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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1481 - (show annotations)
Tue May 27 18:24:42 2014 UTC (5 years, 4 months ago) by ph10
File size: 266152 byte(s)
Give error for \x{} and \o{}.
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
51 Version 8.35 04-April-2014
52 --------------------------
53
54 1. A new flag is set, when property checks are present in an XCLASS.
55 When this flag is not set, PCRE can perform certain optimizations
56 such as studying these XCLASS-es.
57
58 2. The auto-possessification of character sets were improved: a normal
59 and an extended character set can be compared now. Furthermore
60 the JIT compiler optimizes more character set checks.
61
62 3. Got rid of some compiler warnings for potentially uninitialized variables
63 that show up only when compiled with -O2.
64
65 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
66 of a match later then the end of the match. The pcretest program was not
67 handling the case sensibly - it was outputting from the start to the next
68 binary zero. It now reports this situation in a message, and outputs the
69 text from the end to the start.
70
71 5. Fast forward search is improved in JIT. Instead of the first three
72 characters, any three characters with fixed position can be searched.
73 Search order: first, last, middle.
74
75 6. Improve character range checks in JIT. Characters are read by an inprecise
76 function now, which returns with an unknown value if the character code is
77 above a certain treshold (e.g: 256). The only limitation is that the value
78 must be bigger than the treshold as well. This function is useful, when
79 the characters above the treshold are handled in the same way.
80
81 7. The macros whose names start with RAWUCHAR are placeholders for a future
82 mode in which only the bottom 21 bits of 32-bit data items are used. To
83 make this more memorable for those maintaining the code, the names have
84 been changed to start with UCHAR21, and an extensive comment has been added
85 to their definition.
86
87 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
88 to the export list in Makefile.am (they were accidentally omitted from the
89 8.34 tarball).
90
91 9. The informational output from pcretest used the phrase "starting byte set"
92 which is inappropriate for the 16-bit and 32-bit libraries. As the output
93 for "first char" and "need char" really means "non-UTF-char", I've changed
94 "byte" to "char", and slightly reworded the output. The documentation about
95 these values has also been (I hope) clarified.
96
97 10. Another JIT related optimization: use table jumps for selecting the correct
98 backtracking path, when more than four alternatives are present inside a
99 bracket.
100
101 11. Empty match is not possible, when the minimum length is greater than zero,
102 and there is no \K in the pattern. JIT should avoid empty match checks in
103 such cases.
104
105 12. In a caseless character class with UCP support, when a character with more
106 than one alternative case was not the first character of a range, not all
107 the alternative cases were added to the class. For example, s and \x{17f}
108 are both alternative cases for S: the class [RST] was handled correctly,
109 but [R-T] was not.
110
111 13. The configure.ac file always checked for pthread support when JIT was
112 enabled. This is not used in Windows, so I have put this test inside a
113 check for the presence of windows.h (which was already tested for).
114
115 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
116 The algorithm provides a way to skip certain starting offsets, and usually
117 faster than linear prefix searches.
118
119 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
120 as for 'fr_FR' and 'french'. For some reason, however, it then used the
121 Windows-specific input and output files, which have 'french' screwed in.
122 So this could never have worked. One of the problems with locales is that
123 they aren't always the same. I have now updated RunTest so that it checks
124 the output of the locale test (test 3) against three different output
125 files, and it allows the test to pass if any one of them matches. With luck
126 this should make the test pass on some versions of Solaris where it was
127 failing. Because of the uncertainty, the script did not used to stop if
128 test 3 failed; it now does. If further versions of a French locale ever
129 come to light, they can now easily be added.
130
131 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
132 there was a message during ./configure, but it did not stop. This now
133 provokes an error. The invalid example in README has been corrected.
134 If a value less than the minimum is given, the minimum value has always
135 been used, but now a warning is given.
136
137 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
138 was a bug in the test system, which is now fixed. Also, the list of various
139 configurations that are tested for each release did not have one with both
140 16/32 bits and --enable-bar-anycrlf. It now does.
141
142 18. pcretest was missing "-C bsr" for displaying the \R default setting.
143
144 19. Little endian PowerPC systems are supported now by the JIT compiler.
145
146 20. The fast forward newline mechanism could enter to an infinite loop on
147 certain invalid UTF-8 input. Although we don't support these cases
148 this issue can be fixed by a performance optimization.
149
150 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
151 not take account if existing stack usage. There is now a new global
152 variable called pcre_stack_guard that can be set to point to an external
153 function to check stack availability. It is called at the start of
154 processing every parenthesized group.
155
156 22. A typo in the code meant that in ungreedy mode the max/min qualifier
157 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
158 not match "ab".
159
160 23. When UTF was disabled, the JIT program reported some incorrect compile
161 errors. These messages are silenced now.
162
163 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
164 compiler.
165
166 25. Change all the temporary files used in RunGrepTest to be different to those
167 used by RunTest so that the tests can be run simultaneously, for example by
168 "make -j check".
169
170
171 Version 8.34 15-December-2013
172 -----------------------------
173
174 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
175 executable memory. Patch inspired by Carsten Klein.
176
177 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
178 this macro is never tested and has no effect, because the work to support
179 coverage involves only compiling and linking options and special targets in
180 the Makefile. The comment in config.h implied that defining the macro would
181 enable coverage support, which is totally false. There was also support for
182 setting this macro in the CMake files (my fault, I just copied it from
183 configure). SUPPORT_GCOV has now been removed.
184
185 3. Make a small performance improvement in strlen16() and strlen32() in
186 pcretest.
187
188 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
189 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
190 these cases). There was also one in pcretest.c.
191
192 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
193
194 6. In UTF mode, the code for checking whether a group could match an empty
195 string (which is used for indefinitely repeated groups to allow for
196 breaking an infinite loop) was broken when the group contained a repeated
197 negated single-character class with a character that occupied more than one
198 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
199 UTF-8 mode). The effect was undefined: the group might or might not be
200 deemed as matching an empty string, or the program might have crashed.
201
202 7. The code for checking whether a group could match an empty string was not
203 recognizing that \h, \H, \v, \V, and \R must match a character.
204
205 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
206 an empty string. If it can, pcretest shows this in its information output.
207
208 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
209 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
210 matched by pcre_exec() without using JIT:
211
212 (a) If the rest of the pattern did not match after a maximal run of
213 grapheme clusters, the code for backing up to try with fewer of them
214 did not always back up over a full grapheme when characters that do not
215 have the modifier quality were involved, e.g. Hangul syllables.
216
217 (b) If the match point in a subject started with modifier character, and
218 there was no match, the code could incorrectly back up beyond the match
219 point, and potentially beyond the first character in the subject,
220 leading to a segfault or an incorrect match result.
221
222 10. A conditional group with an assertion condition could lead to PCRE
223 recording an incorrect first data item for a match if no other first data
224 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
225 first data item, and therefore matched "ca" after "c" instead of at the
226 start.
227
228 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
229 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
230
231 12. The source of pcregrep now includes z/OS-specific code so that it can be
232 compiled for z/OS as part of the special z/OS distribution.
233
234 13. Added the -T and -TM options to pcretest.
235
236 14. The code in pcre_compile.c for creating the table of named capturing groups
237 has been refactored. Instead of creating the table dynamically during the
238 actual compiling pass, the information is remembered during the pre-compile
239 pass (on the stack unless there are more than 20 named groups, in which
240 case malloc() is used) and the whole table is created before the actual
241 compile happens. This has simplified the code (it is now nearly 150 lines
242 shorter) and prepared the way for better handling of references to groups
243 with duplicate names.
244
245 15. A back reference to a named subpattern when there is more than one of the
246 same name now checks them in the order in which they appear in the pattern.
247 The first one that is set is used for the reference. Previously only the
248 first one was inspected. This change makes PCRE more compatible with Perl.
249
250 16. Unicode character properties were updated from Unicode 6.3.0.
251
252 17. The compile-time code for auto-possessification has been refactored, based
253 on a patch by Zoltan Herczeg. It now happens after instead of during
254 compilation. The code is cleaner, and more cases are handled. The option
255 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
256 options in pcretest are provided to set it. It can also be set by
257 (*NO_AUTO_POSSESS) at the start of a pattern.
258
259 18. The character VT has been added to the default ("C" locale) set of
260 characters that match \s and are generally treated as white space,
261 following this same change in Perl 5.18. There is now no difference between
262 "Perl space" and "POSIX space". Whether VT is treated as white space in
263 other locales depends on the locale.
264
265 19. The code for checking named groups as conditions, either for being set or
266 for being recursed, has been refactored (this is related to 14 and 15
267 above). Processing unduplicated named groups should now be as fast at
268 numerical groups, and processing duplicated groups should be faster than
269 before.
270
271 20. Two patches to the CMake build system, by Alexander Barkov:
272
273 (1) Replace the "source" command by "." in CMakeLists.txt because
274 "source" is a bash-ism.
275
276 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
277 without these the CMake build does not work on Solaris.
278
279 21. Perl has changed its handling of \8 and \9. If there is no previously
280 encountered capturing group of those numbers, they are treated as the
281 literal characters 8 and 9 instead of a binary zero followed by the
282 literals. PCRE now does the same.
283
284 22. Following Perl, added \o{} to specify codepoints in octal, making it
285 possible to specify values greater than 0777 and also making them
286 unambiguous.
287
288 23. Perl now gives an error for missing closing braces after \x{... instead of
289 treating the string as literal. PCRE now does the same.
290
291 24. RunTest used to grumble if an inappropriate test was selected explicitly,
292 but just skip it when running all tests. This make it awkward to run ranges
293 of tests when one of them was inappropriate. Now it just skips any
294 inappropriate tests, as it always did when running all tests.
295
296 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
297 character types such as \d or \w, too many callouts were inserted, and the
298 data that they returned was rubbish.
299
300 26. In UCP mode, \s was not matching two of the characters that Perl matches,
301 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
302 were matched by \h. The code has now been refactored so that the lists of
303 the horizontal and vertical whitespace characters used for \h and \v (which
304 are defined only in one place) are now also used for \s.
305
306 27. Add JIT support for the 64 bit TileGX architecture.
307 Patch by Jiong Wang (Tilera Corporation).
308
309 28. Possessive quantifiers for classes (both explicit and automatically
310 generated) now use special opcodes instead of wrapping in ONCE brackets.
311
312 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
313 (because it's meaningless), this was not happening when PCRE_CASELESS was
314 set. Not wrong, but inefficient.
315
316 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
317 properties for \w, \d, etc) is present in a test regex. Otherwise if the
318 test contains no characters greater than 255, Perl doesn't realise it
319 should be using Unicode semantics.
320
321 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
322 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
323 does in Unicode mode.
324
325 32. Added the "forbid" facility to pcretest so that putting tests into the
326 wrong test files can sometimes be quickly detected.
327
328 33. There is now a limit (default 250) on the depth of nesting of parentheses.
329 This limit is imposed to control the amount of system stack used at compile
330 time. It can be changed at build time by --with-parens-nest-limit=xxx or
331 the equivalent in CMake.
332
333 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
334 errors. Perl warns for these when in warning mode, but PCRE has no facility
335 for giving warnings.
336
337 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
338 However, this was not working for (?!) because it is optimized to (*FAIL),
339 for which PCRE does not allow quantifiers. The optimization is now disabled
340 when a quantifier follows (?!). I can't see any use for this, but it makes
341 things uniform.
342
343 36. Perl no longer allows group names to start with digits, so I have made this
344 change also in PCRE. It simplifies the code a bit.
345
346 37. In extended mode, Perl ignores spaces before a + that indicates a
347 possessive quantifier. PCRE allowed a space before the quantifier, but not
348 before the possessive +. It now does.
349
350 38. The use of \K (reset reported match start) within a repeated possessive
351 group such as (a\Kb)*+ was not working.
352
353 40. Document that the same character tables must be used at compile time and
354 run time, and that the facility to pass tables to pcre_exec() and
355 pcre_dfa_exec() is for use only with saved/restored patterns.
356
357 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
358 features for Builds with MSVC:
359
360 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
361 builds). The .rc files can be used to set FileDescription and many other
362 attributes.
363
364 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
365 This allows higher-level build scripts which want .pdb files to avoid
366 hard-coding the exact files needed."
367
368 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
369 mean "start of word" and "end of word", respectively, as a transition aid.
370
371 43. A minimizing repeat of a class containing codepoints greater than 255 in
372 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
373 compiled to use the heap for recursion.
374
375 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
376 is configured.
377
378
379 Version 8.33 28-May-2013
380 ------------------------
381
382 1. Added 'U' to some constants that are compared to unsigned integers, to
383 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
384 variables that are added to signed variables, to ensure the result is
385 signed and can be negated.
386
387 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
388 CMake config files.
389
390 3. Revise the creation of config.h.generic so that all boolean macros are
391 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
392 overriding via -D on the command line possible.
393
394 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
395 to unsigned int is reported to make a quite noticeable speed difference in
396 a specific Windows environment. Testing on Linux did also appear to show
397 some benefit (and it is clearly not harmful). Also fixed the definition of
398 Xop which should be unsigned.
399
400 5. Related to (4), changing the definition of the intermediate variable cc
401 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
402 improvements.
403
404 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
405 unnecessary spaces.
406
407 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
408 and later.
409
410 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
411
412 9. Optimizing fast_forward_start_bits in JIT.
413
414 10. Adding support for callouts in JIT, and fixing some issues revealed
415 during this work. Namely:
416
417 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
418
419 (b) Minimum length was not checked before the matching is started.
420
421 11. The value of capture_last that is passed to callouts was incorrect in some
422 cases when there was a capture on one path that was subsequently abandoned
423 after a backtrack. Also, the capture_last value is now reset after a
424 recursion, since all captures are also reset in this case.
425
426 12. The interpreter no longer returns the "too many substrings" error in the
427 case when an overflowing capture is in a branch that is subsequently
428 abandoned after a backtrack.
429
430 13. In the pathological case when an offset vector of size 2 is used, pcretest
431 now prints out the matched string after a yield of 0 or 1.
432
433 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
434 Only supported by the JIT compiler at the moment.
435
436 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
437
438 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
439 offset of the starting point of the matching process, provided the offsets
440 vector is large enough.
441
442 17. The \A escape now records a lookbehind value of 1, though its execution
443 does not actually inspect the previous character. This is to ensure that,
444 in partial multi-segment matching, at least one character from the old
445 segment is retained when a new segment is processed. Otherwise, if there
446 are no lookbehinds in the pattern, \A might match incorrectly at the start
447 of a new segment.
448
449 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
450
451 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
452 gives some modest performance improvement in 8-bit mode.
453
454 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
455 be expressed in certain programming languages using Universal Character
456 Names.
457
458 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
459 which points out that "non characters" are not "characters that may not
460 appear in Unicode strings" but rather "characters that are reserved for
461 internal use and have only local meaning".
462
463 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
464 there was a conditional group that depended on an assertion, if the
465 assertion was false, the callout that immediately followed the alternation
466 in the condition was skipped when pcre_exec() was used for matching.
467
468 23. Allow an explicit callout to be inserted before an assertion that is the
469 condition for a conditional group, for compatibility with automatic
470 callouts, which always insert a callout at this point.
471
472 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
473 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
474
475 25. (*PRUNE) is now supported by the JIT compiler.
476
477 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
478
479 27. Fix the case where there are two or more SKIPs with arguments that may be
480 ignored.
481
482 28. (*SKIP) is now supported by the JIT compiler.
483
484 29. (*THEN) is now supported by the JIT compiler.
485
486 30. Update RunTest with additional test selector options.
487
488 31. The way PCRE handles backtracking verbs has been changed in two ways.
489
490 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
491 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
492 backtracking. In some cases this makes it more Perl-compatible, but Perl's
493 rather obscure rules do not always do the same thing.
494
495 (2) Previously, backtracking verbs were confined within assertions. This is
496 no longer the case for positive assertions, except for (*ACCEPT). Again,
497 this sometimes improves Perl compatibility, and sometimes does not.
498
499 32. A number of tests that were in test 2 because Perl did things differently
500 have been moved to test 1, because either Perl or PCRE has changed, and
501 these tests are now compatible.
502
503 32. Backtracking control verbs are now handled in the same way in JIT and
504 interpreter.
505
506 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
507 contained a forward subroutine reference caused a compile error.
508
509 34. Auto-detect and optimize limited repetitions in JIT.
510
511 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
512 blocking (*UTF) etc.
513
514 36. In the interpreter, maximizing pattern repetitions for characters and
515 character types now use tail recursion, which reduces stack usage.
516
517 37. The value of the max lookbehind was not correctly preserved if a compiled
518 and saved regex was reloaded on a host of different endianness.
519
520 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
521 of the compiled pattern block, expand the flags field from 16 to 32 bits
522 because it was almost full.
523
524 39. Try madvise first before posix_madvise.
525
526 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
527 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
528 original change locked out all matches of empty strings. This has been
529 changed so that one match of an empty string per line is recognized.
530 Subsequent searches on the same line (for colouring or for --only-matching,
531 for example) do not recognize empty strings.
532
533 41. Applied a user patch to fix a number of spelling mistakes in comments.
534
535 42. Data lines longer than 65536 caused pcretest to crash.
536
537 43. Clarified the data type for length and startoffset arguments for pcre_exec
538 and pcre_dfa_exec in the function-specific man pages, where they were
539 explicitly stated to be in bytes, never having been updated. I also added
540 some clarification to the pcreapi man page.
541
542 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
543 a segmentation fault.
544
545
546 Version 8.32 30-November-2012
547 -----------------------------
548
549 1. Improved JIT compiler optimizations for first character search and single
550 character iterators.
551
552 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
553 Patch by Daniel Richard G.
554
555 3. Single character iterator optimizations in the JIT compiler.
556
557 4. Improved JIT compiler optimizations for character ranges.
558
559 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
560 Reported by Giuseppe D'Angelo.
561
562 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
563 of a line, was being set incorrectly in cases where .* appeared inside
564 atomic brackets at the start of a pattern, or where there was a subsequent
565 *PRUNE or *SKIP.
566
567 7. Improved instruction cache flush for POWER/PowerPC.
568 Patch by Daniel Richard G.
569
570 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
571 grep:
572
573 (a) There is now no limit to the number of patterns to be matched.
574
575 (b) An error is given if a pattern is too long.
576
577 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
578 are now supported.
579
580 (d) --exclude-from and --include-from (multiple use) have been added.
581
582 (e) Exclusions and inclusions now apply to all files and directories, not
583 just to those obtained from scanning a directory recursively.
584
585 (f) Multiple uses of -f and --file-list are now supported.
586
587 (g) In a Windows environment, the default for -d has been changed from
588 "read" (the GNU grep default) to "skip", because otherwise the presence
589 of a directory in the file list provokes an error.
590
591 (h) The documentation has been revised and clarified in places.
592
593 9. Improve the matching speed of capturing brackets.
594
595 10. Changed the meaning of \X so that it now matches a Unicode extended
596 grapheme cluster.
597
598 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
599 out POSIX threads when JIT support is configured.
600
601 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
602
603 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
604 structure could go wrong in environments where size_t is not the same size
605 as int.
606
607 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
608 set.
609
610 15. The EBCDIC support had decayed; later updates to the code had included
611 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
612 general tidy up of EBCDIC-related issues, and the documentation was also
613 not quite right. There is now a test that can be run on ASCII systems to
614 check some of the EBCDIC-related things (but is it not a full test).
615
616 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
617 in a small tidy to the code.
618
619 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
620
621 18. If the --only-matching (-o) option in pcregrep is specified multiple
622 times, each one causes appropriate output. For example, -o1 -o2 outputs the
623 substrings matched by the 1st and 2nd capturing parentheses. A separating
624 string can be specified by --om-separator (default empty).
625
626 19. Improving the first n character searches.
627
628 20. Turn case lists for horizontal and vertical white space into macros so that
629 they are defined only once.
630
631 21. This set of changes together give more compatible Unicode case-folding
632 behaviour for characters that have more than one other case when UCP
633 support is available.
634
635 (a) The Unicode property table now has offsets into a new table of sets of
636 three or more characters that are case-equivalent. The MultiStage2.py
637 script that generates these tables (the pcre_ucd.c file) now scans
638 CaseFolding.txt instead of UnicodeData.txt for character case
639 information.
640
641 (b) The code for adding characters or ranges of characters to a character
642 class has been abstracted into a generalized function that also handles
643 case-independence. In UTF-mode with UCP support, this uses the new data
644 to handle characters with more than one other case.
645
646 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
647 whose other case is greater than 256 are now correctly matched
648 caselessly. Previously, the high codepoint matched the low one, but not
649 vice versa.
650
651 (d) The processing of \h, \H, \v, and \ in character classes now makes use
652 of the new class addition function, using character lists defined as
653 macros alongside the case definitions of 20 above.
654
655 (e) Caseless back references now work with characters that have more than
656 one other case.
657
658 (f) General caseless matching of characters with more than one other case
659 is supported.
660
661 22. Unicode character properties were updated from Unicode 6.2.0
662
663 23. Improved CMake support under Windows. Patch by Daniel Richard G.
664
665 24. Add support for 32-bit character strings, and UTF-32
666
667 25. Major JIT compiler update (code refactoring and bugfixing).
668 Experimental Sparc 32 support is added.
669
670 26. Applied a modified version of Daniel Richard G's patch to create
671 pcre.h.generic and config.h.generic by "make" instead of in the
672 PrepareRelease script.
673
674 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
675 pcre_compile.c when checking for a zero character.
676
677 28. Introducing a native interface for JIT. Through this interface, the compiled
678 machine code can be directly executed. The purpose of this interface is to
679 provide fast pattern matching, so several sanity checks are not performed.
680 However, feature tests are still performed. The new interface provides
681 1.4x speedup compared to the old one.
682
683 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
684 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
685 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
686
687 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
688 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
689 the "old" RFC 2279). Instead, it ended up passing a negative length to
690 pcre_exec().
691
692 31. Add support for GCC's visibility feature to hide internal functions.
693
694 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
695 "unknown -C option" after outputting 0 or 1.
696
697 33. There is now support for generating a code coverage report for the test
698 suite in environments where gcc is the compiler and lcov is installed. This
699 is mainly for the benefit of the developers.
700
701 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
702 unaddressable using valgrind annotations, allowing valgrind to detect
703 invalid memory accesses. This is mainly for the benefit of the developers.
704
705 25. (*UTF) can now be used to start a pattern in any of the three libraries.
706
707 26. Give configure error if --enable-cpp but no C++ compiler found.
708
709
710 Version 8.31 06-July-2012
711 -------------------------
712
713 1. Fixing a wrong JIT test case and some compiler warnings.
714
715 2. Removed a bashism from the RunTest script.
716
717 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
718 to unsigned type, result still unsigned" that was given by an MS compiler
719 on encountering the code "-sizeof(xxx)".
720
721 4. Partial matching support is added to the JIT compiler.
722
723 5. Fixed several bugs concerned with partial matching of items that consist
724 of more than one character:
725
726 (a) /^(..)\1/ did not partially match "aba" because checking references was
727 done on an "all or nothing" basis. This also applied to repeated
728 references.
729
730 (b) \R did not give a hard partial match if \r was found at the end of the
731 subject.
732
733 (c) \X did not give a hard partial match after matching one or more
734 characters at the end of the subject.
735
736 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
737 a partial match for the string "\r".
738
739 (e) When newline was set to CRLF, the metacharacter "." did not recognize
740 a partial match for a CR character at the end of the subject string.
741
742 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
743 running pcretest, the text "(JIT)" added to the output whenever JIT is
744 actually used to run the match.
745
746 7. Individual JIT compile options can be set in pcretest by following -s+[+]
747 or /S+[+] with a digit between 1 and 7.
748
749 8. OP_NOT now supports any UTF character not just single-byte ones.
750
751 9. (*MARK) control verb is now supported by the JIT compiler.
752
753 10. The command "./RunTest list" lists the available tests without actually
754 running any of them. (Because I keep forgetting what they all are.)
755
756 11. Add PCRE_INFO_MAXLOOKBEHIND.
757
758 12. Applied a (slightly modified) user-supplied patch that improves performance
759 when the heap is used for recursion (compiled with --disable-stack-for-
760 recursion). Instead of malloc and free for each heap frame each time a
761 logical recursion happens, frames are retained on a chain and re-used where
762 possible. This sometimes gives as much as 30% improvement.
763
764 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
765 call.
766
767 14. As documented, (*COMMIT) is now confined to within a positive assertion.
768
769 15. It is now possible to link pcretest with libedit as an alternative to
770 libreadline.
771
772 16. (*COMMIT) control verb is now supported by the JIT compiler.
773
774 17. The Unicode data tables have been updated to Unicode 6.1.0.
775
776 18. Added --file-list option to pcregrep.
777
778 19. Added binary file support to pcregrep, including the -a, --binary-files,
779 -I, and --text options.
780
781 20. The madvise function is renamed for posix_madvise for QNX compatibility
782 reasons. Fixed by Giuseppe D'Angelo.
783
784 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
785
786 22. Changed the option for creating symbolic links for 16-bit man pages from
787 -s to -sf so that re-installing does not cause issues.
788
789 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
790
791 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
792 matches in certain environments (the workspace was not being correctly
793 retained). Also added to pcre_dfa_exec() a simple plausibility check on
794 some of the workspace data at the beginning of a restart.
795
796 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
797 was not doing so when it should - probably a typo introduced by SVN 528
798 (change 8.10/14).
799
800 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
801 \w+ when the character tables indicated that \x{c4} was a word character.
802 There were several related cases, all because the tests for doing a table
803 lookup were testing for characters less than 127 instead of 255.
804
805 27. If a pattern contains capturing parentheses that are not used in a match,
806 their slots in the ovector are set to -1. For those that are higher than
807 any matched groups, this happens at the end of processing. In the case when
808 there were back references that the ovector was too small to contain
809 (causing temporary malloc'd memory to be used during matching), and the
810 highest capturing number was not used, memory off the end of the ovector
811 was incorrectly being set to -1. (It was using the size of the temporary
812 memory instead of the true size.)
813
814 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
815 ovector size, it uses memory at the end of the block that it has got.
816
817 29. Check for an overlong MARK name and give an error at compile time. The
818 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
819
820 30. JIT compiler update.
821
822 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
823 Rill for the patch.
824
825 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
826
827 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
828
829 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
830 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
831 was enabled.
832
833 35. Improve JIT code generation for greedy plus quantifier.
834
835 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
836 1 to "aa" instead of to an empty string. The bug affected repeated groups
837 that could potentially match an empty string.
838
839 37. Optimizing single character iterators in JIT.
840
841 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
842 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
843 codepoints that are too big for the mode are faulted, and in a UTF mode,
844 disallowed codepoints are also faulted.
845
846 39. If PCRE was compiled with UTF support, in three places in the DFA
847 matcher there was code that should only have been obeyed in UTF mode, but
848 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
849 processing when bytes with values greater than 127 were present. In 16-bit
850 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
851 both cases the values are those that cannot be the first data item in a UTF
852 character. The three items that might have provoked this were recursions,
853 possessively repeated groups, and atomic groups.
854
855 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
856 and pcregrep, because some OS require shared objects to be explicitly
857 passed to ld, causing the link step to fail if they are not.
858
859 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
860 mode, patterns that started with \h* or \R* might be incorrectly matched.
861
862
863 Version 8.30 04-February-2012
864 -----------------------------
865
866 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
867 name is defined in ctype.h.
868
869 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
870 only in quite long subpatterns.
871
872 3. Removed the function pcre_info(), which has been obsolete and deprecated
873 since it was replaced by pcre_fullinfo() in February 2000.
874
875 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
876 match a (*MARK), and the match failed at the start of the subject, a
877 reference to memory before the start of the subject could occur. This bug
878 was introduced by fix 17 of release 8.21.
879
880 5. A reference to an unset group with zero minimum repetition was giving
881 totally wrong answers (in non-JavaScript-compatibility mode). For example,
882 /(another)?(\1?)test/ matched against "hello world test". This bug was
883 introduced in release 8.13.
884
885 6. Add support for 16-bit character strings (a large amount of work involving
886 many changes and refactorings).
887
888 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
889 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
890 from a file.
891
892 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
893 rounding is not applied in this particular case).
894
895 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
896 if they appear, or are escaped, in patterns.
897
898 10. Get rid of a number of -Wunused-but-set-variable warnings.
899
900 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
901 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
902 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
903 also returns the mark "x". This bug applied to capturing parentheses,
904 non-capturing parentheses, and atomic parentheses. It also applied to some
905 assertions.
906
907 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
908 information out of configure.ac instead of relying on pcre.h.generic, which
909 is not stored in the repository.
910
911 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
912 -lreadline.
913
914 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
915
916 15. Applied Graycode's patch to put the top-level frame on the stack rather
917 than the heap when not using the stack for recursion. This gives a
918 performance improvement in many cases when recursion is not deep.
919
920 16. Experimental code added to "pcretest -C" to output the stack frame size.
921
922
923 Version 8.21 12-Dec-2011
924 ------------------------
925
926 1. Updating the JIT compiler.
927
928 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
929 are added as well.
930
931 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
932 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
933 calling _pcre_jit_exec. Some extra comments are added.
934
935 4. (*MARK) settings inside atomic groups that do not contain any capturing
936 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
937 was introduced by change 18 for 8.20.
938
939 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
940 ECMA-262 standard.
941
942 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
943 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
944 This bug was probably introduced by change 9 of 8.13.
945
946 7. While fixing 6 above, I noticed that a number of other items were being
947 incorrectly rejected as "not fixed length". This arose partly because newer
948 opcodes had not been added to the fixed-length checking code. I have (a)
949 corrected the bug and added tests for these items, and (b) arranged for an
950 error to occur if an unknown opcode is encountered while checking for fixed
951 length instead of just assuming "not fixed length". The items that were
952 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
953 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
954 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
955
956 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
957 being incorrectly compiled and would have given unpredicatble results.
958
959 9. A possessively repeated subpattern with minimum repeat count greater than
960 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
961 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
962 the first (A) could occur when it should not.
963
964 10. Add a cast and remove a redundant test from the code.
965
966 11. JIT should use pcre_malloc/pcre_free for allocation.
967
968 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
969 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
970 is anything other than /usr, -L is still shown).
971
972 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
973
974 14. Perl does not support \N without a following name in a [] class; PCRE now
975 also gives an error.
976
977 15. If a forward reference was repeated with an upper limit of around 2000,
978 it caused the error "internal error: overran compiling workspace". The
979 maximum number of forward references (including repeats) was limited by the
980 internal workspace, and dependent on the LINK_SIZE. The code has been
981 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
982 the default depends on LINK_SIZE. There is a new upper limit (for safety)
983 of around 200,000 forward references. While doing this, I also speeded up
984 the filling in of repeated forward references.
985
986 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
987 incorrectly expecting the subject to contain another "a" after the start.
988
989 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
990 in the match, the SKIP should be ignored. This was not happening; instead
991 the SKIP was being treated as NOMATCH. For patterns such as
992 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
993 tested.
994
995 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
996 now much more compatible with Perl, in particular in cases where the result
997 is a non-match for a non-anchored pattern. For example, if
998 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
999 "m", where previously it did not return a name. A side effect of this
1000 change is that for partial matches, the last encountered mark name is
1001 returned, as for non matches. A number of tests that were previously not
1002 Perl-compatible have been moved into the Perl-compatible test files. The
1003 refactoring has had the pleasing side effect of removing one argument from
1004 the match() function, thus reducing its stack requirements.
1005
1006 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1007 subsequent uses of /S (without +) incorrectly behaved like /S+.
1008
1009 21. Retrieve executable code size support for the JIT compiler and fixing
1010 some warnings.
1011
1012 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1013 not work when the shorter character appeared right at the end of the
1014 subject string.
1015
1016 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1017 systems.
1018
1019 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1020 output it when the /M option is used in pcretest.
1021
1022 25. The CheckMan script was not being included in the distribution. Also, added
1023 an explicit "perl" to run Perl scripts from the PrepareRelease script
1024 because this is reportedly needed in Windows.
1025
1026 26. If study data was being save in a file and studying had not found a set of
1027 "starts with" bytes for the pattern, the data written to the file (though
1028 never used) was taken from uninitialized memory and so caused valgrind to
1029 complain.
1030
1031 27. Updated RunTest.bat as provided by Sheri Pierce.
1032
1033 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1034
1035 29. Computation of memory usage for the table of capturing group names was
1036 giving an unnecessarily large value.
1037
1038
1039 Version 8.20 21-Oct-2011
1040 ------------------------
1041
1042 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1043 a POSIX class. After further experiments with Perl, which convinced me that
1044 Perl has bugs and confusions, a closing square bracket is no longer allowed
1045 in a POSIX name. This bug also affected patterns with classes that started
1046 with full stops.
1047
1048 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1049 captured substring, but while checking the failing first alternative,
1050 substring 1 is temporarily captured. If the output vector supplied to
1051 pcre_exec() was not big enough for this capture, the yield of the function
1052 was still zero ("insufficient space for captured substrings"). This cannot
1053 be totally fixed without adding another stack variable, which seems a lot
1054 of expense for a edge case. However, I have improved the situation in cases
1055 such as /(a)(b)x|abc/ matched against "abc", where the return code
1056 indicates that fewer than the maximum number of slots in the ovector have
1057 been set.
1058
1059 3. Related to (2) above: when there are more back references in a pattern than
1060 slots in the output vector, pcre_exec() uses temporary memory during
1061 matching, and copies in the captures as far as possible afterwards. It was
1062 using the entire output vector, but this conflicts with the specification
1063 that only 2/3 is used for passing back captured substrings. Now it uses
1064 only the first 2/3, for compatibility. This is, of course, another edge
1065 case.
1066
1067 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1068 main code base, and can be used by building with --enable-jit. When this is
1069 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1070 runtime --no-jit option is given.
1071
1072 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1073 ovector, the return from the function was zero, implying that there were
1074 other matches that did not fit. The correct "exactly full" value is now
1075 returned.
1076
1077 6. If a subpattern that was called recursively or as a subroutine contained
1078 (*PRUNE) or any other control that caused it to give a non-standard return,
1079 invalid errors such as "Error -26 (nested recursion at the same subject
1080 position)" or even infinite loops could occur.
1081
1082 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1083 computing the minimum length on reaching *ACCEPT, and so ended up with the
1084 wrong value of 1 rather than 0. Further investigation indicates that
1085 computing a minimum subject length in the presence of *ACCEPT is difficult
1086 (think back references, subroutine calls), and so I have changed the code
1087 so that no minimum is registered for a pattern that contains *ACCEPT.
1088
1089 8. If (*THEN) was present in the first (true) branch of a conditional group,
1090 it was not handled as intended. [But see 16 below.]
1091
1092 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1093 Sheri Pierce.
1094
1095 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1096 the first byte in a match must be "a".
1097
1098 11. Change 17 for 8.13 increased the recursion depth for patterns like
1099 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1100 pattern contains any instances of (*THEN). If it does not, the old
1101 optimizations are restored. It would be nice to do this on a per-group
1102 basis, but at the moment that is not feasible.
1103
1104 12. In some environments, the output of pcretest -C is CRLF terminated. This
1105 broke RunTest's code that checks for the link size. A single white space
1106 character after the value is now allowed for.
1107
1108 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1109 For "fr", it uses the Windows-specific input and output files.
1110
1111 14. If (*THEN) appeared in a group that was called recursively or as a
1112 subroutine, it did not work as intended. [But see next item.]
1113
1114 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1115 pattern fragments (but not containing any | characters). If A and B are
1116 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1117 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1118 D. In other words, Perl considers parentheses that do not contain any |
1119 characters to be part of a surrounding alternative, whereas PCRE was
1120 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1121 differently. PCRE now behaves in the same way as Perl, except in the case
1122 of subroutine/recursion calls such as (?1) which have in any case always
1123 been different (but PCRE had them first :-).
1124
1125 16. Related to 15 above: Perl does not treat the | in a conditional group as
1126 creating alternatives. Such a group is treated in the same way as an
1127 ordinary group without any | characters when processing (*THEN). PCRE has
1128 been changed to match Perl's behaviour.
1129
1130 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1131 RunGrepTest script failed.
1132
1133 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1134 inevitable for groups that contain captures, but it can lead to a lot of
1135 stack use in large patterns. The old behaviour has been restored for atomic
1136 groups that do not contain any capturing parentheses.
1137
1138 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1139 suppress the check for a minimum subject length at run time. (If it was
1140 given to pcre_exec() or pcre_dfa_exec() it did work.)
1141
1142 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1143 fail to work when decoding hex characters in data strings in EBCDIC
1144 environments.
1145
1146 21. It appears that in at least one Mac OS environment, the isxdigit() function
1147 is implemented as a macro that evaluates to its argument more than once,
1148 contravening the C 90 Standard (I haven't checked a later standard). There
1149 was an instance in pcretest which caused it to go wrong when processing
1150 \x{...} escapes in subject strings. The has been rewritten to avoid using
1151 things like p++ in the argument of isxdigit().
1152
1153
1154 Version 8.13 16-Aug-2011
1155 ------------------------
1156
1157 1. The Unicode data tables have been updated to Unicode 6.0.0.
1158
1159 2. Two minor typos in pcre_internal.h have been fixed.
1160
1161 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1162 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1163 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1164
1165 4. There were a number of related bugs in the code for matching backrefences
1166 caselessly in UTF-8 mode when codes for the characters concerned were
1167 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1168 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1169 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1170 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1171 2-byte code at the end of the subject (it thought there wasn't enough data
1172 left).
1173
1174 5. Comprehensive information about what went wrong is now returned by
1175 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1176 as the output vector has at least 2 elements. The offset of the start of
1177 the failing character and a reason code are placed in the vector.
1178
1179 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1180 now returned is for the first byte of the failing character, instead of the
1181 last byte inspected. This is an incompatible change, but I hope it is small
1182 enough not to be a problem. It makes the returned offset consistent with
1183 pcre_exec() and pcre_dfa_exec().
1184
1185 7. pcretest now gives a text phrase as well as the error number when
1186 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1187 failure, the offset and reason code are output.
1188
1189 8. When \R was used with a maximizing quantifier it failed to skip backwards
1190 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1191 back over a single character (\n). This seems wrong (because it treated the
1192 two characters as a single entity when going forwards), conflicts with the
1193 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1194 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1195 has been changed.
1196
1197 9. Some internal refactoring has changed the processing so that the handling
1198 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1199 time (the PCRE_DOTALL option was changed this way some time ago: version
1200 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1201 which was always a bit of a fudge. It also means that there is one less
1202 argument for the match() function, which reduces its stack requirements
1203 slightly. This change also fixes an incompatibility with Perl: the pattern
1204 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1205
1206 10. More internal refactoring has drastically reduced the number of recursive
1207 calls to match() for possessively repeated groups such as (abc)++ when
1208 using pcre_exec().
1209
1210 11. While implementing 10, a number of bugs in the handling of groups were
1211 discovered and fixed:
1212
1213 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1214 (a|)*(?1) gave a compile-time internal error.
1215 ((a|)+)+ did not notice that the outer group could match an empty string.
1216 (^a|^)+ was not marked as anchored.
1217 (.*a|.*)+ was not marked as matching at start or after a newline.
1218
1219 12. Yet more internal refactoring has removed another argument from the match()
1220 function. Special calls to this function are now indicated by setting a
1221 value in a variable in the "match data" data block.
1222
1223 13. Be more explicit in pcre_study() instead of relying on "default" for
1224 opcodes that mean there is no starting character; this means that when new
1225 ones are added and accidentally left out of pcre_study(), testing should
1226 pick them up.
1227
1228 14. The -s option of pcretest has been documented for ages as being an old
1229 synonym of -m (show memory usage). I have changed it to mean "force study
1230 for every regex", that is, assume /S for every regex. This is similar to -i
1231 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1232 using it. It makes it easier to run collections of tests with and without
1233 study enabled, and thereby test pcre_study() more easily. All the standard
1234 tests are now run with and without -s (but some patterns can be marked as
1235 "never study" - see 20 below).
1236
1237 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1238 restoration of the capturing data to the outer values was not happening
1239 correctly.
1240
1241 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1242 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1243 pattern had matched an empty string, and so incorrectly returned a no
1244 match.
1245
1246 17. There was optimizing code for the last branch of non-capturing parentheses,
1247 and also for the obeyed branch of a conditional subexpression, which used
1248 tail recursion to cut down on stack usage. Unfortunately, now that there is
1249 the possibility of (*THEN) occurring in these branches, tail recursion is
1250 no longer possible because the return has to be checked for (*THEN). These
1251 two optimizations have therefore been removed. [But see 8.20/11 above.]
1252
1253 18. If a pattern containing \R was studied, it was assumed that \R always
1254 matched two bytes, thus causing the minimum subject length to be
1255 incorrectly computed because \R can also match just one byte.
1256
1257 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1258 was incorrectly computed.
1259
1260 20. If /S is present twice on a test pattern in pcretest input, it now
1261 *disables* studying, thereby overriding the use of -s on the command line
1262 (see 14 above). This is necessary for one or two tests to keep the output
1263 identical in both cases.
1264
1265 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1266 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1267
1268 22. When an atomic group that contained a capturing parenthesis was
1269 successfully matched, but the branch in which it appeared failed, the
1270 capturing was not being forgotten if a higher numbered group was later
1271 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1272 group 1 to "a", when in fact it should be unset. This applied to multi-
1273 branched capturing and non-capturing groups, repeated or not, and also to
1274 positive assertions (capturing in negative assertions does not happen
1275 in PCRE) and also to nested atomic groups.
1276
1277 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1278 subject after a captured substring, to make it easier to tell which of a
1279 number of identical substrings has been captured.
1280
1281 24. The way atomic groups are processed by pcre_exec() has been changed so that
1282 if they are repeated, backtracking one repetition now resets captured
1283 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1284 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1285 "aaa". Previously, it would have been "a". As part of this code
1286 refactoring, the way recursive calls are handled has also been changed.
1287
1288 25. If an assertion condition captured any substrings, they were not passed
1289 back unless some other capturing happened later. For example, if
1290 (?(?=(a))a) was matched against "a", no capturing was returned.
1291
1292 26. When studying a pattern that contained subroutine calls or assertions,
1293 the code for finding the minimum length of a possible match was handling
1294 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1295 group 1 called group 2 while simultaneously a separate group 2 called group
1296 1). A stack overflow occurred in this case. I have fixed this by limiting
1297 the recursion depth to 10.
1298
1299 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1300 Fortmann. This supports explicit test numbers on the command line, and has
1301 argument validation and error reporting.
1302
1303 28. An instance of \X with an unlimited repeat could fail if at any point the
1304 first character it looked at was a mark character.
1305
1306 29. Some minor code refactoring concerning Unicode properties and scripts
1307 should reduce the stack requirement of match() slightly.
1308
1309 30. Added the '=' option to pcretest to check the setting of unused capturing
1310 slots at the end of the pattern, which are documented as being -1, but are
1311 not included in the return count.
1312
1313 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1314 compiled something random. Now it gives a compile-time error (as does
1315 Perl).
1316
1317 32. A *MARK encountered during the processing of a positive assertion is now
1318 recorded and passed back (compatible with Perl).
1319
1320 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1321 had alternative anchored branches, the search for a second match in a line
1322 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1323 matched the line "0102" twice. The same bug affected patterns that started
1324 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1325 twice.
1326
1327 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1328 does, and because of capturing effects, quantifying parenthesized
1329 assertions may at times be useful. Quantifiers are now allowed for
1330 parenthesized assertions.
1331
1332 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1333
1334 36. \g was being checked for fancy things in a character class, when it should
1335 just be a literal "g".
1336
1337 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1338 appearance of a nested POSIX class supersedes an apparent external class.
1339 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1340 unescaped square brackets may also appear as part of class names. For
1341 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1342 more like Perl. (But see 8.20/1 above.)
1343
1344 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1345 was because it thought it was \N{name}, which is not supported).
1346
1347 39. Add minix to OS list not supporting the -S option in pcretest.
1348
1349 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1350 cannot analyze patterns in sufficient detail to catch mutual recursions
1351 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1352 subgroup is called recursively as a subpattern for a second time at the
1353 same position in the subject string. In previous releases this might have
1354 been caught by the recursion limit, or it might have run out of stack.
1355
1356 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1357 happen only once. PCRE was, however incorrectly giving a compile time error
1358 "recursive call could loop indefinitely" because it cannot analyze the
1359 pattern in sufficient detail. The compile time test no longer happens when
1360 PCRE is compiling a conditional subpattern, but actual runaway loops are
1361 now caught at runtime (see 40 above).
1362
1363 42. It seems that Perl allows any characters other than a closing parenthesis
1364 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1365 has been changed to be the same.
1366
1367 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1368 as not to get warnings when autogen.sh is called. Also changed
1369 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1370
1371 44. To help people who use pcregrep to scan files containing exceedingly long
1372 lines, the following changes have been made:
1373
1374 (a) The default value of the buffer size parameter has been increased from
1375 8K to 20K. (The actual buffer used is three times this size.)
1376
1377 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1378 PCRE is built.
1379
1380 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1381 to be set at run time.
1382
1383 (d) Numerical values in pcregrep options can be followed by K or M, for
1384 example --buffer-size=50K.
1385
1386 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1387 given and the return code is set to 2.
1388
1389 45. Add a pointer to the latest mark to the callout data block.
1390
1391 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1392 partial match of an empty string instead of no match. This was specific to
1393 the use of ".".
1394
1395 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1396 complete match instead of a partial match. This bug was dependent on both
1397 the PCRE_UTF8 and PCRE_DOTALL options being set.
1398
1399 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1400 starting byte set, because \b was not being ignored.
1401
1402
1403 Version 8.12 15-Jan-2011
1404 ------------------------
1405
1406 1. Fixed some typos in the markup of the man pages, and wrote a script that
1407 checks for such things as part of the documentation building process.
1408
1409 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1410 --match-limit and --recursion-limit options (added for 8.11). In
1411 particular, this made one of the standard tests fail. (The integer value
1412 went into the wrong half of a long int.)
1413
1414 3. If the --colour option was given to pcregrep with -v (invert match), it
1415 did strange things, either producing crazy output, or crashing. It should,
1416 of course, ignore a request for colour when reporting lines that do not
1417 match.
1418
1419 4. Another pcregrep bug caused similar problems if --colour was specified with
1420 -M (multiline) and the pattern match finished with a line ending.
1421
1422 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1423 matched in multiline mode, the following line was shown as part of the
1424 match. This seems wrong, so I have changed it.
1425
1426 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1427 the check for further matches in the same line (so they could be coloured)
1428 to overrun the end of the current line. If another match was found, it was
1429 incorrectly shown (and then shown again when found in the next line).
1430
1431 7. If pcregrep was compiled under Windows, there was a reference to the
1432 function pcregrep_exit() before it was defined. I am assuming this was
1433 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1434 reported by a user. I've moved the definition above the reference.
1435
1436
1437 Version 8.11 10-Dec-2010
1438 ------------------------
1439
1440 1. (*THEN) was not working properly if there were untried alternatives prior
1441 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1442 backtracked to try for "b" instead of moving to the next alternative branch
1443 at the same level (in this case, to look for "c"). The Perl documentation
1444 is clear that when (*THEN) is backtracked onto, it goes to the "next
1445 alternative in the innermost enclosing group".
1446
1447 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1448 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1449 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1450 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1451 (*THEN).
1452
1453 3. If \s appeared in a character class, it removed the VT character from
1454 the class, even if it had been included by some previous item, for example
1455 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1456 of \s, but is part of the POSIX "space" class.)
1457
1458 4. A partial match never returns an empty string (because you can always
1459 match an empty string at the end of the subject); however the checking for
1460 an empty string was starting at the "start of match" point. This has been
1461 changed to the "earliest inspected character" point, because the returned
1462 data for a partial match starts at this character. This means that, for
1463 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1464 (previously it gave "no match").
1465
1466 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1467 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1468 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1469 has an implication that the given string is incomplete (because a partial
1470 match is preferred over a full match). For this reason, these items now
1471 give a partial match in this situation. [Aside: previously, the one case
1472 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1473 match rather than a full match, which was wrong by the old rules, but is
1474 now correct.]
1475
1476 6. There was a bug in the handling of #-introduced comments, recognized when
1477 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1478 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1479 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1480 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1481 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1482 places in pcre_compile().
1483
1484 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1485 comments when looking ahead for named forward references to subpatterns,
1486 the only newline sequence it recognized was NL. It now handles newlines
1487 according to the set newline convention.
1488
1489 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1490 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1491 cater for a lack of strerror(). These oversights have been fixed.
1492
1493 9. Added --match-limit and --recursion-limit to pcregrep.
1494
1495 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1496
1497 11. When the -o option was used, pcregrep was setting a return code of 1, even
1498 when matches were found, and --line-buffered was not being honoured.
1499
1500 12. Added an optional parentheses number to the -o and --only-matching options
1501 of pcregrep.
1502
1503 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1504 can match an empty string. The code to do it in pcretest and pcredemo
1505 needed fixing:
1506
1507 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1508 only one byte after an empty string match just before CRLF (this case
1509 just got forgotten; "any" and "anycrlf" were OK).
1510
1511 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1512 mode when an empty string match preceded an ASCII character followed by
1513 a non-ASCII character. (The code for advancing by one character rather
1514 than one byte was nonsense.)
1515
1516 (c) The pcredemo.c sample program did not have any code at all to handle
1517 the cases when CRLF is a valid newline sequence.
1518
1519 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1520 as a starting offset was within the subject string. There is now a new
1521 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1522 negative or greater than the length of the string. In order to test this,
1523 pcretest is extended to allow the setting of negative starting offsets.
1524
1525 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1526 starting offset points to the beginning of a UTF-8 character was
1527 unnecessarily clumsy. I tidied it up.
1528
1529 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1530 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1531
1532 17. Nobody had reported that the --include_dir option, which was added in
1533 release 7.7 should have been called --include-dir (hyphen, not underscore)
1534 for compatibility with GNU grep. I have changed it to --include-dir, but
1535 left --include_dir as an undocumented synonym, and the same for
1536 --exclude-dir, though that is not available in GNU grep, at least as of
1537 release 2.5.4.
1538
1539 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1540 characters from a string of bytes) have been redefined so as not to use
1541 loops, in order to improve performance in some environments. At the same
1542 time, I abstracted some of the common code into auxiliary macros to save
1543 repetition (this should not affect the compiled code).
1544
1545 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1546 compile-time error is now given if \c is not followed by an ASCII
1547 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1548 different, and any byte value is allowed.)
1549
1550 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1551 START_OPTIMIZE option, which is now allowed at compile time - but just
1552 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1553 to pcregrep and other applications that have no direct access to PCRE
1554 options. The new /Y option in pcretest sets this option when calling
1555 pcre_compile().
1556
1557 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1558 back references. Groups containing recursive back references were forced to
1559 be atomic by that change, but in the case of named groups, the amount of
1560 memory required was incorrectly computed, leading to "Failed: internal
1561 error: code overflow". This has been fixed.
1562
1563 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1564 pcretest.c, to avoid build problems in some Borland environments.
1565
1566
1567 Version 8.10 25-Jun-2010
1568 ------------------------
1569
1570 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1571 THEN.
1572
1573 2. (*ACCEPT) was not working when inside an atomic group.
1574
1575 3. Inside a character class, \B is treated as a literal by default, but
1576 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1577 causes the error). The code is unchanged, but I tidied the documentation.
1578
1579 4. Inside a character class, PCRE always treated \R and \X as literals,
1580 whereas Perl faults them if its -w option is set. I have changed PCRE so
1581 that it faults them when PCRE_EXTRA is set.
1582
1583 5. Added support for \N, which always matches any character other than
1584 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1585
1586 6. When compiling pcregrep with newer versions of gcc which may have
1587 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1588 declared with attribute warn_unused_result" were given. Just casting the
1589 result to (void) does not stop the warnings; a more elaborate fudge is
1590 needed. I've used a macro to implement this.
1591
1592 7. Minor change to pcretest.c to avoid a compiler warning.
1593
1594 8. Added four artifical Unicode properties to help with an option to make
1595 \s etc use properties (see next item). The new properties are: Xan
1596 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1597
1598 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1599 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1600 this option. Modified pcretest to add /W to test this facility. Added
1601 REG_UCP to make it available via the POSIX interface.
1602
1603 10. Added --line-buffered to pcregrep.
1604
1605 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1606 studied, and the match started with a letter with a code point greater than
1607 127 whose first byte was different to the first byte of the other case of
1608 the letter, the other case of this starting letter was not recognized
1609 (#976).
1610
1611 12. If a pattern that was studied started with a repeated Unicode property
1612 test, for example, \p{Nd}+, there was the theoretical possibility of
1613 setting up an incorrect bitmap of starting bytes, but fortunately it could
1614 not have actually happened in practice until change 8 above was made (it
1615 added property types that matched character-matching opcodes).
1616
1617 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1618 possible starting bytes for non-anchored patterns.
1619
1620 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1621 \R, and also a number of cases that involve Unicode properties, both
1622 explicit and implicit when PCRE_UCP is set.
1623
1624 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1625 input, it could crash or give wrong results if characters with values
1626 greater than 0xc0 were present in the subject string. (Detail: it assumed
1627 UTF-8 input when processing these items.)
1628
1629 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1630 size_t is 64-bit (#991).
1631
1632 17. Added a check for running out of memory when PCRE is compiled with
1633 --disable-stack-for-recursion (#990).
1634
1635 18. If the last data line in a file for pcretest does not have a newline on
1636 the end, a newline was missing in the output.
1637
1638 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1639 less than 128) in its various bitmaps. However, there is a facility for
1640 generating tables according to the current locale when PCRE is compiled. It
1641 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1642 characters, are recognized by isspace() and therefore were getting set in
1643 these tables, and indeed these tables seem to approximate to ISO 8859. This
1644 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1645 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1646 which of course cannot start UTF-8 characters. I have changed the code so
1647 that only real ASCII characters (less than 128) and the correct starting
1648 bytes for UTF-8 encodings are set for characters greater than 127 when in
1649 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1650 altogether.)
1651
1652 20. Added the /T option to pcretest so as to be able to run tests with non-
1653 standard character tables, thus making it possible to include the tests
1654 used for 19 above in the standard set of tests.
1655
1656 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1657 reference to a subpattern the other side of a comment that contains an
1658 opening parenthesis caused either an internal compiling error, or a
1659 reference to the wrong subpattern.
1660
1661
1662 Version 8.02 19-Mar-2010
1663 ------------------------
1664
1665 1. The Unicode data tables have been updated to Unicode 5.2.0.
1666
1667 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1668 configured.
1669
1670 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1671 original author of that file, following a query about its status.
1672
1673 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1674 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1675
1676 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1677 quantifier applied to a forward-referencing subroutine call, could compile
1678 incorrect code or give the error "internal error: previously-checked
1679 referenced subpattern not found".
1680
1681 6. Both MS Visual Studio and Symbian OS have problems with initializing
1682 variables to point to external functions. For these systems, therefore,
1683 pcre_malloc etc. are now initialized to local functions that call the
1684 relevant global functions.
1685
1686 7. There were two entries missing in the vectors called coptable and poptable
1687 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1688 I've fixed the data, and added a kludgy way of testing at compile time that
1689 the lengths are correct (equal to the number of opcodes).
1690
1691 8. Following on from 7, I added a similar kludge to check the length of the
1692 eint vector in pcreposix.c.
1693
1694 9. Error texts for pcre_compile() are held as one long string to avoid too
1695 much relocation at load time. To find a text, the string is searched,
1696 counting zeros. There was no check for running off the end of the string,
1697 which could happen if a new error number was added without updating the
1698 string.
1699
1700 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1701
1702 11. \K was not working if it appeared in an atomic group or in a group that
1703 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1704 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1705 the assertion is positive, but not if it is negative.
1706
1707 12. Change 11 fortuitously reduced the size of the stack frame used in the
1708 "match()" function of pcre_exec.c by one pointer. Forthcoming
1709 implementation of support for (*MARK) will need an extra pointer on the
1710 stack; I have reserved it now, so that the stack frame size does not
1711 decrease.
1712
1713 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1714 item in branch that calls a recursion is a subroutine call - as in the
1715 second branch in the above example - was incorrectly given the compile-
1716 time error "recursive call could loop indefinitely" because pcre_compile()
1717 was not correctly checking the subroutine for matching a non-empty string.
1718
1719 14. The checks for overrunning compiling workspace could trigger after an
1720 overrun had occurred. This is a "should never occur" error, but it can be
1721 triggered by pathological patterns such as hundreds of nested parentheses.
1722 The checks now trigger 100 bytes before the end of the workspace.
1723
1724 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1725
1726
1727 Version 8.01 19-Jan-2010
1728 ------------------------
1729
1730 1. If a pattern contained a conditional subpattern with only one branch (in
1731 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1732 computed the wrong minimum data length (which is of course zero for such
1733 subpatterns). This could cause incorrect "no match" results.
1734
1735 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1736 the pattern is reset in the first branch, pcre_compile() failed with
1737 "internal error: code overflow at offset...". This happened only when
1738 the reset was to the original external option setting. (An optimization
1739 abstracts leading options settings into an external setting, which was the
1740 cause of this.)
1741
1742 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1743 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1744 assertion pattern did not match (meaning that the assertion was true), it
1745 was incorrectly treated as false if the SKIP had been reached during the
1746 matching. This also applied to assertions used as conditions.
1747
1748 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1749 assertion subpattern, including such a pattern used as a condition,
1750 unpredictable results occurred, instead of the error return
1751 PCRE_ERROR_DFA_UITEM.
1752
1753 5. The C++ GlobalReplace function was not working like Perl for the special
1754 situation when an empty string is matched. It now does the fancy magic
1755 stuff that is necessary.
1756
1757 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1758 removed. (These were left over from very, very early versions of PCRE.)
1759
1760 7. Some cosmetic changes to the code to make life easier when compiling it
1761 as part of something else:
1762
1763 (a) Change DEBUG to PCRE_DEBUG.
1764
1765 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1766 called "current" as "current_branch", to prevent a collision with the
1767 Linux macro when compiled as a kernel module.
1768
1769 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1770 prevent a collision with the Linux macro when compiled as a kernel
1771 module.
1772
1773 8. In pcre_compile() there are some checks for integer overflows that used to
1774 cast potentially large values to (double). This has been changed to that
1775 when building, a check for int64_t is made, and if it is found, it is used
1776 instead, thus avoiding the use of floating point arithmetic. (There is no
1777 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1778 double.
1779
1780 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1781 2005 (difference between two addresses compared to an unsigned value).
1782
1783 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1784 custom one, because of the following reported problem in Windows:
1785
1786 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1787 under Win32.
1788 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1789 therefore missing the function definition.
1790 - The compiler thus generates a "C" signature for the test function.
1791 - The linker fails to find the "C" function.
1792 - PCRE fails to configure if asked to do so against libbz2.
1793
1794 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1795 messages were output:
1796
1797 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1798 rerunning libtoolize, to keep the correct libtool macros in-tree.
1799 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1800
1801 I have done both of these things.
1802
1803 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1804 most of the time, it *can* run out if it is given a pattern that contains a
1805 runaway infinite recursion. I updated the discussion in the pcrestack man
1806 page.
1807
1808 13. Now that we have gone to the x.xx style of version numbers, the minor
1809 version may start with zero. Using 08 or 09 is a bad idea because users
1810 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1811 interpreted as invalid octal numbers. I've updated the previous comment in
1812 configure.ac, and also added a check that gives an error if 08 or 09 are
1813 used.
1814
1815 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1816 causing partial matching to fail when the end of the subject matched \W
1817 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1818
1819 15. There were some discrepancies between the declarations in pcre_internal.h
1820 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1821 their definitions. The declarations used "const uschar *" and the
1822 definitions used USPTR. Even though USPTR is normally defined as "const
1823 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1824 reported that: "This difference in casting confuses some C++ compilers, for
1825 example, SunCC recognizes above declarations as different functions and
1826 generates broken code for hbpcre." I have changed the declarations to use
1827 USPTR.
1828
1829 16. GNU libtool is named differently on some systems. The autogen.sh script now
1830 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1831 (FreeBSD).
1832
1833 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1834 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1835 comment: "Figure out how to create a longlong from a string: strtoll and
1836 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1837 instance, but it only takes 2 args instead of 3!"
1838
1839 18. A subtle bug concerned with back references has been fixed by a change of
1840 specification, with a corresponding code fix. A pattern such as
1841 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1842 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1843 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1844 same bug. Such groups have to be quantified to be useful, or contained
1845 inside another quantified group. (If there's no repetition, the reference
1846 can never match.) The problem arises because, having left the group and
1847 moved on to the rest of the pattern, a later failure that backtracks into
1848 the group uses the captured value from the final iteration of the group
1849 rather than the correct earlier one. I have fixed this in PCRE by forcing
1850 any group that contains a reference to itself to be an atomic group; that
1851 is, there cannot be any backtracking into it once it has completed. This is
1852 similar to recursive and subroutine calls.
1853
1854
1855 Version 8.00 19-Oct-09
1856 ----------------------
1857
1858 1. The table for translating pcre_compile() error codes into POSIX error codes
1859 was out-of-date, and there was no check on the pcre_compile() error code
1860 being within the table. This could lead to an OK return being given in
1861 error.
1862
1863 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1864 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1865 in a Windows environment.
1866
1867 3. The pcregrep --count option prints the count for each file even when it is
1868 zero, as does GNU grep. However, pcregrep was also printing all files when
1869 --files-with-matches was added. Now, when both options are given, it prints
1870 counts only for those files that have at least one match. (GNU grep just
1871 prints the file name in this circumstance, but including the count seems
1872 more useful - otherwise, why use --count?) Also ensured that the
1873 combination -clh just lists non-zero counts, with no names.
1874
1875 4. The long form of the pcregrep -F option was incorrectly implemented as
1876 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1877 but it seems right to fix it, and I didn't think it was worth preserving
1878 the old behaviour.
1879
1880 5. The command line items --regex=pattern and --regexp=pattern were not
1881 recognized by pcregrep, which required --regex pattern or --regexp pattern
1882 (with a space rather than an '='). The man page documented the '=' forms,
1883 which are compatible with GNU grep; these now work.
1884
1885 6. No libpcreposix.pc file was created for pkg-config; there was just
1886 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1887
1888 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1889 when UCP support is not needed, by modifying the Python script that
1890 generates it from Unicode data files. This should not matter if the module
1891 is correctly used as a library, but I received one complaint about 50K of
1892 unwanted data. My guess is that the person linked everything into his
1893 program rather than using a library. Anyway, it does no harm.
1894
1895 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1896 was a minimum greater than 1 for a wide character in a possessive
1897 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1898 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1899 character. Chaos in the form of incorrect output or a compiling loop could
1900 result.
1901
1902 9. The restrictions on what a pattern can contain when partial matching is
1903 requested for pcre_exec() have been removed. All patterns can now be
1904 partially matched by this function. In addition, if there are at least two
1905 slots in the offset vector, the offset of the earliest inspected character
1906 for the match and the offset of the end of the subject are set in them when
1907 PCRE_ERROR_PARTIAL is returned.
1908
1909 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1910 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1911 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1912 and may be more useful for multi-segment matching.
1913
1914 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1915 used to be given if ever the end of the subject was reached; now it is
1916 given only if matching could not proceed because another character was
1917 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1918 string "Z", which now yields "no match" instead of "partial match". In the
1919 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1920 final character ended with (*FAIL).
1921
1922 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1923 if the pattern had a "must contain" character that was already found in the
1924 earlier partial match, unless partial matching was again requested. For
1925 example, with the pattern /dog.(body)?/, the "must contain" character is
1926 "g". If the first part-match was for the string "dog", restarting with
1927 "sbody" failed. This bug has been fixed.
1928
1929 13. The string returned by pcre_dfa_exec() after a partial match has been
1930 changed so that it starts at the first inspected character rather than the
1931 first character of the match. This makes a difference only if the pattern
1932 starts with a lookbehind assertion or \b or \B (\K is not supported by
1933 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1934 matching functions compatible, and I think it's the right thing to do.
1935
1936 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1937 so that the demonstration program is easily available in environments where
1938 PCRE has not been installed from source.
1939
1940 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1941 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1942 library.
1943
1944 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1945 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1946 is not the first non-POSIX option to be added. Clearly some people find
1947 these options useful.
1948
1949 17. If a caller to the POSIX matching function regexec() passes a non-zero
1950 value for nmatch with a NULL value for pmatch, the value of
1951 nmatch is forced to zero.
1952
1953 18. RunGrepTest did not have a test for the availability of the -u option of
1954 the diff command, as RunTest does. It now checks in the same way as
1955 RunTest, and also checks for the -b option.
1956
1957 19. If an odd number of negated classes containing just a single character
1958 interposed, within parentheses, between a forward reference to a named
1959 subpattern and the definition of the subpattern, compilation crashed with
1960 an internal error, complaining that it could not find the referenced
1961 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1962 [The bug was that it was starting one character too far in when skipping
1963 over the character class, thus treating the ] as data rather than
1964 terminating the class. This meant it could skip too much.]
1965
1966 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1967 /g option in pcretest when the pattern contains \K, which makes it possible
1968 to have an empty string match not at the start, even when the pattern is
1969 anchored. Updated pcretest and pcredemo to use this option.
1970
1971 21. If the maximum number of capturing subpatterns in a recursion was greater
1972 than the maximum at the outer level, the higher number was returned, but
1973 with unset values at the outer level. The correct (outer level) value is
1974 now given.
1975
1976 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1977 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1978 make it do so. The string so far is captured, making this feature
1979 compatible with Perl.
1980
1981 23. The tests have been re-organized, adding tests 11 and 12, to make it
1982 possible to check the Perl 5.10 features against Perl 5.10.
1983
1984 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1985 pattern matches a fixed length string. PCRE did not allow this; now it
1986 does. Neither allows recursion.
1987
1988 25. I finally figured out how to implement a request to provide the minimum
1989 length of subject string that was needed in order to match a given pattern.
1990 (It was back references and recursion that I had previously got hung up
1991 on.) This code has now been added to pcre_study(); it finds a lower bound
1992 to the length of subject needed. It is not necessarily the greatest lower
1993 bound, but using it to avoid searching strings that are too short does give
1994 some useful speed-ups. The value is available to calling programs via
1995 pcre_fullinfo().
1996
1997 26. While implementing 25, I discovered to my embarrassment that pcretest had
1998 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1999 study optimizations had never been tested with that matching function.
2000 Oops. What is worse, even when it was passed study data, there was a bug in
2001 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2002 were also very few tests of studied patterns with pcre_dfa_exec().
2003
2004 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2005 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2006 on the other side of the coin, they are no longer allowed to have different
2007 names, because these cannot be distinguished in PCRE, and this has caused
2008 confusion. (This is a difference from Perl.)
2009
2010 28. When duplicate subpattern names are present (necessarily with different
2011 numbers, as required by 27 above), and a test is made by name in a
2012 conditional pattern, either for a subpattern having been matched, or for
2013 recursion in such a pattern, all the associated numbered subpatterns are
2014 tested, and the overall condition is true if the condition is true for any
2015 one of them. This is the way Perl works, and is also more like the way
2016 testing by number works.
2017
2018
2019 Version 7.9 11-Apr-09
2020 ---------------------
2021
2022 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2023 (pcretest), all targets were linked against these libraries. This included
2024 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2025 libraries. This caused unwanted dependencies to be created. This problem
2026 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2027 pcretest is linked with readline.
2028
2029 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2030 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2031 moved outside it again, because FALSE and TRUE are already defined in AIX,
2032 but BOOL is not.
2033
2034 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2035 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2036
2037 4. The pcregrep documentation said spaces were inserted as well as colons (or
2038 hyphens) following file names and line numbers when outputting matching
2039 lines. This is not true; no spaces are inserted. I have also clarified the
2040 wording for the --colour (or --color) option.
2041
2042 5. In pcregrep, when --colour was used with -o, the list of matching strings
2043 was not coloured; this is different to GNU grep, so I have changed it to be
2044 the same.
2045
2046 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2047 each matching line was coloured. Now it goes on to look for further matches
2048 of any of the test patterns, which is the same behaviour as GNU grep.
2049
2050 7. A pattern that could match an empty string could cause pcregrep to loop; it
2051 doesn't make sense to accept an empty string match in pcregrep, so I have
2052 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2053 seems to be how GNU grep behaves. [But see later change 40 for release
2054 8.33.]
2055
2056 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2057 start or after a newline", because the conditional assertion was not being
2058 correctly handled. The rule now is that both the assertion and what follows
2059 in the first alternative must satisfy the test.
2060
2061 9. If auto-callout was enabled in a pattern with a conditional group whose
2062 condition was an assertion, PCRE could crash during matching, both with
2063 pcre_exec() and pcre_dfa_exec().
2064
2065 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2066 used for matching.
2067
2068 11. Unicode property support in character classes was not working for
2069 characters (bytes) greater than 127 when not in UTF-8 mode.
2070
2071 12. Added the -M command line option to pcretest.
2072
2073 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2074
2075 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2076
2077 16. Added comments and documentation about mis-use of no_arg in the C++
2078 wrapper.
2079
2080 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2081 from Martin Jerabek that uses macro names for all relevant character and
2082 string constants.
2083
2084 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2085 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2086 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2087 these, but not everybody uses configure.
2088
2089 19. A conditional group that had only one branch was not being correctly
2090 recognized as an item that could match an empty string. This meant that an
2091 enclosing group might also not be so recognized, causing infinite looping
2092 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2093 with the subject "ab", where knowledge that the repeated group can match
2094 nothing is needed in order to break the loop.
2095
2096 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2097 exec(), but without supplying a callout function, matching went wrong.
2098
2099 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2100 leak if the size of the offset vector was greater than 30. When the vector
2101 is smaller, the saved offsets during recursion go onto a local stack
2102 vector, but for larger vectors malloc() is used. It was failing to free
2103 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2104 error, in fact).
2105
2106 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2107 heapframe that is used only when UTF-8 support is enabled. This caused no
2108 problem, but was untidy.
2109
2110 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2111 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2112 included within another project.
2113
2114 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2115 slightly modified by me:
2116
2117 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2118 not building pcregrep.
2119
2120 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2121 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2122
2123 25. Forward references, both numeric and by name, in patterns that made use of
2124 duplicate group numbers, could behave incorrectly or give incorrect errors,
2125 because when scanning forward to find the reference group, PCRE was not
2126 taking into account the duplicate group numbers. A pattern such as
2127 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2128
2129 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2130 the feature of a custom pointer more persuasive (as requested by a user).
2131
2132 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2133 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2134 pre-defined.
2135
2136 28. Added support for (*UTF8) at the start of a pattern.
2137
2138 29. Arrange for flags added by the "release type" setting in CMake to be shown
2139 in the configuration summary.
2140
2141
2142 Version 7.8 05-Sep-08
2143 ---------------------
2144
2145 1. Replaced UCP searching code with optimized version as implemented for Ad
2146 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2147 stage table and inline lookup instead of a function, giving speed ups of 2
2148 to 5 times on some simple patterns that I tested. Permission was given to
2149 distribute the MultiStage2.py script that generates the tables (it's not in
2150 the tarball, but is in the Subversion repository).
2151
2152 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2153 scripts.
2154
2155 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2156 a group with a zero qualifier. The result of the study could be incorrect,
2157 or the function might crash, depending on the pattern.
2158
2159 4. Caseless matching was not working for non-ASCII characters in back
2160 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2161 It now works when Unicode Property Support is available.
2162
2163 5. In pcretest, an escape such as \x{de} in the data was always generating
2164 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2165 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2166 truncation.
2167
2168 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2169
2170 7. Added two (int) casts to pcregrep when printing the difference of two
2171 pointers, in case they are 64-bit values.
2172
2173 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2174 test 2 if it fails.
2175
2176 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2177 and a #define of that name to empty if it is not externally set. This is to
2178 allow users of MSVC to set it if necessary.
2179
2180 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2181 the convenience functions in the pcre_get.c source file.
2182
2183 11. An option change at the start of a pattern that had top-level alternatives
2184 could cause overwriting and/or a crash. This command provoked a crash in
2185 some environments:
2186
2187 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2188
2189 This potential security problem was recorded as CVE-2008-2371.
2190
2191 12. For a pattern where the match had to start at the beginning or immediately
2192 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2193 pcre_dfa_exec() could read past the end of the passed subject if there was
2194 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2195 pcretest so that it places the subject at the end of its malloc-ed buffer.
2196
2197 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2198 exec() might read past the end of the data buffer in UTF-8 mode.
2199
2200 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2201 the data contained the byte 0x85 as part of a UTF-8 character within its
2202 first line. This applied both to normal and DFA matching.
2203
2204 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2205 /^[^d]*?$/8 failed to match "abc".
2206
2207 16. Added a missing copyright notice to pcrecpp_internal.h.
2208
2209 17. Make it more clear in the documentation that values returned from
2210 pcre_exec() in ovector are byte offsets, not character counts.
2211
2212 18. Tidied a few places to stop certain compilers from issuing warnings.
2213
2214 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2215 supplied by Stefan Weber. I made a further small update for 7.8 because
2216 there is a change of source arrangements: the pcre_searchfuncs.c module is
2217 replaced by pcre_ucd.c.
2218
2219
2220 Version 7.7 07-May-08
2221 ---------------------
2222
2223 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2224 a string to a long long, pretend we don't even have a long long." This is
2225 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2226
2227 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2228 pre-7.6 versions, which defined a global no_arg variable instead of putting
2229 it in the RE class. (See also #8 below.)
2230
2231 3. Remove a line of dead code, identified by coverity and reported by Nuno
2232 Lopes.
2233
2234 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2235
2236 (1) The include/exclude patterns were being applied to the whole pathnames
2237 of files, instead of just to the final components.
2238
2239 (2) If there was more than one level of directory, the subdirectories were
2240 skipped unless they satisfied the include/exclude conditions. This is
2241 inconsistent with GNU grep (and could even be seen as contrary to the
2242 pcregrep specification - which I improved to make it absolutely clear).
2243 The action now is always to scan all levels of directory, and just
2244 apply the include/exclude patterns to regular files.
2245
2246 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2247 --exclude_dir in the tests to avoid scanning .svn directories.
2248
2249 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2250 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2251 doesn't support NULs in patterns.
2252
2253 7. Added some missing "const"s to declarations of static tables in
2254 pcre_compile.c and pcre_dfa_exec.c.
2255
2256 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2257 caused by fix #2 above. (Subsequently also a second patch to fix the
2258 first patch. And a third patch - this was a messy problem.)
2259
2260 9. Applied Craig's patch to remove the use of push_back().
2261
2262 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2263 matching function regexec().
2264
2265 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2266 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2267 references. PCRE supports relative numbers with this syntax (I don't think
2268 Oniguruma does).
2269
2270 12. Previously, a group with a zero repeat such as (...){0} was completely
2271 omitted from the compiled regex. However, this means that if the group
2272 was called as a subroutine from elsewhere in the pattern, things went wrong
2273 (an internal error was given). Such groups are now left in the compiled
2274 pattern, with a new opcode that causes them to be skipped at execution
2275 time.
2276
2277 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2278 to the way PCRE behaves:
2279
2280 (a) A lone ] character is dis-allowed (Perl treats it as data).
2281
2282 (b) A back reference to an unmatched subpattern matches an empty string
2283 (Perl fails the current match path).
2284
2285 (c) A data ] in a character class must be notated as \] because if the
2286 first data character in a class is ], it defines an empty class. (In
2287 Perl it is not possible to have an empty class.) The empty class []
2288 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2289 The negative empty class [^] matches any one character, independently
2290 of the DOTALL setting.
2291
2292 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2293 non-existent subpattern following a character class starting with ']' and
2294 containing () gave an internal compiling error instead of "reference to
2295 non-existent subpattern". Fortunately, when the pattern did exist, the
2296 compiled code was correct. (When scanning forwards to check for the
2297 existence of the subpattern, it was treating the data ']' as terminating
2298 the class, so got the count wrong. When actually compiling, the reference
2299 was subsequently set up correctly.)
2300
2301 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2302 it was being rejected as not supported by pcre_dfa_exec(), even though
2303 other assertions are supported. I have made pcre_dfa_exec() support
2304 (*FAIL).
2305
2306 16. The implementation of 13c above involved the invention of a new opcode,
2307 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2308 cannot be changed at match time, I realized I could make a small
2309 improvement to matching performance by compiling OP_ALLANY instead of
2310 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2311 on the OP_ANY path.
2312
2313 17. Compiling pcretest on Windows with readline support failed without the
2314 following two fixes: (1) Make the unistd.h include conditional on
2315 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2316
2317 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2318 ncurses library to be included for pcretest when ReadLine support is
2319 requested, but also to allow for it to be overridden. This patch came from
2320 Daniel Bergström.
2321
2322 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2323 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2324 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2325 spotting this.
2326
2327
2328 Version 7.6 28-Jan-08
2329 ---------------------
2330
2331 1. A character class containing a very large number of characters with
2332 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2333 overflow.
2334
2335 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2336 HAVE_LONG_LONG is not defined.
2337
2338 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2339 bring them up to date and include new features. This patch includes:
2340
2341 - Fixed PH's badly added libz and libbz2 support.
2342 - Fixed a problem with static linking.
2343 - Added pcredemo. [But later removed - see 7 below.]
2344 - Fixed dftables problem and added an option.
2345 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2346 HAVE_LONG_LONG.
2347 - Added readline support for pcretest.
2348 - Added an listing of the option settings after cmake has run.
2349
2350 4. A user submitted a patch to Makefile that makes it easy to create
2351 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2352 Makefile.am that cause it to include this special target, without
2353 affecting anything else. Note that the same mingw target plus all
2354 the other distribution libraries and programs are now supported
2355 when configuring with CMake (see 6 below) instead of with
2356 Configure/Make.
2357
2358 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2359 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2360 exported in the Windows port". It has not yet been confirmed that the patch
2361 solves the problem, but it does no harm.
2362
2363 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2364 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2365 with CMake, and also correct the comment about stack recursion.
2366
2367 7. Remove the automatic building of pcredemo from the ./configure system and
2368 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2369 of a program that users should build themselves after PCRE is installed, so
2370 building it automatically is not really right. What is more, it gave
2371 trouble in some build environments.
2372
2373 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2374
2375
2376 Version 7.5 10-Jan-08
2377 ---------------------
2378
2379 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2380 values in parens when parsing an RE using the C++ wrapper."
2381
2382 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2383 Characters greater than 255 were excluded from the class instead of being
2384 included.
2385
2386 3. The same bug as (2) above applied to negated POSIX classes such as
2387 [:^space:].
2388
2389 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2390 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2391 I have changed it.
2392
2393 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2394 first named subpattern) and a construct such as (?&a) would reference the
2395 first named subpattern whose name started with "a" (in other words, the
2396 length check was missing). Both these problems are fixed. "Subpattern name
2397 expected" is now given for (?&) (a zero-length name), and this patch also
2398 makes it give the same error for \k'' (previously it complained that that
2399 was a reference to a non-existent subpattern).
2400
2401 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2402 this is right because (?- can be followed by option settings as well as by
2403 digits. I have, however, made the messages clearer.
2404
2405 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2406 than the number used in the conditional) now cause a compile-time error.
2407 This is actually not compatible with Perl, which accepts such patterns, but
2408 treats the conditional as always being FALSE (as PCRE used to), but it
2409 seems to me that giving a diagnostic is better.
2410
2411 8. Change "alphameric" to the more common word "alphanumeric" in comments
2412 and messages.
2413
2414 9. Fix two occurrences of "backslash" in comments that should have been
2415 "backspace".
2416
2417 10. Remove two redundant lines of code that can never be obeyed (their function
2418 was moved elsewhere).
2419
2420 11. The program that makes PCRE's Unicode character property table had a bug
2421 which caused it to generate incorrect table entries for sequences of
2422 characters that have the same character type, but are in different scripts.
2423 It amalgamated them into a single range, with the script of the first of
2424 them. In other words, some characters were in the wrong script. There were
2425 thirteen such cases, affecting characters in the following ranges:
2426
2427 U+002b0 - U+002c1
2428 U+0060c - U+0060d
2429 U+0061e - U+00612
2430 U+0064b - U+0065e
2431 U+0074d - U+0076d
2432 U+01800 - U+01805
2433 U+01d00 - U+01d77
2434 U+01d9b - U+01dbf
2435 U+0200b - U+0200f
2436 U+030fc - U+030fe
2437 U+03260 - U+0327f
2438 U+0fb46 - U+0fbb1
2439 U+10450 - U+1049d
2440
2441 12. The -o option (show only the matching part of a line) for pcregrep was not
2442 compatible with GNU grep in that, if there was more than one match in a
2443 line, it showed only the first of them. It now behaves in the same way as
2444 GNU grep.
2445
2446 13. If the -o and -v options were combined for pcregrep, it printed a blank
2447 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2448 does the same. The return code can be used to tell if there were any
2449 non-matching lines.
2450
2451 14. Added --file-offsets and --line-offsets to pcregrep.
2452
2453 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2454 infinitely looping recursion. The bug was that positive lookaheads were not
2455 being skipped when checking for a possible empty match (negative lookaheads
2456 and both kinds of lookbehind were skipped).
2457
2458 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2459 inclusion of <windows.h> to before rather than after the definition of
2460 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2461
2462 17. Specifying a possessive quantifier with a specific limit for a Unicode
2463 character property caused pcre_compile() to compile bad code, which led at
2464 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2465 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2466 caused the error; without that there was no problem.
2467
2468 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2469
2470 19. Added --enable-pcretest-libreadline.
2471
2472 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2473 RE::GlobalReplace(). As a result, the number of replacements returned was
2474 double what it should be. I removed one of the increments, but Craig sent a
2475 later patch that removed the other one (the right fix) and added unit tests
2476 that check the return values (which was not done before).
2477
2478 21. Several CMake things:
2479
2480 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2481 the names libpcre and libpcreposix, not just pcre and pcreposix.
2482
2483 (2) The above change means that pcretest and pcregrep are now correctly
2484 linked with the newly-built libraries, not previously installed ones.
2485
2486 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2487
2488 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2489 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2490 UTF-8 newline character). The key issue is that the pattern starts .*;
2491 this means that the match must be either at the beginning, or after a
2492 newline. The bug was in the code for advancing after a failed match and
2493 checking that the new position followed a newline. It was not taking
2494 account of UTF-8 characters correctly.
2495
2496 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2497 character classes. PCRE was not treating the sequence [:...:] as a
2498 character class unless the ... were all letters. Perl, however, seems to
2499 allow any characters between [: and :], though of course it rejects as
2500 unknown any "names" that contain non-letters, because all the known class
2501 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2502 for example, whereas PCRE did not - it did not recognize a POSIX character
2503 class. This seemed a bit dangerous, so the code has been changed to be
2504 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2505 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2506 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2507 Perl does, and where it didn't before.
2508
2509 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2510 Windows environments %n is disabled by default.
2511
2512
2513 Version 7.4 21-Sep-07
2514 ---------------------
2515
2516 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2517 means that a class such as [\s] counted as "explicit reference to CR or
2518 LF". That isn't really right - the whole point of the change was to try to
2519 help when there was an actual mention of one of the two characters. So now
2520 the change happens only if \r or \n (or a literal CR or LF) character is
2521 encountered.
2522
2523 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2524 of both had grown to the point where there were only 3 bits left.
2525 Fortunately, there was spare space in the data structure, and so I have
2526 moved the internal flags into a new 16-bit field to free up more option
2527 bits.
2528
2529 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2530 but did not set the internal JCHANGED flag - either of these is enough to
2531 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2532 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2533 start sets both bits.
2534
2535 4. Added options (at build time, compile time, exec time) to change \R from
2536 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2537
2538 5. doc/pcresyntax.html was missing from the distribution.
2539
2540 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2541 compatibility, even though it is no longer used.
2542
2543 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2544 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2545 windows.h file is present (where different names are used). [This was
2546 reversed later after testing - see 16 below.]
2547
2548 8. Changed all #include <config.h> to #include "config.h". There were also
2549 some further <pcre.h> cases that I changed to "pcre.h".
2550
2551 9. When pcregrep was used with the --colour option, it missed the line ending
2552 sequence off the lines that it output.
2553
2554 10. It was pointed out to me that arrays of string pointers cause lots of
2555 relocations when a shared library is dynamically loaded. A technique of
2556 using a single long string with a table of offsets can drastically reduce
2557 these. I have refactored PCRE in four places to do this. The result is
2558 dramatic:
2559
2560 Originally: 290
2561 After changing UCP table: 187
2562 After changing error message table: 43
2563 After changing table of "verbs" 36
2564 After changing table of Posix names 22
2565
2566 Thanks to the folks working on Gregex for glib for this insight.
2567
2568 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2569 unicode-properties was also set.
2570
2571 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2572
2573 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2574 checked only for CRLF.
2575
2576 14. Added casts to pcretest.c to avoid compiler warnings.
2577
2578 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2579
2580 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2581 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2582 entirely. This removes changes made in 7 above.
2583
2584 17. The CMake files have been updated, and there is now more information about
2585 building with CMake in the NON-UNIX-USE document.
2586
2587
2588 Version 7.3 28-Aug-07
2589 ---------------------
2590
2591 1. In the rejigging of the build system that eventually resulted in 7.1, the
2592 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2593 brackets there is not right, since it causes compilers to look for an
2594 installed pcre.h, not the version that is in the source that is being
2595 compiled (which of course may be different). I have changed it back to:
2596
2597 #include "pcre.h"
2598
2599 I have a vague recollection that the change was concerned with compiling in
2600 different directories, but in the new build system, that is taken care of
2601 by the VPATH setting the Makefile.
2602
2603 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2604 when the subject happened to end in the byte 0x85 (e.g. if the last
2605 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2606 characters but of course it shouldn't be taken as a newline when it is part
2607 of another character. The bug was that, for an unlimited repeat of . in
2608 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2609 characters when looking for a newline.
2610
2611 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2612
2613 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2614 in debug output.
2615
2616 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2617 long printing in the pcrecpp unittest when running under MinGW.
2618
2619 6. ESC_K was left out of the EBCDIC table.
2620
2621 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2622 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2623 limit also applies to "virtual nesting" when a pattern is recursive, and in
2624 this case 1000 isn't so big. I have been able to remove this limit at the
2625 expense of backing off one optimization in certain circumstances. Normally,
2626 when pcre_exec() would call its internal match() function recursively and
2627 immediately return the result unconditionally, it uses a "tail recursion"
2628 feature to save stack. However, when a subpattern that can match an empty
2629 string has an unlimited repetition quantifier, it no longer makes this
2630 optimization. That gives it a stack frame in which to save the data for
2631 checking that an empty string has been matched. Previously this was taken
2632 from the 1000-entry workspace that had been reserved. So now there is no
2633 explicit limit, but more stack is used.
2634
2635 8. Applied Daniel's patches to solve problems with the import/export magic
2636 syntax that is required for Windows, and which was going wrong for the
2637 pcreposix and pcrecpp parts of the library. These were overlooked when this
2638 problem was solved for the main library.
2639
2640 9. There were some crude static tests to avoid integer overflow when computing
2641 the size of patterns that contain repeated groups with explicit upper
2642 limits. As the maximum quantifier is 65535, the maximum group length was
2643 set at 30,000 so that the product of these two numbers did not overflow a
2644 32-bit integer. However, it turns out that people want to use groups that
2645 are longer than 30,000 bytes (though not repeat them that many times).
2646 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2647 made it possible to implement the integer overflow checks in a much more
2648 dynamic way, which I have now done. The artificial limitation on group
2649 length has been removed - we now have only the limit on the total length of
2650 the compiled pattern, which depends on the LINK_SIZE setting.
2651
2652 10. Fixed a bug in the documentation for get/copy named substring when
2653 duplicate names are permitted. If none of the named substrings are set, the
2654 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2655 empty string.
2656
2657 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2658 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2659 because the ] is interpreted as the first data character and the
2660 terminating ] is not found. PCRE has been made compatible with Perl in this
2661 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2662 cause memory overwriting.
2663
2664 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2665 string has been matched (to stop an infinite loop). It was not recognizing
2666 a conditional subpattern that could match an empty string if that
2667 subpattern was within another subpattern. For example, it looped when
2668 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2669 condition was not nested. This bug has been fixed.
2670
2671 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2672 past the start of the subject in the presence of bytes with the top bit
2673 set, for example "\x8aBCD".
2674
2675 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2676 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2677
2678 14. Optimized (?!) to (*FAIL).
2679
2680 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2681 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2682 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2683 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2684 does: it's just the validity check that is more restrictive.
2685
2686 16. Inserted checks for integer overflows during escape sequence (backslash)
2687 processing, and also fixed erroneous offset values for syntax errors during
2688 backslash processing.
2689
2690 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2691 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2692
2693 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2694 caused an overrun.
2695
2696 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2697 something other than just ASCII characters) inside a group that had an
2698 unlimited repeat caused a loop at compile time (while checking to see
2699 whether the group could match an empty string).
2700
2701 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2702 [\P{Any}] did so. (Error in the code for printing property names.)
2703
2704 21. An orphan \E inside a character class could cause a crash.
2705
2706 22. A repeated capturing bracket such as (A)? could cause a wild memory
2707 reference during compilation.
2708
2709 23. There are several functions in pcre_compile() that scan along a compiled
2710 expression for various reasons (e.g. to see if it's fixed length for look
2711 behind). There were bugs in these functions when a repeated \p or \P was
2712 present in the pattern. These operators have additional parameters compared
2713 with \d, etc, and these were not being taken into account when moving along
2714 the compiled data. Specifically:
2715
2716 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2717 length.
2718
2719 (b) An item such as \pL+ within a repeated group could cause crashes or
2720 loops.
2721
2722 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2723 "reference to non-existent subpattern" error.
2724
2725 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2726
2727 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2728 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2729
2730 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2731
2732 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2733 character were causing crashes (broken optimization).
2734
2735 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2736 \p or \P) caused a compile-time loop.
2737
2738 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2739 break. For example, the unstudied pattern [\r\n]A does not match the string
2740 "\r\nA" because change 7.0/46 below moves the current point on by two
2741 characters after failing to match at the start. However, the pattern \nA
2742 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2743 the same is true. There doesn't seem any very clean way out of this, but
2744 what I have chosen to do makes the common cases work: PCRE now takes note
2745 of whether there can be an explicit match for \r or \n anywhere in the
2746 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2747 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2748 pattern has explicit CR or LF references.
2749
2750 29. Added (*CR) etc for changing newline setting at start of pattern.
2751
2752
2753 Version 7.2 19-Jun-07
2754 ---------------------
2755
2756 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2757 which is apparently normally available under Windows.
2758
2759 2. Re-jig the pcregrep tests with different newline settings in an attempt
2760 to make them independent of the local environment's newline setting.
2761
2762 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2763
2764 4. Some of the "internals" tests were previously cut out when the link size
2765 was not 2, because the output contained actual offsets. The recent new
2766 "Z" feature of pcretest means that these can be cut out, making the tests
2767 usable with all link sizes.
2768
2769 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2770 stack recursion. This gives a massive performance boost under BSD, but just
2771 a small improvement under Linux. However, it saves one field in the frame
2772 in all cases.
2773
2774 6. Added more features from the forthcoming Perl 5.10:
2775
2776 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2777 recursion call. It refers to the nth most recently opened parentheses.
2778
2779 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2780 to be opened parentheses.
2781
2782 (c) Conditions that refer to capturing parentheses can be specified
2783 relatively, for example, (?(-2)... or (?(+3)...
2784
2785 (d) \K resets the start of the current match so that everything before
2786 is not part of it.
2787
2788 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2789
2790 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2791 reference syntax.
2792
2793 (g) (?| introduces a group in which the numbering of parentheses in each
2794 alternative starts with the same number.
2795
2796 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2797
2798 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2799 PCRE_INFO_JCHANGED.
2800
2801 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2802 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2803 for detecting groups that can match an empty string.
2804
2805 9. A pattern with a very large number of alternatives (more than several
2806 hundred) was running out of internal workspace during the pre-compile
2807 phase, where pcre_compile() figures out how much memory will be needed. A
2808 bit of new cunning has reduced the workspace needed for groups with
2809 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2810 workspace instead of running out of the 4096 that are available.
2811
2812 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2813
2814 11. Applied patch from Google to remove an optimization that didn't quite work.
2815 The report of the bug said:
2816
2817 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2818 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2819 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2820
2821 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2822 it matched the wrong number of bytes.
2823
2824
2825 Version 7.1 24-Apr-07
2826 ---------------------
2827
2828 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2829 that is more "standard", making use of automake and other Autotools. There
2830 is some re-arrangement of the files and adjustment of comments consequent
2831 on this.
2832
2833 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2834 for recursive directory scanning broke on some systems because the files
2835 are not scanned in any specific order and on different systems the order
2836 was different. A call to "sort" has been inserted into RunGrepTest for the
2837 approprate test as a short-term fix. In the longer term there may be an
2838 alternative.
2839
2840 3. I had an email from Eric Raymond about problems translating some of PCRE's
2841 man pages to HTML (despite the fact that I distribute HTML pages, some
2842 people do their own conversions for various reasons). The problems
2843 concerned the use of low-level troff macros .br and .in. I have therefore
2844 removed all such uses from the man pages (some were redundant, some could
2845 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2846 HTML has been updated to handle .nf/.fi and to complain if it encounters
2847 .br or .in.
2848
2849 4. Updated comments in configure.ac that get placed in config.h.in and also
2850 arranged for config.h to be included in the distribution, with the name
2851 config.h.generic, for the benefit of those who have to compile without
2852 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2853
2854 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2855 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2856 makevp.bat for the current PCRE, using the additional files
2857 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2858
2859 6. A Windows user reported a minor discrepancy with test 2, which turned out
2860 to be caused by a trailing space on an input line that had got lost in his
2861 copy. The trailing space was an accident, so I've just removed it.
2862
2863 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2864 that is needed.
2865
2866 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2867 as "const" (a) because they are and (b) because it helps the PHP
2868 maintainers who have recently made a script to detect big data structures
2869 in the php code that should be moved to the .rodata section. I remembered
2870 to update Builducptable as well, so it won't revert if ucptable.h is ever
2871 re-created.
2872
2873 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2874 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2875 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2876 support is not required. This saves 1.5-2K of code, which is important in
2877 some applications.
2878
2879 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2880 so as not to refer to the tables, even though these functions will never be
2881 called when UTF-8 support is disabled. Otherwise there are problems with a
2882 shared library.
2883
2884 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2885
2886 (a) It was defining its arguments as char * instead of void *.
2887
2888 (b) It was assuming that all moves were upwards in memory; this was true
2889 a long time ago when I wrote it, but is no longer the case.
2890
2891 The emulated memove() is provided for those environments that have neither
2892 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2893 is clearly not the case, as these two bugs were recently reported.
2894
2895 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2896 and Detrail to create the HTML documentation, the .txt form of the man
2897 pages, and it removes trailing spaces from listed files. It also creates
2898 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2899 case, it wraps all the #defines with #ifndefs. This script should be run
2900 before "make dist".
2901
2902 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2903 with Unicode property support.
2904
2905 (a) For a maximizing quantifier, if the two different cases of the
2906 character were of different lengths in their UTF-8 codings (there are
2907 some cases like this - I found 11), and the matching function had to
2908 back up over a mixture of the two cases, it incorrectly assumed they
2909 were both the same length.
2910
2911 (b) When PCRE was configured to use the heap rather than the stack for
2912 recursion during matching, it was not correctly preserving the data for
2913 the other case of a UTF-8 character when checking ahead for a match
2914 while processing a minimizing repeat. If the check also involved
2915 matching a wide character, but failed, corruption could cause an
2916 erroneous result when trying to check for a repeat of the original
2917 character.
2918
2919 13. Some tidying changes to the testing mechanism:
2920
2921 (a) The RunTest script now detects the internal link size and whether there
2922 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2923 values substituted by "configure". (The RunGrepTest script already did
2924 this for UTF-8.) The configure.ac script no longer substitutes the
2925 relevant variables.
2926
2927 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2928 with length and offset values. This means that the output is different
2929 for different internal link sizes. Test 2 is skipped for link sizes
2930 other than 2 because of this, bypassing the problem. Unfortunately,
2931 there was also a test in test 3 (the locale tests) that used /B and
2932 failed for link sizes other than 2. Rather than cut the whole test out,
2933 I have added a new /Z option to pcretest that replaces the length and
2934 offset values with spaces. This is now used to make test 3 independent
2935 of link size. (Test 2 will be tidied up later.)
2936
2937 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2938 segmentation fault instead of returning the appropriate error message.
2939
2940 15. In multiline mode when the newline sequence was set to "any", the pattern
2941 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2942 This doesn't seem right; it now treats the CRLF combination as the line
2943 ending, and so does not match in that case. It's only a pattern such as ^$
2944 that would hit this one: something like ^ABC$ would have failed after \r
2945 and then tried again after \r\n.
2946
2947 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2948 in an attempt to make files that differ only in their line terminators
2949 compare equal. This works on Linux.
2950
2951 17. Under certain error circumstances pcregrep might try to free random memory
2952 as it exited. This is now fixed, thanks to valgrind.
2953
2954 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2955 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2956 was because its rules for how to advance for /g after matching an empty
2957 string at the end of a line did not allow for this case. They now check for
2958 it specially.
2959
2960 20. pcretest is supposed to handle patterns and data of any length, by
2961 extending its buffers when necessary. It was getting this wrong when the
2962 buffer for a data line had to be extended.
2963
2964 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2965 CRLF as a newline sequence.
2966
2967 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2968 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2969 I have nevertheless tidied it up.
2970
2971 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2972
2973 24. Added a man page for pcre-config.
2974
2975
2976 Version 7.0 19-Dec-06
2977 ---------------------
2978
2979 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2980 moving to gcc 4.1.1.
2981
2982 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2983 sys/time.h, which is documented as needed for this function. It doesn't
2984 seem to matter on Linux, but it showed up on some releases of OS X.
2985
2986 3. It seems that there are systems where bytes whose values are greater than
2987 127 match isprint() in the "C" locale. The "C" locale should be the
2988 default when a C program starts up. In most systems, only ASCII printing
2989 characters match isprint(). This difference caused the output from pcretest
2990 to vary, making some of the tests fail. I have changed pcretest so that:
2991
2992 (a) When it is outputting text in the compiled version of a pattern, bytes
2993 other than 32-126 are always shown as hex escapes.
2994
2995 (b) When it is outputting text that is a matched part of a subject string,
2996 it does the same, unless a different locale has been set for the match
2997 (using the /L modifier). In this case, it uses isprint() to decide.
2998
2999 4. Fixed a major bug that caused incorrect computation of the amount of memory
3000 required for a compiled pattern when options that changed within the
3001 pattern affected the logic of the preliminary scan that determines the
3002 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3003 that the computed length was too small. The symptoms of this bug were
3004 either the PCRE error "internal error: code overflow" from pcre_compile(),
3005 or a glibc crash with a message such as "pcretest: free(): invalid next
3006 size (fast)". Examples of patterns that provoked this bug (shown in
3007 pcretest format) are:
3008
3009 /(?-x: )/x
3010 /(?x)(?-x: \s*#\s*)/
3011 /((?i)[\x{c0}])/8
3012 /(?i:[\x{c0}])/8
3013
3014 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3015 is now done differently.
3016
3017 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3018 wrapper classes; (b) implement a new function in the C++ scanner that is
3019 more efficient than the old way of doing things because it avoids levels of
3020 recursion in the regex matching; (c) add a paragraph to the documentation
3021 for the FullMatch() function.
3022
3023 6. The escape sequence \n was being treated as whatever was defined as
3024 "newline". Not only was this contrary to the documentation, which states
3025 that \n is character 10 (hex 0A), but it also went horribly wrong when
3026 "newline" was defined as CRLF. This has been fixed.
3027
3028 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3029 was being set to -1 for the "end of line" case (supposedly a value that no
3030 character can have). Though this value is never used (the check for end of
3031 line is "zero bytes in current character"), it caused compiler complaints.
3032 I've changed it to 0xffffffff.
3033
3034 8. In pcre_version.c, the version string was being built by a sequence of
3035 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3036 string (as it is for production releases) called a macro with an empty
3037 argument. The C standard says the result of this is undefined. The gcc
3038 compiler treats it as an empty string (which was what was wanted) but it is
3039 reported that Visual C gives an error. The source has been hacked around to
3040 avoid this problem.
3041
3042 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3043 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3044 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3045 of them did).
3046
3047 10. Originally, pcretest opened its input and output without "b"; then I was
3048 told that "b" was needed in some environments, so it was added for release
3049 5.0 to both the input and output. (It makes no difference on Unix-like
3050 systems.) Later I was told that it is wrong for the input on Windows. I've
3051 now abstracted the modes into two macros, to make it easier to fiddle with
3052 them, and removed "b" from the input mode under Windows.
3053
3054 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3055
3056 12. Added -help and --help to pcretest as an official way of being reminded
3057 of the options.
3058
3059 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3060 and pcrecpp.cc because they annoy compilers at high warning levels.
3061
3062 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3063
3064 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3065 scripts are not C programs :-) and which was not noticed because it works
3066 on Linux.
3067
3068 16. pcretest is supposed to handle any length of pattern and data line (as one
3069 line or as a continued sequence of lines) by extending its input buffer if
3070 necessary. This feature was broken for very long pattern lines, leading to
3071 a string of junk being passed to pcre_compile() if the pattern was longer
3072 than about 50K.
3073
3074 17. I have done a major re-factoring of the way pcre_compile() computes the
3075 amount of memory needed for a compiled pattern. Previously, there was code
3076 that made a preliminary scan of the pattern in order to do this. That was
3077 OK when PCRE was new, but as the facilities have expanded, it has become
3078 harder and harder to keep it in step with the real compile phase, and there
3079 have been a number of bugs (see for example, 4 above). I have now found a
3080 cunning way of running the real compile function in a "fake" mode that
3081 enables it to compute how much memory it would need, while actually only
3082 ever using a few hundred bytes of working memory and without too many
3083 tests of the mode. This should make future maintenance and development
3084 easier. A side effect of this work is that the limit of 200 on the nesting
3085 depth of parentheses has been removed (though this was never a serious
3086 limitation, I suspect). However, there is a downside: pcre_compile() now
3087 runs more slowly than before (30% or more, depending on the pattern). I
3088 hope this isn't a big issue. There is no effect on runtime performance.
3089
3090 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3091 newline (only possible for the last line of a file) and it was a
3092 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3093
3094 19. Added additional timing features to pcretest. (1) The -tm option now times
3095 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3096 separate command line item, by a number that specifies the number of
3097 repeats to use when timing. The default is 50000; this gives better
3098 precision, but takes uncomfortably long for very large patterns.
3099
3100 20. Extended pcre_study() to be more clever in cases where a branch of a
3101 subpattern has no definite first character. For example, (a*|b*)[cd] would
3102 previously give no result from pcre_study(). Now it recognizes that the
3103 first character must be a, b, c, or d.
3104
3105 21. There was an incorrect error "recursive call could loop indefinitely" if
3106 a subpattern (or the entire pattern) that was being tested for matching an
3107 empty string contained only one non-empty item after a nested subpattern.
3108 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3109 incorrectly, because the \d was being skipped in the check.
3110
3111 22. The pcretest program now has a new pattern option /B and a command line
3112 option -b, which is equivalent to adding /B to every pattern. This causes
3113 it to show the compiled bytecode, without the additional information that
3114 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3115 is the same as /B/I).
3116
3117 23. A new optimization is now able automatically to treat some sequences such
3118 as a*b as a*+b. More specifically, if something simple (such as a character
3119 or a simple class like \d) has an unlimited quantifier, and is followed by
3120 something that cannot possibly match the quantified thing, the quantifier
3121 is automatically "possessified".
3122
3123 24. A recursive reference to a subpattern whose number was greater than 39
3124 went wrong under certain circumstances in UTF-8 mode. This bug could also
3125 have affected the operation of pcre_study().
3126
3127 25. Realized that a little bit of performance could be had by replacing
3128 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3129
3130 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3131
3132 27. Possessive quantifiers such as a++ were previously implemented by turning
3133 them into atomic groups such as ($>a+). Now they have their own opcodes,
3134 which improves performance. This includes the automatically created ones
3135 from 23 above.
3136
3137 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3138 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3139 the first matched character to be a colon. This applied both to named and
3140 numbered groups.
3141
3142 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3143
3144 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3145 building PCRE on Windows easier, so I have included it in the distribution.
3146
3147 31. There is now a check in pcretest against a ridiculously large number being
3148 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3149 loop, the loop is abandoned.
3150
3151 32. Forward references to subpatterns in conditions such as (?(2)...) where
3152 subpattern 2 is defined later cause pcre_compile() to search forwards in
3153 the pattern for the relevant set of parentheses. This search went wrong
3154 when there were unescaped parentheses in a character class, parentheses
3155 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3156
3157 33. "Subroutine" calls and backreferences were previously restricted to
3158 referencing subpatterns earlier in the regex. This restriction has now
3159 been removed.
3160
3161 34. Added a number of extra features that are going to be in Perl 5.10. On the
3162 whole, these are just syntactic alternatives for features that PCRE had
3163 previously implemented using the Python syntax or my own invention. The
3164 other formats are all retained for compatibility.
3165
3166 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3167 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3168 also .NET compatible.
3169
3170 (b) A recursion or subroutine call to a named group can now be defined as
3171 (?&name) as well as (?P>name).
3172
3173 (c) A backreference to a named group can now be defined as \k<name> or
3174 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3175 5.10, are also .NET compatible.
3176
3177 (d) A conditional reference to a named group can now use the syntax
3178 (?(<name>) or (?('name') as well as (?(name).
3179
3180 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3181 groups (named and numbered) that are never evaluated inline, but can be
3182 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3183 is always false. There may be only one alternative in such a group.
3184
3185 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3186 as the simple (?(R). The condition is true only if the most recent
3187 recursion is that of the given number or name. It does not search out
3188 through the entire recursion stack.
3189
3190 (g) The escape \gN or \g{N} has been added, where N is a positive or
3191 negative number, specifying an absolute or relative reference.
3192
3193 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3194 some "unreachable code" warnings.
3195
3196 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3197 things, this adds five new scripts.
3198
3199 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3200 There were also incompatibilities regarding the handling of \Q..\E inside
3201 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3202 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3203
3204 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3205 matches an empty string, and forcibly breaks the loop. There were bugs in
3206 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3207 against aaaa the result was just "a" rather than "aaaa", for example. Two
3208 separate and independent bugs (that affected different cases) have been
3209 fixed.
3210
3211 39. Refactored the code to abolish the use of different opcodes for small
3212 capturing bracket numbers. This is a tidy that I avoided doing when I
3213 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3214 The new approach is not only tidier, it makes it possible to reduce the
3215 memory needed to fix the previous bug (38).
3216
3217 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3218 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3219 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3220 mode.
3221
3222 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3223 report.
3224
3225 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3226 copy construction and assignment in the C++ wrapper.
3227
3228 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3229 couple of bugs that could have given wrong results in the "--newline=crlf"
3230 case.
3231
3232 44. Added a number of casts and did some reorganization of signed/unsigned int
3233 variables following suggestions from Dair Grant. Also renamed the variable
3234 "this" as "item" because it is a C++ keyword.
3235
3236 45. Arranged for dftables to add
3237
3238 #include "pcre_internal.h"
3239
3240 to pcre_chartables.c because without it, gcc 4.x may remove the array
3241 definition from the final binary if PCRE is built into a static library and
3242 dead code stripping is activated.
3243
3244 46. For an unanchored pattern, if a match attempt fails at the start of a
3245 newline sequence, and the newline setting is CRLF or ANY, and the next two
3246 characters are CRLF, advance by two characters instead of one.
3247
3248
3249 Version 6.7 04-Jul-06
3250 ---------------------
3251
3252 1. In order to handle tests when input lines are enormously long, pcretest has
3253 been re-factored so that it automatically extends its buffers when
3254 necessary. The code is crude, but this _is_ just a test program. The
3255 default size has been increased from 32K to 50K.
3256
3257 2. The code in pcre_study() was using the value of the re argument before
3258 testing it for NULL. (Of course, in any sensible call of the function, it
3259 won't be NULL.)
3260
3261 3. The memmove() emulation function in pcre_internal.h, which is used on
3262 systems that lack both memmove() and bcopy() - that is, hardly ever -
3263 was missing a "static" storage class specifier.
3264
3265 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3266 containing an extended class (one that cannot be represented by a bitmap
3267 because it contains high-valued characters or Unicode property items, e.g.
3268 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3269 pattern, but PCRE should not loop if you do not (it no longer does).
3270 [Detail: two cases were found: (a) a repeated subpattern containing an
3271 extended class; (b) a recursive reference to a subpattern that followed a
3272 previous extended class. It wasn't skipping over the extended class
3273 correctly when UTF-8 mode was not set.]
3274
3275 5. A negated single-character class was not being recognized as fixed-length
3276 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3277 compile error "lookbehind assertion is not fixed length".
3278
3279 6. The RunPerlTest auxiliary script was showing an unexpected difference
3280 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3281 write a Perl script that can interpret lines of an input file either as
3282 byte characters or as UTF-8, which is what "perltest" was being required to
3283 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3284 can't do is switch easily at run time between having the "use utf8;" pragma
3285 or not. In the end, I fudged it by using the RunPerlTest script to insert
3286 "use utf8;" explicitly for the UTF-8 tests.
3287
3288 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3289 the end of the subject string, contrary to the documentation and to what
3290 Perl does. This was true of both matching functions. Now it matches only at
3291 the start of the subject and immediately after *internal* newlines.
3292
3293 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3294 a pointer to an int instead of a pointer to an unsigned long int. This
3295 caused problems on 64-bit systems.
3296
3297 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3298 instance of the 'standard' template library not being so standard".
3299
3300 10. There was no check on the number of named subpatterns nor the maximum
3301 length of a subpattern name. The product of these values is used to compute
3302 the size of the memory block for a compiled pattern. By supplying a very
3303 long subpattern name and a large number of named subpatterns, the size
3304 computation could be caused to overflow. This is now prevented by limiting
3305 the length of names to 32 characters, and the number of named subpatterns
3306 to 10,000.
3307
3308 11. Subpatterns that are repeated with specific counts have to be replicated in
3309 the compiled pattern. The size of memory for this was computed from the
3310 length of the subpattern and the repeat count. The latter is limited to
3311 65535, but there was no limit on the former, meaning that integer overflow
3312 could in principle occur. The compiled length of a repeated subpattern is
3313 now limited to 30,000 bytes in order to prevent this.
3314
3315 12. Added the optional facility to have named substrings with the same name.
3316
3317 13. Added the ability to use a named substring as a condition, using the
3318 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3319 are numbers (not recommended). Forward references are permitted.
3320
3321 14. Added forward references in named backreferences (if you see what I mean).
3322
3323 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3324 pattern could run off the end of the subject. For example, the pattern
3325 "(?s)(.{1,5})"8 did this with the subject "ab".
3326
3327 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3328 PCRE_CASELESS was set when matching characters that were quantified with ?
3329 or *.
3330
3331 17. A character class other than a single negated character that had a minimum
3332 but no maximum quantifier - for example [ab]{6,} - was not handled
3333 correctly by pce_dfa_exec(). It would match only one character.
3334
3335 18. A valid (though odd) pattern that looked like a POSIX character
3336 class but used an invalid character after [ (for example [[,abc,]]) caused
3337 pcre_compile() to give the error "Failed: internal error: code overflow" or
3338 in some cases to crash with a glibc free() error. This could even happen if
3339 the pattern terminated after [[ but there just happened to be a sequence of
3340 letters, a binary zero, and a closing ] in the memory that followed.
3341
3342 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3343 over the years. Originally (before any Unicode support), just the bottom 8
3344 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3345 output from "man perlunicode" includes this:
3346
3347 The regular expression compiler produces polymorphic opcodes. That
3348 is, the pattern adapts to the data and automatically switches to
3349 the Unicode character scheme when presented with Unicode data--or
3350 instead uses a traditional byte scheme when presented with byte
3351 data.
3352
3353 Sadly, a wide octal escape does not cause a switch, and in a string with
3354 no other multibyte characters, these octal escapes are treated as before.
3355 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3356 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3357 Unicode string.
3358
3359 I have not perpetrated such confusion in PCRE. Up till now, it took just
3360 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3361 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3362 translate to the appropriate multibyte character.
3363
3364 29. Applied some refactoring to reduce the number of warnings from Microsoft
3365 and Borland compilers. This has included removing the fudge introduced
3366 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3367 a warning about an unused variable.
3368
3369 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3370 characters since release 4.0, because Perl (from release 5.004) does not.
3371 [Or at least, is documented not to: some releases seem to be in conflict
3372 with the documentation.] However, when a pattern was studied with
3373 pcre_study() and all its branches started with \s, PCRE still included VT
3374 as a possible starting character. Of course, this did no harm; it just
3375 caused an unnecessary match attempt.
3376
3377 22. Removed a now-redundant internal flag bit that recorded the fact that case
3378 dependency changed within the pattern. This was once needed for "required
3379 byte" processing, but is no longer used. This recovers a now-scarce options
3380 bit. Also moved the least significant internal flag bit to the most-
3381 significant bit of the word, which was not previously used (hangover from
3382 the days when it was an int rather than a uint) to free up another bit for
3383 the future.
3384
3385 23. Added support for CRLF line endings as well as CR and LF. As well as the
3386 default being selectable at build time, it can now be changed at runtime
3387 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3388 specify that it is scanning data with non-default line endings.
3389
3390 24. Changed the definition of CXXLINK to make it agree with the definition of
3391 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3392
3393 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3394 recursions. This makes a big different to stack usage for some patterns.
3395
3396 26. If a subpattern containing a named recursion or subroutine reference such
3397 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3398 the space required for the compiled pattern went wrong and gave too small a
3399 value. Depending on the environment, this could lead to "Failed: internal
3400 error: code overflow at offset 49" or "glibc detected double free or
3401 corruption" errors.
3402
3403 27. Applied patches from Google (a) to support the new newline modes and (b) to
3404 advance over multibyte UTF-8 characters in GlobalReplace.
3405
3406 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3407 difference for some implementation of PCRE in some Windows version.
3408
3409 29. Added some extra testing facilities to pcretest:
3410
3411 \q<number> in a data line sets the "match limit" value
3412 \Q<number> in a data line sets the "match recursion limt" value
3413 -S <number> sets the stack size, where <number> is in megabytes
3414
3415 The -S option isn't available for Windows.
3416
3417
3418 Version 6.6 06-Feb-06
3419 ---------------------
3420
3421 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3422 in pcreposix.h. I have copied the definition from pcre.h.
3423
3424 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3425 because pcre.h is no longer a built file.
3426
3427 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3428 not normally included in the compiled code.
3429
3430
3431 Version 6.5 01-Feb-06
3432 ---------------------
3433
3434 1. When using the partial match feature with pcre_dfa_exec(), it was not
3435 anchoring the second and subsequent partial matches at the new starting
3436 point. This could lead to incorrect results. For example, with the pattern
3437 /1234/, partially matching against "123" and then "a4" gave a match.
3438
3439 2. Changes to pcregrep:
3440
3441 (a) All non-match returns from pcre_exec() were being treated as failures
3442 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3443 error message is output. Some extra information is given for the
3444 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3445 probably the only errors that are likely to be caused by users (by
3446 specifying a regex that has nested indefinite repeats, for instance).
3447 If there are more than 20 of these errors, pcregrep is abandoned.
3448
3449 (b) A binary zero was treated as data while matching, but terminated the
3450 output line if it was written out. This has been fixed: binary zeroes
3451 are now no different to any other data bytes.
3452
3453 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3454 used to set a locale for matching. The --locale=xxxx long option has
3455 been added (no short equivalent) to specify a locale explicitly on the
3456 pcregrep command, overriding the environment variables.
3457
3458 (d) When -B was used with -n, some line numbers in the output were one less
3459 than they should have been.
3460
3461 (e) Added the -o (--only-matching) option.
3462
3463 (f) If -A or -C was used with -c (count only), some lines of context were
3464 accidentally printed for the final match.
3465
3466 (g) Added the -H (--with-filename) option.
3467
3468 (h) The combination of options -rh failed to suppress file names for files
3469 that were found from directory arguments.
3470
3471 (i) Added the -D (--devices) and -d (--directories) options.
3472
3473 (j) Added the -F (--fixed-strings) option.
3474
3475 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3476
3477 (l) Added the --colo(u)r option.
3478
3479 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3480 is not present by default.
3481
3482 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3483 items such as (?R) or (?1), when the recursion could match a number of
3484 alternatives. If it matched one of the alternatives, but subsequently,
3485 outside the recursion, there was a failure, the code tried to back up into
3486 the recursion. However, because of the way PCRE is implemented, this is not
3487 possible, and the result was an incorrect result from the match.
3488
3489 In order to prevent this happening, the specification of recursion has
3490 been changed so that all such subpatterns are automatically treated as
3491 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3492
3493 4. I had overlooked the fact that, in some locales, there are characters for
3494 which isalpha() is true but neither isupper() nor islower() are true. In
3495 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3496 and ordfeminine) are like this. This affected the treatment of \w and \W
3497 when they appeared in character classes, but not when they appeared outside
3498 a character class. The bit map for "word" characters is now created
3499 separately from the results of isalnum() instead of just taking it from the
3500 upper, lower, and digit maps. (Plus the underscore character, of course.)
3501
3502 5. The above bug also affected the handling of POSIX character classes such as
3503 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3504 permanent tables. Instead, the bit maps for such a class were previously
3505 created as the appropriate unions of the upper, lower, and digit bitmaps.
3506 Now they are created by subtraction from the [[:word:]] class, which has
3507 its own bitmap.
3508
3509 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3510 It is created by subtracting the vertical space characters (\x09, \x0a,
3511 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3512 subtraction was done in the overall bitmap for a character class, meaning
3513 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3514 be recognized. This bug has been fixed.
3515
3516 7. Patches from the folks at Google:
3517
3518 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3519 real life, but is still worth protecting against".
3520
3521 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3522 regular expressions".
3523
3524 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3525 have it.
3526
3527 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3528 "configure" and the latter not, in order to fix a problem somebody had
3529 with compiling the Arg class on HP-UX.
3530
3531 (e) Improve the error-handling of the C++ wrapper a little bit.
3532
3533 (f) New tests for checking recursion limiting.
3534
3535 8. The pcre_memmove() function, which is used only if the environment does not
3536 have a standard memmove() function (and is therefore rarely compiled),
3537 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3538 returning a result (though PCRE never actually uses the result).
3539
3540 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3541 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3542 returned instead of calling malloc() with an overflowing number that would
3543 most likely cause subsequent chaos.
3544
3545 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3546
3547 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3548 with this option is matched, the nmatch and pmatch options of regexec() are
3549 ignored.
3550
3551 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3552 provided in case anyone wants to the the POSIX interface with UTF-8
3553 strings.
3554
3555 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3556 C++ linking (needed for some HP-UX environments).
3557
3558 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3559 (unused parameter) and in the pcre_printint() function (omitted "default"
3560 switch label when the default is to do nothing).
3561
3562 15. Added some code to make it possible, when PCRE is compiled as a C++
3563 library, to replace subject pointers for pcre_exec() with a smart pointer
3564 class, thus making it possible to process discontinuous strings.
3565
3566 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3567 much the same function. They were added by different people who were trying
3568 to make PCRE easy to compile on non-Unix systems. It has been suggested
3569 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3570 for compiling on Windows systems. I have therefore replaced it with
3571 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3572 defaults to "extern" for C or "extern C" for C++, which works fine on
3573 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3574 SCOPE with something explicit in config.h. In addition:
3575
3576 (a) pcreposix.h still had just "extern" instead of either of these macros;
3577 I have replaced it with PCRE_DATA_SCOPE.
3578
3579 (b) Functions such as _pcre_xclass(), which are internal to the library,
3580 but external in the C sense, all had PCRE_EXPORT in their definitions.
3581 This is apparently wrong for the Windows case, so I have removed it.
3582 (It makes no difference on Unix-like systems.)
3583
3584 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3585 of recursive calls to match(). This is different to MATCH_LIMIT because
3586 that limits the total number of calls to match(), not all of which increase
3587 the depth of recursion. Limiting the recursion depth limits the amount of
3588 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3589 when PCRE is compiled, and changed at run time. A patch from Google adds
3590 this functionality to the C++ interface.
3591
3592 18. Changes to the handling of Unicode character properties:
3593
3594 (a) Updated the table to Unicode 4.1.0.
3595
3596 (b) Recognize characters that are not in the table as "Cn" (undefined).
3597
3598 (c) I revised the way the table is implemented to a much improved format
3599 which includes recognition of ranges. It now supports the ranges that
3600 are defined in UnicodeData.txt, and it also amalgamates other
3601 characters into ranges. This has reduced the number of entries in the
3602 table from around 16,000 to around 3,000, thus reducing its size
3603 considerably. I realized I did not need to use a tree structure after
3604 all - a binary chop search is just as efficient. Having reduced the
3605 number of entries, I extended their size from 6 bytes to 8 bytes to
3606 allow for more data.
3607
3608 (d) Added support for Unicode script names via properties such as \p{Han}.
3609
3610 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3611 matching that character.
3612
3613 20. When matching a repeated Unicode property with a minimum greater than zero,
3614 (for example \pL{2,}), PCRE could look past the end of the subject if it
3615 reached it while seeking the minimum number of characters. This could
3616 happen only if some of the characters were more than one byte long, because
3617 there is a check for at least the minimum number of bytes.
3618
3619 21. Refactored the implementation of \p and \P so as to be more general, to
3620 allow for more different types of property in future. This has changed the
3621 compiled form incompatibly. Anybody with saved compiled patterns that use
3622 \p or \P will have to recompile them.
3623
3624 22. Added "Any" and "L&" to the supported property types.
3625
3626 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3627 but give a compile time error if the value is greater than 0xff.
3628
3629 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3630 accidentally not being installed or uninstalled.
3631
3632 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3633 made were to insert the current release number. This seemed silly, because
3634 it made things harder for people building PCRE on systems that don't run
3635 "configure". I have turned pcre.h into a distributed file, no longer built
3636 by "configure", with the version identification directly included. There is
3637 no longer a pcre.h.in file.
3638
3639 However, this change necessitated a change to the pcre-config script as
3640 well. It is built from pcre-config.in, and one of the substitutions was the
3641 release number. I have updated configure.ac so that ./configure now finds
3642 the release number by grepping pcre.h.
3643
3644 26. Added the ability to run the tests under valgrind.
3645
3646
3647 Version 6.4 05-Sep-05
3648 ---------------------
3649
3650 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3651 "--" to be printed when multiple files were scanned, even when none of the
3652 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3653 consider it to be a bug, and have restored the previous behaviour.
3654
3655 2. A couple of code tidies to get rid of compiler warnings.
3656
3657 3. The pcretest program used to cheat by referring to symbols in the library
3658 whose names begin with _pcre_. These are internal symbols that are not
3659 really supposed to be visible externally, and in some environments it is
3660 possible to suppress them. The cheating is now confined to including
3661 certain files from the library's source, which is a bit cleaner.
3662
3663 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3664 file's purpose clearer.
3665
3666 5. Reorganized pcre_ucp_findchar().
3667
3668
3669 Version 6.3 15-Aug-05
3670 ---------------------
3671
3672 1. The file libpcre.pc.in did not have general read permission in the tarball.
3673
3674 2. There were some problems when building without C++ support:
3675
3676 (a) If C++ support was not built, "make install" and "make test" still
3677 tried to test it.
3678
3679 (b) There were problems when the value of CXX was explicitly set. Some
3680 changes have been made to try to fix these, and ...
3681
3682 (c) --disable-cpp can now be used to explicitly disable C++ support.
3683
3684 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3685 backslash in a target when C++ was disabled. This confuses some
3686 versions of "make", apparently. Using an intermediate variable solves
3687 this. (Same for CPP_LOBJ.)
3688
3689 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3690 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3691 necessary on certain architectures.
3692
3693 4. Added a setting of -export-symbols-regex to the link command to remove
3694 those symbols that are exported in the C sense, but actually are local
3695 within the library, and not documented. Their names all begin with
3696 "_pcre_". This is not a perfect job, because (a) we have to except some
3697 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3698 available (and never for static libraries). I have made a note to try to
3699 find a way round (a) in the future.
3700
3701
3702 Version 6.2 01-Aug-05
3703 ---------------------
3704
3705 1. There was no test for integer overflow of quantifier values. A construction
3706 such as {1111111111111111} would give undefined results. What is worse, if
3707 a minimum quantifier for a parenthesized subpattern overflowed and became
3708 negative, the calculation of the memory size went wrong. This could have
3709 led to memory overwriting.
3710
3711 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3712
3713 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3714 operating environments where this matters.
3715
3716 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3717 PCRE options from within the C++ wrapper.
3718
3719 5. Named capturing subpatterns were not being correctly counted when a pattern
3720 was compiled. This caused two problems: (a) If there were more than 100
3721 such subpatterns, the calculation of the memory needed for the whole
3722 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3723 back references of the form \12, where the number was greater than 9, were
3724 not recognized as back references, even though there were sufficient
3725 previous subpatterns.
3726
3727 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3728 versions of gcc, e.g. 2.95.4.
3729
3730
3731 Version 6.1 21-Jun-05
3732 ---------------------
3733
3734 1. There was one reference to the variable "posix" in pcretest.c that was not
3735 surrounded by "#if !defined NOPOSIX".
3736
3737 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3738 the cross-check on the old pcre_info() function, for the benefit of the
3739 cut-down version of PCRE that is currently imported into Exim.
3740
3741 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3742 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3743 patterns that start (?i) but I don't think that matters. The use of (?i) is
3744 just an example; this all applies to the other options as well.
3745
3746 4. Since libtool seems to echo the compile commands it is issuing, the output
3747 from "make" can be reduced a bit by putting "@" in front of each libtool
3748 compile command.
3749
3750 5. Patch from the folks at Google for configure.in to be a bit more thorough
3751 in checking for a suitable C++ installation before trying to compile the
3752 C++ stuff. This should fix a reported problem when a compiler was present,
3753 but no suitable headers.
3754
3755 6. The man pages all had just "PCRE" as their title. I have changed them to
3756 be the relevant file name. I have also arranged that these names are
3757 retained in the file doc/pcre.txt, which is a concatenation in text format
3758 of all the man pages except the little individual ones for each function.
3759
3760 7. The NON-UNIX-USE file had not been updated for the different set of source
3761 files that come with release 6. I also added a few comments about the C++
3762 wrapper.
3763
3764
3765 Version 6.0 07-Jun-05
3766 ---------------------
3767
3768 1. Some minor internal re-organization to help with my DFA experiments.
3769
3770 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3771 didn't matter for the library itself when fully configured, but did matter
3772 when compiling without UCP support, or within Exim, where the ucp files are
3773 not imported.
3774
3775 3. Refactoring of the library code to split up the various functions into
3776 different source modules. The addition of the new DFA matching code (see
3777 below) to a single monolithic source would have made it really too
3778 unwieldy, quite apart from causing all the code to be include in a
3779 statically linked application, when only some functions are used. This is
3780 relevant even without the DFA addition now that patterns can be compiled in
3781 one application and matched in another.
3782
3783 The downside of splitting up is that there have to be some external
3784 functions and data tables that are used internally in different modules of
3785 the library but which are not part of the API. These have all had their
3786 names changed to start with "_pcre_" so that they are unlikely to clash
3787 with other external names.
3788
3789 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3790 a different (DFA) algorithm. Although it is slower than the original
3791 function, it does have some advantages for certain types of matching
3792 problem.
3793
3794 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3795 including restarting after a partial match.
3796
3797 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3798 defined when compiling for Windows was sent to me. I have put it into the
3799 code, though I have no means of testing or verifying it.
3800
3801 7. Added the pcre_refcount() auxiliary function.
3802
3803 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3804 match before or at the first newline in the subject string. In pcretest,
3805 the /f option on a pattern can be used to set this.
3806
3807 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3808 would behave wrongly. This has been present in PCRE since release 4.0.
3809
3810 10. A number of changes to the pcregrep command:
3811
3812 (a) Refactored how -x works; insert ^(...)$ instead of setting
3813 PCRE_ANCHORED and checking the length, in preparation for adding
3814 something similar for -w.
3815
3816 (b) Added the -w (match as a word) option.
3817
3818 (c) Refactored the way lines are read and buffered so as to have more
3819 than one at a time available.
3820
3821 (d) Implemented a pcregrep test script.
3822
3823 (e) Added the -M (multiline match) option. This allows patterns to match
3824 over several lines of the subject. The buffering ensures that at least
3825 8K, or the rest of the document (whichever is the shorter) is available
3826 for matching (and similarly the previous 8K for lookbehind assertions).
3827
3828 (f) Changed the --help output so that it now says
3829
3830 -w, --word-regex(p)
3831
3832 instead of two lines, one with "regex" and the other with "regexp"
3833 because that confused at least one person since the short forms are the
3834 same. (This required a bit of code, as the output is generated
3835 automatically from a table. It wasn't just a text change.)
3836
3837 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3838 option but starts with a hyphen. Could be a pattern or a path name
3839 starting with a hyphen, for instance.
3840
3841 (h) "-" can be given as a file name to represent stdin.
3842
3843 (i) When file names are being printed, "(standard input)" is used for
3844 the standard input, for compatibility with GNU grep. Previously
3845 "<stdin>" was used.
3846
3847 (j) The option --label=xxx can be used to supply a name to be used for
3848 stdin when file names are being printed. There is no short form.
3849
3850 (k) Re-factored the options decoding logic because we are going to add
3851 two more options that take data. Such options can now be given in four
3852 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3853
3854 (l) Added the -A, -B, and -C options for requesting that lines of context
3855 around matches be printed.
3856
3857 (m) Added the -L option to print the names of files that do not contain
3858 any matching lines, that is, the complement of -l.
3859
3860 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3861 continue to scan other files.
3862
3863 (o) The -s option was incorrectly implemented. For compatibility with other
3864 greps, it now suppresses the error message for a non-existent or non-
3865 accessible file (but not the return code). There is a new option called
3866 -q that suppresses the output of matching lines, which was what -s was
3867 previously doing.
3868
3869 (p) Added --include and --exclude options to specify files for inclusion
3870 and exclusion when recursing.
3871
3872 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3873 Hopefully, it now does.
3874
3875 12. Missing cast in pcre_study().
3876
3877 13. Added an "uninstall" target to the makefile.
3878
3879 14. Replaced "extern" in the function prototypes in Makefile.in with
3880 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3881 world, but is set differently for Windows.
3882
3883 15. Added a second compiling function called pcre_compile2(). The only
3884 difference is that it has an extra argument, which is a pointer to an
3885 integer error code. When there is a compile-time failure, this is set
3886 non-zero, in addition to the error test pointer being set to point to an
3887 error message. The new argument may be NULL if no error number is required
3888 (but then you may as well call pcre_compile(), which is now just a
3889 wrapper). This facility is provided because some applications need a
3890 numeric error indication, but it has also enabled me to tidy up the way
3891 compile-time errors are handled in the POSIX wrapper.
3892
3893 16. Added VPATH=.libs to the makefile; this should help when building with one
3894 prefix path and installing with another. (Or so I'm told by someone who
3895 knows more about this stuff than I do.)
3896
3897 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3898 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3899 match everything, including newlines. This is not POSIX-compatible, but
3900 somebody wanted the feature. From pcretest it can be activated by using
3901 both the P and the s flags.
3902
3903 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3904
3905 19. libpcre.pc was being incorrectly installed as executable.
3906
3907 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3908 it now also looks for '\r' so that it will work unmodified on Windows.
3909
3910 21. Added Google's contributed C++ wrapper to the distribution.
3911
3912 22. Added some untidy missing memory free() calls in pcretest, to keep
3913 Electric Fence happy when testing.
3914
3915
3916
3917 Version 5.0 13-Sep-04
3918 ---------------------
3919
3920 1. Internal change: literal characters are no longer packed up into items
3921 containing multiple characters in a single byte-string. Each character
3922 is now matched using a separate opcode. However, there may be more than one
3923 byte in the character in UTF-8 mode.
3924
3925 2. The pcre_callout_block structure has two new fields: pattern_position and
3926 next_item_length. These contain the offset in the pattern to the next match
3927 item, and its length, respectively.
3928
3929 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3930 insertion of callouts before each pattern item. Added the /C option to
3931 pcretest to make use of this.
3932
3933 4. On the advice of a Windows user, the lines
3934
3935 #if defined(_WIN32) || defined(WIN32)
3936 _setmode( _fileno( stdout ), 0x8000 );
3937 #endif /* defined(_WIN32) || defined(WIN32) */
3938
3939 have been added to the source of pcretest. This apparently does useful
3940 magic in relation to line terminators.
3941
3942 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3943 for the benefit of those environments where the "b" makes a difference.
3944
3945 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3946 to know about it. I have put a hack into configure.in that adds in code
3947 to set GCC=yes if CC=icc. This seems to end up at a point in the
3948 generated configure script that is early enough to affect the setting of
3949 compiler options, which is what is needed, but I have no means of testing
3950 whether it really works. (The user who reported this had patched the
3951 generated configure script, which of course I cannot do.)
3952
3953 LATER: After change 22 below (new libtool files), the configure script
3954 seems to know about icc (and also ecc). Therefore, I have commented out
3955 this hack in configure.in.
3956
3957 7. Added support for pkg-config (2 patches were sent in).
3958
3959 8. Negated POSIX character classes that used a combination of internal tables
3960 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3961 [[:^ascii]]. Typically, they would match almost any characters. The other
3962 POSIX classes were not broken in this way.
3963
3964 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3965 to find the match, as PCRE was deluded into thinking that the match had to
3966 start at the start point or following a newline. The same bug applied to
3967 patterns with negative forward assertions or any backward assertions
3968 preceding ".*" at the start, unless the pattern required a fixed first
3969 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3970
3971 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3972 starting at the last subject character, bytes beyond the end of the subject
3973 string were read.
3974
3975 11. Renamed the variable "class" as "classbits" to make life easier for C++
3976 users. (Previously there was a macro definition, but it apparently wasn't
3977 enough.)
3978
3979 12. Added the new field "tables" to the extra data so that tables can be passed
3980 in at exec time, or the internal tables can be re-selected. This allows
3981 a compiled regex to be saved and re-used at a later time by a different
3982 program that might have everything at different addresses.
3983
3984 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3985 -R library as well as a -L library.
3986
3987 14. The debugging options of pcretest (-d on the command line or D on a
3988 pattern) showed incorrect output for anything following an extended class
3989 that contained multibyte characters and which was followed by a quantifier.
3990
3991 15. Added optional support for general category Unicode character properties
3992 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3993 support. It adds about 90K to the size of the library. The meanings of the
3994 inbuilt class escapes such as \d and \s have NOT been changed.
3995
3996 16. Updated pcredemo.c to include calls to free() to release the memory for the
3997 compiled pattern.
3998
3999 17. The generated file chartables.c was being created in the source directory
4000 instead of in the building directory. This caused the build to fail if the
4001 source directory was different from the building directory, and was
4002 read-only.
4003
4004 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4005 file. No doubt somebody will tell me if they don't make sense... Also added
4006 Dan Mooney's comments about building on OpenVMS.
4007
4008 19. Added support for partial matching via the PCRE_PARTIAL option for
4009 pcre_exec() and the \P data escape in pcretest.
4010
4011 20. Extended pcretest with 3 new pattern features:
4012
4013 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4014 write the compiled pattern to the file whose name is "rest-of-line".
4015 This is a straight binary dump of the data, with the saved pointer to
4016 the character tables forced to be NULL. The study data, if any, is
4017 written too. After writing, pcretest reads a new pattern.
4018
4019 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4020 compiled pattern from the given file. There must not be any
4021 occurrences of "<" in the file name (pretty unlikely); if there are,
4022 pcretest will instead treat the initial "<" as a pattern delimiter.
4023 After reading in the pattern, pcretest goes on to read data lines as
4024 usual.
4025
4026 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4027 and 16-bit fields in a compiled pattern, to simulate a pattern that
4028 was compiled on a host of opposite endianness.
4029
4030 21. The pcre-exec() function can now cope with patterns that were compiled on
4031 hosts of opposite endianness, with this restriction:
4032
4033 As for any compiled expression that is saved and used later, the tables
4034 pointer field cannot be preserved; the extra_data field in the arguments
4035 to pcre_exec() should be used to pass in a tables address if a value
4036 other than the default internal tables were used at compile time.
4037
4038 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4039 now diagnosed as an error. Previously, most of the time, a negative number
4040 would have been treated as zero, but if in addition "ovector" was passed as
4041 NULL, a crash could occur.
4042
4043 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4044 new versions from the libtool 1.5 distribution (the last one is a copy of
4045 a file called libtool.m4). This seems to have fixed the need to patch
4046 "configure" to support Darwin 1.3 (which I used to do). However, I still
4047 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4048 workstation).
4049
4050 24. Changed the PCRE licence to be the more standard "BSD" licence.
4051
4052
4053 Version 4.5 01-Dec-03
4054 ---------------------
4055
4056 1. There has been some re-arrangement of the code for the match() function so
4057 that it can be compiled in a version that does not call itself recursively.
4058 Instead, it keeps those local variables that need separate instances for
4059 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4060 needs to "recurse". Keeping track of where control must go is done by means
4061 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4062 hide most of the details from the main code, and operates only if
4063 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4064 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4065 operating.
4066
4067 To make it easier for callers to provide specially tailored get/free
4068 functions for this usage, two new functions, pcre_stack_malloc, and
4069 pcre_stack_free, are used. They are always called in strict stacking order,
4070 and the size of block requested is always the same.
4071
4072 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4073 PCRE has been compiled to use the stack or the heap for recursion. The
4074 -C option of pcretest uses this to show which version is compiled.
4075
4076 A new data escape \S, is added to pcretest; it causes the amounts of store
4077 obtained and freed by both kinds of malloc/free at match time to be added
4078 to the output.
4079
4080 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4081 what's available on my current Linux desktop machine.
4082
4083 3. When matching a UTF-8 string, the test for a valid string at the start has
4084 been extended. If start_offset is not zero, PCRE now checks that it points
4085 to a byte that is the start of a UTF-8 character. If not, it returns
4086 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4087 this is necessary because there may be backward assertions in the pattern.
4088 When matching the same subject several times, it may save resources to use
4089 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4090
4091 4. The code for checking the validity of UTF-8 strings has been tightened so
4092 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4093 containing "overlong sequences".
4094
4095 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4096 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4097 should have been "&", but it just so happened that all the cases this let
4098 through by mistake were picked up later in the function.
4099
4100 6. I had used a variable called "isblank" - this is a C99 function, causing
4101 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4102
4103 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4104 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4105 so the version is shown; (c) stop "make test" echoing "./RunTest".
4106
4107 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4108
4109 9. The prototype for memmove() for systems that don't have it was using
4110 size_t, but the inclusion of the header that defines size_t was later. I've
4111 moved the #includes for the C headers earlier to avoid this.
4112
4113 10. Added some adjustments to the code to make it easier to compiler on certain
4114 special systems:
4115
4116 (a) Some "const" qualifiers were missing.
4117 (b) Added the macro EXPORT before all exported functions; by default this
4118 is defined to be empty.
4119 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4120 that it reads its output file name as an argument instead of writing
4121 to the standard output and assuming this can be redirected.
4122
4123 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4124 class containing characters with values greater than 255, PCRE compilation
4125 went into a loop.
4126
4127 12. A recursive reference to a subpattern that was within another subpattern
4128 that had a minimum quantifier of zero caused PCRE to crash. For example,
4129 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4130 recursion. If the recursively-called subpattern itself had a zero repeat,
4131 that was OK.
4132
4133 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4134 buffer into which it was copied (for escape processing) was still set at
4135 1024, so long lines caused crashes.
4136
4137 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4138 "internal error: code overflow...". This applied to any character class
4139 that was followed by a possessive quantifier.
4140
4141 15. Modified the Makefile to add libpcre.la as a prerequisite for
4142 libpcreposix.la because I was told this is needed for a parallel build to
4143 work.
4144
4145 16. If a pattern that contained .* following optional items at the start was
4146 studied, the wrong optimizing data was generated, leading to matching
4147 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4148 matching string must start with a or b or c. The correct conclusion for
4149 this pattern is that a match can start with any character.
4150
4151
4152 Version 4.4 13-Aug-03
4153 ---------------------
4154
4155 1. In UTF-8 mode, a character class containing characters with values between
4156 127 and 255 was not handled correctly if the compiled pattern was studied.
4157 In fixing this, I have also improved the studying algorithm for such
4158 classes (slightly).
4159
4160 2. Three internal functions had redundant arguments passed to them. Removal
4161 might give a very teeny performance improvement.
4162
4163 3. Documentation bug: the value of the capture_top field in a callout is *one
4164 more than* the number of the hightest numbered captured substring.
4165
4166 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4167 in incorrectly linking with a previously installed version. They now link
4168 explicitly with libpcre.la.
4169
4170 5. configure.in no longer needs to recognize Cygwin specially.
4171
4172 6. A problem in pcre.in for Windows platforms is fixed.
4173
4174 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4175 pcretest, it used to include the size of the study block as part of its
4176 output. Unfortunately, the structure contains a field that has a different
4177 size on different hardware architectures. This meant that the tests that
4178 showed this size failed. As the block is currently always of a fixed size,
4179 this information isn't actually particularly useful in pcretest output, so
4180 I have just removed it.
4181
4182 8. Three pre-processor statements accidentally did not start in column 1.
4183 Sadly, there are *still* compilers around that complain, even though
4184 standard C has not required this for well over a decade. Sigh.
4185
4186 9. In pcretest, the code for checking callouts passed small integers in the
4187 callout_data field, which is a void * field. However, some picky compilers
4188 complained about the casts involved for this on 64-bit systems. Now
4189 pcretest passes the address of the small integer instead, which should get
4190 rid of the warnings.
4191
4192 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4193 both compile and run time, and gives an error if an invalid UTF-8 sequence
4194 is found. There is a option for disabling this check in cases where the
4195 string is known to be correct and/or the maximum performance is wanted.
4196
4197 11. In response to a bug report, I changed one line in Makefile.in from
4198
4199 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4200 to
4201 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4202
4203 to look similar to other lines, but I have no way of telling whether this
4204 is the right thing to do, as I do not use Windows. No doubt I'll get told
4205 if it's wrong...
4206
4207
4208 Version 4.3 21-May-03
4209 ---------------------
4210
4211 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4212 Makefile.
4213
4214 2. Some refactoring to improve the quality of the code:
4215
4216 (i) The utf8_table... variables are now declared "const".
4217
4218 (ii) The code for \cx, which used the "case flipping" table to upper case
4219 lower case letters, now just substracts 32. This is ASCII-specific,
4220 but the whole concept of \cx is ASCII-specific, so it seems
4221 reasonable.
4222
4223 (iii) PCRE was using its character types table to recognize decimal and
4224 hexadecimal digits in the pattern. This is silly, because it handles
4225 only 0-9, a-f, and A-F, but the character types table is locale-
4226 specific, which means strange things might happen. A private
4227 table is now used for this - though it costs 256 bytes, a table is
4228 much faster than multiple explicit tests. Of course, the standard
4229 character types table is still used for matching digits in subject
4230 strings against \d.
4231
4232 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4233 ending in _t are). So I've renamed it as ESC_tee.
4234
4235 3. The first argument for regexec() in the POSIX wrapper should have been
4236 defined as "const".
4237
4238 4. Changed pcretest to use malloc() for its buffers so that they can be
4239 Electric Fenced for debugging.
4240
4241 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4242 to read one or more bytes before the start of the subject string. Often this
4243 had no effect on PCRE's behaviour, but in some circumstances it could
4244 provoke a segmentation fault.
4245
4246 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4247 to try to read one or more bytes before the start of the subject string.
4248
4249 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4250 UTF-8 support could misbehave in various ways if the subject string
4251 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4252 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4253 back over UTF-8 characters.)
4254
4255
4256 Version 4.2 14-Apr-03
4257 ---------------------
4258
4259 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4260
4261 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4262 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4263 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4264 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4265 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4266 and BUILD_EXEEXT
4267 Note: automatic setting of the BUILD variables is not yet working
4268 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4269 compile-time but not at link-time
4270 [LINK]: use for linking executables only
4271 make different versions for Windows and non-Windows
4272 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4273 libraries
4274 [LINK_FOR_BUILD]: new variable
4275 [OBJEXT]: use throughout
4276 [EXEEXT]: use throughout
4277 <winshared>: new target
4278 <wininstall>: new target
4279 <dftables.o>: use native compiler
4280 <dftables>: use native linker
4281 <install>: handle Windows platform correctly
4282 <clean>: ditto
4283 <check>: ditto
4284 copy DLL to top builddir before testing
4285
4286 As part of these changes, -no-undefined was removed again. This was reported
4287 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4288 in any case.
4289
4290 3. Some tidies to get rid of compiler warnings:
4291
4292 . In the match_data structure, match_limit was an unsigned long int, whereas
4293 match_call_count was an int. I've made them both unsigned long ints.
4294
4295 . In pcretest the fact that a const uschar * doesn't automatically cast to
4296 a void * provoked a warning.
4297
4298 . Turning on some more compiler warnings threw up some "shadow" variables
4299 and a few more missing casts.
4300
4301 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4302 option, a class that contained a single character with a value between 128
4303 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4304
4305 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4306 option, a class that contained several characters, but with at least one
4307 whose value was between 128 and 255 caused PCRE to crash.
4308
4309
4310 Version 4.1 12-Mar-03
4311 ---------------------
4312
4313 1. Compiling with gcc -pedantic found a couple of places where casts were
4314 needed, and a string in dftables.c that was longer than standard compilers are
4315 required to support.
4316
4317 2. Compiling with Sun's compiler found a few more places where the code could
4318 be tidied up in order to avoid warnings.
4319
4320 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4321 first of these names is deprecated in the latest Autoconf in favour of the name
4322 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4323 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4324 analogy I have changed it to CFLAGS_FOR_BUILD.
4325
4326 4. Added -no-undefined to the linking command in the Makefile, because this is
4327 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4328 linking step for the pcreposix library.
4329
4330 5. PCRE was failing to diagnose the case of two named groups with the same
4331 name.
4332
4333 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4334 literal character that is needed in the subject for a match, and scans along to
4335 ensure that it is present before embarking on the full matching process. This
4336 saves time in cases of nested unlimited repeats that are never going to match.
4337 Problem: the scan can take a lot of time if the subject is very long (e.g.
4338 megabytes), thus penalizing straightforward matches. It is now done only if the
4339 amount of subject to be scanned is less than 1000 bytes.
4340
4341 7. A lesser problem with the same optimization is that it was recording the
4342 first character of an anchored pattern as "needed", thus provoking a search
4343 right along the subject, even when the first match of the pattern was going to
4344 fail. The "needed" character is now not set for anchored patterns, unless it
4345 follows something in the pattern that is of non-fixed length. Thus, it still
4346 fulfils its original purpose of finding quick non-matches in cases of nested
4347 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4348
4349
4350 Version 4.0 17-Feb-03
4351 ---------------------
4352
4353 1. If a comment in an extended regex that started immediately after a meta-item
4354 extended to the end of string, PCRE compiled incorrect data. This could lead to
4355 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4356
4357 2. Moved to autoconf 2.53 and libtool 1.4.2.
4358
4359 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4360 the special perltest8 script is no longer needed - all the tests can be run
4361 from a single perltest script.
4362
4363 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4364 by \s. It has now been removed in PCRE. This means it isn't recognized as
4365 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4366 class [:space:] *does* include VT, thereby creating a mess.
4367
4368 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4369 space and tab.
4370
4371 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4372 its new features into the main test script, reducing the number of scripts.
4373
4374 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4375 were backward compatible, and made the (?i) apply to the whole pattern, as if
4376 /i were given. Now it behaves more logically, and applies the option setting
4377 only to what follows. PCRE has been changed to follow suit. However, if it
4378 finds options settings right at the start of the pattern, it extracts them into
4379 the global options, as before. Thus, they show up in the info data.
4380
4381 8. Added support for the \Q...\E escape sequence. Characters in between are
4382 treated as literals. This is slightly different from Perl in that $ and @ are
4383 also handled as literals inside the quotes. In Perl, they will cause variable
4384 interpolation. Note the following examples:
4385
4386 Pattern PCRE matches Perl matches
4387
4388 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4389 \Qabc\$xyz\E abc\$xyz abc\$xyz
4390 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4391
4392 For compatibility with Perl, \Q...\E sequences are recognized inside character
4393 classes as well as outside them.
4394
4395 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4396 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4397 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4398 signed/unsigned warnings.
4399
4400 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4401 option for pcretest, so I've replaced it by a simple function that does just
4402 that job.
4403
4404 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4405 "pcregrep -".
4406
4407 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4408 Java package. This provides some syntactic sugar for simple cases of what my
4409 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4410 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4411 item, you can use this simplified notation. Note that only makes sense with
4412 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4413 greediness, whatever the setting of the PCRE_UNGREEDY option.
4414
4415 13. A change of greediness default within a pattern was not taking effect at
4416 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4417 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4418 was abstracted outside.
4419
4420 14. PCRE now supports the \G assertion. It is true when the current matching
4421 position is at the start point of the match. This differs from \A when the
4422 starting offset is non-zero. Used with the /g option of pcretest (or similar
4423 code), it works in the same way as it does for Perl's /g option. If all
4424 alternatives of a regex begin with \G, the expression is anchored to the start
4425 match position, and the "anchored" flag is set in the compiled expression.
4426
4427 15. Some bugs concerning the handling of certain option changes within patterns
4428 have been fixed. These applied to options other than (?ims). For example,
4429 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4430 been the other way round. Some of this was related to change 7 above.
4431
4432 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4433 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4434 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4435 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4436
4437 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4438 mode. Unlike ".", it always matches newline, whatever the setting of
4439 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4440 assertions. Perl allows it, but it doesn't (in general) work because it can't
4441 calculate the length of the lookbehind. At least, that's the case for Perl
4442 5.8.0 - I've been told they are going to document that it doesn't work in
4443 future.
4444
4445 18. Added an error diagnosis for escapes that PCRE does not support: these are
4446 \L, \l, \N, \P, \p, \U, \u, and \X.
4447
4448 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4449 reading past the end of the pattern in cases such as /[abcd/.
4450
4451 20. PCRE was getting more memory than necessary for patterns with classes that
4452 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4453
4454 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4455 compiling PCRE for use with Virtual Pascal.
4456
4457 22. Small fix to the Makefile to make it work properly if the build is done
4458 outside the source tree.
4459
4460 23. Added a new extension: a condition to go with recursion. If a conditional
4461 subpattern starts with (?(R) the "true" branch is used if recursion has
4462 happened, whereas the "false" branch is used only at the top level.
4463
4464 24. When there was a very long string of literal characters (over 255 bytes
4465 without UTF support, over 250 bytes with UTF support), the computation of how
4466 much memory was required could be incorrect, leading to segfaults or other
4467 strange effects.
4468
4469 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4470 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4471 there was a subsequent back reference to those brackets. This meant that, for
4472 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4473 possible to check for precisely this case. All we can do is abandon the
4474 optimization if .* occurs inside capturing brackets when there are any back
4475 references whatsoever. (See below for a better fix that came later.)
4476
4477 26. The handling of the optimization for finding the first character of a
4478 non-anchored pattern, and for finding a character that is required later in the
4479 match were failing in some cases. This didn't break the matching; it just
4480 failed to optimize when it could. The way this is done has been re-implemented.
4481
4482 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4483
4484 28. Added a new feature that provides some of the functionality that Perl
4485 provides with (?{...}). The facility is termed a "callout". The way it is done
4486 in PCRE is for the caller to provide an optional function, by setting
4487 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4488 global variable. By default it is unset, which disables all calling out. To get
4489 the function called, the regex must include (?C) at appropriate points. This
4490 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4491 This provides a means of identifying different callout points. When PCRE
4492 reaches such a point in the regex, if pcre_callout has been set, the external
4493 function is called. It is provided with data in a structure called
4494 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4495 matching continues; if it returns a non-zero value, the match at the current
4496 point fails. However, backtracking will occur if possible. [This was changed
4497 later and other features added - see item 49 below.]
4498
4499 29. pcretest is upgraded to test the callout functionality. It provides a
4500 callout function that displays information. By default, it shows the start of
4501 the match and the current position in the text. There are some new data escapes
4502 to vary what happens:
4503
4504 \C+ in addition, show current contents of captured substrings
4505 \C- do not supply a callout function
4506 \C!n return 1 when callout number n is reached
4507 \C!n!m return 1 when callout number n is reached for the mth time
4508
4509 30. If pcregrep was called with the -l option and just a single file name, it
4510 output "<stdin>" if a match was found, instead of the file name.
4511
4512 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4513 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4514 pcre_exec(). This saves a malloc/free per call. The default value of
4515 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4516 when configuring.
4517
4518 32. The default maximum size of a compiled pattern is 64K. There have been a
4519 few cases of people hitting this limit. The code now uses macros to handle the
4520 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4521 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4522 configuring. Tests 2 and 5 work only with 2-byte links because they output
4523 debugging information about compiled patterns.
4524
4525 33. Internal code re-arrangements:
4526
4527 (a) Moved the debugging function for printing out a compiled regex into
4528 its own source file (printint.c) and used #include to pull it into
4529 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4530 separate copies.
4531
4532 (b) Defined the list of op-code names for debugging as a macro in
4533 internal.h so that it is next to the definition of the opcodes.
4534
4535 (c) Defined a table of op-code lengths for simpler skipping along compiled
4536 code. This is again a macro in internal.h so that it is next to the
4537 definition of the opcodes.
4538
4539 34. Added support for recursive calls to individual subpatterns, along the
4540 lines of Robin Houston's patch (but implemented somewhat differently).
4541
4542 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4543 allow it to read and process whole directories in Win32. This code was
4544 contributed by Lionel Fourquaux; it has not been tested by me.
4545
4546 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4547 used to name a group. Names consist of alphanumerics and underscores, and must
4548 be unique. Back references use the syntax (?P=name) and recursive calls use
4549 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4550 numbers. The function pcre_fullinfo() can be used after compilation to extract
4551 a name/number map. There are three relevant calls:
4552
4553 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4554 PCRE_INFO_NAMECOUNT yields the number of entries
4555 PCRE_INFO_NAMETABLE yields a pointer to the map.
4556
4557 The map is a vector of fixed-size entries. The size of each entry depends on
4558 the length of the longest name used. The first two bytes of each entry are the
4559 group number, most significant byte first. There follows the corresponding
4560 name, zero terminated. The names are in alphabetical order.
4561
4562 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4563 case instead of 255. Making it the same both with and without UTF-8 support
4564 means that the same test output works with both.
4565
4566 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4567 calling malloc() with a zero argument.
4568
4569 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4570 optimization. I've improved things by keeping a bitmap of backreferences with
4571 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4572 fact referenced, the optimization can be applied. It is unlikely that a
4573 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4574 the match to follow \n) will appear inside brackets with a number greater than
4575 31, but if it does, any back reference > 31 suppresses the optimization.
4576
4577 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4578 of disabling numbered capturing parentheses. Any opening parenthesis that is
4579 not followed by ? behaves as if it were followed by ?: but named parentheses
4580 can still be used for capturing (and they will acquire numbers in the usual
4581 way).
4582
4583 41. Redesigned the return codes from the match() function into yes/no/error so
4584 that errors can be passed back from deep inside the nested calls. A malloc
4585 failure while inside a recursive subpattern call now causes the
4586 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4587
4588 42. It is now possible to set a limit on the number of times the match()
4589 function is called in a call to pcre_exec(). This facility makes it possible to
4590 limit the amount of recursion and backtracking, though not in a directly
4591 obvious way, because the match() function is used in a number of different
4592 circumstances. The count starts from zero for each position in the subject
4593 string (for non-anchored patterns). The default limit is, for compatibility, a
4594 large number, namely 10 000 000. You can change this in two ways:
4595
4596 (a) When configuring PCRE before making, you can use --with-match-limit=n
4597 to set a default value for the compiled library.
4598
4599 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4600 a different value is set. See 45 below.
4601
4602 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4603
4604 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4605 of things that can be changed at compile time. The first argument specifies
4606 what is wanted and the second points to where the information is to be placed.
4607 The current list of available information is:
4608
4609 PCRE_CONFIG_UTF8
4610
4611 The output is an integer that is set to one if UTF-8 support is available;
4612 otherwise it is set to zero.
4613
4614 PCRE_CONFIG_NEWLINE
4615
4616 The output is an integer that it set to the value of the code that is used for
4617 newline. It is either LF (10) or CR (13).
4618
4619 PCRE_CONFIG_LINK_SIZE
4620
4621 The output is an integer that contains the number of bytes used for internal
4622 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4623
4624 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4625
4626 The output is an integer that contains the threshold above which the POSIX
4627 interface uses malloc() for output vectors. See item 31 above.
4628
4629 PCRE_CONFIG_MATCH_LIMIT
4630
4631 The output is an unsigned integer that contains the default limit of the number
4632 of match() calls in a pcre_exec() execution. See 42 above.
4633
4634 44. pcretest has been upgraded by the addition of the -C option. This causes it
4635 to extract all the available output from the new pcre_config() function, and to
4636 output it. The program then exits immediately.
4637
4638 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4639 order to support additional features. One way would have been to define
4640 pcre_exec2() (for example) with extra arguments, but this would not have been
4641 extensible, and would also have required all calls to the original function to
4642 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4643 is used for passing in "extra" data from pcre_study().
4644
4645 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4646 contains the following fields:
4647
4648 flags a bitmap indicating which of the following fields are set
4649 study_data opaque data from pcre_study()
4650 match_limit a way of specifying a limit on match() calls for a specific
4651 call to pcre_exec()
4652 callout_data data for callouts (see 49 below)
4653
4654 The flag bits are also defined in pcre.h, and are
4655
4656 PCRE_EXTRA_STUDY_DATA
4657 PCRE_EXTRA_MATCH_LIMIT
4658 PCRE_EXTRA_CALLOUT_DATA
4659
4660 The pcre_study() function now returns one of these new pcre_extra blocks, with
4661 the actual study data pointed to by the study_data field, and the
4662 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4663 before. That is, this change is entirely upwards-compatible and requires no
4664 change to existing code.
4665
4666 If you want to pass in additional data to pcre_exec(), you can either place it
4667 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4668 block.
4669
4670 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4671 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4672 times with different match limits, until it finds the minimum value needed for
4673 pcre_exec() to complete. The value is then output. This can be instructive; for
4674 most simple matches the number is quite small, but for pathological cases it
4675 gets very large very quickly.
4676
4677 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4678 returns the size of the data block pointed to by the study_data field in a
4679 pcre_extra block, that is, the value that was passed as the argument to
4680 pcre_malloc() when PCRE was getting memory in which to place the information
4681 created by pcre_study(). The fourth argument should point to a size_t variable.
4682 pcretest has been extended so that this information is shown after a successful
4683 pcre_study() call when information about the compiled regex is being displayed.
4684
4685 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4686 because what follows is always an absolute path. (Later: it turns out that this
4687 is more than cosmetic for MinGW, because it doesn't like empty path
4688 components.)
4689
4690 49. Some changes have been made to the callout feature (see 28 above):
4691
4692 (i) A callout function now has three choices for what it returns:
4693
4694 0 => success, carry on matching
4695 > 0 => failure at this point, but backtrack if possible
4696 < 0 => serious error, return this value from pcre_exec()
4697
4698 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4699 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4700 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4701 use by callout functions. It will never be used by PCRE itself.
4702
4703 (ii) The pcre_extra structure (see 45 above) has a void * field called
4704 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4705 pcre_callout_block structure has a field of the same name. The contents of
4706 the field passed in the pcre_extra structure are passed to the callout
4707 function in the corresponding field in the callout block. This makes it
4708 easier to use the same callout-containing regex from multiple threads. For
4709 testing, the pcretest program has a new data escape
4710
4711 \C*n pass the number n (may be negative) as callout_data
4712
4713 If the callout function in pcretest receives a non-zero value as
4714 callout_data, it returns that value.
4715
4716 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4717 there were some redundant $(CFLAGS) in commands that are now specified as
4718 $(LINK), which already includes $(CFLAGS).
4719
4720 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4721 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4722 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4723 one-byte characters throughout. Note that case-insensitive matching applies
4724 only to characters whose values are less than 256. PCRE doesn't support the
4725 notion of cases for higher-valued characters.
4726
4727 (i) A character class whose characters are all within 0-255 is handled as
4728 a bit map, and the map is inverted for negative classes. Previously, a
4729 character > 255 always failed to match such a class; however it should
4730 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4731
4732 (ii) A negated character class with a single character < 255 is coded as
4733 "not this character" (OP_NOT). This wasn't working properly when the test
4734 character was multibyte, either singly or repeated.
4735
4736 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4737 mode, for example: \x{100}{2,3}.
4738
4739 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4740 singly or repeated) now correctly test multibyte characters. However,
4741 PCRE doesn't recognize any characters with values greater than 255 as
4742 digits, spaces, or word characters. Such characters always match \D, \S,
4743 and \W, and never match \d, \s, or \w.
4744
4745 (v) Classes may now contain characters and character ranges with values
4746 greater than 255. For example: [ab\x{100}-\x{400}].
4747
4748 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4749 PCRE in UTF-8 mode.
4750
4751 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4752 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4753 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4754 value.)
4755
4756 53. The single man page has become too large. I have therefore split it up into
4757 a number of separate man pages. These also give rise to individual HTML pages;
4758 these are now put in a separate directory, and there is an index.html page that
4759 lists them all. Some hyperlinking between the pages has been installed.
4760
4761 54. Added convenience functions for handling named capturing parentheses.
4762
4763 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4764 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4765 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4766 are faulted.
4767
4768 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4769 calling configure. These values are used when compiling the dftables.c program
4770 which is run to generate the source of the default character tables. They
4771 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4772 you will need to set these values.
4773
4774 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4775
4776
4777 Version 3.9 02-Jan-02
4778 ---------------------
4779
4780 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4781
4782 2. If --disable-static was given, the building process failed when trying to
4783 build pcretest and pcregrep. (For some reason it was using libtool to compile
4784 them, which is not right, as they aren't part of the library.)
4785
4786
4787 Version 3.8 18-Dec-01
4788 ---------------------
4789
4790 1. The experimental UTF-8 code was completely screwed up. It was packing the
4791 bytes in the wrong order. How dumb can you get?
4792
4793
4794 Version 3.7 29-Oct-01
4795 ---------------------
4796
4797 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4798 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4799 this didn't happen under Solaris 8, where I normally test things.
4800
4801 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4802 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4803 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4804 this fix an hour or so after the initial 3.7 release.)
4805
4806
4807 Version 3.6 23-Oct-01
4808 ---------------------
4809
4810 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4811 offsets passed as NULL with zero offset count.
4812
4813 2. The config.guess and config.sub files had not been updated when I moved to
4814 the latest autoconf.
4815
4816
4817 Version 3.5 15-Aug-01
4818 ---------------------
4819
4820 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4821 had been forgotten.
4822
4823 2. By using declared but undefined structures, we can avoid using "void"
4824 definitions in pcre.h while keeping the internal definitions of the structures
4825 private.
4826
4827 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4828 user point of view, this means that both static and shared libraries are built
4829 by default, but this can be individually controlled. More of the work of
4830 handling this static/shared cases is now inside libtool instead of PCRE's make
4831 file.
4832
4833 4. The pcretest utility is now installed along with pcregrep because it is
4834 useful for users (to test regexs) and by doing this, it automatically gets
4835 relinked by libtool. The documentation has been turned into a man page, so
4836 there are now .1, .txt, and .html versions in /doc.
4837
4838 5. Upgrades to pcregrep:
4839 (i) Added long-form option names like gnu grep.
4840 (ii) Added --help to list all options with an explanatory phrase.
4841 (iii) Added -r, --recursive to recurse into sub-directories.
4842 (iv) Added -f, --file to read patterns from a file.
4843
4844 6. pcre_exec() was referring to its "code" argument before testing that
4845 argument for NULL (and giving an error if it was NULL).
4846
4847 7. Upgraded Makefile.in to allow for compiling in a different directory from
4848 the source directory.
4849
4850 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4851 options bits, the pointer it was passed was to an int instead of to an unsigned
4852 long int. This mattered only on 64-bit systems.
4853
4854 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4855 generated) instead of pcre.in, which it its source. Also made the same change
4856 in several of the .c files.
4857
4858 10. A new release of gcc defines printf() as a macro, which broke pcretest
4859 because it had an ifdef in the middle of a string argument for printf(). Fixed
4860 by using separate calls to printf().
4861
4862 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4863 script, to force use of CR or LF instead of \n in the source. On non-Unix
4864 systems, the value can be set in config.h.
4865
4866 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4867 absolute limit. Changed the text of the error message to make this clear, and
4868 likewise updated the man page.
4869
4870 13. The limit of 99 on the number of capturing subpatterns has been removed.
4871 The new limit is 65535, which I hope will not be a "real" limit.
4872
4873
4874 Version 3.4 22-Aug-00
4875 ---------------------
4876
4877 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4878
4879 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4880
4881
4882 Version 3.3 01-Aug-00
4883 ---------------------
4884
4885 1. If an octal character was given, but the value was greater than \377, it
4886 was not getting masked to the least significant bits, as documented. This could
4887 lead to crashes in some systems.
4888
4889 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4890 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4891
4892 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4893 These just pass their arguments on to (pcre_free)(), but they are provided
4894 because some uses of PCRE bind it to non-C systems that can call its functions,
4895 but cannot call free() or pcre_free() directly.
4896
4897 4. Add "make test" as a synonym for "make check". Corrected some comments in
4898 the Makefile.
4899
4900 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4901 Makefile.
4902
4903 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4904 command called pgrep for grepping around the active processes.
4905
4906 7. Added the beginnings of support for UTF-8 character strings.
4907
4908 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4909 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4910 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4911 out for the ar command.)
4912
4913
4914 Version 3.2 12-May-00
4915 ---------------------
4916
4917 This is purely a bug fixing release.
4918
4919 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4920 of ZA. This was just one example of several cases that could provoke this bug,
4921 which was introduced by change 9 of version 2.00. The code for breaking
4922 infinite loops after an iteration that matches an empty string was't working
4923 correctly.
4924
4925 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4926 when matched against abbab (for example). After matching an empty string, it
4927 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4928 caused it to match further down the string than it should.
4929
4930 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4931 was there because it doesn't seem to be needed, and it causes trouble on some
4932 systems, as it is not a Standard C header. It has been removed.
4933
4934 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4935 were reported on the Macintosh. The changes were from
4936
4937 while ((c = *(++ptr)) != 0 && c != '\n');
4938 to
4939 while ((c = *(++ptr)) != 0 && c != '\n') ;
4940
4941 Totally extraordinary, but if that's what it takes...
4942
4943 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4944 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4945 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4946 assumes the way PCRE uses memmove() (always moving upwards).
4947
4948 6. PCRE is being used in one environment where strchr() is not available. There
4949 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4950 faster code anyway.
4951
4952
4953 Version 3.1 09-Feb-00
4954 ---------------------
4955
4956 The only change in this release is the fixing of some bugs in Makefile.in for
4957 the "install" target:
4958
4959 (1) It was failing to install pcreposix.h.
4960
4961 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4962
4963
4964 Version 3.0 01-Feb-00
4965 ---------------------
4966
4967 1. Add support for the /+ modifier to perltest (to output $` like it does in
4968 pcretest).
4969
4970 2. Add support for the /g modifier to perltest.
4971
4972 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4973 matches null strings.
4974
4975 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4976 pattern. Perl treats empty patterns specially - it reuses the most recent
4977 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4978 effect.
4979
4980 5. The POSIX interface was broken in that it was just handing over the POSIX
4981 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4982 required a bigger vector, with some working space on the end. This means that
4983 the POSIX wrapper now has to get and free some memory, and copy the results.
4984
4985 6. Added some simple autoconf support, placing the test data and the
4986 documentation in separate directories, re-organizing some of the
4987 information files, and making it build pcre-config (a GNU standard). Also added
4988 libtool support for building PCRE as a shared library, which is now the
4989 default.
4990
4991 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4992 09 are not valid octal constants. Single digits will be used for minor values
4993 less than 10.
4994
4995 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4996 existing programs that set these in the POSIX interface can use PCRE without
4997 modification.
4998
4999 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5000 return all that pcre_info() returns, plus additional data. The pcre_info()
5001 function is retained for compatibility, but is considered to be obsolete.
5002
5003 10. Added experimental recursion feature (?R) to handle one common case that
5004 Perl 5.6 will be able to do with (?p{...}).
5005
5006 11. Added support for POSIX character classes like [:alpha:], which Perl is
5007 adopting.
5008
5009
5010 Version 2.08 31-Aug-99
5011 ----------------------
5012
5013 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5014 trying to match at the startoffset position, but instead was moving forward to
5015 the next newline as if a previous match had failed.
5016
5017 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5018 and could get into a loop if a null string was matched other than at the start
5019 of the subject.
5020
5021 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5022 be distinguished at compile time, and for completeness also added PCRE_DATE.
5023
5024 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5025 in GnuWin32 environments.
5026
5027
5028 Version 2.07 29-Jul-99
5029 ----------------------
5030
5031 1. The documentation is now supplied in plain text form and HTML as well as in
5032 the form of man page sources.
5033
5034 2. C++ compilers don't like assigning (void *) values to other pointer types.
5035 In particular this affects malloc(). Although there is no problem in Standard
5036 C, I've put in casts to keep C++ compilers happy.
5037
5038 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5039 should be (const char *).
5040
5041 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5042 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5043 However, I haven't made this a standard facility. The documentation doesn't
5044 mention it, and the Makefile doesn't support it.
5045
5046 5. The Makefile now contains an "install" target, with editable destinations at
5047 the top of the file. The pcretest program is not installed.
5048
5049 6. pgrep -V now gives the PCRE version number and date.
5050
5051 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5052 causing the entire string to be ignored, instead of just the last character.
5053
5054 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5055 non-matching string, it can take a very, very long time, even for strings of
5056 quite modest length, because of the nested recursion. PCRE now does better in
5057 some of these cases. It does this by remembering the last required literal
5058 character in the pattern, and pre-searching the subject to ensure it is present
5059 before running the real match. In other words, it applies a heuristic to detect
5060 some types of certain failure quickly, and in the above example, if presented
5061 with a string that has no trailing " it gives "no match" very quickly.
5062
5063 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5064 other alternatives are tried instead.
5065
5066
5067 Version 2.06 09-Jun-99
5068 ----------------------
5069
5070 1. Change pcretest's output for amount of store used to show just the code
5071 space, because the remainder (the data block) varies in size between 32-bit and
5072 64-bit systems.
5073
5074 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5075 start matching at. This allows lookbehinds to work when searching for multiple
5076 occurrences in a string.
5077
5078 3. Added additional options to pcretest for testing multiple occurrences:
5079
5080 /+ outputs the rest of the string that follows a match
5081 /g loops for multiple occurrences, using the new startoffset argument
5082 /G loops for multiple occurrences by passing an incremented pointer
5083
5084 4. PCRE wasn't doing the "first character" optimization for patterns starting
5085 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5086 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5087 the letter 'x'. On long subject strings, this gives a significant speed-up.
5088
5089
5090 Version 2.05 21-Apr-99
5091 ----------------------
5092
5093 1. Changed the type of magic_number from int to long int so that it works
5094 properly on 16-bit systems.
5095
5096 2. Fixed a bug which caused patterns starting with .* not to work correctly
5097 when the subject string contained newline characters. PCRE was assuming
5098 anchoring for such patterns in all cases, which is not correct because .* will
5099 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5100 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5101 must be retried after every newline in the subject.
5102
5103
5104 Version 2.04 18-Feb-99
5105 ----------------------
5106
5107 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5108 computation of the store needed to hold the pattern was incorrect (too large).
5109 If such patterns were nested a few deep, this could multiply and become a real
5110 problem.
5111
5112 2. Added /M option to pcretest to show the memory requirement of a specific
5113 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5114
5115 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5116 compiled in such a way that the backtracking after subsequent failure was
5117 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5118 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5119
5120
5121 Version 2.03 02-Feb-99
5122 ----------------------
5123
5124 1. Fixed typo and small mistake in man page.
5125
5126 2. Added 4th condition (GPL supersedes if conflict) and created separate
5127 LICENCE file containing the conditions.
5128
5129 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5130 Perl, that is the internal \ allows the delimiter to be included in the
5131 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5132 the final delimiter, add \ to the end of the pattern (to test the error).
5133
5134 4. Added the convenience functions for extracting substrings after a successful
5135 match. Updated pcretest to make it able to test these functions.
5136
5137
5138 Version 2.02 14-Jan-99
5139 ----------------------
5140
5141 1. Initialized the working variables associated with each extraction so that
5142 their saving and restoring doesn't refer to uninitialized store.
5143
5144 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5145 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5146 fix the problem.
5147
5148 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5149 calls, and wasn't printing the correct value for compiling calls. Increased the
5150 default value of LOOPREPEAT, and the number of significant figures in the
5151 times.
5152
5153 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5154
5155 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5156 a building problem on Windows NT with a FAT file system.
5157
5158
5159 Version 2.01 21-Oct-98
5160 ----------------------
5161
5162 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5163 to character tables built by pcre_maketables() in the current locale. If NULL
5164 is passed, the default tables are used.
5165
5166
5167 Version 2.00 24-Sep-98
5168 ----------------------
5169
5170 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5171 it any more.
5172
5173 2. Allow quantification of (?>) groups, and make it work correctly.
5174
5175 3. The first character computation wasn't working for (?>) groups.
5176
5177 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5178 end of the subject) and add 5.005's \z, which really does match only at the
5179 very end of the subject.
5180
5181 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5182
5183 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5184 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5185 localized options. All options to pcre_study() were also removed.
5186
5187 7. Add other new features from 5.005:
5188
5189 $(?<= positive lookbehind
5190 $(?<! negative lookbehind
5191 (?imsx-imsx) added the unsetting capability
5192 such a setting is global if at outer level; local otherwise
5193 (?imsx-imsx:) non-capturing groups with option setting
5194 (?(cond)re|re) conditional pattern matching
5195
5196 A backreference to itself in a repeated group matches the previous
5197 captured string.
5198
5199 8. General tidying up of studying (both automatic and via "study")
5200 consequential on the addition of new assertions.
5201
5202 9. As in 5.005, unlimited repeated groups that could match an empty substring
5203 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5204 runtime if any iteration does actually match an empty substring.
5205
5206 10. Include the RunTest script in the distribution.
5207
5208 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5209 discrepancies, some of which were old and were also with respect to 5.004. They
5210 have now been fixed.
5211
5212
5213 Version 1.09 28-Apr-98
5214 ----------------------
5215
5216 1. A negated single character class followed by a quantifier with a minimum
5217 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5218 program crashes, or just wrong answers. This did not apply to negated classes
5219 containing more than one character, or to minima other than one.
5220
5221
5222 Version 1.08 27-Mar-98
5223 ----------------------
5224
5225 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5226
5227 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5228 latter must appear before anything that relies on it in the pattern.
5229
5230
5231 Version 1.07 16-Feb-98
5232 ----------------------
5233
5234 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5235 repeat of a potentially empty string).
5236
5237
5238 Version 1.06 23-Jan-98
5239 ----------------------
5240
5241 1. Added Markus Oberhumer's little patches for C++.
5242
5243 2. Literal strings longer than 255 characters were broken.
5244
5245
5246 Version 1.05 23-Dec-97
5247 ----------------------
5248
5249 1. Negated character classes containing more than one character were failing if
5250 PCRE_CASELESS was set at run time.
5251
5252
5253 Version 1.04 19-Dec-97
5254 ----------------------
5255
5256 1. Corrected the man page, where some "const" qualifiers had been omitted.
5257
5258 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5259 input syntax.
5260
5261 3. Fixed memory leak which occurred when a regex with back references was
5262 matched with an offsets vector that wasn't big enough. The temporary memory
5263 that is used in this case wasn't being freed if the match failed.
5264
5265 4. Tidied pcretest to ensure it frees memory that it gets.
5266
5267 5. Temporary memory was being obtained in the case where the passed offsets
5268 vector was exactly big enough.
5269
5270 6. Corrected definition of offsetof() from change 5 below.
5271
5272 7. I had screwed up change 6 below and broken the rules for the use of
5273 setjmp(). Now fixed.
5274
5275
5276 Version 1.03 18-Dec-97
5277 ----------------------
5278
5279 1. A erroneous regex with a missing opening parenthesis was correctly
5280 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5281 on some systems.
5282
5283 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5284 it was reported that one broken compiler failed on the former because "code" is
5285 also an independent variable.
5286
5287 3. The erroneous regex a[]b caused an array overrun reference.
5288
5289 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5290 fail on data ending with that character. (It was going on too far, and checking
5291 the next character, typically a binary zero.) This was specific to the
5292 optimized code for single-character negative classes.
5293
5294 5. Added a contributed patch from the TIN world which does the following:
5295
5296 + Add an undef for memmove, in case the the system defines a macro for it.
5297
5298 + Add a definition of offsetof(), in case there isn't one. (I don't know
5299 the reason behind this - offsetof() is part of the ANSI standard - but
5300 it does no harm).
5301
5302 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5303 most of the places where whitespace preceded '#'. I have given up and
5304 allowed the remaining 2 cases to be at the margin.
5305
5306 + Rename some variables in pcre to eliminate shadowing. This seems very
5307 pedantic, but does no harm, of course.
5308
5309 6. Moved the call to setjmp() into its own function, to get rid of warnings
5310 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5311
5312 7. Constructs such as \d{8,} were compiling into the equivalent of
5313 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5314 outcome, but in this particular case used more store than had been allocated,
5315 which caused the bug to be discovered because it threw up an internal error.
5316
5317 8. The debugging code in both pcre and pcretest for outputting the compiled
5318 form of a regex was going wrong in the case of back references followed by
5319 curly-bracketed repeats.
5320
5321
5322 Version 1.02 12-Dec-97
5323 ----------------------
5324
5325 1. Typos in pcre.3 and comments in the source fixed.
5326
5327 2. Applied a contributed patch to get rid of places where it used to remove
5328 'const' from variables, and fixed some signed/unsigned and uninitialized
5329 variable warnings.
5330
5331 3. Added the "runtest" target to Makefile.
5332
5333 4. Set default compiler flag to -O2 rather than just -O.
5334
5335
5336 Version 1.01 19-Nov-97
5337 ----------------------
5338
5339 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5340 like /([ab]*)*/, that is, for classes with more than one character in them.
5341
5342 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5343 as /((?>a*))*/ (a PCRE_EXTRA facility).
5344
5345
5346 Version 1.00 18-Nov-97
5347 ----------------------
5348
5349 1. Added compile-time macros to support systems such as SunOS4 which don't have
5350 memmove() or strerror() but have other things that can be used instead.
5351
5352 2. Arranged that "make clean" removes the executables.
5353
5354
5355 Version 0.99 27-Oct-97
5356 ----------------------
5357
5358 1. Fixed bug in code for optimizing classes with only one character. It was
5359 initializing a 32-byte map regardless, which could cause it to run off the end
5360 of the memory it had got.
5361
5362 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5363
5364
5365 Version 0.98 22-Oct-97
5366 ----------------------
5367
5368 1. Fixed bug in code for handling temporary memory usage when there are more
5369 back references than supplied space in the ovector. This could cause segfaults.
5370
5371
5372 Version 0.97 21-Oct-97
5373 ----------------------
5374
5375 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5376
5377 2. Optimized negated single characters not to use a bit map.
5378
5379 3. Brought error texts together as macro definitions; clarified some of them;
5380 fixed one that was wrong - it said "range out of order" when it meant "invalid
5381 escape sequence".
5382
5383 4. Changed some char * arguments to const char *.
5384
5385 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5386
5387 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5388 pcretest.
5389
5390
5391 Version 0.96 16-Oct-97
5392 ----------------------
5393
5394 1. Added a simple "pgrep" utility to the distribution.
5395
5396 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5397 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5398 where "ddd" means "one or more decimal digits".
5399
5400 3. Fixed serious bug. If a pattern had a back reference, but the call to
5401 pcre_exec() didn't supply a large enough ovector to record the related
5402 identifying subpattern, the match always failed. PCRE now remembers the number
5403 of the largest back reference, and gets some temporary memory in which to save
5404 the offsets during matching if necessary, in order to ensure that
5405 backreferences always work.
5406
5407 4. Increased the compatibility with Perl in a number of ways:
5408
5409 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5410 to request this handling. The option can be set at compile or exec time.
5411
5412 (b) $ matches before a terminating newline by default; an option
5413 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5414 mode). The option can be set at compile or exec time.
5415
5416 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5417 the same as Perl's. If the decimal number it represents is less than 10
5418 or there aren't that many previous left capturing parentheses, an octal
5419 escape is read. Inside a character class, it's always an octal escape,
5420 even if it is a single digit.
5421
5422 (d) An escaped but undefined alphabetic character is taken as a literal,
5423 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5424 escapes.
5425
5426 (e) {0} is now permitted. (The previous item is removed from the compiled
5427 pattern).
5428
5429 5. Changed all the names of code files so that the basic parts are no longer
5430 than 10 characters, and abolished the teeny "globals.c" file.
5431
5432 6. Changed the handling of character classes; they are now done with a 32-byte
5433 bit map always.
5434
5435 7. Added the -d and /D options to pcretest to make it possible to look at the
5436 internals of compilation without having to recompile pcre.
5437
5438
5439 Version 0.95 23-Sep-97
5440 ----------------------
5441
5442 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5443 \x20 at the start of a run of normal characters. These were being treated as
5444 real characters, instead of the source characters being re-checked.
5445
5446
5447 Version 0.94 18-Sep-97
5448 ----------------------
5449
5450 1. The functions are now thread-safe, with the caveat that the global variables
5451 containing pointers to malloc() and free() or alternative functions are the
5452 same for all threads.
5453
5454 2. Get pcre_study() to generate a bitmap of initial characters for non-
5455 anchored patterns when this is possible, and use it if passed to pcre_exec().
5456
5457
5458 Version 0.93 15-Sep-97
5459 ----------------------
5460
5461 1. /(b)|(:+)/ was computing an incorrect first character.
5462
5463 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5464 but not actually doing anything yet.
5465
5466 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5467 as Perl does (e.g. [-az] or [az-]).
5468
5469 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5470 all possible positions.
5471
5472 5. Split up into different modules to avoid including unneeded functions in a
5473 compiled binary. However, compile and exec are still in one module. The "study"
5474 function is split off.
5475
5476 6. The character tables are now in a separate module whose source is generated
5477 by an auxiliary program - but can then be edited by hand if required. There are
5478 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5479 toupper() in the code.
5480
5481 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5482 make them global. Abolish the function for setting them, as the caller can now
5483 set them directly.
5484
5485
5486 Version 0.92 11-Sep-97
5487 ----------------------
5488
5489 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5490 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5491
5492 2. Caseless matching was not working in character classes if the characters in
5493 the pattern were in upper case.
5494
5495 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5496
5497 4. Make PCRE_ANCHORED public and accept as a compile option.
5498
5499 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5500 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5501 pass them.
5502
5503 6. Give an error if bad option bits passed at compile or run time.
5504
5505 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5506 pcretest to cause it to pass that flag.
5507
5508 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5509 options, and the first character, if set.
5510
5511 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5512
5513
5514 Version 0.91 10-Sep-97
5515 ----------------------
5516
5517 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5518 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5519
5520 2. PCRE was looping on encountering an indefinitely repeated back reference to
5521 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5522 Perl does - treats the match as successful.
5523
5524 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5