Minor doc fix; update tests by adding comments to each file, and generally

tidying.


git-svn-id: svn://vcs.exim.org/pcre/code/trunk@450 2f5784b3-3f2a-0410-8824-cb99058d5e15

24 files changed, 171 insertions, 113 deletions

diff --git a/testdata/testinput1 b/testdata/testinput1
index 8b0caa4..6cacbb5 100644
--- a/testdata/testinput1
+++ b/testdata/testinput1
@@ -1,3 +1,6 @@
+/-- This set of tests is for features that are compatible with all versions of
+    Perl 5, in non-UTF-8 mode. --/
+
 /the quick brown fox/
     the quick brown fox
     The quick brown FOX
@@ -4064,4 +4067,4 @@
 /^%((?(?=[a])[^%])|b)*%$/
     %ab%
 
-/ End of testinput1 /
+/-- End of testinput1 --/
diff --git a/testdata/testinput10 b/testdata/testinput10
index 726a389..420c274 100644
--- a/testdata/testinput10
+++ b/testdata/testinput10
@@ -121,4 +121,4 @@ are all themselves checked in other tests. --/
 
 /[^\xaa]/8BM
 
-/ End of testinput10 /
+/-- End of testinput10 --/
diff --git a/testdata/testinput11 b/testdata/testinput11
index 0518b68..1a08cd1 100644
--- a/testdata/testinput11
+++ b/testdata/testinput11
@@ -1,3 +1,5 @@
+/-- These tests are for the Perl 5.10 features that PCRE supports. --/
+
 /\H\h\V\v/
     X X\x0a
     X\x09X\x0b
@@ -250,4 +252,4 @@
     ** Failers
     AD 
         
-/ End of testinput11 /
+/-- End of testinput11 --/
diff --git a/testdata/testinput12 b/testdata/testinput12
index ae2ef10..91f31e3 100644
--- a/testdata/testinput12
+++ b/testdata/testinput12
@@ -1,3 +1,6 @@
+/-- These tests for Unicode property support test PCRE's API and show some of
+    the compiled code. They are not Perl-compatible. --/
+
 /[\p{L}]/DZ
 
 /[\p{^L}]/DZ
@@ -155,9 +158,9 @@
     \x{c0}
     \x{e0} 
 
-/The next two are special cases where the lengths of the different cases of the 
-same character differ. The first went wrong with heap frame storage; the 2nd
-was broken in all cases./
+/-- The next two are special cases where the lengths of the different cases of
+the same character differ. The first went wrong with heap frame storage; the
+second was broken in all cases. --/
 
 /^\x{023a}+?(\x{0130}+)/8i
   \x{023a}\x{2c65}\x{0130}
@@ -185,9 +188,9 @@ was broken in all cases./
     \x{c0}
     \x{e0} 
 
-/The next two should be Perl-compatible, but it fails to match \x{e0}. PCRE
+/-- The next two should be Perl-compatible, but it fails to match \x{e0}. PCRE
 will match it only with UCP support, because without that it has no notion
-of case for anything other than the ASCII letters. / 
+of case for anything other than the ASCII letters. --/ 
 
 /((?i)[\x{c0}])/8
     \x{c0}
@@ -197,3 +200,4 @@ of case for anything other than the ASCII letters. /
     \x{c0}
     \x{e0} 
     
+/-- End of testinput12 --/
diff --git a/testdata/testinput2 b/testdata/testinput2
index 1e29593..ec5d621 100644
--- a/testdata/testinput2
+++ b/testdata/testinput2
@@ -1,3 +1,14 @@
+/-- This set of tests is not Perl-compatible. It checks on special features
+    of PCRE's API, error diagnostics, and the compiled code of some patterns.
+    It also checks the non-Perl syntax the PCRE supports (Python, .NET, 
+    Oniguruma). Finally, there are some tests where PCRE and Perl differ, 
+    either because PCRE can't be compatible, or there is potential Perl 
+    bug. --/  
+  
+/-- Originally, the Perl 5.10 things were in here too, but now I have separated
+    many (most?) of them out into test 11. However, there may still be some
+    that were overlooked. --/   
+
 /(a)b|/I
 
 /abc/I
@@ -2842,4 +2853,4 @@ a random value. /Ix
 /^X(?7)(a)(?|(b)|(q)(r)(s))(c)(d)(Y)/
     XYabcdY
 
-/ End of testinput2 /
+/-- End of testinput2 --/
diff --git a/testdata/testinput3 b/testdata/testinput3
index 1376c21..6fea2f5 100644
--- a/testdata/testinput3
+++ b/testdata/testinput3
@@ -1,3 +1,7 @@
+/-- This set of tests checks local-specific features, using the fr_FR locale. 
+    It is not Perl-compatible. There is different version called wintestinput3
+  f  or use on Windows, where the locale is called "french". --/
+
 /^[\w]+/
     *** Failers
     École
@@ -88,4 +92,4 @@
     
 /[[:alpha:]][[:lower:]][[:upper:]]/DZLfr_FR 
 
-/ End of testinput3 /
+/-- End of testinput3 --/
diff --git a/testdata/testinput4 b/testdata/testinput4
index e5f6194..12f4c7e 100644
--- a/testdata/testinput4
+++ b/testdata/testinput4
@@ -1,7 +1,6 @@
-/-- Do not use the \x{} construct except with patterns that have the --/
-/-- /8 option set, because PCRE doesn't recognize them as UTF-8 unless --/
-/-- that option is set. However, the latest Perls recognize them always. --/
-
+/-- This set of tests if for UTF-8 support, excluding Unicode properties. It is
+    compatible with all versions of Perl 5. --/
+   
 /a.b/8
     acb
     a\x7fb
@@ -623,4 +622,22 @@
 
 /(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8
 
-/ End of testinput4 /
+/^[a\x{c0}]b/8
+    \x{c0}b
+    
+/^([a\x{c0}]*?)aa/8
+    a\x{c0}aaaa/ 
+
+/^([a\x{c0}]*?)aa/8
+    a\x{c0}aaaa/ 
+    a\x{c0}a\x{c0}aaa/ 
+
+/^([a\x{c0}]*)aa/8
+    a\x{c0}aaaa/ 
+    a\x{c0}a\x{c0}aaa/ 
+
+/^([a\x{c0}]*)a\x{c0}/8
+    a\x{c0}aaaa/ 
+    a\x{c0}a\x{c0}aaa/ 
+
+/-- End of testinput4 --/
diff --git a/testdata/testinput5 b/testdata/testinput5
index 5d3030f..c1a21ba 100644
--- a/testdata/testinput5
+++ b/testdata/testinput5
@@ -1,3 +1,6 @@
+/-- This set of tests checks the API, internals, and non-Perl stuff for UTF-8
+    support, excluding Unicode properties. --/
+
 /\x{100}/8DZ
 
 /\x{1000}/8DZ
@@ -53,30 +56,6 @@
 /.{3,5}?/DZ8
     \x{212ab}\x{212ab}\x{212ab}\x{861}
 
-/-- These tests are here rather than in testinput4 because Perl 5.6 has some
-problems with UTF-8 support, in the area of \x{..} where the value is < 255. 
-It grumbles about invalid UTF-8 strings. --/
-
-/^[a\x{c0}]b/8
-    \x{c0}b
-    
-/^([a\x{c0}]*?)aa/8
-    a\x{c0}aaaa/ 
-
-/^([a\x{c0}]*?)aa/8
-    a\x{c0}aaaa/ 
-    a\x{c0}a\x{c0}aaa/ 
-
-/^([a\x{c0}]*)aa/8
-    a\x{c0}aaaa/ 
-    a\x{c0}a\x{c0}aaa/ 
-
-/^([a\x{c0}]*)a\x{c0}/8
-    a\x{c0}aaaa/ 
-    a\x{c0}a\x{c0}aaa/ 
-    
-/-- --/ 
-    
 /(?<=\C)X/8
     Should produce an error diagnostic
     
@@ -741,4 +720,4 @@ can't tell the difference.) --/
     the cat\P
     the cat\P\P
 
-/ End of testinput5 /
+/-- End of testinput5 --/
diff --git a/testdata/testinput6 b/testdata/testinput6
index f9cc5f7..f4249da 100644
--- a/testdata/testinput6
+++ b/testdata/testinput6
@@ -1,3 +1,7 @@
+/-- This set of tests is for Unicode property support. It is compatible with
+    Perl 5.10, but not 5.8 because it tests some extra properties that are
+    not in the earlier release. --/ 
+
 /^\pC\pL\pM\pN\pP\pS\pZ</8
     \x7f\x{c0}\x{30f}\x{660}\x{66c}\x{f01}\x{1680}<
     \np\x{300}9!\$ < 
@@ -752,4 +756,4 @@
 /[\p{Lu}\x20]+/
     \x41\x20\x50\xC2\x54\xC9\x20\x54\x4F\x44\x41\x59
 
-/ End of testinput6 /
+/-- End of testinput6 --/
diff --git a/testdata/testinput7 b/testdata/testinput7
index c32d336..7aafc95 100644
--- a/testdata/testinput7
+++ b/testdata/testinput7
@@ -1,3 +1,6 @@
+/-- This set of tests check the DFA matching functionality of pcre_dfa_exec().
+    The -dfa flag must be used with pcretest when running it. --/
+     
 /abc/
     abc
     
@@ -4486,4 +4489,4 @@
 /(?=C)/g+
     ABCDECBA
 
-/ End of testinput7 /
+/-- End of testinput7 --/
diff --git a/testdata/testinput8 b/testdata/testinput8
index 11884ad..e879e81 100644
--- a/testdata/testinput8
+++ b/testdata/testinput8
@@ -1,6 +1,6 @@
-/-- Do not use the \x{} construct except with patterns that have the --/
-/-- /8 option set, because PCRE doesn't recognize them as UTF-8 unless --/
-/-- that option is set. However, the latest Perls recognize them always. --/
+/-- This set of tests checks UTF-8 support with the DFA matching functionality
+    of pcre_dfa_exec(). The -dfa flag must be used with pcretest when running 
+    it. --/
 
 /\x{100}ab/8
   \x{100}ab
@@ -667,4 +667,4 @@
 /X/8f<any> 
     A\x{1ec5}ABCXYZ
 
-/ End of testinput 8 / 
+/-- End of testinput8 --/ 
diff --git a/testdata/testinput9 b/testdata/testinput9
index ab04455..590420c 100644
--- a/testdata/testinput9
+++ b/testdata/testinput9
@@ -1,3 +1,7 @@
+/-- This set of tests check Unicode property support with the DFA matching 
+    functionality of pcre_dfa_exec(). The -dfa flag must be used with pcretest
+    when running it. --/
+
 /\pL\P{Nd}/8
     AB
     *** Failers
@@ -843,4 +847,4 @@
     ** Failers 
     \x{1d79}\x{a77d} 
 
-/ End / 
+/-- End of testinput9 --/ 
diff --git a/testdata/testoutput1 b/testdata/testoutput1
index 81b0cb8..a2a6dd4 100644
--- a/testdata/testoutput1
+++ b/testdata/testoutput1
@@ -1,3 +1,6 @@
+/-- This set of tests is for features that are compatible with all versions of
+    Perl 5, in non-UTF-8 mode. --/
+
 /the quick brown fox/
     the quick brown fox
  0: the quick brown fox
@@ -6646,4 +6649,4 @@ No match
  0: %ab%
  1: 
 
-/ End of testinput1 /
+/-- End of testinput1 --/
diff --git a/testdata/testoutput10 b/testdata/testoutput10
index 4eaaa39..a0db640 100644
--- a/testdata/testoutput10
+++ b/testdata/testoutput10
@@ -666,4 +666,4 @@ Memory allocation (code space): 40
  39     End
 ------------------------------------------------------------------
 
-/ End of testinput10 /
+/-- End of testinput10 --/
diff --git a/testdata/testoutput11 b/testdata/testoutput11
index b2f0d7d..f7b54ac 100644
--- a/testdata/testoutput11
+++ b/testdata/testoutput11
@@ -1,3 +1,5 @@
+/-- These tests are for the Perl 5.10 features that PCRE supports. --/
+
 /\H\h\V\v/
     X X\x0a
  0: X X\x0a
@@ -518,4 +520,4 @@ No match
     AD 
 No match
         
-/ End of testinput11 /
+/-- End of testinput11 --/
diff --git a/testdata/testoutput12 b/testdata/testoutput12
index 454e2e1..70ca7a3 100644
--- a/testdata/testoutput12
+++ b/testdata/testoutput12
@@ -1,3 +1,6 @@
+/-- These tests for Unicode property support test PCRE's API and show some of
+    the compiled code. They are not Perl-compatible. --/
+
 /[\p{L}]/DZ
 ------------------------------------------------------------------
         Bra
@@ -399,9 +402,9 @@ No match
     \x{e0} 
  0: \x{e0}
 
-/The next two are special cases where the lengths of the different cases of the 
-same character differ. The first went wrong with heap frame storage; the 2nd
-was broken in all cases./
+/-- The next two are special cases where the lengths of the different cases of
+the same character differ. The first went wrong with heap frame storage; the
+second was broken in all cases. --/
 
 /^\x{023a}+?(\x{0130}+)/8i
   \x{023a}\x{2c65}\x{0130}
@@ -447,9 +450,9 @@ was broken in all cases./
     \x{e0} 
  0: \x{e0}
 
-/The next two should be Perl-compatible, but it fails to match \x{e0}. PCRE
+/-- The next two should be Perl-compatible, but it fails to match \x{e0}. PCRE
 will match it only with UCP support, because without that it has no notion
-of case for anything other than the ASCII letters. / 
+of case for anything other than the ASCII letters. --/ 
 
 /((?i)[\x{c0}])/8
     \x{c0}
@@ -465,3 +468,4 @@ of case for anything other than the ASCII letters. /
     \x{e0} 
  0: \x{e0}
     
+/-- End of testinput12 --/
diff --git a/testdata/testoutput2 b/testdata/testoutput2
index 393556e..67301ea 100644
--- a/testdata/testoutput2
+++ b/testdata/testoutput2
@@ -1,3 +1,14 @@
+/-- This set of tests is not Perl-compatible. It checks on special features
+    of PCRE's API, error diagnostics, and the compiled code of some patterns.
+    It also checks the non-Perl syntax the PCRE supports (Python, .NET, 
+    Oniguruma). Finally, there are some tests where PCRE and Perl differ, 
+    either because PCRE can't be compatible, or there is potential Perl 
+    bug. --/  
+  
+/-- Originally, the Perl 5.10 things were in here too, but now I have separated
+    many (most?) of them out into test 11. However, there may still be some
+    that were overlooked. --/   
+
 /(a)b|/I
 Capturing subpattern count = 1
 No options
@@ -9834,4 +9845,4 @@ No match
  6: d
  7: Y
 
-/ End of testinput2 /
+/-- End of testinput2 --/
diff --git a/testdata/testoutput3 b/testdata/testoutput3
index 28b1c3a..ce18ed2 100644
--- a/testdata/testoutput3
+++ b/testdata/testoutput3
@@ -1,3 +1,7 @@
+/-- This set of tests checks local-specific features, using the fr_FR locale. 
+    It is not Perl-compatible. There is different version called wintestinput3
+  f  or use on Windows, where the locale is called "french". --/
+
 /^[\w]+/
     *** Failers
 No match
@@ -160,4 +164,4 @@ No options
 No first char
 No need char
 
-/ End of testinput3 /
+/-- End of testinput3 --/
diff --git a/testdata/testoutput4 b/testdata/testoutput4
index e67a958..128afe4 100644
--- a/testdata/testoutput4
+++ b/testdata/testoutput4
@@ -1,9 +1,6 @@
-/-- Do not use the \x{} construct except with patterns that have the --/
-/-- /8 option set, because PCRE doesn't recognize them as UTF-8 unless --/
-No match
-/-- that option is set. However, the latest Perls recognize them always. --/
-No match
-
+/-- This set of tests if for UTF-8 support, excluding Unicode properties. It is
+    compatible with all versions of Perl 5. --/
+   
 /a.b/8
     acb
  0: acb
@@ -1089,4 +1086,37 @@ No match
 
 /(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8
 
-/ End of testinput4 /
+/^[a\x{c0}]b/8
+    \x{c0}b
+ 0: \x{c0}b
+    
+/^([a\x{c0}]*?)aa/8
+    a\x{c0}aaaa/ 
+ 0: a\x{c0}aa
+ 1: a\x{c0}
+
+/^([a\x{c0}]*?)aa/8
+    a\x{c0}aaaa/ 
+ 0: a\x{c0}aa
+ 1: a\x{c0}
+    a\x{c0}a\x{c0}aaa/ 
+ 0: a\x{c0}a\x{c0}aa
+ 1: a\x{c0}a\x{c0}
+
+/^([a\x{c0}]*)aa/8
+    a\x{c0}aaaa/ 
+ 0: a\x{c0}aaaa
+ 1: a\x{c0}aa
+    a\x{c0}a\x{c0}aaa/ 
+ 0: a\x{c0}a\x{c0}aaa
+ 1: a\x{c0}a\x{c0}a
+
+/^([a\x{c0}]*)a\x{c0}/8
+    a\x{c0}aaaa/ 
+ 0: a\x{c0}
+ 1: 
+    a\x{c0}a\x{c0}aaa/ 
+ 0: a\x{c0}a\x{c0}
+ 1: a\x{c0}
+
+/-- End of testinput4 --/
diff --git a/testdata/testoutput5 b/testdata/testoutput5
index 71f18bf..ded5edb 100644
--- a/testdata/testoutput5
+++ b/testdata/testoutput5
@@ -1,3 +1,6 @@
+/-- This set of tests checks the API, internals, and non-Perl stuff for UTF-8
+    support, excluding Unicode properties. --/
+
 /\x{100}/8DZ
 ------------------------------------------------------------------
         Bra
@@ -274,45 +277,6 @@ No need char
     \x{212ab}\x{212ab}\x{212ab}\x{861}
  0: \x{212ab}\x{212ab}\x{212ab}
 
-/-- These tests are here rather than in testinput4 because Perl 5.6 has some
-problems with UTF-8 support, in the area of \x{..} where the value is < 255. 
-It grumbles about invalid UTF-8 strings. --/
-
-/^[a\x{c0}]b/8
-    \x{c0}b
- 0: \x{c0}b
-    
-/^([a\x{c0}]*?)aa/8
-    a\x{c0}aaaa/ 
- 0: a\x{c0}aa
- 1: a\x{c0}
-
-/^([a\x{c0}]*?)aa/8
-    a\x{c0}aaaa/ 
- 0: a\x{c0}aa
- 1: a\x{c0}
-    a\x{c0}a\x{c0}aaa/ 
- 0: a\x{c0}a\x{c0}aa
- 1: a\x{c0}a\x{c0}
-
-/^([a\x{c0}]*)aa/8
-    a\x{c0}aaaa/ 
- 0: a\x{c0}aaaa
- 1: a\x{c0}aa
-    a\x{c0}a\x{c0}aaa/ 
- 0: a\x{c0}a\x{c0}aaa
- 1: a\x{c0}a\x{c0}a
-
-/^([a\x{c0}]*)a\x{c0}/8
-    a\x{c0}aaaa/ 
- 0: a\x{c0}
- 1: 
-    a\x{c0}a\x{c0}aaa/ 
- 0: a\x{c0}a\x{c0}
- 1: a\x{c0}
-    
-/-- --/ 
-    
 /(?<=\C)X/8
 Failed: \C not allowed in lookbehind assertion at offset 6
 
@@ -2067,4 +2031,4 @@ Partial match: \x{123}X\x{123}\x{123}\x{123}\x{123}
     the cat\P\P
 Partial match: the cat
 
-/ End of testinput5 /
+/-- End of testinput5 --/
diff --git a/testdata/testoutput6 b/testdata/testoutput6
index 722a73c..4245d6a 100644
--- a/testdata/testoutput6
+++ b/testdata/testoutput6
@@ -1,3 +1,7 @@
+/-- This set of tests is for Unicode property support. It is compatible with
+    Perl 5.10, but not 5.8 because it tests some extra properties that are
+    not in the earlier release. --/ 
+
 /^\pC\pL\pM\pN\pP\pS\pZ</8
     \x7f\x{c0}\x{30f}\x{660}\x{66c}\x{f01}\x{1680}<
  0: \x{7f}\x{c0}\x{30f}\x{660}\x{66c}\x{f01}\x{1680}<
@@ -1289,4 +1293,4 @@ No match
     \x41\x20\x50\xC2\x54\xC9\x20\x54\x4F\x44\x41\x59
  0: A P\xc2T\xc9 TODAY
 
-/ End of testinput6 /
+/-- End of testinput6 --/
diff --git a/testdata/testoutput7 b/testdata/testoutput7
index cc8b91d..b9f9bb0 100644
--- a/testdata/testoutput7
+++ b/testdata/testoutput7
@@ -1,3 +1,6 @@
+/-- This set of tests check the DFA matching functionality of pcre_dfa_exec().
+    The -dfa flag must be used with pcretest when running it. --/
+     
 /abc/
     abc
  0: abc
@@ -7469,4 +7472,4 @@ Partial match: +ab
  0: 
  0+ CBA
 
-/ End of testinput7 /
+/-- End of testinput7 --/
diff --git a/testdata/testoutput8 b/testdata/testoutput8
index af10c4a..d991095 100644
--- a/testdata/testoutput8
+++ b/testdata/testoutput8
@@ -1,8 +1,6 @@
-/-- Do not use the \x{} construct except with patterns that have the --/
-/-- /8 option set, because PCRE doesn't recognize them as UTF-8 unless --/
-No match
-/-- that option is set. However, the latest Perls recognize them always. --/
-No match
+/-- This set of tests checks UTF-8 support with the DFA matching functionality
+    of pcre_dfa_exec(). The -dfa flag must be used with pcretest when running 
+    it. --/
 
 /\x{100}ab/8
   \x{100}ab
@@ -1288,4 +1286,4 @@ No match
     A\x{1ec5}ABCXYZ
  0: X
 
-/ End of testinput 8 / 
+/-- End of testinput8 --/ 
diff --git a/testdata/testoutput9 b/testdata/testoutput9
index 04e3873..851e3f0 100644
--- a/testdata/testoutput9
+++ b/testdata/testoutput9
@@ -1,3 +1,7 @@
+/-- This set of tests check Unicode property support with the DFA matching 
+    functionality of pcre_dfa_exec(). The -dfa flag must be used with pcretest
+    when running it. --/
+
 /\pL\P{Nd}/8
     AB
  0: AB
@@ -1670,4 +1674,4 @@ No match
     \x{1d79}\x{a77d} 
 No match
 
-/ End / 
+/-- End of testinput9 --/