/[pcre]/code/trunk/ChangeLog
ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1496 - (show annotations) (download)
Fri Jul 18 08:24:35 2014 UTC (7 days, 20 hours ago) by ph10
File size: 268175 byte(s)
Fix typos in ChangeLog.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.36 xx-xxx-2014
5 ------------------------
6
7 1. Got rid of some compiler warnings in the C++ modules that were shown up by
8 -Wmissing-field-initializers and -Wunused-parameter.
9
10 2. The tests for quantifiers being too big (greater than 65535) were being
11 applied after reading the number, and stupidly assuming that integer
12 overflow would give a negative number. The tests are now applied as the
13 numbers are read.
14
15 3. Tidy code in pcre_exec.c where two branches that used to be different are
16 now the same.
17
18 4. The JIT compiler did not generate match limit checks for certain
19 bracketed expressions with quantifiers. This may lead to exponential
20 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
21 issue should be resolved now.
22
23 5. Fixed an issue, which occures when nested alternatives are optimized
24 with table jumps.
25
26 6. Inserted two casts and changed some ints to size_t in the light of some
27 reported 64-bit compiler warnings (Bugzilla 1477).
28
29 7. Fixed a bug concerned with zero-minimum possessive groups that could match
30 an empty string, which sometimes were behaving incorrectly in the
31 interpreter (though correctly in the JIT matcher). This pcretest input is
32 an example:
33
34 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
35 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
36
37 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
38 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
39 for an empty string was breaking the inner loop and carrying on at a lower
40 level, when possessive repeated groups should always return to a higher
41 level as they have no backtrack points in them. The empty string test now
42 occurs at the outer level.
43
44 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
45 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
46
47 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
48 doesn't).
49
50 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
51 to hold a pattern to be incorrectly computed (too small) when there were
52 named back references to duplicated names. This could cause "internal
53 error: code overflow" or "double free or corruption" or other memory
54 handling errors.
55
56 11. When named subpatterns had the same prefixes, back references could be
57 confused. For example, in this pattern:
58
59 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
60
61 the reference to 'Name' was incorrectly treated as a reference to a
62 duplicate name.
63
64 12. A pattern such as /^s?c/mi8 where the optional character has more than
65 one "other case" was incorrectly compiled such that it would only try to
66 match starting at "c".
67
68 13. When a pattern starting with \s was studied, VT was not included in the
69 list of possible starting characters; this should have been part of the
70 8.34/18 patch.
71
72 14. If a character class started [\Qx]... where x is any character, the class
73 was incorrectly terminated at the ].
74
75 15. If a pattern that started with a caseless match for a character with more
76 than one "other case" was studied, PCRE did not set up the starting code
77 unit bit map for the list of possible characters. Now it does. This is an
78 optimization improvement, not a bug fix.
79
80 16. The Unicode data tables have been updated to Unicode 7.0.0.
81
82 17. Fixed a number of memory leaks in pcregrep.
83
84 18. Avoid a compiler warning (from some compilers) for a function call with
85 a cast that removes "const" from an lvalue by using an intermediate
86 variable (to which the compiler does not object).
87
88 19. Incorrect code was compiled if a group that contained an internal recursive
89 back reference was optional (had quantifier with a minimum of zero). This
90 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
91 caused segmentation faults because of stack overflows at compile time.
92
93
94 Version 8.35 04-April-2014
95 --------------------------
96
97 1. A new flag is set, when property checks are present in an XCLASS.
98 When this flag is not set, PCRE can perform certain optimizations
99 such as studying these XCLASS-es.
100
101 2. The auto-possessification of character sets were improved: a normal
102 and an extended character set can be compared now. Furthermore
103 the JIT compiler optimizes more character set checks.
104
105 3. Got rid of some compiler warnings for potentially uninitialized variables
106 that show up only when compiled with -O2.
107
108 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
109 of a match later then the end of the match. The pcretest program was not
110 handling the case sensibly - it was outputting from the start to the next
111 binary zero. It now reports this situation in a message, and outputs the
112 text from the end to the start.
113
114 5. Fast forward search is improved in JIT. Instead of the first three
115 characters, any three characters with fixed position can be searched.
116 Search order: first, last, middle.
117
118 6. Improve character range checks in JIT. Characters are read by an inprecise
119 function now, which returns with an unknown value if the character code is
120 above a certain threshold (e.g: 256). The only limitation is that the value
121 must be bigger than the threshold as well. This function is useful when
122 the characters above the threshold are handled in the same way.
123
124 7. The macros whose names start with RAWUCHAR are placeholders for a future
125 mode in which only the bottom 21 bits of 32-bit data items are used. To
126 make this more memorable for those maintaining the code, the names have
127 been changed to start with UCHAR21, and an extensive comment has been added
128 to their definition.
129
130 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
131 to the export list in Makefile.am (they were accidentally omitted from the
132 8.34 tarball).
133
134 9. The informational output from pcretest used the phrase "starting byte set"
135 which is inappropriate for the 16-bit and 32-bit libraries. As the output
136 for "first char" and "need char" really means "non-UTF-char", I've changed
137 "byte" to "char", and slightly reworded the output. The documentation about
138 these values has also been (I hope) clarified.
139
140 10. Another JIT related optimization: use table jumps for selecting the correct
141 backtracking path, when more than four alternatives are present inside a
142 bracket.
143
144 11. Empty match is not possible, when the minimum length is greater than zero,
145 and there is no \K in the pattern. JIT should avoid empty match checks in
146 such cases.
147
148 12. In a caseless character class with UCP support, when a character with more
149 than one alternative case was not the first character of a range, not all
150 the alternative cases were added to the class. For example, s and \x{17f}
151 are both alternative cases for S: the class [RST] was handled correctly,
152 but [R-T] was not.
153
154 13. The configure.ac file always checked for pthread support when JIT was
155 enabled. This is not used in Windows, so I have put this test inside a
156 check for the presence of windows.h (which was already tested for).
157
158 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
159 The algorithm provides a way to skip certain starting offsets, and usually
160 faster than linear prefix searches.
161
162 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
163 as for 'fr_FR' and 'french'. For some reason, however, it then used the
164 Windows-specific input and output files, which have 'french' screwed in.
165 So this could never have worked. One of the problems with locales is that
166 they aren't always the same. I have now updated RunTest so that it checks
167 the output of the locale test (test 3) against three different output
168 files, and it allows the test to pass if any one of them matches. With luck
169 this should make the test pass on some versions of Solaris where it was
170 failing. Because of the uncertainty, the script did not used to stop if
171 test 3 failed; it now does. If further versions of a French locale ever
172 come to light, they can now easily be added.
173
174 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
175 there was a message during ./configure, but it did not stop. This now
176 provokes an error. The invalid example in README has been corrected.
177 If a value less than the minimum is given, the minimum value has always
178 been used, but now a warning is given.
179
180 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
181 was a bug in the test system, which is now fixed. Also, the list of various
182 configurations that are tested for each release did not have one with both
183 16/32 bits and --enable-bar-anycrlf. It now does.
184
185 18. pcretest was missing "-C bsr" for displaying the \R default setting.
186
187 19. Little endian PowerPC systems are supported now by the JIT compiler.
188
189 20. The fast forward newline mechanism could enter to an infinite loop on
190 certain invalid UTF-8 input. Although we don't support these cases
191 this issue can be fixed by a performance optimization.
192
193 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
194 not take account if existing stack usage. There is now a new global
195 variable called pcre_stack_guard that can be set to point to an external
196 function to check stack availability. It is called at the start of
197 processing every parenthesized group.
198
199 22. A typo in the code meant that in ungreedy mode the max/min qualifier
200 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
201 not match "ab".
202
203 23. When UTF was disabled, the JIT program reported some incorrect compile
204 errors. These messages are silenced now.
205
206 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
207 compiler.
208
209 25. Change all the temporary files used in RunGrepTest to be different to those
210 used by RunTest so that the tests can be run simultaneously, for example by
211 "make -j check".
212
213
214 Version 8.34 15-December-2013
215 -----------------------------
216
217 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
218 executable memory. Patch inspired by Carsten Klein.
219
220 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
221 this macro is never tested and has no effect, because the work to support
222 coverage involves only compiling and linking options and special targets in
223 the Makefile. The comment in config.h implied that defining the macro would
224 enable coverage support, which is totally false. There was also support for
225 setting this macro in the CMake files (my fault, I just copied it from
226 configure). SUPPORT_GCOV has now been removed.
227
228 3. Make a small performance improvement in strlen16() and strlen32() in
229 pcretest.
230
231 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
232 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
233 these cases). There was also one in pcretest.c.
234
235 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
236
237 6. In UTF mode, the code for checking whether a group could match an empty
238 string (which is used for indefinitely repeated groups to allow for
239 breaking an infinite loop) was broken when the group contained a repeated
240 negated single-character class with a character that occupied more than one
241 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
242 UTF-8 mode). The effect was undefined: the group might or might not be
243 deemed as matching an empty string, or the program might have crashed.
244
245 7. The code for checking whether a group could match an empty string was not
246 recognizing that \h, \H, \v, \V, and \R must match a character.
247
248 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
249 an empty string. If it can, pcretest shows this in its information output.
250
251 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
252 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
253 matched by pcre_exec() without using JIT:
254
255 (a) If the rest of the pattern did not match after a maximal run of
256 grapheme clusters, the code for backing up to try with fewer of them
257 did not always back up over a full grapheme when characters that do not
258 have the modifier quality were involved, e.g. Hangul syllables.
259
260 (b) If the match point in a subject started with modifier character, and
261 there was no match, the code could incorrectly back up beyond the match
262 point, and potentially beyond the first character in the subject,
263 leading to a segfault or an incorrect match result.
264
265 10. A conditional group with an assertion condition could lead to PCRE
266 recording an incorrect first data item for a match if no other first data
267 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
268 first data item, and therefore matched "ca" after "c" instead of at the
269 start.
270
271 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
272 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
273
274 12. The source of pcregrep now includes z/OS-specific code so that it can be
275 compiled for z/OS as part of the special z/OS distribution.
276
277 13. Added the -T and -TM options to pcretest.
278
279 14. The code in pcre_compile.c for creating the table of named capturing groups
280 has been refactored. Instead of creating the table dynamically during the
281 actual compiling pass, the information is remembered during the pre-compile
282 pass (on the stack unless there are more than 20 named groups, in which
283 case malloc() is used) and the whole table is created before the actual
284 compile happens. This has simplified the code (it is now nearly 150 lines
285 shorter) and prepared the way for better handling of references to groups
286 with duplicate names.
287
288 15. A back reference to a named subpattern when there is more than one of the
289 same name now checks them in the order in which they appear in the pattern.
290 The first one that is set is used for the reference. Previously only the
291 first one was inspected. This change makes PCRE more compatible with Perl.
292
293 16. Unicode character properties were updated from Unicode 6.3.0.
294
295 17. The compile-time code for auto-possessification has been refactored, based
296 on a patch by Zoltan Herczeg. It now happens after instead of during
297 compilation. The code is cleaner, and more cases are handled. The option
298 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
299 options in pcretest are provided to set it. It can also be set by
300 (*NO_AUTO_POSSESS) at the start of a pattern.
301
302 18. The character VT has been added to the default ("C" locale) set of
303 characters that match \s and are generally treated as white space,
304 following this same change in Perl 5.18. There is now no difference between
305 "Perl space" and "POSIX space". Whether VT is treated as white space in
306 other locales depends on the locale.
307
308 19. The code for checking named groups as conditions, either for being set or
309 for being recursed, has been refactored (this is related to 14 and 15
310 above). Processing unduplicated named groups should now be as fast at
311 numerical groups, and processing duplicated groups should be faster than
312 before.
313
314 20. Two patches to the CMake build system, by Alexander Barkov:
315
316 (1) Replace the "source" command by "." in CMakeLists.txt because
317 "source" is a bash-ism.
318
319 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
320 without these the CMake build does not work on Solaris.
321
322 21. Perl has changed its handling of \8 and \9. If there is no previously
323 encountered capturing group of those numbers, they are treated as the
324 literal characters 8 and 9 instead of a binary zero followed by the
325 literals. PCRE now does the same.
326
327 22. Following Perl, added \o{} to specify codepoints in octal, making it
328 possible to specify values greater than 0777 and also making them
329 unambiguous.
330
331 23. Perl now gives an error for missing closing braces after \x{... instead of
332 treating the string as literal. PCRE now does the same.
333
334 24. RunTest used to grumble if an inappropriate test was selected explicitly,
335 but just skip it when running all tests. This make it awkward to run ranges
336 of tests when one of them was inappropriate. Now it just skips any
337 inappropriate tests, as it always did when running all tests.
338
339 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
340 character types such as \d or \w, too many callouts were inserted, and the
341 data that they returned was rubbish.
342
343 26. In UCP mode, \s was not matching two of the characters that Perl matches,
344 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
345 were matched by \h. The code has now been refactored so that the lists of
346 the horizontal and vertical whitespace characters used for \h and \v (which
347 are defined only in one place) are now also used for \s.
348
349 27. Add JIT support for the 64 bit TileGX architecture.
350 Patch by Jiong Wang (Tilera Corporation).
351
352 28. Possessive quantifiers for classes (both explicit and automatically
353 generated) now use special opcodes instead of wrapping in ONCE brackets.
354
355 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
356 (because it's meaningless), this was not happening when PCRE_CASELESS was
357 set. Not wrong, but inefficient.
358
359 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
360 properties for \w, \d, etc) is present in a test regex. Otherwise if the
361 test contains no characters greater than 255, Perl doesn't realise it
362 should be using Unicode semantics.
363
364 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
365 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
366 does in Unicode mode.
367
368 32. Added the "forbid" facility to pcretest so that putting tests into the
369 wrong test files can sometimes be quickly detected.
370
371 33. There is now a limit (default 250) on the depth of nesting of parentheses.
372 This limit is imposed to control the amount of system stack used at compile
373 time. It can be changed at build time by --with-parens-nest-limit=xxx or
374 the equivalent in CMake.
375
376 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
377 errors. Perl warns for these when in warning mode, but PCRE has no facility
378 for giving warnings.
379
380 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
381 However, this was not working for (?!) because it is optimized to (*FAIL),
382 for which PCRE does not allow quantifiers. The optimization is now disabled
383 when a quantifier follows (?!). I can't see any use for this, but it makes
384 things uniform.
385
386 36. Perl no longer allows group names to start with digits, so I have made this
387 change also in PCRE. It simplifies the code a bit.
388
389 37. In extended mode, Perl ignores spaces before a + that indicates a
390 possessive quantifier. PCRE allowed a space before the quantifier, but not
391 before the possessive +. It now does.
392
393 38. The use of \K (reset reported match start) within a repeated possessive
394 group such as (a\Kb)*+ was not working.
395
396 40. Document that the same character tables must be used at compile time and
397 run time, and that the facility to pass tables to pcre_exec() and
398 pcre_dfa_exec() is for use only with saved/restored patterns.
399
400 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
401 features for Builds with MSVC:
402
403 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
404 builds). The .rc files can be used to set FileDescription and many other
405 attributes.
406
407 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
408 This allows higher-level build scripts which want .pdb files to avoid
409 hard-coding the exact files needed."
410
411 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
412 mean "start of word" and "end of word", respectively, as a transition aid.
413
414 43. A minimizing repeat of a class containing codepoints greater than 255 in
415 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
416 compiled to use the heap for recursion.
417
418 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
419 is configured.
420
421
422 Version 8.33 28-May-2013
423 ------------------------
424
425 1. Added 'U' to some constants that are compared to unsigned integers, to
426 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
427 variables that are added to signed variables, to ensure the result is
428 signed and can be negated.
429
430 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
431 CMake config files.
432
433 3. Revise the creation of config.h.generic so that all boolean macros are
434 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
435 overriding via -D on the command line possible.
436
437 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
438 to unsigned int is reported to make a quite noticeable speed difference in
439 a specific Windows environment. Testing on Linux did also appear to show
440 some benefit (and it is clearly not harmful). Also fixed the definition of
441 Xop which should be unsigned.
442
443 5. Related to (4), changing the definition of the intermediate variable cc
444 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
445 improvements.
446
447 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
448 unnecessary spaces.
449
450 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
451 and later.
452
453 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
454
455 9. Optimizing fast_forward_start_bits in JIT.
456
457 10. Adding support for callouts in JIT, and fixing some issues revealed
458 during this work. Namely:
459
460 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
461
462 (b) Minimum length was not checked before the matching is started.
463
464 11. The value of capture_last that is passed to callouts was incorrect in some
465 cases when there was a capture on one path that was subsequently abandoned
466 after a backtrack. Also, the capture_last value is now reset after a
467 recursion, since all captures are also reset in this case.
468
469 12. The interpreter no longer returns the "too many substrings" error in the
470 case when an overflowing capture is in a branch that is subsequently
471 abandoned after a backtrack.
472
473 13. In the pathological case when an offset vector of size 2 is used, pcretest
474 now prints out the matched string after a yield of 0 or 1.
475
476 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
477 Only supported by the JIT compiler at the moment.
478
479 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
480
481 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
482 offset of the starting point of the matching process, provided the offsets
483 vector is large enough.
484
485 17. The \A escape now records a lookbehind value of 1, though its execution
486 does not actually inspect the previous character. This is to ensure that,
487 in partial multi-segment matching, at least one character from the old
488 segment is retained when a new segment is processed. Otherwise, if there
489 are no lookbehinds in the pattern, \A might match incorrectly at the start
490 of a new segment.
491
492 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
493
494 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
495 gives some modest performance improvement in 8-bit mode.
496
497 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
498 be expressed in certain programming languages using Universal Character
499 Names.
500
501 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
502 which points out that "non characters" are not "characters that may not
503 appear in Unicode strings" but rather "characters that are reserved for
504 internal use and have only local meaning".
505
506 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
507 there was a conditional group that depended on an assertion, if the
508 assertion was false, the callout that immediately followed the alternation
509 in the condition was skipped when pcre_exec() was used for matching.
510
511 23. Allow an explicit callout to be inserted before an assertion that is the
512 condition for a conditional group, for compatibility with automatic
513 callouts, which always insert a callout at this point.
514
515 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
516 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
517
518 25. (*PRUNE) is now supported by the JIT compiler.
519
520 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
521
522 27. Fix the case where there are two or more SKIPs with arguments that may be
523 ignored.
524
525 28. (*SKIP) is now supported by the JIT compiler.
526
527 29. (*THEN) is now supported by the JIT compiler.
528
529 30. Update RunTest with additional test selector options.
530
531 31. The way PCRE handles backtracking verbs has been changed in two ways.
532
533 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
534 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
535 backtracking. In some cases this makes it more Perl-compatible, but Perl's
536 rather obscure rules do not always do the same thing.
537
538 (2) Previously, backtracking verbs were confined within assertions. This is
539 no longer the case for positive assertions, except for (*ACCEPT). Again,
540 this sometimes improves Perl compatibility, and sometimes does not.
541
542 32. A number of tests that were in test 2 because Perl did things differently
543 have been moved to test 1, because either Perl or PCRE has changed, and
544 these tests are now compatible.
545
546 32. Backtracking control verbs are now handled in the same way in JIT and
547 interpreter.
548
549 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
550 contained a forward subroutine reference caused a compile error.
551
552 34. Auto-detect and optimize limited repetitions in JIT.
553
554 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
555 blocking (*UTF) etc.
556
557 36. In the interpreter, maximizing pattern repetitions for characters and
558 character types now use tail recursion, which reduces stack usage.
559
560 37. The value of the max lookbehind was not correctly preserved if a compiled
561 and saved regex was reloaded on a host of different endianness.
562
563 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
564 of the compiled pattern block, expand the flags field from 16 to 32 bits
565 because it was almost full.
566
567 39. Try madvise first before posix_madvise.
568
569 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
570 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
571 original change locked out all matches of empty strings. This has been
572 changed so that one match of an empty string per line is recognized.
573 Subsequent searches on the same line (for colouring or for --only-matching,
574 for example) do not recognize empty strings.
575
576 41. Applied a user patch to fix a number of spelling mistakes in comments.
577
578 42. Data lines longer than 65536 caused pcretest to crash.
579
580 43. Clarified the data type for length and startoffset arguments for pcre_exec
581 and pcre_dfa_exec in the function-specific man pages, where they were
582 explicitly stated to be in bytes, never having been updated. I also added
583 some clarification to the pcreapi man page.
584
585 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
586 a segmentation fault.
587
588
589 Version 8.32 30-November-2012
590 -----------------------------
591
592 1. Improved JIT compiler optimizations for first character search and single
593 character iterators.
594
595 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
596 Patch by Daniel Richard G.
597
598 3. Single character iterator optimizations in the JIT compiler.
599
600 4. Improved JIT compiler optimizations for character ranges.
601
602 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
603 Reported by Giuseppe D'Angelo.
604
605 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
606 of a line, was being set incorrectly in cases where .* appeared inside
607 atomic brackets at the start of a pattern, or where there was a subsequent
608 *PRUNE or *SKIP.
609
610 7. Improved instruction cache flush for POWER/PowerPC.
611 Patch by Daniel Richard G.
612
613 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
614 grep:
615
616 (a) There is now no limit to the number of patterns to be matched.
617
618 (b) An error is given if a pattern is too long.
619
620 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
621 are now supported.
622
623 (d) --exclude-from and --include-from (multiple use) have been added.
624
625 (e) Exclusions and inclusions now apply to all files and directories, not
626 just to those obtained from scanning a directory recursively.
627
628 (f) Multiple uses of -f and --file-list are now supported.
629
630 (g) In a Windows environment, the default for -d has been changed from
631 "read" (the GNU grep default) to "skip", because otherwise the presence
632 of a directory in the file list provokes an error.
633
634 (h) The documentation has been revised and clarified in places.
635
636 9. Improve the matching speed of capturing brackets.
637
638 10. Changed the meaning of \X so that it now matches a Unicode extended
639 grapheme cluster.
640
641 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
642 out POSIX threads when JIT support is configured.
643
644 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
645
646 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
647 structure could go wrong in environments where size_t is not the same size
648 as int.
649
650 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
651 set.
652
653 15. The EBCDIC support had decayed; later updates to the code had included
654 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
655 general tidy up of EBCDIC-related issues, and the documentation was also
656 not quite right. There is now a test that can be run on ASCII systems to
657 check some of the EBCDIC-related things (but is it not a full test).
658
659 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
660 in a small tidy to the code.
661
662 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
663
664 18. If the --only-matching (-o) option in pcregrep is specified multiple
665 times, each one causes appropriate output. For example, -o1 -o2 outputs the
666 substrings matched by the 1st and 2nd capturing parentheses. A separating
667 string can be specified by --om-separator (default empty).
668
669 19. Improving the first n character searches.
670
671 20. Turn case lists for horizontal and vertical white space into macros so that
672 they are defined only once.
673
674 21. This set of changes together give more compatible Unicode case-folding
675 behaviour for characters that have more than one other case when UCP
676 support is available.
677
678 (a) The Unicode property table now has offsets into a new table of sets of
679 three or more characters that are case-equivalent. The MultiStage2.py
680 script that generates these tables (the pcre_ucd.c file) now scans
681 CaseFolding.txt instead of UnicodeData.txt for character case
682 information.
683
684 (b) The code for adding characters or ranges of characters to a character
685 class has been abstracted into a generalized function that also handles
686 case-independence. In UTF-mode with UCP support, this uses the new data
687 to handle characters with more than one other case.
688
689 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
690 whose other case is greater than 256 are now correctly matched
691 caselessly. Previously, the high codepoint matched the low one, but not
692 vice versa.
693
694 (d) The processing of \h, \H, \v, and \ in character classes now makes use
695 of the new class addition function, using character lists defined as
696 macros alongside the case definitions of 20 above.
697
698 (e) Caseless back references now work with characters that have more than
699 one other case.
700
701 (f) General caseless matching of characters with more than one other case
702 is supported.
703
704 22. Unicode character properties were updated from Unicode 6.2.0
705
706 23. Improved CMake support under Windows. Patch by Daniel Richard G.
707
708 24. Add support for 32-bit character strings, and UTF-32
709
710 25. Major JIT compiler update (code refactoring and bugfixing).
711 Experimental Sparc 32 support is added.
712
713 26. Applied a modified version of Daniel Richard G's patch to create
714 pcre.h.generic and config.h.generic by "make" instead of in the
715 PrepareRelease script.
716
717 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
718 pcre_compile.c when checking for a zero character.
719
720 28. Introducing a native interface for JIT. Through this interface, the compiled
721 machine code can be directly executed. The purpose of this interface is to
722 provide fast pattern matching, so several sanity checks are not performed.
723 However, feature tests are still performed. The new interface provides
724 1.4x speedup compared to the old one.
725
726 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
727 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
728 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
729
730 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
731 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
732 the "old" RFC 2279). Instead, it ended up passing a negative length to
733 pcre_exec().
734
735 31. Add support for GCC's visibility feature to hide internal functions.
736
737 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
738 "unknown -C option" after outputting 0 or 1.
739
740 33. There is now support for generating a code coverage report for the test
741 suite in environments where gcc is the compiler and lcov is installed. This
742 is mainly for the benefit of the developers.
743
744 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
745 unaddressable using valgrind annotations, allowing valgrind to detect
746 invalid memory accesses. This is mainly for the benefit of the developers.
747
748 25. (*UTF) can now be used to start a pattern in any of the three libraries.
749
750 26. Give configure error if --enable-cpp but no C++ compiler found.
751
752
753 Version 8.31 06-July-2012
754 -------------------------
755
756 1. Fixing a wrong JIT test case and some compiler warnings.
757
758 2. Removed a bashism from the RunTest script.
759
760 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
761 to unsigned type, result still unsigned" that was given by an MS compiler
762 on encountering the code "-sizeof(xxx)".
763
764 4. Partial matching support is added to the JIT compiler.
765
766 5. Fixed several bugs concerned with partial matching of items that consist
767 of more than one character:
768
769 (a) /^(..)\1/ did not partially match "aba" because checking references was
770 done on an "all or nothing" basis. This also applied to repeated
771 references.
772
773 (b) \R did not give a hard partial match if \r was found at the end of the
774 subject.
775
776 (c) \X did not give a hard partial match after matching one or more
777 characters at the end of the subject.
778
779 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
780 a partial match for the string "\r".
781
782 (e) When newline was set to CRLF, the metacharacter "." did not recognize
783 a partial match for a CR character at the end of the subject string.
784
785 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
786 running pcretest, the text "(JIT)" added to the output whenever JIT is
787 actually used to run the match.
788
789 7. Individual JIT compile options can be set in pcretest by following -s+[+]
790 or /S+[+] with a digit between 1 and 7.
791
792 8. OP_NOT now supports any UTF character not just single-byte ones.
793
794 9. (*MARK) control verb is now supported by the JIT compiler.
795
796 10. The command "./RunTest list" lists the available tests without actually
797 running any of them. (Because I keep forgetting what they all are.)
798
799 11. Add PCRE_INFO_MAXLOOKBEHIND.
800
801 12. Applied a (slightly modified) user-supplied patch that improves performance
802 when the heap is used for recursion (compiled with --disable-stack-for-
803 recursion). Instead of malloc and free for each heap frame each time a
804 logical recursion happens, frames are retained on a chain and re-used where
805 possible. This sometimes gives as much as 30% improvement.
806
807 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
808 call.
809
810 14. As documented, (*COMMIT) is now confined to within a positive assertion.
811
812 15. It is now possible to link pcretest with libedit as an alternative to
813 libreadline.
814
815 16. (*COMMIT) control verb is now supported by the JIT compiler.
816
817 17. The Unicode data tables have been updated to Unicode 6.1.0.
818
819 18. Added --file-list option to pcregrep.
820
821 19. Added binary file support to pcregrep, including the -a, --binary-files,
822 -I, and --text options.
823
824 20. The madvise function is renamed for posix_madvise for QNX compatibility
825 reasons. Fixed by Giuseppe D'Angelo.
826
827 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
828
829 22. Changed the option for creating symbolic links for 16-bit man pages from
830 -s to -sf so that re-installing does not cause issues.
831
832 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
833
834 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
835 matches in certain environments (the workspace was not being correctly
836 retained). Also added to pcre_dfa_exec() a simple plausibility check on
837 some of the workspace data at the beginning of a restart.
838
839 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
840 was not doing so when it should - probably a typo introduced by SVN 528
841 (change 8.10/14).
842
843 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
844 \w+ when the character tables indicated that \x{c4} was a word character.
845 There were several related cases, all because the tests for doing a table
846 lookup were testing for characters less than 127 instead of 255.
847
848 27. If a pattern contains capturing parentheses that are not used in a match,
849 their slots in the ovector are set to -1. For those that are higher than
850 any matched groups, this happens at the end of processing. In the case when
851 there were back references that the ovector was too small to contain
852 (causing temporary malloc'd memory to be used during matching), and the
853 highest capturing number was not used, memory off the end of the ovector
854 was incorrectly being set to -1. (It was using the size of the temporary
855 memory instead of the true size.)
856
857 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
858 ovector size, it uses memory at the end of the block that it has got.
859
860 29. Check for an overlong MARK name and give an error at compile time. The
861 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
862
863 30. JIT compiler update.
864
865 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
866 Rill for the patch.
867
868 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
869
870 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
871
872 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
873 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
874 was enabled.
875
876 35. Improve JIT code generation for greedy plus quantifier.
877
878 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
879 1 to "aa" instead of to an empty string. The bug affected repeated groups
880 that could potentially match an empty string.
881
882 37. Optimizing single character iterators in JIT.
883
884 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
885 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
886 codepoints that are too big for the mode are faulted, and in a UTF mode,
887 disallowed codepoints are also faulted.
888
889 39. If PCRE was compiled with UTF support, in three places in the DFA
890 matcher there was code that should only have been obeyed in UTF mode, but
891 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
892 processing when bytes with values greater than 127 were present. In 16-bit
893 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
894 both cases the values are those that cannot be the first data item in a UTF
895 character. The three items that might have provoked this were recursions,
896 possessively repeated groups, and atomic groups.
897
898 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
899 and pcregrep, because some OS require shared objects to be explicitly
900 passed to ld, causing the link step to fail if they are not.
901
902 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
903 mode, patterns that started with \h* or \R* might be incorrectly matched.
904
905
906 Version 8.30 04-February-2012
907 -----------------------------
908
909 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
910 name is defined in ctype.h.
911
912 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
913 only in quite long subpatterns.
914
915 3. Removed the function pcre_info(), which has been obsolete and deprecated
916 since it was replaced by pcre_fullinfo() in February 2000.
917
918 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
919 match a (*MARK), and the match failed at the start of the subject, a
920 reference to memory before the start of the subject could occur. This bug
921 was introduced by fix 17 of release 8.21.
922
923 5. A reference to an unset group with zero minimum repetition was giving
924 totally wrong answers (in non-JavaScript-compatibility mode). For example,
925 /(another)?(\1?)test/ matched against "hello world test". This bug was
926 introduced in release 8.13.
927
928 6. Add support for 16-bit character strings (a large amount of work involving
929 many changes and refactorings).
930
931 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
932 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
933 from a file.
934
935 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
936 rounding is not applied in this particular case).
937
938 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
939 if they appear, or are escaped, in patterns.
940
941 10. Get rid of a number of -Wunused-but-set-variable warnings.
942
943 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
944 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
945 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
946 also returns the mark "x". This bug applied to capturing parentheses,
947 non-capturing parentheses, and atomic parentheses. It also applied to some
948 assertions.
949
950 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
951 information out of configure.ac instead of relying on pcre.h.generic, which
952 is not stored in the repository.
953
954 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
955 -lreadline.
956
957 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
958
959 15. Applied Graycode's patch to put the top-level frame on the stack rather
960 than the heap when not using the stack for recursion. This gives a
961 performance improvement in many cases when recursion is not deep.
962
963 16. Experimental code added to "pcretest -C" to output the stack frame size.
964
965
966 Version 8.21 12-Dec-2011
967 ------------------------
968
969 1. Updating the JIT compiler.
970
971 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
972 are added as well.
973
974 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
975 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
976 calling _pcre_jit_exec. Some extra comments are added.
977
978 4. (*MARK) settings inside atomic groups that do not contain any capturing
979 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
980 was introduced by change 18 for 8.20.
981
982 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
983 ECMA-262 standard.
984
985 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
986 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
987 This bug was probably introduced by change 9 of 8.13.
988
989 7. While fixing 6 above, I noticed that a number of other items were being
990 incorrectly rejected as "not fixed length". This arose partly because newer
991 opcodes had not been added to the fixed-length checking code. I have (a)
992 corrected the bug and added tests for these items, and (b) arranged for an
993 error to occur if an unknown opcode is encountered while checking for fixed
994 length instead of just assuming "not fixed length". The items that were
995 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
996 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
997 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
998
999 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1000 being incorrectly compiled and would have given unpredicatble results.
1001
1002 9. A possessively repeated subpattern with minimum repeat count greater than
1003 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1004 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1005 the first (A) could occur when it should not.
1006
1007 10. Add a cast and remove a redundant test from the code.
1008
1009 11. JIT should use pcre_malloc/pcre_free for allocation.
1010
1011 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1012 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1013 is anything other than /usr, -L is still shown).
1014
1015 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1016
1017 14. Perl does not support \N without a following name in a [] class; PCRE now
1018 also gives an error.
1019
1020 15. If a forward reference was repeated with an upper limit of around 2000,
1021 it caused the error "internal error: overran compiling workspace". The
1022 maximum number of forward references (including repeats) was limited by the
1023 internal workspace, and dependent on the LINK_SIZE. The code has been
1024 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1025 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1026 of around 200,000 forward references. While doing this, I also speeded up
1027 the filling in of repeated forward references.
1028
1029 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1030 incorrectly expecting the subject to contain another "a" after the start.
1031
1032 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1033 in the match, the SKIP should be ignored. This was not happening; instead
1034 the SKIP was being treated as NOMATCH. For patterns such as
1035 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1036 tested.
1037
1038 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1039 now much more compatible with Perl, in particular in cases where the result
1040 is a non-match for a non-anchored pattern. For example, if
1041 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1042 "m", where previously it did not return a name. A side effect of this
1043 change is that for partial matches, the last encountered mark name is
1044 returned, as for non matches. A number of tests that were previously not
1045 Perl-compatible have been moved into the Perl-compatible test files. The
1046 refactoring has had the pleasing side effect of removing one argument from
1047 the match() function, thus reducing its stack requirements.
1048
1049 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1050 subsequent uses of /S (without +) incorrectly behaved like /S+.
1051
1052 21. Retrieve executable code size support for the JIT compiler and fixing
1053 some warnings.
1054
1055 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1056 not work when the shorter character appeared right at the end of the
1057 subject string.
1058
1059 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1060 systems.
1061
1062 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1063 output it when the /M option is used in pcretest.
1064
1065 25. The CheckMan script was not being included in the distribution. Also, added
1066 an explicit "perl" to run Perl scripts from the PrepareRelease script
1067 because this is reportedly needed in Windows.
1068
1069 26. If study data was being save in a file and studying had not found a set of
1070 "starts with" bytes for the pattern, the data written to the file (though
1071 never used) was taken from uninitialized memory and so caused valgrind to
1072 complain.
1073
1074 27. Updated RunTest.bat as provided by Sheri Pierce.
1075
1076 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1077
1078 29. Computation of memory usage for the table of capturing group names was
1079 giving an unnecessarily large value.
1080
1081
1082 Version 8.20 21-Oct-2011
1083 ------------------------
1084
1085 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1086 a POSIX class. After further experiments with Perl, which convinced me that
1087 Perl has bugs and confusions, a closing square bracket is no longer allowed
1088 in a POSIX name. This bug also affected patterns with classes that started
1089 with full stops.
1090
1091 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1092 captured substring, but while checking the failing first alternative,
1093 substring 1 is temporarily captured. If the output vector supplied to
1094 pcre_exec() was not big enough for this capture, the yield of the function
1095 was still zero ("insufficient space for captured substrings"). This cannot
1096 be totally fixed without adding another stack variable, which seems a lot
1097 of expense for a edge case. However, I have improved the situation in cases
1098 such as /(a)(b)x|abc/ matched against "abc", where the return code
1099 indicates that fewer than the maximum number of slots in the ovector have
1100 been set.
1101
1102 3. Related to (2) above: when there are more back references in a pattern than
1103 slots in the output vector, pcre_exec() uses temporary memory during
1104 matching, and copies in the captures as far as possible afterwards. It was
1105 using the entire output vector, but this conflicts with the specification
1106 that only 2/3 is used for passing back captured substrings. Now it uses
1107 only the first 2/3, for compatibility. This is, of course, another edge
1108 case.
1109
1110 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1111 main code base, and can be used by building with --enable-jit. When this is
1112 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1113 runtime --no-jit option is given.
1114
1115 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1116 ovector, the return from the function was zero, implying that there were
1117 other matches that did not fit. The correct "exactly full" value is now
1118 returned.
1119
1120 6. If a subpattern that was called recursively or as a subroutine contained
1121 (*PRUNE) or any other control that caused it to give a non-standard return,
1122 invalid errors such as "Error -26 (nested recursion at the same subject
1123 position)" or even infinite loops could occur.
1124
1125 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1126 computing the minimum length on reaching *ACCEPT, and so ended up with the
1127 wrong value of 1 rather than 0. Further investigation indicates that
1128 computing a minimum subject length in the presence of *ACCEPT is difficult
1129 (think back references, subroutine calls), and so I have changed the code
1130 so that no minimum is registered for a pattern that contains *ACCEPT.
1131
1132 8. If (*THEN) was present in the first (true) branch of a conditional group,
1133 it was not handled as intended. [But see 16 below.]
1134
1135 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1136 Sheri Pierce.
1137
1138 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1139 the first byte in a match must be "a".
1140
1141 11. Change 17 for 8.13 increased the recursion depth for patterns like
1142 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1143 pattern contains any instances of (*THEN). If it does not, the old
1144 optimizations are restored. It would be nice to do this on a per-group
1145 basis, but at the moment that is not feasible.
1146
1147 12. In some environments, the output of pcretest -C is CRLF terminated. This
1148 broke RunTest's code that checks for the link size. A single white space
1149 character after the value is now allowed for.
1150
1151 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1152 For "fr", it uses the Windows-specific input and output files.
1153
1154 14. If (*THEN) appeared in a group that was called recursively or as a
1155 subroutine, it did not work as intended. [But see next item.]
1156
1157 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1158 pattern fragments (but not containing any | characters). If A and B are
1159 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1160 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1161 D. In other words, Perl considers parentheses that do not contain any |
1162 characters to be part of a surrounding alternative, whereas PCRE was
1163 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1164 differently. PCRE now behaves in the same way as Perl, except in the case
1165 of subroutine/recursion calls such as (?1) which have in any case always
1166 been different (but PCRE had them first :-).
1167
1168 16. Related to 15 above: Perl does not treat the | in a conditional group as
1169 creating alternatives. Such a group is treated in the same way as an
1170 ordinary group without any | characters when processing (*THEN). PCRE has
1171 been changed to match Perl's behaviour.
1172
1173 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1174 RunGrepTest script failed.
1175
1176 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1177 inevitable for groups that contain captures, but it can lead to a lot of
1178 stack use in large patterns. The old behaviour has been restored for atomic
1179 groups that do not contain any capturing parentheses.
1180
1181 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1182 suppress the check for a minimum subject length at run time. (If it was
1183 given to pcre_exec() or pcre_dfa_exec() it did work.)
1184
1185 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1186 fail to work when decoding hex characters in data strings in EBCDIC
1187 environments.
1188
1189 21. It appears that in at least one Mac OS environment, the isxdigit() function
1190 is implemented as a macro that evaluates to its argument more than once,
1191 contravening the C 90 Standard (I haven't checked a later standard). There
1192 was an instance in pcretest which caused it to go wrong when processing
1193 \x{...} escapes in subject strings. The has been rewritten to avoid using
1194 things like p++ in the argument of isxdigit().
1195
1196
1197 Version 8.13 16-Aug-2011
1198 ------------------------
1199
1200 1. The Unicode data tables have been updated to Unicode 6.0.0.
1201
1202 2. Two minor typos in pcre_internal.h have been fixed.
1203
1204 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1205 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1206 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1207
1208 4. There were a number of related bugs in the code for matching backrefences
1209 caselessly in UTF-8 mode when codes for the characters concerned were
1210 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1211 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1212 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1213 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1214 2-byte code at the end of the subject (it thought there wasn't enough data
1215 left).
1216
1217 5. Comprehensive information about what went wrong is now returned by
1218 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1219 as the output vector has at least 2 elements. The offset of the start of
1220 the failing character and a reason code are placed in the vector.
1221
1222 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1223 now returned is for the first byte of the failing character, instead of the
1224 last byte inspected. This is an incompatible change, but I hope it is small
1225 enough not to be a problem. It makes the returned offset consistent with
1226 pcre_exec() and pcre_dfa_exec().
1227
1228 7. pcretest now gives a text phrase as well as the error number when
1229 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1230 failure, the offset and reason code are output.
1231
1232 8. When \R was used with a maximizing quantifier it failed to skip backwards
1233 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1234 back over a single character (\n). This seems wrong (because it treated the
1235 two characters as a single entity when going forwards), conflicts with the
1236 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1237 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1238 has been changed.
1239
1240 9. Some internal refactoring has changed the processing so that the handling
1241 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1242 time (the PCRE_DOTALL option was changed this way some time ago: version
1243 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1244 which was always a bit of a fudge. It also means that there is one less
1245 argument for the match() function, which reduces its stack requirements
1246 slightly. This change also fixes an incompatibility with Perl: the pattern
1247 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1248
1249 10. More internal refactoring has drastically reduced the number of recursive
1250 calls to match() for possessively repeated groups such as (abc)++ when
1251 using pcre_exec().
1252
1253 11. While implementing 10, a number of bugs in the handling of groups were
1254 discovered and fixed:
1255
1256 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1257 (a|)*(?1) gave a compile-time internal error.
1258 ((a|)+)+ did not notice that the outer group could match an empty string.
1259 (^a|^)+ was not marked as anchored.
1260 (.*a|.*)+ was not marked as matching at start or after a newline.
1261
1262 12. Yet more internal refactoring has removed another argument from the match()
1263 function. Special calls to this function are now indicated by setting a
1264 value in a variable in the "match data" data block.
1265
1266 13. Be more explicit in pcre_study() instead of relying on "default" for
1267 opcodes that mean there is no starting character; this means that when new
1268 ones are added and accidentally left out of pcre_study(), testing should
1269 pick them up.
1270
1271 14. The -s option of pcretest has been documented for ages as being an old
1272 synonym of -m (show memory usage). I have changed it to mean "force study
1273 for every regex", that is, assume /S for every regex. This is similar to -i
1274 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1275 using it. It makes it easier to run collections of tests with and without
1276 study enabled, and thereby test pcre_study() more easily. All the standard
1277 tests are now run with and without -s (but some patterns can be marked as
1278 "never study" - see 20 below).
1279
1280 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1281 restoration of the capturing data to the outer values was not happening
1282 correctly.
1283
1284 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1285 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1286 pattern had matched an empty string, and so incorrectly returned a no
1287 match.
1288
1289 17. There was optimizing code for the last branch of non-capturing parentheses,
1290 and also for the obeyed branch of a conditional subexpression, which used
1291 tail recursion to cut down on stack usage. Unfortunately, now that there is
1292 the possibility of (*THEN) occurring in these branches, tail recursion is
1293 no longer possible because the return has to be checked for (*THEN). These
1294 two optimizations have therefore been removed. [But see 8.20/11 above.]
1295
1296 18. If a pattern containing \R was studied, it was assumed that \R always
1297 matched two bytes, thus causing the minimum subject length to be
1298 incorrectly computed because \R can also match just one byte.
1299
1300 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1301 was incorrectly computed.
1302
1303 20. If /S is present twice on a test pattern in pcretest input, it now
1304 *disables* studying, thereby overriding the use of -s on the command line
1305 (see 14 above). This is necessary for one or two tests to keep the output
1306 identical in both cases.
1307
1308 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1309 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1310
1311 22. When an atomic group that contained a capturing parenthesis was
1312 successfully matched, but the branch in which it appeared failed, the
1313 capturing was not being forgotten if a higher numbered group was later
1314 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1315 group 1 to "a", when in fact it should be unset. This applied to multi-
1316 branched capturing and non-capturing groups, repeated or not, and also to
1317 positive assertions (capturing in negative assertions does not happen
1318 in PCRE) and also to nested atomic groups.
1319
1320 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1321 subject after a captured substring, to make it easier to tell which of a
1322 number of identical substrings has been captured.
1323
1324 24. The way atomic groups are processed by pcre_exec() has been changed so that
1325 if they are repeated, backtracking one repetition now resets captured
1326 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1327 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1328 "aaa". Previously, it would have been "a". As part of this code
1329 refactoring, the way recursive calls are handled has also been changed.
1330
1331 25. If an assertion condition captured any substrings, they were not passed
1332 back unless some other capturing happened later. For example, if
1333 (?(?=(a))a) was matched against "a", no capturing was returned.
1334
1335 26. When studying a pattern that contained subroutine calls or assertions,
1336 the code for finding the minimum length of a possible match was handling
1337 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1338 group 1 called group 2 while simultaneously a separate group 2 called group
1339 1). A stack overflow occurred in this case. I have fixed this by limiting
1340 the recursion depth to 10.
1341
1342 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1343 Fortmann. This supports explicit test numbers on the command line, and has
1344 argument validation and error reporting.
1345
1346 28. An instance of \X with an unlimited repeat could fail if at any point the
1347 first character it looked at was a mark character.
1348
1349 29. Some minor code refactoring concerning Unicode properties and scripts
1350 should reduce the stack requirement of match() slightly.
1351
1352 30. Added the '=' option to pcretest to check the setting of unused capturing
1353 slots at the end of the pattern, which are documented as being -1, but are
1354 not included in the return count.
1355
1356 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1357 compiled something random. Now it gives a compile-time error (as does
1358 Perl).
1359
1360 32. A *MARK encountered during the processing of a positive assertion is now
1361 recorded and passed back (compatible with Perl).
1362
1363 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1364 had alternative anchored branches, the search for a second match in a line
1365 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1366 matched the line "0102" twice. The same bug affected patterns that started
1367 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1368 twice.
1369
1370 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1371 does, and because of capturing effects, quantifying parenthesized
1372 assertions may at times be useful. Quantifiers are now allowed for
1373 parenthesized assertions.
1374
1375 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1376
1377 36. \g was being checked for fancy things in a character class, when it should
1378 just be a literal "g".
1379
1380 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1381 appearance of a nested POSIX class supersedes an apparent external class.
1382 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1383 unescaped square brackets may also appear as part of class names. For
1384 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1385 more like Perl. (But see 8.20/1 above.)
1386
1387 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1388 was because it thought it was \N{name}, which is not supported).
1389
1390 39. Add minix to OS list not supporting the -S option in pcretest.
1391
1392 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1393 cannot analyze patterns in sufficient detail to catch mutual recursions
1394 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1395 subgroup is called recursively as a subpattern for a second time at the
1396 same position in the subject string. In previous releases this might have
1397 been caught by the recursion limit, or it might have run out of stack.
1398
1399 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1400 happen only once. PCRE was, however incorrectly giving a compile time error
1401 "recursive call could loop indefinitely" because it cannot analyze the
1402 pattern in sufficient detail. The compile time test no longer happens when
1403 PCRE is compiling a conditional subpattern, but actual runaway loops are
1404 now caught at runtime (see 40 above).
1405
1406 42. It seems that Perl allows any characters other than a closing parenthesis
1407 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1408 has been changed to be the same.
1409
1410 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1411 as not to get warnings when autogen.sh is called. Also changed
1412 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1413
1414 44. To help people who use pcregrep to scan files containing exceedingly long
1415 lines, the following changes have been made:
1416
1417 (a) The default value of the buffer size parameter has been increased from
1418 8K to 20K. (The actual buffer used is three times this size.)
1419
1420 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1421 PCRE is built.
1422
1423 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1424 to be set at run time.
1425
1426 (d) Numerical values in pcregrep options can be followed by K or M, for
1427 example --buffer-size=50K.
1428
1429 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1430 given and the return code is set to 2.
1431
1432 45. Add a pointer to the latest mark to the callout data block.
1433
1434 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1435 partial match of an empty string instead of no match. This was specific to
1436 the use of ".".
1437
1438 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1439 complete match instead of a partial match. This bug was dependent on both
1440 the PCRE_UTF8 and PCRE_DOTALL options being set.
1441
1442 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1443 starting byte set, because \b was not being ignored.
1444
1445
1446 Version 8.12 15-Jan-2011
1447 ------------------------
1448
1449 1. Fixed some typos in the markup of the man pages, and wrote a script that
1450 checks for such things as part of the documentation building process.
1451
1452 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1453 --match-limit and --recursion-limit options (added for 8.11). In
1454 particular, this made one of the standard tests fail. (The integer value
1455 went into the wrong half of a long int.)
1456
1457 3. If the --colour option was given to pcregrep with -v (invert match), it
1458 did strange things, either producing crazy output, or crashing. It should,
1459 of course, ignore a request for colour when reporting lines that do not
1460 match.
1461
1462 4. Another pcregrep bug caused similar problems if --colour was specified with
1463 -M (multiline) and the pattern match finished with a line ending.
1464
1465 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1466 matched in multiline mode, the following line was shown as part of the
1467 match. This seems wrong, so I have changed it.
1468
1469 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1470 the check for further matches in the same line (so they could be coloured)
1471 to overrun the end of the current line. If another match was found, it was
1472 incorrectly shown (and then shown again when found in the next line).
1473
1474 7. If pcregrep was compiled under Windows, there was a reference to the
1475 function pcregrep_exit() before it was defined. I am assuming this was
1476 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1477 reported by a user. I've moved the definition above the reference.
1478
1479
1480 Version 8.11 10-Dec-2010
1481 ------------------------
1482
1483 1. (*THEN) was not working properly if there were untried alternatives prior
1484 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1485 backtracked to try for "b" instead of moving to the next alternative branch
1486 at the same level (in this case, to look for "c"). The Perl documentation
1487 is clear that when (*THEN) is backtracked onto, it goes to the "next
1488 alternative in the innermost enclosing group".
1489
1490 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1491 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1492 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1493 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1494 (*THEN).
1495
1496 3. If \s appeared in a character class, it removed the VT character from
1497 the class, even if it had been included by some previous item, for example
1498 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1499 of \s, but is part of the POSIX "space" class.)
1500
1501 4. A partial match never returns an empty string (because you can always
1502 match an empty string at the end of the subject); however the checking for
1503 an empty string was starting at the "start of match" point. This has been
1504 changed to the "earliest inspected character" point, because the returned
1505 data for a partial match starts at this character. This means that, for
1506 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1507 (previously it gave "no match").
1508
1509 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1510 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1511 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1512 has an implication that the given string is incomplete (because a partial
1513 match is preferred over a full match). For this reason, these items now
1514 give a partial match in this situation. [Aside: previously, the one case
1515 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1516 match rather than a full match, which was wrong by the old rules, but is
1517 now correct.]
1518
1519 6. There was a bug in the handling of #-introduced comments, recognized when
1520 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1521 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1522 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1523 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1524 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1525 places in pcre_compile().
1526
1527 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1528 comments when looking ahead for named forward references to subpatterns,
1529 the only newline sequence it recognized was NL. It now handles newlines
1530 according to the set newline convention.
1531
1532 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1533 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1534 cater for a lack of strerror(). These oversights have been fixed.
1535
1536 9. Added --match-limit and --recursion-limit to pcregrep.
1537
1538 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1539
1540 11. When the -o option was used, pcregrep was setting a return code of 1, even
1541 when matches were found, and --line-buffered was not being honoured.
1542
1543 12. Added an optional parentheses number to the -o and --only-matching options
1544 of pcregrep.
1545
1546 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1547 can match an empty string. The code to do it in pcretest and pcredemo
1548 needed fixing:
1549
1550 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1551 only one byte after an empty string match just before CRLF (this case
1552 just got forgotten; "any" and "anycrlf" were OK).
1553
1554 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1555 mode when an empty string match preceded an ASCII character followed by
1556 a non-ASCII character. (The code for advancing by one character rather
1557 than one byte was nonsense.)
1558
1559 (c) The pcredemo.c sample program did not have any code at all to handle
1560 the cases when CRLF is a valid newline sequence.
1561
1562 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1563 as a starting offset was within the subject string. There is now a new
1564 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1565 negative or greater than the length of the string. In order to test this,
1566 pcretest is extended to allow the setting of negative starting offsets.
1567
1568 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1569 starting offset points to the beginning of a UTF-8 character was
1570 unnecessarily clumsy. I tidied it up.
1571
1572 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1573 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1574
1575 17. Nobody had reported that the --include_dir option, which was added in
1576 release 7.7 should have been called --include-dir (hyphen, not underscore)
1577 for compatibility with GNU grep. I have changed it to --include-dir, but
1578 left --include_dir as an undocumented synonym, and the same for
1579 --exclude-dir, though that is not available in GNU grep, at least as of
1580 release 2.5.4.
1581
1582 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1583 characters from a string of bytes) have been redefined so as not to use
1584 loops, in order to improve performance in some environments. At the same
1585 time, I abstracted some of the common code into auxiliary macros to save
1586 repetition (this should not affect the compiled code).
1587
1588 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1589 compile-time error is now given if \c is not followed by an ASCII
1590 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1591 different, and any byte value is allowed.)
1592
1593 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1594 START_OPTIMIZE option, which is now allowed at compile time - but just
1595 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1596 to pcregrep and other applications that have no direct access to PCRE
1597 options. The new /Y option in pcretest sets this option when calling
1598 pcre_compile().
1599
1600 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1601 back references. Groups containing recursive back references were forced to
1602 be atomic by that change, but in the case of named groups, the amount of
1603 memory required was incorrectly computed, leading to "Failed: internal
1604 error: code overflow". This has been fixed.
1605
1606 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1607 pcretest.c, to avoid build problems in some Borland environments.
1608
1609
1610 Version 8.10 25-Jun-2010
1611 ------------------------
1612
1613 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1614 THEN.
1615
1616 2. (*ACCEPT) was not working when inside an atomic group.
1617
1618 3. Inside a character class, \B is treated as a literal by default, but
1619 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1620 causes the error). The code is unchanged, but I tidied the documentation.
1621
1622 4. Inside a character class, PCRE always treated \R and \X as literals,
1623 whereas Perl faults them if its -w option is set. I have changed PCRE so
1624 that it faults them when PCRE_EXTRA is set.
1625
1626 5. Added support for \N, which always matches any character other than
1627 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1628
1629 6. When compiling pcregrep with newer versions of gcc which may have
1630 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1631 declared with attribute warn_unused_result" were given. Just casting the
1632 result to (void) does not stop the warnings; a more elaborate fudge is
1633 needed. I've used a macro to implement this.
1634
1635 7. Minor change to pcretest.c to avoid a compiler warning.
1636
1637 8. Added four artifical Unicode properties to help with an option to make
1638 \s etc use properties (see next item). The new properties are: Xan
1639 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1640
1641 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1642 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1643 this option. Modified pcretest to add /W to test this facility. Added
1644 REG_UCP to make it available via the POSIX interface.
1645
1646 10. Added --line-buffered to pcregrep.
1647
1648 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1649 studied, and the match started with a letter with a code point greater than
1650 127 whose first byte was different to the first byte of the other case of
1651 the letter, the other case of this starting letter was not recognized
1652 (#976).
1653
1654 12. If a pattern that was studied started with a repeated Unicode property
1655 test, for example, \p{Nd}+, there was the theoretical possibility of
1656 setting up an incorrect bitmap of starting bytes, but fortunately it could
1657 not have actually happened in practice until change 8 above was made (it
1658 added property types that matched character-matching opcodes).
1659
1660 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1661 possible starting bytes for non-anchored patterns.
1662
1663 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1664 \R, and also a number of cases that involve Unicode properties, both
1665 explicit and implicit when PCRE_UCP is set.
1666
1667 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1668 input, it could crash or give wrong results if characters with values
1669 greater than 0xc0 were present in the subject string. (Detail: it assumed
1670 UTF-8 input when processing these items.)
1671
1672 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1673 size_t is 64-bit (#991).
1674
1675 17. Added a check for running out of memory when PCRE is compiled with
1676 --disable-stack-for-recursion (#990).
1677
1678 18. If the last data line in a file for pcretest does not have a newline on
1679 the end, a newline was missing in the output.
1680
1681 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1682 less than 128) in its various bitmaps. However, there is a facility for
1683 generating tables according to the current locale when PCRE is compiled. It
1684 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1685 characters, are recognized by isspace() and therefore were getting set in
1686 these tables, and indeed these tables seem to approximate to ISO 8859. This
1687 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1688 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1689 which of course cannot start UTF-8 characters. I have changed the code so
1690 that only real ASCII characters (less than 128) and the correct starting
1691 bytes for UTF-8 encodings are set for characters greater than 127 when in
1692 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1693 altogether.)
1694
1695 20. Added the /T option to pcretest so as to be able to run tests with non-
1696 standard character tables, thus making it possible to include the tests
1697 used for 19 above in the standard set of tests.
1698
1699 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1700 reference to a subpattern the other side of a comment that contains an
1701 opening parenthesis caused either an internal compiling error, or a
1702 reference to the wrong subpattern.
1703
1704
1705 Version 8.02 19-Mar-2010
1706 ------------------------
1707
1708 1. The Unicode data tables have been updated to Unicode 5.2.0.
1709
1710 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1711 configured.
1712
1713 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1714 original author of that file, following a query about its status.
1715
1716 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1717 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1718
1719 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1720 quantifier applied to a forward-referencing subroutine call, could compile
1721 incorrect code or give the error "internal error: previously-checked
1722 referenced subpattern not found".
1723
1724 6. Both MS Visual Studio and Symbian OS have problems with initializing
1725 variables to point to external functions. For these systems, therefore,
1726 pcre_malloc etc. are now initialized to local functions that call the
1727 relevant global functions.
1728
1729 7. There were two entries missing in the vectors called coptable and poptable
1730 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1731 I've fixed the data, and added a kludgy way of testing at compile time that
1732 the lengths are correct (equal to the number of opcodes).
1733
1734 8. Following on from 7, I added a similar kludge to check the length of the
1735 eint vector in pcreposix.c.
1736
1737 9. Error texts for pcre_compile() are held as one long string to avoid too
1738 much relocation at load time. To find a text, the string is searched,
1739 counting zeros. There was no check for running off the end of the string,
1740 which could happen if a new error number was added without updating the
1741 string.
1742
1743 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1744
1745 11. \K was not working if it appeared in an atomic group or in a group that
1746 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1747 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1748 the assertion is positive, but not if it is negative.
1749
1750 12. Change 11 fortuitously reduced the size of the stack frame used in the
1751 "match()" function of pcre_exec.c by one pointer. Forthcoming
1752 implementation of support for (*MARK) will need an extra pointer on the
1753 stack; I have reserved it now, so that the stack frame size does not
1754 decrease.
1755
1756 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1757 item in branch that calls a recursion is a subroutine call - as in the
1758 second branch in the above example - was incorrectly given the compile-
1759 time error "recursive call could loop indefinitely" because pcre_compile()
1760 was not correctly checking the subroutine for matching a non-empty string.
1761
1762 14. The checks for overrunning compiling workspace could trigger after an
1763 overrun had occurred. This is a "should never occur" error, but it can be
1764 triggered by pathological patterns such as hundreds of nested parentheses.
1765 The checks now trigger 100 bytes before the end of the workspace.
1766
1767 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1768
1769
1770 Version 8.01 19-Jan-2010
1771 ------------------------
1772
1773 1. If a pattern contained a conditional subpattern with only one branch (in
1774 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1775 computed the wrong minimum data length (which is of course zero for such
1776 subpatterns). This could cause incorrect "no match" results.
1777
1778 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1779 the pattern is reset in the first branch, pcre_compile() failed with
1780 "internal error: code overflow at offset...". This happened only when
1781 the reset was to the original external option setting. (An optimization
1782 abstracts leading options settings into an external setting, which was the
1783 cause of this.)
1784
1785 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1786 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1787 assertion pattern did not match (meaning that the assertion was true), it
1788 was incorrectly treated as false if the SKIP had been reached during the
1789 matching. This also applied to assertions used as conditions.
1790
1791 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1792 assertion subpattern, including such a pattern used as a condition,
1793 unpredictable results occurred, instead of the error return
1794 PCRE_ERROR_DFA_UITEM.
1795
1796 5. The C++ GlobalReplace function was not working like Perl for the special
1797 situation when an empty string is matched. It now does the fancy magic
1798 stuff that is necessary.
1799
1800 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1801 removed. (These were left over from very, very early versions of PCRE.)
1802
1803 7. Some cosmetic changes to the code to make life easier when compiling it
1804 as part of something else:
1805
1806 (a) Change DEBUG to PCRE_DEBUG.
1807
1808 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1809 called "current" as "current_branch", to prevent a collision with the
1810 Linux macro when compiled as a kernel module.
1811
1812 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1813 prevent a collision with the Linux macro when compiled as a kernel
1814 module.
1815
1816 8. In pcre_compile() there are some checks for integer overflows that used to
1817 cast potentially large values to (double). This has been changed to that
1818 when building, a check for int64_t is made, and if it is found, it is used
1819 instead, thus avoiding the use of floating point arithmetic. (There is no
1820 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1821 double.
1822
1823 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1824 2005 (difference between two addresses compared to an unsigned value).
1825
1826 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1827 custom one, because of the following reported problem in Windows:
1828
1829 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1830 under Win32.
1831 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1832 therefore missing the function definition.
1833 - The compiler thus generates a "C" signature for the test function.
1834 - The linker fails to find the "C" function.
1835 - PCRE fails to configure if asked to do so against libbz2.
1836
1837 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1838 messages were output:
1839
1840 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1841 rerunning libtoolize, to keep the correct libtool macros in-tree.
1842 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1843
1844 I have done both of these things.
1845
1846 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1847 most of the time, it *can* run out if it is given a pattern that contains a
1848 runaway infinite recursion. I updated the discussion in the pcrestack man
1849 page.
1850
1851 13. Now that we have gone to the x.xx style of version numbers, the minor
1852 version may start with zero. Using 08 or 09 is a bad idea because users
1853 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1854 interpreted as invalid octal numbers. I've updated the previous comment in
1855 configure.ac, and also added a check that gives an error if 08 or 09 are
1856 used.
1857
1858 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1859 causing partial matching to fail when the end of the subject matched \W
1860 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1861
1862 15. There were some discrepancies between the declarations in pcre_internal.h
1863 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1864 their definitions. The declarations used "const uschar *" and the
1865 definitions used USPTR. Even though USPTR is normally defined as "const
1866 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1867 reported that: "This difference in casting confuses some C++ compilers, for
1868 example, SunCC recognizes above declarations as different functions and
1869 generates broken code for hbpcre." I have changed the declarations to use
1870 USPTR.
1871
1872 16. GNU libtool is named differently on some systems. The autogen.sh script now
1873 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1874 (FreeBSD).
1875
1876 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1877 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1878 comment: "Figure out how to create a longlong from a string: strtoll and
1879 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1880 instance, but it only takes 2 args instead of 3!"
1881
1882 18. A subtle bug concerned with back references has been fixed by a change of
1883 specification, with a corresponding code fix. A pattern such as
1884 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1885 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1886 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1887 same bug. Such groups have to be quantified to be useful, or contained
1888 inside another quantified group. (If there's no repetition, the reference
1889 can never match.) The problem arises because, having left the group and
1890 moved on to the rest of the pattern, a later failure that backtracks into
1891 the group uses the captured value from the final iteration of the group
1892 rather than the correct earlier one. I have fixed this in PCRE by forcing
1893 any group that contains a reference to itself to be an atomic group; that
1894 is, there cannot be any backtracking into it once it has completed. This is
1895 similar to recursive and subroutine calls.
1896
1897
1898 Version 8.00 19-Oct-09
1899 ----------------------
1900
1901 1. The table for translating pcre_compile() error codes into POSIX error codes
1902 was out-of-date, and there was no check on the pcre_compile() error code
1903 being within the table. This could lead to an OK return being given in
1904 error.
1905
1906 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1907 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1908 in a Windows environment.
1909
1910 3. The pcregrep --count option prints the count for each file even when it is
1911 zero, as does GNU grep. However, pcregrep was also printing all files when
1912 --files-with-matches was added. Now, when both options are given, it prints
1913 counts only for those files that have at least one match. (GNU grep just
1914 prints the file name in this circumstance, but including the count seems
1915 more useful - otherwise, why use --count?) Also ensured that the
1916 combination -clh just lists non-zero counts, with no names.
1917
1918 4. The long form of the pcregrep -F option was incorrectly implemented as
1919 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1920 but it seems right to fix it, and I didn't think it was worth preserving
1921 the old behaviour.
1922
1923 5. The command line items --regex=pattern and --regexp=pattern were not
1924 recognized by pcregrep, which required --regex pattern or --regexp pattern
1925 (with a space rather than an '='). The man page documented the '=' forms,
1926 which are compatible with GNU grep; these now work.
1927
1928 6. No libpcreposix.pc file was created for pkg-config; there was just
1929 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1930
1931 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1932 when UCP support is not needed, by modifying the Python script that
1933 generates it from Unicode data files. This should not matter if the module
1934 is correctly used as a library, but I received one complaint about 50K of
1935 unwanted data. My guess is that the person linked everything into his
1936 program rather than using a library. Anyway, it does no harm.
1937
1938 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1939 was a minimum greater than 1 for a wide character in a possessive
1940 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1941 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1942 character. Chaos in the form of incorrect output or a compiling loop could
1943 result.
1944
1945 9. The restrictions on what a pattern can contain when partial matching is
1946 requested for pcre_exec() have been removed. All patterns can now be
1947 partially matched by this function. In addition, if there are at least two
1948 slots in the offset vector, the offset of the earliest inspected character
1949 for the match and the offset of the end of the subject are set in them when
1950 PCRE_ERROR_PARTIAL is returned.
1951
1952 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1953 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1954 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1955 and may be more useful for multi-segment matching.
1956
1957 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1958 used to be given if ever the end of the subject was reached; now it is
1959 given only if matching could not proceed because another character was
1960 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1961 string "Z", which now yields "no match" instead of "partial match". In the
1962 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1963 final character ended with (*FAIL).
1964
1965 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1966 if the pattern had a "must contain" character that was already found in the
1967 earlier partial match, unless partial matching was again requested. For
1968 example, with the pattern /dog.(body)?/, the "must contain" character is
1969 "g". If the first part-match was for the string "dog", restarting with
1970 "sbody" failed. This bug has been fixed.
1971
1972 13. The string returned by pcre_dfa_exec() after a partial match has been
1973 changed so that it starts at the first inspected character rather than the
1974 first character of the match. This makes a difference only if the pattern
1975 starts with a lookbehind assertion or \b or \B (\K is not supported by
1976 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1977 matching functions compatible, and I think it's the right thing to do.
1978
1979 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1980 so that the demonstration program is easily available in environments where
1981 PCRE has not been installed from source.
1982
1983 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1984 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1985 library.
1986
1987 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1988 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1989 is not the first non-POSIX option to be added. Clearly some people find
1990 these options useful.
1991
1992 17. If a caller to the POSIX matching function regexec() passes a non-zero
1993 value for nmatch with a NULL value for pmatch, the value of
1994 nmatch is forced to zero.
1995
1996 18. RunGrepTest did not have a test for the availability of the -u option of
1997 the diff command, as RunTest does. It now checks in the same way as
1998 RunTest, and also checks for the -b option.
1999
2000 19. If an odd number of negated classes containing just a single character
2001 interposed, within parentheses, between a forward reference to a named
2002 subpattern and the definition of the subpattern, compilation crashed with
2003 an internal error, complaining that it could not find the referenced
2004 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2005 [The bug was that it was starting one character too far in when skipping
2006 over the character class, thus treating the ] as data rather than
2007 terminating the class. This meant it could skip too much.]
2008
2009 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2010 /g option in pcretest when the pattern contains \K, which makes it possible
2011 to have an empty string match not at the start, even when the pattern is
2012 anchored. Updated pcretest and pcredemo to use this option.
2013
2014 21. If the maximum number of capturing subpatterns in a recursion was greater
2015 than the maximum at the outer level, the higher number was returned, but
2016 with unset values at the outer level. The correct (outer level) value is
2017 now given.
2018
2019 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2020 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2021 make it do so. The string so far is captured, making this feature
2022 compatible with Perl.
2023
2024 23. The tests have been re-organized, adding tests 11 and 12, to make it
2025 possible to check the Perl 5.10 features against Perl 5.10.
2026
2027 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2028 pattern matches a fixed length string. PCRE did not allow this; now it
2029 does. Neither allows recursion.
2030
2031 25. I finally figured out how to implement a request to provide the minimum
2032 length of subject string that was needed in order to match a given pattern.
2033 (It was back references and recursion that I had previously got hung up
2034 on.) This code has now been added to pcre_study(); it finds a lower bound
2035 to the length of subject needed. It is not necessarily the greatest lower
2036 bound, but using it to avoid searching strings that are too short does give
2037 some useful speed-ups. The value is available to calling programs via
2038 pcre_fullinfo().
2039
2040 26. While implementing 25, I discovered to my embarrassment that pcretest had
2041 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2042 study optimizations had never been tested with that matching function.
2043 Oops. What is worse, even when it was passed study data, there was a bug in
2044 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2045 were also very few tests of studied patterns with pcre_dfa_exec().
2046
2047 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2048 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2049 on the other side of the coin, they are no longer allowed to have different
2050 names, because these cannot be distinguished in PCRE, and this has caused
2051 confusion. (This is a difference from Perl.)
2052
2053 28. When duplicate subpattern names are present (necessarily with different
2054 numbers, as required by 27 above), and a test is made by name in a
2055 conditional pattern, either for a subpattern having been matched, or for
2056 recursion in such a pattern, all the associated numbered subpatterns are
2057 tested, and the overall condition is true if the condition is true for any
2058 one of them. This is the way Perl works, and is also more like the way
2059 testing by number works.
2060
2061
2062 Version 7.9 11-Apr-09
2063 ---------------------
2064
2065 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2066 (pcretest), all targets were linked against these libraries. This included
2067 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2068 libraries. This caused unwanted dependencies to be created. This problem
2069 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2070 pcretest is linked with readline.
2071
2072 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2073 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2074 moved outside it again, because FALSE and TRUE are already defined in AIX,
2075 but BOOL is not.
2076
2077 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2078 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2079
2080 4. The pcregrep documentation said spaces were inserted as well as colons (or
2081 hyphens) following file names and line numbers when outputting matching
2082 lines. This is not true; no spaces are inserted. I have also clarified the
2083 wording for the --colour (or --color) option.
2084
2085 5. In pcregrep, when --colour was used with -o, the list of matching strings
2086 was not coloured; this is different to GNU grep, so I have changed it to be
2087 the same.
2088
2089 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2090 each matching line was coloured. Now it goes on to look for further matches
2091 of any of the test patterns, which is the same behaviour as GNU grep.
2092
2093 7. A pattern that could match an empty string could cause pcregrep to loop; it
2094 doesn't make sense to accept an empty string match in pcregrep, so I have
2095 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2096 seems to be how GNU grep behaves. [But see later change 40 for release
2097 8.33.]
2098
2099 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2100 start or after a newline", because the conditional assertion was not being
2101 correctly handled. The rule now is that both the assertion and what follows
2102 in the first alternative must satisfy the test.
2103
2104 9. If auto-callout was enabled in a pattern with a conditional group whose
2105 condition was an assertion, PCRE could crash during matching, both with
2106 pcre_exec() and pcre_dfa_exec().
2107
2108 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2109 used for matching.
2110
2111 11. Unicode property support in character classes was not working for
2112 characters (bytes) greater than 127 when not in UTF-8 mode.
2113
2114 12. Added the -M command line option to pcretest.
2115
2116 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2117
2118 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2119
2120 16. Added comments and documentation about mis-use of no_arg in the C++
2121 wrapper.
2122
2123 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2124 from Martin Jerabek that uses macro names for all relevant character and
2125 string constants.
2126
2127 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2128 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2129 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2130 these, but not everybody uses configure.
2131
2132 19. A conditional group that had only one branch was not being correctly
2133 recognized as an item that could match an empty string. This meant that an
2134 enclosing group might also not be so recognized, causing infinite looping
2135 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2136 with the subject "ab", where knowledge that the repeated group can match
2137 nothing is needed in order to break the loop.
2138
2139 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2140 exec(), but without supplying a callout function, matching went wrong.
2141
2142 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2143 leak if the size of the offset vector was greater than 30. When the vector
2144 is smaller, the saved offsets during recursion go onto a local stack
2145 vector, but for larger vectors malloc() is used. It was failing to free
2146 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2147 error, in fact).
2148
2149 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2150 heapframe that is used only when UTF-8 support is enabled. This caused no
2151 problem, but was untidy.
2152
2153 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2154 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2155 included within another project.
2156
2157 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2158 slightly modified by me:
2159
2160 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2161 not building pcregrep.
2162
2163 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2164 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2165
2166 25. Forward references, both numeric and by name, in patterns that made use of
2167 duplicate group numbers, could behave incorrectly or give incorrect errors,
2168 because when scanning forward to find the reference group, PCRE was not
2169 taking into account the duplicate group numbers. A pattern such as
2170 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2171
2172 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2173 the feature of a custom pointer more persuasive (as requested by a user).
2174
2175 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2176 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2177 pre-defined.
2178
2179 28. Added support for (*UTF8) at the start of a pattern.
2180
2181 29. Arrange for flags added by the "release type" setting in CMake to be shown
2182 in the configuration summary.
2183
2184
2185 Version 7.8 05-Sep-08
2186 ---------------------
2187
2188 1. Replaced UCP searching code with optimized version as implemented for Ad
2189 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2190 stage table and inline lookup instead of a function, giving speed ups of 2
2191 to 5 times on some simple patterns that I tested. Permission was given to
2192 distribute the MultiStage2.py script that generates the tables (it's not in
2193 the tarball, but is in the Subversion repository).
2194
2195 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2196 scripts.
2197
2198 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2199 a group with a zero qualifier. The result of the study could be incorrect,
2200 or the function might crash, depending on the pattern.
2201
2202 4. Caseless matching was not working for non-ASCII characters in back
2203 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2204 It now works when Unicode Property Support is available.
2205
2206 5. In pcretest, an escape such as \x{de} in the data was always generating
2207 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2208 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2209 truncation.
2210
2211 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2212
2213 7. Added two (int) casts to pcregrep when printing the difference of two
2214 pointers, in case they are 64-bit values.
2215
2216 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2217 test 2 if it fails.
2218
2219 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2220 and a #define of that name to empty if it is not externally set. This is to
2221 allow users of MSVC to set it if necessary.
2222
2223 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2224 the convenience functions in the pcre_get.c source file.
2225
2226 11. An option change at the start of a pattern that had top-level alternatives
2227 could cause overwriting and/or a crash. This command provoked a crash in
2228 some environments:
2229
2230 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2231
2232 This potential security problem was recorded as CVE-2008-2371.
2233
2234 12. For a pattern where the match had to start at the beginning or immediately
2235 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2236 pcre_dfa_exec() could read past the end of the passed subject if there was
2237 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2238 pcretest so that it places the subject at the end of its malloc-ed buffer.
2239
2240 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2241 exec() might read past the end of the data buffer in UTF-8 mode.
2242
2243 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2244 the data contained the byte 0x85 as part of a UTF-8 character within its
2245 first line. This applied both to normal and DFA matching.
2246
2247 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2248 /^[^d]*?$/8 failed to match "abc".
2249
2250 16. Added a missing copyright notice to pcrecpp_internal.h.
2251
2252 17. Make it more clear in the documentation that values returned from
2253 pcre_exec() in ovector are byte offsets, not character counts.
2254
2255 18. Tidied a few places to stop certain compilers from issuing warnings.
2256
2257 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2258 supplied by Stefan Weber. I made a further small update for 7.8 because
2259 there is a change of source arrangements: the pcre_searchfuncs.c module is
2260 replaced by pcre_ucd.c.
2261
2262
2263 Version 7.7 07-May-08
2264 ---------------------
2265
2266 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2267 a string to a long long, pretend we don't even have a long long." This is
2268 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2269
2270 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2271 pre-7.6 versions, which defined a global no_arg variable instead of putting
2272 it in the RE class. (See also #8 below.)
2273
2274 3. Remove a line of dead code, identified by coverity and reported by Nuno
2275 Lopes.
2276
2277 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2278
2279 (1) The include/exclude patterns were being applied to the whole pathnames
2280 of files, instead of just to the final components.
2281
2282 (2) If there was more than one level of directory, the subdirectories were
2283 skipped unless they satisfied the include/exclude conditions. This is
2284 inconsistent with GNU grep (and could even be seen as contrary to the
2285 pcregrep specification - which I improved to make it absolutely clear).
2286 The action now is always to scan all levels of directory, and just
2287 apply the include/exclude patterns to regular files.
2288
2289 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2290 --exclude_dir in the tests to avoid scanning .svn directories.
2291
2292 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2293 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2294 doesn't support NULs in patterns.
2295
2296 7. Added some missing "const"s to declarations of static tables in
2297 pcre_compile.c and pcre_dfa_exec.c.
2298
2299 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2300 caused by fix #2 above. (Subsequently also a second patch to fix the
2301 first patch. And a third patch - this was a messy problem.)
2302
2303 9. Applied Craig's patch to remove the use of push_back().
2304
2305 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2306 matching function regexec().
2307
2308 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2309 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2310 references. PCRE supports relative numbers with this syntax (I don't think
2311 Oniguruma does).
2312
2313 12. Previously, a group with a zero repeat such as (...){0} was completely
2314 omitted from the compiled regex. However, this means that if the group
2315 was called as a subroutine from elsewhere in the pattern, things went wrong
2316 (an internal error was given). Such groups are now left in the compiled
2317 pattern, with a new opcode that causes them to be skipped at execution
2318 time.
2319
2320 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2321 to the way PCRE behaves:
2322
2323 (a) A lone ] character is dis-allowed (Perl treats it as data).
2324
2325 (b) A back reference to an unmatched subpattern matches an empty string
2326 (Perl fails the current match path).
2327
2328 (c) A data ] in a character class must be notated as \] because if the
2329 first data character in a class is ], it defines an empty class. (In
2330 Perl it is not possible to have an empty class.) The empty class []
2331 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2332 The negative empty class [^] matches any one character, independently
2333 of the DOTALL setting.
2334
2335 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2336 non-existent subpattern following a character class starting with ']' and
2337 containing () gave an internal compiling error instead of "reference to
2338 non-existent subpattern". Fortunately, when the pattern did exist, the
2339 compiled code was correct. (When scanning forwards to check for the
2340 existence of the subpattern, it was treating the data ']' as terminating
2341 the class, so got the count wrong. When actually compiling, the reference
2342 was subsequently set up correctly.)
2343
2344 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2345 it was being rejected as not supported by pcre_dfa_exec(), even though
2346 other assertions are supported. I have made pcre_dfa_exec() support
2347 (*FAIL).
2348
2349 16. The implementation of 13c above involved the invention of a new opcode,
2350 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2351 cannot be changed at match time, I realized I could make a small
2352 improvement to matching performance by compiling OP_ALLANY instead of
2353 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2354 on the OP_ANY path.
2355
2356 17. Compiling pcretest on Windows with readline support failed without the
2357 following two fixes: (1) Make the unistd.h include conditional on
2358 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2359
2360 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2361 ncurses library to be included for pcretest when ReadLine support is
2362 requested, but also to allow for it to be overridden. This patch came from
2363 Daniel Bergström.
2364
2365 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2366 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2367 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2368 spotting this.
2369
2370
2371 Version 7.6 28-Jan-08
2372 ---------------------
2373
2374 1. A character class containing a very large number of characters with
2375 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2376 overflow.
2377
2378 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2379 HAVE_LONG_LONG is not defined.
2380
2381 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2382 bring them up to date and include new features. This patch includes:
2383
2384 - Fixed PH's badly added libz and libbz2 support.
2385 - Fixed a problem with static linking.
2386 - Added pcredemo. [But later removed - see 7 below.]
2387 - Fixed dftables problem and added an option.
2388 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2389 HAVE_LONG_LONG.
2390 - Added readline support for pcretest.
2391 - Added an listing of the option settings after cmake has run.
2392
2393 4. A user submitted a patch to Makefile that makes it easy to create
2394 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2395 Makefile.am that cause it to include this special target, without
2396 affecting anything else. Note that the same mingw target plus all
2397 the other distribution libraries and programs are now supported
2398 when configuring with CMake (see 6 below) instead of with
2399 Configure/Make.
2400
2401 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2402 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2403 exported in the Windows port". It has not yet been confirmed that the patch
2404 solves the problem, but it does no harm.
2405
2406 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2407 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2408 with CMake, and also correct the comment about stack recursion.
2409
2410 7. Remove the automatic building of pcredemo from the ./configure system and
2411 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2412 of a program that users should build themselves after PCRE is installed, so
2413 building it automatically is not really right. What is more, it gave
2414 trouble in some build environments.
2415
2416 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2417
2418
2419 Version 7.5 10-Jan-08
2420 ---------------------
2421
2422 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2423 values in parens when parsing an RE using the C++ wrapper."
2424
2425 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2426 Characters greater than 255 were excluded from the class instead of being
2427 included.
2428
2429 3. The same bug as (2) above applied to negated POSIX classes such as
2430 [:^space:].
2431
2432 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2433 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2434 I have changed it.
2435
2436 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2437 first named subpattern) and a construct such as (?&a) would reference the
2438 first named subpattern whose name started with "a" (in other words, the
2439 length check was missing). Both these problems are fixed. "Subpattern name
2440 expected" is now given for (?&) (a zero-length name), and this patch also
2441 makes it give the same error for \k'' (previously it complained that that
2442 was a reference to a non-existent subpattern).
2443
2444 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2445 this is right because (?- can be followed by option settings as well as by
2446 digits. I have, however, made the messages clearer.
2447
2448 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2449 than the number used in the conditional) now cause a compile-time error.
2450 This is actually not compatible with Perl, which accepts such patterns, but
2451 treats the conditional as always being FALSE (as PCRE used to), but it
2452 seems to me that giving a diagnostic is better.
2453
2454 8. Change "alphameric" to the more common word "alphanumeric" in comments
2455 and messages.
2456
2457 9. Fix two occurrences of "backslash" in comments that should have been
2458 "backspace".
2459
2460 10. Remove two redundant lines of code that can never be obeyed (their function
2461 was moved elsewhere).
2462
2463 11. The program that makes PCRE's Unicode character property table had a bug
2464 which caused it to generate incorrect table entries for sequences of
2465 characters that have the same character type, but are in different scripts.
2466 It amalgamated them into a single range, with the script of the first of
2467 them. In other words, some characters were in the wrong script. There were
2468 thirteen such cases, affecting characters in the following ranges:
2469
2470 U+002b0 - U+002c1
2471 U+0060c - U+0060d
2472 U+0061e - U+00612
2473 U+0064b - U+0065e
2474 U+0074d - U+0076d
2475 U+01800 - U+01805
2476 U+01d00 - U+01d77
2477 U+01d9b - U+01dbf
2478 U+0200b - U+0200f
2479 U+030fc - U+030fe
2480 U+03260 - U+0327f
2481 U+0fb46 - U+0fbb1
2482 U+10450 - U+1049d
2483
2484 12. The -o option (show only the matching part of a line) for pcregrep was not
2485 compatible with GNU grep in that, if there was more than one match in a
2486 line, it showed only the first of them. It now behaves in the same way as
2487 GNU grep.
2488
2489 13. If the -o and -v options were combined for pcregrep, it printed a blank
2490 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2491 does the same. The return code can be used to tell if there were any
2492 non-matching lines.
2493
2494 14. Added --file-offsets and --line-offsets to pcregrep.
2495
2496 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2497 infinitely looping recursion. The bug was that positive lookaheads were not
2498 being skipped when checking for a possible empty match (negative lookaheads
2499 and both kinds of lookbehind were skipped).
2500
2501 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2502 inclusion of <windows.h> to before rather than after the definition of
2503 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2504
2505 17. Specifying a possessive quantifier with a specific limit for a Unicode
2506 character property caused pcre_compile() to compile bad code, which led at
2507 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2508 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2509 caused the error; without that there was no problem.
2510
2511 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2512
2513 19. Added --enable-pcretest-libreadline.
2514
2515 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2516 RE::GlobalReplace(). As a result, the number of replacements returned was
2517 double what it should be. I removed one of the increments, but Craig sent a
2518 later patch that removed the other one (the right fix) and added unit tests
2519 that check the return values (which was not done before).
2520
2521 21. Several CMake things:
2522
2523 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2524 the names libpcre and libpcreposix, not just pcre and pcreposix.
2525
2526 (2) The above change means that pcretest and pcregrep are now correctly
2527 linked with the newly-built libraries, not previously installed ones.
2528
2529 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2530
2531 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2532 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2533 UTF-8 newline character). The key issue is that the pattern starts .*;
2534 this means that the match must be either at the beginning, or after a
2535 newline. The bug was in the code for advancing after a failed match and
2536 checking that the new position followed a newline. It was not taking
2537 account of UTF-8 characters correctly.
2538
2539 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2540 character classes. PCRE was not treating the sequence [:...:] as a
2541 character class unless the ... were all letters. Perl, however, seems to
2542 allow any characters between [: and :], though of course it rejects as
2543 unknown any "names" that contain non-letters, because all the known class
2544 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2545 for example, whereas PCRE did not - it did not recognize a POSIX character
2546 class. This seemed a bit dangerous, so the code has been changed to be
2547 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2548 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2549 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2550 Perl does, and where it didn't before.
2551
2552 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2553 Windows environments %n is disabled by default.
2554
2555
2556 Version 7.4 21-Sep-07
2557 ---------------------
2558
2559 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2560 means that a class such as [\s] counted as "explicit reference to CR or
2561 LF". That isn't really right - the whole point of the change was to try to
2562 help when there was an actual mention of one of the two characters. So now
2563 the change happens only if \r or \n (or a literal CR or LF) character is
2564 encountered.
2565
2566 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2567 of both had grown to the point where there were only 3 bits left.
2568 Fortunately, there was spare space in the data structure, and so I have
2569 moved the internal flags into a new 16-bit field to free up more option
2570 bits.
2571
2572 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2573 but did not set the internal JCHANGED flag - either of these is enough to
2574 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2575 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2576 start sets both bits.
2577
2578 4. Added options (at build time, compile time, exec time) to change \R from
2579 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2580
2581 5. doc/pcresyntax.html was missing from the distribution.
2582
2583 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2584 compatibility, even though it is no longer used.
2585
2586 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2587 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2588 windows.h file is present (where different names are used). [This was
2589 reversed later after testing - see 16 below.]
2590
2591 8. Changed all #include <config.h> to #include "config.h". There were also
2592 some further <pcre.h> cases that I changed to "pcre.h".
2593
2594 9. When pcregrep was used with the --colour option, it missed the line ending
2595 sequence off the lines that it output.
2596
2597 10. It was pointed out to me that arrays of string pointers cause lots of
2598 relocations when a shared library is dynamically loaded. A technique of
2599 using a single long string with a table of offsets can drastically reduce
2600 these. I have refactored PCRE in four places to do this. The result is
2601 dramatic:
2602
2603 Originally: 290
2604 After changing UCP table: 187
2605 After changing error message table: 43
2606 After changing table of "verbs" 36
2607 After changing table of Posix names 22
2608
2609 Thanks to the folks working on Gregex for glib for this insight.
2610
2611 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2612 unicode-properties was also set.
2613
2614 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2615
2616 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2617 checked only for CRLF.
2618
2619 14. Added casts to pcretest.c to avoid compiler warnings.
2620
2621 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2622
2623 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2624 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2625 entirely. This removes changes made in 7 above.
2626
2627 17. The CMake files have been updated, and there is now more information about
2628 building with CMake in the NON-UNIX-USE document.
2629
2630
2631 Version 7.3 28-Aug-07
2632 ---------------------
2633
2634 1. In the rejigging of the build system that eventually resulted in 7.1, the
2635 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2636 brackets there is not right, since it causes compilers to look for an
2637 installed pcre.h, not the version that is in the source that is being
2638 compiled (which of course may be different). I have changed it back to:
2639
2640 #include "pcre.h"
2641
2642 I have a vague recollection that the change was concerned with compiling in
2643 different directories, but in the new build system, that is taken care of
2644 by the VPATH setting the Makefile.
2645
2646 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2647 when the subject happened to end in the byte 0x85 (e.g. if the last
2648 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2649 characters but of course it shouldn't be taken as a newline when it is part
2650 of another character. The bug was that, for an unlimited repeat of . in
2651 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2652 characters when looking for a newline.
2653
2654 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2655
2656 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2657 in debug output.
2658
2659 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2660 long printing in the pcrecpp unittest when running under MinGW.
2661
2662 6. ESC_K was left out of the EBCDIC table.
2663
2664 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2665 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2666 limit also applies to "virtual nesting" when a pattern is recursive, and in
2667 this case 1000 isn't so big. I have been able to remove this limit at the
2668 expense of backing off one optimization in certain circumstances. Normally,
2669 when pcre_exec() would call its internal match() function recursively and
2670 immediately return the result unconditionally, it uses a "tail recursion"
2671 feature to save stack. However, when a subpattern that can match an empty
2672 string has an unlimited repetition quantifier, it no longer makes this
2673 optimization. That gives it a stack frame in which to save the data for
2674 checking that an empty string has been matched. Previously this was taken
2675 from the 1000-entry workspace that had been reserved. So now there is no
2676 explicit limit, but more stack is used.
2677
2678 8. Applied Daniel's patches to solve problems with the import/export magic
2679 syntax that is required for Windows, and which was going wrong for the
2680 pcreposix and pcrecpp parts of the library. These were overlooked when this
2681 problem was solved for the main library.
2682
2683 9. There were some crude static tests to avoid integer overflow when computing
2684 the size of patterns that contain repeated groups with explicit upper
2685 limits. As the maximum quantifier is 65535, the maximum group length was
2686 set at 30,000 so that the product of these two numbers did not overflow a
2687 32-bit integer. However, it turns out that people want to use groups that
2688 are longer than 30,000 bytes (though not repeat them that many times).
2689 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2690 made it possible to implement the integer overflow checks in a much more
2691 dynamic way, which I have now done. The artificial limitation on group
2692 length has been removed - we now have only the limit on the total length of
2693 the compiled pattern, which depends on the LINK_SIZE setting.
2694
2695 10. Fixed a bug in the documentation for get/copy named substring when
2696 duplicate names are permitted. If none of the named substrings are set, the
2697 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2698 empty string.
2699
2700 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2701 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2702 because the ] is interpreted as the first data character and the
2703 terminating ] is not found. PCRE has been made compatible with Perl in this
2704 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2705 cause memory overwriting.
2706
2707 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2708 string has been matched (to stop an infinite loop). It was not recognizing
2709 a conditional subpattern that could match an empty string if that
2710 subpattern was within another subpattern. For example, it looped when
2711 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2712 condition was not nested. This bug has been fixed.
2713
2714 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2715 past the start of the subject in the presence of bytes with the top bit
2716 set, for example "\x8aBCD".
2717
2718 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2719 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2720
2721 14. Optimized (?!) to (*FAIL).
2722
2723 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2724 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2725 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2726 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2727 does: it's just the validity check that is more restrictive.
2728
2729 16. Inserted checks for integer overflows during escape sequence (backslash)
2730 processing, and also fixed erroneous offset values for syntax errors during
2731 backslash processing.
2732
2733 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2734 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2735
2736 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2737 caused an overrun.
2738
2739 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2740 something other than just ASCII characters) inside a group that had an
2741 unlimited repeat caused a loop at compile time (while checking to see
2742 whether the group could match an empty string).
2743
2744 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2745 [\P{Any}] did so. (Error in the code for printing property names.)
2746
2747 21. An orphan \E inside a character class could cause a crash.
2748
2749 22. A repeated capturing bracket such as (A)? could cause a wild memory
2750 reference during compilation.
2751
2752 23. There are several functions in pcre_compile() that scan along a compiled
2753 expression for various reasons (e.g. to see if it's fixed length for look
2754 behind). There were bugs in these functions when a repeated \p or \P was
2755 present in the pattern. These operators have additional parameters compared
2756 with \d, etc, and these were not being taken into account when moving along
2757 the compiled data. Specifically:
2758
2759 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2760 length.
2761
2762 (b) An item such as \pL+ within a repeated group could cause crashes or
2763 loops.
2764
2765 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2766 "reference to non-existent subpattern" error.
2767
2768 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2769
2770 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2771 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2772
2773 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2774
2775 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2776 character were causing crashes (broken optimization).
2777
2778 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2779 \p or \P) caused a compile-time loop.
2780
2781 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2782 break. For example, the unstudied pattern [\r\n]A does not match the string
2783 "\r\nA" because change 7.0/46 below moves the current point on by two
2784 characters after failing to match at the start. However, the pattern \nA
2785 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2786 the same is true. There doesn't seem any very clean way out of this, but
2787 what I have chosen to do makes the common cases work: PCRE now takes note
2788 of whether there can be an explicit match for \r or \n anywhere in the
2789 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2790 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2791 pattern has explicit CR or LF references.
2792
2793 29. Added (*CR) etc for changing newline setting at start of pattern.
2794
2795
2796 Version 7.2 19-Jun-07
2797 ---------------------
2798
2799 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2800 which is apparently normally available under Windows.
2801
2802 2. Re-jig the pcregrep tests with different newline settings in an attempt
2803 to make them independent of the local environment's newline setting.
2804
2805 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2806
2807 4. Some of the "internals" tests were previously cut out when the link size
2808 was not 2, because the output contained actual offsets. The recent new
2809 "Z" feature of pcretest means that these can be cut out, making the tests
2810 usable with all link sizes.
2811
2812 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2813 stack recursion. This gives a massive performance boost under BSD, but just
2814 a small improvement under Linux. However, it saves one field in the frame
2815 in all cases.
2816
2817 6. Added more features from the forthcoming Perl 5.10:
2818
2819 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2820 recursion call. It refers to the nth most recently opened parentheses.
2821
2822 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2823 to be opened parentheses.
2824
2825 (c) Conditions that refer to capturing parentheses can be specified
2826 relatively, for example, (?(-2)... or (?(+3)...
2827
2828 (d) \K resets the start of the current match so that everything before
2829 is not part of it.
2830
2831 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2832
2833 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2834 reference syntax.
2835
2836 (g) (?| introduces a group in which the numbering of parentheses in each
2837 alternative starts with the same number.
2838
2839 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2840
2841 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2842 PCRE_INFO_JCHANGED.
2843
2844 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2845 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2846 for detecting groups that can match an empty string.
2847
2848 9. A pattern with a very large number of alternatives (more than several
2849 hundred) was running out of internal workspace during the pre-compile
2850 phase, where pcre_compile() figures out how much memory will be needed. A
2851 bit of new cunning has reduced the workspace needed for groups with
2852 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2853 workspace instead of running out of the 4096 that are available.
2854
2855 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2856
2857 11. Applied patch from Google to remove an optimization that didn't quite work.
2858 The report of the bug said:
2859
2860 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2861 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2862 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2863
2864 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2865 it matched the wrong number of bytes.
2866
2867
2868 Version 7.1 24-Apr-07
2869 ---------------------
2870
2871 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2872 that is more "standard", making use of automake and other Autotools. There
2873 is some re-arrangement of the files and adjustment of comments consequent
2874 on this.
2875
2876 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2877 for recursive directory scanning broke on some systems because the files
2878 are not scanned in any specific order and on different systems the order
2879 was different. A call to "sort" has been inserted into RunGrepTest for the
2880 approprate test as a short-term fix. In the longer term there may be an
2881 alternative.
2882
2883 3. I had an email from Eric Raymond about problems translating some of PCRE's
2884 man pages to HTML (despite the fact that I distribute HTML pages, some
2885 people do their own conversions for various reasons). The problems
2886 concerned the use of low-level troff macros .br and .in. I have therefore
2887 removed all such uses from the man pages (some were redundant, some could
2888 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2889 HTML has been updated to handle .nf/.fi and to complain if it encounters
2890 .br or .in.
2891
2892 4. Updated comments in configure.ac that get placed in config.h.in and also
2893 arranged for config.h to be included in the distribution, with the name
2894 config.h.generic, for the benefit of those who have to compile without
2895 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2896
2897 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2898 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2899 makevp.bat for the current PCRE, using the additional files
2900 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2901
2902 6. A Windows user reported a minor discrepancy with test 2, which turned out
2903 to be caused by a trailing space on an input line that had got lost in his
2904 copy. The trailing space was an accident, so I've just removed it.
2905
2906 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2907 that is needed.
2908
2909 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2910 as "const" (a) because they are and (b) because it helps the PHP
2911 maintainers who have recently made a script to detect big data structures
2912 in the php code that should be moved to the .rodata section. I remembered
2913 to update Builducptable as well, so it won't revert if ucptable.h is ever
2914 re-created.
2915
2916 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2917 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2918 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2919 support is not required. This saves 1.5-2K of code, which is important in
2920 some applications.
2921
2922 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2923 so as not to refer to the tables, even though these functions will never be
2924 called when UTF-8 support is disabled. Otherwise there are problems with a
2925 shared library.
2926
2927 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2928
2929 (a) It was defining its arguments as char * instead of void *.
2930
2931 (b) It was assuming that all moves were upwards in memory; this was true
2932 a long time ago when I wrote it, but is no longer the case.
2933
2934 The emulated memove() is provided for those environments that have neither
2935 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2936 is clearly not the case, as these two bugs were recently reported.
2937
2938 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2939 and Detrail to create the HTML documentation, the .txt form of the man
2940 pages, and it removes trailing spaces from listed files. It also creates
2941 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2942 case, it wraps all the #defines with #ifndefs. This script should be run
2943 before "make dist".
2944
2945 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2946 with Unicode property support.
2947
2948 (a) For a maximizing quantifier, if the two different cases of the
2949 character were of different lengths in their UTF-8 codings (there are
2950 some cases like this - I found 11), and the matching function had to
2951 back up over a mixture of the two cases, it incorrectly assumed they
2952 were both the same length.
2953
2954 (b) When PCRE was configured to use the heap rather than the stack for
2955 recursion during matching, it was not correctly preserving the data for
2956 the other case of a UTF-8 character when checking ahead for a match
2957 while processing a minimizing repeat. If the check also involved
2958 matching a wide character, but failed, corruption could cause an
2959 erroneous result when trying to check for a repeat of the original
2960 character.
2961
2962 13. Some tidying changes to the testing mechanism:
2963
2964 (a) The RunTest script now detects the internal link size and whether there
2965 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2966 values substituted by "configure". (The RunGrepTest script already did
2967 this for UTF-8.) The configure.ac script no longer substitutes the
2968 relevant variables.
2969
2970 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2971 with length and offset values. This means that the output is different
2972 for different internal link sizes. Test 2 is skipped for link sizes
2973 other than 2 because of this, bypassing the problem. Unfortunately,
2974 there was also a test in test 3 (the locale tests) that used /B and
2975 failed for link sizes other than 2. Rather than cut the whole test out,
2976 I have added a new /Z option to pcretest that replaces the length and
2977 offset values with spaces. This is now used to make test 3 independent
2978 of link size. (Test 2 will be tidied up later.)
2979
2980 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2981 segmentation fault instead of returning the appropriate error message.
2982
2983 15. In multiline mode when the newline sequence was set to "any", the pattern
2984 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2985 This doesn't seem right; it now treats the CRLF combination as the line
2986 ending, and so does not match in that case. It's only a pattern such as ^$
2987 that would hit this one: something like ^ABC$ would have failed after \r
2988 and then tried again after \r\n.
2989
2990 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2991 in an attempt to make files that differ only in their line terminators
2992 compare equal. This works on Linux.
2993
2994 17. Under certain error circumstances pcregrep might try to free random memory
2995 as it exited. This is now fixed, thanks to valgrind.
2996
2997 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2998 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2999 was because its rules for how to advance for /g after matching an empty
3000 string at the end of a line did not allow for this case. They now check for
3001 it specially.
3002
3003 20. pcretest is supposed to handle patterns and data of any length, by
3004 extending its buffers when necessary. It was getting this wrong when the
3005 buffer for a data line had to be extended.
3006
3007 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3008 CRLF as a newline sequence.
3009
3010 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3011 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3012 I have nevertheless tidied it up.
3013
3014 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3015
3016 24. Added a man page for pcre-config.
3017
3018
3019 Version 7.0 19-Dec-06
3020 ---------------------
3021
3022 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3023 moving to gcc 4.1.1.
3024
3025 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3026 sys/time.h, which is documented as needed for this function. It doesn't
3027 seem to matter on Linux, but it showed up on some releases of OS X.
3028
3029 3. It seems that there are systems where bytes whose values are greater than
3030 127 match isprint() in the "C" locale. The "C" locale should be the
3031 default when a C program starts up. In most systems, only ASCII printing
3032 characters match isprint(). This difference caused the output from pcretest
3033 to vary, making some of the tests fail. I have changed pcretest so that:
3034
3035 (a) When it is outputting text in the compiled version of a pattern, bytes
3036 other than 32-126 are always shown as hex escapes.
3037
3038 (b) When it is outputting text that is a matched part of a subject string,
3039 it does the same, unless a different locale has been set for the match
3040 (using the /L modifier). In this case, it uses isprint() to decide.
3041
3042 4. Fixed a major bug that caused incorrect computation of the amount of memory
3043 required for a compiled pattern when options that changed within the
3044 pattern affected the logic of the preliminary scan that determines the
3045 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3046 that the computed length was too small. The symptoms of this bug were
3047 either the PCRE error "internal error: code overflow" from pcre_compile(),
3048 or a glibc crash with a message such as "pcretest: free(): invalid next
3049 size (fast)". Examples of patterns that provoked this bug (shown in
3050 pcretest format) are:
3051
3052 /(?-x: )/x
3053 /(?x)(?-x: \s*#\s*)/
3054 /((?i)[\x{c0}])/8
3055 /(?i:[\x{c0}])/8
3056
3057 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3058 is now done differently.
3059
3060 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3061 wrapper classes; (b) implement a new function in the C++ scanner that is
3062 more efficient than the old way of doing things because it avoids levels of
3063 recursion in the regex matching; (c) add a paragraph to the documentation
3064 for the FullMatch() function.
3065
3066 6. The escape sequence \n was being treated as whatever was defined as
3067 "newline". Not only was this contrary to the documentation, which states
3068 that \n is character 10 (hex 0A), but it also went horribly wrong when
3069 "newline" was defined as CRLF. This has been fixed.
3070
3071 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3072 was being set to -1 for the "end of line" case (supposedly a value that no
3073 character can have). Though this value is never used (the check for end of
3074 line is "zero bytes in current character"), it caused compiler complaints.
3075 I've changed it to 0xffffffff.
3076
3077 8. In pcre_version.c, the version string was being built by a sequence of
3078 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3079 string (as it is for production releases) called a macro with an empty
3080 argument. The C standard says the result of this is undefined. The gcc
3081 compiler treats it as an empty string (which was what was wanted) but it is
3082 reported that Visual C gives an error. The source has been hacked around to
3083 avoid this problem.
3084
3085 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3086 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3087 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3088 of them did).
3089
3090 10. Originally, pcretest opened its input and output without "b"; then I was
3091 told that "b" was needed in some environments, so it was added for release
3092 5.0 to both the input and output. (It makes no difference on Unix-like
3093 systems.) Later I was told that it is wrong for the input on Windows. I've
3094 now abstracted the modes into two macros, to make it easier to fiddle with
3095 them, and removed "b" from the input mode under Windows.
3096
3097 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3098
3099 12. Added -help and --help to pcretest as an official way of being reminded
3100 of the options.
3101
3102 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3103 and pcrecpp.cc because they annoy compilers at high warning levels.
3104
3105 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3106
3107 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3108 scripts are not C programs :-) and which was not noticed because it works
3109 on Linux.
3110
3111 16. pcretest is supposed to handle any length of pattern and data line (as one
3112 line or as a continued sequence of lines) by extending its input buffer if
3113 necessary. This feature was broken for very long pattern lines, leading to
3114 a string of junk being passed to pcre_compile() if the pattern was longer
3115 than about 50K.
3116
3117 17. I have done a major re-factoring of the way pcre_compile() computes the
3118 amount of memory needed for a compiled pattern. Previously, there was code
3119 that made a preliminary scan of the pattern in order to do this. That was
3120 OK when PCRE was new, but as the facilities have expanded, it has become
3121 harder and harder to keep it in step with the real compile phase, and there
3122 have been a number of bugs (see for example, 4 above). I have now found a
3123 cunning way of running the real compile function in a "fake" mode that
3124 enables it to compute how much memory it would need, while actually only
3125 ever using a few hundred bytes of working memory and without too many
3126 tests of the mode. This should make future maintenance and development
3127 easier. A side effect of this work is that the limit of 200 on the nesting
3128 depth of parentheses has been removed (though this was never a serious
3129 limitation, I suspect). However, there is a downside: pcre_compile() now
3130 runs more slowly than before (30% or more, depending on the pattern). I
3131 hope this isn't a big issue. There is no effect on runtime performance.
3132
3133 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3134 newline (only possible for the last line of a file) and it was a
3135 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3136
3137 19. Added additional timing features to pcretest. (1) The -tm option now times
3138 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3139 separate command line item, by a number that specifies the number of
3140 repeats to use when timing. The default is 50000; this gives better
3141 precision, but takes uncomfortably long for very large patterns.
3142
3143 20. Extended pcre_study() to be more clever in cases where a branch of a
3144 subpattern has no definite first character. For example, (a*|b*)[cd] would
3145 previously give no result from pcre_study(). Now it recognizes that the
3146 first character must be a, b, c, or d.
3147
3148 21. There was an incorrect error "recursive call could loop indefinitely" if
3149 a subpattern (or the entire pattern) that was being tested for matching an
3150 empty string contained only one non-empty item after a nested subpattern.
3151 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3152 incorrectly, because the \d was being skipped in the check.
3153
3154 22. The pcretest program now has a new pattern option /B and a command line
3155 option -b, which is equivalent to adding /B to every pattern. This causes
3156 it to show the compiled bytecode, without the additional information that
3157 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3158 is the same as /B/I).
3159
3160 23. A new optimization is now able automatically to treat some sequences such
3161 as a*b as a*+b. More specifically, if something simple (such as a character
3162 or a simple class like \d) has an unlimited quantifier, and is followed by
3163 something that cannot possibly match the quantified thing, the quantifier
3164 is automatically "possessified".
3165
3166 24. A recursive reference to a subpattern whose number was greater than 39
3167 went wrong under certain circumstances in UTF-8 mode. This bug could also
3168 have affected the operation of pcre_study().
3169
3170 25. Realized that a little bit of performance could be had by replacing
3171 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3172
3173 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3174
3175 27. Possessive quantifiers such as a++ were previously implemented by turning
3176 them into atomic groups such as ($>a+). Now they have their own opcodes,
3177 which improves performance. This includes the automatically created ones
3178 from 23 above.
3179
3180 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3181 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3182 the first matched character to be a colon. This applied both to named and
3183 numbered groups.
3184
3185 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3186
3187 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3188 building PCRE on Windows easier, so I have included it in the distribution.
3189
3190 31. There is now a check in pcretest against a ridiculously large number being
3191 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3192 loop, the loop is abandoned.
3193
3194 32. Forward references to subpatterns in conditions such as (?(2)...) where
3195 subpattern 2 is defined later cause pcre_compile() to search forwards in
3196 the pattern for the relevant set of parentheses. This search went wrong
3197 when there were unescaped parentheses in a character class, parentheses
3198 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3199
3200 33. "Subroutine" calls and backreferences were previously restricted to
3201 referencing subpatterns earlier in the regex. This restriction has now
3202 been removed.
3203
3204 34. Added a number of extra features that are going to be in Perl 5.10. On the
3205 whole, these are just syntactic alternatives for features that PCRE had
3206 previously implemented using the Python syntax or my own invention. The
3207 other formats are all retained for compatibility.
3208
3209 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3210 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3211 also .NET compatible.
3212
3213 (b) A recursion or subroutine call to a named group can now be defined as
3214 (?&name) as well as (?P>name).
3215
3216 (c) A backreference to a named group can now be defined as \k<name> or
3217 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3218 5.10, are also .NET compatible.
3219
3220 (d) A conditional reference to a named group can now use the syntax
3221 (?(<name>) or (?('name') as well as (?(name).
3222
3223 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3224 groups (named and numbered) that are never evaluated inline, but can be
3225 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3226 is always false. There may be only one alternative in such a group.
3227
3228 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3229 as the simple (?(R). The condition is true only if the most recent
3230 recursion is that of the given number or name. It does not search out
3231 through the entire recursion stack.
3232
3233 (g) The escape \gN or \g{N} has been added, where N is a positive or
3234 negative number, specifying an absolute or relative reference.
3235
3236 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3237 some "unreachable code" warnings.
3238
3239 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3240 things, this adds five new scripts.
3241
3242 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3243 There were also incompatibilities regarding the handling of \Q..\E inside
3244 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3245 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3246
3247 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3248 matches an empty string, and forcibly breaks the loop. There were bugs in
3249 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3250 against aaaa the result was just "a" rather than "aaaa", for example. Two
3251 separate and independent bugs (that affected different cases) have been
3252 fixed.
3253
3254 39. Refactored the code to abolish the use of different opcodes for small
3255 capturing bracket numbers. This is a tidy that I avoided doing when I
3256 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3257 The new approach is not only tidier, it makes it possible to reduce the
3258 memory needed to fix the previous bug (38).
3259
3260 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3261 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3262 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3263 mode.
3264
3265 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3266 report.
3267
3268 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3269 copy construction and assignment in the C++ wrapper.
3270
3271 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3272 couple of bugs that could have given wrong results in the "--newline=crlf"
3273 case.
3274
3275 44. Added a number of casts and did some reorganization of signed/unsigned int
3276 variables following suggestions from Dair Grant. Also renamed the variable
3277 "this" as "item" because it is a C++ keyword.
3278
3279 45. Arranged for dftables to add
3280
3281 #include "pcre_internal.h"
3282
3283 to pcre_chartables.c because without it, gcc 4.x may remove the array
3284 definition from the final binary if PCRE is built into a static library and
3285 dead code stripping is activated.
3286
3287 46. For an unanchored pattern, if a match attempt fails at the start of a
3288 newline sequence, and the newline setting is CRLF or ANY, and the next two
3289 characters are CRLF, advance by two characters instead of one.
3290
3291
3292 Version 6.7 04-Jul-06
3293 ---------------------
3294
3295 1. In order to handle tests when input lines are enormously long, pcretest has
3296 been re-factored so that it automatically extends its buffers when
3297 necessary. The code is crude, but this _is_ just a test program. The
3298 default size has been increased from 32K to 50K.
3299
3300 2. The code in pcre_study() was using the value of the re argument before
3301 testing it for NULL. (Of course, in any sensible call of the function, it
3302 won't be NULL.)
3303
3304 3. The memmove() emulation function in pcre_internal.h, which is used on
3305 systems that lack both memmove() and bcopy() - that is, hardly ever -
3306 was missing a "static" storage class specifier.
3307
3308 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3309 containing an extended class (one that cannot be represented by a bitmap
3310 because it contains high-valued characters or Unicode property items, e.g.
3311 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3312 pattern, but PCRE should not loop if you do not (it no longer does).
3313 [Detail: two cases were found: (a) a repeated subpattern containing an
3314 extended class; (b) a recursive reference to a subpattern that followed a
3315 previous extended class. It wasn't skipping over the extended class
3316 correctly when UTF-8 mode was not set.]
3317
3318 5. A negated single-character class was not being recognized as fixed-length
3319 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3320 compile error "lookbehind assertion is not fixed length".
3321
3322 6. The RunPerlTest auxiliary script was showing an unexpected difference
3323 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3324 write a Perl script that can interpret lines of an input file either as
3325 byte characters or as UTF-8, which is what "perltest" was being required to
3326 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3327 can't do is switch easily at run time between having the "use utf8;" pragma
3328 or not. In the end, I fudged it by using the RunPerlTest script to insert
3329 "use utf8;" explicitly for the UTF-8 tests.
3330
3331 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3332 the end of the subject string, contrary to the documentation and to what
3333 Perl does. This was true of both matching functions. Now it matches only at
3334 the start of the subject and immediately after *internal* newlines.
3335
3336 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3337 a pointer to an int instead of a pointer to an unsigned long int. This
3338 caused problems on 64-bit systems.
3339
3340 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3341 instance of the 'standard' template library not being so standard".
3342
3343 10. There was no check on the number of named subpatterns nor the maximum
3344 length of a subpattern name. The product of these values is used to compute
3345 the size of the memory block for a compiled pattern. By supplying a very
3346 long subpattern name and a large number of named subpatterns, the size
3347 computation could be caused to overflow. This is now prevented by limiting
3348 the length of names to 32 characters, and the number of named subpatterns
3349 to 10,000.
3350
3351 11. Subpatterns that are repeated with specific counts have to be replicated in
3352 the compiled pattern. The size of memory for this was computed from the
3353 length of the subpattern and the repeat count. The latter is limited to
3354 65535, but there was no limit on the former, meaning that integer overflow
3355 could in principle occur. The compiled length of a repeated subpattern is
3356 now limited to 30,000 bytes in order to prevent this.
3357
3358 12. Added the optional facility to have named substrings with the same name.
3359
3360 13. Added the ability to use a named substring as a condition, using the
3361 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3362 are numbers (not recommended). Forward references are permitted.
3363
3364 14. Added forward references in named backreferences (if you see what I mean).
3365
3366 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3367 pattern could run off the end of the subject. For example, the pattern
3368 "(?s)(.{1,5})"8 did this with the subject "ab".
3369
3370 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3371 PCRE_CASELESS was set when matching characters that were quantified with ?
3372 or *.
3373
3374 17. A character class other than a single negated character that had a minimum
3375 but no maximum quantifier - for example [ab]{6,} - was not handled
3376 correctly by pce_dfa_exec(). It would match only one character.
3377
3378 18. A valid (though odd) pattern that looked like a POSIX character
3379 class but used an invalid character after [ (for example [[,abc,]]) caused
3380 pcre_compile() to give the error "Failed: internal error: code overflow" or
3381 in some cases to crash with a glibc free() error. This could even happen if
3382 the pattern terminated after [[ but there just happened to be a sequence of
3383 letters, a binary zero, and a closing ] in the memory that followed.
3384
3385 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3386 over the years. Originally (before any Unicode support), just the bottom 8
3387 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3388 output from "man perlunicode" includes this:
3389
3390 The regular expression compiler produces polymorphic opcodes. That
3391 is, the pattern adapts to the data and automatically switches to
3392 the Unicode character scheme when presented with Unicode data--or
3393 instead uses a traditional byte scheme when presented with byte
3394 data.
3395
3396 Sadly, a wide octal escape does not cause a switch, and in a string with
3397 no other multibyte characters, these octal escapes are treated as before.
3398 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3399 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3400 Unicode string.
3401
3402 I have not perpetrated such confusion in PCRE. Up till now, it took just
3403 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3404 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3405 translate to the appropriate multibyte character.
3406
3407 29. Applied some refactoring to reduce the number of warnings from Microsoft
3408 and Borland compilers. This has included removing the fudge introduced
3409 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3410 a warning about an unused variable.
3411
3412 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3413 characters since release 4.0, because Perl (from release 5.004) does not.
3414 [Or at least, is documented not to: some releases seem to be in conflict
3415 with the documentation.] However, when a pattern was studied with
3416 pcre_study() and all its branches started with \s, PCRE still included VT
3417 as a possible starting character. Of course, this did no harm; it just
3418 caused an unnecessary match attempt.
3419
3420 22. Removed a now-redundant internal flag bit that recorded the fact that case
3421 dependency changed within the pattern. This was once needed for "required
3422 byte" processing, but is no longer used. This recovers a now-scarce options
3423 bit. Also moved the least significant internal flag bit to the most-
3424 significant bit of the word, which was not previously used (hangover from
3425 the days when it was an int rather than a uint) to free up another bit for
3426 the future.
3427
3428 23. Added support for CRLF line endings as well as CR and LF. As well as the
3429 default being selectable at build time, it can now be changed at runtime
3430 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3431 specify that it is scanning data with non-default line endings.
3432
3433 24. Changed the definition of CXXLINK to make it agree with the definition of
3434 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3435
3436 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3437 recursions. This makes a big different to stack usage for some patterns.
3438
3439 26. If a subpattern containing a named recursion or subroutine reference such
3440 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3441 the space required for the compiled pattern went wrong and gave too small a
3442 value. Depending on the environment, this could lead to "Failed: internal
3443 error: code overflow at offset 49" or "glibc detected double free or
3444 corruption" errors.
3445
3446 27. Applied patches from Google (a) to support the new newline modes and (b) to
3447 advance over multibyte UTF-8 characters in GlobalReplace.
3448
3449 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3450 difference for some implementation of PCRE in some Windows version.
3451
3452 29. Added some extra testing facilities to pcretest:
3453
3454 \q<number> in a data line sets the "match limit" value
3455 \Q<number> in a data line sets the "match recursion limt" value
3456 -S <number> sets the stack size, where <number> is in megabytes
3457
3458 The -S option isn't available for Windows.
3459
3460
3461 Version 6.6 06-Feb-06
3462 ---------------------
3463
3464 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3465 in pcreposix.h. I have copied the definition from pcre.h.
3466
3467 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3468 because pcre.h is no longer a built file.
3469
3470 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3471 not normally included in the compiled code.
3472
3473
3474 Version 6.5 01-Feb-06
3475 ---------------------
3476
3477 1. When using the partial match feature with pcre_dfa_exec(), it was not
3478 anchoring the second and subsequent partial matches at the new starting
3479 point. This could lead to incorrect results. For example, with the pattern
3480 /1234/, partially matching against "123" and then "a4" gave a match.
3481
3482 2. Changes to pcregrep:
3483
3484 (a) All non-match returns from pcre_exec() were being treated as failures
3485 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3486 error message is output. Some extra information is given for the
3487 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3488 probably the only errors that are likely to be caused by users (by
3489 specifying a regex that has nested indefinite repeats, for instance).
3490 If there are more than 20 of these errors, pcregrep is abandoned.
3491
3492 (b) A binary zero was treated as data while matching, but terminated the
3493 output line if it was written out. This has been fixed: binary zeroes
3494 are now no different to any other data bytes.
3495
3496 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3497 used to set a locale for matching. The --locale=xxxx long option has
3498 been added (no short equivalent) to specify a locale explicitly on the
3499 pcregrep command, overriding the environment variables.
3500
3501 (d) When -B was used with -n, some line numbers in the output were one less
3502 than they should have been.
3503
3504 (e) Added the -o (--only-matching) option.
3505
3506 (f) If -A or -C was used with -c (count only), some lines of context were
3507 accidentally printed for the final match.
3508
3509 (g) Added the -H (--with-filename) option.
3510
3511 (h) The combination of options -rh failed to suppress file names for files
3512 that were found from directory arguments.
3513
3514 (i) Added the -D (--devices) and -d (--directories) options.
3515
3516 (j) Added the -F (--fixed-strings) option.
3517
3518 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3519
3520 (l) Added the --colo(u)r option.
3521
3522 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3523 is not present by default.
3524
3525 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3526 items such as (?R) or (?1), when the recursion could match a number of
3527 alternatives. If it matched one of the alternatives, but subsequently,
3528 outside the recursion, there was a failure, the code tried to back up into
3529 the recursion. However, because of the way PCRE is implemented, this is not
3530 possible, and the result was an incorrect result from the match.
3531
3532 In order to prevent this happening, the specification of recursion has
3533 been changed so that all such subpatterns are automatically treated as
3534 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3535
3536 4. I had overlooked the fact that, in some locales, there are characters for
3537 which isalpha() is true but neither isupper() nor islower() are true. In
3538 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3539 and ordfeminine) are like this. This affected the treatment of \w and \W
3540 when they appeared in character classes, but not when they appeared outside
3541 a character class. The bit map for "word" characters is now created
3542 separately from the results of isalnum() instead of just taking it from the
3543 upper, lower, and digit maps. (Plus the underscore character, of course.)
3544
3545 5. The above bug also affected the handling of POSIX character classes such as
3546 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3547 permanent tables. Instead, the bit maps for such a class were previously
3548 created as the appropriate unions of the upper, lower, and digit bitmaps.
3549 Now they are created by subtraction from the [[:word:]] class, which has
3550 its own bitmap.
3551
3552 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3553 It is created by subtracting the vertical space characters (\x09, \x0a,
3554 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3555 subtraction was done in the overall bitmap for a character class, meaning
3556 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3557 be recognized. This bug has been fixed.
3558
3559 7. Patches from the folks at Google:
3560
3561 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3562 real life, but is still worth protecting against".
3563
3564 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3565 regular expressions".
3566
3567 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3568 have it.
3569
3570 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3571 "configure" and the latter not, in order to fix a problem somebody had
3572 with compiling the Arg class on HP-UX.
3573
3574 (e) Improve the error-handling of the C++ wrapper a little bit.
3575
3576 (f) New tests for checking recursion limiting.
3577
3578 8. The pcre_memmove() function, which is used only if the environment does not
3579 have a standard memmove() function (and is therefore rarely compiled),
3580 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3581 returning a result (though PCRE never actually uses the result).
3582
3583 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3584 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3585 returned instead of calling malloc() with an overflowing number that would
3586 most likely cause subsequent chaos.
3587
3588 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3589
3590 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3591 with this option is matched, the nmatch and pmatch options of regexec() are
3592 ignored.
3593
3594 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3595 provided in case anyone wants to the the POSIX interface with UTF-8
3596 strings.
3597
3598 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3599 C++ linking (needed for some HP-UX environments).
3600
3601 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3602 (unused parameter) and in the pcre_printint() function (omitted "default"
3603 switch label when the default is to do nothing).
3604
3605 15. Added some code to make it possible, when PCRE is compiled as a C++
3606 library, to replace subject pointers for pcre_exec() with a smart pointer
3607 class, thus making it possible to process discontinuous strings.
3608
3609 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3610 much the same function. They were added by different people who were trying
3611 to make PCRE easy to compile on non-Unix systems. It has been suggested
3612 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3613 for compiling on Windows systems. I have therefore replaced it with
3614 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3615 defaults to "extern" for C or "extern C" for C++, which works fine on
3616 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3617 SCOPE with something explicit in config.h. In addition:
3618
3619 (a) pcreposix.h still had just "extern" instead of either of these macros;
3620 I have replaced it with PCRE_DATA_SCOPE.
3621
3622 (b) Functions such as _pcre_xclass(), which are internal to the library,
3623 but external in the C sense, all had PCRE_EXPORT in their definitions.
3624 This is apparently wrong for the Windows case, so I have removed it.
3625 (It makes no difference on Unix-like systems.)
3626
3627 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3628 of recursive calls to match(). This is different to MATCH_LIMIT because
3629 that limits the total number of calls to match(), not all of which increase
3630 the depth of recursion. Limiting the recursion depth limits the amount of
3631 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3632 when PCRE is compiled, and changed at run time. A patch from Google adds
3633 this functionality to the C++ interface.
3634
3635 18. Changes to the handling of Unicode character properties:
3636
3637 (a) Updated the table to Unicode 4.1.0.
3638
3639 (b) Recognize characters that are not in the table as "Cn" (undefined).
3640
3641 (c) I revised the way the table is implemented to a much improved format
3642 which includes recognition of ranges. It now supports the ranges that
3643 are defined in UnicodeData.txt, and it also amalgamates other
3644 characters into ranges. This has reduced the number of entries in the
3645 table from around 16,000 to around 3,000, thus reducing its size
3646 considerably. I realized I did not need to use a tree structure after
3647 all - a binary chop search is just as efficient. Having reduced the
3648 number of entries, I extended their size from 6 bytes to 8 bytes to
3649 allow for more data.
3650
3651 (d) Added support for Unicode script names via properties such as \p{Han}.
3652
3653 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3654 matching that character.
3655
3656 20. When matching a repeated Unicode property with a minimum greater than zero,
3657 (for example \pL{2,}), PCRE could look past the end of the subject if it
3658 reached it while seeking the minimum number of characters. This could
3659 happen only if some of the characters were more than one byte long, because
3660 there is a check for at least the minimum number of bytes.
3661
3662 21. Refactored the implementation of \p and \P so as to be more general, to
3663 allow for more different types of property in future. This has changed the
3664 compiled form incompatibly. Anybody with saved compiled patterns that use
3665 \p or \P will have to recompile them.
3666
3667 22. Added "Any" and "L&" to the supported property types.
3668
3669 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3670 but give a compile time error if the value is greater than 0xff.
3671
3672 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3673 accidentally not being installed or uninstalled.
3674
3675 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3676 made were to insert the current release number. This seemed silly, because
3677 it made things harder for people building PCRE on systems that don't run
3678 "configure". I have turned pcre.h into a distributed file, no longer built
3679 by "configure", with the version identification directly included. There is
3680 no longer a pcre.h.in file.
3681
3682 However, this change necessitated a change to the pcre-config script as
3683 well. It is built from pcre-config.in, and one of the substitutions was the
3684 release number. I have updated configure.ac so that ./configure now finds
3685 the release number by grepping pcre.h.
3686
3687 26. Added the ability to run the tests under valgrind.
3688
3689
3690 Version 6.4 05-Sep-05
3691 ---------------------
3692
3693 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3694 "--" to be printed when multiple files were scanned, even when none of the
3695 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3696 consider it to be a bug, and have restored the previous behaviour.
3697
3698 2. A couple of code tidies to get rid of compiler warnings.
3699
3700 3. The pcretest program used to cheat by referring to symbols in the library
3701 whose names begin with _pcre_. These are internal symbols that are not
3702 really supposed to be visible externally, and in some environments it is
3703 possible to suppress them. The cheating is now confined to including
3704 certain files from the library's source, which is a bit cleaner.
3705
3706 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3707 file's purpose clearer.
3708
3709 5. Reorganized pcre_ucp_findchar().
3710
3711
3712 Version 6.3 15-Aug-05
3713 ---------------------
3714
3715 1. The file libpcre.pc.in did not have general read permission in the tarball.
3716
3717 2. There were some problems when building without C++ support:
3718
3719 (a) If C++ support was not built, "make install" and "make test" still
3720 tried to test it.
3721
3722 (b) There were problems when the value of CXX was explicitly set. Some
3723 changes have been made to try to fix these, and ...
3724
3725 (c) --disable-cpp can now be used to explicitly disable C++ support.
3726
3727 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3728 backslash in a target when C++ was disabled. This confuses some
3729 versions of "make", apparently. Using an intermediate variable solves
3730 this. (Same for CPP_LOBJ.)
3731
3732 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3733 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3734 necessary on certain architectures.
3735
3736 4. Added a setting of -export-symbols-regex to the link command to remove
3737 those symbols that are exported in the C sense, but actually are local
3738 within the library, and not documented. Their names all begin with
3739 "_pcre_". This is not a perfect job, because (a) we have to except some
3740 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3741 available (and never for static libraries). I have made a note to try to
3742 find a way round (a) in the future.
3743
3744
3745 Version 6.2 01-Aug-05
3746 ---------------------
3747
3748 1. There was no test for integer overflow of quantifier values. A construction
3749 such as {1111111111111111} would give undefined results. What is worse, if
3750 a minimum quantifier for a parenthesized subpattern overflowed and became
3751 negative, the calculation of the memory size went wrong. This could have
3752 led to memory overwriting.
3753
3754 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3755
3756 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3757 operating environments where this matters.
3758
3759 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3760 PCRE options from within the C++ wrapper.
3761
3762 5. Named capturing subpatterns were not being correctly counted when a pattern
3763 was compiled. This caused two problems: (a) If there were more than 100
3764 such subpatterns, the calculation of the memory needed for the whole
3765 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3766 back references of the form \12, where the number was greater than 9, were
3767 not recognized as back references, even though there were sufficient
3768 previous subpatterns.
3769
3770 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3771 versions of gcc, e.g. 2.95.4.
3772
3773
3774 Version 6.1 21-Jun-05
3775 ---------------------
3776
3777 1. There was one reference to the variable "posix" in pcretest.c that was not
3778 surrounded by "#if !defined NOPOSIX".
3779
3780 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3781 the cross-check on the old pcre_info() function, for the benefit of the
3782 cut-down version of PCRE that is currently imported into Exim.
3783
3784 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3785 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3786 patterns that start (?i) but I don't think that matters. The use of (?i) is
3787 just an example; this all applies to the other options as well.
3788
3789 4. Since libtool seems to echo the compile commands it is issuing, the output
3790 from "make" can be reduced a bit by putting "@" in front of each libtool
3791 compile command.
3792
3793 5. Patch from the folks at Google for configure.in to be a bit more thorough
3794 in checking for a suitable C++ installation before trying to compile the
3795 C++ stuff. This should fix a reported problem when a compiler was present,
3796 but no suitable headers.
3797
3798 6. The man pages all had just "PCRE" as their title. I have changed them to
3799 be the relevant file name. I have also arranged that these names are
3800 retained in the file doc/pcre.txt, which is a concatenation in text format
3801 of all the man pages except the little individual ones for each function.
3802
3803 7. The NON-UNIX-USE file had not been updated for the different set of source
3804 files that come with release 6. I also added a few comments about the C++
3805 wrapper.
3806
3807
3808 Version 6.0 07-Jun-05
3809 ---------------------
3810
3811 1. Some minor internal re-organization to help with my DFA experiments.
3812
3813 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3814 didn't matter for the library itself when fully configured, but did matter
3815 when compiling without UCP support, or within Exim, where the ucp files are
3816 not imported.
3817
3818 3. Refactoring of the library code to split up the various functions into
3819 different source modules. The addition of the new DFA matching code (see
3820 below) to a single monolithic source would have made it really too
3821 unwieldy, quite apart from causing all the code to be include in a
3822 statically linked application, when only some functions are used. This is
3823 relevant even without the DFA addition now that patterns can be compiled in
3824 one application and matched in another.
3825
3826 The downside of splitting up is that there have to be some external
3827 functions and data tables that are used internally in different modules of
3828 the library but which are not part of the API. These have all had their
3829 names changed to start with "_pcre_" so that they are unlikely to clash
3830 with other external names.
3831
3832 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3833 a different (DFA) algorithm. Although it is slower than the original
3834 function, it does have some advantages for certain types of matching
3835 problem.
3836
3837 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3838 including restarting after a partial match.
3839
3840 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3841 defined when compiling for Windows was sent to me. I have put it into the
3842 code, though I have no means of testing or verifying it.
3843
3844 7. Added the pcre_refcount() auxiliary function.
3845
3846 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3847 match before or at the first newline in the subject string. In pcretest,
3848 the /f option on a pattern can be used to set this.
3849
3850 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3851 would behave wrongly. This has been present in PCRE since release 4.0.
3852
3853 10. A number of changes to the pcregrep command:
3854
3855 (a) Refactored how -x works; insert ^(...)$ instead of setting
3856 PCRE_ANCHORED and checking the length, in preparation for adding
3857 something similar for -w.
3858
3859 (b) Added the -w (match as a word) option.
3860
3861 (c) Refactored the way lines are read and buffered so as to have more
3862 than one at a time available.
3863
3864 (d) Implemented a pcregrep test script.
3865
3866 (e) Added the -M (multiline match) option. This allows patterns to match
3867 over several lines of the subject. The buffering ensures that at least
3868 8K, or the rest of the document (whichever is the shorter) is available
3869 for matching (and similarly the previous 8K for lookbehind assertions).
3870
3871 (f) Changed the --help output so that it now says
3872
3873 -w, --word-regex(p)
3874
3875 instead of two lines, one with "regex" and the other with "regexp"
3876 because that confused at least one person since the short forms are the
3877 same. (This required a bit of code, as the output is generated
3878 automatically from a table. It wasn't just a text change.)
3879
3880 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3881 option but starts with a hyphen. Could be a pattern or a path name
3882 starting with a hyphen, for instance.
3883
3884 (h) "-" can be given as a file name to represent stdin.
3885
3886 (i) When file names are being printed, "(standard input)" is used for
3887 the standard input, for compatibility with GNU grep. Previously
3888 "<stdin>" was used.
3889
3890 (j) The option --label=xxx can be used to supply a name to be used for
3891 stdin when file names are being printed. There is no short form.
3892
3893 (k) Re-factored the options decoding logic because we are going to add
3894 two more options that take data. Such options can now be given in four
3895 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3896
3897 (l) Added the -A, -B, and -C options for requesting that lines of context
3898 around matches be printed.
3899
3900 (m) Added the -L option to print the names of files that do not contain
3901 any matching lines, that is, the complement of -l.
3902
3903 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3904 continue to scan other files.
3905
3906 (o) The -s option was incorrectly implemented. For compatibility with other
3907 greps, it now suppresses the error message for a non-existent or non-
3908 accessible file (but not the return code). There is a new option called
3909 -q that suppresses the output of matching lines, which was what -s was
3910 previously doing.
3911
3912 (p) Added --include and --exclude options to specify files for inclusion
3913 and exclusion when recursing.
3914
3915 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3916 Hopefully, it now does.
3917
3918 12. Missing cast in pcre_study().
3919
3920 13. Added an "uninstall" target to the makefile.
3921
3922 14. Replaced "extern" in the function prototypes in Makefile.in with
3923 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3924 world, but is set differently for Windows.
3925
3926 15. Added a second compiling function called pcre_compile2(). The only
3927 difference is that it has an extra argument, which is a pointer to an
3928 integer error code. When there is a compile-time failure, this is set
3929 non-zero, in addition to the error test pointer being set to point to an
3930 error message. The new argument may be NULL if no error number is required
3931 (but then you may as well call pcre_compile(), which is now just a
3932 wrapper). This facility is provided because some applications need a
3933 numeric error indication, but it has also enabled me to tidy up the way
3934 compile-time errors are handled in the POSIX wrapper.
3935
3936 16. Added VPATH=.libs to the makefile; this should help when building with one
3937 prefix path and installing with another. (Or so I'm told by someone who
3938 knows more about this stuff than I do.)
3939
3940 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3941 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3942 match everything, including newlines. This is not POSIX-compatible, but
3943 somebody wanted the feature. From pcretest it can be activated by using
3944 both the P and the s flags.
3945
3946 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3947
3948 19. libpcre.pc was being incorrectly installed as executable.
3949
3950 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3951 it now also looks for '\r' so that it will work unmodified on Windows.
3952
3953 21. Added Google's contributed C++ wrapper to the distribution.
3954
3955 22. Added some untidy missing memory free() calls in pcretest, to keep
3956 Electric Fence happy when testing.
3957
3958
3959
3960 Version 5.0 13-Sep-04
3961 ---------------------
3962
3963 1. Internal change: literal characters are no longer packed up into items
3964 containing multiple characters in a single byte-string. Each character
3965 is now matched using a separate opcode. However, there may be more than one
3966 byte in the character in UTF-8 mode.
3967
3968 2. The pcre_callout_block structure has two new fields: pattern_position and
3969 next_item_length. These contain the offset in the pattern to the next match
3970 item, and its length, respectively.
3971
3972 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3973 insertion of callouts before each pattern item. Added the /C option to
3974 pcretest to make use of this.
3975
3976 4. On the advice of a Windows user, the lines
3977
3978 #if defined(_WIN32) || defined(WIN32)
3979 _setmode( _fileno( stdout ), 0x8000 );
3980 #endif /* defined(_WIN32) || defined(WIN32) */
3981
3982 have been added to the source of pcretest. This apparently does useful
3983 magic in relation to line terminators.
3984
3985 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3986 for the benefit of those environments where the "b" makes a difference.
3987
3988 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3989 to know about it. I have put a hack into configure.in that adds in code
3990 to set GCC=yes if CC=icc. This seems to end up at a point in the
3991 generated configure script that is early enough to affect the setting of
3992 compiler options, which is what is needed, but I have no means of testing
3993 whether it really works. (The user who reported this had patched the
3994 generated configure script, which of course I cannot do.)
3995
3996 LATER: After change 22 below (new libtool files), the configure script
3997 seems to know about icc (and also ecc). Therefore, I have commented out
3998 this hack in configure.in.
3999
4000 7. Added support for pkg-config (2 patches were sent in).
4001
4002 8. Negated POSIX character classes that used a combination of internal tables
4003 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4004 [[:^ascii]]. Typically, they would match almost any characters. The other
4005 POSIX classes were not broken in this way.
4006
4007 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4008 to find the match, as PCRE was deluded into thinking that the match had to
4009 start at the start point or following a newline. The same bug applied to
4010 patterns with negative forward assertions or any backward assertions
4011 preceding ".*" at the start, unless the pattern required a fixed first
4012 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4013
4014 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4015 starting at the last subject character, bytes beyond the end of the subject
4016 string were read.
4017
4018 11. Renamed the variable "class" as "classbits" to make life easier for C++
4019 users. (Previously there was a macro definition, but it apparently wasn't
4020 enough.)
4021
4022 12. Added the new field "tables" to the extra data so that tables can be passed
4023 in at exec time, or the internal tables can be re-selected. This allows
4024 a compiled regex to be saved and re-used at a later time by a different
4025 program that might have everything at different addresses.
4026
4027 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4028 -R library as well as a -L library.
4029
4030 14. The debugging options of pcretest (-d on the command line or D on a
4031 pattern) showed incorrect output for anything following an extended class
4032 that contained multibyte characters and which was followed by a quantifier.
4033
4034 15. Added optional support for general category Unicode character properties
4035 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4036 support. It adds about 90K to the size of the library. The meanings of the
4037 inbuilt class escapes such as \d and \s have NOT been changed.
4038
4039 16. Updated pcredemo.c to include calls to free() to release the memory for the
4040 compiled pattern.
4041
4042 17. The generated file chartables.c was being created in the source directory
4043 instead of in the building directory. This caused the build to fail if the
4044 source directory was different from the building directory, and was
4045 read-only.
4046
4047 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4048 file. No doubt somebody will tell me if they don't make sense... Also added
4049 Dan Mooney's comments about building on OpenVMS.
4050
4051 19. Added support for partial matching via the PCRE_PARTIAL option for
4052 pcre_exec() and the \P data escape in pcretest.
4053
4054 20. Extended pcretest with 3 new pattern features:
4055
4056 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4057 write the compiled pattern to the file whose name is "rest-of-line".
4058 This is a straight binary dump of the data, with the saved pointer to
4059 the character tables forced to be NULL. The study data, if any, is
4060 written too. After writing, pcretest reads a new pattern.
4061
4062 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4063 compiled pattern from the given file. There must not be any
4064 occurrences of "<" in the file name (pretty unlikely); if there are,
4065 pcretest will instead treat the initial "<" as a pattern delimiter.
4066 After reading in the pattern, pcretest goes on to read data lines as
4067 usual.
4068
4069 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4070 and 16-bit fields in a compiled pattern, to simulate a pattern that
4071 was compiled on a host of opposite endianness.
4072
4073 21. The pcre-exec() function can now cope with patterns that were compiled on
4074 hosts of opposite endianness, with this restriction:
4075
4076 As for any compiled expression that is saved and used later, the tables
4077 pointer field cannot be preserved; the extra_data field in the arguments
4078 to pcre_exec() should be used to pass in a tables address if a value
4079 other than the default internal tables were used at compile time.
4080
4081 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4082 now diagnosed as an error. Previously, most of the time, a negative number
4083 would have been treated as zero, but if in addition "ovector" was passed as
4084 NULL, a crash could occur.
4085
4086 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4087 new versions from the libtool 1.5 distribution (the last one is a copy of
4088 a file called libtool.m4). This seems to have fixed the need to patch
4089 "configure" to support Darwin 1.3 (which I used to do). However, I still
4090 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4091 workstation).
4092
4093 24. Changed the PCRE licence to be the more standard "BSD" licence.
4094
4095
4096 Version 4.5 01-Dec-03
4097 ---------------------
4098
4099 1. There has been some re-arrangement of the code for the match() function so
4100 that it can be compiled in a version that does not call itself recursively.
4101 Instead, it keeps those local variables that need separate instances for
4102 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4103 needs to "recurse". Keeping track of where control must go is done by means
4104 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4105 hide most of the details from the main code, and operates only if
4106 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4107 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4108 operating.
4109
4110 To make it easier for callers to provide specially tailored get/free
4111 functions for this usage, two new functions, pcre_stack_malloc, and
4112 pcre_stack_free, are used. They are always called in strict stacking order,
4113 and the size of block requested is always the same.
4114
4115 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4116 PCRE has been compiled to use the stack or the heap for recursion. The
4117 -C option of pcretest uses this to show which version is compiled.
4118
4119 A new data escape \S, is added to pcretest; it causes the amounts of store
4120 obtained and freed by both kinds of malloc/free at match time to be added
4121 to the output.
4122
4123 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4124 what's available on my current Linux desktop machine.
4125
4126 3. When matching a UTF-8 string, the test for a valid string at the start has
4127 been extended. If start_offset is not zero, PCRE now checks that it points
4128 to a byte that is the start of a UTF-8 character. If not, it returns
4129 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4130 this is necessary because there may be backward assertions in the pattern.
4131 When matching the same subject several times, it may save resources to use
4132 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4133
4134 4. The code for checking the validity of UTF-8 strings has been tightened so
4135 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4136 containing "overlong sequences".
4137
4138 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4139 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4140 should have been "&", but it just so happened that all the cases this let
4141 through by mistake were picked up later in the function.
4142
4143 6. I had used a variable called "isblank" - this is a C99 function, causing
4144 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4145
4146 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4147 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4148 so the version is shown; (c) stop "make test" echoing "./RunTest".
4149
4150 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4151
4152 9. The prototype for memmove() for systems that don't have it was using
4153 size_t, but the inclusion of the header that defines size_t was later. I've
4154 moved the #includes for the C headers earlier to avoid this.
4155
4156 10. Added some adjustments to the code to make it easier to compiler on certain
4157 special systems:
4158
4159 (a) Some "const" qualifiers were missing.
4160 (b) Added the macro EXPORT before all exported functions; by default this
4161 is defined to be empty.
4162 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4163 that it reads its output file name as an argument instead of writing
4164 to the standard output and assuming this can be redirected.
4165
4166 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4167 class containing characters with values greater than 255, PCRE compilation
4168 went into a loop.
4169
4170 12. A recursive reference to a subpattern that was within another subpattern
4171 that had a minimum quantifier of zero caused PCRE to crash. For example,
4172 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4173 recursion. If the recursively-called subpattern itself had a zero repeat,
4174 that was OK.
4175
4176 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4177 buffer into which it was copied (for escape processing) was still set at
4178 1024, so long lines caused crashes.
4179
4180 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4181 "internal error: code overflow...". This applied to any character class
4182 that was followed by a possessive quantifier.
4183
4184 15. Modified the Makefile to add libpcre.la as a prerequisite for
4185 libpcreposix.la because I was told this is needed for a parallel build to
4186 work.
4187
4188 16. If a pattern that contained .* following optional items at the start was
4189 studied, the wrong optimizing data was generated, leading to matching
4190 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4191 matching string must start with a or b or c. The correct conclusion for
4192 this pattern is that a match can start with any character.
4193
4194
4195 Version 4.4 13-Aug-03
4196 ---------------------
4197
4198 1. In UTF-8 mode, a character class containing characters with values between
4199 127 and 255 was not handled correctly if the compiled pattern was studied.
4200 In fixing this, I have also improved the studying algorithm for such
4201 classes (slightly).
4202
4203 2. Three internal functions had redundant arguments passed to them. Removal
4204 might give a very teeny performance improvement.
4205
4206 3. Documentation bug: the value of the capture_top field in a callout is *one
4207 more than* the number of the hightest numbered captured substring.
4208
4209 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4210 in incorrectly linking with a previously installed version. They now link
4211 explicitly with libpcre.la.
4212
4213 5. configure.in no longer needs to recognize Cygwin specially.
4214
4215 6. A problem in pcre.in for Windows platforms is fixed.
4216
4217 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4218 pcretest, it used to include the size of the study block as part of its
4219 output. Unfortunately, the structure contains a field that has a different
4220 size on different hardware architectures. This meant that the tests that
4221 showed this size failed. As the block is currently always of a fixed size,
4222 this information isn't actually particularly useful in pcretest output, so
4223 I have just removed it.
4224
4225 8. Three pre-processor statements accidentally did not start in column 1.
4226 Sadly, there are *still* compilers around that complain, even though
4227 standard C has not required this for well over a decade. Sigh.
4228
4229 9. In pcretest, the code for checking callouts passed small integers in the
4230 callout_data field, which is a void * field. However, some picky compilers
4231 complained about the casts involved for this on 64-bit systems. Now
4232 pcretest passes the address of the small integer instead, which should get
4233 rid of the warnings.
4234
4235 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4236 both compile and run time, and gives an error if an invalid UTF-8 sequence
4237 is found. There is a option for disabling this check in cases where the
4238 string is known to be correct and/or the maximum performance is wanted.
4239
4240 11. In response to a bug report, I changed one line in Makefile.in from
4241
4242 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4243 to
4244 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4245
4246 to look similar to other lines, but I have no way of telling whether this
4247 is the right thing to do, as I do not use Windows. No doubt I'll get told
4248 if it's wrong...
4249
4250
4251 Version 4.3 21-May-03
4252 ---------------------
4253
4254 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4255 Makefile.
4256
4257 2. Some refactoring to improve the quality of the code:
4258
4259 (i) The utf8_table... variables are now declared "const".
4260
4261 (ii) The code for \cx, which used the "case flipping" table to upper case
4262 lower case letters, now just substracts 32. This is ASCII-specific,
4263 but the whole concept of \cx is ASCII-specific, so it seems
4264 reasonable.
4265
4266 (iii) PCRE was using its character types table to recognize decimal and
4267 hexadecimal digits in the pattern. This is silly, because it handles
4268 only 0-9, a-f, and A-F, but the character types table is locale-
4269 specific, which means strange things might happen. A private
4270 table is now used for this - though it costs 256 bytes, a table is
4271 much faster than multiple explicit tests. Of course, the standard
4272 character types table is still used for matching digits in subject
4273 strings against \d.
4274
4275 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4276 ending in _t are). So I've renamed it as ESC_tee.
4277
4278 3. The first argument for regexec() in the POSIX wrapper should have been
4279 defined as "const".
4280
4281 4. Changed pcretest to use malloc() for its buffers so that they can be
4282 Electric Fenced for debugging.
4283
4284 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4285 to read one or more bytes before the start of the subject string. Often this
4286 had no effect on PCRE's behaviour, but in some circumstances it could
4287 provoke a segmentation fault.
4288
4289 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4290 to try to read one or more bytes before the start of the subject string.
4291
4292 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4293 UTF-8 support could misbehave in various ways if the subject string
4294 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4295 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4296 back over UTF-8 characters.)
4297
4298
4299 Version 4.2 14-Apr-03
4300 ---------------------
4301
4302 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4303
4304 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4305 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4306 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4307 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4308 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4309 and BUILD_EXEEXT
4310 Note: automatic setting of the BUILD variables is not yet working
4311 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4312 compile-time but not at link-time
4313 [LINK]: use for linking executables only
4314 make different versions for Windows and non-Windows
4315 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4316 libraries
4317 [LINK_FOR_BUILD]: new variable
4318 [OBJEXT]: use throughout
4319 [EXEEXT]: use throughout
4320 <winshared>: new target
4321 <wininstall>: new target
4322 <dftables.o>: use native compiler
4323 <dftables>: use native linker
4324 <install>: handle Windows platform correctly
4325 <clean>: ditto
4326 <check>: ditto
4327 copy DLL to top builddir before testing
4328
4329 As part of these changes, -no-undefined was removed again. This was reported
4330 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4331 in any case.
4332
4333 3. Some tidies to get rid of compiler warnings:
4334
4335 . In the match_data structure, match_limit was an unsigned long int, whereas
4336 match_call_count was an int. I've made them both unsigned long ints.
4337
4338 . In pcretest the fact that a const uschar * doesn't automatically cast to
4339 a void * provoked a warning.
4340
4341 . Turning on some more compiler warnings threw up some "shadow" variables
4342 and a few more missing casts.
4343
4344 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4345 option, a class that contained a single character with a value between 128
4346 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4347
4348 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4349 option, a class that contained several characters, but with at least one
4350 whose value was between 128 and 255 caused PCRE to crash.
4351
4352
4353 Version 4.1 12-Mar-03
4354 ---------------------
4355
4356 1. Compiling with gcc -pedantic found a couple of places where casts were
4357 needed, and a string in dftables.c that was longer than standard compilers are
4358 required to support.
4359
4360 2. Compiling with Sun's compiler found a few more places where the code could
4361 be tidied up in order to avoid warnings.
4362
4363 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4364 first of these names is deprecated in the latest Autoconf in favour of the name
4365 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4366 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4367 analogy I have changed it to CFLAGS_FOR_BUILD.
4368
4369 4. Added -no-undefined to the linking command in the Makefile, because this is
4370 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4371 linking step for the pcreposix library.
4372
4373 5. PCRE was failing to diagnose the case of two named groups with the same
4374 name.
4375
4376 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4377 literal character that is needed in the subject for a match, and scans along to
4378 ensure that it is present before embarking on the full matching process. This
4379 saves time in cases of nested unlimited repeats that are never going to match.
4380 Problem: the scan can take a lot of time if the subject is very long (e.g.
4381 megabytes), thus penalizing straightforward matches. It is now done only if the
4382 amount of subject to be scanned is less than 1000 bytes.
4383
4384 7. A lesser problem with the same optimization is that it was recording the
4385 first character of an anchored pattern as "needed", thus provoking a search
4386 right along the subject, even when the first match of the pattern was going to
4387 fail. The "needed" character is now not set for anchored patterns, unless it
4388 follows something in the pattern that is of non-fixed length. Thus, it still
4389 fulfils its original purpose of finding quick non-matches in cases of nested
4390 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4391
4392
4393 Version 4.0 17-Feb-03
4394 ---------------------
4395
4396 1. If a comment in an extended regex that started immediately after a meta-item
4397 extended to the end of string, PCRE compiled incorrect data. This could lead to
4398 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4399
4400 2. Moved to autoconf 2.53 and libtool 1.4.2.
4401
4402 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4403 the special perltest8 script is no longer needed - all the tests can be run
4404 from a single perltest script.
4405
4406 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4407 by \s. It has now been removed in PCRE. This means it isn't recognized as
4408 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4409 class [:space:] *does* include VT, thereby creating a mess.
4410
4411 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4412 space and tab.
4413
4414 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4415 its new features into the main test script, reducing the number of scripts.
4416
4417 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4418 were backward compatible, and made the (?i) apply to the whole pattern, as if
4419 /i were given. Now it behaves more logically, and applies the option setting
4420 only to what follows. PCRE has been changed to follow suit. However, if it
4421 finds options settings right at the start of the pattern, it extracts them into
4422 the global options, as before. Thus, they show up in the info data.
4423
4424 8. Added support for the \Q...\E escape sequence. Characters in between are
4425 treated as literals. This is slightly different from Perl in that $ and @ are
4426 also handled as literals inside the quotes. In Perl, they will cause variable
4427 interpolation. Note the following examples:
4428
4429 Pattern PCRE matches Perl matches
4430
4431 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4432 \Qabc\$xyz\E abc\$xyz abc\$xyz
4433 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4434
4435 For compatibility with Perl, \Q...\E sequences are recognized inside character
4436 classes as well as outside them.
4437
4438 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4439 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4440 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4441 signed/unsigned warnings.
4442
4443 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4444 option for pcretest, so I've replaced it by a simple function that does just
4445 that job.
4446
4447 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4448 "pcregrep -".
4449
4450 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4451 Java package. This provides some syntactic sugar for simple cases of what my
4452 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4453 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4454 item, you can use this simplified notation. Note that only makes sense with
4455 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4456 greediness, whatever the setting of the PCRE_UNGREEDY option.
4457
4458 13. A change of greediness default within a pattern was not taking effect at
4459 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4460 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4461 was abstracted outside.
4462
4463 14. PCRE now supports the \G assertion. It is true when the current matching
4464 position is at the start point of the match. This differs from \A when the
4465 starting offset is non-zero. Used with the /g option of pcretest (or similar
4466 code), it works in the same way as it does for Perl's /g option. If all
4467 alternatives of a regex begin with \G, the expression is anchored to the start
4468 match position, and the "anchored" flag is set in the compiled expression.
4469
4470 15. Some bugs concerning the handling of certain option changes within patterns
4471 have been fixed. These applied to options other than (?ims). For example,
4472 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4473 been the other way round. Some of this was related to change 7 above.
4474
4475 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4476 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4477 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4478 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4479
4480 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4481 mode. Unlike ".", it always matches newline, whatever the setting of
4482 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4483 assertions. Perl allows it, but it doesn't (in general) work because it can't
4484 calculate the length of the lookbehind. At least, that's the case for Perl
4485 5.8.0 - I've been told they are going to document that it doesn't work in
4486 future.
4487
4488 18. Added an error diagnosis for escapes that PCRE does not support: these are
4489 \L, \l, \N, \P, \p, \U, \u, and \X.
4490
4491 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4492 reading past the end of the pattern in cases such as /[abcd/.
4493
4494 20. PCRE was getting more memory than necessary for patterns with classes that
4495 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4496
4497 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4498 compiling PCRE for use with Virtual Pascal.
4499
4500 22. Small fix to the Makefile to make it work properly if the build is done
4501 outside the source tree.
4502
4503 23. Added a new extension: a condition to go with recursion. If a conditional
4504 subpattern starts with (?(R) the "true" branch is used if recursion has
4505 happened, whereas the "false" branch is used only at the top level.
4506
4507 24. When there was a very long string of literal characters (over 255 bytes
4508 without UTF support, over 250 bytes with UTF support), the computation of how
4509 much memory was required could be incorrect, leading to segfaults or other
4510 strange effects.
4511
4512 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4513 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4514 there was a subsequent back reference to those brackets. This meant that, for
4515 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4516 possible to check for precisely this case. All we can do is abandon the
4517 optimization if .* occurs inside capturing brackets when there are any back
4518 references whatsoever. (See below for a better fix that came later.)
4519
4520 26. The handling of the optimization for finding the first character of a
4521 non-anchored pattern, and for finding a character that is required later in the
4522 match were failing in some cases. This didn't break the matching; it just
4523 failed to optimize when it could. The way this is done has been re-implemented.
4524
4525 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4526
4527 28. Added a new feature that provides some of the functionality that Perl
4528 provides with (?{...}). The facility is termed a "callout". The way it is done
4529 in PCRE is for the caller to provide an optional function, by setting
4530 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4531 global variable. By default it is unset, which disables all calling out. To get
4532 the function called, the regex must include (?C) at appropriate points. This
4533 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4534 This provides a means of identifying different callout points. When PCRE
4535 reaches such a point in the regex, if pcre_callout has been set, the external
4536 function is called. It is provided with data in a structure called
4537 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4538 matching continues; if it returns a non-zero value, the match at the current
4539 point fails. However, backtracking will occur if possible. [This was changed
4540 later and other features added - see item 49 below.]
4541
4542 29. pcretest is upgraded to test the callout functionality. It provides a
4543 callout function that displays information. By default, it shows the start of
4544 the match and the current position in the text. There are some new data escapes
4545 to vary what happens:
4546
4547 \C+ in addition, show current contents of captured substrings
4548 \C- do not supply a callout function
4549 \C!n return 1 when callout number n is reached
4550 \C!n!m return 1 when callout number n is reached for the mth time
4551
4552 30. If pcregrep was called with the -l option and just a single file name, it
4553 output "<stdin>" if a match was found, instead of the file name.
4554
4555 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4556 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4557 pcre_exec(). This saves a malloc/free per call. The default value of
4558 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4559 when configuring.
4560
4561 32. The default maximum size of a compiled pattern is 64K. There have been a
4562 few cases of people hitting this limit. The code now uses macros to handle the
4563 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4564 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4565 configuring. Tests 2 and 5 work only with 2-byte links because they output
4566 debugging information about compiled patterns.
4567
4568 33. Internal code re-arrangements:
4569
4570 (a) Moved the debugging function for printing out a compiled regex into
4571 its own source file (printint.c) and used #include to pull it into
4572 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4573 separate copies.
4574
4575 (b) Defined the list of op-code names for debugging as a macro in
4576 internal.h so that it is next to the definition of the opcodes.
4577
4578 (c) Defined a table of op-code lengths for simpler skipping along compiled
4579 code. This is again a macro in internal.h so that it is next to the
4580 definition of the opcodes.
4581
4582 34. Added support for recursive calls to individual subpatterns, along the
4583 lines of Robin Houston's patch (but implemented somewhat differently).
4584
4585 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4586 allow it to read and process whole directories in Win32. This code was
4587 contributed by Lionel Fourquaux; it has not been tested by me.
4588
4589 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4590 used to name a group. Names consist of alphanumerics and underscores, and must
4591 be unique. Back references use the syntax (?P=name) and recursive calls use
4592 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4593 numbers. The function pcre_fullinfo() can be used after compilation to extract
4594 a name/number map. There are three relevant calls:
4595
4596 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4597 PCRE_INFO_NAMECOUNT yields the number of entries
4598 PCRE_INFO_NAMETABLE yields a pointer to the map.
4599
4600 The map is a vector of fixed-size entries. The size of each entry depends on
4601 the length of the longest name used. The first two bytes of each entry are the
4602 group number, most significant byte first. There follows the corresponding
4603 name, zero terminated. The names are in alphabetical order.
4604
4605 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4606 case instead of 255. Making it the same both with and without UTF-8 support
4607 means that the same test output works with both.
4608
4609 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4610 calling malloc() with a zero argument.
4611
4612 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4613 optimization. I've improved things by keeping a bitmap of backreferences with
4614 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4615 fact referenced, the optimization can be applied. It is unlikely that a
4616 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4617 the match to follow \n) will appear inside brackets with a number greater than
4618 31, but if it does, any back reference > 31 suppresses the optimization.
4619
4620 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4621 of disabling numbered capturing parentheses. Any opening parenthesis that is
4622 not followed by ? behaves as if it were followed by ?: but named parentheses
4623 can still be used for capturing (and they will acquire numbers in the usual
4624 way).
4625
4626 41. Redesigned the return codes from the match() function into yes/no/error so
4627 that errors can be passed back from deep inside the nested calls. A malloc
4628 failure while inside a recursive subpattern call now causes the
4629 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4630
4631 42. It is now possible to set a limit on the number of times the match()
4632 function is called in a call to pcre_exec(). This facility makes it possible to
4633 limit the amount of recursion and backtracking, though not in a directly
4634 obvious way, because the match() function is used in a number of different
4635 circumstances. The count starts from zero for each position in the subject
4636 string (for non-anchored patterns). The default limit is, for compatibility, a
4637 large number, namely 10 000 000. You can change this in two ways:
4638
4639 (a) When configuring PCRE before making, you can use --with-match-limit=n
4640 to set a default value for the compiled library.
4641
4642 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4643 a different value is set. See 45 below.
4644
4645 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4646
4647 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4648 of things that can be changed at compile time. The first argument specifies
4649 what is wanted and the second points to where the information is to be placed.
4650 The current list of available information is:
4651
4652 PCRE_CONFIG_UTF8
4653
4654 The output is an integer that is set to one if UTF-8 support is available;
4655 otherwise it is set to zero.
4656
4657 PCRE_CONFIG_NEWLINE
4658
4659 The output is an integer that it set to the value of the code that is used for
4660 newline. It is either LF (10) or CR (13).
4661
4662 PCRE_CONFIG_LINK_SIZE
4663
4664 The output is an integer that contains the number of bytes used for internal
4665 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4666
4667 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4668
4669 The output is an integer that contains the threshold above which the POSIX
4670 interface uses malloc() for output vectors. See item 31 above.
4671
4672 PCRE_CONFIG_MATCH_LIMIT
4673
4674 The output is an unsigned integer that contains the default limit of the number
4675 of match() calls in a pcre_exec() execution. See 42 above.
4676
4677 44. pcretest has been upgraded by the addition of the -C option. This causes it
4678 to extract all the available output from the new pcre_config() function, and to
4679 output it. The program then exits immediately.
4680
4681 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4682 order to support additional features. One way would have been to define
4683 pcre_exec2() (for example) with extra arguments, but this would not have been
4684 extensible, and would also have required all calls to the original function to
4685 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4686 is used for passing in "extra" data from pcre_study().
4687
4688 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4689 contains the following fields:
4690
4691 flags a bitmap indicating which of the following fields are set
4692 study_data opaque data from pcre_study()
4693 match_limit a way of specifying a limit on match() calls for a specific
4694 call to pcre_exec()
4695 callout_data data for callouts (see 49 below)
4696
4697 The flag bits are also defined in pcre.h, and are
4698
4699 PCRE_EXTRA_STUDY_DATA
4700 PCRE_EXTRA_MATCH_LIMIT
4701 PCRE_EXTRA_CALLOUT_DATA
4702
4703 The pcre_study() function now returns one of these new pcre_extra blocks, with
4704 the actual study data pointed to by the study_data field, and the
4705 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4706 before. That is, this change is entirely upwards-compatible and requires no
4707 change to existing code.
4708
4709 If you want to pass in additional data to pcre_exec(), you can either place it
4710 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4711 block.
4712
4713 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4714 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4715 times with different match limits, until it finds the minimum value needed for
4716 pcre_exec() to complete. The value is then output. This can be instructive; for
4717 most simple matches the number is quite small, but for pathological cases it
4718 gets very large very quickly.
4719
4720 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4721 returns the size of the data block pointed to by the study_data field in a
4722 pcre_extra block, that is, the value that was passed as the argument to
4723 pcre_malloc() when PCRE was getting memory in which to place the information
4724 created by pcre_study(). The fourth argument should point to a size_t variable.
4725 pcretest has been extended so that this information is shown after a successful
4726 pcre_study() call when information about the compiled regex is being displayed.
4727
4728 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4729 because what follows is always an absolute path. (Later: it turns out that this
4730 is more than cosmetic for MinGW, because it doesn't like empty path
4731 components.)
4732
4733 49. Some changes have been made to the callout feature (see 28 above):
4734
4735 (i) A callout function now has three choices for what it returns:
4736
4737 0 => success, carry on matching
4738 > 0 => failure at this point, but backtrack if possible
4739 < 0 => serious error, return this value from pcre_exec()
4740
4741 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4742 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4743 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4744 use by callout functions. It will never be used by PCRE itself.
4745
4746 (ii) The pcre_extra structure (see 45 above) has a void * field called
4747 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4748 pcre_callout_block structure has a field of the same name. The contents of
4749 the field passed in the pcre_extra structure are passed to the callout
4750 function in the corresponding field in the callout block. This makes it
4751 easier to use the same callout-containing regex from multiple threads. For
4752 testing, the pcretest program has a new data escape
4753
4754 \C*n pass the number n (may be negative) as callout_data
4755
4756 If the callout function in pcretest receives a non-zero value as
4757 callout_data, it returns that value.
4758
4759 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4760 there were some redundant $(CFLAGS) in commands that are now specified as
4761 $(LINK), which already includes $(CFLAGS).
4762
4763 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4764 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4765 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4766 one-byte characters throughout. Note that case-insensitive matching applies
4767 only to characters whose values are less than 256. PCRE doesn't support the
4768 notion of cases for higher-valued characters.
4769
4770 (i) A character class whose characters are all within 0-255 is handled as
4771 a bit map, and the map is inverted for negative classes. Previously, a
4772 character > 255 always failed to match such a class; however it should
4773 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4774
4775 (ii) A negated character class with a single character < 255 is coded as
4776 "not this character" (OP_NOT). This wasn't working properly when the test
4777 character was multibyte, either singly or repeated.
4778
4779 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4780 mode, for example: \x{100}{2,3}.
4781
4782 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4783 singly or repeated) now correctly test multibyte characters. However,
4784 PCRE doesn't recognize any characters with values greater than 255 as
4785 digits, spaces, or word characters. Such characters always match \D, \S,
4786 and \W, and never match \d, \s, or \w.
4787
4788 (v) Classes may now contain characters and character ranges with values
4789 greater than 255. For example: [ab\x{100}-\x{400}].
4790
4791 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4792 PCRE in UTF-8 mode.
4793
4794 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4795 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4796 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4797 value.)
4798
4799 53. The single man page has become too large. I have therefore split it up into
4800 a number of separate man pages. These also give rise to individual HTML pages;
4801 these are now put in a separate directory, and there is an index.html page that
4802 lists them all. Some hyperlinking between the pages has been installed.
4803
4804 54. Added convenience functions for handling named capturing parentheses.
4805
4806 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4807 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4808 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4809 are faulted.
4810
4811 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4812 calling configure. These values are used when compiling the dftables.c program
4813 which is run to generate the source of the default character tables. They
4814 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4815 you will need to set these values.
4816
4817 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4818
4819
4820 Version 3.9 02-Jan-02
4821 ---------------------
4822
4823 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4824
4825 2. If --disable-static was given, the building process failed when trying to
4826 build pcretest and pcregrep. (For some reason it was using libtool to compile
4827 them, which is not right, as they aren't part of the library.)
4828
4829
4830 Version 3.8 18-Dec-01
4831 ---------------------
4832
4833 1. The experimental UTF-8 code was completely screwed up. It was packing the
4834 bytes in the wrong order. How dumb can you get?
4835
4836
4837 Version 3.7 29-Oct-01
4838 ---------------------
4839
4840 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4841 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4842 this didn't happen under Solaris 8, where I normally test things.
4843
4844 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4845 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4846 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4847 this fix an hour or so after the initial 3.7 release.)
4848
4849
4850 Version 3.6 23-Oct-01
4851 ---------------------
4852
4853 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4854 offsets passed as NULL with zero offset count.
4855
4856 2. The config.guess and config.sub files had not been updated when I moved to
4857 the latest autoconf.
4858
4859
4860 Version 3.5 15-Aug-01
4861 ---------------------
4862
4863 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4864 had been forgotten.
4865
4866 2. By using declared but undefined structures, we can avoid using "void"
4867 definitions in pcre.h while keeping the internal definitions of the structures
4868 private.
4869
4870 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4871 user point of view, this means that both static and shared libraries are built
4872 by default, but this can be individually controlled. More of the work of
4873 handling this static/shared cases is now inside libtool instead of PCRE's make
4874 file.
4875
4876 4. The pcretest utility is now installed along with pcregrep because it is
4877 useful for users (to test regexs) and by doing this, it automatically gets
4878 relinked by libtool. The documentation has been turned into a man page, so
4879 there are now .1, .txt, and .html versions in /doc.
4880
4881 5. Upgrades to pcregrep:
4882 (i) Added long-form option names like gnu grep.
4883 (ii) Added --help to list all options with an explanatory phrase.
4884 (iii) Added -r, --recursive to recurse into sub-directories.
4885 (iv) Added -f, --file to read patterns from a file.
4886
4887 6. pcre_exec() was referring to its "code" argument before testing that
4888 argument for NULL (and giving an error if it was NULL).
4889
4890 7. Upgraded Makefile.in to allow for compiling in a different directory from
4891 the source directory.
4892
4893 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4894 options bits, the pointer it was passed was to an int instead of to an unsigned
4895 long int. This mattered only on 64-bit systems.
4896
4897 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4898 generated) instead of pcre.in, which it its source. Also made the same change
4899 in several of the .c files.
4900
4901 10. A new release of gcc defines printf() as a macro, which broke pcretest
4902 because it had an ifdef in the middle of a string argument for printf(). Fixed
4903 by using separate calls to printf().
4904
4905 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4906 script, to force use of CR or LF instead of \n in the source. On non-Unix
4907 systems, the value can be set in config.h.
4908
4909 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4910 absolute limit. Changed the text of the error message to make this clear, and
4911 likewise updated the man page.
4912
4913 13. The limit of 99 on the number of capturing subpatterns has been removed.
4914 The new limit is 65535, which I hope will not be a "real" limit.
4915
4916
4917 Version 3.4 22-Aug-00
4918 ---------------------
4919
4920 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4921
4922 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4923
4924
4925 Version 3.3 01-Aug-00
4926 ---------------------
4927
4928 1. If an octal character was given, but the value was greater than \377, it
4929 was not getting masked to the least significant bits, as documented. This could
4930 lead to crashes in some systems.
4931
4932 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4933 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4934
4935 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4936 These just pass their arguments on to (pcre_free)(), but they are provided
4937 because some uses of PCRE bind it to non-C systems that can call its functions,
4938 but cannot call free() or pcre_free() directly.
4939
4940 4. Add "make test" as a synonym for "make check". Corrected some comments in
4941 the Makefile.
4942
4943 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4944 Makefile.
4945
4946 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4947 command called pgrep for grepping around the active processes.
4948
4949 7. Added the beginnings of support for UTF-8 character strings.
4950
4951 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4952 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4953 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4954 out for the ar command.)
4955
4956
4957 Version 3.2 12-May-00
4958 ---------------------
4959
4960 This is purely a bug fixing release.
4961
4962 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4963 of ZA. This was just one example of several cases that could provoke this bug,
4964 which was introduced by change 9 of version 2.00. The code for breaking
4965 infinite loops after an iteration that matches an empty string was't working
4966 correctly.
4967
4968 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4969 when matched against abbab (for example). After matching an empty string, it
4970 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4971 caused it to match further down the string than it should.
4972
4973 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4974 was there because it doesn't seem to be needed, and it causes trouble on some
4975 systems, as it is not a Standard C header. It has been removed.
4976
4977 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4978 were reported on the Macintosh. The changes were from
4979
4980 while ((c = *(++ptr)) != 0 && c != '\n');
4981 to
4982 while ((c = *(++ptr)) != 0 && c != '\n') ;
4983
4984 Totally extraordinary, but if that's what it takes...
4985
4986 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4987 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4988 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4989 assumes the way PCRE uses memmove() (always moving upwards).
4990
4991 6. PCRE is being used in one environment where strchr() is not available. There
4992 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4993 faster code anyway.
4994
4995
4996 Version 3.1 09-Feb-00
4997 ---------------------
4998
4999 The only change in this release is the fixing of some bugs in Makefile.in for
5000 the "install" target:
5001
5002 (1) It was failing to install pcreposix.h.
5003
5004 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5005
5006
5007 Version 3.0 01-Feb-00
5008 ---------------------
5009
5010 1. Add support for the /+ modifier to perltest (to output $` like it does in
5011 pcretest).
5012
5013 2. Add support for the /g modifier to perltest.
5014
5015 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5016 matches null strings.
5017
5018 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5019 pattern. Perl treats empty patterns specially - it reuses the most recent
5020 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5021 effect.
5022
5023 5. The POSIX interface was broken in that it was just handing over the POSIX
5024 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5025 required a bigger vector, with some working space on the end. This means that
5026 the POSIX wrapper now has to get and free some memory, and copy the results.
5027
5028 6. Added some simple autoconf support, placing the test data and the
5029 documentation in separate directories, re-organizing some of the
5030 information files, and making it build pcre-config (a GNU standard). Also added
5031 libtool support for building PCRE as a shared library, which is now the
5032 default.
5033
5034 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5035 09 are not valid octal constants. Single digits will be used for minor values
5036 less than 10.
5037
5038 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5039 existing programs that set these in the POSIX interface can use PCRE without
5040 modification.
5041
5042 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5043 return all that pcre_info() returns, plus additional data. The pcre_info()
5044 function is retained for compatibility, but is considered to be obsolete.
5045
5046 10. Added experimental recursion feature (?R) to handle one common case that
5047 Perl 5.6 will be able to do with (?p{...}).
5048
5049 11. Added support for POSIX character classes like [:alpha:], which Perl is
5050 adopting.
5051
5052
5053 Version 2.08 31-Aug-99
5054 ----------------------
5055
5056 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5057 trying to match at the startoffset position, but instead was moving forward to
5058 the next newline as if a previous match had failed.
5059
5060 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5061 and could get into a loop if a null string was matched other than at the start
5062 of the subject.
5063
5064 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5065 be distinguished at compile time, and for completeness also added PCRE_DATE.
5066
5067 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5068 in GnuWin32 environments.
5069
5070
5071 Version 2.07 29-Jul-99
5072 ----------------------
5073
5074 1. The documentation is now supplied in plain text form and HTML as well as in
5075 the form of man page sources.
5076
5077 2. C++ compilers don't like assigning (void *) values to other pointer types.
5078 In particular this affects malloc(). Although there is no problem in Standard
5079 C, I've put in casts to keep C++ compilers happy.
5080
5081 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5082 should be (const char *).
5083
5084 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5085 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5086 However, I haven't made this a standard facility. The documentation doesn't
5087 mention it, and the Makefile doesn't support it.
5088
5089 5. The Makefile now contains an "install" target, with editable destinations at
5090 the top of the file. The pcretest program is not installed.
5091
5092 6. pgrep -V now gives the PCRE version number and date.
5093
5094 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5095 causing the entire string to be ignored, instead of just the last character.
5096
5097 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5098 non-matching string, it can take a very, very long time, even for strings of
5099 quite modest length, because of the nested recursion. PCRE now does better in
5100 some of these cases. It does this by remembering the last required literal
5101 character in the pattern, and pre-searching the subject to ensure it is present
5102 before running the real match. In other words, it applies a heuristic to detect
5103 some types of certain failure quickly, and in the above example, if presented
5104 with a string that has no trailing " it gives "no match" very quickly.
5105
5106 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5107 other alternatives are tried instead.
5108
5109
5110 Version 2.06 09-Jun-99
5111 ----------------------
5112
5113 1. Change pcretest's output for amount of store used to show just the code
5114 space, because the remainder (the data block) varies in size between 32-bit and
5115 64-bit systems.
5116
5117 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5118 start matching at. This allows lookbehinds to work when searching for multiple
5119 occurrences in a string.
5120
5121 3. Added additional options to pcretest for testing multiple occurrences:
5122
5123 /+ outputs the rest of the string that follows a match
5124 /g loops for multiple occurrences, using the new startoffset argument
5125 /G loops for multiple occurrences by passing an incremented pointer
5126
5127 4. PCRE wasn't doing the "first character" optimization for patterns starting
5128 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5129 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5130 the letter 'x'. On long subject strings, this gives a significant speed-up.
5131
5132
5133 Version 2.05 21-Apr-99
5134 ----------------------
5135
5136 1. Changed the type of magic_number from int to long int so that it works
5137 properly on 16-bit systems.
5138
5139 2. Fixed a bug which caused patterns starting with .* not to work correctly
5140 when the subject string contained newline characters. PCRE was assuming
5141 anchoring for such patterns in all cases, which is not correct because .* will
5142 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5143 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5144 must be retried after every newline in the subject.
5145
5146
5147 Version 2.04 18-Feb-99
5148 ----------------------
5149
5150 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5151 computation of the store needed to hold the pattern was incorrect (too large).
5152 If such patterns were nested a few deep, this could multiply and become a real
5153 problem.
5154
5155 2. Added /M option to pcretest to show the memory requirement of a specific
5156 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5157
5158 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5159 compiled in such a way that the backtracking after subsequent failure was
5160 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5161 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5162
5163
5164 Version 2.03 02-Feb-99
5165 ----------------------
5166
5167 1. Fixed typo and small mistake in man page.
5168
5169 2. Added 4th condition (GPL supersedes if conflict) and created separate
5170 LICENCE file containing the conditions.
5171
5172 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5173 Perl, that is the internal \ allows the delimiter to be included in the
5174 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5175 the final delimiter, add \ to the end of the pattern (to test the error).
5176
5177 4. Added the convenience functions for extracting substrings after a successful
5178 match. Updated pcretest to make it able to test these functions.
5179
5180
5181 Version 2.02 14-Jan-99
5182 ----------------------
5183
5184 1. Initialized the working variables associated with each extraction so that
5185 their saving and restoring doesn't refer to uninitialized store.
5186
5187 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5188 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5189 fix the problem.
5190
5191 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5192 calls, and wasn't printing the correct value for compiling calls. Increased the
5193 default value of LOOPREPEAT, and the number of significant figures in the
5194 times.
5195
5196 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5197
5198 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5199 a building problem on Windows NT with a FAT file system.
5200
5201
5202 Version 2.01 21-Oct-98
5203 ----------------------
5204
5205 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5206 to character tables built by pcre_maketables() in the current locale. If NULL
5207 is passed, the default tables are used.
5208
5209
5210 Version 2.00 24-Sep-98
5211 ----------------------
5212
5213 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5214 it any more.
5215
5216 2. Allow quantification of (?>) groups, and make it work correctly.
5217
5218 3. The first character computation wasn't working for (?>) groups.
5219
5220 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5221 end of the subject) and add 5.005's \z, which really does match only at the
5222 very end of the subject.
5223
5224 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5225
5226 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5227 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5228 localized options. All options to pcre_study() were also removed.
5229
5230 7. Add other new features from 5.005:
5231
5232 $(?<= positive lookbehind
5233 $(?<! negative lookbehind
5234 (?imsx-imsx) added the unsetting capability
5235 such a setting is global if at outer level; local otherwise
5236 (?imsx-imsx:) non-capturing groups with option setting
5237 (?(cond)re|re) conditional pattern matching
5238
5239 A backreference to itself in a repeated group matches the previous
5240 captured string.
5241
5242 8. General tidying up of studying (both automatic and via "study")
5243 consequential on the addition of new assertions.
5244
5245 9. As in 5.005, unlimited repeated groups that could match an empty substring
5246 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5247 runtime if any iteration does actually match an empty substring.
5248
5249 10. Include the RunTest script in the distribution.
5250
5251 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5252 discrepancies, some of which were old and were also with respect to 5.004. They
5253 have now been fixed.
5254
5255
5256 Version 1.09 28-Apr-98
5257 ----------------------
5258
5259 1. A negated single character class followed by a quantifier with a minimum
5260 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5261 program crashes, or just wrong answers. This did not apply to negated classes
5262 containing more than one character, or to minima other than one.
5263
5264
5265 Version 1.08 27-Mar-98
5266 ----------------------
5267
5268 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5269
5270 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5271 latter must appear before anything that relies on it in the pattern.
5272
5273
5274 Version 1.07 16-Feb-98
5275 ----------------------
5276
5277 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5278 repeat of a potentially empty string).
5279
5280
5281 Version 1.06 23-Jan-98
5282 ----------------------
5283
5284 1. Added Markus Oberhumer's little patches for C++.
5285
5286 2. Literal strings longer than 255 characters were broken.
5287
5288
5289 Version 1.05 23-Dec-97
5290 ----------------------
5291
5292 1. Negated character classes containing more than one character were failing if
5293 PCRE_CASELESS was set at run time.
5294
5295
5296 Version 1.04 19-Dec-97
5297 ----------------------
5298
5299 1. Corrected the man page, where some "const" qualifiers had been omitted.
5300
5301 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5302 input syntax.
5303
5304 3. Fixed memory leak which occurred when a regex with back references was
5305 matched with an offsets vector that wasn't big enough. The temporary memory
5306 that is used in this case wasn't being freed if the match failed.
5307
5308 4. Tidied pcretest to ensure it frees memory that it gets.
5309
5310 5. Temporary memory was being obtained in the case where the passed offsets
5311 vector was exactly big enough.
5312
5313 6. Corrected definition of offsetof() from change 5 below.
5314
5315 7. I had screwed up change 6 below and broken the rules for the use of
5316 setjmp(). Now fixed.
5317
5318
5319 Version 1.03 18-Dec-97
5320 ----------------------
5321
5322 1. A erroneous regex with a missing opening parenthesis was correctly
5323 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5324 on some systems.
5325
5326 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5327 it was reported that one broken compiler failed on the former because "code" is
5328 also an independent variable.
5329
5330 3. The erroneous regex a[]b caused an array overrun reference.
5331
5332 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5333 fail on data ending with that character. (It was going on too far, and checking
5334 the next character, typically a binary zero.) This was specific to the
5335 optimized code for single-character negative classes.
5336
5337 5. Added a contributed patch from the TIN world which does the following:
5338
5339 + Add an undef for memmove, in case the the system defines a macro for it.
5340
5341 + Add a definition of offsetof(), in case there isn't one. (I don't know
5342 the reason behind this - offsetof() is part of the ANSI standard - but
5343 it does no harm).
5344
5345 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5346 most of the places where whitespace preceded '#'. I have given up and
5347 allowed the remaining 2 cases to be at the margin.
5348
5349 + Rename some variables in pcre to eliminate shadowing. This seems very
5350 pedantic, but does no harm, of course.
5351
5352 6. Moved the call to setjmp() into its own function, to get rid of warnings
5353 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5354
5355 7. Constructs such as \d{8,} were compiling into the equivalent of
5356 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5357 outcome, but in this particular case used more store than had been allocated,
5358 which caused the bug to be discovered because it threw up an internal error.
5359
5360 8. The debugging code in both pcre and pcretest for outputting the compiled
5361 form of a regex was going wrong in the case of back references followed by
5362 curly-bracketed repeats.
5363
5364
5365 Version 1.02 12-Dec-97
5366 ----------------------
5367
5368 1. Typos in pcre.3 and comments in the source fixed.
5369
5370 2. Applied a contributed patch to get rid of places where it used to remove
5371 'const' from variables, and fixed some signed/unsigned and uninitialized
5372 variable warnings.
5373
5374 3. Added the "runtest" target to Makefile.
5375
5376 4. Set default compiler flag to -O2 rather than just -O.
5377
5378
5379 Version 1.01 19-Nov-97
5380 ----------------------
5381
5382 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5383 like /([ab]*)*/, that is, for classes with more than one character in them.
5384
5385 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5386 as /((?>a*))*/ (a PCRE_EXTRA facility).
5387
5388
5389 Version 1.00 18-Nov-97
5390 ----------------------
5391
5392 1. Added compile-time macros to support systems such as SunOS4 which don't have
5393 memmove() or strerror() but have other things that can be used instead.
5394
5395 2. Arranged that "make clean" removes the executables.
5396
5397
5398 Version 0.99 27-Oct-97
5399 ----------------------
5400
5401 1. Fixed bug in code for optimizing classes with only one character. It was
5402 initializing a 32-byte map regardless, which could cause it to run off the end
5403 of the memory it had got.
5404
5405 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5406
5407
5408 Version 0.98 22-Oct-97
5409 ----------------------
5410
5411 1. Fixed bug in code for handling temporary memory usage when there are more
5412 back references than supplied space in the ovector. This could cause segfaults.
5413
5414
5415 Version 0.97 21-Oct-97
5416 ----------------------
5417
5418 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5419
5420 2. Optimized negated single characters not to use a bit map.
5421
5422 3. Brought error texts together as macro definitions; clarified some of them;
5423 fixed one that was wrong - it said "range out of order" when it meant "invalid
5424 escape sequence".
5425
5426 4. Changed some char * arguments to const char *.
5427
5428 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5429
5430 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5431 pcretest.
5432
5433
5434 Version 0.96 16-Oct-97
5435 ----------------------
5436
5437 1. Added a simple "pgrep" utility to the distribution.
5438
5439 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5440 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5441 where "ddd" means "one or more decimal digits".
5442
5443 3. Fixed serious bug. If a pattern had a back reference, but the call to
5444 pcre_exec() didn't supply a large enough ovector to record the related
5445 identifying subpattern, the match always failed. PCRE now remembers the number
5446 of the largest back reference, and gets some temporary memory in which to save
5447 the offsets during matching if necessary, in order to ensure that
5448 backreferences always work.
5449
5450 4. Increased the compatibility with Perl in a number of ways:
5451
5452 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5453 to request this handling. The option can be set at compile or exec time.
5454
5455 (b) $ matches before a terminating newline by default; an option
5456 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5457 mode). The option can be set at compile or exec time.
5458
5459 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5460 the same as Perl's. If the decimal number it represents is less than 10
5461 or there aren't that many previous left capturing parentheses, an octal
5462 escape is read. Inside a character class, it's always an octal escape,
5463 even if it is a single digit.
5464
5465 (d) An escaped but undefined alphabetic character is taken as a literal,
5466 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5467 escapes.
5468
5469 (e) {0} is now permitted. (The previous item is removed from the compiled
5470 pattern).
5471
5472 5. Changed all the names of code files so that the basic parts are no longer
5473 than 10 characters, and abolished the teeny "globals.c" file.
5474
5475 6. Changed the handling of character classes; they are now done with a 32-byte
5476 bit map always.
5477
5478 7. Added the -d and /D options to pcretest to make it possible to look at the
5479 internals of compilation without having to recompile pcre.
5480
5481
5482 Version 0.95 23-Sep-97
5483 ----------------------
5484
5485 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5486 \x20 at the start of a run of normal characters. These were being treated as
5487 real characters, instead of the source characters being re-checked.
5488
5489
5490 Version 0.94 18-Sep-97
5491 ----------------------
5492
5493 1. The functions are now thread-safe, with the caveat that the global variables
5494 containing pointers to malloc() and free() or alternative functions are the
5495 same for all threads.
5496
5497 2. Get pcre_study() to generate a bitmap of initial characters for non-
5498 anchored patterns when this is possible, and use it if passed to pcre_exec().
5499
5500
5501 Version 0.93 15-Sep-97
5502 ----------------------
5503
5504 1. /(b)|(:+)/ was computing an incorrect first character.
5505
5506 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5507 but not actually doing anything yet.
5508
5509 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5510 as Perl does (e.g. [-az] or [az-]).
5511
5512 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5513 all possible positions.
5514
5515 5. Split up into different modules to avoid including unneeded functions in a
5516 compiled binary. However, compile and exec are still in one module. The "study"
5517 function is split off.
5518
5519 6. The character tables are now in a separate module whose source is generated
5520 by an auxiliary program - but can then be edited by hand if required. There are
5521 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5522 toupper() in the code.
5523
5524 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5525 make them global. Abolish the function for setting them, as the caller can now
5526 set them directly.
5527
5528
5529 Version 0.92 11-Sep-97
5530 ----------------------
5531
5532 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5533 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5534
5535 2. Caseless matching was not working in character classes if the characters in
5536 the pattern were in upper case.
5537
5538 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5539
5540 4. Make PCRE_ANCHORED public and accept as a compile option.
5541
5542 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5543 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5544 pass them.
5545
5546 6. Give an error if bad option bits passed at compile or run time.
5547
5548 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5549 pcretest to cause it to pass that flag.
5550
5551 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5552 options, and the first character, if set.
5553
5554 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5555
5556
5557 Version 0.91 10-Sep-97
5558 ----------------------
5559
5560 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5561 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5562
5563 2. PCRE was looping on encountering an indefinitely repeated back reference to
5564 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5565 Perl does - treats the match as successful.
5566
5567 ****

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12