read-only register @samp{.p}.

@cindex top margin
@cindex margin, top
@cindex bottom margin
@cindex margin, bottom
This specifies only the size of the page, not the top and bottom
margins.  Those are not set by GNU @code{troff} directly.  @xref{Traps},
for further information on how to do this.

Negative @code{pl} values are possible also, but not very useful: no
trap is sprung, and each line is output on a single page (thus
suppressing all vertical spacing).

If no argument or an invalid argument is given, @code{pl} sets the page
length to 11@dmn{i}.
@endDefreq

@cindex headers
@cindex footers
@cindex titles
GNU @code{troff} provides several operations that help in setting up top
and bottom titles (also known as headers and footers).

@Defreq {tl, @code{'}@Var{left}@code{'}@Var{center}@code{'}@Var{right}@code{'}}
@cindex title line (@code{tl})
@cindex three-part title (@code{tl})
@cindex page number character (@code{%})
Print a @dfn{title line}.  It consists of three parts: a left-justified
portion, a centered portion, and a right-justified portion.  The
argument separator @samp{'} can be replaced with any character not
occurring in the title line.  The @samp{%} character is replaced with
the current page number.  This character can be changed with the
@code{pc} request (see below).  The delimiter need not be a neutral
apostrophe: @code{tl} accepts the same delimiters as most escape
sequences; see @ref{Delimiters}.  Without an argument, @code{tl} is
ignored.

@itemize @bullet
@item
The line length set by the @code{ll} request is not honoured by
@code{tl}; use the @code{lt} request (described below) instead, to
control line length for text set by @code{tl}.

@item
A title line is not restricted to the top or bottom of a page.

@item
@code{tl} prints the title line immediately, ignoring a partially
collected line (which stays untouched).

@item
It is not an error to omit closing delimiters.  For example,
@w{@samp{.tl /foo}} is equivalent to @w{@samp{.tl /foo///}}: It prints a
title line with the left-justified word @samp{foo}; the centered and
right-justified parts are empty.
@end itemize
@endDefreq

@DefreqList {lt, [@Var{length}]}
@DefreqItem {lt, @t{+}@Var{length}}
@DefreqItem {lt, @t{-}@Var{length}}
@DefregListEndx {.lt}
@cindex length of title line (@code{lt})
@cindex title line, length (@code{lt})
@cindex title line length register (@code{.lt})
The title line is printed using its own line length, which is specified
(or incremented or decremented) with the @code{lt} request.  Initially,
the title line length is set to 6.5@dmn{i}.  If a negative line length
is specified (which is not allowed), @code{gtroff} emits a warning in
category @samp{range} and sets the title line length to zero.  The
default scaling indicator is @samp{m}.  If @code{lt} is called without
an argument, the title length is reset to the previous value before the
last call to @code{lt}.  The current setting is available in the
@code{.lt} read-only register; it is associated with the environment
(@pxref{Environments}).
@endDefreq

@DefreqList {pn, page}
@DefreqItem {pn, @t{+}@Var{page}}
@DefreqItem {pn, @t{-}@Var{page}}
@DefregListEndx {.pn}
@cindex page number (@code{pn})
@cindex number, page (@code{pn})
Change (increase or decrease) the page number of the @emph{next} page.
The only argument is the page number; the request is ignored without a
parameter.

The read-only register @code{.pn} contains the number of the next page:
either the value set by a @code{pn} request, or the number of the
current page plus@tie{}1.
@endDefreq

@Defreq {pc, [@Var{char}]}
@cindex changing the page number character (@code{pc})
@cindex page number character, changing (@code{pc})
@vindex %
Change the page number character (used by the @code{tl} request) to a
different character.  With no argument, this mechanism is disabled.
This doesn't affect the register@tie{}@code{%}.
@endDefreq

@xref{Traps}.


@c =====================================================================

@node Page Control, Fonts and Symbols, Page Layout, GNU troff Reference
@section Page Control
@cindex page control
@cindex control, page

@DefreqList {bp, [@Var{page}]}
@DefreqItem {bp, @t{+}@Var{page}}
@DefreqItem {bp, @t{-}@Var{page}}
@DefregListEndx {%}
@cindex new page (@code{bp})
@cindex page, new (@code{bp})
Stop processing the current page and move to the next page.  This
request causes a break.  It can also take an argument to set (increase,
decrease) the page number of the next page (which becomes the current
page after @code{bp} has finished).  The difference between @code{bp}
and @code{pn} is that @code{pn} does not cause a break or actually
eject a page.  @xref{Page Layout}.
@cindex CSTR@tie{}#54 errata
@cindex CSTR@tie{}#54 erratum, @code{bp} request
This request is incorrectly documented in the @acronym{AT&T}
@code{troff} manual as having a default scaling indicator of @samp{v}.

@Example
.de newpage                         \" define macro
'bp                                 \" begin page
'sp .5i                             \" vertical space
.tl 'left top'center top'right top' \" title
'sp .3i                             \" vertical space
..                                  \" end macro
@endExample

@cindex @code{bp} request, and top-level diversion
@cindex top-level diversion, and @code{bp}
@cindex diversion, top-level, and @code{bp}
@code{bp} has no effect if not called within the top-level diversion
(@pxref{Diversions}).

@cindex page number register (@code{%})
@cindex current page number (@code{%})
The writable register@tie{}@code{%} holds the current page number.

The register @code{.pe} is set to@tie{}1 while @code{bp} is active.
@xref{Page Location Traps}.
@endDefreq

@Defreq {ne, [@Var{space}]}
@cindex orphan lines, preventing with @code{ne}
@cindex conditional page break (@code{ne})
@cindex page break, conditional (@code{ne})
Your text may @slanted{need} a certain amount of vertical space before a
page break occurs.  For instance, you may wish to ensure that the first
output line of a paragraph is not @dfn{orphaned} at the bottom of a
page.  The @code{ne} request tests the amount of distance to the next
page location trap (or the page bottom if none is planted earlier; see
@ref{Page Location Traps}), and breaks the page if less than @var{space}
is available.  The default scaling unit is @samp{v}.  If @var{space} is
not specified, @samp{1v} is assumed.

For example, to require room for at least the first two output lines of
a paragraph, you can do the following.

@Example
.ne 2v
Considering how common illness is,
how tremendous the spiritual change that it brings,
how astonishing,
when the lights of health go down,
the undiscovered countries that are then disclosed,
@endExample

@c XXX: Some of this might be better placed in a revised Chapter 3.
This method is reliable only if no output line is pending when @code{ne}
is invoked.  When macro packages are used, this is often not the case:@:
their paragraphing macros perform the break.  You may need to experiment
with placing the @code{ne} after the paragraphing macro, or @code{br}
and @code{ne} before it.

@code{ne} is also useful to force grouping of section headings with
their subsequent paragraphs, or tables with their captions and/or
explanations.  Macro packages often use @code{ne} with diversions to
implement keeps and displays; see @ref{Diversions}.  They may also offer
parameters for widow and orphan management.
@endDefreq

@DefreqList {sv, [@Var{space}]}
@DefreqListEndx {os, }
@cindex @code{ne} request, comparison with @code{sv}
The @code{sv} request is similar to @code{ne}, but @slanted{saves} the
specified vertical space.  If @var{space} is available before the next
page location trap (or the page bottom if none is planted earlier; see
@ref{Page Location Traps}), the space is output immediately.  Any
partially collected line is ignored.  Otherwise, the vertical space is
saved.  Output the space on demand with @code{os}.
@cindex @code{sv} request, and no-space mode
@cindex @code{os} request, and no-space mode
Both @code{sv} and @code{os} ignore no-space mode (recall
@ref{Manipulating Spacing}).  While the @code{sv} request allows
negative values for @var{space}, @code{os} ignores them.  The default
scaling unit is @samp{v}.  If @var{space} is not specified, @samp{1v} is
assumed.
@endDefreq

@Defreg {nl}
@cindex vertical drawing position (@code{nl})
@cindex vertical position, drawing (@code{nl})
@cindex drawing position, vertical (@code{nl})
@code{nl} interpolates or sets the vertical drawing position.  When the
formatter starts and first page transition hasn't happened yet,
@code{nl} is negative.  If a header trap has been planted on the page
(typically at vertical position @code{0}), you can assign a negative
value to @code{nl} to spring it if that page has already started
(@pxref{Page Location Traps}).

@Example
.de header
.  sp
.  tl ''Goldbach Solution''
.  sp
..
.
First page.
@r{@dots{}}
.bp
.wh 0 header \" plant header trap at top of page
.nr nl (-1)
Second page.
@r{@dots{}}
@endExample
@Example
    @result{} First page.
    @result{}
    @result{} @r{@dots{}}
    @result{}
    @result{}                   Goldbach Solution
    @result{}
    @result{} Second page.
    @result{}
    @result{} @r{@dots{}}
@endExample

@noindent
Without resetting @code{nl} to a negative value, the trap just planted
would be active beginning with the @emph{next} page, not the current
one.

@xref{Diversions}, for a comparison of @code{nl} with the @code{.h} and
@code{.d} registers.
@endDefreg

@codequotebacktick off
@codequoteundirected off


@c =====================================================================

@node Fonts and Symbols, Manipulating Type Size and Vertical Spacing, Page Control, GNU troff Reference
@section Fonts and Symbols
@cindex fonts

@code{gtroff} can switch fonts at any point in the text.

The basic set of fonts is @samp{R}, @samp{I}, @samp{B}, and @samp{BI}.
These are Times roman, italic, bold, and bold-italic.  For non-terminal
devices, there is also at least one symbol font that contains various
special symbols (Greek, mathematics).

@menu
* Changing Fonts::
* Font Families::
* Font Positions::
* Using Symbols::
* Character Classes::
* Special Fonts::
* Artificial Fonts::
* Ligatures and Kerning::
* Dummy Characters::
@end menu

@c ---------------------------------------------------------------------

@node Changing Fonts, Font Families, Fonts and Symbols, Fonts and Symbols
@subsection Changing Fonts
@cindex fonts

@DefreqList {ft, [@Var{font}]}
@DefescItemx {\\f, , f, }
@DefescItem {\\f, (, fn, }
@DefescItem {\\f, [, font, ]}
@DefregListEndx {.sty}
@cindex changing fonts (@code{ft}, @code{\f})
@cindex fonts, changing (@code{ft}, @code{\f})
@cindex @code{sty} request, and changing fonts
@cindex @code{fam} request, and changing fonts
@cindex @code{\F}, and changing fonts
@kindex styles
@kindex family
@pindex DESC
The @code{ft} request and the @code{\f} escape change the current font
to @var{font} (one-character name@tie{}@var{f}, two-character name
@var{fn}).

If @var{font} is a style name (as set with the @code{sty} request or
with the @code{styles} command in the @file{DESC} file), use it within
the current font family (as set with the @code{fam} request, the
@code{\F} escape, or the @code{family} command in the @file{DESC} file).

It is not possible to switch to a font with the name @samp{DESC}
(whereas this name could be used as a style name; however, this is not
recommended).

@cindex previous font (@code{ft}, @code{\f[]}, @code{\fP})
@cindex font, previous (@code{ft}, @code{\f[]}, @code{\fP})
With no argument or using @samp{P} as an argument, @code{ft} switches
to the previous font.  Use @code{\f[]} to do this with an escape
sequence.  The old syntax forms @code{\fP} or @code{\f[P]} are also
supported.

Fonts are generally specified as uppercase strings, which are usually
1@tie{}to 4 characters representing an abbreviation or acronym of the
font name.  This is no limitation, just a convention.

The example below produces two identical lines.

@Example
eggs, bacon,
.ft B
spam
.ft
and sausage.

eggs, bacon, \fBspam\fP and sausage.
@endExample

@code{\f} doesn't produce an input token in GNU @code{troff}.  As a
consequence, it can be used in requests like @code{mc} (which expects a
single character as an argument) to change the font on the fly:

@Example
.mc \f[I]x\f[]
@endExample

The current style name is available in the read-only string-valued
register @samp{.sty}; it is associated with the environment
(@pxref{Environments}).  If the current font isn't a style,
interpolating @samp{.sty} produces nothing.

@xref{Font Positions}, for an alternative syntax.
@endDefreq

@Defreq {ftr, f [@Var{g}]}
@cindex font translation (@code{ftr})
@cindex @code{ft} request, and font translations
@cindex @code{ul} request, and font translations
@cindex @code{bd} request, and font translations
@cindex @code{\f}, and font translations
@cindex @code{cs} request, and font translations
@cindex @code{tkf} request, and font translations
@cindex @code{special} request, and font translations
@cindex @code{fspecial} request, and font translations
@cindex @code{fp} request, and font translations
@cindex @code{sty} request, and font translations
@cindex @code{if} request, and font translations
@cindex @code{ie} request, and font translations
@cindex @code{while} request, and font translations
Translate font@tie{}@var{f} to font@tie{}@var{g}.  Whenever a font
named@tie{}@var{f} is referred to in a @code{\f} escape sequence, in the
@code{F} and @code{S} conditional operators, or in the @code{ft},
@code{ul}, @code{bd}, @code{cs}, @code{tkf}, @code{special},
@code{fspecial}, @code{fp}, or @code{sty} requests, font@tie{}@var{g} is
used.  If @var{g} is missing or equal to@tie{}@var{f} the translation is
undone.

Font translations cannot be chained.

@Example
.ftr XXX TR
.ftr XXX YYY
.ft XXX
    @error{} warning: can't find font 'XXX'
@endExample
@endDefreq

@DefreqList {fzoom, f [@Var{zoom}]}
@DefregListEndx {.zoom}
@cindex magnification of a font (@code{fzoom})
@cindex font, magnification (@code{fzoom})
@cindex zoom factor of a font (@code{fzoom})
@cindex factor, zoom, of a font (@code{fzoom})
@cindex font, zoom factor (@code{fzoom})
@cindex optical size of a font
@cindex font, optical size
@cindex size, optical, of a font
Set magnification of font@tie{}@var{f} to factor @var{zoom}, which must
be a non-negative integer multiple of 1/1000th.  This request is useful
to adjust the optical size of a font in relation to the others.  In the
example below, font @code{CR} is magnified by 10% (the zoom factor is
thus 1.1).

@Example
.fam P
.fzoom CR 1100
.ps 12
Palatino and \f[CR]Courier\f[]
@endExample

A missing or zero value of @var{zoom} is the same as a value of 1000,
which means no magnification.  @var{f}@tie{}must be a real font name,
not a style.

The magnification of a font is completely transparent to GNU
@code{troff}; a change of the zoom factor doesn't cause any effect
except that the dimensions of glyphs, (word) spaces, kerns, etc., of the
affected font are adjusted accordingly.

The zoom factor of the current font is available in the read-only
register @samp{.zoom}, in multiples of 1/1000th.  It returns zero if
there is no magnification.
@endDefreq

@c ---------------------------------------------------------------------

@node Font Families, Font Positions, Changing Fonts, Fonts and Symbols
@subsection Font Families
@cindex font families
@cindex families, font
@cindex font styles
@cindex styles, font

Due to the variety of fonts available, @code{gtroff} has added the
concept of @dfn{font families} and @dfn{font styles}.  The fonts are
specified as the concatenation of the font family and style.  Specifying
a font without the family part causes @code{gtroff} to use that style of
the current family.

@cindex PostScript fonts
@cindex fonts, PostScript
Currently, fonts for the devices @option{-Tps}, @option{-Tpdf},
@option{-Tdvi}, @option{-Tlj4}, @option{-Tlbp}, and the X11 fonts are
set up to this mechanism.  By default, @code{gtroff} uses the Times
family with the four styles @samp{R}, @samp{I}, @samp{B}, and @samp{BI}.

This way, it is possible to use the basic four fonts and to select a
different font family on the command line (@pxref{Groff Options}).

@DefreqList {fam, [@Var{family}]}
@DefregItemx {.fam}
@DefescItemx {\\F, , f, }
@DefescItem {\\F, (, fm, }
@DefescItem {\\F, [, family, ]}
@DefregListEndx {.fn}
@cindex changing font family (@code{fam}, @code{\F})
@cindex font family, changing (@code{fam}, @code{\F})
Set the font family to @var{family} (one-character name@tie{}@var{f},
two-character name @var{fm}).  If no argument is given, switch to the
previous font family, or the default family if there is none.  Use
@samp{\F[]} to do this with an escape sequence; @samp{\FP} selects font
family @samp{P} instead.  The initial font family is @samp{T} (Times),
but can be overridden by the output device description file---@xref{DESC
File Format}.  The current font family is available in the read-only
string-valued register @code{.fam}; it is associated with the
environment (@pxref{Environments}).

@Example
spam,
.fam H    \" helvetica family
spam,     \" used font is family H + style R = HR
.ft B     \" family H + style B = font HB
spam,
.fam T    \" times family
spam,     \" used font is family T + style B = TB
.ft AR    \" font AR (not a style)
baked beans,
.ft R     \" family T + style R = font TR
and spam.
@endExample

@code{\F} doesn't produce an input token in GNU @code{troff}.  As a
consequence, it can be used in requests like @code{mc} (which expects
a single character as an argument) to change the font family on the fly.

@Example
.mc \F[P]x\F[]
@endExample

The read-only string-valued register @code{.fn} contains the current
@dfn{real font name} of the current font.  If the current font is a
style, the value of @code{\n[.fn]} is the proper concatenation of family
and style name.
@endDefreq

@Defreq {sty, n style}
@cindex changing font style (@code{sty})
@cindex font style, changing (@code{sty})
@cindex @code{cs} request, and font styles
@cindex @code{bd} request, and font styles
@cindex @code{tkf} request, and font styles
@cindex @code{uf} request, and font styles
@cindex @code{fspecial} request, and font styles
Associate @var{style} with font position@tie{}@var{n}.  A font position
can be associated either with a font or with a style.  The current font
is the index of a font position and so is also either a font or a style.
If it is a style, the font that is actually used is the font whose name
is the concatenation of the name of the current family and the name of
the current style.  For example, if the current font is@tie{}1 and font
position@tie{}1 is associated with style @samp{R} and the current font
family is @samp{T}, then font @samp{TR} is used.  If the current font is
not a style, then the current family is ignored.  If the requests
@code{cs}, @code{bd}, @code{tkf}, @code{uf}, or @code{fspecial} are
applied to a style, they are instead applied to the member of the
current family corresponding to that style.

@var{n}@tie{}must be a non-negative integer.

@pindex DESC
@kindex styles
The default family can be set with the @option{-f} option (@pxref{Groff
Options}).  The @code{styles} command in the @file{DESC} file controls
which font positions (if any) are initially associated with styles
rather than fonts.  For example, the default setting for PostScript
fonts

@Example
styles R I B BI
@endExample

@noindent
is equivalent to

@Example
.sty 1 R
.sty 2 I
.sty 3 B
.sty 4 BI
@endExample

@code{fam} and @code{\F} always check whether the current font position
is valid; this can give surprising results if the current font position
is associated with a style.

In the following example, we want to access the PostScript font
@code{FooBar} from the font family @code{Foo}:

@Example
.sty \n[.fp] Bar
.fam Foo
    @error{} warning: can't find font 'FooR'
@endExample

@noindent
The default font position at startup is@tie{}1; for the PostScript
device, this is associated with style @samp{R}, so @code{gtroff} tries
to open @code{FooR}.

A solution to this problem is to use a dummy font like the following:

@Example
.fp 0 dummy TR    \" set up dummy font at position 0
.sty \n[.fp] Bar  \" register style `Bar'
.ft 0             \" switch to font at position 0
.fam Foo          \" activate family `Foo'
.ft Bar           \" switch to font `FooBar'
@endExample

@xref{Font Positions}.
@endDefreq

@c ---------------------------------------------------------------------

@codequotebacktick on
@codequoteundirected on

@node Font Positions, Using Symbols, Font Families, Fonts and Symbols
@subsection Font Positions
@cindex font positions
@cindex positions, font

To support typeface indirection through styles, and for compatibility
with @acronym{AT&T} @code{troff}, the formatter maintains a list of font
@dfn{positions} at which fonts required by a document are @dfn{mounted}.
An output device's description file @file{DESC} typically configures a
set of pre-mounted fonts; see @ref{Device and Font Description Files}.
A font need not be explicitly mounted before it is selected; GNU
@code{troff} will search @env{GROFF_FONT_PATH} for it by name and mount
it at the first free mounting position on demand.

@DefreqList {fp, pos font [@Var{external-name}]}
@DefregItemx {.f}
@DefregListEndx {.fp}
@cindex mounting font (@code{fp})
@cindex font, mounting (@code{fp})
Mount the font named @var{font} at position @var{pos}, a non-negative
integer.  This numeric index can then be referred to with font selection
instructions.  When the formatter starts up, it reads the output
device's description to mount an initial set of faces, and selects font
position@tie{}1.  Position@tie{}0 is unused by default.  Unless the
@var{external-name} argument is given, the name @var{font} should be
that of a font description file stored in a directory corresponding to
the selected output device.

@cindex font position register (@code{.f})
The position of the currently selected font (or style) is available in
the read-only register @samp{.f}.  It is associated with the environment
(@pxref{Environments}).

The value of @code{.f} can be copied to another register to save the
current font or style for later recall.

@Example
.nr saved-font \n[.f]
.ft B
@r{@dots{}} text text text @r{@dots{}}
.ft \n[saved-font]
@endExample

@cindex next free font position register (@code{.fp})
The index of the next (nonzero) free font position is available in the
read-only register @samp{.fp}.
@cindex @file{DESC} file, and font mounting
Fonts not listed in the @file{DESC} file are automatically mounted at
position @samp{\n[.fp]} when selected.  When mounting a font at a
position explicitly with the @code{fp} request, this same practice
should be followed, although GNU @code{troff} does not enforce this
strictly.

If there is no third argument, the font description is sought in the
file system under the name @var{font}.

@c The third argument was a late revision to device-independent troff.
@c It wasn't in the "Unix 4.0" version of CSTR #54 (January 1981), which
@c featured Kernighan's device-independent rewrite, but appeared by the
@c time of its 1992 revision.
The optional third argument @var{external-name} is the name of a font
description file in the operating environment.  GNU @code{troff} refuses
to traverse directories to locate the font description.  The second
argument @var{font} is then the internal name of the font, which
identifies it to GNU @code{troff} once it has been mounted.  This
feature enables font names to be aliased, which can be necessary in
compatibility mode since AT&T @code{troff} syntax affords no means of
identifying fonts with names longer than two characters, like @samp{TBI}
or @samp{ZCMI}, in a font selection escape sequence.
@xref{Compatibility Mode}.  You can also alias fonts on mounting for
convenience or abstraction.

@Example
.ft SC ZCMI \" "script" font
Send a \f(SChand-written\fP thank-you note.
.ft Emph I
.ft Strong B
Are \f[Emph]these names\f[] \f[Strong]comfortable\f[]?
@endExample
@endDefreq

The @code{ft} request and @code{\f} escape sequence accept mounting
positions in the place of font or style names.

@DefreqList {ft, nnn}
@DefescItemx {\\f, , n, }
@DefescItem {\\f, (, nn, }
@DefescListEnd {\\f, [, nnn, ]}
@cindex changing font position (@code{\f})
@cindex font position, changing (@code{\f})
@cindex @code{sty} request, and font positions
@cindex @code{fam} request, and font positions
@cindex @code{\F}, and font positions
@kindex styles
@kindex family
@pindex DESC
Select font position @var{nnn} (one-digit position@tie{}@var{n},
two-digit position @var{nn}), which must be a non-negative integer.  If
@var{nnn} is associated with a style (as set with the @code{sty} request
or with the @code{styles} command in the @file{DESC} file), use it
within the current font family (as set with the @code{fam} request, the
@code{\F} escape sequence, or the @code{family} command in the
@file{DESC} file).

@Example
this is font 1
.ft 2
this is font 2
.ft                   \" switch back to font 1
.ft 3
this is font 3
.ft
this is font 1 again
@endExample

@xref{Changing Fonts}, for font selection by name.
@endDefreq

@c ---------------------------------------------------------------------

@node Using Symbols, Character Classes, Font Positions, Fonts and Symbols
@subsection Using Symbols
@cindex using symbols
@cindex symbols, using

@cindex glyph
@cindex character
@cindex glyph, distinguished from character
@cindex character, distinguished from glyph
@cindex ligature
A @dfn{glyph} is a graphical representation of a @dfn{character}.  While
a character is an abstraction of semantic information, a glyph is
something that can be seen on screen or paper.  A character has many
possible representation forms (for example, the character `A' can be
written in an upright or slanted typeface, producing distinct
glyphs).  Sometimes, a sequence of characters map to a single glyph:@:
this is a @dfn{ligature}---the most common is `fi'.

Space characters never become glyphs in GNU @code{troff}.  If not
discarded (as when trailing on text lines), they are represented by
horizontal motions in the output.

@cindex symbol
@cindex special fonts
@kindex fonts
@pindex DESC
@cindex @code{special} request, and glyph search order
@cindex @code{fspecial} request, and glyph search order
A @dfn{symbol} is simply a named glyph.  Within @code{gtroff}, all glyph
names of a particular font are defined in its font file.  If the user
requests a glyph not available in this font, @code{gtroff} looks up an
ordered list of @dfn{special fonts}.  By default, the PostScript output
device supports the two special fonts @samp{SS} (slanted symbols) and
@samp{S} (symbols) (the former is looked up before the latter).  Other
output devices use different names for special fonts.  Fonts mounted
with the @code{fonts} keyword in the @file{DESC} file are globally
available.  To install additional special fonts locally (i.e., for a
particular font), use the @code{fspecial} request.

Here are the exact rules how @code{gtroff} searches a given symbol:

@itemize @bullet
@item
If the symbol has been defined with the @code{char} request, use it.
This hides a symbol with the same name in the current font.

@item
Check the current font.

@item
If the symbol has been defined with the @code{fchar} request, use it.

@item
Check whether the current font has a font-specific list of special
fonts; test all fonts in the order of appearance in the last
@code{fspecial} call if appropriate.

@item
If the symbol has been defined with the @code{fschar} request for the
current font, use it.

@item
Check all fonts in the order of appearance in the last @code{special}
call.

@item
If the symbol has been defined with the @code{schar} request, use it.

@item
As a last resort, consult all fonts loaded up to now for special fonts
and check them, starting with the lowest font number.  This can
sometimes lead to surprising results since the @code{fonts} line in
the @file{DESC} file often contains empty positions, which are filled
later on.  For example, consider the following:

@Example
fonts 3 0 0 FOO
@endExample

@noindent
This mounts font @code{foo} at font position@tie{}3.  We assume that
@code{FOO} is a special font, containing glyph @code{foo}, and that no
font has been loaded yet.  The line

@Example
.fspecial BAR BAZ
@endExample

@noindent
makes font @code{BAZ} special only if font @code{BAR} is active.  We
further assume that @code{BAZ} is really a special font, i.e., the font
description file contains the @code{special} keyword, and that it also
contains glyph @code{foo} with a special shape fitting to font
@code{BAR}.  After executing @code{fspecial}, font @code{BAR} is loaded
at font position@tie{}1, and @code{BAZ} at position@tie{}2.

We now switch to a new font @code{XXX}, trying to access glyph
@code{foo} that is assumed to be missing.  There are neither
font-specific special fonts for @code{XXX} nor any other fonts made
special with the @code{special} request, so @code{gtroff} starts the
search for special fonts in the list of already mounted fonts, with
increasing font positions.  Consequently, it finds @code{BAZ} before
@code{FOO} even for @code{XXX}, which is not the intended behaviour.
@end itemize

@xref{Device and Font Description Files}, and @ref{Special Fonts}, for
more details.

@cindex list of available glyphs (@cite{groff_char@r{(7)}} man page)
@cindex available glyphs, list (@cite{groff_char@r{(7)}} man page)
@cindex glyphs, available, list (@cite{groff_char@r{(7)}} man page)
The list of available symbols is device dependent; see the
@cite{groff_char@r{(7)}} man page for a complete list of all glyphs.
For example, say

@Example
man -Tdvi groff_char > groff_char.dvi
@endExample

@noindent
for a list using the default DVI fonts (not all versions of the
@code{man} program support the @option{-T} option).  If you want to use
an additional macro package to change the used fonts, @code{groff} must
be called directly:

@Example
groff -Tdvi -mec -man groff_char.7 > groff_char.dvi
@endExample

@cindex composite glyph names
@cindex glyph names, composite
@cindex @code{groff} glyph list (GGL)
@cindex GGL (@code{groff} glyph list)
@cindex Adobe Glyph List (AGL)
Glyph names not listed in @cite{groff_char@r{(7)}} are derived
algorithmically, using a simplified version of the Adobe Glyph List
(AGL) algorithm, which is described in
@uref{https://github.com@//adobe-type-tools@//agl-aglfn}.  The (frozen)
set of glyph names that can't be derived algorithmically is called the
@dfn{@code{groff} glyph list (GGL)}.

@itemize @bullet
@item
A glyph for Unicode character U+@var{XXXX}[@var{X}[@var{X}]], which is
not a composite character is named
@code{u@var{XXXX}@r{[}@var{X}@r{[}@var{X}@r{]]}}.  @var{X} must be an
uppercase hexadecimal digit.  Examples: @code{u1234}, @code{u008E},
@code{u12DB8}.  The largest Unicode value is 0x10FFFF.  There must be at
least four @code{X} digits; if necessary, add leading zeroes (after the
@samp{u}).  No zero padding is allowed for character codes greater than
0xFFFF.  Surrogates (i.e., Unicode values greater than 0xFFFF
represented with character codes from the surrogate area U+D800-U+DFFF)
are not allowed either.

@item
A glyph representing more than a single input character is named

@display
@samp{u} @var{component1} @samp{_} @var{component2} @samp{_} @var{component3} @dots{}
@end display

@noindent
Example: @code{u0045_0302_0301}.

For simplicity, all Unicode characters that are composites must be
maximally decomposed to NFD;@footnote{This is ``Normalization Form D''
as documented in Unicode Standard Annex #15
(@uref{https://unicode.org@//reports@//tr15/}).} for example,
@code{u00CA_0301} is not a valid glyph name since U+00CA (@sc{latin
capital letter e with circumflex}) can be further decomposed into U+0045
(@sc{latin capital letter e}) and U+0302 (@sc{combining circumflex
accent}).  @code{u0045_0302_0301} is thus the glyph name for U+1EBE,
@sc{latin capital letter e with circumflex and acute}.

@item
groff maintains a table to decompose all algorithmically derived glyph
names that are composites itself.  For example, @code{u0100} (@sc{latin
letter a with macron}) is automatically decomposed into
@code{u0041_0304}.  Additionally, a glyph name of the GGL is preferred
to an algorithmically derived glyph name; @code{groff} also
automatically does the mapping.  Example: The glyph @code{u0045_0302} is
mapped to @code{^E}.

@item
glyph names of the GGL can't be used in composite glyph names; for
example, @code{^E_u0301} is invalid.
@end itemize

@DefescList {\\, (, nm, }
@DefescItem {\\, [, name, ]}
@DefescListEnd {\\, [, base-glyph combining-component @dots{}, ]}
@esindex \(
@esindex \[
Typeset a special character @var{name} (two-character name @var{nm}) or
a composite glyph consisting of @var{base-glyph} overlaid with one or
more @var{combining-component}s.  For example, @samp{\[A ho]} is a
capital letter ``A'' with a ``hook accent'' (ogonek).

There is no special syntax for one-character names---the analogous form
@samp{\@var{n}} would collide with other escape sequences.  However, the
four escape sequences @code{\'}, @code{\-}, @code{\_}, and @code{\`},
are translated on input to the special character escape sequences
@code{\[aa]}, @code{\[-]}, @code{\[ul]}, and @code{\[ga]}, respectively.

A special character name of length one is not the same thing as an
ordinary character: that is, the character @code{a} is not the same as
@code{\[a]}.

If @var{name} is undefined, a warning in category @samp{char} is
produced and the escape is ignored.  @xref{Warnings}, for information
about the enablement and suppression of warnings.

GNU @code{troff} resolves @code{\[@r{@dots{}}]} with more than a single
component as follows:

@itemize @bullet
@item
Any component that is found in the GGL is converted to the
@code{u@var{XXXX}} form.

@item
Any component @code{u@var{XXXX}} that is found in the list of
decomposable glyphs is decomposed.

@item
The resulting elements are then concatenated with @samp{_} in between,
dropping the leading @samp{u} in all elements but the first.
@end itemize

No check for the existence of any component (similar to @code{tr}
request) is done.

Examples:

@table @code
@item \[A ho]
@samp{A} maps to @code{u0041}, @samp{ho} maps to @code{u02DB}, thus the
final glyph name would be @code{u0041_02DB}.  Note this is not the
expected result: The ogonek glyph @samp{ho} is a spacing ogonek, but for
a proper composite a non-spacing ogonek (U+0328) is necessary.  Looking
into the file @file{composite.tmac} one can find @w{@samp{.composite ho
u0328}}, which changes the mapping of @samp{ho} while a composite glyph
name is constructed, causing the final glyph name to be
@code{u0041_0328}.

@item \[^E u0301]
@itemx \[^E aa]
@itemx \[E a^ aa]
@itemx \[E ^ @code{'}]
@samp{^E} maps to @code{u0045_0302}, thus the final glyph name is
@code{u0045_0302_0301} in all forms (assuming proper calls of the
@code{composite} request).
@end table

It is not possible to define glyphs with names like @w{@samp{A ho}}
within a @code{groff} font file.  This is not really a limitation;
instead, you have to define @code{u0041_0328}.
@endDefesc

@Defesc {\\C, @code{'}, xxx, @code{'}}
@cindex named character (@code{\C})
@cindex character, named (@code{\C})
Typeset the glyph named @var{xxx}.@footnote{@code{\C} is actually a
misnomer since it accesses an output glyph.}  Normally it is more
convenient to use @code{\[@var{xxx}]}, but @code{\C} has the advantage
that it is compatible with newer versions of @acronym{AT&T} @code{troff}
and is available in compatibility mode.
@endDefesc

@Defreq {composite, from to}
@pindex composite.tmac
Map glyph name @var{from} to glyph name @var{to} if it is used in
@code{\[...]} with more than one component.  See above for examples.

This mapping is based on glyph names only; no check for the existence of
either glyph is done.

A set of default mappings for many accents can be found in the file
@file{composite.tmac}, which is loaded at startup.
@endDefreq

@Defesc {\\N, @code{'}, n, @code{'}}
@cindex numbered glyph (@code{\N})
@cindex glyph, numbered (@code{\N})
@cindex @code{char} request, used with @code{\N}
@cindex Unicode
Typeset the glyph with code@tie{}@var{n} in the current font
(@code{n}@tie{}is @emph{not} the input character code).  The number
@var{n}@tie{}can be any non-negative decimal integer.  Most devices only
have glyphs with codes between 0 and@tie{}255; the Unicode output device
uses codes in the range 0--65535.  If the current font does not contain
a glyph with that code, special fonts are @emph{not} searched.  The
@code{\N} escape sequence can be conveniently used in conjunction with
the @code{char} request:

@Example
.char \[phone] \f[ZD]\N'37'
@endExample

@noindent
@pindex DESC
@cindex unnamed glyphs
@cindex glyphs, unnamed
The code of each glyph is given in the fourth column in the font
description file after the @code{charset} command.  It is possible to
include unnamed glyphs in the font description file by using a name of
@samp{---}; the @code{\N} escape sequence is the only way to use these.

No kerning is applied to glyphs accessed with @code{\N}.
@endDefesc

Some escape sequences directly map onto special glyphs.

@Defesc {\@code{'}, , , }
A backslash followed by the apostrophe character,
@acronym{ASCII} character @code{0x27} (@acronym{EBCDIC} character
@code{0x7D}), is a synonym for @code{\[aa]}, the acute accent.
@endDefesc

@Defesc {\@code{`}, , , }
A backslash followed by @acronym{ASCII} character @code{0x60}
(@acronym{EBCDIC} character @code{0x79} [usually]), is a synonym for
@code{\[ga]}, the grave accent.
@endDefesc

@Defesc {\\-, , , }
A backslash followed by a dash is a synonym for @code{\[-]}, the minus
sign.
@endDefesc

@Defesc {\\_, , , }
A backslash followed by an underscore is a synonym for @code{\[ul]}, the
underrule.  On @code{troff}-mode (typesetter) devices it may be
font-invariant and drawn lower than the underscore glyph @samp{_}.
@endDefesc

@Defreq {cflags, n c1 c2 @dots{}}
@cindex glyph properties (@code{cflags})
@cindex character properties (@code{cflags})
@cindex properties of glyphs (@code{cflags})
@cindex properties of characters (@code{cflags})
Assign properties encoded by the number @var{n} to characters @var{c1},
@var{c2}, and so on.

Input characters, including special characters introduced by an escape,
have certain properties associated with them.@footnote{Output glyphs
don't have such properties.  For GNU @code{troff}, a glyph is a
box numbered with an index into a font, a given height above and depth
below the baseline, and a width---nothing more.}  These properties can
be modified with this request.  The first argument is the sum of the
desired flags and the remaining arguments are the characters to be
assigned those properties.  Spaces between the @var{cn} arguments are
optional.  Any argument @var{cn} can be a character class defined with
the @code{class} request rather than an individual character.
@xref{Character Classes}.

The non-negative integer @var{n} is the sum of any of the following.
Some combinations are nonsensical, such as @samp{33} (1 + 32).

@table @code
@item 1
@cindex end-of-sentence characters
@cindex characters, end-of-sentence
Recognize the character as ending a sentence if followed by a newline
or two spaces.  Initially, characters @samp{.?!} have this property.

@item 2
@cindex hyphenating characters
@cindex characters, hyphenation
Enable breaks before the character.  A line is not broken at a character
with this property unless the characters on each side both have non-zero
hyphenation codes.  This exception can be overridden by adding 64.
Initially, no characters have this property.

@item 4
@cindex @code{\-} glyph, and @code{cflags}
@cindex @code{hy} glyph, and @code{cflags}
@cindex @code{em} glyph, and @code{cflags}
Enable breaks after the character.  A line is not broken at a character
with this property unless the characters on each side both have non-zero
hyphenation codes.  This exception can be overridden by adding 64.
Initially, characters @samp{\-\[hy]\[em]} have this property.

@item 8
@cindex overlapping characters
@cindex characters, overlapping
@cindex @code{ul} glyph, and @code{cflags}
@cindex @code{rn} glyph, and @code{cflags}
@cindex @code{ru} glyph, and @code{cflags}
@cindex @code{radicalex} glyph, and @code{cflags}
@cindex @code{sqrtex} glyph, and @code{cflags}
Mark the glyph associated with this character as overlapping other
instances of itself horizontally.  Initially, characters
@samp{\[ul]\[rn]\[ru]\[radicalex]\[sqrtex]} have this property.

@item 16
@cindex @code{br} glyph, and @code{cflags}
Mark the glyph associated with this character as overlapping other
instances of itself vertically.  Initially, the character @samp{\[br]}
has this property.

@item 32
@cindex transparent characters
@cindex character, transparent
@cindex @code{"}, at end of sentence
@cindex @code{'}, at end of sentence
@cindex @code{)}, at end of sentence
@cindex @code{]}, at end of sentence
@cindex @code{*}, at end of sentence
@cindex @code{dg} glyph, at end of sentence
@cindex @code{dd} glyph, at end of sentence
@cindex @code{rq} glyph, at end of sentence
@cindex @code{cq} glyph, at end of sentence
Mark the character as transparent for the purpose of end-of-sentence
recognition.  In other words, an end-of-sentence character followed by
any number of characters with this property is treated as the end of a
sentence if followed by a newline or two spaces.  This is the same as
having a zero space factor in @TeX{}.  Initially, characters
@samp{"')]*\[dg]\[dd]\[rq]\[cq]} have this property.

@item 64
Ignore hyphenation codes of the surrounding characters.  Use this in
combination with values 2 and@tie{}4 (initially, no characters have this
property).

For example, if you need an automatic break point after the en-dash in
numeric ranges like ``3000--5000'', insert

@Example
.cflags 68 \[en]
@endExample

@noindent
into your document.  Note, however, that this can lead to bad layout if
done without thinking; in most situations, a better solution instead of
changing the @code{cflags} value is to insert @code{\:} right after the
hyphen at the places that really need a break point.
@end table

The remaining values were implemented for East Asian language support;
those who use alphabetic scripts exclusively can disregard them.

@table @code
@item 128
Prohibit a line break before the character, but allow a line break after
the character.  This works only in combination with flags 256 and 512
and has no effect otherwise.  Initially, no characters have this
property.

@item 256
Prohibit a line break after the character, but allow a line break before
the character.  This works only in combination with flags 128 and 512
and has no effect otherwise.  Initially, no characters have this
property.

@item 512
Allow line break before or after the character.  This works only in
combination with flags 128 and 256 and has no effect otherwise.
Initially, no characters have this property.
@end table

In contrast to values 2 and@tie{}4, the values 128, 256, and 512 work
pairwise.  If, for example, the left character has value 512, and the
right character 128, no break will be automatically inserted between
them.  If we use value@tie{}6 instead for the left character, a break
after the character can't be suppressed since the neighboring character
on the right doesn't get examined.
@endDefreq

@DefreqList {char, c [@Var{contents}]}
@DefreqItemx {fchar, c [@Var{contents}]}
@DefreqItemx {fschar, f c [@Var{contents}]}
@DefreqListEndx {schar, c [@Var{contents}]}
@cindex defining character (@code{char})
@cindex defining fallback character (@code{fchar}, @code{fschar}, @code{schar})
@cindex character, defining (@code{char})
@cindex character, defining fallback (@code{fchar}, @code{fschar}, @code{schar})
@cindex fallback character, defining (@code{fchar}, @code{fschar}, @code{schar})
@cindex creating new characters (@code{char})
@cindex defining symbol (@code{char})
@cindex symbol, defining (@code{char})
@cindex defining glyph (@code{char})
@cindex glyph, defining (@code{char})
@cindex escape character, while defining glyph
@cindex character, escape, while defining glyph
@cindex @code{tr} request, and glyph definitions
@cindex @code{cp} request, and glyph definitions
@cindex @code{rc} request, and glyph definitions
@cindex @code{lc} request, and glyph definitions
@cindex @code{\l}, and glyph definitions
@cindex @code{\L}, and glyph definitions
@cindex @code{\&}, and glyph definitions
@cindex @code{\e}, and glyph definitions
@cindex @code{hcode} request, and glyph definitions
Define a new character or glyph@tie{}@var{c} to be @var{contents}, which
can be empty.  More precisely, @code{char} defines a @code{groff} object
(or redefines an existing one) that is accessed with the
name@tie{}@var{c} on input, and produces @var{contents} on output.
Every time glyph@tie{}@var{c} needs to be printed, @var{contents} is
processed in a temporary environment and the result is wrapped up into a
single object.  Compatibility mode is turned off and the escape
character is set to@tie{}@code{\} while @var{contents} is processed.
Any emboldening, constant spacing, or track kerning is applied to this
object rather than to individual glyphs in @var{contents}.

An object defined by these requests can be used just like a normal glyph
provided by the output device.  In particular, other characters can be
translated to it with the @code{tr} or @code{trin} requests; it can be
made the leader character with the @code{lc} request; repeated patterns
can be drawn with it using the @code{\l} and @code{\L} escape sequences;
and words containing@tie{}@var{c} can be hyphenated correctly if the
@code{hcode} request is used to give the object a hyphenation code.

There is a special anti-recursion feature: use of the object within its
own definition is handled like a normal character (not
defined with @code{char}).

The @code{tr} and @code{trin} requests take precedence if @code{char}
accesses the same symbol.

@Example
.tr XY
X
    @result{} Y
.char X Z
X
    @result{} Y
.tr XX
X
    @result{} Z
@endExample

The @code{fchar} request defines a fallback glyph: @code{gtroff} only
checks for glyphs defined with @code{fchar} if it cannot find the glyph
in the current font.  @code{gtroff} carries out this test before
checking special fonts.

@code{fschar} defines a fallback glyph for font@tie{}@var{f}:
@code{gtroff} checks for glyphs defined with @code{fschar} after the
list of fonts declared as font-specific special fonts with the
@code{fspecial} request, but before the list of fonts declared as global
special fonts with the @code{special} request.

Finally, the @code{schar} request defines a global fallback glyph:
@code{gtroff} checks for glyphs defined with @code{schar} after the list
of fonts declared as global special fonts with the @code{special}
request, but before the already mounted special fonts.

@xref{Character Classes}.
@endDefreq

@DefreqList {rchar, c @dots{}}
@DefreqListEndx {rfschar, f c @dots{}}
@cindex removing glyph definition (@code{rchar}, @code{rfschar})
@cindex glyph, removing definition (@code{rchar}, @code{rfschar})
@cindex fallback glyph, removing definition (@code{rchar}, @code{rfschar})
Remove definition of each ordinary or special character @var{c},
undoing the effect of a @code{char}, @code{fchar}, or @code{schar}
request.  Those supplied by font description files cannot be removed.
Spaces and tabs may separate @var{c}@tie{}arguments.

The request @code{rfschar} removes glyph definitions defined with
@code{fschar} for font@tie{}@var{f}.
@endDefreq

@c ---------------------------------------------------------------------

@node Character Classes, Special Fonts, Using Symbols, Fonts and Symbols
@subsection Character Classes
@cindex character classes
@cindex classes, character

Classes are particularly useful for East Asian languages such as
Chinese, Japanese, and Korean, where the number of needed characters is
much larger than in European languages, and where large sets of
characters share the same properties.

@Defreq {class, name c1 c2 @dots{}}
@cindex character class (@code{class})
@cindex defining character class (@code{class})
@cindex class of characters (@code{class})
Define a character class (or simply ``class'') @var{name} comprising
the characters @var{c1}, @var{c2}, and so on.

A class thus defined can then be referred to in lieu of listing all the
characters within it.  Currently, only the @code{cflags} request can
handle references to character classes.

In the request's simplest form, each @var{cn} is a character (or special
character).

@Example
.class [quotes] ' \[aq] \[dq] \[oq] \[cq] \[lq] \[rq]
@endExample

Since class and glyph names share the same name space, it is recommended
to start and end the class name with @code{[} and @code{]},
respectively, to avoid collisions with existing character names defined
by GNU @code{troff} or the user (with @code{char} and related requests).
This practice applies the presence of @code{]} in the class name to
prevent the use of the special character escape form
@code{\[@r{@dots{}}]}, thus you must use the @code{\C} escape to access
a class with such a name.

@cindex GGL (@code{groff} glyph list)
@cindex @code{groff} glyph list (GGL)
You can also use a character range notation consisting of a
start character followed by @samp{-} and then an end character.
Internally, GNU @code{troff} converts these two symbol names to
Unicode code points (according to the @code{groff} glyph list [GGL]),
which then give the start and end value of the range.  If that fails,
the class definition is skipped.

Furthermore, classes can be nested.

@Example
.class [prepunct] , : ; > @}
.class [prepunctx] \C'[prepunct]' \[u2013]-\[u2016]
@endExample

@noindent
The class @samp{[prepunctx]} thus contains the contents of the class
@code{[prepunct]} as defined above (the set @samp{, : ; > @}}), and
characters in the range between @code{U+2013} and @code{U+2016}.

If you want to include @samp{-} in a class, it must be the first
character value in the argument list, otherwise it gets misinterpreted
as part of the range syntax.

It is not possible to use class names as end points of range
definitions.

A typical use of the @code{class} request is to control line-breaking
and hyphenation rules as defined by the @code{cflags} request.  For
example, to inhibit line breaks before the characters belonging to the
@code{prepunctx} class defined in the previous example, you can write
the following.

@Example
.cflags 2 \C'[prepunctx]'
@endExample

@noindent
See the @code{cflags} request in @ref{Using Symbols}, for more details.
@endDefreq

@c ---------------------------------------------------------------------

@node Special Fonts, Artificial Fonts, Character Classes, Fonts and Symbols
@subsection Special Fonts
@cindex special fonts
@cindex fonts, special

Special fonts are those that @code{gtroff} searches when it cannot find
the requested glyph in the current font.  The Symbol font is usually a
special font.

@code{gtroff} provides the following two requests to add more special
fonts.  @xref{Using Symbols}, for a detailed description of the glyph
searching mechanism in @code{gtroff}.

Usually, only non-TTY devices have special fonts.

@DefreqList {special, [@Var{s1} @Var{s2} @dots{}]}
@DefreqListEndx {fspecial, f [@Var{s1} @Var{s2} @dots{}]}
@kindex fonts
@pindex DESC
Use the @code{special} request to define special fonts.  Initially, this
list is empty.

Use the @code{fspecial} request to designate special fonts only when
font@tie{}@var{f} is active.  Initially, this list is empty.

Previous calls to @code{special} or @code{fspecial} are overwritten;
without arguments, the particular list of special fonts is set to empty.
Special fonts are searched in the order they appear as arguments.

All fonts that appear in a call to @code{special} or @code{fspecial}
are loaded.

@xref{Using Symbols}, for the exact search order of glyphs.
@endDefreq

@c ---------------------------------------------------------------------

@node Artificial Fonts, Ligatures and Kerning, Special Fonts, Fonts and Symbols
@subsection Artificial Fonts
@cindex artificial fonts
@cindex fonts, artificial

There are a number of requests and escape sequences for artificially
creating fonts.  These are largely vestiges of the days when output
devices did not have a wide variety of fonts, and when @code{nroff} and
@code{troff} were separate programs.  Most of them are no longer
necessary in GNU @code{troff}.  Nevertheless, they are supported.

@DefescList {\\H, @code{'}, height, @code{'}}
@DefescItem {\\H, @code{'}, @t{+}height, @code{'}}
@DefescItem {\\H, @code{'}, @t{-}height, @code{'}}
@DefregListEndx {.height}
@cindex changing the font height (@code{\H})
@cindex font height, changing (@code{\H})
@cindex height, font, changing (@code{\H})
Change (increment, decrement) the height of the current font, but not
the width.  If @var{height} is zero, restore the original height.
Default scaling indicator is @samp{z}.

The read-only register @code{.height} contains the font height as set by
@code{\H}.

Currently, only the @option{-Tps} and @option{-Tpdf} devices support
this feature.

@code{\H} doesn't produce an input token in GNU @code{troff}.  As a
consequence, it can be used in requests like @code{mc} (which expects
a single character as an argument) to change the font on the fly:

@Example
.mc \H'+5z'x\H'0'
@endExample

In compatibility mode, @code{gtroff} behaves differently: If an
increment or decrement is used, it is always taken relative to the
current type size and not relative to the previously selected font
height.  Thus,

@Example
.cp 1
\H'+5'test \H'+5'test
@endExample

@noindent
prints the word @samp{test} twice with the same font height (five points
larger than the current font size).
@endDefesc

@DefescList {\\S, @code{'}, slant, @code{'}}
@DefregListEndx {.slant}
@cindex changing the font slant (@code{\S})
@cindex font slant, changing (@code{\S})
@cindex slant, font, changing (@code{\S})
Slant the current font by @var{slant} degrees.  Positive values slant to
the right.  Only integer values are possible.

The read-only register @code{.slant} contains the font slant as set by
@code{\S}.

Currently, only the @option{-Tps} and @option{-Tpdf} devices support
this feature.

@code{\S} doesn't produce an input token in GNU @code{troff}.  As a
consequence, it can be used in requests like @code{mc} (which expects
a single character as an argument) to change the font on the fly:

@Example
.mc \S'20'x\S'0'
@endExample

@cindex CSTR@tie{}#54 errata
@cindex CSTR@tie{}#54 erratum, @code{\S} escape
This escape is incorrectly documented in the @acronym{AT&T}
@code{troff} manual; the slant is always set to an absolute value.
@endDefesc

@Defreq {ul, [@Var{lines}]}
@cindex underlining (@code{ul})
The @code{ul} request normally underlines subsequent lines if a TTY
output device is used.  Otherwise, the lines are printed in italics
(only the term `underlined' is used in the following).  The single
argument is the quantity of input lines to be underlined; with no
argument, the next line is underlined.  If @var{lines} is zero or
negative, stop the effects of @code{ul} (if it was active).  Requests
and empty lines do not count for computing the number of underlined
input lines, even if they produce some output like @code{tl}.  Lines
inserted by macros (e.g., invoked by a trap) do count.

At the beginning of @code{ul}, the current font is stored and the
underline font is activated.  Within the span of a @code{ul} request, it
is possible to change fonts, but after the last line affected by
@code{ul} the saved font is restored.

This number of lines still to be underlined is associated with the
environment (@pxref{Environments}).  The underline font can be changed
with the @code{uf} request.

@c XXX @xref should be changed to grotty

@c @xref{@code{troff} and @code{nroff} Modes}, for a discussion of how
@c underlining is implemented for terminal output devices, and what
@c problems can arise.

The @code{ul} request does not underline spaces.
@endDefreq

@Defreq {cu, [@Var{lines}]}
@cindex continuous underlining (@code{cu})
@cindex underlining, continuous (@code{cu})
The @code{cu} request is similar to @code{ul} but underlines spaces as
well (if a TTY output device is used).
@endDefreq

@Defreq {uf, font}
@cindex underline font (@code{uf})
@cindex font for underlining (@code{uf})
Set the underline font (globally) used by @code{ul} and @code{cu}.  By
default, this is the font at position@tie{}2.  @var{font} can be either
a non-negative font position or the name of a font.
@endDefreq

@DefreqList {bd, font [@Var{offset}]}
@DefreqItem {bd, font1 font2 [@Var{offset}]}
@DefregListEndx {.b}
@cindex imitating boldface (@code{bd})
@cindex boldface, imitating (@code{bd})
Embolden @var{font} by overstriking its glyphs offset by @var{offset}
units minus one.

Two syntax forms are available.

@itemize @bullet
@item
Imitate a bold font unconditionally.  The first argument specifies the
font to embolden, and the second is the number of basic units, minus
one, by which the two glyphs are offset.  If the second argument is
missing, emboldening is turned off.

@var{font} can be either a non-negative font position or the name of a
font.

@var{offset} is available in the @code{.b} read-only register if a
special font is active; in the @code{bd} request, its default unit is
@samp{u}.

@cindex @code{fspecial} request, and imitating bold
@kindex special
@cindex embolding of special fonts
@cindex special fonts, emboldening
@item
Imitate a bold form conditionally.  Embolden @var{font1} by @var{offset}
only if font @var{font2} is the current font.  This request can be
issued repeatedly to set up different emboldening values for different
current fonts.  If the second argument is missing, emboldening is turned
off for this particular current font.

This affects special fonts only (either set up with the @code{special}
command in font files or with the @code{fspecial} request).
@end itemize
@endDefreq

@Defreq {cs, font [@Var{width} [@Var{em-size}]]}
@cindex constant glyph space mode (@code{cs})
@cindex mode for constant glyph space (@code{cs})
@cindex glyph, constant space
@cindex @code{ps} request, and constant glyph space mode
Switch to and from @dfn{constant glyph space mode}.  If activated, the
width of every glyph is @math{@var{width}/36} ems.  The em size is given
absolutely by @var{em-size}; if this argument is missing, the em value
is taken from the current font size (as set with the @code{ps} request)
when the font is effectively in use.  Without second and third argument,
constant glyph space mode is deactivated.

Default scaling indicator for @var{em-size} is @samp{z}; @var{width} is
an integer.
@endDefreq

@c ---------------------------------------------------------------------

@node Ligatures and Kerning, Dummy Characters, Artificial Fonts, Fonts and Symbols
@subsection Ligatures and Kerning
@cindex ligatures and kerning
@cindex kerning and ligatures

Ligatures are groups of characters that are run together, i.e, producing
a single glyph.  For example, the letters `f' and `i' can form a
ligature `fi' as in the word `file'.  This produces a cleaner look
(albeit subtle) to the printed output.  Usually, ligatures are not
available in fonts for TTY output devices.

Most PostScript fonts support the fi and fl ligatures.  The C/A/T
typesetter that was the target of @acronym{AT&T} @code{troff} also
supported `ff', `ffi', and `ffl' ligatures.  Advanced typesetters or
`expert' fonts may include ligatures for `ft' and `ct', although GNU
@code{troff} does not support these (yet).

Only the current font is checked for ligatures and kerns; neither
special fonts nor special charcters defined with the @code{char} request
(and its siblings) are taken into account.

@DefreqList {lg, [@Var{flag}]}
@DefregListEndx {.lg}
@cindex activating ligatures (@code{lg})
@cindex ligatures, activating (@code{lg})
@cindex ligatures enabled register (@code{.lg})
Switch the ligature mechanism on or off; if the parameter is non-zero or
missing, ligatures are enabled, otherwise disabled.  Default is on.  The
current ligature mode can be found in the read-only register @code{.lg}
(set to 1 or@tie{}2 if ligatures are enabled, 0@tie{}otherwise).

Setting the ligature mode to@tie{}2 enables the two-character ligatures
(fi, fl, and ff) and disables the three-character ligatures (ffi and
ffl).
@endDefreq

@dfn{Pairwise kerning} is another subtle typesetting mechanism that
modifies the distance between a glyph pair to improve readability.  In
most cases (but not always) the distance is decreased.
@iftex
For example, compare the combination of the letters `V' and `A'.  With
kerning, `VA' is printed.  Without kerning it appears as `V@w{}A'.
@end iftex
Typewriter-like fonts and fonts for terminals where all glyphs have the
same width don't use kerning.

@DefreqList {kern, [@Var{flag}]}
@DefregListEndx {.kern}
@cindex activating kerning (@code{kern})
@cindex kerning, activating (@code{kern})
@cindex kerning enabled register (@code{.kern})
Switch kerning on or off.  If the parameter is non-zero or missing,
enable pairwise kerning, otherwise disable it.  The read-only register
@code{.kern} is set to@tie{}1 if pairwise kerning is enabled,
0@tie{}otherwise.

@cindex dummy character (@code{\&}), effect on kerning
@cindex character, dummy (@code{\&}), effect on kerning
If the font description file contains pairwise kerning information,
glyphs from that font are kerned.  Kerning between two glyphs can be
inhibited by placing @code{\&} between them: @samp{V\&A}.

@xref{Font Description File Format}.
@endDefreq

@cindex track kerning
@cindex kerning, track
@dfn{Track kerning} expands or reduces the space between glyphs.  This
can be handy, for example, if you need to squeeze a long word onto a
single line or spread some text to fill a narrow column.  It must be
used with great care since it is usually considered bad typography if
the reader notices the effect.

@Defreq {tkf, f s1 n1 s2 n2}
@cindex activating track kerning (@code{tkf})
@cindex track kerning, activating (@code{tkf})
Enable track kerning for font@tie{}@var{f}.  If the current font
is@tie{}@var{f} the width of every glyph is increased by an amount
between @var{n1} and @var{n2} (@var{n1}, @var{n2} can be negative); if
the current type size is less than or equal to @var{s1} the width is
increased by @var{n1}; if it is greater than or equal to @var{s2} the
width is increased by @var{n2}; if the type size is greater than or
equal to @var{s1} and less than or equal to @var{s2} the increase in
width is a linear function of the type size.

The default scaling indicator is @samp{z} for @var{s1} and @var{s2},
@samp{p} for @var{n1} and @var{n2}.

The track kerning amount is added even to the rightmost glyph in a line;
for large values it is thus recommended to increase the line length by
the same amount to compensate.
@endDefreq

Sometimes, when typesetting letters of different fonts, more or less
space at such boundaries is needed.  Two escape sequences help with
this.

@Defesc {\\/, , , }
@cindex italic correction (@code{\/})
@cindex correction, italic (@code{\/})
@cindex correction between oblique and upright glyph (@code{\/}, @code{\,})
@cindex roman glyph, correction after italic glyph (@code{\/})
@cindex upright glyph, correction after oblique glyph (@code{\/})
Apply an @dfn{italic correction}:@: modify the spacing of the preceding
glyph so that the distance between it and the following glyph is correct
if the latter is of upright shape.  For example, if an
italic@tie{}@samp{f} is followed immediately by a roman right
parenthesis, then in many fonts the top right portion of
the@tie{}@samp{f} overlaps the top left of the right parenthesis, which
is ugly.  Use this escape sequence whenever an oblique glyph is
immediately followed by an upright glyph without any intervening space.
@endDefesc

@Defesc {\\\,, , , }
@cindex left italic correction (@code{\,})
@cindex correction, left italic (@code{\,})
@cindex correction between upright and oblique glyph (@code{\/}, @code{\,})
@cindex roman glyph, correction before italic glyph (@code{\,})
@cindex upright glyph, correction before oblique glyph (@code{\,})
Apply a @dfn{left italic correction}:@: modify the spacing of the
following glyph so that the distance between it and the preceding
glyph is correct if the latter is of upright shape.  For example,
if a roman left parenthesis is immediately followed by an
italic@tie{}@samp{f}, then in many fonts the bottom left portion of
the@tie{}@samp{f} overlaps the bottom of the left parenthesis, which is
ugly.  Use this escape sequence whenever an upright glyph is followed
immediately by an oblique glyph without any intervening space.
@endDefesc

@c TODO: Move this node earlier in the text due to dummy characters'
@c multifarious effects.
@node Dummy Characters,  , Ligatures and Kerning, Fonts and Symbols
@subsection Dummy Characters

As discussed in @ref{Requests and Macros}, the first character on an
input line is treated specially.  Further, formatting a glyph has many
consequences on formatter state (@pxref{Environments}).  Occasionally,
we want to escape this context or embrace some of those consequences
without actually rendering a glyph to the output.

@Defesc {\\&, , , }
@cindex dummy character (@code{\&})
@cindex character, dummy (@code{\&})
Interpolate a dummy character, which is constitutive of output but
invisible.@footnote{Opinions of this escape sequence's name abound.
``Zero-width space'' is a popular misnomer:@: @code{roff} formatters do
not treat it like a space.  Ossanna called it a ``non-printing,
zero-width character'', but the character causes @emph{output} even
though it does not ``print''.  If no output line is pending, the dummy
character starts one.  Contrast an empty input document with one
containing only @code{\&} (and a newline).  The former produces no
output; the latter, a blank page.}  Its presence alters the
interpretation context of a subsequent input character, and enjoys
several applications.

@itemize @bullet
@item
It prevents insertion of extra space after an end-of-sentence character
(@pxref{Sentences}).

@Example
Test.
Test.
    @result{} Test.  Test.
Test.\&
Test.
    @result{} Test. Test.
@endExample

@item
It prevents interpretation of a control character at the beginning of an
input line.

@Example
.Test
    @error{} warning: macro 'Test' not defined
\&.Test
    @result{} .Test
@endExample

@item
It prevents kerning between two glyphs.

@iftex
@c can't use @Example...@endExample here
@example
@group
VA
    @result{} @r{VA}
V\&A
    @result{} @r{V@w{}A}
@end group
@end example
@end iftex

@item
It permits the @code{tr} request to remap a character to nothing
(@pxref{Character Translations}).
@end itemize

The dummy character escape sequence sees use in macro definitions as a
means of ensuring that arguments are treated as text even if they begin
with spaces or control characters.

@Example
.de HD \" typeset a simple bold heading
.  sp
.  ft B
\&\\$1 \" exercise: remove the \&
.  ft
.  sp
..
.HD .\|.\|.\|surprised?
@endExample
@endDefesc

One way to think about the dummy character is to imagine placing the
symbol @samp{&} in the input at a certain location; if doing so has all
the side effects on formatting that you desire except for sticking an
ugly ampersand in the midst of your text, the dummy character is what
you want in its place.

@Defesc {\\), , , }
@cindex transparent dummy character (@code{\)})
@cindex character, transparent dummy (@code{\)})
@cindex dummy character, transparent (@code{\)})
Interpolate a @slanted{transparent} dummy character---one that is
transparent to end-of-sentence detection.  It behaves as @code{\&},
except that @code{\&} is treated as letters and numerals normally are
after @samp{.}, @samp{?} and @samp{!}; @code{\&} cancels end-of-sentence
detection, and @code{\)} does not.
@c This feature seems too weak to me; see Savannah #60571.  -- GBR

@Example
.de Suffix-&
.  nop \&\\$1
..
.
.de Suffix-)
.  nop \)\\$1
..
.
Here's a sentence.\c
.Suffix-& '
Another one.\c
.Suffix-) '
And a third.
    @result{} Here's a sentence.' Another one.'  And a third.
@endExample
@endDefesc


@c =====================================================================

@c TODO: Move the troff and nroff mode stuff here.  Try to keep stuff
@c that isn't ignored in nroff above this point, and stuff for
@c typesetters below, until we hit the programming/advanced concepts.
@c XXX: Thorny issue: nroff/terminal devices ignore type size but
@c _honor_ vertical spacing (to within their crude vertical motion
@c quanta).

@node Manipulating Type Size and Vertical Spacing, Colors, Fonts and Symbols, GNU troff Reference
@section Manipulating Type Size and Vertical Spacing
@cindex manipulating type size and vertical spacing

@cindex text baseline
@cindex baseline, text
@cindex type size
@cindex size, size
@cindex vertical spacing
@cindex spacing, vertical
These concepts were introduced in @ref{Page Geometry}.  The height of a
font's tallest glyph is one em, which is equal to the type size in
points.@footnote{In text fonts, the tallest glyphs are typically
parentheses.  Unfortunately, in many cases the actual dimensions of the
glyphs in a font do not closely match its declared type size!  For
example, in the standard PostScript font families, 10-point Times sets
better with 9-point Helvetica and 11-point Courier than if all three
were used at 10@tie{}points.}  A vertical spacing of less than 120% of
the type size can make a document hard to read.  Larger proportions can
be useful to spread the text for annotations or proofreader's marks.  By
default, GNU @code{troff} uses 10@tie{}point type on 12@tie{}point
spacing.
@cindex leading
Typographers call the difference between type size and vertical spacing
@dfn{leading}.@footnote{Pronounce ``leading'' to rhyme with
``sledding''; it refers to the use of lead metal (Latin: @emph{plumbum})
in traditional typesetting.}

@menu
* Changing the Type Size::
* Changing the Vertical Spacing::
* Using Fractional Type Sizes::
@end menu

@c ---------------------------------------------------------------------

@node Changing the Type Size, Changing the Vertical Spacing, Manipulating Type Size and Vertical Spacing, Manipulating Type Size and Vertical Spacing
@subsection Changing the Type Size

@DefreqList {ps, [@Var{size}]}
@DefreqItem {ps, @t{+}@Var{size}}
@DefreqItem {ps, @t{-}@Var{size}}
@DefescItemx {\\s, , size, }
@DefregListEndx {.s}
@cindex changing type sizes (@code{ps}, @code{\s})
@cindex type sizes, changing (@code{ps}, @code{\s})
@cindex point sizes, changing (@code{ps}, @code{\s})
Use the @code{ps} request or the @code{\s} escape sequence to change
(increase, decrease) the type size (in scaled points).  Specify
@var{size} as either an absolute type size, or as a relative change from
the current size.  @code{ps} with no argument restores the previous
size.  The @code{ps} request's default scaling unit is @samp{z}.  If the
requested size is non-positive, it is set to 1@dmn{u}.

@cindex type size registers (@code{.s}, @code{.ps})
@cindex point size registers (@code{.s}, @code{.ps})
The read-only string-valued register @code{.s} interpolates the type
size in points as a decimal fraction; it is associated with the
environment (@pxref{Environments}).  To obtain the type size in scaled
points, interpolate the @code{.ps} register instead (@pxref{Using
Fractional Type Sizes}).

@Example
snap, snap,
.ps +2
grin, grin,
.ps +2
wink, wink, \s+2nudge, nudge,\s+8 say no more!
.ps 10
@endExample

The @code{\s} escape sequence supports a variety of syntax forms.

@table @code
@item \s@var{n}
Set the type size to @var{n}@tie{}points.  @var{n}@tie{}must be a single
digit.  If @var{n}@tie{}is 0, restore the previous size.

@item \s+@var{n}
@itemx \s-@var{n}
Increase or decrease the type size by @var{n}@tie{}points.
@var{n}@tie{}must be exactly one digit.

@item \s(@var{nn}
Set the type size to @var{nn}@tie{}points.  @var{nn} must be exactly two
digits.

@item \s+(@var{nn}
@itemx \s-(@var{nn}
@itemx \s(+@var{nn}
@itemx \s(-@var{nn}
Increase or decrease the type size by @var{nn}@tie{}points.  @var{nn}
must be exactly two digits.
@end table

@xref{Using Fractional Type Sizes}, for further syntactical forms of the
@code{\s} escape sequence that additionally accept decimal fractions.
@endDefreq

The @code{\s} escape sequence affects the environment immediately and
doesn't produce an input token.  Consequently, it can be used in
requests like @code{mc}, which expects a single character as an
argument, to change the type size on the fly.

@Example
.mc \s[20]x\s[0]
@endExample

@Defreq {sizes, s1 s2 @dots{} sn [@t{0}]}
Some devices may permit only certain type sizes, in which case GNU
@code{troff} rounds to the nearest permissible size.  The @file{DESC}
file normally specifies which type sizes are allowed by the device.

Use the @code{sizes} request to change the permissible sizes for the
output device.  Arguments are in scaled points; @xref{Using Fractional
Type Sizes}.  Each can be a single type size (such as @samp{12000}), or
a range of sizes (such as @samp{4000-72000}).  You can optionally end
the list with a zero.
@endDefreq

@need 1000
@node Changing the Vertical Spacing, Using Fractional Type Sizes, Changing the Type Size, Manipulating Type Size and Vertical Spacing
@subsection Changing the Vertical Spacing

@DefreqList {vs, [@Var{space}]}
@DefreqItem {vs, @t{+}@Var{space}}
@DefreqItem {vs, @t{-}@Var{space}}
@DefregListEndx {.v}
@cindex changing vertical line spacing (@code{vs})
@cindex vertical line spacing, changing (@code{vs})
@cindex vertical line spacing register (@code{.v})
Change (increase, decrease) the vertical spacing by @var{space}.  The
default scaling unit is @samp{p}.  If @code{vs} is called without an
argument, the vertical spacing is reset to the previous value before the
last call to @code{vs}.
@cindex @code{.V} register, and @code{vs}
GNU @code{gtroff} emits a warning in category @samp{range} if
@var{space} is negative; the vertical spacing is then set to the
smallest possible positive value, the vertical motion quantum (as found
in the @code{.V} register).

@w{@samp{.vs 0}} isn't saved in a diversion since it doesn't result in
a vertical motion.  You must explicitly issue this request before
calling the diversion.

The read-only register @code{.v} contains the vertical spacing; it is
associated with the environment (@pxref{Environments}).
@endDefreq

@cindex vertical line spacing, effective value
@noindent
When a break occurs, GNU @code{troff} performs the following procedure.

@itemize @bullet
@item
@cindex extra pre-vertical line space (@code{\x})
@cindex line space, extra pre-vertical (@code{\x})
Move the drawing position vertically by the @dfn{extra pre-vertical line
space}, the minimum of all negative @code{\x} escape sequence arguments
in the pending output line.

@item
Move the drawing position vertically by the vertical line spacing.

@item
Write out the pending output line.

@item
@cindex extra post-vertical line space (@code{\x})
@cindex line space, extra post-vertical (@code{\x})
Move the drawing position vertically by the @dfn{extra post-vertical line
space}, the maximum of all positive @code{\x} escape sequence arguments
in the line that has just been output.

@item
@cindex post-vertical line spacing
@cindex line spacing, post-vertical (@code{pvs})
Move the drawing position vertically by the @dfn{post-vertical line
spacing} (see below).
@end itemize

@cindex double-spacing (@code{vs}, @code{pvs})
Prefer @code{vs} or @code{pvs} over @code{ls} to produce double-spaced
documents.  @code{vs} and @code{pvs} have finer granularity than
@code{ls}; moreover, some preprocessors assume single spacing.
@xref{Manipulating Spacing}, regarding the @code{\x} escape sequence and
the @code{ls} request.

@DefreqList {pvs, [@Var{space}]}
@DefreqItem {pvs, @t{+}@Var{space}}
@DefreqItem {pvs, @t{-}@Var{space}}
@DefregListEndx {.pvs}
@cindex @code{ls} request, alternative to (@code{pvs})
@cindex post-vertical line spacing, changing (@code{pvs})
@cindex post-vertical line spacing register (@code{.pvs})
Change (increase, decrease) the post-vertical spacing by @var{space}.
The default scaling unit is @samp{p}.  If @code{pvs} is called without
an argument, the post-vertical spacing is reset to the previous value
before the last call to @code{pvs}.  GNU @code{troff} emits a warning in
category @samp{range} if @var{space} is negative; the post-vertical
spacing is then set to zero.

The read-only register @code{.pvs} contains the post-vertical spacing;
it is associated with the environment (@pxref{Environments}).
@endDefreq

@c ---------------------------------------------------------------------

@c BEGIN Keep (roughly) parallel with subsection "Fractional type sizes
@c and new scaling units" of groff_diff(7).
@node Using Fractional Type Sizes,  , Changing the Type Size, Manipulating Type Size and Vertical Spacing
@subsection Using Fractional Type Sizes
@cindex fractional type sizes
@cindex fractional point sizes
@cindex type sizes, fractional
@cindex point sizes, fractional
@cindex sizes, fractional type

AT&T @code{troff} interpreted all type size measurements in points.
Combined with integer arithmetic, this design choice made it impossible
to support, for instance, ten and a half-point type.  In GNU
@code{troff}, an output device can select a scaling factor that
subdivides a point into ``scaled points''.  A type size expressed in
scaled points can thus represent a non-integral type size.

@cindex @code{s} scaling unit
@cindex unit, scaling, @code{s}
@cindex scaling unit @code{s}
@cindex @code{z} scaling unit
@cindex unit, scaling, @code{z}
@cindex scaling unit @code{z}
@cindex @code{ps} request, with fractional type sizes
@cindex @code{cs} request, with fractional type sizes
@cindex @code{tkf} request, with fractional type sizes
@cindex @code{\H}, with fractional type sizes
@cindex @code{\s}, with fractional type sizes
A @dfn{scaled point} is equal to @math{1/@var{sizescale}} points, where
@var{sizescale} is specified in the device description file @file{DESC},
and defaults to@tie{}1.@footnote{@xref{Device and Font Description
Files}.}  Requests and escape sequences in GNU @code{troff} interpret
arguments that represent a type size in scaled points, which the
formatter multiplies by @var{sizescale} and converts to an integer.
Arguments treated in this way comprise those to the escape sequences
@code{\H} and @code{\s}, to the request @code{ps}, the third argument to
the @code{cs} request, and the second and fourth arguments to the
@code{tkf} request.  Scaled points may be specified explicitly with the
@code{z} scaling unit.

For example, if @var{sizescale} is@tie{}1000, then a scaled point is one
thousandth of a point.  The request @samp{.ps 10.5} is synonymous with
@samp{.ps 10.5z} and sets the type size to 10,500@tie{}scaled points, or
10.5@tie{}points.  Consequently, in GNU @code{troff}, the register
@code{.s} can interpolate a non-integral type size.

@Defreg {.ps}
This read-only register interpolates the type size in scaled points; it
is associated with the environment (@pxref{Environments}).
@endDefreg

It makes no sense to use the @samp{z} scaling unit in a numeric
expression whose default scaling unit is neither @samp{u} nor @samp{z},
so GNU @code{troff} disallows this.  Similarly, it is nonsensical to use
a scaling unit other than @samp{z} or @samp{u} in a numeric expression
whose default scaling unit is @samp{z}, and so GNU @code{troff}
disallows this as well.

Another GNU @code{troff} scaling unit, @samp{s}, multiplies by the
number of basic units in a scaled point.  Thus, @samp{\n[.ps]s} is equal
to @samp{1m} by definition.  Do not confuse the @samp{s} and @samp{z}
scaling units.
@c END Keep (roughly) parallel with subsection "Fractional type sizes
@c and new scaling units" of groff_diff(7).

@DefregList {.psr}
@DefregListEndx {.sr}
@cindex last-requested type size registers (@code{.psr}, @code{.sr})
@cindex type size registers, last-requested (@code{.psr}, @code{.sr})
@cindex last-requested point size registers (@code{.psr}, @code{.sr})
@cindex point size registers, last-requested (@code{.psr}, @code{.sr})
@cindex @code{.ps} register, in comparison with @code{.psr}
@cindex @code{.s} register, in comparison with @code{.sr}
Output devices may be limited in the type sizes they can employ.  The
@code{.s} and @code{.ps} registers represent the type size selected by
the output driver as it understands a device's capability.  The last
@emph{requested} type size is interpolated in scaled points by the
read-only register @code{.psr} and in points as a decimal fraction by
the read-only string-valued register @code{.sr}.  Both are associated
with the environment (@pxref{Environments}).

For example, if a type size of 10.95 points is requested, and the
nearest size permitted by a @code{sizes} request (or a @code{sizescale}
directive in the device's @file{DESC} file) is 11 points, the latter
value is used by the output driver.
@endDefreg

The @code{\s} escape sequence offers the following syntax forms that
work with fractional type sizes and accept scaling units.  You may of
course give them integral arguments.  The delimited forms need not use
the neutral apostrophe; see @ref{Delimiters}.

@table @code
@item \s[@var{n}]
@itemx \s'@var{n}'
Set the type size to @var{n}@tie{}scaled points; @var{n}@tie{}is a
numeric expression with a default scaling unit of @samp{z}.

@item \s[+@var{n}]
@itemx \s[-@var{n}]
@itemx \s+[@var{n}]
@itemx \s-[@var{n}]
@itemx \s'+@var{n}'
@itemx \s'-@var{n}'
@itemx \s+'@var{n}'
@itemx \s-'@var{n}'
Increase or decrease the type size by @var{n}@tie{}scaled points;
@var{n}@tie{}is a numeric expression (which may start with a minus sign)
with a default scaling unit of @samp{z}.
@end table


@c =====================================================================

@c BEGIN Keep (roughly) parallel with section "Colors" of groff(7).
@node Colors, Strings, Manipulating Type Size and Vertical Spacing, GNU troff Reference
@section Colors
@cindex colors

@cindex stroke color
@cindex color, stroke
@cindex fill color
@cindex color, fill
GNU @code{troff} supports color output with a variety of color spaces
and up to 16 bits per channel.  Some devices, particularly terminals,
may be more limited.  When color support is enabled, two colors are
current at any given time: the @dfn{stroke color}, with which glyphs,
rules (lines), and geometric objects like circles and polygons are
drawn, and the @dfn{fill color}, which can be used to paint the interior
of a closed geometric figure.

@DefreqList {color, [@Var{n}]}
@DefregListEndx {.color}
If @var{n} is missing or non-zero, enable the output of color-related
device-independent output commands (this is the default); otherwise,
disable them.  This request sets a global flag; it does not produce an
input token (@pxref{Gtroff Internals}).

The read-only register @code{.color} is@tie{}1 if colors are enabled,
0@tie{}otherwise.

Color can also be disabled with the @option{-c} command-line option.
@endDefreq

@Defreq {defcolor, ident scheme color-component @dots{}}
Define a color named @var{ident}.  @var{scheme} selects a color space
and determines the quantity of required @var{color-component}s; it must
be one of @samp{rgb} (three components), @samp{cmy} (three), @samp{cmyk}
(four), or @samp{gray} (one).  @samp{grey} is accepted as a synonym of
@samp{gray}.  The color components can be encoded as a single
hexadecimal value starting with @samp{#} or @samp{##}.  The former
indicates that each component is in the range 0--255 (0--FF), the latter
the range 0--65,535 (0--FFFF).

@Example
.defcolor half gray #7f
.defcolor pink rgb #FFC0CB
.defcolor magenta rgb  ##ffff0000ffff
@endExample

@cindex @code{f} scaling unit
@cindex unit, scaling, @code{f}
@cindex scaling unit @code{f}
Alternatively, each color component can be specified as a decimal
fraction in the range 0--1, interpreted using a default scaling
indicator of@tie{}@code{f}, which multiplies its value by 65,536 (but
clamps it at 65,535).

@Example
.defcolor gray50 rgb 0.5 0.5 0.5
.defcolor darkgreen rgb 0.1f 0.5f 0.2f
@endExample
@endDefreq

@cindex default color
@cindex color, default
Each output device has a color named @samp{default}, which cannot be
redefined.  A device's default stroke and fill colors are not
necessarily the same.  For the @code{dvi}, @code{html}, @code{pdf},
@code{ps}, and @code{xhtml} output devices, GNU @code{troff}
automatically loads a macro file defining many color names at startup.
By the same mechanism, the devices supported by @code{grotty} recognize
the eight standard ISO@tie{}6429/EMCA-48 color names.@footnote{also
known vulgarly as ``ANSI colors''}

@DefreqList {gcolor, [@Var{color}]}
@DefescItemx {\\m, , c, }
@DefescItem {\\m, (, co, }
@DefescItem {\\m, [, color, ]}
@DefregListEndx {.m}
Set the stroke color to @var{color}.

@Example
.gcolor red
The next words
.gcolor
\m[red]are in red\m[]
and these words are in the previous color.
@endExample

The escape sequence @code{\m[]} restores the previous stroke color, as
does a @code{gcolor} request without an argument.

@cindex stroke color name register (@code{.m})
@cindex name, stroke color, register (@code{.m})
@cindex color name, stroke, register (@code{.m})
The name of the current stroke color is available in the read-only
string-valued register @samp{.m}; it is associated with the environment
(@pxref{Environments}).  It interpolates nothing when the stroke color
is the default.

@code{\m} doesn't produce an input token in GNU @code{troff}
(@pxref{Gtroff Internals}).  It therefore can be used in requests like
@code{mc} (which expects a single character as an argument) to change
the color on the fly:

@Example
.mc \m[red]x\m[]
@endExample
@endDefesc

@DefreqList {fcolor, [@Var{color}]}
@DefescItemx {\\M, , c, }
@DefescItem {\\M, (, co, }
@DefescItem {\\M, [, color, ]}
@DefregListEndx {.M}
Set fill color for objects drawn with @code{\D'@dots{}'} escape
sequences.

Create an ellipse with a red interior as follows.

@Example
\M[red]\h'0.5i'\D'E 2i 1i'\M[]
@endExample

The escape sequence @code{\M[]} restores the previous fill color, as
does an @code{fcolor} request without an argument.

@cindex background color name register (@code{.M})
@cindex name, background color, register (@code{.M})
@cindex color name, background, register (@code{.M})
@cindex fill color name register (@code{.M})
@cindex name, fill color, register (@code{.M})
@cindex color name, fill, register (@code{.M})
The name of the current fill color is available in the read-only
string-valued register @samp{.M}; it is associated with the environment
(@pxref{Environments}).  It interpolates nothing when the fill color
is the default.

@code{\M} doesn't produce an input token in GNU @code{troff}.
@endDefesc
@c END Keep (roughly) parallel with section "Colors" of groff(7).


@c =====================================================================

@c BEGIN Keep (roughly) parallel with section "Strings" of groff(7).
@node Strings, Conditionals and Loops, Colors, GNU troff Reference
@section Strings
@cindex strings

GNU @code{troff} supports strings primarily for user convenience.
Conventionally, if one would define a macro only to interpolate a small
amount of text, without invoking requests or calling any other macros,
one defines a string instead.  Only one string is predefined by the
language.

@Defstr {.T}
@stindex .T
@cindex output device name string (@code{.T})
Contains the name of the output device (for example, @samp{utf8} or
@samp{pdf}).
@endDefmpstr

The @code{ds} request creates a string with a specified name and
contents and the @code{\*} escape sequence dereferences its name,
interpolating its contents.  If the string named by the @code{\*} escape
sequence does not exist, it is defined as empty, nothing is
interpolated, and a warning in category @samp{mac} is emitted.
@xref{Warnings}, for information about the enablement and suppression of
warnings.

@DefreqList {ds, name [@Var{contents}]}
@DefreqItemx {ds1, name [@Var{contents}]}
@DefescItemx {\\*, , n, }
@DefescItem {\\*, (, nm, }
@c XXX: Can't mark the parameters with @Var because @Var gets called
@c recursively if we do.
@c @DefescListEnd {\\*, [, name [@Var{arg1} @Var{arg2} @dots{}], ]}
@DefescListEnd {\\*, [, name @sansserif{[}arg1 arg2 @dots{}@sansserif{]}, ]}
@cindex string interpolation (@code{\*})
@cindex string expansion (@code{\*})
@cindex interpolation of strings (@code{\*})
@cindex expansion of strings (@code{\*})
@cindex string arguments
@cindex arguments, of strings
Define a string called @var{name} with contents @var{contents}.  If
@var{name} already exists as an alias, the target of the alias is
redefined; see @code{als} and @code{rm} below.  If @code{ds} is called
with only one argument, @var{name} is defined as an empty string.
Otherwise, GNU @code{troff} stores @var{contents} in copy
mode.@footnote{@xref{Copy Mode}.}

The @code{\*} escape sequence interpolates a previously defined string
variable @var{name} (one-character name@tie{}@var{n}, two-character name
@var{nm}).  The bracketed interpolation form accepts arguments that are
handled as macro arguments are; recall @ref{Calling Macros}.  In
contrast to macro calls, however, if a closing bracket @samp{]} occurs
in a string argument, that argument must be enclosed in double quotes.
@code{\*} is interpreted even in copy mode.  When defining strings,
argument interpolations must be escaped if they are to reference
parameters from the calling context; @xref{Parameters}.

@Example
.ds cite (\\$1, \\$2)
Gray codes are explored in \*[cite Morgan 1998].
    @result{} Gray codes are explored in (Morgan, 1998).
@endExample

@c TODO: Consider examples of recursive string calls, particularly where
@c one interpolation is constructed from the argument of an enclosing
@c macro, to illustrate ".ds a \$1 \\$1".
@c
@c @Example
@c .ds a \\$1 wildebeest
@c .ds b big, \*[a hairy]
@c I see a \*[b].
@c     @result{} I see a big, hairy wildebeest.
@c @endExample

@cindex trailing spaces in string definitions and appendments
@cindex comments, with @code{ds}
@cindex @code{ds} request, and comments
@strong{Caution:@:} Unlike other requests, the second argument to the
@code{ds} request consumes the remainder of the input line, including
trailing spaces.  This means that comments on a line with such a request
can introduce unwanted space into a string when they are set off from
the material they annotate, as is conventional.

@Example
.ds H2O H\v'+.3m'\s'-2'2\v'-.3m'\s0O \" water
@endExample

@noindent
Instead, place the comment on another line or put the comment escape
sequence immediately adjacent to the last character of the string.

@Example
.ds H2O H\v'+.3m'\s'-2'2\v'-.3m'\s0O\" water
@endExample

Ending string definitions (and appendments) with a comment, even an
empty one, prevents unwanted space from creeping into them during source
document maintenance.

@Example
.ds author Alice Pleasance Liddell\"
.ds empty \" might be appended to later with .as
@endExample

@cindex trailing double quotes in strings
@cindex double quotes, trailing, in strings
@cindex @code{ds} request, and double quotes
@cindex leading spaces with @code{ds}
@cindex spaces with @code{ds}
@cindex @code{ds} request, and leading spaces
An initial neutral double quote @code{"} in @var{contents} is stripped
to allow embedding of leading spaces.  Any other @code{"} is interpreted
literally, but it is wise to use the special character escape sequence
@code{\[dq]} instead if the string might be interpolated as part of a
macro argument; see @ref{Calling Macros}.

@c Examples should be more accessible than Unix nerd stuff like this,
@c but in general document authors shouldn't want to use "straight"
@c double quotes for ordinary prose anyway.  Also, 56 chars is as fat
@c as these examples can get and not overrun the right margin in PDF.
@Example
.ds salutation "         Yours in a white wine sauce,\"
.ds c-var-defn "  char mydate[]=\[dq]2020-07-29\[dq];\"
@endExample

@cindex multi-line strings
@cindex strings, multi-line
@cindex newline character, in strings, escaping
@cindex escaping newline characters, in strings
Strings are not limited to a single input line of text.
@code{\@key{RET}} works just as it does elsewhere.  The resulting string
is stored @emph{without} the newlines.  Care is therefore required when
interpolating strings while filling is disabled.

@Example
.ds foo This string contains \
text on multiple lines \
of input.
@endExample

It is not possible to embed a newline in a string that will be
interpreted as such when the string is interpolated.  To achieve that
effect, use @code{\*} to interpolate a macro instead; see @ref{Punning
Names}.

Because strings are similar to macros, they too can be defined so as to
suppress AT&T @code{troff} compatibility mode when used; see
@ref{Writing Macros} and @ref{Compatibility Mode}.  The @code{ds1}
request defines a string such that compatibility mode is off when the
string is later interpolated.  To be more precise, a @dfn{compatibility
save} input token is inserted at the beginning of the string, and a
@dfn{compatibility restore} input token at the end.

@Example
.nr xxx 12345
.ds aa The value of xxx is \\n[xxx].
.ds1 bb The value of xxx is \\n[xxx].
.
.cp 1
.
\*(aa
    @error{} warning: register '[' not defined
    @result{} The value of xxx is 0xxx].
\*(bb
    @result{} The value of xxx is 12345.
@endExample
@endDefreq

@DefreqList {as, name [@Var{contents}]}
@DefreqListEndx {as1, name [@Var{contents}]}
@cindex appending to a string (@code{as})
@cindex string, appending (@code{as})
The @code{as} request is similar to @code{ds} but appends @var{contents}
to the string stored as @var{name} instead of redefining it.  If
@var{name} doesn't exist yet, it is created.  If @code{as} is called
with only one argument, no operation is performed (beyond dereferencing
the string).

@Example
.as salutation " with shallots, onions and garlic,\"
@endExample

The @code{as1} request is similar to @code{as}, but compatibility mode
is switched off when the appended portion of the string is later
interpolated.  To be more precise, a @dfn{compatibility save} input
token is inserted at the beginning of the appended string, and a
@dfn{compatibility restore} input token at the end.
@endDefreq

Several requests exist to perform rudimentary string operations.
Strings can be queried (@code{length}) and modified (@code{chop},
@code{substring}, @code{stringup}, @code{stringdown}), and their names
can be manipulated through renaming, removal, and aliasing (@code{rn},
@code{rm}, @code{als}).

@Defreq {length, reg anything}
@cindex length of a string (@code{length})
@cindex string, length of (@code{length})
@cindex @code{length} request, and copy mode
@cindex copy mode, and @code{length} request
@cindex mode, copy, and @code{length} request
Compute the number of characters of @var{anything} and store the count
in the register @var{reg}.  If @var{reg} doesn't exist, it is created.
@var{anything} is read in copy mode.

@Example
.ds xxx abcd\h'3i'efgh
.length yyy \*[xxx]
\n[yyy]
    @result{} 14
@endExample
@endDefreq

@Defreq {chop, object}
Remove the last character from the macro, string, or diversion named
@var{object}.  This is useful for removing the newline from the end of a
diversion that is to be interpolated as a string.  This request can be
used repeatedly on the same @var{object}; see @ref{Gtroff Internals},
for details on nodes inserted additionally by GNU @code{troff}.
@endDefreq

@Defreq {substring, str start [@Var{end}]}
@cindex substring (@code{substring})
Replace the string named @var{str} with its substring bounded by the
indices @var{start} and @var{end}, inclusively.  The first character in
the string has index@tie{}0.  If @var{end} is omitted, it is implicitly
set to the largest valid value (the string length minus one).  Negative
indices count backward from the end of the string:@: the last character
has index@tie{}@minus{}1, the character before the last has
index@tie{}@minus{}2, and so on.

@Example
.ds xxx abcdefgh
.substring xxx 1 -4
\*[xxx]
    @result{} bcde
.substring xxx 2
\*[xxx]
    @result{} de
@endExample
@endDefreq

@DefreqList {stringdown, str}
@DefreqListEndx {stringup, str}
@cindex case-transforming a string (@code{stringdown}, @code{stringup})
@cindex uppercasing a string (@code{stringup})
@cindex lowercasing a string (@code{stringdown})
@cindex up-casing a string (@code{stringup})
@cindex down-casing a string (@code{stringdown})
Alter the string named @var{str} by replacing each of its bytes with its
lowercase (@code{stringdown}) or uppercase (@code{stringup}) version (if
one exists).  Special characters in the string will often transform in
the expected way due to the regular naming convention for accented
characters.  When they do not, use substrings and/or catenation.

@Example
.ds resume R\['e]sum\['e]
\*[resume]
.stringdown resume
\*[resume]
.stringup resume
\*[resume]
    @result{} Résumé résumé RÉSUMÉ
@endExample
@endDefreq

(In practice, we would end the @code{ds} request with a comment escape
@code{\"} to prevent space from creeping into the definition during
source document maintenance.)

@Defreq {rn, old new}
@cindex renaming request (@code{rn})
@cindex request, renaming (@code{rn})
@cindex renaming macro (@code{rn})
@cindex macro, renaming (@code{rn})
@cindex renaming string (@code{rn})
@cindex string, renaming (@code{rn})
@cindex renaming diversion (@code{rn})
@cindex diversion, renaming (@code{rn})
Rename the request, macro, diversion, or string @var{old} to @var{new}.
@endDefreq

@Defreq {rm, name}
@cindex removing request (@code{rm})
@cindex request, removing (@code{rm})
@cindex removing macro (@code{rm})
@cindex macro, removing (@code{rm})
@cindex removing string (@code{rm})
@cindex string, removing (@code{rm})
@cindex removing diversion (@code{rm})
@cindex diversion, removing (@code{rm})
Remove the request, macro, diversion, or string @var{name}.  GNU
@code{troff} treats subsequent invocations as if the name had never
been defined.
@endDefreq

@anchor{als}
@Defreq {als, new old}
@cindex alias, string, creating (@code{als})
@cindex alias, macro, creating (@code{als})
@cindex alias, diversion, creating (@code{als})
@cindex creating alias, for string (@code{als})
@cindex creating alias, for macro (@code{als})
@cindex creating alias, for diversion (@code{als})
@cindex string, creating alias for (@code{als})
@cindex macro, creating alias for (@code{als})
@cindex diversion, creating alias for (@code{als})
Create an alias @var{new} for the existing request, string, macro, or
diversion object named @var{old}, causing the names to refer to the same
stored object.  If @var{old} is undefined, a warning in category
@samp{mac} is produced, and the request is ignored.  @xref{Warnings},
for information about the enablement and suppression of warnings.

To understand how the @code{als} request works, consider two different
storage pools:@: one for objects (macros, strings, etc.), and another
for names.  As soon as an object is defined, GNU @code{troff} adds it to
the object pool, adds its name to the name pool, and creates a link
between them.  When @code{als} creates an alias, it adds a new name to
the name pool that gets linked to the same object as the old name.

Now consider this example.

@Example
.de foo
..
.
.als bar foo
.
.de bar
.  foo
..
.
.bar
    @error{} input stack limit exceeded (probable infinite
    @error{} loop)
@endExample

@noindent
In the above, @code{bar} remains an @emph{alias}---another name
for---the object referred to by @code{foo}, which the second @code{de}
request replaces.  Alternatively, imagine that the @code{de} request
@emph{dereferences} its argument before replacing it.  Either way, the
result of calling @code{bar} is a recursive loop that finally leads to
an error.  @xref{Writing Macros}.

@cindex alias, string, removing (@code{rm})
@cindex alias, macro, removing (@code{rm})
@cindex alias, diversion, removing (@code{rm})
@cindex removing alias, for string (@code{rm})
@cindex removing alias, for macro (@code{rm})
@cindex removing alias, for diversion (@code{rm})
@cindex string, removing alias for (@code{rm})
@cindex macro, removing alias for (@code{rm})
@cindex diversion, removing alias for (@code{rm})
To remove an alias, call @code{rm} on its name.  The object itself is
not destroyed until it has no more names.

When a request, macro, string, or diversion is aliased, redefinitions
and appendments ``write through'' alias names.  To replace an alias with
a separately defined object, you must use the @code{rm} request on its
name first.
@endDefreq
@c END Keep (roughly) parallel with section "Strings" of groff(7).


@c =====================================================================

@node Conditionals and Loops, Writing Macros, Strings, GNU troff Reference
@section Conditionals and Loops
@cindex conditionals and loops
@cindex loops and conditionals

@code{groff} has @code{if} and @code{while} control structures like
other languages.  However, the syntax for grouping multiple input lines
in the branches or bodies of these structures is unusual.

@menu
* Operators in Conditionals::
* if-then::
* if-else::
* Conditional Blocks::
* while::
@end menu

@c ---------------------------------------------------------------------

@c BEGIN Keep (roughly) parallel with subsection "Conditional
@c expressions" of groff(7).
@node Operators in Conditionals, if-then, Conditionals and Loops, Conditionals and Loops
@subsection Operators in Conditionals

@cindex @code{if} request, operators to use with
@cindex @code{ie} request, operators to use with
@cindex @code{while} request, operators to use with
@cindex conditional expressions
@cindex expressions, conditional
In @code{if}, @code{ie}, and @code{while} requests, in addition to the
numeric expressions described in @ref{Numeric Expressions}, several
Boolean operators are available; the members of this expanded class are
termed @dfn{conditional expressions}.

@table @code
@item c @var{glyph}
True if @var{glyph} is available, where @var{glyph} is an ordinary
character, a special character @samp{\(@var{xx}} or @samp{\[@var{xxx}]},
@samp{\N'@var{xxx}'}, or has been defined by any of the @code{char},
@code{fchar}, @code{fschar}, or @code{schar} requests.

@item d @var{name}
True if a string, macro, diversion, or request called @var{name} exists.

@item e
True if the current page is even-numbered.

@item F @var{font}
True if @var{font} exists.  @var{font} is handled as if it were opened
with the @code{ft} request (that is, font translation and styles are
applied), without actually mounting it.

@item m @var{color}
True if @var{color} is defined.

@item n
@cindex conditional output for terminal (TTY)
@cindex TTY, conditional output for
@cindex terminal, conditional output for
True if the document is being processed in @code{nroff} mode.
@xref{@code{troff} and @code{nroff} Modes}.

@item o
True if the current page is odd-numbered.

@item r @var{register}
True if @var{register} exists.

@item S @var{style}
True if @var{style} is available for the current font family.  Font
translation is applied.

@item t
True if the document is being processed in @code{troff} mode.
@xref{@code{troff} and @code{nroff} Modes}.

@item v
Always false.  This condition is recognized only for compatibility with
certain other @code{troff} implementations.@footnote{This refers to
@code{vtroff}, a translator that would convert the C/A/T output from
early-vintage @acronym{AT&T} @code{troff} to a form suitable for
Versatec and Benson-Varian plotters.}
@end table

If the first argument to an @code{if}, @code{ie}, or @code{while}
request begins with a non-alphanumeric character apart from @code{!}
(see below); it performs an @slanted{output comparison test}.
@footnote{Strictly, letters not otherwise recognized @emph{are} treated
as output comparison delimiters.  For portability, it is wise to avoid
using letters not in the list above; for example, Plan@tie{}9
@code{troff} uses @samp{h} to test a mode it calls @code{htmlroff}, and
GNU @code{troff} may provide additional operators in the future.}

@cindex output comparison operator
@table @code
@item @code{'}@var{xxx}@code{'}@var{yyy}@code{'}
True if formatting the comparands @var{xxx} and @var{yyy} produces the
same output commands.  The delimiter need not be a neutral apostrophe:
the output comparison operator accepts the same delimiters as most
escape sequences; see @ref{Delimiters}.  This @dfn{output comparison
operator} formats @var{xxx} and @var{yyy} in separate environments;
after the comparison, the resulting data are discarded.

@Example
.ie "|"\fR|\fP" true
.el false
    @result{} true
@endExample

@noindent
The resulting glyph properties, including font family, style, size, and
slant, must match, but not necessarily the requests and/or escape
sequences used to obtain them.  In the previous example, @samp{|} and
@samp{\fR|\fP} result in @samp{|} glyphs in the same typefaces at the
same positions, so the comparands are equal.  If @samp{.ft@tie{}I} had
been added before the @samp{.ie}, they would differ: the first @samp{|}
would produce an italic @samp{|}, not a roman one.  Motions must match
in orientation and magnitude to within the applicable horizontal and
vertical motion quanta of the device, after rounding.  @samp{.if
"\u\d"\v'0'"} is false even though both comparands result in zero net
motion, because motions are not interpreted or optimized but sent as-is
to the output.@footnote{Because formatting of the comparands takes place
in a dummy environment, vertical motions within them cannot spring
traps.}  On the other hand, @samp{.if "\d"\v'0.5m'"} is true, because
@code{\d} is defined as a downward motion of one-half em.@footnote{All
of this is to say that the lists of output nodes created by formatting
@var{xxx} and @var{yyy} must be identical.  @xref{Gtroff Internals}.}

@cindex string comparison
@cindex comparison of strings
Surround the comparands with @code{\?} to avoid formatting them; this
causes them to be compared character by character, as with string
comparisons in other programming languages.

@Example
.ie "\?|\?"\?\fR|\fP\?" true
.el false
    @result{} false
@endExample

@cindex @code{\?}, and copy mode
@cindex copy mode, and @code{\?}
@cindex mode, copy, and @code{\?}
@noindent
Since comparands protected with @code{\?} are read in copy mode
(@pxref{Copy Mode}), they need not even be valid @code{groff} syntax.
The escape character is still lexically recognized, however, and
consumes the next character.

@Example
.ds a \[
.ds b \[
.if '\?\*a\?'\?\*b\?' a and b equivalent
.if '\?\\?'\?\\?' backslashes equivalent
    @result{} a and b equivalent
@c slack lines for pagination control
@c    @error{} warning: missing closing delimiter in
@c    @error{} conditional expression (got newline)
@endExample
@end table

The above operators can't be combined with most others, but a leading
@samp{!}, not followed immediately by spaces or tabs, complements an
expression.

@Example
.nr x 1
.ie !r x register x is not defined
.el      register x is defined
    @result{} register x is defined
@endExample

Spaces and tabs are optional immediately after the @samp{c}, @samp{d},
@samp{F}, @samp{m}, @samp{r}, and @samp{S} operators, but right after
@samp{!}, they end the predicate and the conditional evaluates
true.@footnote{This bizarre behavior maintains compatibility with
@acronym{AT&T} @code{troff}.}

@Example
.nr x 1
.ie ! r x register x is not defined
.el       register x is defined
    @result{} r x register x is not defined
@endExample

@noindent
The unexpected @samp{r x} in the output is a clue that our conditional
was not interpreted as we planned, but matters may not always be so
obvious.
@c END Keep (roughly) parallel with subsection "Conditional expressions"
@c of groff(7).

@c ---------------------------------------------------------------------

@node if-then, if-else, Operators in Conditionals, Conditionals and Loops
@subsection if-then
@cindex if-then

@Defreq {if, cond-expr anything}
Evaluate the conditional expression @var{cond-expr}, and if it evaluates
true (or to a positive value), interpret the remainder of the line
@var{anything} as if it were an input line.  Recall from @ref{Invoking
Requests} that any quantity of spaces between arguments to requests
serves only to separate them; leading spaces in @var{anything} are thus
not seen.  @var{anything} effectively @emph{cannot} be omitted; if
@var{cond-expr} is true and @var{anything} is empty, the newline at the
end of the control line is interpreted as a blank input line (and
therefore a blank text line).

@Example
super\c
tanker
.nr force-word-break 1
super\c
.if ((\n[force-word-break] = 1) & \n[.int])
tanker
    @result{} supertanker super tanker
@endExample
@endDefreq

@Defreq {nop, anything}
Interpret @var{anything} as if it were an input line.  This is similar
to @samp{.if@tie{}1}.  @code{nop} is not really ``no operation''; its
argument @emph{is} processed---unconditionally.  It can be used to cause
text lines to share indentation with surrounding control lines.

@Example
.als real-MAC MAC
.de wrapped-MAC
.  tm MAC: called with arguments \\$@@
.  nop \\*[real-MAC]\\
..
.als MAC wrapped-MAC
\# Later...
.als MAC real-MAC
@endExample

In the above, we've used aliasing, @code{nop}, and the interpolation of
a macro as a string to interpose a wrapper around the macro @samp{MAC}
(perhaps to debug it).
@endDefreq

@c ---------------------------------------------------------------------

@node if-else, while, Operators in Conditionals, Conditionals and Loops
@subsection if-else
@cindex if-else

@DefreqList {ie, cond-expr anything}
@DefreqListEndx {el, anything}
Use the @code{ie} and @code{el} requests to write an if-then-else.  The
first request is the ``if'' part and the latter is the ``else'' part.
Unusually among programming languages, any number of non-conditional
requests may be interposed between the @code{ie} branch and the
@code{el} branch.

@Example
.nr a 0
.ie \na a is nonzero.
.nr a +1
.el a was not positive but is now \na.
    @result{} a was not positive but is now 1.
@endExample

Another way in which @code{el} is an ordinary request is that it does
not lexically ``bind'' more tightly to its @code{ie} counterpart than it
does to any other request.  This fact can surprise C programmers.

@Example
.nr a 1
.nr z 0
.ie \nz \
.  ie \na a is true
.  el     a is false
.el z is false
    @error{} warning: unbalanced 'el' request
    @result{} a is false
@endExample

@c Turn the following into a proper @{x,}ref if the conditional blocks
@c node is relocated elsewhere--but consider if it is wise to do so.
To conveniently nest conditionals, keep reading.

@endDefreq

@c ---------------------------------------------------------------------

@node Conditional Blocks, while, Operators in Conditionals, Conditionals and Loops
@subsection Conditional Blocks
@cindex conditional blocks
@cindex blocks, conditional

@DefescList {\@\{, , , }
@DefescListEnd {\@\}, , , }
@esindex \@{
@esindex \@}
@cindex begin of conditional block (@code{\@{})
@cindex end of conditional block (@code{\@}})
@cindex conditional block, begin (@code{\@{})
@cindex conditional block, end (@code{\@}})
@cindex block, conditional, begin (@code{\@{})
@cindex block, conditional, end (@code{\@}})
@cindex brace escape sequences (@code{\@{}, @code{\@}})
@cindex escape sequences, brace (@code{\@{}, @code{\@}})
@cindex opening brace escape sequence (@code{\@}})
@cindex closing brace escape sequence (@code{\@})}
@cindex brace escape sequence, opening (@code{\@})}
@cindex brace escape sequence, closing (@code{\@})}
It is frequently desirable for a control structure to govern more than
one request, macro call, text line, or a combination of the foregoing.
The opening and closing brace escape sequences @code{\@{} and @code{\@}}
perform such grouping; such @dfn{conditional blocks} can be nested.
Brace escape sequences outside of control structures have no meaning and
produce no output.

@code{\@{} should appear (after optional spaces and tabs) immediately
subsequent to the request's conditional expression.  @code{\@}} should
appear on a line with other occurrences of itself as necessary to match
@code{\@{} sequences.  It can be preceded by a control character,
spaces, and tabs.  Input after any quantity of @code{\@}} sequences on
the same line is processed only if all of the preceding conditions to
which they correspond are true.  Furthermore, a @code{\@}} closing the
body of a @code{while} request (discussed below) must be the last such
escape sequence on an input line.

@Example
A
.if 0 \@{ B
C
D
\@}E
F
    @result{} A F
@endExample

@Example
N
.if 1 \@{ O
.  if 0 \@{ P
Q
R\@} S\@} T
U
    @result{} N O U
@endExample

The above behavior may challenge the intuition; it was implemented to
retain compatibility with @acronym{AT&T} @code{troff}.  For clarity, it
is idiomatic to end input lines with @code{\@{}, followed by
@code{\@key{RET}} if desired to prevent the newline from being
interpreted as a blank text line, and to precede @code{\@}} on an input
line with nothing more than a control character, spaces, tabs, and other
instances of itself.

@Example
.de DEBUG
debug =
.ie \\$1 \@{\
ON,
development
\@}
.el \@{\
OFF,
production
\@}
version
..
.DEBUG 0
.br
.DEBUG 1
@endExample

Try omitting the @code{\@key{RET}}s from the foregoing example and see
how the output changes.  Remember that, as noted above, after a true
conditional (or after the @code{el} request if its counterpart @code{ie}
condition was false) the remainder of the input line is interpreted
@emph{as if it were on an input line by itself}.
@endDefesc

We can use @code{ie}, @code{el}, and conditional blocks to simulate the
multi-way ``switch'' or ``case'' control structures of other languages.
The following example is adapted from the @code{groff} @file{man}
package.  Indentation is used to clarify the logic.

@Example
.\" Simulate switch/case in roff.
.      ie '\\$2'1' .ds title General Commands\"
.el \@{.ie '\\$2'2' .ds title System Calls\"
.el \@{.ie '\\$2'3' .ds title Library Functions\"
.el \@{.ie '\\$2'4' .ds title Kernel Interfaces\"
.el \@{.ie '\\$2'5' .ds title File Formats\"
.el \@{.ie '\\$2'6' .ds title Games\"
.el \@{.ie '\\$2'7' .ds title Miscellaneous Information\"
.el \@{.ie '\\$2'8' .ds title System Management\"
.el \@{.ie '\\$2'9' .ds title Kernel Development\"
.el                .ds title \" empty
.\@}\@}\@}\@}\@}\@}\@}\@}
@endExample

@c ---------------------------------------------------------------------

@node while,  , if-else, Conditionals and Loops
@subsection while
@cindex while

@code{groff} provides a looping construct:@: the @code{while} request.
Its syntax matches the @code{if} request.

@cindex body, of a while request
@Defreq {while, cond-expr anything}
Evaluate the conditional expression @var{cond-expr}, and repeatedly
execute @var{anything} unless and until @var{cond-expr} evaluates false.
@var{anything}, which is often a conditional block, is referred to as
the @code{while} request's @dfn{body}.

@Example
.nr a 0 1
.while (\na < 9) \@{\
\n+a,
.\@}
\n+a
    @result{} 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
@endExample

@cindex @code{de} request, and @code{while}
GNU @code{troff} treats the body of a @code{while} request similarly to
that of a @code{de} request (albeit one not read in copy
mode@footnote{@xref{Copy Mode}.}), but stores it under an internal name
and deletes it when the loop finishes.  The operation of a macro
containing a @code{while} request can slow significantly if the
@code{while} body is large.  Each time the macro is executed, the
@code{while} body is parsed and stored again.

@Example
.de xxx
.  nr num 10
.  while (\\n[num] > 0) \@{\
.    \" many lines of code
.    nr num -1
.  \@}
..
@endExample

@cindex recursive macros
@cindex macros, recursive
@noindent
An often better solution---and one that is more portable, since
@acronym{AT&T} @code{troff} lacked the @code{while} request---is to
instead write a recursive macro.  It will be parsed only
once.@footnote{unless you redefine it}

@Example
.de yyy
.  if (\\n[num] > 0) \@{\
.    \" many lines of code
.    nr num -1
.    yyy
.  \@}
..
.
.de xxx
.  nr num 10
.  yyy
..
@endExample

@noindent
To prevent infinite loops, the default number of available recursion
levels is 1,000 or somewhat less.@footnote{``somewhat less'' because
things other than macro calls can be on the input stack}  You can
disable this protective measure, or raise the limit, by setting the
@code{slimit} register.  @xref{Debugging}.

As noted above, if a @code{while} body begins with a conditional block,
its closing brace must end an input line.

@Example
.if 1 \@{\
.  nr a 0 1
.  while (\n[a] < 10) \@{\
.    nop \n+[a]
.\@}\@}
    @error{} unbalanced brace escape sequences
@endExample
@endDefreq

@Defreq {break, }
@cindex @code{while} request, confusing with @code{br}
@cindex @code{break} request, in a @code{while} loop
@cindex @code{continue} request, in a @code{while} loop
Exit a @code{while} loop.  Do not confuse this request with a
typographical break or the @code{br} request.
@endDefreq

@Defreq {continue, }
Skip the remainder of a @code{while} loop's body, immediately starting
the next iteration.
@endDefreq


@c =====================================================================

@node Writing Macros, Page Motions, Conditionals and Loops, GNU troff Reference
@section Writing Macros
@cindex writing macros
@cindex macros, writing

A @dfn{macro} is a stored collection of text and control lines that can
be interpolated multiple times.  Use macros to define common operations.
Macros are called in the same way that requests are invoked.  While
requests exist for the purpose of creating macros, simply calling an
undefined macro, or interpolating it as a string, will cause it to be
defined as empty.  @xref{Identifiers}.

@Defreq {de, name [@Var{end}]}
Define a macro @var{name}, replacing the definition of any existing
request, macro, string, or diversion called @var{name}.  If
@var{name} already exists as an alias, the target of the alias is
redefined; recall @ref{Strings}.  GNU @code{troff} enters copy
mode,@footnote{@xref{Copy Mode}.} storing subsequent input lines as the
macro definition.  If the optional second argument is not specified, the
definition ends with the control line @samp{..} (two dots).
Alternatively, @var{end} identifies a macro whose call syntax at the
start of a control line ends the definition of @var{name}; @var{end} is
then called normally.  A macro definition must end in the same
conditional block (if any) in which it began (@pxref{Conditional
Blocks}).  Spaces or tabs are permitted after the control character in
the line containing this ending token (either @samp{.} or
@samp{@var{end}}), but a tab immediately after the token prevents its
recognition as the end of a macro definition.  The macro @var{end} can
be called with arguments.@footnote{While it is possible to define and
call a macro @samp{.}, you can't use it as an end macro: during a macro
definition, @samp{..} is never handled as calling @samp{.}, even if
@samp{.de @var{name} .} explicitly precedes it.}
@c
@c @Example
@c .de .
@c (dot macro)
@c ..
@c .
@c ..    \" This calls macro '.'!
@c .de m1 .
@c (m1 macro)
@c ..    \" This does not.
@c .m1
@c     @result{} (dot macro) (m1 macro)
@c @endExample

Here is a small example macro called @samp{P} that causes a break and
inserts some vertical space.  It could be used to separate paragraphs.

@Example
.de P
.  br
.  sp .8v
..
@endExample

We can define one macro within another.  Attempting to nest @samp{..}
naïvely will end the outer definition because the inner definition
isn't interpreted as such until the outer macro is later interpolated.
We can use an end macro instead.  Each level of nesting should use a
unique end macro.

An end macro need not be defined until it is called.  This fact enables
a nested macro definition to begin inside one macro and end inside
another.  Consider the following example.@footnote{Its structure is
adapted from, and isomorphic to, part of a solution by Tadziu Hoffman to
the problem of reflowing text multiple times to find an optimal
configuration for it.
@uref{https://lists.gnu.org/archive/html/groff/2008-12/msg00006.html}}

@Example
.de m1
.  de m2 m3
you
..
.de m3
Hello,
Joe.
..
.de m4
do
..
.m1
know?
.  m3
What
.m4
.m2
    @result{} Hello, Joe.  What do you know?
@endExample

@noindent
A nested macro definition @emph{can} be terminated with @samp{..} and
nested macros @emph{can} reuse end macros, but these control lines must
be escaped multiple times for each level of nesting.  The necessity of
this escaping and the utility of nested macro definitions will become
clearer when we employ macro parameters and consider the behavior of
copy mode in detail.
@endDefreq

@code{de} defines a macro that inherits the compatibility mode
enablement status of its context (@pxref{Implementation Differences}).
Often it is desirable to make a macro that uses @code{groff} features
callable from contexts where compatibility mode is on; for instance,
when writing extensions to a historical macro package.  To achieve this,
compatibility mode needs to be switched off while such a macro is
interpreted---without disturbing that state when it is finished.

@Defreq {de1, name [@Var{end}]}
The @code{de1} request defines a macro to be interpreted with
compatibility mode disabled.  When @var{name} is called, compatibility
mode enablement status is saved; it is restored when the call completes.
Observe the extra backlash before the interpolation of register
@samp{xxx}; we'll explore this subject in @ref{Copy Mode}.

@Example
.nr xxx 12345
.de aa
The value of xxx is \\n[xxx].
.  br
..
.de1 bb
The value of xxx is \\n[xxx].
..
.cp 1
.aa
    @error{} warning: register '[' not defined
    @result{} The value of xxx is 0xxx].
.bb
    @result{} The value of xxx is 12345.
@endExample
@endDefreq

@DefreqList {dei, name [@Var{end}]}
@DefreqListEndx {dei1, name [@Var{end}]}
The @code{dei} request defines a macro with its name and end
macro indirected through strings.  That is, it interpolates strings
named @var{name} and @var{end} before performing the definition.

The following examples are equivalent.

@Example
.ds xx aa
.ds yy bb
.dei xx yy
@endExample

@Example
.de aa bb
@endExample

The @code{dei1} request bears the same relationship to @code{dei} as
@code{de1} does to @code{de}; it temporarily turns compatibility mode
off when @var{name} is called.
@endDefreq

@DefreqList {am, name [@Var{end}]}
@DefreqItemx {am1, name [@Var{end}]}
@DefreqItemx {ami, name [@Var{end}]}
@DefreqListEndx {ami1, name [@Var{end}]}
@cindex appending to a macro (@code{am})
@cindex macro, appending to (@code{am})
@code{am} appends subsequent input lines to macro @var{name}, extending
its definition, and otherwise working as @code{de} does.

To make the previously defined @samp{P} macro set indented instead of
block paragraphs, add the necessary code to the existing macro.

@Example
.am P
.ti +5n
..
@endExample

The other requests are analogous to their @samp{de} counterparts.  The
@code{am1} request turns off compatibility mode during interpretation of
the appendment.  The @code{ami} request appends indirectly, meaning that
strings @var{name} and @var{end} are interpolated with the resulting
names used before appending.  The @code{ami1} request is similar to
@code{ami}, disabling compatibility mode during interpretation of the
appended lines.
@endDefreq

@pindex trace.tmac
Using @file{trace.tmac}, you can trace calls to @code{de},
@code{de1}, @code{am}, and @code{am1}.  You can also use the
@code{backtrace} request at any point desired to troubleshoot tricky
spots (@pxref{Debugging}).

@xref{Strings}, for the @code{als}, @code{rm}, and @code{rn} requests to
create an alias of, remove, and rename a macro, respectively.

@cindex object creation
Macro identifiers share their name space with requests, strings, and
diversions; see @ref{Identifiers}.  The @code{am}, @code{as}, @code{da},
@code{de}, @code{di}, and @code{ds} requests (together with their
variants) create a new object only if the name of the macro, diversion,
or string is currently undefined or if it is defined as a request;
normally, they modify the value of an existing object.  @xref{als,,the
description of the @code{als} request}, for pitfalls when redefining a
macro that is aliased.

@Defreq {return, [@Var{anything}]}
Exit a macro, immediately returning to the caller.  If called with an
argument @var{anything}, exit twice---the current macro and the macro
one level higher.  This is used to define a wrapper macro for
@code{return} in @file{trace.tmac}.
@endDefreq

@menu
* Parameters::
* Copy Mode::
@end menu

@c ---------------------------------------------------------------------

@node Parameters, Copy Mode, Writing Macros, Writing Macros
@subsection Parameters
@cindex parameters

Macro calls and string interpolations optionally accept a list of
arguments; recall @ref{Calling Macros}.  At the time such an
interpolation takes place, these @dfn{parameters} can be examined using
a register and a variety of escape sequences starting with @samp{\$}.
All such escape sequences are interpreted even in copy mode, a fact we
shall motivate and explain below (@pxref{Copy Mode}).

@Defreg {.$}
@cindex parameter count register (@code{.$})
The count of parameters available to a macro or string is kept in this
read-only register.  The @code{shift} request can change its value.
@endDefreg

Any individual parameter can be accessed by its position in the list of
arguments to the macro call, numbered from left to right starting at 1,
with one of the following escape sequences.

@DefescList {\\$, , n, }
@DefescItem {\\$, (, nn, }
@DefescListEnd {\\$, [, nnn, ]}
Interpolate the @var{n}th, @var{nn}th, or @var{nnn}th parameter.  The
first form expects only a single digit (1@leq{}@var{n}@leq{}9)), the
second two digits (01@leq{}@var{nn}@leq{}99)), and the third any
positive integer @var{nnn}.  Macros and strings accept an unlimited
number of parameters.
@endDefesc

@Defreq {shift, [@Var{n}]}
Shift the parameters @var{n} places (1@tie{}by default).  This is a
``left shift'': what was parameter@tie{}@var{i} becomes parameter
@math{@var{i}-@var{n}}.  The parameters formerly in positions 1
to@tie{}@var{n} are no longer available.  Shifting by a non-positive
amount performs no operation.  The register @code{.$} is adjusted
accordingly.
@endDefreq

@cindex copy mode, and macro parameters
@cindex mode, copy, and macro parameters
@cindex macro, parameters (@code{\$})
@cindex parameters, macro (@code{\$})
In practice, parameter interpolations are usually seen prefixed with an
extra escape character.  This is because the @code{\$} family of escape
sequences is interpreted even in copy mode.@footnote{If they were not,
parameter interpolations would be similar to command-line
parameters---fixed for the entire duration of a @code{roff} program's
run.  The advantage of interpolating @code{\$} escape sequences even in
copy mode is that they can interpolate different contents from one call
to the next, like function parameters in a procedural language.  The
additional escape character is the price of this power.}

@DefescList {\\$*, , , }
@DefescItemx {\\$@@, , , }
@DefescListEndx {\\$^, , , }
In some cases it is convenient to interpolate all of the parameters at
once (to pass them to a request, for instance).  The @code{\$*} escape
concatenates the parameters, separating them with spaces.  @code{\$@@}
is similar, concatenating the parameters, surrounding each with double
quotes and separating them with spaces.  If not in compatibility mode,
the interpolation depth of double quotes is preserved (@pxref{Calling
Macros}).  @code{\$^} interpolates all parameters as if they were
arguments to the @code{ds} request.

@Example
.de foo
. tm $1='\\$1'
. tm $2='\\$2'
. tm $*='\\$*'
. tm $@@='\\$@@'
. tm $^='\\$^'
..
.foo " This is a "test"
    @error{} $1=' This is a '
    @error{} $2='test"'
    @error{} $*=' This is a  test"'
    @error{} $@@='" This is a " "test""'
    @error{} $^='" This is a "test"'
@endExample

@code{\$*} is useful when writing a macro that doesn't need to
distinguish its arguments, or even to not interpret them; examples
include macros that produce diagnostic messages by wrapping the
@code{tm} or @code{ab} requests.  Use @code{\$@@} when writing a macro
that may need to shift its parameters and/or wrap a macro or request
that finds the count significant.  If in doubt, prefer @code{\$@@} to
@code{\$*}.  An application of @code{\$^} is seen in @file{trace.tmac},
which redefines some requests and macros for debugging purposes.
@endDefesc

@Defesc {\\$0, , , }
@cindex macro name register (@code{\$0})
@cindex @code{als} request, and @code{\$0}
Interpolate the name by which the macro being interpreted was called.
The @code{als} request can cause a macro to have more than one name.
Applying string interpolation to a macro does not change this name.

@Example
.de foo
.  tm \\$0
..
.als bar foo
.
.de aaa
.  foo
..
.de bbb
.  bar
..
.de ccc
\\*[foo]\\
..
.de ddd
\\*[bar]\\
..
.
.aaa
    @error{} foo
.bbb
    @error{} bar
.ccc
    @error{} ccc
.ddd
    @error{} ddd
@endExample
@endDefesc

@c ---------------------------------------------------------------------

@node Copy Mode, , Parameters, Writing Macros
@subsection Copy Mode
@cindex copy mode
@cindex copy mode
@cindex mode, copy
@cindex mode, copy

@cindex @code{\n}, when reading text for a macro
@cindex @code{\$}, when reading text for a macro
@cindex @code{\*}, when reading text for a macro
@cindex \@key{RET}, when reading text for a macro
When GNU @code{troff} processes certain requests, most importantly those
which define or append to a macro or string, it does so in @dfn{copy
mode}: it copies the characters of the definition into a dedicated
storage region, interpolating the escape sequences @code{\n}, @code{\g},
@code{\$}, @code{\*}, @code{\V}, and @code{\?} normally; interpreting
@code{\@key{RET}} immediately; discarding comments @code{\"} and
@code{\#}; interpolating the current leader, escape, or tab character
with @code{\a}, @code{\e}, and @code{\t}, respectively; and storing all
other escape sequences in an encoded form.

@cindex interpretation mode
@cindex mode, interpretation
The complement of copy mode---a @code{roff} formatter's behavior when
not defining or appending to a macro, string, or diversion---where all
macros are interpolated, requests invoked, and valid escape sequences
processed immediately upon recognition, can be termed
@dfn{interpretation mode}.

@Defesc {\\\\, , , }
The escape character, @code{\} by default, can escape itself.  This
enables you to control whether a given @code{\n}, @code{\g}, @code{\$},
@code{\*}, @code{\V}, or @code{\?} escape sequence is interpreted at the
time the macro containing it is defined, or later when the macro is
called.@footnote{Compare this to the @code{\def} and @code{\edef}
commands in @TeX{}.}

@Example
.nr x 20
.de y
.nr x 10
\&\nx
\&\\nx
..
.y
    @result{} 20 10
@endExample

You can think of @code{\\} as a ``delayed'' backslash; it is the escape
character followed by a backslash from which the escape character has
removed its special meaning.  Consequently, @samp{\\} is not an escape
sequence in the usual sense.  In any escape sequence @samp{\@var{X}}
that GNU @code{troff} does not recognize, the escape character is
ignored and @var{X} is output.  An unrecognized escape sequence causes
a warning in category @samp{escape}, with two exceptions---@samp{\\} is
the first.
@endDefesc

@cindex @code{\\}, when reading text for a macro
@Defesc {\\., , , }
@code{\.} escapes the control character.  It is similar to @code{\\} in
that it isn't a true escape sequence.  It is used to permit nested macro
definitions to end without a named macro call to conclude them.  Without
a syntax for escaping the control character, this would not be possible.

@Example
.de m1
foo
.
.  de m2
bar
\\..
.
..
.m1
.m2
    @result{} foo bar
@endExample

@noindent
The first backslash is consumed while the macro is read, and the second
is interpreted when macro @code{m1} is called.
@endDefesc

@code{roff} documents should not use the @code{\\} or @code{\.}
character sequences outside of copy mode; they serve only to obfuscate
the input.  Use @code{\e} to represent the escape character,
@code{\[rs]} to obtain a backslash glyph, and @code{\&} before @samp{.}
and @samp{'} where GNU @code{troff} expects them as control characters
if you mean to use them literally (recall @ref{Requests and Macros}).

Macro definitions can be nested to arbitrary depth.  The mechanics of
parsing the escape character have significant consequences for this
practice.

@Example
.de M1
\\$1
.  de M2
\\\\$1
.    de M3
\\\\\\\\$1
\\\\..
.    M3 hand.
\\..
.  M2 of
..
This understeer is getting
.M1 out
    @result{} This understeer is getting out of hand.
@endExample

Each escape character is interpreted twice---once in copy mode, when the
macro is defined, and once in interpretation mode, when the macro is
called.  As seen above, this fact leads to exponential growth in the
quantity of escape characters required to delay interpolation of
@code{\n}, @code{\g}, @code{\$}, @code{\*}, @code{\V}, and @code{\?} at
each nesting level, which can be daunting.  GNU @code{troff} offers a
solution.

@Defesc {\\E, , , }
@code{\E} represents an escape character that is not interpreted in copy
mode.  You can use it to ease the writing of nested macro definitions.

@Example
.de M1
.  nop \E$1
.  de M2
.    nop \E$1
.    de M3
.      nop \E$1
\\\\..
.    M3 better.
\\..
.  M2 bit
..
This vehicle handles
.M1 a
    @result{} This vehicle handles a bit better.
@endExample

Observe that because @code{\.} is not a true escape sequence, we can't
use @code{\E} to keep @samp{..} from ending a macro definition
prematurely.  If the multiplicity of backslashes complicates
maintenance, use end macros.

@code{\E} is also convenient to define strings containing escape
sequences that need to work when used in copy mode (for example, as
macro arguments), or which will be interpolated at varying macro nesting
depths.  We might define strings to begin and end superscripting
as follows.@footnote{These are lightly adapted from the @code{groff}
implementation of the @file{ms} macros.}

@Example
.ds @{ \v'-.9m\s'\En[.s]*7u/10u'+.7m'
.ds @} \v'-.7m\s0+.9m'
@endExample

When the @code{ec} request is used to redefine the escape character,
@code{\E} also makes it easier to distinguish the semantics of an escape
character from the other meaning(s) its character might have.  Consider
the use of an unusual escape character, @samp{-}.

@Example
.nr a 1
.ec -
.de xx
--na
..
.xx
    @result{} -na
@endExample

@noindent
This result may surprise you; some people expect @samp{1} to be output
since register @samp{a} has clearly been defined with that value.  What
has happened?  The robotic replacement of @samp{\} with @samp{-} has led
us astray.  You might recognize the sequence @samp{--} more readily with
the default escape character as @samp{\-}, the special character escape
sequence for the minus sign glyph.

@Example
.nr a 1
.ec -
.de xx
-Ena
..
.xx
    @result{} 1
@endExample
@endDefesc


@c =====================================================================

@node Page Motions, Drawing Requests, Writing Macros, GNU troff Reference
@section Page Motions
@cindex page motions
@cindex motions, page

@xref{Manipulating Spacing}, for a discussion of the most commonly used
request for vertical motion, @code{sp}, which spaces downward by one
vee.

@DefreqList {mk, [@Var{reg}]}
@DefreqListEndx {rt, [@Var{dist}]}
@cindex marking vertical page location (@code{mk})
@cindex page location, vertical, marking (@code{mk})
@cindex location, vertical, page, marking (@code{mk})
@cindex vertical page location, marking (@code{mk})
@cindex returning to marked vertical page location (@code{rt})
@cindex page location, vertical, returning to marked (@code{rt})
@cindex location, vertical, page, returning to marked (@code{rt})
@cindex vertical page location, returning to marked (@code{rt})
You can @dfn{mark} a location on a page for subsequent @dfn{return}.
@code{mk} takes an argument, a register name in which to store the
current page location.  If given no argument, it stores the location in
an internal register.  This location can be used later by the @code{rt}
or the @code{sp} requests (or the @code{\v} escape).

The @code{rt} request returns @emph{upward} to the location marked with
the last @code{mk} request.  If used with an argument, it returns to a
vertical position@tie{}@var{dist} from the top of the page (no previous
call to @code{mk} is necessary in this case).  The default scaling
indicator is @samp{v}.

If a page break occurs between a @code{mk} request and its matching
@code{rt} request, the @code{rt} request is silently ignored.

A simple implementation of a macro to set text in two columns follows.

@Example
.nr column-length 1.5i
.nr column-gap 4m
.nr bottom-margin 1m
.
.de 2c
.  br
.  mk
.  ll \\n[column-length]u
.  wh -\\n[bottom-margin]u 2c-trap
.  nr right-side 0
..
.
.de 2c-trap
.  ie \\n[right-side] \@{\
.    nr right-side 0
.    po -(\\n[column-length]u + \\n[column-gap]u)
.    \" remove trap
.    wh -\\n[bottom-margin]u
.  \@}
.  el \@{\
.    \" switch to right side
.    nr right-side 1
.    po +(\\n[column-length]u + \\n[column-gap]u)
.    rt
.  \@}
..
@endExample

Now let us apply our two-column macro.

@Example
.pl 1.5i
.ll 4i
This is a small test that shows how the
rt request works in combination with mk.

.2c
Starting here, text is typeset in two columns.
Note that this implementation isn't robust
and thus not suited for a real two-column
macro.
    @result{} This is a small test that shows how the
    @result{} rt request works in combination with mk.
    @result{}
    @result{} Starting  here,    isn't    robust
    @result{} text is typeset    and   thus  not
    @result{} in two columns.    suited  for   a
    @result{} Note that  this    real two-column
    @result{} implementation     macro.
@endExample
@endDefreq

Several escape sequences enable fine control of movement about the page.

@Defesc {\\v, @code{'}, expr, @code{'}}
@cindex vertical motion (@code{\v})
@cindex motion, vertical (@code{\v})
Vertically move the drawing position.  @var{expr} indicates the
magnitude of motion: positive is downward and and negative upward.  The
default scaling unit is @samp{v}.  The motion is relative to the current
drawing position unless @var{expr} begins with the boundary-relative
motion operator @samp{|}.  @xref{Numeric Expressions}.

Text processing continues at the new drawing position; usually, vertical
motions should be in balanced pairs to avoid a confusing page layout.

@code{\v} will not spring a vertical position trap.  This can be useful;
for example, consider a page bottom trap macro that prints a marker in
the margin to indicate continuation of a footnote.  @xref{Traps}.
@endDefesc

A few escape sequences that produce vertical motion are unusual.  They
are thought to originate early in AT&T @code{nroff} history to achieve
super- and subscripting by half-line motions on line printers and
teletypewriters before the phototypesetter made more precise positioning
available.  They are reckoned in ems---not vees---to maintain continuity
with their original purpose of moving relative to the size of the type
rather than the distance between text baselines (vees).@footnote{At the
@code{grops} defaults of 10-point type on 12-point vertical spacing, the
difference between half a vee and half an em can be subtle:@: large
spacings like @samp{.vs .5i} make it obvious.}

@DefescList {\\r, , , }
@DefescItemx {\\u, , , }
@DefescListEndx {\\d, , , }
Move upward@tie{}1@dmn{m}, upward@tie{}.5@dmn{m}, and
downward@tie{}.5@dmn{m}, respectively.
@endDefesc

@noindent
Let us see these escape sequences in use.

@Example
Obtain 100 cm\u3\d of \ka\d\092\h'|\nau'\r233\dU.
@endExample

In the foregoing we have paired @code{\u} and @code{\d} to typeset a
superscript, and later a full em negative (``reverse'') motion to place
a superscript above a subscript.  A numeral-width horizontal motion
escape sequence aligns the proton and nucleon numbers, while @code{\k}
marks a horizontal position to which @code{\h} returns so that we could
stack them.  (We shall discuss these horizontal motion escape sequences
presently.)  In serious applications, we often want to alter the type
size of the -scripts and to fine-tune the vertical motions, as the
@code{groff} @file{ms} package does with its super- and subscripting
string definitions.

@Defesc {\\h, @code{'}, expr, @code{'}}
@cindex inserting horizontal space (@code{\h})
@cindex horizontal space (@code{\h})
@cindex space, horizontal (@code{\h})
@cindex horizontal motion (@code{\h})
@cindex motion, horizontal (@code{\h})
Horizontally move the drawing position.  @var{expr} indicates the
magnitude of motion: positive is rightward and negative leftward.  The
default scaling unit is @samp{m}.  The motion is relative to the current
drawing position unless @var{expr} begins with the boundary-relative
motion operator @samp{|}.  @xref{Numeric Expressions}.
@endDefesc

The following string definition sets the @TeX{}
logo.@footnote{@xref{Strings}, for an explanation of the trailing
@samp{\"}.}

@Example
.ds TeX T\h'-.1667m'\v'.224m'E\v'-.224m'\h'-.125m'X\"
@endExample

There are a number of special-case escape sequences for horizontal
motion.

@Defesc {\\@key{SP}, , , }
@cindex space, unbreakable and unadjustable (@code{\@key{SP}})
@cindex unbreakable and unadjustable space (@code{\@key{SP}})
@cindex unadjustable and unbreakable space (@code{\@key{SP}})
@c We use the following notation in our man pages; Texinfo is bound to
@c the GNU Emacs dialect.
@esindex \@slanted{space}
Move right one word space.  (The input is a backslash followed by a
space.)  This escape sequence can be thought of as a non-adjustable,
unbreakable space.  Usually you want @code{\~} instead; see
@ref{Manipulating Filling and Adjustment}.
@endDefesc

@cindex thin space (@code{\|})
@cindex space, thin (@code{\|})
@Defesc {\\|, , , }
Move one-sixth @dmn{em} to the right on typesetting output devices.  If
a glyph named @samp{\|} is defined in the current font, its width is
used instead, even on terminal output devices.
@endDefesc

@cindex hair space (@code{\^})
@cindex space, hair (@code{\^})
@Defesc {\\^, , , }
Move one-twelfth @dmn{em} to the right on typesetting output devices.
If a glyph named @samp{\^} is defined in the current font, its width is
used instead, even on terminal output devices.
@endDefesc

@Defesc {\\0, , , }
@cindex space, width of a digit (numeral) (@code{\0})
@cindex digit-width space (@code{\0})
@cindex figure space (@code{\0})
@cindex numeral-width space (@code{\0})
Move right by the width of a numeral in the current font.
@endDefesc

Horizontal motions are not discarded at the end of an output line as
word spaces are.  @xref{Breaking}.

@DefescList {\\w, @code{'}, anything, @code{'}}
@DefregItemx {st}
@DefregItemx {sb}
@DefregItemx {rst}
@DefregItemx {rsb}
@DefregItemx {ct}
@DefregItemx {ssc}
@DefregListEndx {skw}
@cindex width escape (@code{\w})
Interpolate the width of @var{anything} in basic units.  This escape
sequence allows several properties of formatted output to be measured
without writing it out.

@Example
The length of the string 'abc' is \w'abc'u.
    @result{} The length of the string 'abc' is 72u.
@endExample

@cindex dummy environment, used by @code{\w} escape sequence
@cindex environment, dummy, used by @code{\w} escape sequence
@var{anything} is processed in a dummy environment:@: this means that
font and type size changes, for example, may occur within it without
affecting subsequent output.

@need 500
After each use, @code{\w} sets several registers.

@cindex CSTR@tie{}#54 errata
@cindex CSTR@tie{}#54 erratum, @code{sb} register
@cindex CSTR@tie{}#54 erratum, @code{st} register
@table @code
@item st
@itemx sb
The maximum vertical displacements of the text baseline above and below,
respectively.  The sign convention is opposite that of relative vertical
motions; that is, depth below the (original) baseline is negative.
These registers are incorrectly documented in the @acronym{AT&T}
@code{troff} manual as ``the highest and lowest extent of [the argument
to @code{\w}] relative to the baseline''.

@item rst
@itemx rsb
Like @code{st} and @code{sb}, but taking account of the heights and
depths of glyphs.  In other words, these registers store the highest and
lowest vertical positions attained by @var{anything}, doing what
@acronym{AT&T} @code{troff} documented @code{st} and @code{sb} as doing.

@item ct
Characterizes the geometry of glyphs occurring in @var{anything}.

@table @asis
@item 0
only short glyphs, no descenders or tall glyphs

@item 1
at least one descender

@item 2
at least one tall glyph

@item 3
at least one each of a descender and a tall glyph
@end table

@item ssc
The amount of horizontal space (possibly negative) that should be added
to the last glyph before a subscript.

@item skw
How far to right of the center of the last glyph in the @code{\w}
argument, the center of an accent from a roman font should be placed
over that glyph.
@end table
@endDefesc

@DefescList {\\k, , p, }
@DefescItem {\\k, (, ps, }
@DefescListEnd {\\k, [, position, ]}
@cindex saving horizontal input line position (@code{\k})
@cindex horizontal input line position, saving (@code{\k})
@cindex input line position, horizontal, saving (@code{\k})
@cindex position, horizontal input line, saving (@code{\k})
@cindex line, input, horizontal position, saving (@code{\k})
Store the current horizontal position in the @emph{input} line in a
register with the name @var{position} (one-character name@tie{}@var{p},
two-character name @var{ps}).  Use this, for example, to return to the
beginning of a string for highlighting or other decoration.
@endDefesc

@Defreg {hp}
@cindex horizontal input line position register (@code{hp})
@cindex input line, horizontal position, register (@code{hp})
@cindex position, horizontal, in input line, register (@code{hp})
@cindex line, input, horizontal position, register (@code{hp})
The current horizontal position at the input line.
@endDefreg

@Defreg {.k}
@cindex horizontal output line position register (@code{.k})
@cindex output line, horizontal position, register (@code{.k})
@cindex position, horizontal, in output line, register (@code{.k})
@cindex line, output, horizontal position, register (@code{.k})
A read-only register containing the current horizontal output position
(relative to the current indentation).
@endDefreg

@Defesc {\\o, @code{'}, abc, @code{'}}
@cindex overstriking glyphs (@code{\o})
@cindex glyphs, overstriking (@code{\o})
Overstrike glyphs @var{a}, @var{b}, @var{c}, @dots{}; the glyphs are
centered, and the resulting spacing is the largest width of the affected
glyphs.
@endDefesc

@Defesc {\\z, , c, }
@cindex zero-width printing (@code{\z}, @code{\Z})
@cindex printing, zero-width (@code{\z}, @code{\Z})
Print glyph @var{c} with zero width, i.e., without spacing.  Use this to
overstrike glyphs left-aligned.
@endDefesc

@Defesc {\\Z, @code{'}, anything, @code{'}}
@cindex zero-width printing (@code{\z}, @code{\Z})
@cindex printing, zero-width (@code{\z}, @code{\Z})
Save the drawing position, format @var{anything}, then restore it.  The
argument may not contain tabs or leaders.

An example of a strike-through macro follows.

@Example
.de ST
.nr ww \w'\\$1'
\Z@@\v'-.25m'\l'\\n[ww]u'@@\\$1
..
.
This is
.ST "a test"
an actual emergency!
@endExample
@endDefesc

@codequotebacktick off
@codequoteundirected off


@c =====================================================================

@node Drawing Requests, Traps, Page Motions, GNU troff Reference
@section Drawing Requests
@cindex drawing requests
@cindex requests for drawing

@code{gtroff} provides a number of ways to draw lines and other figures
on the page.  Used in combination with the page motion commands
(@pxref{Page Motions}), a wide variety of figures can be drawn.
However, for complex drawings these operations can be quite cumbersome,
and it may be wise to use graphic preprocessors like @code{gpic} or
@code{ggrn}.

All drawing is done via escape sequences.

@DefescList {\\l, @code{'}, l, @code{'}}
@DefescListEnd {\\l, @code{'}, lc, @code{'}}
@cindex drawing horizontal lines (@code{\l})
@cindex horizontal line, drawing (@code{\l})
@cindex line, horizontal, drawing (@code{\l})
Draw a line horizontally.  @var{l} is the length of the line to be
drawn.  If it is positive, start the line at the current location and
draw to the right; its end point is the new current location.  Negative
values are handled differently: The line starts at the current location
and draws to the left, but the current location doesn't move.

@var{l} can also be specified absolutely (i.e., with a leading
@samp{|}), which draws back to the beginning of the input line.  Default
scaling indicator is @samp{m}.

@cindex underscore glyph (@code{\[ru]})
@cindex glyph, underscore (@code{\[ru]})
@cindex line drawing glyph
@cindex glyph, for line drawing
The optional second parameter@tie{}@var{c} is a glyph to draw the line
with.  If this second argument is not specified, @code{gtroff} uses the
underscore glyph, @code{\[ru]}.

@cindex dummy character (@code{\&}), effect on @code{\l} escape
@cindex character, dummy (@code{\&}), effect on @code{\l} escape
To separate the two arguments (to prevent @code{gtroff} from
interpreting a drawing glyph as a scaling indicator if the glyph is
represented by a single character) use @code{\&}.

@Example
.de textbox
\[br]\\$*\[br]\l'|0\[rn]'\l'|0\[ul]'
..
@endExample

@noindent
The above works by outputting a box rule (a vertical line), then the
text given as an argument and then another box rule.  Finally, the
line-drawing escape sequences both draw from the current location to the
beginning of the @emph{input} line---this works because the line length
is negative, not moving the current point.
@endDefesc

@DefescList {\\L, @code{'}, l, @code{'}}
@DefescListEnd {\\L, @code{'}, lg, @code{'}}
@cindex drawing vertical lines (@code{\L})
@cindex vertical line drawing (@code{\L})
@cindex line, vertical, drawing (@code{\L})
@cindex line drawing glyph
@cindex glyph for line drawing
@cindex box rule glyph (@code{\[br]})
@cindex glyph, box rule (@code{\[br]})
Draw vertical lines.  Its parameters are similar to the @code{\l}
escape, except that the default scaling indicator is @samp{v}.  The
movement is downward for positive values, and upward for negative
values.  The default glyph is the box rule glyph, @code{\[br]}.  As with
the vertical motion escape sequences, text processing blindly continues
where the line ends.

@Example
This is a \L'3v'test.
@endExample

@noindent
Here is the result, produced with @code{grotty}.

@Example
This is a
          |
          |
          |test.
@endExample
@endDefesc

@Defesc {\\D, @code{'}, command arg @dots{}, @code{'}}
The @code{\D} escape provides a variety of drawing functions.  On
character devices, only vertical and horizontal lines are supported
within @code{grotty}; other devices may only support a subset of the
available drawing functions.

The default scaling indicator for all subcommands of @code{\D} is
@samp{m} for horizontal distances and @samp{v} for vertical ones.
Exceptions are @samp{\D'f @dots{}'} and @samp{\D't @dots{}'}, which use
@code{u} as the default, and @samp{\D'F@var{x} @dots{}'}, which
arguments are treated similar to the @code{defcolor} request.

@table @code
@item \D'l @var{dx} @var{dy}'
@cindex line, drawing (@samp{\D'l @dots{}'})
@cindex drawing a line (@samp{\D'l @dots{}'})
Draw a line from the current location to the relative point specified by
(@var{dx},@var{dy}), where positive values mean right and down,
respectively.  The end point of the line is the new current location.

The following example is a macro for creating a box around a text
string; for simplicity, the box margin is taken as a fixed value,
0.2@dmn{m}.

@Example
.de TEXTBOX
.  nr @@wd \w'\\$1'
\h'.2m'\
\h'-.2m'\v'(.2m - \\n[rsb]u)'\
\D'l 0 -(\\n[rst]u - \\n[rsb]u + .4m)'\
\D'l (\\n[@@wd]u + .4m) 0'\
\D'l 0 (\\n[rst]u - \\n[rsb]u + .4m)'\
\D'l -(\\n[@@wd]u + .4m) 0'\
\h'.2m'\v'-(.2m - \\n[rsb]u)'\
\\$1\
\h'.2m'
..
@endExample

@noindent
First, the width of the string is stored in register @code{@@wd}.  Then,
four lines are drawn to form a box, properly offset by the box margin.
The registers @code{rst} and @code{rsb} are set by the @code{\w} escape,
containing the largest height and depth of the whole string.

@item \D'c @var{d}'
@cindex circle, drawing (@samp{\D'c @dots{}'})
@cindex drawing a circle (@samp{\D'c @dots{}'})
Draw a circle with a diameter of@tie{}@var{d} with the leftmost point at
the current position.  After drawing, the current location is positioned
at the rightmost point of the circle.

@item \D'C @var{d}'
@cindex circle, solid, drawing (@samp{\D'C @dots{}'})
@cindex drawing a solid circle (@samp{\D'C @dots{}'})
@cindex solid circle, drawing (@samp{\D'C @dots{}'})
Draw a solid circle with the same parameters and behaviour as an
outlined circle.  No outline is drawn.

@item \D'e @var{x} @var{y}'
@cindex drawing an ellipse (@samp{\D'e @dots{}'})
@cindex ellipse, drawing (@samp{\D'e @dots{}'})
Draw an ellipse with a horizontal diameter of @var{x} and a vertical
diameter of @var{y} with the leftmost point at the current position.
After drawing, the current location is positioned at the rightmost point
of the ellipse.

@item \D'E @var{x} @var{y}'
@cindex ellipse, solid, drawing (@samp{\D'E @dots{}'})
@cindex drawing a solid ellipse (@samp{\D'E @dots{}'})
@cindex solid ellipse, drawing (@samp{\D'E @dots{}'})
Draw a solid ellipse with the same parameters and behaviour as an
outlined ellipse.  No outline is drawn.

@item \D'a @var{dx1} @var{dy1} @var{dx2} @var{dy2}'
@cindex arc, drawing (@samp{\D'a @dots{}'})
@cindex drawing an arc (@samp{\D'a @dots{}'})
Draw an arc clockwise from the current location through the two
specified relative locations (@var{dx1},@var{dy1}) and
(@var{dx2},@var{dy2}).  The coordinates of the first point are relative
to the current position, and the coordinates of the second point are
relative to the first point.  After drawing, the current position is
moved to the final point of the arc.

@item \D'~ @var{dx1} @var{dy1} @var{dx2} @var{dy2} @dots{}'
@cindex drawing a spline (@samp{\D'~ @dots{}'})
@cindex spline, drawing (@samp{\D'~ @dots{}'})
Draw a spline from the current location to the relative point
(@var{dx1},@var{dy1}) and then to (@var{dx2},@var{dy2}), and so on.  The
current position is moved to the terminal point of the drawn curve.

@item \D'f @var{n}'
@cindex gray shading (@samp{\D'f @dots{}'})
@cindex shading filled objects (@samp{\D'f @dots{}'})
Set the shade of gray to be used for filling solid objects
to@tie{}@var{n}; @var{n}@tie{}must be an integer between 0
and@tie{}1000, where 0 corresponds solid white and 1000 to solid black,
and values in between correspond to intermediate shades of gray.  This
applies only to solid circles, solid ellipses, and solid polygons.  By
default, a level of 1000 is used.

Nonintuitively, the current point is moved horizontally to the
right by@tie{}@var{n}.

@cindex @samp{\D'f @dots{}'} and horizontal motion quantum
Don't use this command!  It has the serious drawback that it is always
rounded to the next integer multiple of the horizontal motion quantum
(the value of the @code{hor} keyword in the @file{DESC} file).  Use
@code{\M} (@pxref{Colors}) or @samp{\D'Fg @dots{}'} instead.

@item \D'p @var{dx1} @var{dy1} @var{dx2} @var{dy2} @dots{}'
@cindex drawing a polygon (@samp{\D'p @dots{}'})
@cindex polygon, drawing (@samp{\D'p @dots{}'})
Draw a polygon from the current location to the relative position
(@var{dx1},@var{dy1}) and then to (@var{dx2},@var{dy2}) and so on.  When
the specified data points are exhausted, a line is drawn back to the
starting point.  The current position is changed by adding the sum of
all arguments with odd index to the actual horizontal position and the
even ones to the vertical position.

@item \D'P @var{dx1} @var{dy1} @var{dx2} @var{dy2} @dots{}'
@cindex polygon, solid, drawing (@samp{\D'P @dots{}'})
@cindex drawing a solid polygon (@samp{\D'P @dots{}'})
@cindex solid polygon, drawing (@samp{\D'P @dots{}'})
Draw a solid polygon with the same parameters and behaviour as an
outlined polygon.  No outline is drawn.

Here a better variant of the box macro to fill the box with some color.
The box must be drawn before the text since colors in GNU @code{troff}
are not transparent; the filled polygon would hide the text completely.

@Example
.de TEXTBOX
.  nr @@wd \w'\\$1'
\h'.2m'\
\h'-.2m'\v'(.2m - \\n[rsb]u)'\
\M[lightcyan]\
\D'P 0 -(\\n[rst]u - \\n[rsb]u + .4m) \
     (\\n[@@wd]u + .4m) 0 \
     0 (\\n[rst]u - \\n[rsb]u + .4m) \
     -(\\n[@@wd]u + .4m) 0'\
\h'.2m'\v'-(.2m - \\n[rsb]u)'\
\M[]\
\\$1\
\h'.2m'
..
@endExample

If you want a filled polygon that has exactly the same size as an
unfilled one, you must draw both an unfilled and a filled polygon.  A
filled polygon is always smaller than an unfilled one because the latter
uses straight lines with a given line thickness to connect the polygon's
corners, while the former simply fills the area defined by the
coordinates.

@Example
\h'1i'\v'1i'\
\# increase line thickness
\Z'\D't 5p''\
\# draw unfilled polygon
\Z'\D'p 3 3 -6 0''\
\# draw filled polygon
\Z'\D'P 3 3 -6 0''
@endExample

@item \D't @var{n}'
@cindex line thickness (@samp{\D't @dots{}'})
@cindex thickness of lines (@samp{\D't @dots{}'})
Set the current line thickness to @var{n}@tie{}basic units.  A value
of zero selects the smallest available line thickness.  A negative value
makes the line thickness proportional to the current type size (this is
the default behaviour of @acronym{AT&T} @code{troff}).

Nonintuitively, the current point is moved horizontally to the
right by@tie{}@var{n}.

@item \D'F@var{scheme} @var{color_components}'
@cindex unnamed fill colors (@code{\D'F@dots{}'})
@cindex fill colors, unnamed (@code{\D'F@dots{}'})
@cindex colors, fill, unnamed (@code{\D'F@dots{}'})
Change current fill color.  @var{scheme} is a single letter denoting the
color scheme: @samp{r} (rgb), @samp{c} (cmy), @samp{k} (cmyk), @samp{g}
(gray), or @samp{d} (default color).  The color components use exactly
the same syntax as in the @code{defcolor} request (@pxref{Colors}); the
command @code{\D'Fd'} doesn't take an argument.

@emph{No} position changing!

Examples:

@Example
\D'Fg .3'      \" same gray as \D'f 700'
\D'Fr #0000ff' \" blue
@endExample
@end table
@endDefesc

@xref{Graphics Commands}.

@codequotebacktick on
@codequoteundirected on

@Defesc {\\b, @code{'}, contents, @code{'}}
@cindex pile, glyph (@code{\b})
@cindex glyph pile (@code{\b})
@cindex stacking glyphs (@code{\b})
Pile and center a sequence of glyphs vertically on the output line.
@dfn{Piling} vertically stacks glyphs corresponding to each character in
@var{contents}, read from left to right, and placed from top to bottom.
GNU @code{troff} separates the glyphs vertically by 1@dmn{m}, and the
pile itself is centered 0.5@dmn{m} above the text baseline.  The
horizontal drawing position is then advanced by the width of the widest
glyph in the pile.

@cindex @code{\b}, limitations of
@cindex limitations of @code{\b} escape sequence
This rather inflexible positioning algorithm doesn't work with the
@code{dvi} output device since its bracket pieces vary in height.
Instead, use the @code{geqn} preprocessor.

@xref{Manipulating Spacing}, to see how to adjust the vertical spacing
of the output line with the @code{\x} escape sequence.

The idiomatic use of @code{\b} is for building large brackets and
braces, hence its name.  We might construct a large opening brace as
follows.

@Example
\b'\[lt]\[bv]\[lk]\[bv]\[lb]'
@endExample
@endDefesc


@c =====================================================================

@node Deferring Output, Traps, Drawing Requests, GNU troff Reference
@section Deferring Output
@cindex deferred output

@cindex environment
@cindex diversion
@cindex trap
A few @code{roff} language elements are generally not used in simple
documents, but arise as page layouts become more sophisticated and
demanding.  @dfn{Environments} collect formatting parameters like line
length and typeface.  A @dfn{diversion} stores formatted output for
later use.  A @dfn{trap} is a condition on the input or output, tested
automatically by the formatter, that is associated with a macro, causing
it to be called when that condition is fulfilled.

Footnote support often exercises all three of the foregoing features.  A
simple implementation might work as follows.  A pair of macros is
defined: one starts a footnote and the other ends it.  The author calls
the first macro where a footnote marker is desired.  The macro
establishes a diversion so that the footnote text is collected at the
place in the body text where its corresponding marker appears.  An
environment is created for the footnote so that it is set at a smaller
typeface.  The footnote text is formatted in the diversion using that
environment, but it does not yet appear in the output.  The document
author calls the footnote end macro, which returns to the previous
environment and ends the diversion.  Later, after much more body text in
the document, a trap, set a small distance above the page bottom, is
sprung.  The macro called by the trap draws a line across the page and
emits the stored diversion.  Thus, the footnote is rendered.

Diversions and traps make the text formatting process non-linear.  Let
us imagine a set of text lines or paragraphs labelled @samp{A},
@samp{B}, and so on.  If we set up a trap that produces text @samp{T}
(as a page footer, say), and we also use a diversion to store the
formatted text @samp{D}, then a document with input text in the order
@samp{A B C D E F} might render as @samp{A B C E T F}.  The diversion
@samp{D} will never be output if we do not call for it.

Environments of themselves are not a source of non-linearity in document
formatting:@: environment switches have immediate effect.  One could
always write a macro to change as many formatting parameters as desired
with a single convenient call.  But because diversions can be nested and
macros called by traps that are sprung by other trap-called macros, they
may be called upon in varying contexts.  For example, consider a page
header that is always to be set in Helvetica.  A document that uses
Times for most of its body text, but Courier for displayed code
examples, poses a challenge if a page break occurs in the middle of a
code display; if the header trap assumes that the ``previous font'' is
always Times, the rest of the example will be formatted in the wrong
typeface.  One could carefully save all formatting parameters upon
entering the trap and restore them upon leaving it, but this is verbose,
error-prone, and not future-proof as the @code{groff} language develops.
Environments save us considerable effort.

@c =====================================================================

@c BEGIN Keep (roughly) parallel with subsection "Traps" of groff(7).
@node Traps, Diversions, Deferring Output, GNU troff Reference
@section Traps
@cindex traps

@dfn{Traps} are locations in the output or conditions on the input that,
when reached or fulfilled, call a specified macro.  These traps can
occur at a given location on the page, at a given location in the
current diversion (together, these are known as @slanted{vertical
position traps}), at a blank line, at a line with leading space
characters, after a quantity of input lines, or at the end of input.
Macros called by traps are passed no arguments.
@cindex planting a trap
@cindex trap, planting
Setting a trap is also called @dfn{planting} one.
@cindex trap, springing
@cindex springing a trap
It is said that a trap is @dfn{sprung} if its condition is fulfilled.
@c END Keep (roughly) parallel with subsection "Traps" of groff(7).

@menu
* Vertical Position Traps::
* Diversion Traps::
* Input Line Traps::
* Blank Line Traps::
* Leading Space Traps::
* End-of-input Traps::
@end menu

@c ---------------------------------------------------------------------

@node Vertical Position Traps, Page Location Traps, Traps, Traps
@subsection Vertical Position Traps
@cindex vertical position traps
@cindex traps, vertical position

@dfn{Vertical position traps} perform an action when GNU @code{troff}
reaches or passes a certain vertical location on the output page or in a
diversion.  Their applications include setting page headers and footers,
body text in multiple columns, and footnotes.

@DefreqList {vpt, [@Var{flag}]}
@DefregListEndx {.vpt}
@cindex enabling vertical position traps (@code{vpt})
@cindex vertical position traps, enabling (@code{vpt})
@cindex vertical position trap enable register (@code{.vpt})
Enable vertical position traps if @var{flag} is non-zero or absent;
disable them otherwise.  Vertical position traps are those set by the
@code{wh} request or by @code{dt} within a diversion.  The parameter
that controls whether vertical position traps are enabled is global.
Initially, vertical position traps are enabled.  The current value is
stored in the @code{.vpt} read-only register.

A page can't be ejected if @code{vpt} is set to zero.
@endDefreq

@menu
* Page Location Traps::
* Diversion Traps::
@end menu

@c ---------------------------------------------------------------------

@node Page Location Traps, Diversion Traps, Vertical Position Traps, Vertical Position Traps
@subsubsection Page Location Traps
@cindex page location traps
@cindex traps, page location

@Defreq {wh, dist [@Var{name}]}
Call macro @var{name} when the vertical position @var{dist} on the page
is reached or passed in the downward direction.  The default scaling
unit is @samp{v}.  Non-negative values for @var{dist} set the trap
relative to the top of the page; negative values set the trap relative
to the bottom of the page.  An existing @emph{visible} trap (see below)
at @var{dist} is removed; this is @code{wh}'s sole function if
@var{name} is missing.

A trap is sprung only if it is @dfn{visible}, meaning that its location
is reachable on the page@footnote{A trap planted at @samp{20i} or
@samp{-30i} will not be sprung on a page of length @samp{11i}.} and it
is not hidden by another trap at the same location already planted
there.

@cindex page headers
@cindex page footers
@cindex headers
@cindex footers
An example of how a macro package might set headers and footers follows.

@Example
.de hd                \" page header
'  sp .5i
.  tl '\\*[Title]''\\*[Date]'
'  sp .3i
..
.
.de fo                \" page footer
'  sp 1v
.  tl ''%''
'  bp
..
.
.wh 0   hd            \" trap at top of the page
.wh -1i fo            \" trap one inch from bottom
@endExample

A trap above the top or at or below the bottom of the page can be made
visible by either moving it into the page area or increasing the page
length so that the trap is on the page.  Negative trap values always use
the @emph{current} page length; they are not converted to an absolute
vertical position.
@cindex page location traps, debugging
@cindex debugging page location traps
We can use the @code{ptr} request to dump our page location traps to the
standard error stream (@pxref{Debugging}).  Their positions are reported
in basic units appropriate to the device; an @code{nroff} device example
follows.

@Example
.pl 5i
.wh -1i xx
.ptr
    @error{} xx      -240
.pl 100i
.ptr
    @error{} xx      -240
@endExample

It is possible to have more than one trap at the same location (although
only one at a time can be visible); to achieve this, the traps must be
defined at different locations, then moved to the same place with the
@code{ch} request.  In the following example, the many empty lines
caused by the @code{bp} request are not shown in the output.

@Example
.de a
.  nop a
..
.de b
.  nop b
..
.de c
.  nop c
..
.
.wh 1i a
.wh 2i b
.wh 3i c
.bp
    @result{} a b c
@endExample
@Example
.ch b 1i
.ch c 1i
.bp
    @result{} a
@endExample
@Example
.ch a 0.5i
.bp
    @result{} a b
@endExample
@endDefreq

@Defreg {.t}
@cindex distance to next vertical position trap register (@code{.t})
@cindex trap, distance to next vertical position, register (@code{.t})
The read-only register @code{.t} holds the distance to the next vertical
position trap.  If there are no traps between the current position and
the bottom of the page, it contains the distance to the page bottom.
Within a diversion, in the absence of a diversion trap, this distance is
the largest representable integer in basic units---effectively infinite.
@endDefreg

@Defreq {ch, name [@Var{dist}]}
@cindex changing trap location (@code{ch})
@cindex trap, changing location (@code{ch})
Change the location of a trap by moving macro @var{name} to new location
@var{dist}, or by unplanting it altogether if @var{dist} is absent.  The
default scaling unit is @samp{v}.  Parameters to @code{ch} are specified
in the opposite order from @code{wh}.  If @var{name} is the earliest
planted macro of multiple traps at the same location, (re)moving it from
that location exposes the macro next least recently planted at the same
place.@footnote{It may help to think of each trap location as
maintaining a queue; @code{wh} operates on the head of the queue, and
@code{ch} operates on its tail.  Only the trap at the head of the queue
is visible.}

Changing a trap's location is useful for building up footnotes in a
diversion to allow more space at the bottom of the page for them.

@c XXX

@ignore
@Example
... (simplified) footnote example ...
@endExample
@end ignore
@endDefreq

The same macro can be installed simultaneously at multiple locations;
however, only the earliest-planted instance---that has not yet been
deleted with @code{wh}---will be moved by @code{ch}.  The following
example (using an @code{nroff} device) illustrates this
behavior.@footnote{@dots{}which is compatible with Heirloom Doctools
@code{troff}.}  Blank lines have been elided from the output.

@Example
.de T
Trap sprung at \\n(nlu.
.br
..
.wh 1i T
.wh 2i T
foo
.sp 11i
.bp
.ch T 4i
bar
.sp 11i
.bp
.ch T 5i
baz
.sp 11i
.bp
.wh 5i
.ch T 6i
qux
.sp 11i
@endExample
@Example
    @result{} foo
    @result{} Trap sprung at 240u.
    @result{} Trap sprung at 480u.
    @result{} bar
    @result{} Trap sprung at 480u.
    @result{} Trap sprung at 960u.
    @result{} baz
    @result{} Trap sprung at 480u.
    @result{} Trap sprung at 1200u.
    @result{} qux
    @result{} Trap sprung at 1440u.
@endExample

@Defreg {.ne}
The read-only register @code{.ne} contains the amount of space that was
needed in the last @code{ne} request that caused a trap to be sprung;
it is useful in conjunction with the @code{.trunc} register.  @xref{Page
Control}.

Since the @code{.ne} register is set only by traps, it doesn't make
sense to interpolate it outside of macros called by traps.
@endDefreg

@Defreg {.trunc}
@cindex @code{ne} request, and the @code{.trunc} register
@cindex truncated vertical space register (@code{.trunc})
A read-only register containing the amount of vertical space truncated
from an @code{sp} request by the most recently sprung vertical
position trap, or, if the trap was sprung by an @code{ne} request,
minus the amount of vertical motion produced by the @code{ne}
request.  In other words, at the point a trap is sprung, it
represents the difference of what the vertical position would have
been but for the trap, and what the vertical position actually is.

Since the @code{.trunc} register is set only by traps, it doesn't make
sense to interpolate it outside of macros called by traps.
@endDefreg

@Defreg {.pe}
@cindex @code{bp} request, and traps (@code{.pe})
@cindex traps, sprung by @code{bp} request (@code{.pe})
@cindex page ejecting register (@code{.pe})
A read-only register containing@tie{}1 while a page is being ejected
with the @code{bp} request (or by the end of input), and
0@tie{}otherwise.

In the following example, only the second call to@tie{}@code{x} is
caused by @code{bp}.

@Example
.de x
\&.pe=\\n[.pe]
.br
..
.wh 1v x
.wh 4v x
A line.
.br
Another line.
.br
    @result{} A line.
       .pe=0
       Another line.

       .pe=1
@endExample
@endDefreg

@cindex diversions, and traps
@cindex traps, and diversions
An important fact to consider while designing macros is that diversions
and traps do not interact normally.  For example, if a trap calls a
header macro (while outputting a diversion) that tries to change the
font on the current page, the effect is not visible before the diversion
has completely been printed (except for input protected with @code{\!}
or @code{\?}) since the data in the diversion is already formatted.  In
most cases, this is not the expected behaviour.

@c ---------------------------------------------------------------------

@node Diversion Traps, Input Line Traps, Page Location Traps, Vertical Position Traps
@subsubsection Diversion Traps
@cindex diversion traps
@cindex traps, diversion

@Defreq {dt, [@Var{dist} @Var{name}]}
@cindex @code{.t} register, and diversions
@cindex setting diversion trap (@code{dt})
@cindex diversion trap, setting (@code{dt})
@cindex trap, diversion, setting (@code{dt})
Set a trap @emph{within} a diversion at location @var{dist}, which is
interpreted relative to diversion rather than page boundaries.  There
exists only a single diversion trap per diversion.  If invoked with
fewer than two arguments, any diversion trap in the current diversion is
removed.  The register @code{.t} works within diversions.  It is an
error to invoke @code{dt} in the top-level diversion.
@xref{Diversions}.
@endDefreq

@c ---------------------------------------------------------------------

@node Input Line Traps, Blank Line Traps, Diversion Traps, Traps
@subsection Input Line Traps
@cindex input line traps
@cindex traps, input line

@DefreqList {it, [@Var{n} @Var{name}]}
@DefreqListEndx {itc, [@Var{n} @Var{name}]}
@cindex setting input line trap (@code{it}, @code{itc})
@cindex input line trap, setting (@code{it}, @code{itc})
@cindex trap, input line, setting (@code{it}, @code{itc})
@cindex clearing input line trap (@code{it}, @code{itc})
@cindex input line trap, clearing (@code{it}, @code{itc})
@cindex trap, input line, clearing (@code{it}, @code{itc})
Set an input line trap, calling macro @var{name} after processing the
next @var{n}@tie{}productive input lines (@pxref{Manipulating Filling
and Adjustment}).  Any existing input line trap in the environment is
replaced.  Without arguments, @code{it} and @code{itc} clear any
input line trap that has not yet sprung.

Consider a macro @samp{.ST @var{s n}} which sets the next
@var{n}@tie{}input lines in the font style@tie{}@var{s}.

@Example
.de ST \" Use style $1 for next $2 text lines.
.  it \\$2 ES
.  ft \\$1
..
.de ES \" end ST
.  ft R
..
.ST I 1
oblique
face
.ST I 1
oblique\c
face
    @result{} @i{oblique} face @i{oblique}face @
@r{(second ``face'' upright)}
@endExample

@cindex input line traps and interrupted lines (@code{itc})
@cindex interrupted lines and input line traps (@code{itc})
@cindex traps, input line, and interrupted lines (@code{itc})
@cindex lines, interrupted, and input line traps (@code{itc})
Unlike the @code{ce} and @code{rj} requests, @code{it} counts lines
interrupted with the @code{\c} escape sequence separately (@pxref{Line
Continuation}); @code{itc} does not.  To see the difference, let's
change the previous example to use @code{itc} instead.

@Example
@r{@dots{}}
.  itc \\$2 ES
@r{@dots{}}
    @result{} @i{oblique} face @i{obliqueface} @
@r{(second ``face'' oblique)}
@endExample

You can think of the @code{ce} and @code{rj} requests as implicitly
creating an input line trap with @code{itc} that schedules a break when
the trap is sprung.

@Example
.de BR
.  br
.  @slanted{internal: disable centering-without-filling}
..
.
.de ce
.  if \\n[.br] .br
.  itc \\$1 BR
.  @slanted{internal: enable centering-without-filling}
..
@endExample

@need 500
Let us consider in more detail the sorts of input lines that are or are
not ``productive''.

@Example
.de Trap
TRAP SPRUNG
..
.de Mac
.if r a \l'5n'
..
.it 2 Trap
.
foo
.Mac
bar
baz
.it 1 Trap
.sp \" moves, but does not write or draw
qux
.itc 1 Trap
\h'5n'\c \" moves, but does not write or draw
jat
@endExample

@noindent
When @samp{Trap} gets called depends on whether the @samp{a} register is
defined; the control line with the @code{if} request may or may not
produce written output.  We also see that the spacing request @code{sp},
while certainly affecting the output, does not spring the input line
trap.  Similarly, the horizontal motion escape sequence @code{\h} also
affected the output, but was not ``written''.  Observe that we had to
follow it with @code{\c} and use @code{itc} to prevent the newline at
the end of the text line from causing a word break, which, like an
ordinary space character, counts as written output.

@Example
$ groff -Tascii input-trap-example.groff
    @result{} foo bar TRAP SPRUNG baz
    @result{}
    @result{} qux TRAP SPRUNG      jat TRAP SPRUNG
$ groff -Tascii -ra1 input-trap-example.groff
    @result{} foo _____ TRAP SPRUNG bar baz
    @result{}
    @result{} qux TRAP SPRUNG      jat TRAP SPRUNG
@endExample
@endDefreq

Input line traps are associated with the environment
(@pxref{Environments}); switching to another environment suspends the
current input line trap, and going back resumes it, restoring the count
of qualifying lines enumerated in that environment.

@c ---------------------------------------------------------------------

@node Blank Line Traps, Leading Space Traps, Input Line Traps, Traps
@subsection Blank Line Traps
@cindex blank line traps
@cindex traps, blank line

@Defreq {blm, [@Var{name}]}
@cindex blank line macro (@code{blm})
Set a blank line trap, calling the macro @var{name} when GNU
@code{troff} encounters a blank line in an input file, instead of the
usual behavior (@pxref{Breaking}).  A line consisting only of spaces is
also treated as blank and subject to this trap.  If no argument is
supplied, the default blank line behavior is (re-)established.
@endDefreq

@c ---------------------------------------------------------------------

@node Leading Space Traps, End-of-input Traps, Blank Line Traps, Traps
@subsection Leading Space Traps
@cindex leading space traps
@cindex traps, leading space

@DefreqList {lsm, [@Var{name}]}
@DefregItemx {lsn}
@DefregListEndx {lss}
@cindex leading spaces macro (@code{lsm})
Set a leading space trap, calling the macro @var{name} when GNU
@code{troff} encounters leading spaces in an input line; the implicit
line break that normally happens in this case is suppressed.  If no
argument is supplied, the default leading space behavior is
(re-)established (@pxref{Breaking}).

The count of leading spaces on an input line is stored in register
@code{lsn}, and the amount of corresponding horizontal motion in
register @code{lss}, irrespective of whether a leading space trap is
set.  When it is, the leading spaces are removed from the input line,
and no motion is produced before calling @var{name}.

@c XXX The following discussion does not seem to be correct; leading
@c space traps don't "see" _any_ input tokens.  Nothing on the line is
@c passed to it as arguments, and tokens after leading spaces are
@c processed normally after the designated macro is interpolated. XXX
@c
@c The first thing a leading space macro sees is a token.  However, some
@c escape sequences, like @code{\f} and @code{\m}, are handled on the
@c fly (@pxref{Gtroff Internals} for a complete list) without creating a
@c token at all.  Consider a line that starts with two spaces followed
@c by @samp{\fIfoo}.  After skipping the spaces, @samp{\fI} is handled
@c as well such that @code{groff}'s current font is set to @code{I}, but
@c the leading space macro sees only @samp{foo} without the preceding
@c @samp{\fI}.  If the macro should see the font escape, you have to
@c ``protect'' it with something that creates a token, like the
@c dummy character; for example, @samp{\&\fIfoo}.
@endDefreq

@c ---------------------------------------------------------------------

@node End-of-input Traps,  , Leading Space Traps, Traps
@subsection End-of-input Traps
@cindex end-of-input traps
@cindex traps, end-of-input

@Defreq {em, [@Var{name}]}
@cindex setting end-of-input trap (@code{em})
@cindex end-of-input trap, setting (@code{em})
@cindex trap, end-of-input, setting (@code{em})
@cindex end-of-input macro (@code{em})
@cindex macro, end-of-input (@code{em})
Set a trap at the end of input, calling macro @var{name} after the last
line of the last input file has been processed.  If no argument is
given, any existing end-of-input trap is removed.

For example, if the document had to have a section at the bottom of the
last page for someone to approve it, the @code{em} request could be
used.

@Example
.de approval
\c
.  ne 3v
.  sp (\\n[.t]u - 3v)
.  in +4i
.  lc _
.  br
Approved:\t\a
.  sp
Date:\t\t\a
..
.
.em approval
@endExample

The @code{\c} in the above example needs explanation.  For historical
reasons (compatibility with @acronym{AT&T} @code{troff}), the
end-of-input macro exits as soon as it causes a page break if no
partially collected line remains.@footnote{While processing an
end-of-input macro, the formatter assumes that the next page break must
be the last; it goes into ``sudden death overtime''.}

Let us assume that there is no @code{\c} in the above @code{approval}
macro, that the page is full, and last output line has been broken with,
say, a @code{br} request.  Because there is no more room, a @code{ne}
request at this point causes a page ejection, which in turn makes
@code{troff} exit immediately as just described.  In most situations,
this is not desired; people generally want to format the input after
@code{ne}.

To force processing of the whole end-of-input macro independently of
this behavior, it is thus advisable to (invisibly) ensure the existence
of a partially collected line (@code{\c}) whenever there is a chance
that a page break can happen.  In the above example, invoking the
@code{ne} request ensures that there is room for the subsequent
formatted output on the same page, so we need insert @code{\c} only
once.

The next example shows how to append three lines, then start a new page
unconditionally.  Since @w{@samp{.ne 1}} doesn't give the desired
effect---there is always one line available or we are already at the
beginning of the next page---we temporarily increase the page length by
one line so that we can use @w{@samp{.ne 2}}.

@Example
.de EM
.pl +1v
\c
.ne 2
line one
.br
\c
.ne 2
line two
.br
\c
.ne 2
line three
.br
.pl -1v
\c
'bp
..
.em EM
@endExample

This specific feature affects only the first potential page break caused
by the end-of-input macro; further page breaks emitted by the macro are
handled normally.

Another possible use of the @code{em} request is to make GNU
@code{troff} emit a single large page instead of multiple pages.  For
example, one may want to produce a long plain text file for reading
in a terminal or emulator without page footers and headers interrupting
the body of the document.  One approach is to set the page length at the
beginning of the document to a very large value to hold all the
text,@footnote{Another, taken by the @code{groff} @code{man} macros, is
to intercept @code{ne} requests and wrap @code{bp} ones.} and
automatically adjust it to the exact height of the document after the
text has been output.

@Example
.de adjust-page-length
.  br
.  pl \\n[nl]u \" \n[nl]: current vertical position
..
.
.de single-page-mode
.  pl 99999
.  em adjust-page-length
..
.
.\" Activate the above code if configured.
.if \n[do-continuous-rendering] \
.  single-page-mode
@endExample

Since only one end-of-input trap exists and another macro package may
already use it, care must be taken not to break the mechanism.  A simple
solution would be to append the above macro to the macro package's
end-of-input macro using the @code{am} request.
@endDefreq


@c =====================================================================

@c BEGIN Keep (roughly) parallel with subsection "Diversions" of
@c groff(7).
@node Diversions, Punning Names, Traps, GNU troff Reference
@section Diversions
@cindex diversions

In @code{roff} systems it is possible to format text as if for output,
but instead of writing it immediately, one can @dfn{divert} the
formatted text into a named storage area.  It is retrieved later by
specifying its name after a control character.   The same name space is
used for such @slanted{diversions} as for strings and macros; see
@ref{Identifiers}.  Such text is sometimes said to be ``stored in a
macro'', but this coinage obscures the important distinction between
macros and strings on one hand and diversions on the other; the former
store @emph{unformatted} input text, and the latter capture
@emph{formatted} output.  Diversions also do not interpret arguments.
Applications of diversions include ``keeps'' (preventing a page break
from occurring at an inconvenient place by forcing a set of output lines
to be set as a group), footnotes, tables of contents, and indices.
@cindex top-level diversion
@cindex diversion, top-level
For orthogonality it is said that GNU @code{troff} is in the
@dfn{top-level diversion} if no diversion is active (that is, formatted
output is being ``diverted'' immediately to the output device).

Dereferencing an undefined diversion will create an empty one of that
name and cause a warning in category @samp{mac} to be emitted.
@xref{Warnings}, for information about the enablement and suppression of
warnings.  A diversion does not exist for the purpose of testing with
the @code{d} conditional operator until its initial definition ends
(@pxref{Operators in Conditionals}).  The following requests are used to
create and alter diversions.
@c END Keep (roughly) parallel with subsection "Diversions" of groff(7).

@DefreqList {di, [@Var{name}]}
@DefreqListEndx {da, [@Var{name}]}
@cindex beginning diversion (@code{di})
@cindex diversion, beginning (@code{di})
@cindex ending diversion (@code{di})
@cindex diversion, ending (@code{di})
@cindex appending to a diversion (@code{da})
@cindex diversion, appending (@code{da})
Start collecting formatted output in a diversion called @var{name}.  The
@code{da} request appends to a diversion called @var{name}, creating it
if necessary.  If @var{name} already exists as an alias, the target of
the alias is replaced or appended to; recall @ref{Strings}.  The pending
output line is diverted as well.  Switching to another environment (with
the @code{ev} request) before invoking @code{di} or @code{da} avoids
including any pending output line in the diversion; see
@ref{Environments}.

Invoking @code{di} or @code{da} without an argument stops diverting
output to the diversion named by the most recent corresponding request.
If @code{di} or @code{da} is called without an argument when there is no
current diversion, a warning in category @samp{di} is produced.
@xref{Warnings}, for information about the enablement and suppression
of warnings.

@Example
Before the diversion.
.di yyy
In the diversion.
.br
.di
After the diversion.
.br
    @result{} After the diversion.
.yyy
    @result{} Before the diversion.  In the diversion.
@endExample
@endDefreq

@cindex box (diversion type)
Because it is often desirable to exclude the partially collected line
from a diversion, @code{groff} supports an alternative form of diversion
known as a @dfn{box}.

@DefreqList {box, [@Var{name}]}
@DefreqListEndx {boxa, [@Var{name}]}
Divert (or append) output to @var{name}, similarly to the @code{di} and
@code{da} requests, respectively.  Any output line pending when these
requests are invoked is @emph{not} included in the box.  Calling
@code{box} or @code{boxa} without an argument stops diverting output to
the box named by the most recent corresponding request; a pending output
line inside a diversion is discarded.

@Example
Before the box.
.box xxx
In the box.
.br
Hidden treasure.
.box
After the box.
.br
    @result{} Before the box.  After the box.
.xxx
    @result{} In the box.
@endExample
@endDefreq

Apart from pending output line inclusion and the request names that
populate them, boxes are handled exactly as diversions are.  All of the
following @code{groff} language elements can be used with them
interchangeably.

@DefregList {.z}
@DefregListEndx {.d}
@cindex @code{nl} register, and @code{.d}
@cindex nested diversions
@cindex diversion, nested
@cindex diversion name register (@code{.z})
@cindex vertical position in diversion register (@code{.d})
@cindex position, vertical, in diversion, register (@code{.d})
@cindex diversion, vertical position in, register (@code{.d})
Diversions may be nested.  The read-only string-valued register
@code{.z} contains the name of the current diversion.  The read-only
register @code{.d} contains the current vertical place in the diversion.
If the input text is not being diverted, @code{.d} reports the same
location as the register @code{nl}.
@endDefreg

@Defreg {.h}
@cindex high-water mark register (@code{.h})
@cindex mark, high-water, register (@code{.h})
@cindex position of lowest text line (@code{.h})
@cindex text line, position of lowest (@code{.h})
The read-only register @code{.h} stores the @dfn{high-water mark} on the
current page or in the current diversion.  It corresponds to the text
baseline of the lowest line on the page.@footnote{Thus, the ``water''
gets ``higher'' proceeding @emph{down} the page.}

@Example
.tm .h==\n[.h], nl==\n[nl]
    @result{} .h==0, nl==-1
This is a test.
.br
.sp 2
.tm .h==\n[.h], nl==\n[nl]
    @result{} .h==40, nl==120
@endExample

@cindex @code{.h} register, difference to @code{nl}
@cindex @code{nl} register, difference to @code{.h}
@noindent
As implied by the example, vertical motion does not produce text
baselines and thus does not increase the value interpolated by
@samp{\n[.h]}.
@endDefreg

@DefregList {dn}
@DefregListEndx {dl}
@cindex @code{dn} register, and @code{da} (@code{boxa})
@cindex @code{dl} register, and @code{da} (@code{boxa})
@cindex @code{da} request, and @code{dn} (@code{dl})
@cindex @code{boxa} request, and @code{dn} (@code{dl})
After completing a diversion, the writable registers @code{dn} and
@code{dl} contain its vertical and horizontal sizes.  Only the lines
just processed are counted: for the computation of @code{dn} and
@code{dl}, the requests @code{da} and @code{boxa} are handled as if
@code{di} and @code{box} had been used, respectively---lines that have
been already stored in the diversion (box) are not taken into account.

@Example
.\" Center text both horizontally and vertically.
.\" Macro .(c starts centering mode; .)c terminates it.
.
.\" Disable the escape character with .eo so that we
.\" don't have to double backslashes on the "\n"s.
.eo
.de (c
.  br
.  ev (c
.  evc 0
.  in 0
.  nf
.  di @@c
..
@endExample
@Example
.de )c
.  br
.  ev
.  di
.  nr @@s (((\n[.t]u - \n[dn]u) / 2u) - 1v)
.  sp \n[@@s]u
.  ce 1000
.  @@c
.  ce 0
.  sp \n[@@s]u
.  br
.  fi
.  rr @@s
.  rm @@c
..
.ec
@endExample
@endDefreg

@DefescList {\\!, , anything, }
@DefescListEndx {\\?, , anything, \\?}
@cindex transparent output (@code{\!}, @code{\?})
@cindex output, transparent (@code{\!}, @code{\?})
@dfn{Transparently} embed @var{anything} into the current diversion,
preventing requests, macro calls, and escape sequences from being
interpreted when read into a diversion.  This is useful for preventing
them from taking effect until the diverted text is actually output.  The
@code{\!} escape sequence transparently embeds input up to and including
the end of the line.  The @code{\?} escape sequence transparently embeds
input until its own next occurrence.

@cindex @code{\?}, and copy mode
@cindex copy mode, and @code{\?}
@cindex mode, copy, and @code{\?}
@cindex @code{\!}, and copy mode
@cindex copy mode, and @code{\!}
@cindex mode, copy, and @code{\!}
@noindent
@var{anything} may not contain newlines; use @code{\!} by itself to
embed newlines in a diversion.  The escape sequence @code{\?} is also
recognized in copy mode and turned into a single internal code; it is
this code that terminates @var{anything}.  Thus the following example
prints@tie{}4.

@Example
.nr x 1
.nf
.di d
\?\\?\\\\?\\\\\\\\nx\\\\?\\?\?
.di
.nr x 2
.di e
.d
.di
.nr x 3
.di f
.e
.di
.nr x 4
.f
@endExample

Both escape sequences read the data in copy mode.

@cindex @code{\!}, in top-level diversion
@cindex top-level diversion, and @code{\!}
@cindex diversion, top-level, and @code{\!}
If @code{\!} is used in the top-level diversion, its argument is
directly embedded into GNU @code{troff}'s intermediate output.  This can
be used, for example, to control a postprocessor that processes the data
before it is sent to an output driver.

@cindex @code{\?}, in top-level diversion
@cindex top-level diversion, and @code{\?}
@cindex diversion, top-level, and @code{\?}
The @code{\?} escape used in the top-level diversion produces no output
at all; its argument is simply ignored.
@endDefesc

@cindex @code{\!}, and @code{output} request
@cindex @code{output} request, and @code{\!}
@cindex @code{output} request, and copy mode
@cindex copy mode, and @code{output} request
@cindex mode, copy, and @code{output} request
@Defreq {output, contents}
Emit @var{contents} directly to GNU @code{troff}'s intermediate output
(subject to copy mode interpretation); this is similar to @code{\!} used
at the top level.  An initial neutral double quote in @var{contents} is
stripped to allow embedding of leading spaces.

This request can't be used before the first page has started---if you
get an error, simply insert @code{.br} before the @code{output} request.

Use with caution!  It is normally only needed for mark-up used by a
postprocessor that does something with the output before sending it to
the output device, filtering out @var{contents} again.
@endDefreq

@Defreq {asciify, div}
@cindex unformatting diversions (@code{asciify})
@cindex diversion, unformatting (@code{asciify})
@cindex @code{trin} request, and @code{asciify}
@dfn{Unformat} the diversion @var{div} in a way such that Unicode basic
Latin (@acronym{ASCII}) characters, characters translated with the
@code{trin} request, space characters, and some escape sequences, that
were formatted and diverted into @var{div} are treated like ordinary
input characters when @var{div} is reread.  Doing so can be useful in
conjunction with the @code{writem} request.  @code{asciify} can be also
used for gross hacks; for example, the following sets
register@tie{}@code{n} to@tie{}1.

@Example
.tr @@.
.di x
@@nr n 1
.br
.di
.tr @@@@
.asciify x
.x
@endExample

@code{asciify} cannot return all items in a diversion to their source
equivalent: nodes such as those produced by the @code{\N} escape
sequence will remain nodes, so the result cannot be guaranteed to be a
pure string.  @xref{Copy Mode}.  Glyph parameters such as the type face
and size are not preserved; use @code{unformat} to achieve that.
@endDefreq

@Defreq {unformat, div}
Like @code{asciify}, unformat the diversion @var{div}.  However,
@code{unformat} handles only tabs and spaces between words, the latter
usually arising from spaces or newlines in the input.  Tabs are treated
as input tokens, and spaces become adjustable again.  The vertical sizes
of lines are not preserved, but glyph information (font, type size,
space width, and so on) is retained.
@endDefreq


@c =====================================================================

@node Punning Names, Environments, Diversions, GNU troff Reference
@section Punning Names
@cindex diversions

Macros, strings, and diversions share a name space; recall
@ref{Identifiers}.  Internally, the same mechanism is used to store
them.  You can thus call a macro with string interpolation syntax and
vice versa.

@Example
.de subject
Typesetting
..
.de predicate
rewards attention to detail
..
\*[subject] \*[predicate].
Truly.
    @result{} Typesetting
    @result{}  rewards attention to detail Truly.
@endExample

@noindent
What went wrong?  Strings don't contain newlines, but macros do.  String
interpolation placed a newline at the end of @samp{\*[subject]}, and the
next thing on the input was a space.  Then when @samp{\*[predicate]} was
interpolated, it was followed by the empty request @samp{.} on a line by
itself.  If we want to use macros as strings, we must take interpolation
behavior into account.

@Example
.de subject
Typesetting\\
..
.de predicate
rewards attention to detail\\
..
\*[subject] \*[predicate].
Truly.
    @result{} Typesetting rewards attention to detail.  Truly.
@endExample

@noindent
By ending each text line of the macros with an escaped
@code{\@key{RET}}, we get the desired effect (@pxref{Line
Continuation}).@footnote{The backslash is doubled.  @xref{Copy Mode}.}
What would have happened if we had used only one backslash at a time
instead?

Interpolating a string does not hide existing macro arguments.  We can
also place the escaped newline outside the string interpolation instead
of within the string definition.  Thus, in a macro, a more efficient way
of doing

@Example
.xx \\$@@
@endExample

@noindent
is

@Example
\\*[xx]\\
@endExample

@noindent
The latter calling syntax doesn't change the value of @code{\$0}, which
is then inherited from the calling macro (@pxref{Parameters}).

Diversions can be also called with string syntax.  It is sometimes
convenient to copy one-line diversions to a string.

@Example
.di xx
the
.ft I
interpolation system
.ft
.br
.di
.ds yy This is a test of \*(xx\c
\*(yy.
    @result{} This is a test of the @i{interpolation system}.
@endExample

@noindent
As the previous example shows, it is possible to store formatted output
in strings.  The @code{\c} escape sequence prevents the subsequent
newline from being interpreted as a break (again,
@pxref{Line Continuation}).

Copying multi-output line diversions produces unexpected results.

@Example
.di xxx
a funny
.br
test
.br
.di
.ds yyy This is \*[xxx]\c
\*[yyy].
    @result{} test This is a funny.
@endExample

Usually, it is not predictable whether a diversion contains one or more
output lines, so this mechanism should be avoided.  With @acronym{AT&T}
@code{troff}, this was the only solution to strip off a final newline
from a diversion.  Another disadvantage is that the spaces in the copied
string are already formatted, preventing their adjustment.  This can
cause ugly results.

@cindex stripping final newline in diversions
@cindex diversion, stripping final newline
@cindex final newline, stripping in diversions
@cindex newline, final, stripping in diversions
@cindex horizontal space, unformatting
@cindex space, horizontal, unformatting
@cindex unformatting horizontal space
A clean solution to this problem is available in GNU @code{troff}, using
the requests @code{chop} to remove the final newline of a diversion, and
@code{unformat} to make the horizontal spaces adjustable again.

@Example
.box xxx
a funny
.br
test
.br
.box
.chop xxx
.unformat xxx
This is \*[xxx].
    @result{} This is a funny test.
@endExample

@xref{Gtroff Internals}.

@c =====================================================================

@c BEGIN Keep parallel with section "Environments" of groff(7).
@node Environments, Suppressing Output, Diversions, GNU troff Reference
@section Environments
@cindex environments

As discussed in @ref{Deferring Output}, environments store most of the
parameters that control text processing.  A default environment named
@samp{0} exists when GNU @code{troff} starts up; it is modified by
formatting-related requests and escape sequences.

@cindex stack
You can create new environments and switch among them.  Only one is
current at any given time.  Active environments are managed using a
@dfn{stack}, a data structure supporting ``push'' and ``pop''
operations.  The current environment is at the top of the stack.
The same environment name can be pushed onto the stack multiple times,
possibly interleaved with others.  Popping the environment stack does
not destroy the current environment; it remains accessible by name and
can be made current again by pushing it at any time.  Environments
cannot be renamed or deleted, and can only be modified when current.  To
inspect the environment stack, use the @code{pev} request; see
@ref{Debugging}.

Environments store the following information.

@itemize @bullet
@item
a partially collected line, if any

@item
data about the most recently output glyph and line (registers
@code{.cdp}, @code{.cht}, @code{.csk}, @code{.n}, @code{.w})

@item
typeface parameters (size, family, style, height and slant, inter-word
and inter-sentence space sizes)

@item
page parameters (line length, title length, vertical spacing, line
spacing, indentation, line numbering, centering, right-alignment,
underlining, hyphenation parameters)

@item
filling enablement; adjustment enablement and mode

@item
tab stops; tab, leader, escape, control, no-break control, hyphenation,
and margin characters

@item
input line traps

@item
stroke and fill colors
@end itemize
@c END Keep parallel with section "Environments" of groff(7).

@DefreqList {ev, [@Var{ident}]}
@DefregListEndx {.ev}
@cindex switching environments (@code{ev})
@cindex environment, switching (@code{ev})
@cindex environment number/name register (@code{.ev})
Enter the environment @var{ident}, which is created if it does not
already exist, using the same parameters as for the default environment
used at startup.  With no argument, GNU @code{troff} switches to the
previous environment.

Invoking @code{ev} with an argument puts environment @var{ident} onto
the top of the environment stack.  (If it isn't already present in the
stack, this is a proper push.)  Without an argument, @code{ev} pops the
environment stack, making the previous environment current.  It is an
error to pop the environment stack with no previous environment
available.  The read-only string-valued register @code{.ev} contains the
name of the current environment---the one at the top of the stack.

@Example
.ev footnote-env
.fam N
.ps 6
.vs 8
.ll -.5i
.ev

@r{@dots{}}

.ev footnote-env
\[dg] Observe the smaller text and vertical spacing.
.ev
@endExample

We can familiarize ourselves with stack behavior by wrapping the
@code{ev} request with a macro that reports the contents of the
@code{.ev} register to the standard error stream.

@Example
.de EV
.  ev \\$1
.  tm environment is now \\n[.ev]
..
.
.EV foo
.EV bar
.EV
.EV baz
.EV
.EV
.EV
@endExample

@Example
    @error{} environment is now foo
    @error{} environment is now bar
    @error{} environment is now foo
    @error{} environment is now baz
    @error{} environment is now foo
    @error{} environment is now 0
    @error{} error: environment stack underflow
    @error{} environment is now 0
@endExample

@endDefreq

@Defreq {evc, environment}
@cindex copying environment (@code{evc})
@cindex environment, copying (@code{evc})
Copy the contents of @var{environment} to the current environment.

The following environment data are not copied.

@itemize @bullet
@item
a partially collected line, if present;

@item
the interruption status of the previous input line (due to use of the
@code{\c} escape sequence);

@item
the count of remaining lines to center, to right-justify, or to
underline (with or without underlined spaces)---these are set to zero;

@item
the activation status of temporary indentation;

@item
input line traps and their associated data;

@item
the activation status of line numbering (which can be reactivated with
@w{@samp{.nm +0}}); and

@item
the count of consecutive hyphenated lines (set to zero).
@end itemize
@endDefreq

@DefregList {.w}
@DefregItemx {.cht}
@DefregItemx {.cdp}
@DefregListEndx {.csk}
@cindex environment, dimensions of last glyph (@code{.w}, @code{.cht}, @code{.cdp}, @code{.csk})
@cindex width, of last glyph (@code{.w})
@cindex height, of last glyph (@code{.cht})
@cindex depth, of last glyph (@code{.cdp})
@cindex skew, of last glyph (@code{.csk})
@cindex last glyph, dimensions (@code{.w}, @code{.cht}, @code{.cdp}, @code{.csk})
@cindex glyph, last, dimensions (@code{.w}, @code{.cht}, @code{.cdp}, @code{.csk})
The @code{\n[.w]} register contains the width of the last glyph
formatted in the environment.

The @code{\n[.cht]} register contains the height of the last glyph
formatted in the environment.

The @code{\n[.cdp]} register contains the depth of the last glyph
formatted in the environment.  It is positive for glyphs extending below
the baseline.

The @code{\n[.csk]} register contains the @dfn{skew} (how far to the
right of the glyph's center that GNU @code{troff} should place an
accent) of the last glyph formatted in the environment.
@endDefreg

@Defreg {.n}
@cindex environment, previous line length (@code{.n})
@cindex line length, previous (@code{.n})
@cindex length of previous line (@code{.n})
@cindex previous line length (@code{.n})
The @code{\n[.n]} register contains the length of the previous output
line emitted in the environment.
@endDefreg

@codequotebacktick off
@codequoteundirected off


@c =====================================================================

@node Suppressing Output, Colors, Environments, GNU troff Reference
@section Suppressing Output

@Defesc {\\O, [, num, ]}
@cindex suppressing output (@code{\O})
@cindex output, suppressing (@code{\O})
Suppress GNU @code{troff} output of glyphs and geometric primitives.
The sequences @code{\O2}, @code{\O3}, @code{\O4}, and @code{\O5} are
intended for internal use by @code{grohtml}.

@table @samp
@item \O0
Disable the emission of glyphs and geometric primitives to the output
driver, provided that this sequence occurs at the outermost level
(see @code{\O3} and @code{\04} below).  Horizontal motions corresponding
to non-overstruck glyph widths still occur.

@item \O1
Enable the emission of glyphs and geometric primitives to the output
driver, provided that this sequence occurs at the outermost level.
@end table

@vindex opminx
@vindex opminy
@vindex opmaxx
@vindex opmaxy
@code{\O0} and @code{\O1} also reset the four registers @code{opminx},
@code{opminy}, @code{opmaxx}, and @code{opmaxy} to @minus{}1.  These
four registers mark the top left and bottom right hand corners of a box
encompassing all written glyphs.

@table @samp
@item \O2
Provided that this sequence occurs at the outermost level, enable
emission of glyphs and geometric primitives, and write to the standard
error stream the page number and values of the four aforementioned
registers encompassing glyphs written since the last interpolation of a
@code{\O} sequence, as well as the page offset, line length, image file
name (if any), horizontal and vertical device motion quanta, and input
file name.  Numeric values are in basic units.

@item \O3
Begin a nesting level.  This is an internal mechanism for @code{grohtml}
while producing images.  At startup, @code{gtroff} is at the outermost
level.  These sequences are generated when processing the source
document with @command{pre-grohtml}, which uses @command{gtroff} with
the @code{ps} output device, Ghostscript, and the PNM tools to produce
images in PNG format.  They start a new page if the device is not
@code{html} or @code{xhtml}, to reduce the possibility of images
crossing a page boundary.

@item \O4
End a nesting level.
@end table

@table @samp
@item \O[5@var{P}@var{file}]
Provided that this sequence occurs at the outermost level, write the
name @code{file} to the standard error stream at position @var{P}, which
must be one of @code{l}, @code{r}, @code{c}, or@tie{}@code{i},
corresponding to left, right, centered, and inline alignments within the
document, respectively.  @var{file} is a name associated with the
production of the next image.
@end table
@endDefesc

@Defreg {.O}
@cindex suppression nesting level register
@cindex nesting level, suppression, register
@cindex level, suppression nesting, register
Output suppression nesting level applied by @code{\O3} and @code{\O4}
escape sequences.
@endDefreg

@c =====================================================================

@codequotebacktick on
@codequoteundirected on

@c TODO: Rename this node to "Operating Environment Access" or similar,
@c and move the date/time, process ID, etc., read-only registers here.
@node I/O, Postprocessor Access, Suppressing Output, GNU troff Reference
@section I/O
@cindex i/o
@cindex input and output requests
@cindex requests for input and output
@cindex output and input requests

@code{gtroff} has several requests for including files:

@DefreqList {so, file}
@DefreqListEndx {soquiet, file}
@cindex including a file (@code{so})
@cindex file, inclusion (@code{so})
Replace the @code{so} request's control line with the contents of the
file named by the argument, ``sourcing'' it.  @var{file} is sought in
the directories specified by @option{-I} command-line option.  If
@var{file} does not exist, a warning in category @samp{file} is produced
and the request has no further effect.  @xref{Warnings}, for
information about the enablement and suppression of warnings.

@code{so} can be useful for large documents; e.g., allowing each chapter
of a book to be kept in a separate file.  However, files interpolated
with @code{so} are not preprocessed; to overcome this limitation, see
the @cite{gsoelim@r{(1)}} man page.

Since GNU @code{troff} replaces the entire control line with the
contents of a file, it matters whether @code{file} is terminated with a
newline or not.  Assume that file @file{xxx} contains only the word
@samp{foo} without a trailing newline.

@Example
$ printf 'foo' > xxx

The situation is
.so xxx
bar.
    @result{} The situation is foobar.
@endExample

@code{soquiet} works the same way, except that no warning diagnostic is
issued if @var{file} does not exist.
@endDefreq

@Defreq {pso, command}
Read the standard output from the specified @var{command} and include
it in place of the @code{pso} request.

@cindex safer mode
@cindex mode, safer
@cindex unsafe mode
@cindex mode, unsafe
It is an error to use this request in safer mode, which is the
default.  Invoke GNU @code{troff} or a front end with the @option{-U}
option to enable unsafe mode.

The comment regarding a final newline for the @code{so} request is valid
for @code{pso} also.
@endDefreq

@DefreqList {mso, file}
@DefreqListEndx {msoquiet, file}
Identical to the @code{so} and @code{soquiet} requests, respectively,
except that @code{gtroff} searches for the specified @var{file} in the
same directories as macro files for the @option{-m} command-line option.
If the file name to be included has the form @file{@var{name}.tmac} and
it isn't found, these requests try to include @file{tmac.@var{name}} and
vice versa.
@endDefreq

@DefreqList {trf, file}
@DefreqListEndx {cf, file}
@cindex transparent output (@code{cf}, @code{trf})
@cindex output, transparent (@code{cf}, @code{trf})
@cindex @code{cf} request, and copy mode
@cindex copy mode, and @code{cf} request
@cindex mode, copy, and @code{cf} request
@cindex @code{trf} request, and copy mode
@cindex copy mode, and @code{trf} request
@cindex mode, copy, and @code{trf} request
Transparently output the contents of @var{file}.  Each line is output as
if it were preceded by @code{\!}; however, the lines are @emph{not}
subject to copy mode interpretation.  If the file does not end with a
newline, @code{trf} adds one.  Both requests cause a break.

When used in a diversion, these requests embed a node (@pxref{Gtroff
Internals}) in it that, when reread, causes the contents of @var{file}
to be transparently copied to the output.  In @acronym{AT&T}
@code{troff}, the contents of @var{file} are immediately copied to the
output regardless of whether there is a current diversion; this
behaviour is so anomalous that it must be considered a bug.

@cindex @code{trf} request, and invalid characters
@cindex characters, invalid for @code{trf} request
@cindex invalid characters for @code{trf} request
While @code{cf} copies the contents of @var{file} completely
unprocessed, @code{trf} disallows characters such as NUL that are not
valid @code{gtroff} input characters (@pxref{Identifiers}).

For @code{cf}, within a diversion, ``completely unprocessed'' means that
each line of a file to be inserted is handled as if it were preceded by
@code{\!\\!}.

To define a macro@tie{}@code{x} containing the contents of
file@tie{}@file{f}, use

@Example
.ev 1
.di x
.trf f
.di
.ev
@endExample

@noindent
The calls to @code{ev} prevent the partially collected output line
from becoming part of the diversion (@pxref{Diversions}).
@endDefreq

@Defreq {nx, [@Var{file}]}
@cindex processing next file (@code{nx})
@cindex file, processing next (@code{nx})
@cindex next file, processing (@code{nx})
Force @code{gtroff} to continue processing of the file specified as an
argument.  If no argument is given, immediately jump to the end of file.
@endDefreq

@Defreq {rd, [@Var{prompt} [@Var{arg1} @Var{arg2} @dots{}]]}
@cindex reading from standard input (@code{rd})
@cindex standard input, reading from (@code{rd})
@cindex input, standard, reading from (@code{rd})
Read from standard input, and include what is read as though it were
part of the input file.  Text is read until a blank line is encountered.

If standard input is a TTY input device (keyboard), write @var{prompt}
to standard error, followed by a colon (or send BEL for a beep if no
argument is given).

Arguments after @var{prompt} are available for the input.  For example,
the line

@Example
.rd data foo bar
@endExample

with the input @w{@samp{This is \$2.}} prints

@Example
This is bar.
@endExample
@endDefreq

@cindex form letters
@cindex letters, form
Using the @code{nx} and @code{rd} requests, it is easy to set up form
letters.  The form letter template is constructed like this, putting the
following lines into a file called @file{repeat.let}:

@Example
.ce
\*(td
.sp 2
.nf
.rd
.sp
.rd
.fi
Body of letter.
.bp
.nx repeat.let
@endExample

@cindex @code{ex} request, used with @code{nx} and @code{rd}
@noindent
When this is run, a file containing the following lines should be
redirected in.  Requests included in this file are executed as though
they were part of the form letter.  The last block of input is the
@code{ex} request, which tells GNU @code{troff} to stop processing.  If
this were not there, @code{troff} would not know when to stop.

@Example
Trent A. Fisher
708 NW 19th Av., #202
Portland, OR  97209

Dear Trent,

Len Adollar
4315 Sierra Vista
San Diego, CA  92103

Dear Mr. Adollar,

.ex
@endExample

@Defreq {pi, pipe}
Pipe the output of @code{gtroff} to the shell command(s) specified by
@var{pipe}.  This request must occur before @code{gtroff} has a chance
to print anything.

@cindex safer mode
@cindex mode, safer
@cindex unsafe mode
@cindex mode, unsafe
It is an error to use this request in safer mode, which is the
default.  Invoke GNU @code{troff} or a front end with the @option{-U}
option to enable unsafe mode.

Multiple calls to @code{pi} are allowed, acting as a chain.  For
example,

@Example
.pi foo
.pi bar
...
@endExample

is the same as @w{@samp{.pi foo | bar}}.

@cindex @code{groff}, and @code{pi} request
@cindex @code{pi} request, and @code{groff}
The intermediate output format of GNU @code{troff} is piped to the
specified commands.  Consequently, calling @code{groff} without the
@option{-Z} option normally causes a fatal error.
@endDefreq

@cindex system commands, running
@cindex running system commands
@DefreqList {sy, cmds}
@DefregListEndx {systat}
Execute the shell command(s) specified by @var{cmds}.  The output is not
saved anywhere, so it is up to the user to do so.

@cindex safer mode
@cindex mode, safer
@cindex unsafe mode
@cindex mode, unsafe
It is an error to use this request in safer mode; this is the default.
Give GNU @code{troff} or a front end program the @option{-U} option to
enable unsafe mode.

The following code fragment introduces the current time into a document.

@pindex perl
@Example
.sy perl -e 'printf ".nr H %d\\n.nr M %d\\n.nr S %d\\n",\
             (localtime(time))[2,1,0]' > /tmp/x\n[$$]
.so /tmp/x\n[$$]
.sy rm /tmp/x\n[$$]
\nH:\nM:\nS
@endExample

@noindent
This works by having the Perl script (run by @code{sy}) write
@code{nr} requests that set the registers @code{H}, @code{M}, and
@code{S} to a temporary file.  The @code{roff} document then reads the
temporary file using the @code{so} request.

@cindex time, formatting
@cindex formatting the time
The registers @code{seconds}, @code{minutes}, and @code{hours},
initialized at startup of GNU @code{troff}, should satisfy most
requirements.  Use the @code{af} request to format their values for
output.

@Example
.af hours 00
.af minutes 00
.af seconds 00
\n[hours]:\n[minutes]:\n[seconds]
    @result{} 02:17:54
@endExample

@cindex @code{system()} return value register (@code{systat})
The writable register @code{systat} contains the return value of the
@code{system()} function executed by the last @code{sy} request.
@endDefreq

@DefreqList {open, stream file}
@DefreqListEndx {opena, stream file}
@cindex opening file (@code{open})
@cindex file, opening (@code{open})
@cindex appending to a file (@code{opena})
@cindex file, appending to (@code{opena})
Open the specified @var{file} for writing and associates the specified
@var{stream} with it.

The @code{opena} request is like @code{open}, but if the file exists,
append to it instead of truncating it.

@cindex safer mode
@cindex mode, safer
@cindex unsafe mode
@cindex mode, unsafe
It is an error to use these requests in safer mode; this is the default.
Give GNU @code{troff} or a front end program the @option{-U} option to
enable unsafe mode.
@endDefreq

@DefreqList {write, stream data}
@DefreqListEndx {writec, stream data}
@cindex copy mode, and @code{write} request
@cindex @code{write} request, and copy mode
@cindex mode, copy, and @code{write} request
@cindex copy mode, and @code{writec} request
@cindex @code{writec} request, and copy mode
@cindex mode, copy, and @code{writec} request
@cindex writing to file (@code{write}, @code{writec})
@cindex file, writing to (@code{write}, @code{writec})
Write to the file associated with the specified @var{stream}.  The
stream must previously have been the subject of an open request.  The
remainder of the line is interpreted as the @code{ds} request reads its
second argument: an initial neutral double quote in @var{contents} is
stripped to allow embedding of leading spaces, and it is read in copy
mode.

The @code{writec} request is like @code{write}, but only @code{write}
appends a newline to the data.
@endDefreq

@Defreq {writem, stream xx}
@cindex @code{asciify} request, and @code{writem}
Write the contents of the macro or string @var{xx} to the file
associated with the specified @var{stream}.

@cindex @code{writem} request, and copy mode
@cindex copy mode, and @code{writem} request
@cindex mode, copy, and @code{writem} request
@var{xx} is read in copy mode, i.e., already formatted elements are
ignored.  Consequently, diversions must be unformatted with the
@code{asciify} request before calling @code{writem}.  Usually, this
means a loss of information.
@endDefreq

@Defreq {close, stream}
@cindex closing file (@code{close})
@cindex file, closing (@code{close})
Close the specified @var{stream}; the stream is no longer an acceptable
argument to the @code{write} request.

Here a simple macro to write an index entry.

@Example
.open idx test.idx
.
.de IX
.  write idx \\n[%] \\$*
..
.
.IX test entry
.
.close idx
@endExample
@endDefreq

@DefescList {\\V, , e, }
@DefescItem {\\V, (, ev, }
@DefescListEnd {\\V, [, env, ]}
@cindex @code{\V}, and copy mode
@cindex copy mode, and @code{\V}
@cindex mode, copy, and @code{\V}
Interpolate the contents of the specified environment variable @var{env}
(one-character name@tie{}@var{e}, two-character name @var{ev}) as
returned by the function @code{getenv}.  @code{\V} is interpreted even
in copy mode (@pxref{Copy Mode}).
@endDefesc


@c =====================================================================

@node Postprocessor Access, Miscellaneous, I/O, GNU troff Reference
@section Postprocessor Access
@cindex postprocessor access
@cindex access to postprocessor

Two escape sequences and two requests enable documents to pass
information directly to a postprocessor.  These are useful for
exercising device-specific capabilities that the @code{groff} language
does not abstract or generalize; examples include the embedding of
hyperlinks and image files.  Device-specific functions are documented in
each output driver's man page, such as @cite{gropdf@r{(1)}},
@cite{grops@r{(1)}}, or @cite{grotty@r{(1)}}.

@DefreqList {device, xxx @r{@dots{}}}
@DefescListEndx {\\X, @code{'}, xxx @r{@dots{}}, @code{'}}
Embed all @var{xxx} arguments into GNU @code{troff} output as parameters
to a device control command @w{@samp{x X}}.  The meaning and
interpretation of such parameters is determined by the output driver or
other postprocessor.

@cindex @code{device} request, and copy mode
@cindex copy mode, and @code{device} request
@cindex mode, copy, and @code{device} request
The @code{device} request processes its arguments in copy mode
(@pxref{Copy Mode}).  An initial neutral double quote in @var{contents}
is stripped to allow embedding of leading spaces.
@cindex @code{\&}, in @code{\X}
@cindex @code{\)}, in @code{\X}
@cindex @code{\%}, in @code{\X}
@ifnotinfo
@cindex @code{\:}, in @code{\X}
@end ifnotinfo
@ifinfo
@cindex @code{\@r{<colon>}}, in @code{\X}
@end ifinfo
By contrast, within @code{\X} arguments, the escape sequences @code{\&},
@code{\)}, @code{\%}, and @code{\:} are ignored; @code{\@key{SP}} and
@code{\~} are converted to single space characters; and @code{\\} has
its escape character stripped.  So that the basic Latin subset of the
Unicode character set@footnote{that is, ISO@tie{}646:1991-IRV or,
popularly, ``US-ASCII''} can be reliably encoded in device control
commands, seven special character escape sequences (@samp{\-},
@samp{\aq}, @samp{\dq}, @samp{\ga}, @samp{\ha}, @samp{\rs}, and
@samp{\ti},) are mapped to basic Latin glyphs; see the
@cite{groff_char@r{(7)}} man page.  The use of any other escape sequence
in @code{\X} arguments is normally an error.

@kindex use_charnames_in_special
@cindex @file{DESC} file, and @code{use_charnames_in_special} keyword
@cindex @code{\X}, and special characters
If the @code{use_charnames_in_special} directive appears in the output
device's @file{DESC} file, the use of special character escape sequences
is @emph{not} an error; they are simply output verbatim (with the
exception of the seven mapped to Unicode basic Latin characters,
discussed above).  @code{use_charnames_in_special} is currently employed
only by @code{grohtml}.
@endDefesc

@DefreqList {devicem, name}
@DefescItemx {\\Y, , n, }
@DefescItem {\\Y, (, nm, }
@DefescListEnd {\\Y, [, name, ]}
This is approximately equivalent to @samp{\X'\*[@var{name}]'}
(one-character name@tie{}@var{n}, two-character name @var{nm}).
However, the contents of the string or macro @var{name} are not
interpreted; also it is permitted for @var{name} to have been defined as
a macro and thus contain newlines (it is not permitted for the argument
to @code{\X} to contain newlines).  The inclusion of newlines requires
an extension to the @acronym{AT&T} @code{troff} output format, and
confuses drivers that do not know about this extension (@pxref{Device
Control Commands}).
@endDefesc

@DefreqList {tag, name}
@DefreqListEndx {taga, name}
Reserved for internal use.
@endDefreq

@codequotebacktick off
@codequoteundirected off


@c =====================================================================

@node Miscellaneous, Gtroff Internals, Postprocessor Access, GNU troff Reference
@section Miscellaneous

This section documents parts of @code{gtroff} that cannot (yet) be
categorized elsewhere in this manual.

@DefreqList {nm, [@Var{start} [@Var{inc} [@Var{space} [@Var{indent}]]]]}
@DefregItemx {ln}
@DefregListEndx {.nm}
@cindex printing line numbers (@code{nm})
@cindex line numbers, printing (@code{nm})
@cindex numbers, line, printing (@code{nm})
Print line numbers.  @var{start} is the line number of the @emph{next}
output line.  @var{inc} indicates which line numbers are printed.  For
example, the value@tie{}5 means to emit only line numbers that are
multiples of@tie{}5; this defaults to@tie{}1.  @var{space} is the space
to be left between the number and the text; this defaults to one digit
space.  The fourth argument is the indentation of the line numbers,
defaulting to zero.  Both @var{space} and @var{indent} are given as
multiples of digit spaces; they can be negative also.  Without any
arguments, line numbers are turned off.

@code{gtroff} reserves three digit spaces for the line number (which is
printed right-justified) plus the amount given by @var{indent}; the
output lines are concatenated to the line numbers, separated by
@var{space}, and @emph{without} reducing the line length.  Depending on
the value of the horizontal page offset (as set with the @code{po}
request), line numbers that are longer than the reserved space stick
out to the left, or the whole line is moved to the right.

Parameters corresponding to missing arguments are not changed; any
non-digit argument (to be more precise, any argument starting with a
character valid as a delimiter for identifiers) is also treated as
missing.

If line numbering has been disabled with a call to @code{nm} without an
argument, it can be reactivated with @samp{.nm +0}, using the previously
active line numbering parameters.

The parameters of @code{nm} are associated with the environment
(@pxref{Environments}).

@cindex output line number register (@code{ln})
@cindex line number, output, register (@code{ln})
While line numbering is enabled, the output line number register
@code{ln} is updated as each line is output, even if no line number is
formatted with it because it is being skipped (it is not a multiple of
@var{inc}) or because numbering is suppressed (see the @code{nn} request
below).

The @code{.nm} register tracks the enablement status of line numbering.
Temporary suspension of numbering with the @code{nn} request does
@emph{not} alter its value.

@Example
.po 1m
.ll 2i
This test shows how line numbering works with groff.
.nm 999
This test shows how line numbering works with groff.
.br
.nm xxx 3 2
.ll -\w'0'u
This test shows how line numbering works with groff.
.nn 2
This test shows how line numbering works with groff.
@endExample

@noindent
The result is as follows.

@Example
 This  test shows how
 line numbering works
 999 with   groff.   This
1000 test shows how  line
1001 numbering works with
1002 groff.
      This test shows how
      line      numbering
 works  with  groff.
 This test shows how
1005  line      numbering
      works with groff.
@endExample
@endDefreq

@codequotebacktick on
@codequoteundirected on

@DefreqList {nn, [@Var{skip}]}
@DefregListEndx {.nn}
Suppress numbering of the next @var{skip} output lines that would
otherwise be numbered.  The default is@tie{}1.  @code{nn} can be invoked
when line numbering is not active; suppression of numbering will take
effect for @var{skip} lines once @code{nm} enables it.

The @code{.nn} register stores the count of output lines still to have
their numbering suppressed.

This count is associated with the environment (@pxref{Environments}).
@endDefreq

To test whether the current output line will be numbered, you must check
both the @code{.nm} and @code{.nn} registers.

@Example
  .de is-numbered
  .  nop This line
  .  ie (\\n[.nm] & (1-\\n[.nn])) IS
  .  el                           ISN'T
  .  nop numbered.
  .  br
  ..
  Test line numbering.
  .is-numbered
  .nm 1
  .nn 2
  .is-numbered
  .is-numbered
  .is-numbered
  .nm
  .is-numbered
@endExample

@noindent
The output lines correctly report their numbering status.

@Example
  Test line numbering.  This line ISN'T numbered.
  This line ISN'T numbered.
  This line ISN'T numbered.
    1 This line IS numbered.
  This line ISN'T numbered.
@endExample

@codequotebacktick off
@codequoteundirected off

@Defreq {mc, glyph [@Var{dist}]}
@cindex margin glyph (@code{mc})
@cindex glyph, for margins (@code{mc})
Print a @dfn{margin character} to the right of the
text.@footnote{@dfn{Margin character} is a misnomer since it is an
output glyph.}  The first argument is the glyph to be printed.  The
second argument is the distance away from the right margin.  If missing,
the previously set value is used; the default is 10@tie{}points.  For
text lines that are too long (that is, longer than the text length plus
@var{dist}), the margin character is directly appended to the lines.

With no arguments the margin character is turned off.  If this occurs
before a break, no margin character is printed.

For compatibility with @acronym{AT&T} @code{troff}, a call to @code{mc}
to set the margin character can't be undone immediately; at least one
line gets a margin character.  Thus

@Example
.ll 1i
.mc \[br]
.mc
xxx
.br
xxx
@endExample

@noindent
produces

@Example
xxx        |
xxx
@endExample

@cindex @code{tl} request, and @code{mc}
For empty lines and lines produced by the @code{tl} request no margin
character is emitted.

The margin character is associated with the environment
(@pxref{Environments}).

@pindex nrchbar
@pindex changebar
This is quite useful for indicating text that has changed, and, in fact,
there are programs available for doing this (they are called
@code{nrchbar} and @code{changebar} and can be found in any
@samp{comp.sources.unix} archive).

@Example
.ll 3i
.mc |
This paragraph is highlighted with a margin
character.
.sp
Vertical space isn't marked.
.br
\&
.br
But we can fake it with `\&'.
@endExample

Result:

@Example
This  paragraph is highlighted |
with a margin character.       |

Vertical space isn't marked.   |
                               |
But we can fake it with `\&'.  |
@endExample
@endDefreq

@DefreqList {psbb, file}
@DefregItemx {llx}
@DefregItemx {lly}
@DefregItemx {urx}
@DefregListEndx {ury}
@cindex PostScript, bounding box
@cindex bounding box
Retrieve the bounding box of the PostScript image found in @var{file},
which must conform to Adobe's @dfn{Document Structuring Conventions}
(DSC).  The request searches for a @code{%%BoundingBox} comment and
extracts the bounding box values into the registers @code{llx},
@code{lly}, @code{urx}, and @code{ury}.  If an error occurs (for
example, @code{psbb} cannot find the @code{%%BoundingBox} comment), it
sets the four registers to zero.

The search path for @var{file} can be controlled with the @option{-I}
command-line option.
@endDefreq


@c =====================================================================

@node Gtroff Internals, Debugging, Miscellaneous, GNU troff Reference
@section @code{gtroff} Internals

@cindex input token
@cindex token, input
@cindex output node
@cindex node, output
@code{gtroff} processes input in three steps.  One or more input
characters are converted to an @dfn{input token}.@footnote{Except the
escape sequences @code{\f}, @code{\F}, @code{\H}, @code{\m}, @code{\M},
@code{\R}, @code{\s}, and @code{\S}, which are processed immediately if
not in copy mode.}  Then, one or more input tokens are converted to
an @dfn{output node}.  Finally, output nodes are converted to the
intermediate output language understood by all output devices.

Actually, before step one happens, @code{gtroff} converts certain escape
sequences into reserved input characters (not accessible by the user);
such reserved characters are used for other internal processing also --
this is the very reason why not all characters are valid input.
@xref{Identifiers}, for more on this topic.

For example, the input string @samp{fi\[:u]} is converted into a
character token @samp{f}, a character token @samp{i}, and a special
token @samp{:u} (representing u@tie{}umlaut).  Later on, the character
tokens @samp{f} and @samp{i} are merged to a single output node
representing the ligature glyph @samp{fi} (provided the current font has
a glyph for this ligature); the same happens with @samp{:u}.  All output
glyph nodes are `processed', which means that they are invariably
associated with a given font, font size, advance width, etc.  During the
formatting process, @code{gtroff} itself adds various nodes to control
the data flow.

Macros, diversions, and strings collect elements in two chained lists: a
list of input tokens that have been passed unprocessed, and a list of
output nodes.  Consider the following diversion.

@Example
.di xxx
a
\!b
c
.br
.di
@endExample

@noindent
It contains these elements.

@multitable {@i{vertical size node}} {token list} {element number}
@item node list               @tab token list @tab element number

@item @i{line start node}     @tab ---        @tab 1
@item @i{glyph node @code{a}} @tab ---        @tab 2
@item @i{word space node}     @tab ---        @tab 3
@item ---                     @tab @code{b}   @tab 4
@item ---                     @tab @code{\n}  @tab 5
@item @i{glyph node @code{c}} @tab ---        @tab 6
@item @i{vertical size node}  @tab ---        @tab 7
@item @i{vertical size node}  @tab ---        @tab 8
@item ---                     @tab @code{\n}  @tab 9
@end multitable

@cindex @code{\v}, internal representation
@noindent
Elements 1, 7, and@tie{}8 are inserted by @code{gtroff}; the latter two
(which are always present) specify the vertical extent of the last line,
possibly modified by @code{\x}.  The @code{br} request finishes the
pending output line, inserting a newline input token, which is
subsequently converted to a space when the diversion is reread.  Note
that the word space node has a fixed width that isn't adjustable
anymore.  To convert horizontal space nodes back to input tokens, use
the @code{unformat} request.

Macros only contain elements in the token list (and the node list is
empty); diversions and strings can contain elements in both lists.

Note that the @code{chop} request simply reduces the number of elements
in a macro, string, or diversion by one.  Exceptions are
@dfn{compatibility save} and @dfn{compatibility ignore} input tokens,
which are ignored.  The @code{substring} request also ignores those
input tokens.

Some requests like @code{tr} or @code{cflags} work on glyph identifiers
only; this means that the associated glyph can be changed without
destroying this association.  This can be very helpful for substituting
glyphs.  In the following example, we assume that glyph @samp{foo} isn't
available by default, so we provide a substitution using the
@code{fchar} request and map it to input character @samp{x}.

@Example
.fchar \[foo] foo
.tr x \[foo]
@endExample

@noindent
Now let us assume that we install an additional special font @samp{bar}
that has glyph @samp{foo}.

@Example
.special bar
.rchar \[foo]
@endExample

@noindent
Since glyphs defined with @code{fchar} are searched before glyphs in
special fonts, we must call @code{rchar} to remove the definition of the
fallback glyph.  Anyway, the translation is still active; @samp{x} now
maps to the real glyph @samp{foo}.

@cindex compatibility mode, and parameters
@cindex mode, compatibility, and parameters
@cindex arguments, and compatibility mode
@cindex parameters, and compatibility mode
@cindex macro arguments, and compatibility mode
@cindex request arguments, and compatibility mode
Macro and request arguments preserve compatibility mode enablement.

@Example
.cp 1     \" switch to compatibility mode
.de xx
\\$1
..
.cp 0     \" switch compatibility mode off
.xx caf\['e]
    @result{} café
@endExample

@noindent
Since compatibility mode is enabled while @code{de} is invoked, the
macro @code{xx} enables compatibility mode when it is called.  Argument
@code{$1} can still be handled properly because it inherits the
compatibility mode enablement status that was active at the point where
@code{xx} was called.

After interpolation of the parameters, the compatibility save and
restore tokens are removed.


@c =====================================================================

@c BEGIN Keep parallel with section "Debugging" of groff(7).
@node Debugging, Implementation Differences, Gtroff Internals, GNU troff Reference
@section Debugging
@cindex debugging

@flushright
@slanted{Standard troff voodoo, just put a power of two backslashes in
front of it until it works and if you still have problems add a \c.}
--- Ron Natalie
@c https://minnie.tuhs.org/pipermail/tuhs/2021-February/023137.html
@end flushright

GNU @code{troff} is not the easiest language to debug, in part thanks to
its design features of recursive interpolation and the use of
multi-stage pipeline processing in the surrounding system.  Nevertheless
there exist several features useful for troubleshooting.

Preprocessors use the @code{lf} request to preserve the identity of the
line numbers and names of input files.  GNU @code{troff} emits a variety
of error diagnostics and supports several categories of warning; the
output of these can be selectively suppressed.  A trace of the
formatter's input processing stack can be emitted when errors or
warnings occur by means of GNU @code{troff}'s @option{-b} option, or
produced on demand with the @code{backtrace} request.  The @code{tm}
and related requests can be used to emit customized diagnostic messages
or for instrumentation while troubleshooting.  The @code{ex} and
@code{ab} requests cause early termination with successful and error
exit codes respectively, to halt further processing when continuing
would be fruitless.  Examine the state of the formatter with requests
that write lists of defined names (macros, strings, and diversions),
environments, registers, and page location traps to the standard error
stream.
@c END Keep parallel with section "Debugging" of groff(7).

@Defreq {lf, line [@Var{file}]}
@pindex soelim
@cindex multi-file documents
@cindex documents, multi-file
@cindex setting input line number (@code{lf})
@cindex input line number, setting (@code{lf})
@cindex number, input line, setting (@code{lf})
Set the input line number (and, optionally, the file name) GNU
@code{troff} shall use for error and warning messages.  @var{line} is
the input line number of the @emph{next} line.  Without an argument, the
request is ignored.

@code{lf}'s primary purpose is to aid the debugging of documents that
undergo preprocessing.  Programs like @command{tbl} that transform input
in their own languages intoto @code{roff} requests use it so that any
diagnostic messages emitted by @code{troff} correspond to the source
document.
@endDefreq

@DefreqList {tm, message}
@DefreqItemx {tm1, message}
@DefreqListEndx {tmc, message}
@cindex printing to stderr (@code{tm}, @code{tm1}, @code{tmc})
@cindex stderr, printing to (@code{tm}, @code{tm1}, @code{tmc})
Send @var{message}, which consumes the remainder of the input line and
cannot contain special characters, to the standard error stream.

The @code{tm} request ignores leading spaces of @var{contents};
@code{tm1} handles its argument similarly to the @code{ds} request: an
initial neutral double quote is stripped to allow embedding of leading
spaces.

The @code{tmc} request is similar to @code{tm1} but does not append a
newline (as is done in @code{tm} and @code{tm1}).
@endDefreq

@Defreq {ab, [@Var{message}]}
@cindex aborting (@code{ab})
Write any @var{message} to the standard error stream (like @code{tm})
and then abort GNU @code{troff}; that is, stop processing and terminate
with a failure status.
@endDefreq

@Defreq {ex, }
@cindex @code{ex} request, use in debugging
@cindex exiting (@code{ex})
Exit GNU @code{troff}; that is, stop processing and terminate with a
successful status.  To stop processing only the current file, use the
@code{nx} request; see @ref{I/O}.
@endDefreq

When doing something involved, it is useful to leave the debugging
statements in the code and have them turned on by a command-line flag.

@Example
.if \n[DB] .tm debugging output
@endExample

@noindent
To activate such statements, use the @option{-r} option to set the
register.

@Example
groff -rDB=1 file
@endExample

If it is known in advance that there are many errors and no useful
output, GNU @code{troff} can be forced to suppress formatted output with
the @option{-z} option.

@Defreq {pev, }
@cindex dumping environments (@code{pev})
@cindex environments, dumping (@code{pev})
Report the state of the current environment followed by that of all
other environments to the standard error stream.
@endDefreq

@Defreq {pm, }
@cindex dumping symbol table (@code{pm})
@cindex symbol table, dumping (@code{pm})
Report, to the standard error stream, the names of all defined macros,
strings, and diversions with their sizes in bytes.  Since GNU
@code{troff} sometimes adds nodes by itself, the returned sizes can be
larger than expected.
@endDefreq

@Defreq {pnr, }
@cindex dumping registers (@code{pnr})
@cindex registers, dumping (@code{pnr})
Report the names and contents of all currently defined registers to the
standard error stream.
@endDefreq

@Defreq {ptr, }
@cindex dumping page location traps (@code{ptr})
@cindex listing page location traps (@code{ptr})
@cindex traps, page location, dumping (@code{ptr})
@cindex traps, page location, listing (@code{ptr})
Report the names and positions of all page location traps to the
standard error stream.  Empty slots in the list, where a trap has been
planted but subsequently (re)moved, are printed as well.
@c "because they can affect the priority of subsequently planted traps."
@c XXX Is that right?  It's useful to print the empty slots, I think,
@c but a trap planted in an "empty" slot with .wh will become active.
@c The slot seems to act as an immobile dummy list head, but does not
@c change the basic list semantics.  .wh plants a trap at the head of
@c the trap list at a location, and .ch plants a trap at the tail.
@endDefreq

@Defreq {fl, }
@cindex flush output (@code{fl})
@cindex output, flush (@code{fl})
@cindex interactive use of @code{gtroff}
@cindex @code{gtroff}, interactive use
Instruct @code{gtroff} to flush its output immediately.  The intent is
for interactive use, but this behaviour is currently not implemented in
@code{gtroff}.  Contrary to Unix @code{troff}, TTY output is sent to a
device driver also (@code{grotty}), making it non-trivial to communicate
interactively.

This request causes a line break.
@endDefreq

@Defreq {backtrace, }
@cindex backtrace of input stack (@code{backtrace})
@cindex input stack, backtrace (@code{backtrace})
Write the state of the input stack to the standard error stream.

Consider the following in a file @file{test}.

@codequoteundirected on
@Example
.de xxx
.  backtrace
..
.de yyy
.  xxx
..
.
.yyy
    @error{} troff: backtrace: 'test':2: macro 'xxx'
    @error{} troff: backtrace: 'test':5: macro 'yyy'
    @error{} troff: backtrace: file 'test':8
@endExample
@codequoteundirected off

The @option{-b} option of GNU @code{troff} causes a backtrace to be
generated on each error or warning.  Some warnings have to be enabled;
@xref{Warnings}.
@endDefreq

@Defreg {slimit}
@cindex input stack, setting limit
If greater than@tie{}0, sets the maximum quantity of objects on GNU
@code{troff}'s internal input stack.  If less than or equal to@tie{}0,
there is no limit: recursion can continue until program memory is
exhausted.  The default is 1,000.
@endDefreg

@Defreq {warnscale, si}
Set the scaling indicator used in warnings to @var{si}.  Valid values
for @var{si} are @samp{u}, @samp{i}, @samp{c}, @samp{p}, and @samp{P}.
At startup, it is set to @samp{i}.
@endDefreq

@Defreq {spreadwarn, [@Var{limit}]}
Emit a @code{break} warning if the additional space inserted for each
space between words in an output line adjusted to both margins with
@w{@samp{.ad b}} is larger than or equal to @var{limit}.  A negative
value is treated as zero; an absent argument toggles the warning on and
off without changing @var{limit}.  The default scaling indicator is
@samp{m}.  At startup, @code{spreadwarn} is inactive and @var{limit} is
3@dmn{m}.

For example,

@Example
.spreadwarn 0.2m
@endExample

@noindent
causes a warning if @code{break} warnings are not suppressed and
@code{gtroff} must add 0.2@dmn{m} or more for each inter-word space in a
line.  @xref{Warnings}.
@endDefreq

@cindex warnings
GNU @code{troff} has command-line options for reporting warnings
(@option{-w}) and backtraces (@option{-b}) when a warning or an error
occurs.

@DefreqList {warn, [@Var{n}]}
@DefregListEndx {.warn}
@cindex warning level (@code{warn})
Select the categories, or ``types'', of reported warnings.
@var{n}@tie{}is the sum of the numeric codes associated with each
warning category that is to be enabled; all other categories are
disabled.  The categories and their associated codes are listed in
@ref{Warnings}.  For example, @samp{.warn 0} disables all warnings, and
@samp{.warn 1} disables all warnings except those about missing glyphs.
If no argument is given, all warning categories are enabled.

The read-only register @code{.warn} contains the sum of the numeric
codes of enabled warning categories.
@endDefreq

@menu
* Warnings::
@end menu

@c ---------------------------------------------------------------------

@c BEGIN Keep parallel with section "Warnings" of troff(1).
@c Caveat: the man page sorts them by name, not number.
@node Warnings,  , Debugging, Debugging
@subsection Warnings
@cindex warnings

Warning diagnostics emitted by GNU @code{troff} are divided into named,
numbered categories.  The name associated with each warning category is
used by the @option{-w} and @option{-W} options.  Each category is also
assigned a power of two; the sum of enabled category values is used by
the @code{warn} request and the @code{.warn} register.

Warnings of each category are produced under the following
circumstances.

@cindex categories, warning
@cindex warning categories
@table @samp
@item char
@itemx 1
An undefined glyph was requested for output.@footnote{@code{char} is a
misnomer since it reports missing glyphs---there are no ``missing''
input characters, only invalid ones.}  This category is enabled by
default.

@item number
@itemx 2
An invalid numeric expression was encountered.  This category is enabled
by default.
@xref{Numeric Expressions}.

@item break
@itemx 4
@cindex filling, and @code{break} warnings
@cindex mode, fill, and @code{break} warnings
A filled output line could not be broken such that its length was less
than the output line length @samp{\n[.l]}.  This category is enabled by
default.

@item delim
@itemx 8
The closing delimiter in an escape sequence was missing or mismatched.

@item el
@itemx 16
@cindex @code{ie} request, and warnings
@cindex @code{el} request, and warnings
The @code{el} request was encountered with no prior corresponding
@code{ie} request.  @xref{if-else}.

@item scale
@itemx 32
A scaling unit inappropriate to its context was used in a numeric
expression.

@item range
@itemx 64
A numeric expression was out of range for its context.

@item syntax
@itemx 128
A self-contradictory hyphenation mode was requested; an empty or
incomplete numeric expression was encountered; an operand to a numeric
operator was missing; an attempt was made to define a recursive, empty,
or nonsensical character class; or a @code{groff} extension conditional
expression operator was used while in compatibility mode.

@item di
@itemx 256
@cindex @code{di} request, and warnings
@cindex @code{da} request, and warnings
@cindex @code{box} request, and warnings
@cindex @code{boxa} request, and warnings
A @code{di}, @code{da}, @code{box}, or @code{boxa} request was invoked
without an argument when there was no current diversion.

@item mac
@itemx 512
@cindex @code{de}, @code{de1}, @code{dei} requests, and warnings
@cindex @code{am}, @code{am1}, @code{ami} requests, and warnings
@cindex @code{ds}, @code{ds1} requests, and warnings
@cindex @code{as}, @code{as1} requests, and warnings
@cindex @code{di} request, and warnings
@cindex @code{da} request, and warnings
@cindex @code{box}, @code{boxa} requests, and warnings
@cindex @code{\*}, and warnings
An undefined string, macro, or diversion was used.  When such an object
is dereferenced, an empty one of that name is automatically created.
So, unless it is later deleted, at most one warning is given for each.

This warning is also emitted upon an attempt to move an unplanted trap
macro (@pxref{Page Location Traps}).  In such cases, the unplanted macro
is @emph{not} dereferenced, so it is not created if it does not exist.

@item reg
@itemx 1024
@cindex @code{nr} request, and warnings
@cindex @code{\R}, and warnings
@cindex @code{\n}, and warnings
An undefined register was used.  When an undefined register is
dereferenced, it is automatically defined with a value of@tie{}0.  So,
unless it is later deleted, at most one warning is given for each.

@item tab
@itemx 2048
@cindex @code{\t}, and warnings
A tab character was encountered where a number was expected, or appeared
in an unquoted macro argument.

@item right-brace
@itemx 4096
@cindex @code{\@}}, and warnings
A right brace escape sequence @code{\@}} was encountered where a number
was expected.

@item missing
@itemx 8192
A request was invoked with a mandatory argument absent.

@item input
@itemx 16384
An invalid character occurred on the input stream.

@item escape
@itemx 32768
An unsupported escape sequence was encountered.

@item space
@itemx 65536
@cindex compatibility mode
A space was missing between a request or macro and its argument.  This
warning is produced when an undefined name longer than two characters is
encountered and the first two characters of the name constitute a
defined name.  No request is invoked, no macro called, and an empty
macro is not defined.  This category is enabled by default.  It never
occurs in compatibility mode.

@item font
@itemx 131072
A non-existent font was selected, or the selection was ignored because a
font selection escape sequence was used after the output line
continuation escape sequence on an input line.  This category is enabled
by default.

@item ig
@itemx 262144
An invalid escape sequence occurred in input ignored using the @code{ig}
request.  This warning category diagnoses a condition that is an error
when it occurs in non-ignored input.

@item color
@itemx 524288
An undefined color was selected, an attempt was made to define a color
using an unrecognized color space, an invalid component in a color
definition was encountered, or an attempt was made to redefine a default
color.

@item file
@itemx 1048576
An attempt was made to load a file that does not exist.  This category
is enabled by default.
@end table

Two warning names group other warning categories for convenience.

@table @samp
@item all
All warning categories except @samp{di}, @samp{mac} and @samp{reg}.
This shorthand is intended to produce all warnings that are useful with
macro packages written for @acronym{AT&T} @code{troff} and its
descendants, which have less fastidious diagnostics than GNU
@code{troff}.

@item w
All warning categories.  Authors of documents and macro packages
targeting @code{groff} are encouraged to use this setting.
@end table
@c END Keep parallel with section "Warnings" of troff(1).

@c =====================================================================

@codequotebacktick on
@codequoteundirected on

@node Implementation Differences, Safer Mode, Debugging, GNU troff Reference
@section Implementation Differences
@cindex implementation differences
@cindex differences in implementation
@cindex incompatibilities with @acronym{AT&T} @code{troff}

GNU @code{troff} has a number of features that cause incompatibilities
with documents written for other versions of @code{troff}.  Some GNU
extensions to @code{troff} have become supported by other
implementations.

@menu
* Safer Mode::
* Compatibility Mode::
* Other Differences::
@end menu

@c ---------------------------------------------------------------------

@node Safer Mode, Compatibility Mode, Implementation Differences, Implementation Differences
@subsection Safer Mode
@cindex safer mode
@cindex mode, safer

@cindex @code{pi} request, disabled by default
@cindex @code{sy} request, disabled by default
The formatter operates in ``safer'' mode by default; to mitigate risks
from untrusted input documents, the @code{pi} and @code{sy} requests are
disabled.  GNU @code{troff}'s @option{-U} option enables ``unsafe
mode'', restoring their function and enabling additional @code{groff}
extension requests, @code{open}, @code{opena}, and @code{pso}.
@xref{I/O}.

@c ---------------------------------------------------------------------

@node Compatibility Mode, Safer Mode, Other Differences, Implementation Differences
@subsection Compatibility Mode
@cindex compatibility mode
@cindex mode, compatibility

@cindex long names
@cindex names, long
@cindex @code{\*}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\n}, incompatibilities with @acronym{AT&T} @code{troff}
Long identifier names may be GNU @code{troff}'s most obvious innovation.
@acronym{AT&T} @code{troff} interprets @samp{.dsabcd} as defining a
string @samp{ab} with contents @samp{cd}.  Normally, GNU @code{troff}
interprets this as a call of a macro named @code{dsabcd}.
@acronym{AT&T} @code{troff} also interprets @samp{\*[} and @samp{\n[} as
an interpolation of a string or register, respectively, named @samp{[}.
In GNU @code{troff}, however, the @samp{[} is normally interpreted as
delimiting a long name.  In compatibility mode, GNU @code{troff}
interprets names in the traditional way; they thus can be two characters
long at most.

@DefreqList {cp, [@Var{n}]}
@DefregListEndx {.C}
If @var{n} is missing or non-zero, turn on compatibility mode;
otherwise, turn it off.

The read-only register @code{.C} is@tie{}1 if compatibility mode is on,
0@tie{}otherwise.

Compatibility mode can be also turned on with the @option{-C}
command-line option.
@endDefreq

@DefreqList {do, name}
@DefregListEndx {.cp}
The @code{do} request interprets the string, request, diversion, or
macro @var{name} (along with any further arguments) with compatibility
mode disabled.  Compatibility mode is restored (only if it was active)
when the @emph{expansion} of @var{name} is interpreted; that is, the
restored compatibility state applies to the contents of the macro,
string, or diversion @var{name} as well as data read from files or pipes
if @var{name} is any of the @code{so}, @code{soquiet}, @code{mso},
@code{msoquiet}, or @code{pso} requests.

The following example illustrates several aspects of @code{do} behavior.

@Example
.de mac1
FOO
..
.de1 mac2
groff
.mac1
..
.de mac3
compatibility
.mac1
..
.de ma
\\$1
..
.cp 1
.do mac1
.do mac2 \" mac2, defined with .de1, calls "mac1"
.do mac3 \" mac3 calls "ma" with argument "c1"
.do mac3 \[ti] \" groff syntax accepted in .do arguments
    @result{} FOO groff FOO compatibility c1 ~
@endExample

The read-only register @code{.cp}, meaningful only when dereferenced
from a @code{do} request, is@tie{}1 if compatibility mode was on when
the @code{do} request was encountered, and 0@tie{}if it was not.  This
register is specialized and may require a statement of rationale.

When writing macro packages or documents that use GNU @code{troff}
features and which may be mixed with other packages or documents that do
not---common scenarios include serial processing of man pages or use of
the @code{so} or @code{mso} requests---you may desire correct operation
regardless of compatibility mode enablement in the surrounding context.
It may occur to you to save the existing value of @samp{\n(.C} into a
register, say, @samp{_C}, at the beginning of your file, turn
compatibility mode off with @samp{.cp 0}, then restore it from that
register at the end with @samp{.cp \n(_C}.  At the same time, a modular
design of a document or macro package may lead you to multiple layers of
inclusion.  You cannot use the same register name everywhere lest you
``clobber'' the value from a preceding or enclosing context.  The
two-character register name space of @acronym{AT&T} @code{troff} is
confining and mnemonically challenging; you may wish to use the more
capacious name space of GNU @code{troff}.  However, attempting @samp{.nr
_my_saved_C \n(.C} will not work in compatibility mode; the register
name is too long.  ``This is exactly what @code{do} is for,'' you think,
@samp{.do nr _my_saved_C \n(.C}.  The foregoing will always save zero to
your register, because @code{do} turns compatibility mode @emph{off}
while it interprets its argument list.

@need 375 @c 250 < x < 500
To robustly save compatibility mode before switching it off, use

@Example
.do nr _my_saved_C \n[.cp]
.cp 0
@endExample

at the beginning of your file, followed by

@Example
.cp \n[_my_saved_C]
.do rr _my_saved_C
@endExample

at the end.  As in the C language, we all have to share one big
name space, so choose a register name that is unlikely to collide with
other uses.
@endDefreq

@cindex input level in delimited arguments
@cindex interpolation depth in delimited arguments
@cindex delimited arguments, incompatibilities with @acronym{AT&T} @code{troff}
Normally, GNU @code{troff} preserves the interpolation depth in
delimited arguments, but not in compatibility mode.

@Example
.ds xx '
\w'abc\*(xxdef'
    @result{} 168 @r{(normal mode on a terminal device)}
    @result{} 72def' @r{(compatibility mode on a terminal device)}
@endExample

@cindex @code{\f}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\H}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\s}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\S}, incompatibilities with @acronym{AT&T} @code{troff}
Furthermore, the escape sequences @code{\f}, @code{\H}, @code{\m},
@code{\M}, @code{\R}, @code{\s}, and @code{\S} are transparent for the
purpose of recognizing a control character at the beginning of a line
only in compatibility mode.  For example, this code produces bold output
in both cases, but the text differs.

@Example
.de xx
Hello!
..
\fB.xx\fP
    @result{} .xx @r{(normal mode)}
    @result{} Hello! @r{(compatibility mode)}
@endExample

@cindex @code{\s}, incompatibilities with @acronym{AT&T} @code{troff}
Normally, the syntax form @code{\s}@var{n} accepts only a single
character (a digit) for @var{n}, consistently with other forms that
originated in @acronym{AT&T} @code{troff}, like @code{\*}, @code{\$},
@code{\f}, @code{\g}, @code{\k}, @code{\n}, and @code{\z}.  In
compatibility mode only, a non-zero@tie{}@var{n} must be in the range
4--39.  Legacy documents relying upon this quirk of parsing@footnote{The
Graphic Systems C/A/T phototypesetter (the original device target for
@acronym{AT&T} @code{troff}) supported only a few discrete type sizes
in the range 6--36 points, so Ossanna contrived a special case in the
parser to do what the user must have meant.  Kernighan warned of this in
the 1992 revision of CSTR@tie{}#54 (§2.3), and more recently, McIlroy
referred to it as a ``living fossil''.} should be migrated to another
@code{\s} form.

@c ---------------------------------------------------------------------

@node Other Differences, , Compatibility Mode, Implementation Differences
@subsection Other Differences

@code{groff} request names unrecognized by other @code{troff}
implementations will likely be ignored by them; escape sequences that
are @code{groff} extensions are liable to be interpreted as if the
escape character were not present.
@cindex @code{\~}, incompatibilities with @acronym{AT&T} @code{troff}
For example, the adjustable, non-breaking escape sequence @code{\~}
@c BEGIN Keep in sync with groff_diff(7) and groff_man_style(7).
is also supported by Heirloom Doctools @code{troff} 050915 (September
2005), @code{mandoc} 1.9.5 (2009-09-21), @code{neatroff} (commit
1c6ab0f6e, 2016-09-13), and Plan@tie{}9 from User Space @code{troff}
(commit 93f8143600, 2022-08-12), but not by Solaris or Documenter's
Workbench @code{troff}s.
@c as of this writing, 2022-08-13
@c END Keep in sync with groff_diff(7) and groff_man_style(7).
@xref{Manipulating Filling and Adjustment}.

@cindex @code{\A}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\|}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\^}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\&}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\@{}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\@}}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\@key{SP}}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\'}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\`}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\-}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\_}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\!}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\%}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\c}, incompatibilities with @acronym{AT&T} @code{troff}
GNU @code{troff} does not allow the use of the escape sequences
@code{\|}, @code{\^}, @code{\&}, @code{\@{}, @code{\@}},
@code{\@key{SP}}, @code{\'}, @code{\`}, @code{\-}, @code{\_}, @code{\!},
@code{\%}, and @code{\c} in identifiers; @acronym{AT&T} @code{troff}
does.  The @code{\A} escape sequence (@pxref{Identifiers}) may be
helpful in avoiding use of these escape sequences in names.

@cindex adjustment to both margins, difference from @acronym{AT&T} @code{troff}
@cindex rivers
When adjusting to both margins, @acronym{AT&T} @code{troff} at first
adjusts spaces starting from the right; GNU @code{troff} begins from
the left.  Both implementations adjust spaces from opposite ends on
alternating output lines in this adjustment mode to prevent ``rivers''
in the text.

@cindex hyphenation, incompatibilities with @acronym{AT&T} @code{troff}
GNU @code{troff} does not always hyphenate words as @acronym{AT&T}
@code{troff} does.  The @acronym{AT&T} implementation uses a set of
hard-coded rules specific to English, while GNU @code{troff} uses
language-specific hyphenation pattern files derived from @TeX{}.
Furthermore, in old versions of @code{troff} there was a limited amount
of space to store hyphenation exceptions (arguments to the @code{hw}
request); GNU @code{troff} has no such restriction.

@cindex output device name string (@code{.T}), in other implementations
GNU @code{troff} predefines a string @code{.T} containing the argument
given to the @option{-T} command-line option, namely the current output
device (for example, @samp{pdf} or @samp{utf8}).  The existence of this
string is a common feature of post-CSTR@tie{}#54
@code{troff}s@footnote{DWB@tie{}3.3, Solaris, Heirloom Doctools, and
Plan@tie{}9 @code{troff} all support it.} but valid values are specific
to each implementation.

@cindex removal of read-only registers, incompatibility with @acronym{AT&T} @code{troff}
@cindex register, read-only, removal, incompatibility with @acronym{AT&T} @code{troff}
@cindex read-only register removal, incompatibility with @acronym{AT&T} @code{troff}
@acronym{AT&T} @code{troff} ignored attempts to remove read-only
registers; GNU @code{troff} honors such requests.  @xref{Built-in
Registers}.

@cindex output device usage register (@code{.T}), incompatibility with @acronym{AT&T} @code{troff}
The (read-only) register @code{.T} interpolates@tie{}1 if GNU
@code{troff} is called with the @option{-T} command-line option, and
0@tie{}otherwise.  This behavior differs from AT&T @code{troff}, which
interpolated@tie{}1 only if @code{nroff} was the formatter and was
called with @option{-T}.

@cindex @code{lf} request, incompatibilities with @acronym{AT&T} @code{troff}
@acronym{AT&T} @code{troff} and other implementations handle the
@code{lf} request differently.  For them, its @var{line} argument
changes the line number of the @emph{current} line.

@cindex environment availability and naming, incompatibilities with
@acronym{AT&T} @code{troff} had only environments named @samp{0},
@samp{1}, and @samp{2}.  In GNU @code{troff}, any number of environments
may exist, using any valid identifiers for their names
(@pxref{Identifiers}.)

@cindex fractional point sizes
@cindex fractional type sizes
@cindex point sizes, fractional
@cindex type sizes, fractional
@cindex sizes, fractional
@cindex @code{ps} request, incompatibilities with @acronym{AT&T} @code{troff}
Fractional type sizes cause one noteworthy incompatibility.  In
@acronym{AT&T} @code{troff} the @code{ps} request ignores scale
indicators and thus @samp{.ps 10u} sets the type size to
10@tie{}points, whereas in GNU @code{troff} it sets the type size to
10@tie{}@emph{scaled} points.  @xref{Using Fractional Type Sizes}.

@cindex @code{ab} request, incompatibilities with @acronym{AT&T} @code{troff}
The @code{ab} request differs from @acronym{AT&T} @code{troff}:
GNU @code{troff} writes no message to the standard error stream if no
arguments are given, and it exits with a failure status instead of a
successful one.

@cindex @code{bp} request, incompatibilities with @acronym{AT&T} @code{troff}
The @code{bp} request differs from @acronym{AT&T} @code{troff}:
GNU @code{troff} does not accept a scaling indicator on the argument, a
page number; the former (somewhat uselessly) does.

@cindex @code{pm} request, incompatibilities with @acronym{AT&T} @code{troff}
The @code{pm} request differs from @acronym{AT&T} @code{troff}:
GNU @code{troff} reports the sizes of macros, strings, and diversions in
bytes and ignores an argument to report only the sum of the sizes.

@cindex @code{ss} request, incompatibilities with @acronym{AT&T} @code{troff}
Unlike @acronym{AT&T} @code{troff}, GNU @code{troff} does not ignore the
@code{ss} request if the output is a terminal device; instead, the
values of minimal inter-word and additional inter-sentence space are
each rounded down to the nearest multiple of@tie{}12.

@cindex @code{bd} request, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{cs} request, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{tr} request, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{fp} request, incompatibilities with @acronym{AT&T} @code{troff}
@cindex input characters and output glyphs, compatibility with @acronym{AT&T} @code{troff}
@cindex output glyphs, and input characters, compatibility with @acronym{AT&T} @code{troff}
@cindex characters, input, and output glyphs, compatibility with @acronym{AT&T} @code{troff}
@cindex glyphs, output, and input characters, compatibility with @acronym{AT&T} @code{troff}
In GNU @code{troff} there is a fundamental difference between
(unformatted) characters and (formatted) glyphs.  Everything that
affects how a glyph is output is stored with the glyph node; once a
glyph node has been constructed, it is unaffected by any subsequent
requests that are executed, including @code{bd}, @code{cs}, @code{tkf},
@code{tr}, or @code{fp} requests.  Normally, glyphs are constructed from
characters immediately before the glyph is added to an output line.
Macros, diversions, and strings are all, in fact, the same type of
object; they contain a sequence of intermixed character and glyph nodes.
Special characters transform from one to the other:@: before being added
to the output, they behave as characters; afterward, they are glyphs.  A
glyph node does not behave like a character node when it is processed by
a macro:@: it does not inherit any of the special properties that the
character from which it was constructed might have had.  For example,
the input

@Example
.di x
\\\\
.br
.di
.x
@endExample

@noindent
produces @samp{\\} in GNU @code{troff}.  Each pair of backslashes
becomes one backslash @emph{glyph}; the resulting backslashes are thus
not interpreted as escape @emph{characters} when they are reread as the
diversion is output.  @acronym{AT&T} @code{troff} @emph{would} interpret
them as escape characters when rereading them and end up printing one
@samp{\}.

@cindex printing backslash (@code{\\}, @code{\e}, @code{\E}, @code{\[rs]})
@cindex backslash, printing (@code{\\}, @code{\e}, @code{\E}, @code{\[rs]})
@cindex @code{\e}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\!}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex @code{\?}, incompatibilities with @acronym{AT&T} @code{troff}
@cindex transparent output, incompatibilities with @acronym{AT&T} @code{troff}
@cindex output, transparent, incompatibilities with @acronym{AT&T} @code{troff}
One correct way to obtain a printable backslash in most documents is to
use the @code{\e} escape sequence; this always prints a single instance
of the current escape character,@footnote{Naturally, if you've changed
the escape character, you need to prefix the @code{e} with whatever it
is---and you'll likely get something other than a backslash in the
output.} regardless of whether or not it is used in a diversion; it
also works in both GNU @code{troff} and @acronym{AT&T} @code{troff}.

The other correct way, appropriate in contexts independent of the
backslash's common use as a @code{troff} escape character---perhaps in
discussion of character sets or other programming languages---is
the character escape @code{\(rs} or @code{\[rs]}, for ``reverse
solidus'', from its name in the @acronym{ECMA-6} (@acronym{ISO/IEC} 646)
standard.@footnote{The @code{rs} special character identifier was not
defined in @acronym{AT&T} @code{troff}'s font description files, but is
in those if its lineal descendant, Heirloom Doctools @code{troff}, as of
the latter's 060716 release (July 2006).}

To store an escape sequence in a diversion that is interpreted when the
diversion is reread, either use the traditional @code{\!} transparent
output facility, or, if this is unsuitable, the new @code{\?} escape
sequence.  @xref{Diversions} and @ref{Gtroff Internals}.

In the somewhat pathological case where a diversion exists containing a
partially collected line and a partially collected line at the top-level
diversion has never existed, @acronym{AT&T} @code{troff} will output the
partially collected line at the end of input; GNU @code{troff} will not.

@codequotebacktick off
@codequoteundirected off



@c =====================================================================
@c =====================================================================

@node File Formats, Copying This Manual, GNU troff Reference, Top
@chapter File Formats
@cindex file formats
@cindex formats, file

All files read and written by @code{gtroff} are text files.  The
following two sections describe their format.

@menu
* gtroff Output::
* Device and Font Description Files::
@end menu


@c =====================================================================

@c BEGIN TODO: Make parallel with groff_out(5).
@node gtroff Output, Device and Font Description Files, File Formats, File Formats
@section @code{gtroff} Output
@cindex @code{gtroff}, output
@cindex output, @code{gtroff}

This section describes the @code{groff} intermediate output format
produced by GNU @code{troff}.

As @code{groff} is a wrapper program around GNU @code{troff} and
automatically calls an output driver (or ``postprocessor''), this output
does not show up normally.  This is why it is called
@emph{intermediate}.  @code{groff} provides the option @option{-Z} to
inhibit postprocessing such that the produced intermediate output is
sent to standard output just as it is when calling GNU @code{troff}
directly.

@cindex @code{troff} output
@cindex output, @code{troff}
@cindex intermediate output
@cindex output, intermediate
Here, the term @dfn{troff output} describes what is output by
GNU @code{troff}, while @dfn{intermediate output} refers to the language
that is accepted by the parser that prepares this output for the output
drivers.  This parser handles whitespace more flexibly than
@acronym{AT&T}'s implementation and implements obsolete elements for
compatibility; otherwise, both formats are the same.@footnote{The parser
and postprocessor for intermediate output can be found in the file@*
@file{@var{groff-source-dir}/src/libs/libdriver/input.cpp}.}

The main purpose of the intermediate output concept is to facilitate the
development of postprocessors by providing a common programming
interface for all devices.  It has a language of its own that is
completely different from the @code{gtroff} language.  While the
@code{gtroff} language is a high-level programming language for text
processing, the intermediate output language is a kind of low-level
assembler language by specifying all positions on the page for writing
and drawing.

The intermediate output produced by @code{gtroff} is fairly readable,
while output from @acronym{AT&T} @code{troff} is rather hard to
understand because of strange habits that are still supported, but not
used any longer by @code{gtroff}.

@menu
* Language Concepts::
* Command Reference::
* Intermediate Output Examples::
* Output Language Compatibility::
@end menu

@c ---------------------------------------------------------------------

@node Language Concepts, Command Reference, gtroff Output, gtroff Output
@subsection Language Concepts

The fundamental operation of the GNU @code{troff} formatter is the
translation of the @code{groff} input language into a device-independent
form primarily concerned with what has to be written or drawn at
specific positions on the output device.  This language is simple and
imperative.  In the following discussion, the term @dfn{command} always
refers to this intermediate output language, and never to the
@code{groff} language intended for direct use by document authors.
Intermediate output commands comprise several categories: glyph output;
font, color, and text size selection; motion of the printing position;
page advancement; drawing of geometric primitives; and device control
commands, a catch-all for operations not easily classified as any of the
foregoing, such as directives to start and stop output, identify the
intended output device, or place URL hyperlinks in supported output
formats.

@menu
* Separation::
* Argument Units::
* Document Parts::
@end menu

@node Separation, Argument Units, Language Concepts, Language Concepts
@subsubsection Separation

@acronym{AT&T} @code{troff} output has strange requirements regarding
whitespace.  The @code{gtroff} output parser, however, is more tolerant,
making whitespace maximally optional.  Such characters, i.e., the tab,
space, and newline, always have a syntactical meaning.  They are never
printable because spacing within the output is always done by
positioning commands.

Any sequence of space or tab characters is treated as a single
@dfn{syntactical space}.  It separates commands and arguments, but is
only required when there would occur a clashing between the command code
and the arguments without the space.  Most often, this happens when
variable-length command names, arguments, argument lists, or command
clusters meet.  Commands and arguments with a known, fixed length need
not be separated by syntactical space.

A line break is a syntactical element, too.  Every command argument can
be followed by whitespace, a comment, or a newline character.  Thus a
@dfn{syntactical line break} is defined to consist of optional
syntactical space that is optionally followed by a comment, and a
newline character.

The normal commands, those for positioning and text, consist of a single
letter taking a fixed number of arguments.  For historical reasons, the
parser allows stacking of such commands on the same line, but
fortunately, in @code{gtroff}'s intermediate output, every command with
at least one argument is followed by a line break, thus providing
excellent readability.

The other commands---those for drawing and device controlling---have a
more complicated structure; some recognize long command names, and some
take a variable number of arguments.  So all @samp{D} and @samp{x}
commands were designed to request a syntactical line break after their
last argument.  Only one command, @w{@samp{x X}}, has an argument that
can span several input lines; all other commands must have all of
their arguments on the same line as the command, i.e., the arguments may
not be split by a line break.

Empty lines (these are lines containing only space and/or a comment),
can occur everywhere.  They are just ignored.

@node Argument Units, Document Parts, Separation, Language Concepts
@subsubsection Argument Units

Some commands take integer arguments that are assumed to represent
values in a measurement unit, but the letter for the corresponding scale
indicator is not written with the output command arguments.  Most
commands assume the scaling indicator @samp{u}, the basic unit of the
device, some use @samp{z}, the scaled point unit of the device, while
others, such as the color commands, expect plain integers.

Single characters can have the eighth bit set, as can the names of
fonts and special characters.  The names of characters and fonts can be
of arbitrary length.  A character that is to be printed is always in
the current font.

A string argument is always terminated by the next whitespace character
(space, tab, or newline); an embedded @samp{#} character is regarded as
part of the argument, not as the beginning of a comment command.  An
integer argument is already terminated by the next non-digit character,
which then is regarded as the first character of the next argument or
command.

@node Document Parts,  , Argument Units, Language Concepts
@subsubsection Document Parts

A correct intermediate output document consists of two parts, the
@dfn{prologue} and the @dfn{body}.

The task of the prologue is to set the general device parameters using
three exactly specified commands.  @code{gtroff}'s prologue is
guaranteed to consist of the following three lines (in that order):

@Example
x T @var{device}
x res @var{n} @var{h} @var{v}
x init
@endExample

@noindent
with the arguments set as outlined in @ref{Device Control Commands}.
The parser for the intermediate output format is able to interpret
additional whitespace and comments as well even in the prologue.

The body is the main section for processing the document data.
Syntactically, it is a sequence of any commands different from the ones
used in the prologue.  Processing is terminated as soon as the first
@w{@samp{x stop}} command is encountered; the last line of any
@code{gtroff} intermediate output always contains such a command.

Semantically, the body is page oriented.  A new page is started by a
@samp{p} command.  Positioning, writing, and drawing commands are always
done within the current page, so they cannot occur before the first
@samp{p} command.  Absolute positioning (by the @samp{H} and @samp{V}
commands) is done relative to the current page; all other positioning is
done relative to the current location within this page.

@c ---------------------------------------------------------------------

@node Command Reference, Intermediate Output Examples, Language Concepts, gtroff Output
@subsection Command Reference

This section describes all intermediate output commands, both from
@acronym{AT&T} @code{troff} as well as the @code{gtroff} extensions.

@menu
* Comment Command::
* Simple Commands::
* Graphics Commands::
* Device Control Commands::
* Obsolete Command::
@end menu

@node Comment Command, Simple Commands, Command Reference, Command Reference
@subsubsection Comment Command

@table @code
@item #@var{anything}@angles{end of line}
A comment.  Ignore any characters from the @samp{#} character up to the
next newline character.

This command is the only possibility for commenting in the intermediate
output.  Each comment can be preceded by arbitrary syntactical space;
every command can be terminated by a comment.
@end table

@node Simple Commands, Graphics Commands, Comment Command, Command Reference
@subsubsection Simple Commands

The commands in this subsection have a command code consisting of a
single character, taking a fixed number of arguments.  Most of them are
commands for positioning and text writing.  These commands are tolerant
of whitespace.  Optionally, syntactical space can be inserted before,
after, and between the command letter and its arguments.  All of these
commands are stackable; i.e., they can be preceded by other simple
commands or followed by arbitrary other commands on the same line.  A
separating syntactical space is only necessary when two integer
arguments would clash or if the preceding argument ends with a string
argument.

@table @code
@ignore
.if (\n[@USE_ENV_STACK] == 1) \{\
.command {
Open a new environment by copying the actual device configuration data
to the environment stack.
.
The current environment is setup by the device specification and
manipulated by the setting commands.
.
.
.command }
Close the actual environment (opened by a preceding
.BR { \~command)
and restore the previous environment from the environment
stack as the actual device configuration data.
.
\}              \" endif @USE_ENV_STACK
@end ignore

@item C @var{xxx}@angles{whitespace}
Print a special character named @var{xxx}.  The trailing syntactical
space or line break is necessary to allow glyph names of arbitrary
length.  The glyph is printed at the current print position; the glyph's
size is read from the font file.  The print position is not changed.

@item c @var{g}
Print glyph@tie{}@var{g} at the current print
position;@footnote{@samp{c} is actually a misnomer since it outputs a
glyph.} the glyph's size is read from the font file.  The print position
is not changed.

@item f @var{n}
Set font to font number@tie{}@var{n} (a non-negative integer).

@item H @var{n}
Move right to the absolute vertical position@tie{}@var{n} (a
non-negative integer in basic units @samp{u} relative to left edge of
current page.

@item h @var{n}
Move @var{n} (a non-negative integer) basic units @samp{u} horizontally
to the right.  The @acronym{AT&T} @code{troff} manual allows negative
values for @var{n} also, but GNU @code{troff} doesn't use them.

@item m @var{color-scheme} @r{[}@var{component} @dots{}@r{]}
Set the color for text (glyphs), line drawing, and the outline of
graphic objects using different color schemes; the analogous command
for the filling color of graphic objects is @samp{DF}.  The color
components are specified as integer arguments between 0 and 65535.  The
number of color components and their meaning vary for the different
color schemes.  These commands are generated by @code{gtroff}'s escape
sequence @code{\m}.  No position changing.  These commands are a
@code{gtroff} extension.

@table @code
@item mc @var{cyan} @var{magenta} @var{yellow}
Set color using the CMY color scheme, having the 3@tie{}color components
@var{cyan}, @var{magenta}, and @var{yellow}.

@item md
Set color to the default color value (black in most cases).  No
component arguments.

@item mg @var{gray}
Set color to the shade of gray given by the argument, an integer between
0 (black) and 65535 (white).

@item mk @var{cyan} @var{magenta} @var{yellow} @var{black}
Set color using the CMYK color scheme, having the 4@tie{}color
components @var{cyan}, @var{magenta}, @var{yellow}, and @var{black}.

@item mr @var{red} @var{green} @var{blue}
Set color using the RGB color scheme, having the 3@tie{}color components
@var{red}, @var{green}, and @var{blue}.
@end table

@item N @var{n}
Print glyph with index@tie{}@var{n} (a non-negative integer) of the
current font.  This command is a @code{gtroff} extension.

@item n @var{b} @var{a}
Inform the device about a line break, but no positioning is done by this
command.  In @acronym{AT&T} @code{troff}, the integer arguments @var{b}
and@tie{}@var{a} informed about the space before and after the current
line to make the intermediate output more human readable without
performing any action.  In @code{groff}, they are just ignored, but they
must be provided for compatibility reasons.

@item p @var{n}
Begin a new page in the output.  The page number is set to@tie{}@var{n}.
This page is completely independent of pages formerly processed even if
those have the same page number.  The vertical position on the output is
automatically set to@tie{}0.  All positioning, writing, and drawing is
always done relative to a page, so a @samp{p} command must be issued
before any of these commands.

@item s @var{n}
Set type size to @var{n}@tie{}scaled points (this is unit @samp{z}).
@acronym{AT&T} @code{troff} used the unit points (@samp{p}) instead.
@xref{Output Language Compatibility}.

@item t @var{xxx}@angles{whitespace}
@itemx t @var{xxx} @var{dummy-arg}@angles{whitespace}
Print a word, i.e., a sequence of characters @var{xxx} representing
output glyphs which names are single characters, terminated by a space
character or a line break; an optional second integer argument is
ignored (this allows the formatter to generate an even number of
arguments).  The first glyph should be printed at the current position,
the current horizontal position should then be increased by the width of
the first glyph, and so on for each glyph.  The widths of the glyphs are
read from the font file, scaled for the current type size, and rounded
to a multiple of the horizontal motion quantum.  Special characters
cannot be printed using this command (use the @samp{C} command for
special characters).  This command is a @code{gtroff} extension; it is
only used for devices whose @file{DESC} file contains the
@code{tcommand} keyword (@pxref{DESC File Format}).

@item u @var{n} @var{xxx}@angles{whitespace}
Print word with track kerning.  This is the same as the @samp{t} command
except that after printing each glyph, the current horizontal position
is increased by the sum of the width of that glyph and@tie{}@var{n} (an
integer in basic units @samp{u}).  This command is a @code{gtroff}
extension; it is only used for devices whose @file{DESC} file contains
the @code{tcommand} keyword (@pxref{DESC File Format}).

@item V @var{n}
Move down to the absolute vertical position@tie{}@var{n} (a non-negative
integer in basic units @samp{u}) relative to upper edge of current page.

@item v @var{n}
Move @var{n}@tie{}basic units @samp{u} down (@var{n} is a non-negative
integer).  The @acronym{AT&T} @code{troff} manual allows negative values
for @var{n} also, but GNU @code{troff} doesn't use them.

@item w
Describe an adjustable space. This performs no action; it is present for
documentary purposes.  The spacing itself must be performed explicitly
by a move command.
@end table

@node Graphics Commands, Device Control Commands, Simple Commands, Command Reference
@subsubsection Graphics Commands

Each graphics or drawing command in the intermediate output starts with
the letter @samp{D}, followed by one or two characters that specify a
subcommand; this is followed by a fixed or variable number of integer
arguments that are separated by a single space character.  A @samp{D}
command may not be followed by another command on the same line (apart
from a comment), so each @samp{D} command is terminated by a syntactical
line break.

@code{gtroff} output follows the classical spacing rules (no space
between command and subcommand, all arguments are preceded by a single
space character), but the parser allows optional space between the
command letters and makes the space before the first argument optional.
As usual, each space can be any sequence of tab and space characters.

Some graphics commands can take a variable number of arguments.  In this
case, they are integers representing a size measured in basic units
@samp{u}.  The arguments called @var{h1}, @var{h2}, @dots{}, @var{hn}
stand for horizontal distances where positive means right, negative
left.  The arguments called @var{v1}, @var{v2}, @dots{}, @var{vn} stand
for vertical distances where positive means down, negative up.  All
these distances are offsets relative to the current location.

Each graphics command directly corresponds to a similar @code{gtroff}
@code{\D} escape sequence.  @xref{Drawing Requests}.

Unknown @samp{D} commands are assumed to be device-specific.  Its
arguments are parsed as strings; the whole information is then sent to
the postprocessor.

In the following command reference, the syntax element @angles{line
break} means a syntactical line break as defined above.

@table @code
@item D~ @var{h1} @var{v1} @var{h2} @var{v2} @dots{} @var{hn} @var{vn}@angles{line break}
Draw B-spline from current position to offset (@var{h1},@var{v1}), then
to offset (@var{h2},@var{v2}), if given, etc., up to
(@var{hn},@var{vn}).  This command takes a variable number of argument
pairs; the current position is moved to the terminal point of the drawn
curve.

@item Da @var{h1} @var{v1} @var{h2} @var{v2}@angles{line break}
Draw arc from current position to
(@var{h1},@var{v1})@math{+}(@var{h2},@var{v2}) with center at
(@var{h1},@var{v1}); then move the current position to the final point
of the arc.

@item DC @var{d}@angles{line break}
@itemx DC @var{d} @var{dummy-arg}@angles{line break}
Draw a solid circle using the current fill color with
diameter@tie{}@var{d} (integer in basic units @samp{u}) with leftmost
point at the current position; then move the current position to the
rightmost point of the circle.  An optional second integer argument is
ignored (this allows the formatter to generate an even number of
arguments).  This command is a @code{gtroff} extension.

@item Dc @var{d}@angles{line break}
Draw circle line with diameter@tie{}@var{d} (integer in basic units
@samp{u}) with leftmost point at the current position; then move the
current position to the rightmost point of the circle.

@item DE @var{h} @var{v}@angles{line break}
Draw a solid ellipse in the current fill color with a horizontal
diameter of@tie{}@var{h} and a vertical diameter of@tie{}@var{v} (both
integers in basic units @samp{u}) with the leftmost point at the current
position; then move to the rightmost point of the ellipse.  This command
is a @code{gtroff} extension.

@item De @var{h} @var{v}@angles{line break}
Draw an outlined ellipse with a horizontal diameter of@tie{}@var{h} and
a vertical diameter of@tie{}@var{v} (both integers in basic units
@samp{u}) with the leftmost point at current position; then move to the
rightmost point of the ellipse.

@item DF @var{color-scheme} @r{[}@var{component} @dots{}@r{]}@angles{line break}
Set fill color for solid drawing objects using different color schemes;
the analogous command for setting the color of text, line graphics, and
the outline of graphic objects is @samp{m}.  The color components are
specified as integer arguments between 0 and 65535.  The number of color
components and their meaning vary for the different color schemes.
These commands are generated by @code{gtroff}'s escape sequences
@samp{\D'F @dots{}'} and @code{\M} (with no other corresponding
graphics commands).  No position changing.  This command is a
@code{gtroff} extension.

@table @code
@item DFc @var{cyan} @var{magenta} @var{yellow}@angles{line break}
Set fill color for solid drawing objects using the CMY color scheme,
having the 3@tie{}color components @var{cyan}, @var{magenta}, and
@var{yellow}.

@item DFd@angles{line break}
Set fill color for solid drawing objects to the default fill color value
(black in most cases).  No component arguments.

@item DFg @var{gray}@angles{line break}
Set fill color for solid drawing objects to the shade of gray given by
the argument, an integer between 0 (black) and 65535 (white).

@item DFk @var{cyan} @var{magenta} @var{yellow} @var{black}@angles{line break}
Set fill color for solid drawing objects using the CMYK color scheme,
having the 4@tie{}color components @var{cyan}, @var{magenta},
@var{yellow}, and @var{black}.

@item DFr @var{red} @var{green} @var{blue}@angles{line break}
Set fill color for solid drawing objects using the RGB color scheme,
having the 3@tie{}color components @var{red}, @var{green}, and
@var{blue}.
@end table

@item Df @var{n}@angles{line break}
The argument@tie{}@var{n} must be an integer in the range @math{-32767}
to 32767.

@table @asis
@item @math{0 @leq{} @var{n} @leq{} 1000}
Set the color for filling solid drawing objects to a shade of gray,
where 0 corresponds to solid white, 1000 (the default) to solid black,
and values in between to intermediate shades of gray; this is obsoleted
by command @samp{DFg}.

@item @math{@var{n} < 0} or @math{@var{n} > 1000}
Set the filling color to the color that is currently being used for the
text and the outline, see command @samp{m}.  For example, the command
sequence

@Example
mg 0 0 65535
Df -1
@endExample

@noindent
sets all colors to blue.
@end table

@noindent
No position changing.  This command is a @code{gtroff} extension.

@item Dl @var{h} @var{v}@angles{line break}
Draw line from current position to offset (@var{h},@var{v}) (integers in
basic units @samp{u}); then set current position to the end of the drawn
line.

@item Dp @var{h1} @var{v1} @var{h2} @var{v2} @dots{} @var{hn} @var{vn}@angles{line break}
Draw a polygon line from current position to offset (@var{h1},@var{v1}),
from there to offset (@var{h2},@var{v2}), etc., up to offset
(@var{hn},@var{vn}), and from there back to the starting position.  For
historical reasons, the position is changed by adding the sum of all
arguments with odd index to the actual horizontal position and the even
ones to the vertical position.  Although this doesn't make sense it is
kept for compatibility.
@ignore
As the polygon is closed, the end of drawing is the starting point, so
the position doesn't change.
@end ignore
This command is a @code{gtroff} extension.

@item DP @var{h1} @var{v1} @var{h2} @var{v2} @dots{} @var{hn} @var{vn}@angles{line break}
Draw a solid polygon in the current fill color rather than an outlined
polygon, using the same arguments and positioning as the corresponding
@samp{Dp} command.
@ignore
No position changing.
@end ignore
This command is a @code{gtroff} extension.

@item Dt @var{n}@angles{line break}
Set the current line thickness to@tie{}@var{n} (an integer in basic
units @samp{u}) if @math{@var{n}>0}; if @math{@var{n}=0} select the
smallest available line thickness; if @math{@var{n}<0} set the line
thickness proportional to the type size (this is the default before the
first @samp{Dt} command was specified).  For historical reasons, the
horizontal position is changed by adding the argument to the actual
horizontal position, while the vertical position is not changed.
Although this doesn't make sense it is kept for compatibility.
@ignore
No position changing.
@end ignore
This command is a @code{gtroff} extension.
@end table

@node Device Control Commands, Obsolete Command, Graphics Commands, Command Reference
@subsubsection Device Control Commands

Each device control command starts with the letter @samp{x}, followed by
a space character (optional or arbitrary space or tab in @code{gtroff})
and a subcommand letter or word; each argument (if any) must be preceded
by a syntactical space.  All @samp{x} commands are terminated by a
syntactical line break; no device control command can be followed by
another command on the same line (except a comment).

The subcommand is basically a single letter, but to increase
readability, it can be written as a word, i.e., an arbitrary sequence of
characters terminated by the next tab, space, or newline character.  All
characters of the subcommand word but the first are simply ignored.  For
example, @code{gtroff} outputs the initialization command @w{@samp{x i}}
as @w{@samp{x init}} and the resolution command @w{@samp{x r}} as
@w{@samp{x res}}.

In the following, the syntax element @angles{line break} means a
syntactical line break (@pxref{Separation}).

@table @code
@item xF @var{name}@angles{line break}
The @samp{F} stands for @var{Filename}.

Use @var{name} as the intended name for the current file in error
reports.  This is useful for remembering the original file name when
@code{gtroff} uses an internal piping mechanism.  The input file is not
changed by this command.  This command is a @code{gtroff} extension.

@item xf @var{n} @var{s}@angles{line break}
The @samp{f} stands for @var{font}.

Mount font position@tie{}@var{n} (a non-negative integer) with font
named@tie{}@var{s} (a text word).  @xref{Font Positions}.

@item xH @var{n}@angles{line break}
The @samp{H} stands for @var{Height}.

Set glyph height to@tie{}@var{n} (a positive integer in scaled points
@samp{z}).  @acronym{AT&T} @code{troff} uses the unit points (@samp{p})
instead.  @xref{Output Language Compatibility}.

@item xi@angles{line break}
The @samp{i} stands for @var{init}.

Initialize device.  This is the third command of the prologue.

@item xp@angles{line break}
The @samp{p} stands for @var{pause}.

Parsed but ignored.  The @acronym{AT&T} @code{troff} manual documents
this command as

@display
pause device, can be restarted
@end display

but GNU @code{troff} output drivers do nothing with this command.

@item xr @var{n} @var{h} @var{v}@angles{line break}
The @samp{r} stands for @var{resolution}.

Resolution is@tie{}@var{n}, while @var{h} is the minimal horizontal
motion, and @var{v} the minimal vertical motion possible with this
device; all arguments are positive integers in basic units @samp{u} per
inch.  This is the second command of the prologue.

@item xS @var{n}@angles{line break}
The @samp{S} stands for @var{Slant}.

Set slant to@tie{}@var{n} (an integer in basic units @samp{u}).

@item xs@angles{line break}
The @samp{s} stands for @var{stop}.

Terminates the processing of the current file; issued as the last
command of any intermediate @code{troff} output.

@item xt@angles{line break}
The @samp{t} stands for @var{trailer}.

Generate trailer information, if any.  In GNU @code{troff}, this is
ignored.

@item xT @var{xxx}@angles{line break}
The @samp{T} stands for @var{Typesetter}.

Set the name of the output driver to @var{xxx}, a sequence of
non-whitespace characters terminated by whitespace.  The possible names
correspond to those of @code{groff}'s @option{-T} option.  This is the
first command of the prologue.

@item xu @var{n}@angles{line break}
The @samp{u} stands for @var{underline}.

Configure underlining of spaces.  If @var{n} is@tie{}1, start
underlining of spaces; if @var{n} is@tie{}0, stop underlining of spaces.
This is needed for the @code{cu} request in @code{nroff} mode and is
ignored otherwise.  This command is a @code{gtroff} extension.

@item xX @var{anything}@angles{line break}
The @samp{x} stands for @var{X-escape}.

Send string @var{anything} uninterpreted to the device.  If the line
following this command starts with a @samp{+} character this line is
interpreted as a continuation line in the following sense.  The @samp{+}
is ignored, but a newline character is sent instead to the device, the
rest of the line is sent uninterpreted.  The same applies to all
following lines until the first character of a line is not a @samp{+}
character.  This command is generated by the @code{gtroff} escape
sequence @code{\X}.  The line-continuing feature is a @code{gtroff}
extension.
@end table

@node Obsolete Command,  , Device Control Commands, Command Reference
@subsubsection Obsolete Command
In @acronym{AT&T} @code{troff} output, the writing of a single glyph is
mostly done by a very strange command that combines a horizontal move
and a single character giving the glyph name.  It doesn't have a command
code, but is represented by a 3-character argument consisting of exactly
2@tie{}digits and a character.

@table @asis
@item @var{dd}@var{g}
Move right @var{dd} (exactly two decimal digits) basic units @samp{u},
then print glyph@tie{}@var{g} (represented as a single character).

In GNU @code{troff}, arbitrary syntactical space around and within this
command is allowed.  Only when a preceding command on the same line ends
with an argument of variable length is a separating space obligatory.
In @acronym{AT&T} @code{troff}, large clusters of these and other
commands are used, mostly without spaces; this made such output almost
unreadable.
@end table

For modern high-resolution devices, this command does not make sense
because the width of the glyphs can become much larger than two decimal
digits.  In @code{gtroff}, this is only used for the devices @code{X75},
@code{X75-12}, @code{X100}, and @code{X100-12}.  For other devices, the
commands @samp{t} and @samp{u} provide a better functionality.

@c ---------------------------------------------------------------------

@codequotebacktick on
@codequoteundirected on

@node Intermediate Output Examples, Output Language Compatibility, Command Reference, gtroff Output
@subsection Intermediate Output Examples

This section presents the intermediate output generated from the same
input for three different devices.  The input is the sentence @samp{hell
world} fed into @code{gtroff} on the command line.

@table @asis
@item High-resolution device @code{ps}

This is the standard output of @code{gtroff} if no @option{-T} option is
given.

@example
@group
shell> echo "hell world" | groff -Z -T ps

x T ps
x res 72000 1 1
x init
@end group
p1
x font 5 TR
f5
s10000
V12000
H72000
thell
wh2500
tw
H96620
torld
n12000 0
@group
x trailer
V792000
x stop
@end group
@end example

@noindent
This output can be fed into @code{grops} to get its representation as a
PostScript file.

@item Low-resolution device @code{latin1}

This is similar to the high-resolution device except that the
positioning is done at a minor scale.  Some comments (lines starting
with @samp{#}) were added for clarification; they were not generated by
the formatter.

@example
@group
shell> echo "hell world" | groff -Z -T latin1

# prologue
x T latin1
x res 240 24 40
x init
@end group
# begin a new page
p1
# font setup
x font 1 R
f1
s10
# initial positioning on the page
V40
H0
# write text 'hell'
thell
# inform about space, and issue a horizontal jump
wh24
# write text 'world'
tworld
# announce line break, but do nothing because...
n40 0
@group
# ...the end of the document has been reached
x trailer
V2640
x stop
@end group
@end example

@noindent
This output can be fed into @code{grotty} to get a formatted text
document.

@item @acronym{AT&T} @code{troff} output
Since a computer monitor has a much lower resolution than modern
printers, the intermediate output for X11 devices can use the
jump-and-write command with its 2-digit displacements.

@example
@group
shell> echo "hell world" | groff -Z -T X100

x T X100
x res 100 1 1
x init
@end group
p1
x font 5 TR
f5
s10
V16
H100
# write text with jump-and-write commands
ch07e07l03lw06w11o07r05l03dh7
n16 0
@group
x trailer
V1100
x stop
@end group
@end example

@noindent
This output can be fed into @code{xditview} or @code{gxditview} for
displaying in@tie{}X.

Due to the obsolete jump-and-write command, the text clusters in the
@acronym{AT&T} @code{troff} output are almost unreadable.
@end table

@codequotebacktick off
@codequoteundirected off

@c ---------------------------------------------------------------------

@node Output Language Compatibility,  , Intermediate Output Examples, gtroff Output
@subsection Output Language Compatibility

The intermediate output language of @acronym{AT&T} @code{troff} was
first documented in @cite{A Typesetter-independent TROFF}, by Brian
Kernighan, and by 1992 the @acronym{AT&T} @code{troff} manual was
updated to incorprate a description of it.

The GNU @code{troff} intermediate output format is compatible with this
specification except for the following features.

@itemize @bullet
@item
The classical quasi-device independence is not yet implemented.

@item
The old hardware was very different from what we use today.  So the
@code{groff} devices are also fundamentally different from the ones
in @acronym{AT&T} @code{troff}.  For example, the @acronym{AT&T}
PostScript device is called @code{post} and has a resolution of only 720
units per inch, suitable for printers 20 years ago, while @code{groff}'s
@code{ps} device has a resolution of 72000 units per inch.  Maybe, by
implementing some rescaling mechanism similar to the classical
quasi-device independence, @code{groff} could emulate @acronym{AT&T}'s
@code{post} device.

@item
The B-spline command @samp{D~} is correctly handled by the intermediate
output parser, but the drawing routines aren't implemented in some of
the postprocessor programs.

@item
The argument of the commands @samp{s} and @w{@samp{x H}} has the
implicit unit scaled point @samp{z} in @code{gtroff}, while
@acronym{AT&T} @code{troff} has point (@samp{p}).  This isn't an
incompatibility but a compatible extension, for both units coincide for
all devices without a @code{sizescale} parameter in the @file{DESC}
file, including all postprocessors from @acronym{AT&T} and
@code{groff}'s text devices.  The few @code{groff} devices with a
@code{sizescale} parameter either do not exist for @acronym{AT&T}
@code{troff}, have a different name, or seem to have a different
resolution.  So conflicts are very unlikely.

@item
The position changing after the commands @samp{Dp}, @samp{DP}, and
@samp{Dt} is illogical, but as old versions of @code{gtroff} used this
feature it is kept for compatibility reasons.

@ignore
Temporarily, there existed some confusion on the positioning after the
@samp{D} commands that are @code{groff} extensions.  This has been
clarified by establishing the classical rule for all @code{groff}
drawing commands:

@itemize
@item
The position after a graphic object has been drawn is at its end; for
circles and ellipses, the `end' is at the right side.

@item
From this, the positionings specified for the drawing commands above
follow quite naturally.
@end itemize
@end ignore

@end itemize
@c END TODO: Make parallel with groff_out(5).


@c =====================================================================

@codequotebacktick on
@codequoteundirected on

@c BEGIN Keep parallel with groff_font(5).
@node Device and Font Description Files,  , gtroff Output, File Formats
@section Device and Font Description Files
@cindex font files
@cindex files, font

The @code{groff} font and output device description formats are slight
extensions of those used by @acronym{AT&T} device-independent
@code{troff}.  In distinction to the @acronym{AT&T} implementation,
@code{groff} lacks a binary format; all files are text
files.@footnote{Plan@tie{}9 @code{troff} has also abandoned the binary
format.}  The device and font description files for a device @var{name}
are stored in a @file{dev@var{name}} directory.  The device description
file is called @file{DESC}, and, for each font supported by the device,
a font description file is called@tie{}@file{@var{f}}, where
@var{f}@tie{}is usually an abbreviation of a font's name and/or style.
For example, the @code{ps} (PostScript) device has @code{groff} font
description files for Times roman (@file{TR}) and Zapf Chancery Medium
italic (@file{ZCMI}), among many others, while the @code{utf8} device
(for terminal emulators) has only font descriptions for the roman,
italic, bold, and bold-italic styles (@file{R}, @file{I}, @file{B}, and
@file{BI}, respectively).

Device and font description files are read both by the formatter, GNU
@code{troff}, and by output drivers.  The programs delegate these files'
processing to an internal library, @file{libgroff}, ensuring their
consistent interpretation.

@menu
* DESC File Format::
* Font Description File Format::
@end menu

@c ---------------------------------------------------------------------

@node DESC File Format, Font Description File Format, Device and Font Description Files, Device and Font Description Files
@subsection @file{DESC} File Format
@cindex @file{DESC} file format
@cindex font description file format
@cindex format of font description file

The @file{DESC} file contains a series of directives; each begins a
line.  Their order is not important, with two exceptions: (1) the
@code{res} directive must precede any @code{papersize} directive; and
(2) the @code{charset} directive must come last (if at all).  If a
directive name is repeated, later entries in the file override previous
ones (except that the paper dimensions are computed based on the
@code{res} directive last seen when @code{papersize} is encountered).
Spaces and/or tabs separate words and are ignored at line boundaries.
@cindex comments in device description files
@cindex device description files, comments
@kindex #
Comments start with the @samp{#} character and extend to the end of a
line.  Empty lines are ignored.

@table @code
@item family @var{fam}
@kindex family
The default font family is @var{fam}.

@item fonts @var{n} @var{F1} @r{@dots{}} @var{Fn}
@kindex fonts
Fonts @var{F1}, @dots{}, @var{Fn} are mounted at font positions
@var{m}+1, @dots{}, @var{m}+@var{n} where @var{m} is the number of
@code{styles} (see below).  This directive may extend over more than one
line.  A font name of@tie{}@code{0} causes no font to be mounted at the
corresponding position.

@item hor @var{n}
@kindex hor
@cindex horizontal motion quantum
@cindex motion quantum, horizontal
@cindex quantum, horizontal motion
@cindex horizontal resolution
@cindex resolution, horizontal
The horizontal motion quantum is @var{n}@tie{}basic units.  All
horizontal quantities are rounded to multiples of@tie{}@var{n}.

@item image_generator @var{program}
@kindex image_generator
@cindex PostScript, PNG image generation
@cindex PNG image generation from PostScript
Use @var{program} to generate PNG images from PostScript input.  Under
GNU/Linux, this is usually @code{gs}, but under other systems (notably
Cygwin) it might be set to another name.  The @code{grohtml} driver uses
this directive.

@item paperlength @var{n}
@kindex paperlength
The vertical dimension of the output medium is @var{n}@tie{}basic units
(deprecated: use @code{papersize} instead).

@item papersize @var{format-or-dimension-pair-or-file-name} @r{@dots{}}
@kindex papersize
The dimensions of the output medium are as according to the
argument, which is either a standard paper format, a pair of dimensions,
or the name of a plain text file containing either of the foregoing.

Recognized paper formats are the ISO and DIN formats
@code{A0}--@code{A7}, @code{B0}--@code{B7}, @code{C0}--@code{C7},
@code{D0}--@code{D7}; the U.S.@: paper types @code{letter},
@code{legal}, @code{tabloid}, @code{ledger}, @code{statement}, and
@code{executive}; and the envelope formats @code{com10}, @code{monarch},
and @code{DL}.  Matching is performed without regard for lettercase.

Alternatively, the argument can be a custom paper format in the format
@code{@var{length},@var{width}} (with no spaces before or after the
comma).  Both @var{length} and @var{width} must have a unit appended;
valid units are @samp{i} for inches, @samp{c} for centimeters, @samp{p}
for points, and @samp{P} for picas.  Example: @samp{12c,235p}.  An
argument that starts with a digit is always treated as a custom paper
format.

Finally, the argument can be a file name (e.g., @file{/etc/papersize});
if the file can be opened, the first line is read and a match attempted
against each of the other forms.  No comment syntax is supported.

More than one argument can be specified;
each is scanned in turn and the first valid paper specification used.

@item paperwidth @var{n}
@kindex paperwidth
The horizontal dimension of the output medium is @var{n}@tie{}basic
units (deprecated: use @code{papersize} instead).

@item pass_filenames
@kindex pass_filenames
Direct GNU @code{troff} to emit the name of the source file being
processed.  This is achieved with the intermediate output command
@samp{x F}, which @code{grohtml} interprets.

@item postpro @var{program}
@kindex postpro
Use @var{program} as the postprocessor.

@item prepro @var{program}
@kindex prepro
Use @var{program} as a preprocessor.  The @code{html} and @code{xhtml}
output devices use this directive.

@item print @var{program}
@kindex print
Use @var{program} as a spooler program for printing.  If omitted, the
@option{-l} and @option{-L} options of @code{groff} are ignored.

@item res @var{n}
@kindex res
@cindex device resolution
@cindex resolution, device
The device resolution is @var{n}@tie{}basic units per inch.

@item sizes @var{s1} @r{@dots{}} @var{sn} 0
@kindex sizes
The device has fonts at @var{s1}, @dots{}, @var{sn} scaled points (see
below).  The list of sizes must be terminated by@tie{}@code{0}.  Each
@var{si} can also be a range of sizes @var{m}--@var{n}.  The list can
extend over more than one line.

@item sizescale @var{n}
@kindex sizescale
A typographical point is subdivided into @var{n}@tie{}scaled points.
The default is@tie{}@code{1}.  @xref{Using Fractional Type Sizes}.

@item styles @var{S1} @r{@dots{}} @var{Sm}
@kindex styles
The first@tie{}@var{m} font mounting positions are associated with
styles @var{S1}, @dots{}, @var{Sm}.

@item tcommand
@kindex tcommand
The postprocessor can handle the @samp{t} and @samp{u} intermediate
output commands.

@item unicode
@kindex unicode
The output device supports the complete Unicode repertoire.  This
directive is useful only for devices that produce character entities
instead of glyphs.

If @code{unicode} is present, no @code{charset} section is required in
the font description files since the Unicode handling built into
@code{groff} is used.  However, if there are entries in a font
description file's @code{charset} section, they either override the
default mappings for those particular characters or add new mappings
(normally for composite characters).

The @code{utf8}, @code{html}, and @code{xhtml} output devices use this
directive.

@item unitwidth @var{n}
@kindex unitwidth
Quantities in the font description files are in basic units for fonts
whose type size is @var{n}@tie{}scaled points.

@item unscaled_charwidths
@kindex unscaled_charwidths
Make the font handling module always return unscaled character widths.
The @code{grohtml} driver uses this directive.

@item use_charnames_in_special
@kindex use_charnames_in_special
GNU @code{troff} should encode special characters inside device control
commands; see @ref{Postprocessor Access}.  The @code{grohtml} driver
uses this directive.

@item vert @var{n}
@kindex vert
@cindex vertical motion quantum
@cindex motion quantum, vertical
@cindex quantum, vertical motion
@cindex vertical resolution
@cindex resolution, vertical
The vertical motion quantum is @var{n}@tie{}basic units.  All vertical
quantities are rounded to multiples of@tie{}@var{n}.

@item charset
@kindex charset
This line and everything following it in the file are ignored.  It is
recognized for compatibility with other @code{troff} implementations.
In GNU @code{troff}, character set repertoire is described on a
per-font basis.
@end table

@kindex spare1
@kindex spare2
@kindex biggestfont
GNU @code{troff} recognizes but ignores the directives @code{spare1},
@code{spare2}, and @code{biggestfont}.

The @code{res}, @code{unitwidth}, @code{fonts}, and @code{sizes} lines
are mandatory.  Directives not listed above are ignored by GNU
@code{troff} but may be used by postprocessors to obtain further
information about the device.

@c ---------------------------------------------------------------------

@node Font Description File Format,  , DESC File Format, Device and Font Description Files
@subsection Font Description File Format
@cindex font file, format
@cindex font description file, format
@cindex format of font files
@cindex format of font description files

On typesetting output devices, each font is typically available at
multiple sizes.  While paper measurements in the device description file
are in absolute units, measurements applicable to fonts must be
proportional to the type size.  @code{groff} achieves this using the
precedent set by @acronym{AT&T} device-independent @code{troff}: one
font size is chosen as a norm, and all others are scaled linearly
relative to that basis.  The ``unit width'' is the number of basic units
per point when the font is rendered at this nominal size.

For instance, @code{groff}'s @code{lbp} device uses a @code{unitwidth}
of@tie{}800.  Its Times roman font @samp{TR} has a @code{spacewidth}
of@tie{}833; this is also the width of its comma, period, centered
period, and mathematical asterisk, while its @samp{M} is 2,963 basic
units.  Thus, an @samp{M} on the @code{lbp} device is 2,963 basic units
wide at a notional type size of 800@tie{}points.@footnote{800-point type
is not practical for most purposes, but using it enables the quantities
in the font description files to be expressed as integers.}

A font description file has two sections.  The first is a sequence of
directives, and is parsed similarly to the @file{DESC} file described
above.  Except for the directive names that begin the second section,
their ordering is immaterial.  Later directives of the same name
override earlier ones, spaces and tabs are handled in the same way,
@cindex comments in font description files
@cindex font description files, comments
@kindex #
and the same comment syntax is supported.  Empty lines are ignored
throughout.

@table @code
@item name @var{f}
@kindex name
The name of the font is@tie{}@var{f}.  @samp{DESC} is an invalid font
name.  Simple integers are valid, but their use is
discouraged.@footnote{@code{groff} requests and escape sequences
interpret non-negative font names as mounting positions instead.
Further, a font named @samp{0} cannot be automatically mounted by the
@code{fonts} directive of a @file{DESC} file.}

@item spacewidth @var{n}
@kindex spacewidth
The width of an unadjusted inter-word space is @var{n}@tie{}basic units.
@end table

The directives above must appear in the first section; those below are
optional.

@table @code
@item slant @var{n}
@kindex slant
The font's glyphs have a slant of @var{n}@tie{}degrees; a positive
@var{n} slants in the direction of text flow.

@item ligatures @var{lig1} @r{@dots{}} @var{lign} @r{[}0@r{]}
@kindex ligatures
Glyphs @var{lig1}, @dots{}, @var{lign} are ligatures; possible ligatures
are @samp{ff}, @samp{fi}, @samp{fl}, @samp{ffi} and @samp{ffl}.  For
compatibility with other @code{troff} implementations, the list of
ligatures may be terminated with a@tie{}@code{0}.  The list of ligatures
must not extend over more than one line.

@item special
@cindex special fonts
@kindex special
The font is @dfn{special}: when a glyph is requested that is not present
in the current font, it is sought in any mounted fonts that bear this
property.
@end table

Other directives in this section are ignored by GNU @code{troff}, but
may be used by postprocessors to obtain further information about the
font.

The second section contains one or two subsections.  These can appear in
either order; the first one encountered commences the second section.
Each starts with a directive on a line by itself.  A @code{charset}
subsection is mandatory unless the associated @file{DESC} file contains
the @code{unicode} directive.  Another subsection, @code{kernpairs},
is optional.

@kindex charset
The directive @code{charset} starts the character set
subsection.@footnote{For typesetter devices, this directive is misnamed
since it starts a list of glyphs, not characters.}  It precedes a series
of glyph descriptions, one per line.  Each such glyph description
comprises a set of fields separated by spaces or tabs and organized as
follows.

@quotation
@var{name} @var{metrics} @var{type} @var{code} [@var{entity-name}]
[@code{--} @var{comment}]
@end quotation

@cindex 8-bit input
@cindex input, 8-bit
@cindex accessing unnamed glyphs with @code{\N}
@cindex unnamed glyphs, accessing with @code{\N}
@cindex characters, unnamed, accessing with @code{\N}
@cindex glyphs, unnamed, accessing with @code{\N}
@kindex ---
@noindent
@var{name} identifies the glyph:
@c XXX: Move this footnote to a more general discussion since it is
@c applicable to the groff system overall.
@c
@c @footnote{The distinction between input, characters, and output,
@c glyphs, is not clearly separated in the terminology of @code{groff};
@c for example, the @code{char} request should be called @code{glyph}
@c since it defines an output entity.}
if @var{name} is a printable character@tie{}@var{c}, it corresponds to
the @code{troff} ordinary character@tie{}@var{c}.  If @var{name} is a
multi-character sequence not beginning with @code{\}, it corresponds to
the GNU @code{troff} special character escape sequence
@samp{\[@var{name}]}.  A name consisting of three minus signs,
@samp{---}, is special and indicates that the glyph is unnamed: such
glyphs can be accessed only by the @code{\N} escape sequence in
@code{troff}.  A special character named @samp{---} can still be defined
using @code{char} and similar requests.  The @var{name} @samp{\-}
defines the minus sign glyph.  Finally, @var{name} can be the
unbreakable one-sixth and one-twelfth space escape sequences, @code{\|}
and @code{\^} (``thin'' and ``hair'' spaces, respectively), in which
case only the width metric described below is interpreted; a font can
thus customize the widths of these spaces.
@c XXX: For exhaustivity purposes...you can define "\whatever", which
@c has to be accessed with \C'\\whatever' or \[\\whatever], but the
@c parser matches predefined escape sequences before looking up special
@c characters.  Most such definitions are inaccessible from the
@c language, because nearly every '\x', where 'x' is a Unicode basic
@c Latin character, is a predefined groff escape sequence.
@c
@c XXX: Commented out because the charXXX feature is very legacy, and as
@c noted below, discouraged in font description files.
@c
@c GNU @code{troff} supports 8-bit input characters; however some
@c utilities have difficulties with eight-bit characters.  For this
@c reason, there is a convention that the entity name @samp{char@var{n}}
@c is equivalent to the single input character whose code
@c is@tie{}@var{n}.  For example, @samp{char163} would be equivalent to
@c the character with code@tie{}163, which is the pounds sterling sign
@c in the ISO@tie{}@w{Latin-1} character set.  You shouldn't use
@c @samp{char@var{n}} entities in font description files since they are
@c related to input, not output.  Otherwise, you get hard-coded
@c connections between input and output encoding, which prevents use of
@c different (input) character sets.

The form of the @var{metrics} field is as follows.

@display
@group
@var{width}[@code{,}[@var{height}[@code{,}[@var{depth}[@code{,}[@var{italic-correction}
  [@code{,}[@var{left-italic-correction}[@code{,}[@var{subscript-correction}]]]]]]]]]]
@end group
@end display

@noindent
There must not be any spaces, tabs, or newlines between these
@dfn{subfields} (which have been split here into two lines only for
better legibility).  The subfields are in basic units expressed as
decimal integers.  Unspecified subfields default to@tie{}@code{0}.
Since there is no associated binary format, these values are not
required to fit into the C language data type @samp{char} as they are in
@acronym{AT&T} device-independent @code{troff}.

The @var{width} subfield gives the width of the glyph.  The @var{height}
subfield gives the height of the glyph (upward is positive); if a glyph
does not extend above the baseline, it should be given a zero height,
rather than a negative height.  The @var{depth} subfield gives the depth
of the glyph, that is, the distance below the baseline to which the
glyph extends (downward is positive); if a glyph does not extend below
the baseline, it should be given a zero depth, rather than a negative
depth.  Italic corrections are relevant to glyphs in italic or oblique
styles.  The @var{italic-correction} is the amount of space that should
be added after an oblique glyph to be followed immediately by an upright
glyph.  The @var{left-italic-correction} is the amount of space that
should be added before an oblique glyph to be preceded immediately by an
upright glyph.  The @var{subscript-correction} is the amount of space
that should be added after an oblique glyph to be followed by a
subscript; it should be less than the italic correction.

For fonts used with typesetting devices, the @var{type} field gives a
featural description of the glyph: it is a bit mask recording whether
the glyph is an ascender, descender, both, or neither.  When a @code{\w}
escape sequence is interpolated, these values are bitwise or-ed
together for each glyph and stored in the @code{nr} register.  In font
descriptions for terminal devices, all glyphs might have a type of zero,
regardless of their appearance.

@table @code
@item 0
means the glyph lies entirely between the baseline and a horizontal line
at the ``x-height'' of the font; typical examples are @samp{a},
@samp{c}, and @samp{x};

@item 1
means the glyph descends below the baseline, like @samp{p};

@item 2
means the glyph ascends above the font's x-height, like @samp{A} or
@samp{b}; and

@item 3
means the glyph is both an ascender and a descender---this is true of
parentheses in some fonts.
@end table

The @var{code} field gives a numeric identifier that the postprocessor
uses to render the glyph.  The glyph can be specified to @code{troff}
using this code by means of the @code{\N} escape sequence.  @var{code}
can be any integer.@footnote{that is, any integer parsable by the C
standard library's @code{strotol} function}

The @var{entity-name} field defines an identifier for the glyph that the
postprocessor uses to print the GNU @code{troff} glyph @var{name}.  This
field is optional; it was introduced so that the @code{grohtml} output
driver could encode its character set.  For example, the glyph
@samp{\[Po]} is represented by @samp{&pound;} in @acronym{HTML} 4.0.
For efficiency, these data are now compiled directly into
@code{grohtml}.  @code{grops} uses the field to build sub-encoding
arrays for PostScript fonts containing more than 256 glyphs.  Anything
on the line after the @var{entity-name} field or @samp{--} is ignored.

A line in the @code{charset} section can also have the form

@Example
@var{name} "
@endExample

@noindent
identifying @var{name} as another name for the glyph mentioned in the
preceding line.  Such aliases can be chained.

@kindex kernpairs
The directive @code{kernpairs} starts a list of kerning adjustments to
be made to adjacent glyph pairs from this font.  It contains a sequence
of lines formatted as follows.

@Example
@var{g1} @var{g2} @var{n}
@endExample

@noindent
The foregoing means that when glyph @var{g1} is typeset immediately
before @var{g2}, the space between them should be increased
by@tie{}@var{n}.  Most kerning pairs should have a negative value
for@tie{}@var{n}.
@c END Keep parallel with groff_font(5).

@codequotebacktick off
@codequoteundirected off



@c =====================================================================
@c =====================================================================

@node Copying This Manual, Request Index, Font Description File Format, Top
@appendix Copying This Manual

@include fdl.texi



@c =====================================================================
@c =====================================================================

@c This'll do us for the rest of the file...
@codequotebacktick on
@codequoteundirected on

@node Request Index, Escape Sequence Index, Copying This Manual, Top
@appendix Request Index

Request names appear without a leading control character; the defaults
are @code{.} for the regular control character and @code{'} for the
no-break control character.

@printindex rq



@c =====================================================================
@c =====================================================================

@node Escape Sequence Index, Operator Index, Request Index, Top
@appendix Escape Sequence Index

The escape character, @code{\} by default, is always followed by at
least one more input character, making an escape @emph{sequence}.  Any
input token @code{\@var{X}} with @var{X} not in the list below emits a
warning and interpolates glyph @var{X}.  Note the entries for @code{\.},
which may be obscured by the leader dots, and for @code{\@key{RET}} and
@code{\@key{SP}}, which are sorted alphabetically, not by code point
order.

@printindex es



@c =====================================================================
@c =====================================================================

@node Operator Index, Register Index, Escape Sequence Index, Top
@appendix Operator Index

@printindex op



@c =====================================================================
@c =====================================================================

@node Register Index, Macro Index, Operator Index, Top
@appendix Register Index

The macro package or program a specific register belongs to is appended
in brackets.

A register name@tie{}@code{x} consisting of exactly one character can be
accessed as @samp{\nx}.  A register name @code{xx} consisting of exactly
two characters can be accessed as @samp{\n(xx}.  Register names
@code{xxx} of any length can be accessed as @samp{\n[xxx]}.

@printindex vr



@c =====================================================================
@c =====================================================================

@node Macro Index, String Index, Register Index, Top
@appendix Macro Index

The macro package a specific macro belongs to is appended in brackets.
They appear without the leading control character (normally @samp{.}).

@printindex ma



@c =====================================================================
@c =====================================================================

@node String Index, File Keyword Index, Macro Index, Top
@appendix String Index

The macro package or program a that defines or uses each string is
appended in brackets. (Only one string, @code{.T}, is defined by the
@code{troff} formatter itself.)  @xref{Strings}.


@printindex st



@c =====================================================================
@c =====================================================================

@node File Keyword Index, Program and File Index, String Index, Top
@appendix File Keyword Index

@printindex ky



@c =====================================================================
@c =====================================================================

@node Program and File Index, Concept Index, File Keyword Index, Top
@appendix Program and File Index

@printindex pg



@c =====================================================================
@c =====================================================================

@node Concept Index,  , Program and File Index, Top
@appendix Concept Index

@printindex cp


@bye

@c Local Variables:
@c mode: texinfo
@c coding: latin-1
@c fill-column: 72
@c End:
@c vim: set textwidth=72: