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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1480 - (show annotations)
Tue May 27 18:02:51 2014 UTC (5 years, 4 months ago) by ph10
File size: 266052 byte(s)
Error occurred while calculating annotation data.
Fix auto-possessification bug.
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
48 Version 8.35 04-April-2014
49 --------------------------
50
51 1. A new flag is set, when property checks are present in an XCLASS.
52 When this flag is not set, PCRE can perform certain optimizations
53 such as studying these XCLASS-es.
54
55 2. The auto-possessification of character sets were improved: a normal
56 and an extended character set can be compared now. Furthermore
57 the JIT compiler optimizes more character set checks.
58
59 3. Got rid of some compiler warnings for potentially uninitialized variables
60 that show up only when compiled with -O2.
61
62 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
63 of a match later then the end of the match. The pcretest program was not
64 handling the case sensibly - it was outputting from the start to the next
65 binary zero. It now reports this situation in a message, and outputs the
66 text from the end to the start.
67
68 5. Fast forward search is improved in JIT. Instead of the first three
69 characters, any three characters with fixed position can be searched.
70 Search order: first, last, middle.
71
72 6. Improve character range checks in JIT. Characters are read by an inprecise
73 function now, which returns with an unknown value if the character code is
74 above a certain treshold (e.g: 256). The only limitation is that the value
75 must be bigger than the treshold as well. This function is useful, when
76 the characters above the treshold are handled in the same way.
77
78 7. The macros whose names start with RAWUCHAR are placeholders for a future
79 mode in which only the bottom 21 bits of 32-bit data items are used. To
80 make this more memorable for those maintaining the code, the names have
81 been changed to start with UCHAR21, and an extensive comment has been added
82 to their definition.
83
84 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
85 to the export list in Makefile.am (they were accidentally omitted from the
86 8.34 tarball).
87
88 9. The informational output from pcretest used the phrase "starting byte set"
89 which is inappropriate for the 16-bit and 32-bit libraries. As the output
90 for "first char" and "need char" really means "non-UTF-char", I've changed
91 "byte" to "char", and slightly reworded the output. The documentation about
92 these values has also been (I hope) clarified.
93
94 10. Another JIT related optimization: use table jumps for selecting the correct
95 backtracking path, when more than four alternatives are present inside a
96 bracket.
97
98 11. Empty match is not possible, when the minimum length is greater than zero,
99 and there is no \K in the pattern. JIT should avoid empty match checks in
100 such cases.
101
102 12. In a caseless character class with UCP support, when a character with more
103 than one alternative case was not the first character of a range, not all
104 the alternative cases were added to the class. For example, s and \x{17f}
105 are both alternative cases for S: the class [RST] was handled correctly,
106 but [R-T] was not.
107
108 13. The configure.ac file always checked for pthread support when JIT was
109 enabled. This is not used in Windows, so I have put this test inside a
110 check for the presence of windows.h (which was already tested for).
111
112 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
113 The algorithm provides a way to skip certain starting offsets, and usually
114 faster than linear prefix searches.
115
116 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
117 as for 'fr_FR' and 'french'. For some reason, however, it then used the
118 Windows-specific input and output files, which have 'french' screwed in.
119 So this could never have worked. One of the problems with locales is that
120 they aren't always the same. I have now updated RunTest so that it checks
121 the output of the locale test (test 3) against three different output
122 files, and it allows the test to pass if any one of them matches. With luck
123 this should make the test pass on some versions of Solaris where it was
124 failing. Because of the uncertainty, the script did not used to stop if
125 test 3 failed; it now does. If further versions of a French locale ever
126 come to light, they can now easily be added.
127
128 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
129 there was a message during ./configure, but it did not stop. This now
130 provokes an error. The invalid example in README has been corrected.
131 If a value less than the minimum is given, the minimum value has always
132 been used, but now a warning is given.
133
134 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
135 was a bug in the test system, which is now fixed. Also, the list of various
136 configurations that are tested for each release did not have one with both
137 16/32 bits and --enable-bar-anycrlf. It now does.
138
139 18. pcretest was missing "-C bsr" for displaying the \R default setting.
140
141 19. Little endian PowerPC systems are supported now by the JIT compiler.
142
143 20. The fast forward newline mechanism could enter to an infinite loop on
144 certain invalid UTF-8 input. Although we don't support these cases
145 this issue can be fixed by a performance optimization.
146
147 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
148 not take account if existing stack usage. There is now a new global
149 variable called pcre_stack_guard that can be set to point to an external
150 function to check stack availability. It is called at the start of
151 processing every parenthesized group.
152
153 22. A typo in the code meant that in ungreedy mode the max/min qualifier
154 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
155 not match "ab".
156
157 23. When UTF was disabled, the JIT program reported some incorrect compile
158 errors. These messages are silenced now.
159
160 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
161 compiler.
162
163 25. Change all the temporary files used in RunGrepTest to be different to those
164 used by RunTest so that the tests can be run simultaneously, for example by
165 "make -j check".
166
167
168 Version 8.34 15-December-2013
169 -----------------------------
170
171 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
172 executable memory. Patch inspired by Carsten Klein.
173
174 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
175 this macro is never tested and has no effect, because the work to support
176 coverage involves only compiling and linking options and special targets in
177 the Makefile. The comment in config.h implied that defining the macro would
178 enable coverage support, which is totally false. There was also support for
179 setting this macro in the CMake files (my fault, I just copied it from
180 configure). SUPPORT_GCOV has now been removed.
181
182 3. Make a small performance improvement in strlen16() and strlen32() in
183 pcretest.
184
185 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
186 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
187 these cases). There was also one in pcretest.c.
188
189 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
190
191 6. In UTF mode, the code for checking whether a group could match an empty
192 string (which is used for indefinitely repeated groups to allow for
193 breaking an infinite loop) was broken when the group contained a repeated
194 negated single-character class with a character that occupied more than one
195 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
196 UTF-8 mode). The effect was undefined: the group might or might not be
197 deemed as matching an empty string, or the program might have crashed.
198
199 7. The code for checking whether a group could match an empty string was not
200 recognizing that \h, \H, \v, \V, and \R must match a character.
201
202 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
203 an empty string. If it can, pcretest shows this in its information output.
204
205 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
206 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
207 matched by pcre_exec() without using JIT:
208
209 (a) If the rest of the pattern did not match after a maximal run of
210 grapheme clusters, the code for backing up to try with fewer of them
211 did not always back up over a full grapheme when characters that do not
212 have the modifier quality were involved, e.g. Hangul syllables.
213
214 (b) If the match point in a subject started with modifier character, and
215 there was no match, the code could incorrectly back up beyond the match
216 point, and potentially beyond the first character in the subject,
217 leading to a segfault or an incorrect match result.
218
219 10. A conditional group with an assertion condition could lead to PCRE
220 recording an incorrect first data item for a match if no other first data
221 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
222 first data item, and therefore matched "ca" after "c" instead of at the
223 start.
224
225 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
226 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
227
228 12. The source of pcregrep now includes z/OS-specific code so that it can be
229 compiled for z/OS as part of the special z/OS distribution.
230
231 13. Added the -T and -TM options to pcretest.
232
233 14. The code in pcre_compile.c for creating the table of named capturing groups
234 has been refactored. Instead of creating the table dynamically during the
235 actual compiling pass, the information is remembered during the pre-compile
236 pass (on the stack unless there are more than 20 named groups, in which
237 case malloc() is used) and the whole table is created before the actual
238 compile happens. This has simplified the code (it is now nearly 150 lines
239 shorter) and prepared the way for better handling of references to groups
240 with duplicate names.
241
242 15. A back reference to a named subpattern when there is more than one of the
243 same name now checks them in the order in which they appear in the pattern.
244 The first one that is set is used for the reference. Previously only the
245 first one was inspected. This change makes PCRE more compatible with Perl.
246
247 16. Unicode character properties were updated from Unicode 6.3.0.
248
249 17. The compile-time code for auto-possessification has been refactored, based
250 on a patch by Zoltan Herczeg. It now happens after instead of during
251 compilation. The code is cleaner, and more cases are handled. The option
252 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
253 options in pcretest are provided to set it. It can also be set by
254 (*NO_AUTO_POSSESS) at the start of a pattern.
255
256 18. The character VT has been added to the default ("C" locale) set of
257 characters that match \s and are generally treated as white space,
258 following this same change in Perl 5.18. There is now no difference between
259 "Perl space" and "POSIX space". Whether VT is treated as white space in
260 other locales depends on the locale.
261
262 19. The code for checking named groups as conditions, either for being set or
263 for being recursed, has been refactored (this is related to 14 and 15
264 above). Processing unduplicated named groups should now be as fast at
265 numerical groups, and processing duplicated groups should be faster than
266 before.
267
268 20. Two patches to the CMake build system, by Alexander Barkov:
269
270 (1) Replace the "source" command by "." in CMakeLists.txt because
271 "source" is a bash-ism.
272
273 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
274 without these the CMake build does not work on Solaris.
275
276 21. Perl has changed its handling of \8 and \9. If there is no previously
277 encountered capturing group of those numbers, they are treated as the
278 literal characters 8 and 9 instead of a binary zero followed by the
279 literals. PCRE now does the same.
280
281 22. Following Perl, added \o{} to specify codepoints in octal, making it
282 possible to specify values greater than 0777 and also making them
283 unambiguous.
284
285 23. Perl now gives an error for missing closing braces after \x{... instead of
286 treating the string as literal. PCRE now does the same.
287
288 24. RunTest used to grumble if an inappropriate test was selected explicitly,
289 but just skip it when running all tests. This make it awkward to run ranges
290 of tests when one of them was inappropriate. Now it just skips any
291 inappropriate tests, as it always did when running all tests.
292
293 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
294 character types such as \d or \w, too many callouts were inserted, and the
295 data that they returned was rubbish.
296
297 26. In UCP mode, \s was not matching two of the characters that Perl matches,
298 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
299 were matched by \h. The code has now been refactored so that the lists of
300 the horizontal and vertical whitespace characters used for \h and \v (which
301 are defined only in one place) are now also used for \s.
302
303 27. Add JIT support for the 64 bit TileGX architecture.
304 Patch by Jiong Wang (Tilera Corporation).
305
306 28. Possessive quantifiers for classes (both explicit and automatically
307 generated) now use special opcodes instead of wrapping in ONCE brackets.
308
309 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
310 (because it's meaningless), this was not happening when PCRE_CASELESS was
311 set. Not wrong, but inefficient.
312
313 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
314 properties for \w, \d, etc) is present in a test regex. Otherwise if the
315 test contains no characters greater than 255, Perl doesn't realise it
316 should be using Unicode semantics.
317
318 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
319 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
320 does in Unicode mode.
321
322 32. Added the "forbid" facility to pcretest so that putting tests into the
323 wrong test files can sometimes be quickly detected.
324
325 33. There is now a limit (default 250) on the depth of nesting of parentheses.
326 This limit is imposed to control the amount of system stack used at compile
327 time. It can be changed at build time by --with-parens-nest-limit=xxx or
328 the equivalent in CMake.
329
330 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
331 errors. Perl warns for these when in warning mode, but PCRE has no facility
332 for giving warnings.
333
334 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
335 However, this was not working for (?!) because it is optimized to (*FAIL),
336 for which PCRE does not allow quantifiers. The optimization is now disabled
337 when a quantifier follows (?!). I can't see any use for this, but it makes
338 things uniform.
339
340 36. Perl no longer allows group names to start with digits, so I have made this
341 change also in PCRE. It simplifies the code a bit.
342
343 37. In extended mode, Perl ignores spaces before a + that indicates a
344 possessive quantifier. PCRE allowed a space before the quantifier, but not
345 before the possessive +. It now does.
346
347 38. The use of \K (reset reported match start) within a repeated possessive
348 group such as (a\Kb)*+ was not working.
349
350 40. Document that the same character tables must be used at compile time and
351 run time, and that the facility to pass tables to pcre_exec() and
352 pcre_dfa_exec() is for use only with saved/restored patterns.
353
354 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
355 features for Builds with MSVC:
356
357 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
358 builds). The .rc files can be used to set FileDescription and many other
359 attributes.
360
361 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
362 This allows higher-level build scripts which want .pdb files to avoid
363 hard-coding the exact files needed."
364
365 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
366 mean "start of word" and "end of word", respectively, as a transition aid.
367
368 43. A minimizing repeat of a class containing codepoints greater than 255 in
369 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
370 compiled to use the heap for recursion.
371
372 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
373 is configured.
374
375
376 Version 8.33 28-May-2013
377 ------------------------
378
379 1. Added 'U' to some constants that are compared to unsigned integers, to
380 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
381 variables that are added to signed variables, to ensure the result is
382 signed and can be negated.
383
384 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
385 CMake config files.
386
387 3. Revise the creation of config.h.generic so that all boolean macros are
388 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
389 overriding via -D on the command line possible.
390
391 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
392 to unsigned int is reported to make a quite noticeable speed difference in
393 a specific Windows environment. Testing on Linux did also appear to show
394 some benefit (and it is clearly not harmful). Also fixed the definition of
395 Xop which should be unsigned.
396
397 5. Related to (4), changing the definition of the intermediate variable cc
398 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
399 improvements.
400
401 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
402 unnecessary spaces.
403
404 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
405 and later.
406
407 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
408
409 9. Optimizing fast_forward_start_bits in JIT.
410
411 10. Adding support for callouts in JIT, and fixing some issues revealed
412 during this work. Namely:
413
414 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
415
416 (b) Minimum length was not checked before the matching is started.
417
418 11. The value of capture_last that is passed to callouts was incorrect in some
419 cases when there was a capture on one path that was subsequently abandoned
420 after a backtrack. Also, the capture_last value is now reset after a
421 recursion, since all captures are also reset in this case.
422
423 12. The interpreter no longer returns the "too many substrings" error in the
424 case when an overflowing capture is in a branch that is subsequently
425 abandoned after a backtrack.
426
427 13. In the pathological case when an offset vector of size 2 is used, pcretest
428 now prints out the matched string after a yield of 0 or 1.
429
430 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
431 Only supported by the JIT compiler at the moment.
432
433 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
434
435 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
436 offset of the starting point of the matching process, provided the offsets
437 vector is large enough.
438
439 17. The \A escape now records a lookbehind value of 1, though its execution
440 does not actually inspect the previous character. This is to ensure that,
441 in partial multi-segment matching, at least one character from the old
442 segment is retained when a new segment is processed. Otherwise, if there
443 are no lookbehinds in the pattern, \A might match incorrectly at the start
444 of a new segment.
445
446 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
447
448 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
449 gives some modest performance improvement in 8-bit mode.
450
451 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
452 be expressed in certain programming languages using Universal Character
453 Names.
454
455 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
456 which points out that "non characters" are not "characters that may not
457 appear in Unicode strings" but rather "characters that are reserved for
458 internal use and have only local meaning".
459
460 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
461 there was a conditional group that depended on an assertion, if the
462 assertion was false, the callout that immediately followed the alternation
463 in the condition was skipped when pcre_exec() was used for matching.
464
465 23. Allow an explicit callout to be inserted before an assertion that is the
466 condition for a conditional group, for compatibility with automatic
467 callouts, which always insert a callout at this point.
468
469 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
470 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
471
472 25. (*PRUNE) is now supported by the JIT compiler.
473
474 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
475
476 27. Fix the case where there are two or more SKIPs with arguments that may be
477 ignored.
478
479 28. (*SKIP) is now supported by the JIT compiler.
480
481 29. (*THEN) is now supported by the JIT compiler.
482
483 30. Update RunTest with additional test selector options.
484
485 31. The way PCRE handles backtracking verbs has been changed in two ways.
486
487 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
488 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
489 backtracking. In some cases this makes it more Perl-compatible, but Perl's
490 rather obscure rules do not always do the same thing.
491
492 (2) Previously, backtracking verbs were confined within assertions. This is
493 no longer the case for positive assertions, except for (*ACCEPT). Again,
494 this sometimes improves Perl compatibility, and sometimes does not.
495
496 32. A number of tests that were in test 2 because Perl did things differently
497 have been moved to test 1, because either Perl or PCRE has changed, and
498 these tests are now compatible.
499
500 32. Backtracking control verbs are now handled in the same way in JIT and
501 interpreter.
502
503 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
504 contained a forward subroutine reference caused a compile error.
505
506 34. Auto-detect and optimize limited repetitions in JIT.
507
508 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
509 blocking (*UTF) etc.
510
511 36. In the interpreter, maximizing pattern repetitions for characters and
512 character types now use tail recursion, which reduces stack usage.
513
514 37. The value of the max lookbehind was not correctly preserved if a compiled
515 and saved regex was reloaded on a host of different endianness.
516
517 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
518 of the compiled pattern block, expand the flags field from 16 to 32 bits
519 because it was almost full.
520
521 39. Try madvise first before posix_madvise.
522
523 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
524 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
525 original change locked out all matches of empty strings. This has been
526 changed so that one match of an empty string per line is recognized.
527 Subsequent searches on the same line (for colouring or for --only-matching,
528 for example) do not recognize empty strings.
529
530 41. Applied a user patch to fix a number of spelling mistakes in comments.
531
532 42. Data lines longer than 65536 caused pcretest to crash.
533
534 43. Clarified the data type for length and startoffset arguments for pcre_exec
535 and pcre_dfa_exec in the function-specific man pages, where they were
536 explicitly stated to be in bytes, never having been updated. I also added
537 some clarification to the pcreapi man page.
538
539 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
540 a segmentation fault.
541
542
543 Version 8.32 30-November-2012
544 -----------------------------
545
546 1. Improved JIT compiler optimizations for first character search and single
547 character iterators.
548
549 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
550 Patch by Daniel Richard G.
551
552 3. Single character iterator optimizations in the JIT compiler.
553
554 4. Improved JIT compiler optimizations for character ranges.
555
556 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
557 Reported by Giuseppe D'Angelo.
558
559 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
560 of a line, was being set incorrectly in cases where .* appeared inside
561 atomic brackets at the start of a pattern, or where there was a subsequent
562 *PRUNE or *SKIP.
563
564 7. Improved instruction cache flush for POWER/PowerPC.
565 Patch by Daniel Richard G.
566
567 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
568 grep:
569
570 (a) There is now no limit to the number of patterns to be matched.
571
572 (b) An error is given if a pattern is too long.
573
574 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
575 are now supported.
576
577 (d) --exclude-from and --include-from (multiple use) have been added.
578
579 (e) Exclusions and inclusions now apply to all files and directories, not
580 just to those obtained from scanning a directory recursively.
581
582 (f) Multiple uses of -f and --file-list are now supported.
583
584 (g) In a Windows environment, the default for -d has been changed from
585 "read" (the GNU grep default) to "skip", because otherwise the presence
586 of a directory in the file list provokes an error.
587
588 (h) The documentation has been revised and clarified in places.
589
590 9. Improve the matching speed of capturing brackets.
591
592 10. Changed the meaning of \X so that it now matches a Unicode extended
593 grapheme cluster.
594
595 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
596 out POSIX threads when JIT support is configured.
597
598 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
599
600 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
601 structure could go wrong in environments where size_t is not the same size
602 as int.
603
604 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
605 set.
606
607 15. The EBCDIC support had decayed; later updates to the code had included
608 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
609 general tidy up of EBCDIC-related issues, and the documentation was also
610 not quite right. There is now a test that can be run on ASCII systems to
611 check some of the EBCDIC-related things (but is it not a full test).
612
613 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
614 in a small tidy to the code.
615
616 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
617
618 18. If the --only-matching (-o) option in pcregrep is specified multiple
619 times, each one causes appropriate output. For example, -o1 -o2 outputs the
620 substrings matched by the 1st and 2nd capturing parentheses. A separating
621 string can be specified by --om-separator (default empty).
622
623 19. Improving the first n character searches.
624
625 20. Turn case lists for horizontal and vertical white space into macros so that
626 they are defined only once.
627
628 21. This set of changes together give more compatible Unicode case-folding
629 behaviour for characters that have more than one other case when UCP
630 support is available.
631
632 (a) The Unicode property table now has offsets into a new table of sets of
633 three or more characters that are case-equivalent. The MultiStage2.py
634 script that generates these tables (the pcre_ucd.c file) now scans
635 CaseFolding.txt instead of UnicodeData.txt for character case
636 information.
637
638 (b) The code for adding characters or ranges of characters to a character
639 class has been abstracted into a generalized function that also handles
640 case-independence. In UTF-mode with UCP support, this uses the new data
641 to handle characters with more than one other case.
642
643 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
644 whose other case is greater than 256 are now correctly matched
645 caselessly. Previously, the high codepoint matched the low one, but not
646 vice versa.
647
648 (d) The processing of \h, \H, \v, and \ in character classes now makes use
649 of the new class addition function, using character lists defined as
650 macros alongside the case definitions of 20 above.
651
652 (e) Caseless back references now work with characters that have more than
653 one other case.
654
655 (f) General caseless matching of characters with more than one other case
656 is supported.
657
658 22. Unicode character properties were updated from Unicode 6.2.0
659
660 23. Improved CMake support under Windows. Patch by Daniel Richard G.
661
662 24. Add support for 32-bit character strings, and UTF-32
663
664 25. Major JIT compiler update (code refactoring and bugfixing).
665 Experimental Sparc 32 support is added.
666
667 26. Applied a modified version of Daniel Richard G's patch to create
668 pcre.h.generic and config.h.generic by "make" instead of in the
669 PrepareRelease script.
670
671 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
672 pcre_compile.c when checking for a zero character.
673
674 28. Introducing a native interface for JIT. Through this interface, the compiled
675 machine code can be directly executed. The purpose of this interface is to
676 provide fast pattern matching, so several sanity checks are not performed.
677 However, feature tests are still performed. The new interface provides
678 1.4x speedup compared to the old one.
679
680 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
681 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
682 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
683
684 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
685 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
686 the "old" RFC 2279). Instead, it ended up passing a negative length to
687 pcre_exec().
688
689 31. Add support for GCC's visibility feature to hide internal functions.
690
691 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
692 "unknown -C option" after outputting 0 or 1.
693
694 33. There is now support for generating a code coverage report for the test
695 suite in environments where gcc is the compiler and lcov is installed. This
696 is mainly for the benefit of the developers.
697
698 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
699 unaddressable using valgrind annotations, allowing valgrind to detect
700 invalid memory accesses. This is mainly for the benefit of the developers.
701
702 25. (*UTF) can now be used to start a pattern in any of the three libraries.
703
704 26. Give configure error if --enable-cpp but no C++ compiler found.
705
706
707 Version 8.31 06-July-2012
708 -------------------------
709
710 1. Fixing a wrong JIT test case and some compiler warnings.
711
712 2. Removed a bashism from the RunTest script.
713
714 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
715 to unsigned type, result still unsigned" that was given by an MS compiler
716 on encountering the code "-sizeof(xxx)".
717
718 4. Partial matching support is added to the JIT compiler.
719
720 5. Fixed several bugs concerned with partial matching of items that consist
721 of more than one character:
722
723 (a) /^(..)\1/ did not partially match "aba" because checking references was
724 done on an "all or nothing" basis. This also applied to repeated
725 references.
726
727 (b) \R did not give a hard partial match if \r was found at the end of the
728 subject.
729
730 (c) \X did not give a hard partial match after matching one or more
731 characters at the end of the subject.
732
733 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
734 a partial match for the string "\r".
735
736 (e) When newline was set to CRLF, the metacharacter "." did not recognize
737 a partial match for a CR character at the end of the subject string.
738
739 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
740 running pcretest, the text "(JIT)" added to the output whenever JIT is
741 actually used to run the match.
742
743 7. Individual JIT compile options can be set in pcretest by following -s+[+]
744 or /S+[+] with a digit between 1 and 7.
745
746 8. OP_NOT now supports any UTF character not just single-byte ones.
747
748 9. (*MARK) control verb is now supported by the JIT compiler.
749
750 10. The command "./RunTest list" lists the available tests without actually
751 running any of them. (Because I keep forgetting what they all are.)
752
753 11. Add PCRE_INFO_MAXLOOKBEHIND.
754
755 12. Applied a (slightly modified) user-supplied patch that improves performance
756 when the heap is used for recursion (compiled with --disable-stack-for-
757 recursion). Instead of malloc and free for each heap frame each time a
758 logical recursion happens, frames are retained on a chain and re-used where
759 possible. This sometimes gives as much as 30% improvement.
760
761 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
762 call.
763
764 14. As documented, (*COMMIT) is now confined to within a positive assertion.
765
766 15. It is now possible to link pcretest with libedit as an alternative to
767 libreadline.
768
769 16. (*COMMIT) control verb is now supported by the JIT compiler.
770
771 17. The Unicode data tables have been updated to Unicode 6.1.0.
772
773 18. Added --file-list option to pcregrep.
774
775 19. Added binary file support to pcregrep, including the -a, --binary-files,
776 -I, and --text options.
777
778 20. The madvise function is renamed for posix_madvise for QNX compatibility
779 reasons. Fixed by Giuseppe D'Angelo.
780
781 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
782
783 22. Changed the option for creating symbolic links for 16-bit man pages from
784 -s to -sf so that re-installing does not cause issues.
785
786 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
787
788 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
789 matches in certain environments (the workspace was not being correctly
790 retained). Also added to pcre_dfa_exec() a simple plausibility check on
791 some of the workspace data at the beginning of a restart.
792
793 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
794 was not doing so when it should - probably a typo introduced by SVN 528
795 (change 8.10/14).
796
797 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
798 \w+ when the character tables indicated that \x{c4} was a word character.
799 There were several related cases, all because the tests for doing a table
800 lookup were testing for characters less than 127 instead of 255.
801
802 27. If a pattern contains capturing parentheses that are not used in a match,
803 their slots in the ovector are set to -1. For those that are higher than
804 any matched groups, this happens at the end of processing. In the case when
805 there were back references that the ovector was too small to contain
806 (causing temporary malloc'd memory to be used during matching), and the
807 highest capturing number was not used, memory off the end of the ovector
808 was incorrectly being set to -1. (It was using the size of the temporary
809 memory instead of the true size.)
810
811 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
812 ovector size, it uses memory at the end of the block that it has got.
813
814 29. Check for an overlong MARK name and give an error at compile time. The
815 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
816
817 30. JIT compiler update.
818
819 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
820 Rill for the patch.
821
822 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
823
824 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
825
826 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
827 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
828 was enabled.
829
830 35. Improve JIT code generation for greedy plus quantifier.
831
832 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
833 1 to "aa" instead of to an empty string. The bug affected repeated groups
834 that could potentially match an empty string.
835
836 37. Optimizing single character iterators in JIT.
837
838 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
839 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
840 codepoints that are too big for the mode are faulted, and in a UTF mode,
841 disallowed codepoints are also faulted.
842
843 39. If PCRE was compiled with UTF support, in three places in the DFA
844 matcher there was code that should only have been obeyed in UTF mode, but
845 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
846 processing when bytes with values greater than 127 were present. In 16-bit
847 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
848 both cases the values are those that cannot be the first data item in a UTF
849 character. The three items that might have provoked this were recursions,
850 possessively repeated groups, and atomic groups.
851
852 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
853 and pcregrep, because some OS require shared objects to be explicitly
854 passed to ld, causing the link step to fail if they are not.
855
856 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
857 mode, patterns that started with \h* or \R* might be incorrectly matched.
858
859
860 Version 8.30 04-February-2012
861 -----------------------------
862
863 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
864 name is defined in ctype.h.
865
866 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
867 only in quite long subpatterns.
868
869 3. Removed the function pcre_info(), which has been obsolete and deprecated
870 since it was replaced by pcre_fullinfo() in February 2000.
871
872 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
873 match a (*MARK), and the match failed at the start of the subject, a
874 reference to memory before the start of the subject could occur. This bug
875 was introduced by fix 17 of release 8.21.
876
877 5. A reference to an unset group with zero minimum repetition was giving
878 totally wrong answers (in non-JavaScript-compatibility mode). For example,
879 /(another)?(\1?)test/ matched against "hello world test". This bug was
880 introduced in release 8.13.
881
882 6. Add support for 16-bit character strings (a large amount of work involving
883 many changes and refactorings).
884
885 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
886 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
887 from a file.
888
889 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
890 rounding is not applied in this particular case).
891
892 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
893 if they appear, or are escaped, in patterns.
894
895 10. Get rid of a number of -Wunused-but-set-variable warnings.
896
897 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
898 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
899 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
900 also returns the mark "x". This bug applied to capturing parentheses,
901 non-capturing parentheses, and atomic parentheses. It also applied to some
902 assertions.
903
904 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
905 information out of configure.ac instead of relying on pcre.h.generic, which
906 is not stored in the repository.
907
908 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
909 -lreadline.
910
911 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
912
913 15. Applied Graycode's patch to put the top-level frame on the stack rather
914 than the heap when not using the stack for recursion. This gives a
915 performance improvement in many cases when recursion is not deep.
916
917 16. Experimental code added to "pcretest -C" to output the stack frame size.
918
919
920 Version 8.21 12-Dec-2011
921 ------------------------
922
923 1. Updating the JIT compiler.
924
925 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
926 are added as well.
927
928 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
929 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
930 calling _pcre_jit_exec. Some extra comments are added.
931
932 4. (*MARK) settings inside atomic groups that do not contain any capturing
933 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
934 was introduced by change 18 for 8.20.
935
936 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
937 ECMA-262 standard.
938
939 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
940 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
941 This bug was probably introduced by change 9 of 8.13.
942
943 7. While fixing 6 above, I noticed that a number of other items were being
944 incorrectly rejected as "not fixed length". This arose partly because newer
945 opcodes had not been added to the fixed-length checking code. I have (a)
946 corrected the bug and added tests for these items, and (b) arranged for an
947 error to occur if an unknown opcode is encountered while checking for fixed
948 length instead of just assuming "not fixed length". The items that were
949 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
950 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
951 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
952
953 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
954 being incorrectly compiled and would have given unpredicatble results.
955
956 9. A possessively repeated subpattern with minimum repeat count greater than
957 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
958 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
959 the first (A) could occur when it should not.
960
961 10. Add a cast and remove a redundant test from the code.
962
963 11. JIT should use pcre_malloc/pcre_free for allocation.
964
965 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
966 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
967 is anything other than /usr, -L is still shown).
968
969 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
970
971 14. Perl does not support \N without a following name in a [] class; PCRE now
972 also gives an error.
973
974 15. If a forward reference was repeated with an upper limit of around 2000,
975 it caused the error "internal error: overran compiling workspace". The
976 maximum number of forward references (including repeats) was limited by the
977 internal workspace, and dependent on the LINK_SIZE. The code has been
978 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
979 the default depends on LINK_SIZE. There is a new upper limit (for safety)
980 of around 200,000 forward references. While doing this, I also speeded up
981 the filling in of repeated forward references.
982
983 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
984 incorrectly expecting the subject to contain another "a" after the start.
985
986 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
987 in the match, the SKIP should be ignored. This was not happening; instead
988 the SKIP was being treated as NOMATCH. For patterns such as
989 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
990 tested.
991
992 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
993 now much more compatible with Perl, in particular in cases where the result
994 is a non-match for a non-anchored pattern. For example, if
995 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
996 "m", where previously it did not return a name. A side effect of this
997 change is that for partial matches, the last encountered mark name is
998 returned, as for non matches. A number of tests that were previously not
999 Perl-compatible have been moved into the Perl-compatible test files. The
1000 refactoring has had the pleasing side effect of removing one argument from
1001 the match() function, thus reducing its stack requirements.
1002
1003 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1004 subsequent uses of /S (without +) incorrectly behaved like /S+.
1005
1006 21. Retrieve executable code size support for the JIT compiler and fixing
1007 some warnings.
1008
1009 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1010 not work when the shorter character appeared right at the end of the
1011 subject string.
1012
1013 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1014 systems.
1015
1016 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1017 output it when the /M option is used in pcretest.
1018
1019 25. The CheckMan script was not being included in the distribution. Also, added
1020 an explicit "perl" to run Perl scripts from the PrepareRelease script
1021 because this is reportedly needed in Windows.
1022
1023 26. If study data was being save in a file and studying had not found a set of
1024 "starts with" bytes for the pattern, the data written to the file (though
1025 never used) was taken from uninitialized memory and so caused valgrind to
1026 complain.
1027
1028 27. Updated RunTest.bat as provided by Sheri Pierce.
1029
1030 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1031
1032 29. Computation of memory usage for the table of capturing group names was
1033 giving an unnecessarily large value.
1034
1035
1036 Version 8.20 21-Oct-2011
1037 ------------------------
1038
1039 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1040 a POSIX class. After further experiments with Perl, which convinced me that
1041 Perl has bugs and confusions, a closing square bracket is no longer allowed
1042 in a POSIX name. This bug also affected patterns with classes that started
1043 with full stops.
1044
1045 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1046 captured substring, but while checking the failing first alternative,
1047 substring 1 is temporarily captured. If the output vector supplied to
1048 pcre_exec() was not big enough for this capture, the yield of the function
1049 was still zero ("insufficient space for captured substrings"). This cannot
1050 be totally fixed without adding another stack variable, which seems a lot
1051 of expense for a edge case. However, I have improved the situation in cases
1052 such as /(a)(b)x|abc/ matched against "abc", where the return code
1053 indicates that fewer than the maximum number of slots in the ovector have
1054 been set.
1055
1056 3. Related to (2) above: when there are more back references in a pattern than
1057 slots in the output vector, pcre_exec() uses temporary memory during
1058 matching, and copies in the captures as far as possible afterwards. It was
1059 using the entire output vector, but this conflicts with the specification
1060 that only 2/3 is used for passing back captured substrings. Now it uses
1061 only the first 2/3, for compatibility. This is, of course, another edge
1062 case.
1063
1064 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1065 main code base, and can be used by building with --enable-jit. When this is
1066 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1067 runtime --no-jit option is given.
1068
1069 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1070 ovector, the return from the function was zero, implying that there were
1071 other matches that did not fit. The correct "exactly full" value is now
1072 returned.
1073
1074 6. If a subpattern that was called recursively or as a subroutine contained
1075 (*PRUNE) or any other control that caused it to give a non-standard return,
1076 invalid errors such as "Error -26 (nested recursion at the same subject
1077 position)" or even infinite loops could occur.
1078
1079 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1080 computing the minimum length on reaching *ACCEPT, and so ended up with the
1081 wrong value of 1 rather than 0. Further investigation indicates that
1082 computing a minimum subject length in the presence of *ACCEPT is difficult
1083 (think back references, subroutine calls), and so I have changed the code
1084 so that no minimum is registered for a pattern that contains *ACCEPT.
1085
1086 8. If (*THEN) was present in the first (true) branch of a conditional group,
1087 it was not handled as intended. [But see 16 below.]
1088
1089 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1090 Sheri Pierce.
1091
1092 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1093 the first byte in a match must be "a".
1094
1095 11. Change 17 for 8.13 increased the recursion depth for patterns like
1096 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1097 pattern contains any instances of (*THEN). If it does not, the old
1098 optimizations are restored. It would be nice to do this on a per-group
1099 basis, but at the moment that is not feasible.
1100
1101 12. In some environments, the output of pcretest -C is CRLF terminated. This
1102 broke RunTest's code that checks for the link size. A single white space
1103 character after the value is now allowed for.
1104
1105 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1106 For "fr", it uses the Windows-specific input and output files.
1107
1108 14. If (*THEN) appeared in a group that was called recursively or as a
1109 subroutine, it did not work as intended. [But see next item.]
1110
1111 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1112 pattern fragments (but not containing any | characters). If A and B are
1113 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1114 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1115 D. In other words, Perl considers parentheses that do not contain any |
1116 characters to be part of a surrounding alternative, whereas PCRE was
1117 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1118 differently. PCRE now behaves in the same way as Perl, except in the case
1119 of subroutine/recursion calls such as (?1) which have in any case always
1120 been different (but PCRE had them first :-).
1121
1122 16. Related to 15 above: Perl does not treat the | in a conditional group as
1123 creating alternatives. Such a group is treated in the same way as an
1124 ordinary group without any | characters when processing (*THEN). PCRE has
1125 been changed to match Perl's behaviour.
1126
1127 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1128 RunGrepTest script failed.
1129
1130 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1131 inevitable for groups that contain captures, but it can lead to a lot of
1132 stack use in large patterns. The old behaviour has been restored for atomic
1133 groups that do not contain any capturing parentheses.
1134
1135 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1136 suppress the check for a minimum subject length at run time. (If it was
1137 given to pcre_exec() or pcre_dfa_exec() it did work.)
1138
1139 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1140 fail to work when decoding hex characters in data strings in EBCDIC
1141 environments.
1142
1143 21. It appears that in at least one Mac OS environment, the isxdigit() function
1144 is implemented as a macro that evaluates to its argument more than once,
1145 contravening the C 90 Standard (I haven't checked a later standard). There
1146 was an instance in pcretest which caused it to go wrong when processing
1147 \x{...} escapes in subject strings. The has been rewritten to avoid using
1148 things like p++ in the argument of isxdigit().
1149
1150
1151 Version 8.13 16-Aug-2011
1152 ------------------------
1153
1154 1. The Unicode data tables have been updated to Unicode 6.0.0.
1155
1156 2. Two minor typos in pcre_internal.h have been fixed.
1157
1158 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1159 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1160 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1161
1162 4. There were a number of related bugs in the code for matching backrefences
1163 caselessly in UTF-8 mode when codes for the characters concerned were
1164 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1165 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1166 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1167 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1168 2-byte code at the end of the subject (it thought there wasn't enough data
1169 left).
1170
1171 5. Comprehensive information about what went wrong is now returned by
1172 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1173 as the output vector has at least 2 elements. The offset of the start of
1174 the failing character and a reason code are placed in the vector.
1175
1176 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1177 now returned is for the first byte of the failing character, instead of the
1178 last byte inspected. This is an incompatible change, but I hope it is small
1179 enough not to be a problem. It makes the returned offset consistent with
1180 pcre_exec() and pcre_dfa_exec().
1181
1182 7. pcretest now gives a text phrase as well as the error number when
1183 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1184 failure, the offset and reason code are output.
1185
1186 8. When \R was used with a maximizing quantifier it failed to skip backwards
1187 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1188 back over a single character (\n). This seems wrong (because it treated the
1189 two characters as a single entity when going forwards), conflicts with the
1190 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1191 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1192 has been changed.
1193
1194 9. Some internal refactoring has changed the processing so that the handling
1195 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1196 time (the PCRE_DOTALL option was changed this way some time ago: version
1197 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1198 which was always a bit of a fudge. It also means that there is one less
1199 argument for the match() function, which reduces its stack requirements
1200 slightly. This change also fixes an incompatibility with Perl: the pattern
1201 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1202
1203 10. More internal refactoring has drastically reduced the number of recursive
1204 calls to match() for possessively repeated groups such as (abc)++ when
1205 using pcre_exec().
1206
1207 11. While implementing 10, a number of bugs in the handling of groups were
1208 discovered and fixed:
1209
1210 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1211 (a|)*(?1) gave a compile-time internal error.
1212 ((a|)+)+ did not notice that the outer group could match an empty string.
1213 (^a|^)+ was not marked as anchored.
1214 (.*a|.*)+ was not marked as matching at start or after a newline.
1215
1216 12. Yet more internal refactoring has removed another argument from the match()
1217 function. Special calls to this function are now indicated by setting a
1218 value in a variable in the "match data" data block.
1219
1220 13. Be more explicit in pcre_study() instead of relying on "default" for
1221 opcodes that mean there is no starting character; this means that when new
1222 ones are added and accidentally left out of pcre_study(), testing should
1223 pick them up.
1224
1225 14. The -s option of pcretest has been documented for ages as being an old
1226 synonym of -m (show memory usage). I have changed it to mean "force study
1227 for every regex", that is, assume /S for every regex. This is similar to -i
1228 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1229 using it. It makes it easier to run collections of tests with and without
1230 study enabled, and thereby test pcre_study() more easily. All the standard
1231 tests are now run with and without -s (but some patterns can be marked as
1232 "never study" - see 20 below).
1233
1234 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1235 restoration of the capturing data to the outer values was not happening
1236 correctly.
1237
1238 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1239 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1240 pattern had matched an empty string, and so incorrectly returned a no
1241 match.
1242
1243 17. There was optimizing code for the last branch of non-capturing parentheses,
1244 and also for the obeyed branch of a conditional subexpression, which used
1245 tail recursion to cut down on stack usage. Unfortunately, now that there is
1246 the possibility of (*THEN) occurring in these branches, tail recursion is
1247 no longer possible because the return has to be checked for (*THEN). These
1248 two optimizations have therefore been removed. [But see 8.20/11 above.]
1249
1250 18. If a pattern containing \R was studied, it was assumed that \R always
1251 matched two bytes, thus causing the minimum subject length to be
1252 incorrectly computed because \R can also match just one byte.
1253
1254 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1255 was incorrectly computed.
1256
1257 20. If /S is present twice on a test pattern in pcretest input, it now
1258 *disables* studying, thereby overriding the use of -s on the command line
1259 (see 14 above). This is necessary for one or two tests to keep the output
1260 identical in both cases.
1261
1262 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1263 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1264
1265 22. When an atomic group that contained a capturing parenthesis was
1266 successfully matched, but the branch in which it appeared failed, the
1267 capturing was not being forgotten if a higher numbered group was later
1268 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1269 group 1 to "a", when in fact it should be unset. This applied to multi-
1270 branched capturing and non-capturing groups, repeated or not, and also to
1271 positive assertions (capturing in negative assertions does not happen
1272 in PCRE) and also to nested atomic groups.
1273
1274 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1275 subject after a captured substring, to make it easier to tell which of a
1276 number of identical substrings has been captured.
1277
1278 24. The way atomic groups are processed by pcre_exec() has been changed so that
1279 if they are repeated, backtracking one repetition now resets captured
1280 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1281 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1282 "aaa". Previously, it would have been "a". As part of this code
1283 refactoring, the way recursive calls are handled has also been changed.
1284
1285 25. If an assertion condition captured any substrings, they were not passed
1286 back unless some other capturing happened later. For example, if
1287 (?(?=(a))a) was matched against "a", no capturing was returned.
1288
1289 26. When studying a pattern that contained subroutine calls or assertions,
1290 the code for finding the minimum length of a possible match was handling
1291 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1292 group 1 called group 2 while simultaneously a separate group 2 called group
1293 1). A stack overflow occurred in this case. I have fixed this by limiting
1294 the recursion depth to 10.
1295
1296 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1297 Fortmann. This supports explicit test numbers on the command line, and has
1298 argument validation and error reporting.
1299
1300 28. An instance of \X with an unlimited repeat could fail if at any point the
1301 first character it looked at was a mark character.
1302
1303 29. Some minor code refactoring concerning Unicode properties and scripts
1304 should reduce the stack requirement of match() slightly.
1305
1306 30. Added the '=' option to pcretest to check the setting of unused capturing
1307 slots at the end of the pattern, which are documented as being -1, but are
1308 not included in the return count.
1309
1310 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1311 compiled something random. Now it gives a compile-time error (as does
1312 Perl).
1313
1314 32. A *MARK encountered during the processing of a positive assertion is now
1315 recorded and passed back (compatible with Perl).
1316
1317 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1318 had alternative anchored branches, the search for a second match in a line
1319 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1320 matched the line "0102" twice. The same bug affected patterns that started
1321 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1322 twice.
1323
1324 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1325 does, and because of capturing effects, quantifying parenthesized
1326 assertions may at times be useful. Quantifiers are now allowed for
1327 parenthesized assertions.
1328
1329 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1330
1331 36. \g was being checked for fancy things in a character class, when it should
1332 just be a literal "g".
1333
1334 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1335 appearance of a nested POSIX class supersedes an apparent external class.
1336 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1337 unescaped square brackets may also appear as part of class names. For
1338 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1339 more like Perl. (But see 8.20/1 above.)
1340
1341 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1342 was because it thought it was \N{name}, which is not supported).
1343
1344 39. Add minix to OS list not supporting the -S option in pcretest.
1345
1346 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1347 cannot analyze patterns in sufficient detail to catch mutual recursions
1348 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1349 subgroup is called recursively as a subpattern for a second time at the
1350 same position in the subject string. In previous releases this might have
1351 been caught by the recursion limit, or it might have run out of stack.
1352
1353 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1354 happen only once. PCRE was, however incorrectly giving a compile time error
1355 "recursive call could loop indefinitely" because it cannot analyze the
1356 pattern in sufficient detail. The compile time test no longer happens when
1357 PCRE is compiling a conditional subpattern, but actual runaway loops are
1358 now caught at runtime (see 40 above).
1359
1360 42. It seems that Perl allows any characters other than a closing parenthesis
1361 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1362 has been changed to be the same.
1363
1364 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1365 as not to get warnings when autogen.sh is called. Also changed
1366 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1367
1368 44. To help people who use pcregrep to scan files containing exceedingly long
1369 lines, the following changes have been made:
1370
1371 (a) The default value of the buffer size parameter has been increased from
1372 8K to 20K. (The actual buffer used is three times this size.)
1373
1374 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1375 PCRE is built.
1376
1377 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1378 to be set at run time.
1379
1380 (d) Numerical values in pcregrep options can be followed by K or M, for
1381 example --buffer-size=50K.
1382
1383 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1384 given and the return code is set to 2.
1385
1386 45. Add a pointer to the latest mark to the callout data block.
1387
1388 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1389 partial match of an empty string instead of no match. This was specific to
1390 the use of ".".
1391
1392 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1393 complete match instead of a partial match. This bug was dependent on both
1394 the PCRE_UTF8 and PCRE_DOTALL options being set.
1395
1396 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1397 starting byte set, because \b was not being ignored.
1398
1399
1400 Version 8.12 15-Jan-2011
1401 ------------------------
1402
1403 1. Fixed some typos in the markup of the man pages, and wrote a script that
1404 checks for such things as part of the documentation building process.
1405
1406 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1407 --match-limit and --recursion-limit options (added for 8.11). In
1408 particular, this made one of the standard tests fail. (The integer value
1409 went into the wrong half of a long int.)
1410
1411 3. If the --colour option was given to pcregrep with -v (invert match), it
1412 did strange things, either producing crazy output, or crashing. It should,
1413 of course, ignore a request for colour when reporting lines that do not
1414 match.
1415
1416 4. Another pcregrep bug caused similar problems if --colour was specified with
1417 -M (multiline) and the pattern match finished with a line ending.
1418
1419 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1420 matched in multiline mode, the following line was shown as part of the
1421 match. This seems wrong, so I have changed it.
1422
1423 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1424 the check for further matches in the same line (so they could be coloured)
1425 to overrun the end of the current line. If another match was found, it was
1426 incorrectly shown (and then shown again when found in the next line).
1427
1428 7. If pcregrep was compiled under Windows, there was a reference to the
1429 function pcregrep_exit() before it was defined. I am assuming this was
1430 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1431 reported by a user. I've moved the definition above the reference.
1432
1433
1434 Version 8.11 10-Dec-2010
1435 ------------------------
1436
1437 1. (*THEN) was not working properly if there were untried alternatives prior
1438 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1439 backtracked to try for "b" instead of moving to the next alternative branch
1440 at the same level (in this case, to look for "c"). The Perl documentation
1441 is clear that when (*THEN) is backtracked onto, it goes to the "next
1442 alternative in the innermost enclosing group".
1443
1444 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1445 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1446 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1447 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1448 (*THEN).
1449
1450 3. If \s appeared in a character class, it removed the VT character from
1451 the class, even if it had been included by some previous item, for example
1452 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1453 of \s, but is part of the POSIX "space" class.)
1454
1455 4. A partial match never returns an empty string (because you can always
1456 match an empty string at the end of the subject); however the checking for
1457 an empty string was starting at the "start of match" point. This has been
1458 changed to the "earliest inspected character" point, because the returned
1459 data for a partial match starts at this character. This means that, for
1460 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1461 (previously it gave "no match").
1462
1463 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1464 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1465 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1466 has an implication that the given string is incomplete (because a partial
1467 match is preferred over a full match). For this reason, these items now
1468 give a partial match in this situation. [Aside: previously, the one case
1469 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1470 match rather than a full match, which was wrong by the old rules, but is
1471 now correct.]
1472
1473 6. There was a bug in the handling of #-introduced comments, recognized when
1474 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1475 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1476 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1477 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1478 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1479 places in pcre_compile().
1480
1481 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1482 comments when looking ahead for named forward references to subpatterns,
1483 the only newline sequence it recognized was NL. It now handles newlines
1484 according to the set newline convention.
1485
1486 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1487 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1488 cater for a lack of strerror(). These oversights have been fixed.
1489
1490 9. Added --match-limit and --recursion-limit to pcregrep.
1491
1492 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1493
1494 11. When the -o option was used, pcregrep was setting a return code of 1, even
1495 when matches were found, and --line-buffered was not being honoured.
1496
1497 12. Added an optional parentheses number to the -o and --only-matching options
1498 of pcregrep.
1499
1500 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1501 can match an empty string. The code to do it in pcretest and pcredemo
1502 needed fixing:
1503
1504 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1505 only one byte after an empty string match just before CRLF (this case
1506 just got forgotten; "any" and "anycrlf" were OK).
1507
1508 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1509 mode when an empty string match preceded an ASCII character followed by
1510 a non-ASCII character. (The code for advancing by one character rather
1511 than one byte was nonsense.)
1512
1513 (c) The pcredemo.c sample program did not have any code at all to handle
1514 the cases when CRLF is a valid newline sequence.
1515
1516 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1517 as a starting offset was within the subject string. There is now a new
1518 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1519 negative or greater than the length of the string. In order to test this,
1520 pcretest is extended to allow the setting of negative starting offsets.
1521
1522 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1523 starting offset points to the beginning of a UTF-8 character was
1524 unnecessarily clumsy. I tidied it up.
1525
1526 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1527 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1528
1529 17. Nobody had reported that the --include_dir option, which was added in
1530 release 7.7 should have been called --include-dir (hyphen, not underscore)
1531 for compatibility with GNU grep. I have changed it to --include-dir, but
1532 left --include_dir as an undocumented synonym, and the same for
1533 --exclude-dir, though that is not available in GNU grep, at least as of
1534 release 2.5.4.
1535
1536 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1537 characters from a string of bytes) have been redefined so as not to use
1538 loops, in order to improve performance in some environments. At the same
1539 time, I abstracted some of the common code into auxiliary macros to save
1540 repetition (this should not affect the compiled code).
1541
1542 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1543 compile-time error is now given if \c is not followed by an ASCII
1544 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1545 different, and any byte value is allowed.)
1546
1547 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1548 START_OPTIMIZE option, which is now allowed at compile time - but just
1549 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1550 to pcregrep and other applications that have no direct access to PCRE
1551 options. The new /Y option in pcretest sets this option when calling
1552 pcre_compile().
1553
1554 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1555 back references. Groups containing recursive back references were forced to
1556 be atomic by that change, but in the case of named groups, the amount of
1557 memory required was incorrectly computed, leading to "Failed: internal
1558 error: code overflow". This has been fixed.
1559
1560 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1561 pcretest.c, to avoid build problems in some Borland environments.
1562
1563
1564 Version 8.10 25-Jun-2010
1565 ------------------------
1566
1567 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1568 THEN.
1569
1570 2. (*ACCEPT) was not working when inside an atomic group.
1571
1572 3. Inside a character class, \B is treated as a literal by default, but
1573 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1574 causes the error). The code is unchanged, but I tidied the documentation.
1575
1576 4. Inside a character class, PCRE always treated \R and \X as literals,
1577 whereas Perl faults them if its -w option is set. I have changed PCRE so
1578 that it faults them when PCRE_EXTRA is set.
1579
1580 5. Added support for \N, which always matches any character other than
1581 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1582
1583 6. When compiling pcregrep with newer versions of gcc which may have
1584 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1585 declared with attribute warn_unused_result" were given. Just casting the
1586 result to (void) does not stop the warnings; a more elaborate fudge is
1587 needed. I've used a macro to implement this.
1588
1589 7. Minor change to pcretest.c to avoid a compiler warning.
1590
1591 8. Added four artifical Unicode properties to help with an option to make
1592 \s etc use properties (see next item). The new properties are: Xan
1593 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1594
1595 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1596 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1597 this option. Modified pcretest to add /W to test this facility. Added
1598 REG_UCP to make it available via the POSIX interface.
1599
1600 10. Added --line-buffered to pcregrep.
1601
1602 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1603 studied, and the match started with a letter with a code point greater than
1604 127 whose first byte was different to the first byte of the other case of
1605 the letter, the other case of this starting letter was not recognized
1606 (#976).
1607
1608 12. If a pattern that was studied started with a repeated Unicode property
1609 test, for example, \p{Nd}+, there was the theoretical possibility of
1610 setting up an incorrect bitmap of starting bytes, but fortunately it could
1611 not have actually happened in practice until change 8 above was made (it
1612 added property types that matched character-matching opcodes).
1613
1614 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1615 possible starting bytes for non-anchored patterns.
1616
1617 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1618 \R, and also a number of cases that involve Unicode properties, both
1619 explicit and implicit when PCRE_UCP is set.
1620
1621 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1622 input, it could crash or give wrong results if characters with values
1623 greater than 0xc0 were present in the subject string. (Detail: it assumed
1624 UTF-8 input when processing these items.)
1625
1626 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1627 size_t is 64-bit (#991).
1628
1629 17. Added a check for running out of memory when PCRE is compiled with
1630 --disable-stack-for-recursion (#990).
1631
1632 18. If the last data line in a file for pcretest does not have a newline on
1633 the end, a newline was missing in the output.
1634
1635 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1636 less than 128) in its various bitmaps. However, there is a facility for
1637 generating tables according to the current locale when PCRE is compiled. It
1638 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1639 characters, are recognized by isspace() and therefore were getting set in
1640 these tables, and indeed these tables seem to approximate to ISO 8859. This
1641 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1642 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1643 which of course cannot start UTF-8 characters. I have changed the code so
1644 that only real ASCII characters (less than 128) and the correct starting
1645 bytes for UTF-8 encodings are set for characters greater than 127 when in
1646 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1647 altogether.)
1648
1649 20. Added the /T option to pcretest so as to be able to run tests with non-
1650 standard character tables, thus making it possible to include the tests
1651 used for 19 above in the standard set of tests.
1652
1653 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1654 reference to a subpattern the other side of a comment that contains an
1655 opening parenthesis caused either an internal compiling error, or a
1656 reference to the wrong subpattern.
1657
1658
1659 Version 8.02 19-Mar-2010
1660 ------------------------
1661
1662 1. The Unicode data tables have been updated to Unicode 5.2.0.
1663
1664 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1665 configured.
1666
1667 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1668 original author of that file, following a query about its status.
1669
1670 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1671 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1672
1673 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1674 quantifier applied to a forward-referencing subroutine call, could compile
1675 incorrect code or give the error "internal error: previously-checked
1676 referenced subpattern not found".
1677
1678 6. Both MS Visual Studio and Symbian OS have problems with initializing
1679 variables to point to external functions. For these systems, therefore,
1680 pcre_malloc etc. are now initialized to local functions that call the
1681 relevant global functions.
1682
1683 7. There were two entries missing in the vectors called coptable and poptable
1684 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1685 I've fixed the data, and added a kludgy way of testing at compile time that
1686 the lengths are correct (equal to the number of opcodes).
1687
1688 8. Following on from 7, I added a similar kludge to check the length of the
1689 eint vector in pcreposix.c.
1690
1691 9. Error texts for pcre_compile() are held as one long string to avoid too
1692 much relocation at load time. To find a text, the string is searched,
1693 counting zeros. There was no check for running off the end of the string,
1694 which could happen if a new error number was added without updating the
1695 string.
1696
1697 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1698
1699 11. \K was not working if it appeared in an atomic group or in a group that
1700 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1701 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1702 the assertion is positive, but not if it is negative.
1703
1704 12. Change 11 fortuitously reduced the size of the stack frame used in the
1705 "match()" function of pcre_exec.c by one pointer. Forthcoming
1706 implementation of support for (*MARK) will need an extra pointer on the
1707 stack; I have reserved it now, so that the stack frame size does not
1708 decrease.
1709
1710 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1711 item in branch that calls a recursion is a subroutine call - as in the
1712 second branch in the above example - was incorrectly given the compile-
1713 time error "recursive call could loop indefinitely" because pcre_compile()
1714 was not correctly checking the subroutine for matching a non-empty string.
1715
1716 14. The checks for overrunning compiling workspace could trigger after an
1717 overrun had occurred. This is a "should never occur" error, but it can be
1718 triggered by pathological patterns such as hundreds of nested parentheses.
1719 The checks now trigger 100 bytes before the end of the workspace.
1720
1721 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1722
1723
1724 Version 8.01 19-Jan-2010
1725 ------------------------
1726
1727 1. If a pattern contained a conditional subpattern with only one branch (in
1728 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1729 computed the wrong minimum data length (which is of course zero for such
1730 subpatterns). This could cause incorrect "no match" results.
1731
1732 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1733 the pattern is reset in the first branch, pcre_compile() failed with
1734 "internal error: code overflow at offset...". This happened only when
1735 the reset was to the original external option setting. (An optimization
1736 abstracts leading options settings into an external setting, which was the
1737 cause of this.)
1738
1739 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1740 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1741 assertion pattern did not match (meaning that the assertion was true), it
1742 was incorrectly treated as false if the SKIP had been reached during the
1743 matching. This also applied to assertions used as conditions.
1744
1745 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1746 assertion subpattern, including such a pattern used as a condition,
1747 unpredictable results occurred, instead of the error return
1748 PCRE_ERROR_DFA_UITEM.
1749
1750 5. The C++ GlobalReplace function was not working like Perl for the special
1751 situation when an empty string is matched. It now does the fancy magic
1752 stuff that is necessary.
1753
1754 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1755 removed. (These were left over from very, very early versions of PCRE.)
1756
1757 7. Some cosmetic changes to the code to make life easier when compiling it
1758 as part of something else:
1759
1760 (a) Change DEBUG to PCRE_DEBUG.
1761
1762 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1763 called "current" as "current_branch", to prevent a collision with the
1764 Linux macro when compiled as a kernel module.
1765
1766 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1767 prevent a collision with the Linux macro when compiled as a kernel
1768 module.
1769
1770 8. In pcre_compile() there are some checks for integer overflows that used to
1771 cast potentially large values to (double). This has been changed to that
1772 when building, a check for int64_t is made, and if it is found, it is used
1773 instead, thus avoiding the use of floating point arithmetic. (There is no
1774 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1775 double.
1776
1777 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1778 2005 (difference between two addresses compared to an unsigned value).
1779
1780 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1781 custom one, because of the following reported problem in Windows:
1782
1783 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1784 under Win32.
1785 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1786 therefore missing the function definition.
1787 - The compiler thus generates a "C" signature for the test function.
1788 - The linker fails to find the "C" function.
1789 - PCRE fails to configure if asked to do so against libbz2.
1790
1791 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1792 messages were output:
1793
1794 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1795 rerunning libtoolize, to keep the correct libtool macros in-tree.
1796 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1797
1798 I have done both of these things.
1799
1800 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1801 most of the time, it *can* run out if it is given a pattern that contains a
1802 runaway infinite recursion. I updated the discussion in the pcrestack man
1803 page.
1804
1805 13. Now that we have gone to the x.xx style of version numbers, the minor
1806 version may start with zero. Using 08 or 09 is a bad idea because users
1807 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1808 interpreted as invalid octal numbers. I've updated the previous comment in
1809 configure.ac, and also added a check that gives an error if 08 or 09 are
1810 used.
1811
1812 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1813 causing partial matching to fail when the end of the subject matched \W
1814 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1815
1816 15. There were some discrepancies between the declarations in pcre_internal.h
1817 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1818 their definitions. The declarations used "const uschar *" and the
1819 definitions used USPTR. Even though USPTR is normally defined as "const
1820 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1821 reported that: "This difference in casting confuses some C++ compilers, for
1822 example, SunCC recognizes above declarations as different functions and
1823 generates broken code for hbpcre." I have changed the declarations to use
1824 USPTR.
1825
1826 16. GNU libtool is named differently on some systems. The autogen.sh script now
1827 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1828 (FreeBSD).
1829
1830 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1831 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1832 comment: "Figure out how to create a longlong from a string: strtoll and
1833 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1834 instance, but it only takes 2 args instead of 3!"
1835
1836 18. A subtle bug concerned with back references has been fixed by a change of
1837 specification, with a corresponding code fix. A pattern such as
1838 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1839 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1840 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1841 same bug. Such groups have to be quantified to be useful, or contained
1842 inside another quantified group. (If there's no repetition, the reference
1843 can never match.) The problem arises because, having left the group and
1844 moved on to the rest of the pattern, a later failure that backtracks into
1845 the group uses the captured value from the final iteration of the group
1846 rather than the correct earlier one. I have fixed this in PCRE by forcing
1847 any group that contains a reference to itself to be an atomic group; that
1848 is, there cannot be any backtracking into it once it has completed. This is
1849 similar to recursive and subroutine calls.
1850
1851
1852 Version 8.00 19-Oct-09
1853 ----------------------
1854
1855 1. The table for translating pcre_compile() error codes into POSIX error codes
1856 was out-of-date, and there was no check on the pcre_compile() error code
1857 being within the table. This could lead to an OK return being given in
1858 error.
1859
1860 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1861 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1862 in a Windows environment.
1863
1864 3. The pcregrep --count option prints the count for each file even when it is
1865 zero, as does GNU grep. However, pcregrep was also printing all files when
1866 --files-with-matches was added. Now, when both options are given, it prints
1867 counts only for those files that have at least one match. (GNU grep just
1868 prints the file name in this circumstance, but including the count seems
1869 more useful - otherwise, why use --count?) Also ensured that the
1870 combination -clh just lists non-zero counts, with no names.
1871
1872 4. The long form of the pcregrep -F option was incorrectly implemented as
1873 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1874 but it seems right to fix it, and I didn't think it was worth preserving
1875 the old behaviour.
1876
1877 5. The command line items --regex=pattern and --regexp=pattern were not
1878 recognized by pcregrep, which required --regex pattern or --regexp pattern
1879 (with a space rather than an '='). The man page documented the '=' forms,
1880 which are compatible with GNU grep; these now work.
1881
1882 6. No libpcreposix.pc file was created for pkg-config; there was just
1883 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1884
1885 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1886 when UCP support is not needed, by modifying the Python script that
1887 generates it from Unicode data files. This should not matter if the module
1888 is correctly used as a library, but I received one complaint about 50K of
1889 unwanted data. My guess is that the person linked everything into his
1890 program rather than using a library. Anyway, it does no harm.
1891
1892 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1893 was a minimum greater than 1 for a wide character in a possessive
1894 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1895 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1896 character. Chaos in the form of incorrect output or a compiling loop could
1897 result.
1898
1899 9. The restrictions on what a pattern can contain when partial matching is
1900 requested for pcre_exec() have been removed. All patterns can now be
1901 partially matched by this function. In addition, if there are at least two
1902 slots in the offset vector, the offset of the earliest inspected character
1903 for the match and the offset of the end of the subject are set in them when
1904 PCRE_ERROR_PARTIAL is returned.
1905
1906 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1907 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1908 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1909 and may be more useful for multi-segment matching.
1910
1911 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1912 used to be given if ever the end of the subject was reached; now it is
1913 given only if matching could not proceed because another character was
1914 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1915 string "Z", which now yields "no match" instead of "partial match". In the
1916 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1917 final character ended with (*FAIL).
1918
1919 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1920 if the pattern had a "must contain" character that was already found in the
1921 earlier partial match, unless partial matching was again requested. For
1922 example, with the pattern /dog.(body)?/, the "must contain" character is
1923 "g". If the first part-match was for the string "dog", restarting with
1924 "sbody" failed. This bug has been fixed.
1925
1926 13. The string returned by pcre_dfa_exec() after a partial match has been
1927 changed so that it starts at the first inspected character rather than the
1928 first character of the match. This makes a difference only if the pattern
1929 starts with a lookbehind assertion or \b or \B (\K is not supported by
1930 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1931 matching functions compatible, and I think it's the right thing to do.
1932
1933 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1934 so that the demonstration program is easily available in environments where
1935 PCRE has not been installed from source.
1936
1937 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1938 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1939 library.
1940
1941 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1942 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1943 is not the first non-POSIX option to be added. Clearly some people find
1944 these options useful.
1945
1946 17. If a caller to the POSIX matching function regexec() passes a non-zero
1947 value for nmatch with a NULL value for pmatch, the value of
1948 nmatch is forced to zero.
1949
1950 18. RunGrepTest did not have a test for the availability of the -u option of
1951 the diff command, as RunTest does. It now checks in the same way as
1952 RunTest, and also checks for the -b option.
1953
1954 19. If an odd number of negated classes containing just a single character
1955 interposed, within parentheses, between a forward reference to a named
1956 subpattern and the definition of the subpattern, compilation crashed with
1957 an internal error, complaining that it could not find the referenced
1958 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1959 [The bug was that it was starting one character too far in when skipping
1960 over the character class, thus treating the ] as data rather than
1961 terminating the class. This meant it could skip too much.]
1962
1963 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1964 /g option in pcretest when the pattern contains \K, which makes it possible
1965 to have an empty string match not at the start, even when the pattern is
1966 anchored. Updated pcretest and pcredemo to use this option.
1967
1968 21. If the maximum number of capturing subpatterns in a recursion was greater
1969 than the maximum at the outer level, the higher number was returned, but
1970 with unset values at the outer level. The correct (outer level) value is
1971 now given.
1972
1973 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1974 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1975 make it do so. The string so far is captured, making this feature
1976 compatible with Perl.
1977
1978 23. The tests have been re-organized, adding tests 11 and 12, to make it
1979 possible to check the Perl 5.10 features against Perl 5.10.
1980
1981 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1982 pattern matches a fixed length string. PCRE did not allow this; now it
1983 does. Neither allows recursion.
1984
1985 25. I finally figured out how to implement a request to provide the minimum
1986 length of subject string that was needed in order to match a given pattern.
1987 (It was back references and recursion that I had previously got hung up
1988 on.) This code has now been added to pcre_study(); it finds a lower bound
1989 to the length of subject needed. It is not necessarily the greatest lower
1990 bound, but using it to avoid searching strings that are too short does give
1991 some useful speed-ups. The value is available to calling programs via
1992 pcre_fullinfo().
1993
1994 26. While implementing 25, I discovered to my embarrassment that pcretest had
1995 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1996 study optimizations had never been tested with that matching function.
1997 Oops. What is worse, even when it was passed study data, there was a bug in
1998 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1999 were also very few tests of studied patterns with pcre_dfa_exec().
2000
2001 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2002 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2003 on the other side of the coin, they are no longer allowed to have different
2004 names, because these cannot be distinguished in PCRE, and this has caused
2005 confusion. (This is a difference from Perl.)
2006
2007 28. When duplicate subpattern names are present (necessarily with different
2008 numbers, as required by 27 above), and a test is made by name in a
2009 conditional pattern, either for a subpattern having been matched, or for
2010 recursion in such a pattern, all the associated numbered subpatterns are
2011 tested, and the overall condition is true if the condition is true for any
2012 one of them. This is the way Perl works, and is also more like the way
2013 testing by number works.
2014
2015
2016 Version 7.9 11-Apr-09
2017 ---------------------
2018
2019 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2020 (pcretest), all targets were linked against these libraries. This included
2021 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2022 libraries. This caused unwanted dependencies to be created. This problem
2023 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2024 pcretest is linked with readline.
2025
2026 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2027 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2028 moved outside it again, because FALSE and TRUE are already defined in AIX,
2029 but BOOL is not.
2030
2031 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2032 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2033
2034 4. The pcregrep documentation said spaces were inserted as well as colons (or
2035 hyphens) following file names and line numbers when outputting matching
2036 lines. This is not true; no spaces are inserted. I have also clarified the
2037 wording for the --colour (or --color) option.
2038
2039 5. In pcregrep, when --colour was used with -o, the list of matching strings
2040 was not coloured; this is different to GNU grep, so I have changed it to be
2041 the same.
2042
2043 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2044 each matching line was coloured. Now it goes on to look for further matches
2045 of any of the test patterns, which is the same behaviour as GNU grep.
2046
2047 7. A pattern that could match an empty string could cause pcregrep to loop; it
2048 doesn't make sense to accept an empty string match in pcregrep, so I have
2049 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2050 seems to be how GNU grep behaves. [But see later change 40 for release
2051 8.33.]
2052
2053 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2054 start or after a newline", because the conditional assertion was not being
2055 correctly handled. The rule now is that both the assertion and what follows
2056 in the first alternative must satisfy the test.
2057
2058 9. If auto-callout was enabled in a pattern with a conditional group whose
2059 condition was an assertion, PCRE could crash during matching, both with
2060 pcre_exec() and pcre_dfa_exec().
2061
2062 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2063 used for matching.
2064
2065 11. Unicode property support in character classes was not working for
2066 characters (bytes) greater than 127 when not in UTF-8 mode.
2067
2068 12. Added the -M command line option to pcretest.
2069
2070 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2071
2072 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2073
2074 16. Added comments and documentation about mis-use of no_arg in the C++
2075 wrapper.
2076
2077 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2078 from Martin Jerabek that uses macro names for all relevant character and
2079 string constants.
2080
2081 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2082 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2083 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2084 these, but not everybody uses configure.
2085
2086 19. A conditional group that had only one branch was not being correctly
2087 recognized as an item that could match an empty string. This meant that an
2088 enclosing group might also not be so recognized, causing infinite looping
2089 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2090 with the subject "ab", where knowledge that the repeated group can match
2091 nothing is needed in order to break the loop.
2092
2093 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2094 exec(), but without supplying a callout function, matching went wrong.
2095
2096 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2097 leak if the size of the offset vector was greater than 30. When the vector
2098 is smaller, the saved offsets during recursion go onto a local stack
2099 vector, but for larger vectors malloc() is used. It was failing to free
2100 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2101 error, in fact).
2102
2103 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2104 heapframe that is used only when UTF-8 support is enabled. This caused no
2105 problem, but was untidy.
2106
2107 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2108 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2109 included within another project.
2110
2111 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2112 slightly modified by me:
2113
2114 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2115 not building pcregrep.
2116
2117 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2118 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2119
2120 25. Forward references, both numeric and by name, in patterns that made use of
2121 duplicate group numbers, could behave incorrectly or give incorrect errors,
2122 because when scanning forward to find the reference group, PCRE was not
2123 taking into account the duplicate group numbers. A pattern such as
2124 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2125
2126 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2127 the feature of a custom pointer more persuasive (as requested by a user).
2128
2129 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2130 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2131 pre-defined.
2132
2133 28. Added support for (*UTF8) at the start of a pattern.
2134
2135 29. Arrange for flags added by the "release type" setting in CMake to be shown
2136 in the configuration summary.
2137
2138
2139 Version 7.8 05-Sep-08
2140 ---------------------
2141
2142 1. Replaced UCP searching code with optimized version as implemented for Ad
2143 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2144 stage table and inline lookup instead of a function, giving speed ups of 2
2145 to 5 times on some simple patterns that I tested. Permission was given to
2146 distribute the MultiStage2.py script that generates the tables (it's not in
2147 the tarball, but is in the Subversion repository).
2148
2149 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2150 scripts.
2151
2152 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2153 a group with a zero qualifier. The result of the study could be incorrect,
2154 or the function might crash, depending on the pattern.
2155
2156 4. Caseless matching was not working for non-ASCII characters in back
2157 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2158 It now works when Unicode Property Support is available.
2159
2160 5. In pcretest, an escape such as \x{de} in the data was always generating
2161 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2162 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2163 truncation.
2164
2165 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2166
2167 7. Added two (int) casts to pcregrep when printing the difference of two
2168 pointers, in case they are 64-bit values.
2169
2170 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2171 test 2 if it fails.
2172
2173 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2174 and a #define of that name to empty if it is not externally set. This is to
2175 allow users of MSVC to set it if necessary.
2176
2177 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2178 the convenience functions in the pcre_get.c source file.
2179
2180 11. An option change at the start of a pattern that had top-level alternatives
2181 could cause overwriting and/or a crash. This command provoked a crash in
2182 some environments:
2183
2184 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2185
2186 This potential security problem was recorded as CVE-2008-2371.
2187
2188 12. For a pattern where the match had to start at the beginning or immediately
2189 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2190 pcre_dfa_exec() could read past the end of the passed subject if there was
2191 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2192 pcretest so that it places the subject at the end of its malloc-ed buffer.
2193
2194 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2195 exec() might read past the end of the data buffer in UTF-8 mode.
2196
2197 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2198 the data contained the byte 0x85 as part of a UTF-8 character within its
2199 first line. This applied both to normal and DFA matching.
2200
2201 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2202 /^[^d]*?$/8 failed to match "abc".
2203
2204 16. Added a missing copyright notice to pcrecpp_internal.h.
2205
2206 17. Make it more clear in the documentation that values returned from
2207 pcre_exec() in ovector are byte offsets, not character counts.
2208
2209 18. Tidied a few places to stop certain compilers from issuing warnings.
2210
2211 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2212 supplied by Stefan Weber. I made a further small update for 7.8 because
2213 there is a change of source arrangements: the pcre_searchfuncs.c module is
2214 replaced by pcre_ucd.c.
2215
2216
2217 Version 7.7 07-May-08
2218 ---------------------
2219
2220 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2221 a string to a long long, pretend we don't even have a long long." This is
2222 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2223
2224 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2225 pre-7.6 versions, which defined a global no_arg variable instead of putting
2226 it in the RE class. (See also #8 below.)
2227
2228 3. Remove a line of dead code, identified by coverity and reported by Nuno
2229 Lopes.
2230
2231 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2232
2233 (1) The include/exclude patterns were being applied to the whole pathnames
2234 of files, instead of just to the final components.
2235
2236 (2) If there was more than one level of directory, the subdirectories were
2237 skipped unless they satisfied the include/exclude conditions. This is
2238 inconsistent with GNU grep (and could even be seen as contrary to the
2239 pcregrep specification - which I improved to make it absolutely clear).
2240 The action now is always to scan all levels of directory, and just
2241 apply the include/exclude patterns to regular files.
2242
2243 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2244 --exclude_dir in the tests to avoid scanning .svn directories.
2245
2246 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2247 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2248 doesn't support NULs in patterns.
2249
2250 7. Added some missing "const"s to declarations of static tables in
2251 pcre_compile.c and pcre_dfa_exec.c.
2252
2253 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2254 caused by fix #2 above. (Subsequently also a second patch to fix the
2255 first patch. And a third patch - this was a messy problem.)
2256
2257 9. Applied Craig's patch to remove the use of push_back().
2258
2259 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2260 matching function regexec().
2261
2262 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2263 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2264 references. PCRE supports relative numbers with this syntax (I don't think
2265 Oniguruma does).
2266
2267 12. Previously, a group with a zero repeat such as (...){0} was completely
2268 omitted from the compiled regex. However, this means that if the group
2269 was called as a subroutine from elsewhere in the pattern, things went wrong
2270 (an internal error was given). Such groups are now left in the compiled
2271 pattern, with a new opcode that causes them to be skipped at execution
2272 time.
2273
2274 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2275 to the way PCRE behaves:
2276
2277 (a) A lone ] character is dis-allowed (Perl treats it as data).
2278
2279 (b) A back reference to an unmatched subpattern matches an empty string
2280 (Perl fails the current match path).
2281
2282 (c) A data ] in a character class must be notated as \] because if the
2283 first data character in a class is ], it defines an empty class. (In
2284 Perl it is not possible to have an empty class.) The empty class []
2285 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2286 The negative empty class [^] matches any one character, independently
2287 of the DOTALL setting.
2288
2289 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2290 non-existent subpattern following a character class starting with ']' and
2291 containing () gave an internal compiling error instead of "reference to
2292 non-existent subpattern". Fortunately, when the pattern did exist, the
2293 compiled code was correct. (When scanning forwards to check for the
2294 existence of the subpattern, it was treating the data ']' as terminating
2295 the class, so got the count wrong. When actually compiling, the reference
2296 was subsequently set up correctly.)
2297
2298 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2299 it was being rejected as not supported by pcre_dfa_exec(), even though
2300 other assertions are supported. I have made pcre_dfa_exec() support
2301 (*FAIL).
2302
2303 16. The implementation of 13c above involved the invention of a new opcode,
2304 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2305 cannot be changed at match time, I realized I could make a small
2306 improvement to matching performance by compiling OP_ALLANY instead of
2307 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2308 on the OP_ANY path.
2309
2310 17. Compiling pcretest on Windows with readline support failed without the
2311 following two fixes: (1) Make the unistd.h include conditional on
2312 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2313
2314 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2315 ncurses library to be included for pcretest when ReadLine support is
2316 requested, but also to allow for it to be overridden. This patch came from
2317 Daniel Bergström.
2318
2319 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2320 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2321 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2322 spotting this.
2323
2324
2325 Version 7.6 28-Jan-08
2326 ---------------------
2327
2328 1. A character class containing a very large number of characters with
2329 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2330 overflow.
2331
2332 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2333 HAVE_LONG_LONG is not defined.
2334
2335 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2336 bring them up to date and include new features. This patch includes:
2337
2338 - Fixed PH's badly added libz and libbz2 support.
2339 - Fixed a problem with static linking.
2340 - Added pcredemo. [But later removed - see 7 below.]
2341 - Fixed dftables problem and added an option.
2342 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2343 HAVE_LONG_LONG.
2344 - Added readline support for pcretest.
2345 - Added an listing of the option settings after cmake has run.
2346
2347 4. A user submitted a patch to Makefile that makes it easy to create
2348 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2349 Makefile.am that cause it to include this special target, without
2350 affecting anything else. Note that the same mingw target plus all
2351 the other distribution libraries and programs are now supported
2352 when configuring with CMake (see 6 below) instead of with
2353 Configure/Make.
2354
2355 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2356 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2357 exported in the Windows port". It has not yet been confirmed that the patch
2358 solves the problem, but it does no harm.
2359
2360 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2361 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2362 with CMake, and also correct the comment about stack recursion.
2363
2364 7. Remove the automatic building of pcredemo from the ./configure system and
2365 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2366 of a program that users should build themselves after PCRE is installed, so
2367 building it automatically is not really right. What is more, it gave
2368 trouble in some build environments.
2369
2370 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2371
2372
2373 Version 7.5 10-Jan-08
2374 ---------------------
2375
2376 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2377 values in parens when parsing an RE using the C++ wrapper."
2378
2379 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2380 Characters greater than 255 were excluded from the class instead of being
2381 included.
2382
2383 3. The same bug as (2) above applied to negated POSIX classes such as
2384 [:^space:].
2385
2386 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2387 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2388 I have changed it.
2389
2390 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2391 first named subpattern) and a construct such as (?&a) would reference the
2392 first named subpattern whose name started with "a" (in other words, the
2393 length check was missing). Both these problems are fixed. "Subpattern name
2394 expected" is now given for (?&) (a zero-length name), and this patch also
2395 makes it give the same error for \k'' (previously it complained that that
2396 was a reference to a non-existent subpattern).
2397
2398 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2399 this is right because (?- can be followed by option settings as well as by
2400 digits. I have, however, made the messages clearer.
2401
2402 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2403 than the number used in the conditional) now cause a compile-time error.
2404 This is actually not compatible with Perl, which accepts such patterns, but
2405 treats the conditional as always being FALSE (as PCRE used to), but it
2406 seems to me that giving a diagnostic is better.
2407
2408 8. Change "alphameric" to the more common word "alphanumeric" in comments
2409 and messages.
2410
2411 9. Fix two occurrences of "backslash" in comments that should have been
2412 "backspace".
2413
2414 10. Remove two redundant lines of code that can never be obeyed (their function
2415 was moved elsewhere).
2416
2417 11. The program that makes PCRE's Unicode character property table had a bug
2418 which caused it to generate incorrect table entries for sequences of
2419 characters that have the same character type, but are in different scripts.
2420 It amalgamated them into a single range, with the script of the first of
2421 them. In other words, some characters were in the wrong script. There were
2422 thirteen such cases, affecting characters in the following ranges:
2423
2424 U+002b0 - U+002c1
2425 U+0060c - U+0060d
2426 U+0061e - U+00612
2427 U+0064b - U+0065e
2428 U+0074d - U+0076d
2429 U+01800 - U+01805
2430 U+01d00 - U+01d77
2431 U+01d9b - U+01dbf
2432 U+0200b - U+0200f
2433 U+030fc - U+030fe
2434 U+03260 - U+0327f
2435 U+0fb46 - U+0fbb1
2436 U+10450 - U+1049d
2437
2438 12. The -o option (show only the matching part of a line) for pcregrep was not
2439 compatible with GNU grep in that, if there was more than one match in a
2440 line, it showed only the first of them. It now behaves in the same way as
2441 GNU grep.
2442
2443 13. If the -o and -v options were combined for pcregrep, it printed a blank
2444 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2445 does the same. The return code can be used to tell if there were any
2446 non-matching lines.
2447
2448 14. Added --file-offsets and --line-offsets to pcregrep.
2449
2450 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2451 infinitely looping recursion. The bug was that positive lookaheads were not
2452 being skipped when checking for a possible empty match (negative lookaheads
2453 and both kinds of lookbehind were skipped).
2454
2455 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2456 inclusion of <windows.h> to before rather than after the definition of
2457 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2458
2459 17. Specifying a possessive quantifier with a specific limit for a Unicode
2460 character property caused pcre_compile() to compile bad code, which led at
2461 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2462 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2463 caused the error; without that there was no problem.
2464
2465 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2466
2467 19. Added --enable-pcretest-libreadline.
2468
2469 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2470 RE::GlobalReplace(). As a result, the number of replacements returned was
2471 double what it should be. I removed one of the increments, but Craig sent a
2472 later patch that removed the other one (the right fix) and added unit tests
2473 that check the return values (which was not done before).
2474
2475 21. Several CMake things:
2476
2477 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2478 the names libpcre and libpcreposix, not just pcre and pcreposix.
2479
2480 (2) The above change means that pcretest and pcregrep are now correctly
2481 linked with the newly-built libraries, not previously installed ones.
2482
2483 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2484
2485 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2486 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2487 UTF-8 newline character). The key issue is that the pattern starts .*;
2488 this means that the match must be either at the beginning, or after a
2489 newline. The bug was in the code for advancing after a failed match and
2490 checking that the new position followed a newline. It was not taking
2491 account of UTF-8 characters correctly.
2492
2493 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2494 character classes. PCRE was not treating the sequence [:...:] as a
2495 character class unless the ... were all letters. Perl, however, seems to
2496 allow any characters between [: and :], though of course it rejects as
2497 unknown any "names" that contain non-letters, because all the known class
2498 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2499 for example, whereas PCRE did not - it did not recognize a POSIX character
2500 class. This seemed a bit dangerous, so the code has been changed to be
2501 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2502 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2503 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2504 Perl does, and where it didn't before.
2505
2506 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2507 Windows environments %n is disabled by default.
2508
2509
2510 Version 7.4 21-Sep-07
2511 ---------------------
2512
2513 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2514 means that a class such as [\s] counted as "explicit reference to CR or
2515 LF". That isn't really right - the whole point of the change was to try to
2516 help when there was an actual mention of one of the two characters. So now
2517 the change happens only if \r or \n (or a literal CR or LF) character is
2518 encountered.
2519
2520 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2521 of both had grown to the point where there were only 3 bits left.
2522 Fortunately, there was spare space in the data structure, and so I have
2523 moved the internal flags into a new 16-bit field to free up more option
2524 bits.
2525
2526 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2527 but did not set the internal JCHANGED flag - either of these is enough to
2528 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2529 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2530 start sets both bits.
2531
2532 4. Added options (at build time, compile time, exec time) to change \R from
2533 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2534
2535 5. doc/pcresyntax.html was missing from the distribution.
2536
2537 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2538 compatibility, even though it is no longer used.
2539
2540 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2541 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2542 windows.h file is present (where different names are used). [This was
2543 reversed later after testing - see 16 below.]
2544
2545 8. Changed all #include <config.h> to #include "config.h". There were also
2546 some further <pcre.h> cases that I changed to "pcre.h".
2547
2548 9. When pcregrep was used with the --colour option, it missed the line ending
2549 sequence off the lines that it output.
2550
2551 10. It was pointed out to me that arrays of string pointers cause lots of
2552 relocations when a shared library is dynamically loaded. A technique of
2553 using a single long string with a table of offsets can drastically reduce
2554 these. I have refactored PCRE in four places to do this. The result is
2555 dramatic:
2556
2557 Originally: 290
2558 After changing UCP table: 187
2559 After changing error message table: 43
2560 After changing table of "verbs" 36
2561 After changing table of Posix names 22
2562
2563 Thanks to the folks working on Gregex for glib for this insight.
2564
2565 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2566 unicode-properties was also set.
2567
2568 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2569
2570 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2571 checked only for CRLF.
2572
2573 14. Added casts to pcretest.c to avoid compiler warnings.
2574
2575 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2576
2577 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2578 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2579 entirely. This removes changes made in 7 above.
2580
2581 17. The CMake files have been updated, and there is now more information about
2582 building with CMake in the NON-UNIX-USE document.
2583
2584
2585 Version 7.3 28-Aug-07
2586 ---------------------
2587
2588 1. In the rejigging of the build system that eventually resulted in 7.1, the
2589 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2590 brackets there is not right, since it causes compilers to look for an
2591 installed pcre.h, not the version that is in the source that is being
2592 compiled (which of course may be different). I have changed it back to:
2593
2594 #include "pcre.h"
2595
2596 I have a vague recollection that the change was concerned with compiling in
2597 different directories, but in the new build system, that is taken care of
2598 by the VPATH setting the Makefile.
2599
2600 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2601 when the subject happened to end in the byte 0x85 (e.g. if the last
2602 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2603 characters but of course it shouldn't be taken as a newline when it is part
2604 of another character. The bug was that, for an unlimited repeat of . in
2605 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2606 characters when looking for a newline.
2607
2608 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2609
2610 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2611 in debug output.
2612
2613 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2614 long printing in the pcrecpp unittest when running under MinGW.
2615
2616 6. ESC_K was left out of the EBCDIC table.
2617
2618 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2619 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2620 limit also applies to "virtual nesting" when a pattern is recursive, and in
2621 this case 1000 isn't so big. I have been able to remove this limit at the
2622 expense of backing off one optimization in certain circumstances. Normally,
2623 when pcre_exec() would call its internal match() function recursively and
2624 immediately return the result unconditionally, it uses a "tail recursion"
2625 feature to save stack. However, when a subpattern that can match an empty
2626 string has an unlimited repetition quantifier, it no longer makes this
2627 optimization. That gives it a stack frame in which to save the data for
2628 checking that an empty string has been matched. Previously this was taken
2629 from the 1000-entry workspace that had been reserved. So now there is no
2630 explicit limit, but more stack is used.
2631
2632 8. Applied Daniel's patches to solve problems with the import/export magic
2633 syntax that is required for Windows, and which was going wrong for the
2634 pcreposix and pcrecpp parts of the library. These were overlooked when this
2635 problem was solved for the main library.
2636
2637 9. There were some crude static tests to avoid integer overflow when computing
2638 the size of patterns that contain repeated groups with explicit upper
2639 limits. As the maximum quantifier is 65535, the maximum group length was
2640 set at 30,000 so that the product of these two numbers did not overflow a
2641 32-bit integer. However, it turns out that people want to use groups that
2642 are longer than 30,000 bytes (though not repeat them that many times).
2643 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2644 made it possible to implement the integer overflow checks in a much more
2645 dynamic way, which I have now done. The artificial limitation on group
2646 length has been removed - we now have only the limit on the total length of
2647 the compiled pattern, which depends on the LINK_SIZE setting.
2648
2649 10. Fixed a bug in the documentation for get/copy named substring when
2650 duplicate names are permitted. If none of the named substrings are set, the
2651 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2652 empty string.
2653
2654 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2655 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2656 because the ] is interpreted as the first data character and the
2657 terminating ] is not found. PCRE has been made compatible with Perl in this
2658 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2659 cause memory overwriting.
2660
2661 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2662 string has been matched (to stop an infinite loop). It was not recognizing
2663 a conditional subpattern that could match an empty string if that
2664 subpattern was within another subpattern. For example, it looped when
2665 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2666 condition was not nested. This bug has been fixed.
2667
2668 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2669 past the start of the subject in the presence of bytes with the top bit
2670 set, for example "\x8aBCD".
2671
2672 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2673 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2674
2675 14. Optimized (?!) to (*FAIL).
2676
2677 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2678 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2679 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2680 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2681 does: it's just the validity check that is more restrictive.
2682
2683 16. Inserted checks for integer overflows during escape sequence (backslash)
2684 processing, and also fixed erroneous offset values for syntax errors during
2685 backslash processing.
2686
2687 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2688 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2689
2690 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2691 caused an overrun.
2692
2693 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2694 something other than just ASCII characters) inside a group that had an
2695 unlimited repeat caused a loop at compile time (while checking to see
2696 whether the group could match an empty string).
2697
2698 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2699 [\P{Any}] did so. (Error in the code for printing property names.)
2700
2701 21. An orphan \E inside a character class could cause a crash.
2702
2703 22. A repeated capturing bracket such as (A)? could cause a wild memory
2704 reference during compilation.
2705
2706 23. There are several functions in pcre_compile() that scan along a compiled
2707 expression for various reasons (e.g. to see if it's fixed length for look
2708 behind). There were bugs in these functions when a repeated \p or \P was
2709 present in the pattern. These operators have additional parameters compared
2710 with \d, etc, and these were not being taken into account when moving along
2711 the compiled data. Specifically:
2712
2713 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2714 length.
2715
2716 (b) An item such as \pL+ within a repeated group could cause crashes or
2717 loops.
2718
2719 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2720 "reference to non-existent subpattern" error.
2721
2722 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2723
2724 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2725 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2726
2727 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2728
2729 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2730 character were causing crashes (broken optimization).
2731
2732 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2733 \p or \P) caused a compile-time loop.
2734
2735 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2736 break. For example, the unstudied pattern [\r\n]A does not match the string
2737 "\r\nA" because change 7.0/46 below moves the current point on by two
2738 characters after failing to match at the start. However, the pattern \nA
2739 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2740 the same is true. There doesn't seem any very clean way out of this, but
2741 what I have chosen to do makes the common cases work: PCRE now takes note
2742 of whether there can be an explicit match for \r or \n anywhere in the
2743 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2744 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2745 pattern has explicit CR or LF references.
2746
2747 29. Added (*CR) etc for changing newline setting at start of pattern.
2748
2749
2750 Version 7.2 19-Jun-07
2751 ---------------------
2752
2753 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2754 which is apparently normally available under Windows.
2755
2756 2. Re-jig the pcregrep tests with different newline settings in an attempt
2757 to make them independent of the local environment's newline setting.
2758
2759 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2760
2761 4. Some of the "internals" tests were previously cut out when the link size
2762 was not 2, because the output contained actual offsets. The recent new
2763 "Z" feature of pcretest means that these can be cut out, making the tests
2764 usable with all link sizes.
2765
2766 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2767 stack recursion. This gives a massive performance boost under BSD, but just
2768 a small improvement under Linux. However, it saves one field in the frame
2769 in all cases.
2770
2771 6. Added more features from the forthcoming Perl 5.10:
2772
2773 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2774 recursion call. It refers to the nth most recently opened parentheses.
2775
2776 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2777 to be opened parentheses.
2778
2779 (c) Conditions that refer to capturing parentheses can be specified
2780 relatively, for example, (?(-2)... or (?(+3)...
2781
2782 (d) \K resets the start of the current match so that everything before
2783 is not part of it.
2784
2785 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2786
2787 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2788 reference syntax.
2789
2790 (g) (?| introduces a group in which the numbering of parentheses in each
2791 alternative starts with the same number.
2792
2793 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2794
2795 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2796 PCRE_INFO_JCHANGED.
2797
2798 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2799 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2800 for detecting groups that can match an empty string.
2801
2802 9. A pattern with a very large number of alternatives (more than several
2803 hundred) was running out of internal workspace during the pre-compile
2804 phase, where pcre_compile() figures out how much memory will be needed. A
2805 bit of new cunning has reduced the workspace needed for groups with
2806 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2807 workspace instead of running out of the 4096 that are available.
2808
2809 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2810
2811 11. Applied patch from Google to remove an optimization that didn't quite work.
2812 The report of the bug said:
2813
2814 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2815 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2816 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2817
2818 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2819 it matched the wrong number of bytes.
2820
2821
2822 Version 7.1 24-Apr-07
2823 ---------------------
2824
2825 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2826 that is more "standard", making use of automake and other Autotools. There
2827 is some re-arrangement of the files and adjustment of comments consequent
2828 on this.
2829
2830 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2831 for recursive directory scanning broke on some systems because the files
2832 are not scanned in any specific order and on different systems the order
2833 was different. A call to "sort" has been inserted into RunGrepTest for the
2834 approprate test as a short-term fix. In the longer term there may be an
2835 alternative.
2836
2837 3. I had an email from Eric Raymond about problems translating some of PCRE's
2838 man pages to HTML (despite the fact that I distribute HTML pages, some
2839 people do their own conversions for various reasons). The problems
2840 concerned the use of low-level troff macros .br and .in. I have therefore
2841 removed all such uses from the man pages (some were redundant, some could
2842 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2843 HTML has been updated to handle .nf/.fi and to complain if it encounters
2844 .br or .in.
2845
2846 4. Updated comments in configure.ac that get placed in config.h.in and also
2847 arranged for config.h to be included in the distribution, with the name
2848 config.h.generic, for the benefit of those who have to compile without
2849 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2850
2851 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2852 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2853 makevp.bat for the current PCRE, using the additional files
2854 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2855
2856 6. A Windows user reported a minor discrepancy with test 2, which turned out
2857 to be caused by a trailing space on an input line that had got lost in his
2858 copy. The trailing space was an accident, so I've just removed it.
2859
2860 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2861 that is needed.
2862
2863 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2864 as "const" (a) because they are and (b) because it helps the PHP
2865 maintainers who have recently made a script to detect big data structures
2866 in the php code that should be moved to the .rodata section. I remembered
2867 to update Builducptable as well, so it won't revert if ucptable.h is ever
2868 re-created.
2869
2870 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2871 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2872 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2873 support is not required. This saves 1.5-2K of code, which is important in
2874 some applications.
2875
2876 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2877 so as not to refer to the tables, even though these functions will never be
2878 called when UTF-8 support is disabled. Otherwise there are problems with a
2879 shared library.
2880
2881 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2882
2883 (a) It was defining its arguments as char * instead of void *.
2884
2885 (b) It was assuming that all moves were upwards in memory; this was true
2886 a long time ago when I wrote it, but is no longer the case.
2887
2888 The emulated memove() is provided for those environments that have neither
2889 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2890 is clearly not the case, as these two bugs were recently reported.
2891
2892 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2893 and Detrail to create the HTML documentation, the .txt form of the man
2894 pages, and it removes trailing spaces from listed files. It also creates
2895 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2896 case, it wraps all the #defines with #ifndefs. This script should be run
2897 before "make dist".
2898
2899 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2900 with Unicode property support.
2901
2902 (a) For a maximizing quantifier, if the two different cases of the
2903 character were of different lengths in their UTF-8 codings (there are
2904 some cases like this - I found 11), and the matching function had to
2905 back up over a mixture of the two cases, it incorrectly assumed they
2906 were both the same length.
2907
2908 (b) When PCRE was configured to use the heap rather than the stack for
2909 recursion during matching, it was not correctly preserving the data for
2910 the other case of a UTF-8 character when checking ahead for a match
2911 while processing a minimizing repeat. If the check also involved
2912 matching a wide character, but failed, corruption could cause an
2913 erroneous result when trying to check for a repeat of the original
2914 character.
2915
2916 13. Some tidying changes to the testing mechanism:
2917
2918 (a) The RunTest script now detects the internal link size and whether there
2919 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2920 values substituted by "configure". (The RunGrepTest script already did
2921 this for UTF-8.) The configure.ac script no longer substitutes the
2922 relevant variables.
2923
2924 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2925 with length and offset values. This means that the output is different
2926 for different internal link sizes. Test 2 is skipped for link sizes
2927 other than 2 because of this, bypassing the problem. Unfortunately,
2928 there was also a test in test 3 (the locale tests) that used /B and
2929 failed for link sizes other than 2. Rather than cut the whole test out,
2930 I have added a new /Z option to pcretest that replaces the length and
2931 offset values with spaces. This is now used to make test 3 independent
2932 of link size. (Test 2 will be tidied up later.)
2933
2934 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2935 segmentation fault instead of returning the appropriate error message.
2936
2937 15. In multiline mode when the newline sequence was set to "any", the pattern
2938 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2939 This doesn't seem right; it now treats the CRLF combination as the line
2940 ending, and so does not match in that case. It's only a pattern such as ^$
2941 that would hit this one: something like ^ABC$ would have failed after \r
2942 and then tried again after \r\n.
2943
2944 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2945 in an attempt to make files that differ only in their line terminators
2946 compare equal. This works on Linux.
2947
2948 17. Under certain error circumstances pcregrep might try to free random memory
2949 as it exited. This is now fixed, thanks to valgrind.
2950
2951 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2952 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2953 was because its rules for how to advance for /g after matching an empty
2954 string at the end of a line did not allow for this case. They now check for
2955 it specially.
2956
2957 20. pcretest is supposed to handle patterns and data of any length, by
2958 extending its buffers when necessary. It was getting this wrong when the
2959 buffer for a data line had to be extended.
2960
2961 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2962 CRLF as a newline sequence.
2963
2964 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2965 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2966 I have nevertheless tidied it up.
2967
2968 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2969
2970 24. Added a man page for pcre-config.
2971
2972
2973 Version 7.0 19-Dec-06
2974 ---------------------
2975
2976 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2977 moving to gcc 4.1.1.
2978
2979 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2980 sys/time.h, which is documented as needed for this function. It doesn't
2981 seem to matter on Linux, but it showed up on some releases of OS X.
2982
2983 3. It seems that there are systems where bytes whose values are greater than
2984 127 match isprint() in the "C" locale. The "C" locale should be the
2985 default when a C program starts up. In most systems, only ASCII printing
2986 characters match isprint(). This difference caused the output from pcretest
2987 to vary, making some of the tests fail. I have changed pcretest so that:
2988
2989 (a) When it is outputting text in the compiled version of a pattern, bytes
2990 other than 32-126 are always shown as hex escapes.
2991
2992 (b) When it is outputting text that is a matched part of a subject string,
2993 it does the same, unless a different locale has been set for the match
2994 (using the /L modifier). In this case, it uses isprint() to decide.
2995
2996 4. Fixed a major bug that caused incorrect computation of the amount of memory
2997 required for a compiled pattern when options that changed within the
2998 pattern affected the logic of the preliminary scan that determines the
2999 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3000 that the computed length was too small. The symptoms of this bug were
3001 either the PCRE error "internal error: code overflow" from pcre_compile(),
3002 or a glibc crash with a message such as "pcretest: free(): invalid next
3003 size (fast)". Examples of patterns that provoked this bug (shown in
3004 pcretest format) are:
3005
3006 /(?-x: )/x
3007 /(?x)(?-x: \s*#\s*)/
3008 /((?i)[\x{c0}])/8
3009 /(?i:[\x{c0}])/8
3010
3011 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3012 is now done differently.
3013
3014 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3015 wrapper classes; (b) implement a new function in the C++ scanner that is
3016 more efficient than the old way of doing things because it avoids levels of
3017 recursion in the regex matching; (c) add a paragraph to the documentation
3018 for the FullMatch() function.
3019
3020 6. The escape sequence \n was being treated as whatever was defined as
3021 "newline". Not only was this contrary to the documentation, which states
3022 that \n is character 10 (hex 0A), but it also went horribly wrong when
3023 "newline" was defined as CRLF. This has been fixed.
3024
3025 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3026 was being set to -1 for the "end of line" case (supposedly a value that no
3027 character can have). Though this value is never used (the check for end of
3028 line is "zero bytes in current character"), it caused compiler complaints.
3029 I've changed it to 0xffffffff.
3030
3031 8. In pcre_version.c, the version string was being built by a sequence of
3032 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3033 string (as it is for production releases) called a macro with an empty
3034 argument. The C standard says the result of this is undefined. The gcc
3035 compiler treats it as an empty string (which was what was wanted) but it is
3036 reported that Visual C gives an error. The source has been hacked around to
3037 avoid this problem.
3038
3039 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3040 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3041 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3042 of them did).
3043
3044 10. Originally, pcretest opened its input and output without "b"; then I was
3045 told that "b" was needed in some environments, so it was added for release
3046 5.0 to both the input and output. (It makes no difference on Unix-like
3047 systems.) Later I was told that it is wrong for the input on Windows. I've
3048 now abstracted the modes into two macros, to make it easier to fiddle with
3049 them, and removed "b" from the input mode under Windows.
3050
3051 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3052
3053 12. Added -help and --help to pcretest as an official way of being reminded
3054 of the options.
3055
3056 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3057 and pcrecpp.cc because they annoy compilers at high warning levels.
3058
3059 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3060
3061 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3062 scripts are not C programs :-) and which was not noticed because it works
3063 on Linux.
3064
3065 16. pcretest is supposed to handle any length of pattern and data line (as one
3066 line or as a continued sequence of lines) by extending its input buffer if
3067 necessary. This feature was broken for very long pattern lines, leading to
3068 a string of junk being passed to pcre_compile() if the pattern was longer
3069 than about 50K.
3070
3071 17. I have done a major re-factoring of the way pcre_compile() computes the
3072 amount of memory needed for a compiled pattern. Previously, there was code
3073 that made a preliminary scan of the pattern in order to do this. That was
3074 OK when PCRE was new, but as the facilities have expanded, it has become
3075 harder and harder to keep it in step with the real compile phase, and there
3076 have been a number of bugs (see for example, 4 above). I have now found a
3077 cunning way of running the real compile function in a "fake" mode that
3078 enables it to compute how much memory it would need, while actually only
3079 ever using a few hundred bytes of working memory and without too many
3080 tests of the mode. This should make future maintenance and development
3081 easier. A side effect of this work is that the limit of 200 on the nesting
3082 depth of parentheses has been removed (though this was never a serious
3083 limitation, I suspect). However, there is a downside: pcre_compile() now
3084 runs more slowly than before (30% or more, depending on the pattern). I
3085 hope this isn't a big issue. There is no effect on runtime performance.
3086
3087 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3088 newline (only possible for the last line of a file) and it was a
3089 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3090
3091 19. Added additional timing features to pcretest. (1) The -tm option now times
3092 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3093 separate command line item, by a number that specifies the number of
3094 repeats to use when timing. The default is 50000; this gives better
3095 precision, but takes uncomfortably long for very large patterns.
3096
3097 20. Extended pcre_study() to be more clever in cases where a branch of a
3098 subpattern has no definite first character. For example, (a*|b*)[cd] would
3099 previously give no result from pcre_study(). Now it recognizes that the
3100 first character must be a, b, c, or d.
3101
3102 21. There was an incorrect error "recursive call could loop indefinitely" if
3103 a subpattern (or the entire pattern) that was being tested for matching an
3104 empty string contained only one non-empty item after a nested subpattern.
3105 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3106 incorrectly, because the \d was being skipped in the check.
3107
3108 22. The pcretest program now has a new pattern option /B and a command line
3109 option -b, which is equivalent to adding /B to every pattern. This causes
3110 it to show the compiled bytecode, without the additional information that
3111 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3112 is the same as /B/I).
3113
3114 23. A new optimization is now able automatically to treat some sequences such
3115 as a*b as a*+b. More specifically, if something simple (such as a character
3116 or a simple class like \d) has an unlimited quantifier, and is followed by
3117 something that cannot possibly match the quantified thing, the quantifier
3118 is automatically "possessified".
3119
3120 24. A recursive reference to a subpattern whose number was greater than 39
3121 went wrong under certain circumstances in UTF-8 mode. This bug could also
3122 have affected the operation of pcre_study().
3123
3124 25. Realized that a little bit of performance could be had by replacing
3125 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3126
3127 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3128
3129 27. Possessive quantifiers such as a++ were previously implemented by turning
3130 them into atomic groups such as ($>a+). Now they have their own opcodes,
3131 which improves performance. This includes the automatically created ones
3132 from 23 above.
3133
3134 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3135 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3136 the first matched character to be a colon. This applied both to named and
3137 numbered groups.
3138
3139 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3140
3141 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3142 building PCRE on Windows easier, so I have included it in the distribution.
3143
3144 31. There is now a check in pcretest against a ridiculously large number being
3145 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3146 loop, the loop is abandoned.
3147
3148 32. Forward references to subpatterns in conditions such as (?(2)...) where
3149 subpattern 2 is defined later cause pcre_compile() to search forwards in
3150 the pattern for the relevant set of parentheses. This search went wrong
3151 when there were unescaped parentheses in a character class, parentheses
3152 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3153
3154 33. "Subroutine" calls and backreferences were previously restricted to
3155 referencing subpatterns earlier in the regex. This restriction has now
3156 been removed.
3157
3158 34. Added a number of extra features that are going to be in Perl 5.10. On the
3159 whole, these are just syntactic alternatives for features that PCRE had
3160 previously implemented using the Python syntax or my own invention. The
3161 other formats are all retained for compatibility.
3162
3163 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3164 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3165 also .NET compatible.
3166
3167 (b) A recursion or subroutine call to a named group can now be defined as
3168 (?&name) as well as (?P>name).
3169
3170 (c) A backreference to a named group can now be defined as \k<name> or
3171 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3172 5.10, are also .NET compatible.
3173
3174 (d) A conditional reference to a named group can now use the syntax
3175 (?(<name>) or (?('name') as well as (?(name).
3176
3177 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3178 groups (named and numbered) that are never evaluated inline, but can be
3179 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3180 is always false. There may be only one alternative in such a group.
3181
3182 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3183 as the simple (?(R). The condition is true only if the most recent
3184 recursion is that of the given number or name. It does not search out
3185 through the entire recursion stack.
3186
3187 (g) The escape \gN or \g{N} has been added, where N is a positive or
3188 negative number, specifying an absolute or relative reference.
3189
3190 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3191 some "unreachable code" warnings.
3192
3193 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3194 things, this adds five new scripts.
3195
3196 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3197 There were also incompatibilities regarding the handling of \Q..\E inside
3198 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3199 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3200
3201 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3202 matches an empty string, and forcibly breaks the loop. There were bugs in
3203 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3204 against aaaa the result was just "a" rather than "aaaa", for example. Two
3205 separate and independent bugs (that affected different cases) have been
3206 fixed.
3207
3208 39. Refactored the code to abolish the use of different opcodes for small
3209 capturing bracket numbers. This is a tidy that I avoided doing when I
3210 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3211 The new approach is not only tidier, it makes it possible to reduce the
3212 memory needed to fix the previous bug (38).
3213
3214 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3215 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3216 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3217 mode.
3218
3219 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3220 report.
3221
3222 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3223 copy construction and assignment in the C++ wrapper.
3224
3225 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3226 couple of bugs that could have given wrong results in the "--newline=crlf"
3227 case.
3228
3229 44. Added a number of casts and did some reorganization of signed/unsigned int
3230 variables following suggestions from Dair Grant. Also renamed the variable
3231 "this" as "item" because it is a C++ keyword.
3232
3233 45. Arranged for dftables to add
3234
3235 #include "pcre_internal.h"
3236
3237 to pcre_chartables.c because without it, gcc 4.x may remove the array
3238 definition from the final binary if PCRE is built into a static library and
3239 dead code stripping is activated.
3240
3241 46. For an unanchored pattern, if a match attempt fails at the start of a
3242 newline sequence, and the newline setting is CRLF or ANY, and the next two
3243 characters are CRLF, advance by two characters instead of one.
3244
3245
3246 Version 6.7 04-Jul-06
3247 ---------------------
3248
3249 1. In order to handle tests when input lines are enormously long, pcretest has
3250 been re-factored so that it automatically extends its buffers when
3251 necessary. The code is crude, but this _is_ just a test program. The
3252 default size has been increased from 32K to 50K.
3253
3254 2. The code in pcre_study() was using the value of the re argument before
3255 testing it for NULL. (Of course, in any sensible call of the function, it
3256 won't be NULL.)
3257
3258 3. The memmove() emulation function in pcre_internal.h, which is used on
3259 systems that lack both memmove() and bcopy() - that is, hardly ever -
3260 was missing a "static" storage class specifier.
3261
3262 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3263 containing an extended class (one that cannot be represented by a bitmap
3264 because it contains high-valued characters or Unicode property items, e.g.
3265 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3266 pattern, but PCRE should not loop if you do not (it no longer does).
3267 [Detail: two cases were found: (a) a repeated subpattern containing an
3268 extended class; (b) a recursive reference to a subpattern that followed a
3269 previous extended class. It wasn't skipping over the extended class
3270 correctly when UTF-8 mode was not set.]
3271
3272 5. A negated single-character class was not being recognized as fixed-length
3273 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3274 compile error "lookbehind assertion is not fixed length".
3275
3276 6. The RunPerlTest auxiliary script was showing an unexpected difference
3277 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3278 write a Perl script that can interpret lines of an input file either as
3279 byte characters or as UTF-8, which is what "perltest" was being required to
3280 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3281 can't do is switch easily at run time between having the "use utf8;" pragma
3282 or not. In the end, I fudged it by using the RunPerlTest script to insert
3283 "use utf8;" explicitly for the UTF-8 tests.
3284
3285 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3286 the end of the subject string, contrary to the documentation and to what
3287 Perl does. This was true of both matching functions. Now it matches only at
3288 the start of the subject and immediately after *internal* newlines.
3289
3290 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3291 a pointer to an int instead of a pointer to an unsigned long int. This
3292 caused problems on 64-bit systems.
3293
3294 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3295 instance of the 'standard' template library not being so standard".
3296
3297 10. There was no check on the number of named subpatterns nor the maximum
3298 length of a subpattern name. The product of these values is used to compute
3299 the size of the memory block for a compiled pattern. By supplying a very
3300 long subpattern name and a large number of named subpatterns, the size
3301 computation could be caused to overflow. This is now prevented by limiting
3302 the length of names to 32 characters, and the number of named subpatterns
3303 to 10,000.
3304
3305 11. Subpatterns that are repeated with specific counts have to be replicated in
3306 the compiled pattern. The size of memory for this was computed from the
3307 length of the subpattern and the repeat count. The latter is limited to
3308 65535, but there was no limit on the former, meaning that integer overflow
3309 could in principle occur. The compiled length of a repeated subpattern is
3310 now limited to 30,000 bytes in order to prevent this.
3311
3312 12. Added the optional facility to have named substrings with the same name.
3313
3314 13. Added the ability to use a named substring as a condition, using the
3315 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3316 are numbers (not recommended). Forward references are permitted.
3317
3318 14. Added forward references in named backreferences (if you see what I mean).
3319
3320 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3321 pattern could run off the end of the subject. For example, the pattern
3322 "(?s)(.{1,5})"8 did this with the subject "ab".
3323
3324 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3325 PCRE_CASELESS was set when matching characters that were quantified with ?
3326 or *.
3327
3328 17. A character class other than a single negated character that had a minimum
3329 but no maximum quantifier - for example [ab]{6,} - was not handled
3330 correctly by pce_dfa_exec(). It would match only one character.
3331
3332 18. A valid (though odd) pattern that looked like a POSIX character
3333 class but used an invalid character after [ (for example [[,abc,]]) caused
3334 pcre_compile() to give the error "Failed: internal error: code overflow" or
3335 in some cases to crash with a glibc free() error. This could even happen if
3336 the pattern terminated after [[ but there just happened to be a sequence of
3337 letters, a binary zero, and a closing ] in the memory that followed.
3338
3339 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3340 over the years. Originally (before any Unicode support), just the bottom 8
3341 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3342 output from "man perlunicode" includes this:
3343
3344 The regular expression compiler produces polymorphic opcodes. That
3345 is, the pattern adapts to the data and automatically switches to
3346 the Unicode character scheme when presented with Unicode data--or
3347 instead uses a traditional byte scheme when presented with byte
3348 data.
3349
3350 Sadly, a wide octal escape does not cause a switch, and in a string with
3351 no other multibyte characters, these octal escapes are treated as before.
3352 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3353 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3354 Unicode string.
3355
3356 I have not perpetrated such confusion in PCRE. Up till now, it took just
3357 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3358 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3359 translate to the appropriate multibyte character.
3360
3361 29. Applied some refactoring to reduce the number of warnings from Microsoft
3362 and Borland compilers. This has included removing the fudge introduced
3363 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3364 a warning about an unused variable.
3365
3366 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3367 characters since release 4.0, because Perl (from release 5.004) does not.
3368 [Or at least, is documented not to: some releases seem to be in conflict
3369 with the documentation.] However, when a pattern was studied with
3370 pcre_study() and all its branches started with \s, PCRE still included VT
3371 as a possible starting character. Of course, this did no harm; it just
3372 caused an unnecessary match attempt.
3373
3374 22. Removed a now-redundant internal flag bit that recorded the fact that case
3375 dependency changed within the pattern. This was once needed for "required
3376 byte" processing, but is no longer used. This recovers a now-scarce options
3377 bit. Also moved the least significant internal flag bit to the most-
3378 significant bit of the word, which was not previously used (hangover from
3379 the days when it was an int rather than a uint) to free up another bit for
3380 the future.
3381
3382 23. Added support for CRLF line endings as well as CR and LF. As well as the
3383 default being selectable at build time, it can now be changed at runtime
3384 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3385 specify that it is scanning data with non-default line endings.
3386
3387 24. Changed the definition of CXXLINK to make it agree with the definition of
3388 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3389
3390 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3391 recursions. This makes a big different to stack usage for some patterns.
3392
3393 26. If a subpattern containing a named recursion or subroutine reference such
3394 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3395 the space required for the compiled pattern went wrong and gave too small a
3396 value. Depending on the environment, this could lead to "Failed: internal
3397 error: code overflow at offset 49" or "glibc detected double free or
3398 corruption" errors.
3399
3400 27. Applied patches from Google (a) to support the new newline modes and (b) to
3401 advance over multibyte UTF-8 characters in GlobalReplace.
3402
3403 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3404 difference for some implementation of PCRE in some Windows version.
3405
3406 29. Added some extra testing facilities to pcretest:
3407
3408 \q<number> in a data line sets the "match limit" value
3409 \Q<number> in a data line sets the "match recursion limt" value
3410 -S <number> sets the stack size, where <number> is in megabytes
3411
3412 The -S option isn't available for Windows.
3413
3414
3415 Version 6.6 06-Feb-06
3416 ---------------------
3417
3418 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3419 in pcreposix.h. I have copied the definition from pcre.h.
3420
3421 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3422 because pcre.h is no longer a built file.
3423
3424 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3425 not normally included in the compiled code.
3426
3427
3428 Version 6.5 01-Feb-06
3429 ---------------------
3430
3431 1. When using the partial match feature with pcre_dfa_exec(), it was not
3432 anchoring the second and subsequent partial matches at the new starting
3433 point. This could lead to incorrect results. For example, with the pattern
3434 /1234/, partially matching against "123" and then "a4" gave a match.
3435
3436 2. Changes to pcregrep:
3437
3438 (a) All non-match returns from pcre_exec() were being treated as failures
3439 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3440 error message is output. Some extra information is given for the
3441 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3442 probably the only errors that are likely to be caused by users (by
3443 specifying a regex that has nested indefinite repeats, for instance).
3444 If there are more than 20 of these errors, pcregrep is abandoned.
3445
3446 (b) A binary zero was treated as data while matching, but terminated the
3447 output line if it was written out. This has been fixed: binary zeroes
3448 are now no different to any other data bytes.
3449
3450 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3451 used to set a locale for matching. The --locale=xxxx long option has
3452 been added (no short equivalent) to specify a locale explicitly on the
3453 pcregrep command, overriding the environment variables.
3454
3455 (d) When -B was used with -n, some line numbers in the output were one less
3456 than they should have been.
3457
3458 (e) Added the -o (--only-matching) option.
3459
3460 (f) If -A or -C was used with -c (count only), some lines of context were
3461 accidentally printed for the final match.
3462
3463 (g) Added the -H (--with-filename) option.
3464
3465 (h) The combination of options -rh failed to suppress file names for files
3466 that were found from directory arguments.
3467
3468 (i) Added the -D (--devices) and -d (--directories) options.
3469
3470 (j) Added the -F (--fixed-strings) option.
3471
3472 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3473
3474 (l) Added the --colo(u)r option.
3475
3476 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3477 is not present by default.
3478
3479 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3480 items such as (?R) or (?1), when the recursion could match a number of
3481 alternatives. If it matched one of the alternatives, but subsequently,
3482 outside the recursion, there was a failure, the code tried to back up into
3483 the recursion. However, because of the way PCRE is implemented, this is not
3484 possible, and the result was an incorrect result from the match.
3485
3486 In order to prevent this happening, the specification of recursion has
3487 been changed so that all such subpatterns are automatically treated as
3488 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3489
3490 4. I had overlooked the fact that, in some locales, there are characters for
3491 which isalpha() is true but neither isupper() nor islower() are true. In
3492 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3493 and ordfeminine) are like this. This affected the treatment of \w and \W
3494 when they appeared in character classes, but not when they appeared outside
3495 a character class. The bit map for "word" characters is now created
3496 separately from the results of isalnum() instead of just taking it from the
3497 upper, lower, and digit maps. (Plus the underscore character, of course.)
3498
3499 5. The above bug also affected the handling of POSIX character classes such as
3500 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3501 permanent tables. Instead, the bit maps for such a class were previously
3502 created as the appropriate unions of the upper, lower, and digit bitmaps.
3503 Now they are created by subtraction from the [[:word:]] class, which has
3504 its own bitmap.
3505
3506 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3507 It is created by subtracting the vertical space characters (\x09, \x0a,
3508 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3509 subtraction was done in the overall bitmap for a character class, meaning
3510 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3511 be recognized. This bug has been fixed.
3512
3513 7. Patches from the folks at Google:
3514
3515 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3516 real life, but is still worth protecting against".
3517
3518 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3519 regular expressions".
3520
3521 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3522 have it.
3523
3524 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3525 "configure" and the latter not, in order to fix a problem somebody had
3526 with compiling the Arg class on HP-UX.
3527
3528 (e) Improve the error-handling of the C++ wrapper a little bit.
3529
3530 (f) New tests for checking recursion limiting.
3531
3532 8. The pcre_memmove() function, which is used only if the environment does not
3533 have a standard memmove() function (and is therefore rarely compiled),
3534 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3535 returning a result (though PCRE never actually uses the result).
3536
3537 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3538 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3539 returned instead of calling malloc() with an overflowing number that would
3540 most likely cause subsequent chaos.
3541
3542 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3543
3544 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3545 with this option is matched, the nmatch and pmatch options of regexec() are
3546 ignored.
3547
3548 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3549 provided in case anyone wants to the the POSIX interface with UTF-8
3550 strings.
3551
3552 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3553 C++ linking (needed for some HP-UX environments).
3554
3555 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3556 (unused parameter) and in the pcre_printint() function (omitted "default"
3557 switch label when the default is to do nothing).
3558
3559 15. Added some code to make it possible, when PCRE is compiled as a C++
3560 library, to replace subject pointers for pcre_exec() with a smart pointer
3561 class, thus making it possible to process discontinuous strings.
3562
3563 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3564 much the same function. They were added by different people who were trying
3565 to make PCRE easy to compile on non-Unix systems. It has been suggested
3566 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3567 for compiling on Windows systems. I have therefore replaced it with
3568 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3569 defaults to "extern" for C or "extern C" for C++, which works fine on
3570 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3571 SCOPE with something explicit in config.h. In addition:
3572
3573 (a) pcreposix.h still had just "extern" instead of either of these macros;
3574 I have replaced it with PCRE_DATA_SCOPE.
3575
3576 (b) Functions such as _pcre_xclass(), which are internal to the library,
3577 but external in the C sense, all had PCRE_EXPORT in their definitions.
3578 This is apparently wrong for the Windows case, so I have removed it.
3579 (It makes no difference on Unix-like systems.)
3580
3581 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3582 of recursive calls to match(). This is different to MATCH_LIMIT because
3583 that limits the total number of calls to match(), not all of which increase
3584 the depth of recursion. Limiting the recursion depth limits the amount of
3585 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3586 when PCRE is compiled, and changed at run time. A patch from Google adds
3587 this functionality to the C++ interface.
3588
3589 18. Changes to the handling of Unicode character properties:
3590
3591 (a) Updated the table to Unicode 4.1.0.
3592
3593 (b) Recognize characters that are not in the table as "Cn" (undefined).
3594
3595 (c) I revised the way the table is implemented to a much improved format
3596 which includes recognition of ranges. It now supports the ranges that
3597 are defined in UnicodeData.txt, and it also amalgamates other
3598 characters into ranges. This has reduced the number of entries in the
3599 table from around 16,000 to around 3,000, thus reducing its size
3600 considerably. I realized I did not need to use a tree structure after
3601 all - a binary chop search is just as efficient. Having reduced the
3602 number of entries, I extended their size from 6 bytes to 8 bytes to
3603 allow for more data.
3604
3605 (d) Added support for Unicode script names via properties such as \p{Han}.
3606
3607 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3608 matching that character.
3609
3610 20. When matching a repeated Unicode property with a minimum greater than zero,
3611 (for example \pL{2,}), PCRE could look past the end of the subject if it
3612 reached it while seeking the minimum number of characters. This could
3613 happen only if some of the characters were more than one byte long, because
3614 there is a check for at least the minimum number of bytes.
3615
3616 21. Refactored the implementation of \p and \P so as to be more general, to
3617 allow for more different types of property in future. This has changed the
3618 compiled form incompatibly. Anybody with saved compiled patterns that use
3619 \p or \P will have to recompile them.
3620
3621 22. Added "Any" and "L&" to the supported property types.
3622
3623 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3624 but give a compile time error if the value is greater than 0xff.
3625
3626 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3627 accidentally not being installed or uninstalled.
3628
3629 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3630 made were to insert the current release number. This seemed silly, because
3631 it made things harder for people building PCRE on systems that don't run
3632 "configure". I have turned pcre.h into a distributed file, no longer built
3633 by "configure", with the version identification directly included. There is
3634 no longer a pcre.h.in file.
3635
3636 However, this change necessitated a change to the pcre-config script as
3637 well. It is built from pcre-config.in, and one of the substitutions was the
3638 release number. I have updated configure.ac so that ./configure now finds
3639 the release number by grepping pcre.h.
3640
3641 26. Added the ability to run the tests under valgrind.
3642
3643
3644 Version 6.4 05-Sep-05
3645 ---------------------
3646
3647 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3648 "--" to be printed when multiple files were scanned, even when none of the
3649 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3650 consider it to be a bug, and have restored the previous behaviour.
3651
3652 2. A couple of code tidies to get rid of compiler warnings.
3653
3654 3. The pcretest program used to cheat by referring to symbols in the library
3655 whose names begin with _pcre_. These are internal symbols that are not
3656 really supposed to be visible externally, and in some environments it is
3657 possible to suppress them. The cheating is now confined to including
3658 certain files from the library's source, which is a bit cleaner.
3659
3660 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3661 file's purpose clearer.
3662
3663 5. Reorganized pcre_ucp_findchar().
3664
3665
3666 Version 6.3 15-Aug-05
3667 ---------------------
3668
3669 1. The file libpcre.pc.in did not have general read permission in the tarball.
3670
3671 2. There were some problems when building without C++ support:
3672
3673 (a) If C++ support was not built, "make install" and "make test" still
3674 tried to test it.
3675
3676 (b) There were problems when the value of CXX was explicitly set. Some
3677 changes have been made to try to fix these, and ...
3678
3679 (c) --disable-cpp can now be used to explicitly disable C++ support.
3680
3681 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3682 backslash in a target when C++ was disabled. This confuses some
3683 versions of "make", apparently. Using an intermediate variable solves
3684 this. (Same for CPP_LOBJ.)
3685
3686 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3687 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3688 necessary on certain architectures.
3689
3690 4. Added a setting of -export-symbols-regex to the link command to remove
3691 those symbols that are exported in the C sense, but actually are local
3692 within the library, and not documented. Their names all begin with
3693 "_pcre_". This is not a perfect job, because (a) we have to except some
3694 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3695 available (and never for static libraries). I have made a note to try to
3696 find a way round (a) in the future.
3697
3698
3699 Version 6.2 01-Aug-05
3700 ---------------------
3701
3702 1. There was no test for integer overflow of quantifier values. A construction
3703 such as {1111111111111111} would give undefined results. What is worse, if
3704 a minimum quantifier for a parenthesized subpattern overflowed and became
3705 negative, the calculation of the memory size went wrong. This could have
3706 led to memory overwriting.
3707
3708 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3709
3710 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3711 operating environments where this matters.
3712
3713 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3714 PCRE options from within the C++ wrapper.
3715
3716 5. Named capturing subpatterns were not being correctly counted when a pattern
3717 was compiled. This caused two problems: (a) If there were more than 100
3718 such subpatterns, the calculation of the memory needed for the whole
3719 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3720 back references of the form \12, where the number was greater than 9, were
3721 not recognized as back references, even though there were sufficient
3722 previous subpatterns.
3723
3724 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3725 versions of gcc, e.g. 2.95.4.
3726
3727
3728 Version 6.1 21-Jun-05
3729 ---------------------
3730
3731 1. There was one reference to the variable "posix" in pcretest.c that was not
3732 surrounded by "#if !defined NOPOSIX".
3733
3734 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3735 the cross-check on the old pcre_info() function, for the benefit of the
3736 cut-down version of PCRE that is currently imported into Exim.
3737
3738 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3739 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3740 patterns that start (?i) but I don't think that matters. The use of (?i) is
3741 just an example; this all applies to the other options as well.
3742
3743 4. Since libtool seems to echo the compile commands it is issuing, the output
3744 from "make" can be reduced a bit by putting "@" in front of each libtool
3745 compile command.
3746
3747 5. Patch from the folks at Google for configure.in to be a bit more thorough
3748 in checking for a suitable C++ installation before trying to compile the
3749 C++ stuff. This should fix a reported problem when a compiler was present,
3750 but no suitable headers.
3751
3752 6. The man pages all had just "PCRE" as their title. I have changed them to
3753 be the relevant file name. I have also arranged that these names are
3754 retained in the file doc/pcre.txt, which is a concatenation in text format
3755 of all the man pages except the little individual ones for each function.
3756
3757 7. The NON-UNIX-USE file had not been updated for the different set of source
3758 files that come with release 6. I also added a few comments about the C++
3759 wrapper.
3760
3761
3762 Version 6.0 07-Jun-05
3763 ---------------------
3764
3765 1. Some minor internal re-organization to help with my DFA experiments.
3766
3767 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3768 didn't matter for the library itself when fully configured, but did matter
3769 when compiling without UCP support, or within Exim, where the ucp files are
3770 not imported.
3771
3772 3. Refactoring of the library code to split up the various functions into
3773 different source modules. The addition of the new DFA matching code (see
3774 below) to a single monolithic source would have made it really too
3775 unwieldy, quite apart from causing all the code to be include in a
3776 statically linked application, when only some functions are used. This is
3777 relevant even without the DFA addition now that patterns can be compiled in
3778 one application and matched in another.
3779
3780 The downside of splitting up is that there have to be some external
3781 functions and data tables that are used internally in different modules of
3782 the library but which are not part of the API. These have all had their
3783 names changed to start with "_pcre_" so that they are unlikely to clash
3784 with other external names.
3785
3786 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3787 a different (DFA) algorithm. Although it is slower than the original
3788 function, it does have some advantages for certain types of matching
3789 problem.
3790
3791 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3792 including restarting after a partial match.
3793
3794 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3795 defined when compiling for Windows was sent to me. I have put it into the
3796 code, though I have no means of testing or verifying it.
3797
3798 7. Added the pcre_refcount() auxiliary function.
3799
3800 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3801 match before or at the first newline in the subject string. In pcretest,
3802 the /f option on a pattern can be used to set this.
3803
3804 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3805 would behave wrongly. This has been present in PCRE since release 4.0.
3806
3807 10. A number of changes to the pcregrep command:
3808
3809 (a) Refactored how -x works; insert ^(...)$ instead of setting
3810 PCRE_ANCHORED and checking the length, in preparation for adding
3811 something similar for -w.
3812
3813 (b) Added the -w (match as a word) option.
3814
3815 (c) Refactored the way lines are read and buffered so as to have more
3816 than one at a time available.
3817
3818 (d) Implemented a pcregrep test script.
3819
3820 (e) Added the -M (multiline match) option. This allows patterns to match
3821 over several lines of the subject. The buffering ensures that at least
3822 8K, or the rest of the document (whichever is the shorter) is available
3823 for matching (and similarly the previous 8K for lookbehind assertions).
3824
3825 (f) Changed the --help output so that it now says
3826
3827 -w, --word-regex(p)
3828
3829 instead of two lines, one with "regex" and the other with "regexp"
3830 because that confused at least one person since the short forms are the
3831 same. (This required a bit of code, as the output is generated
3832 automatically from a table. It wasn't just a text change.)
3833
3834 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3835 option but starts with a hyphen. Could be a pattern or a path name
3836 starting with a hyphen, for instance.
3837
3838 (h) "-" can be given as a file name to represent stdin.
3839
3840 (i) When file names are being printed, "(standard input)" is used for
3841 the standard input, for compatibility with GNU grep. Previously
3842 "<stdin>" was used.
3843
3844 (j) The option --label=xxx can be used to supply a name to be used for
3845 stdin when file names are being printed. There is no short form.
3846
3847 (k) Re-factored the options decoding logic because we are going to add
3848 two more options that take data. Such options can now be given in four
3849 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3850
3851 (l) Added the -A, -B, and -C options for requesting that lines of context
3852 around matches be printed.
3853
3854 (m) Added the -L option to print the names of files that do not contain
3855 any matching lines, that is, the complement of -l.
3856
3857 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3858 continue to scan other files.
3859
3860 (o) The -s option was incorrectly implemented. For compatibility with other
3861 greps, it now suppresses the error message for a non-existent or non-
3862 accessible file (but not the return code). There is a new option called
3863 -q that suppresses the output of matching lines, which was what -s was
3864 previously doing.
3865
3866 (p) Added --include and --exclude options to specify files for inclusion
3867 and exclusion when recursing.
3868
3869 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3870 Hopefully, it now does.
3871
3872 12. Missing cast in pcre_study().
3873
3874 13. Added an "uninstall" target to the makefile.
3875
3876 14. Replaced "extern" in the function prototypes in Makefile.in with
3877 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3878 world, but is set differently for Windows.
3879
3880 15. Added a second compiling function called pcre_compile2(). The only
3881 difference is that it has an extra argument, which is a pointer to an
3882 integer error code. When there is a compile-time failure, this is set
3883 non-zero, in addition to the error test pointer being set to point to an
3884 error message. The new argument may be NULL if no error number is required
3885 (but then you may as well call pcre_compile(), which is now just a
3886 wrapper). This facility is provided because some applications need a
3887 numeric error indication, but it has also enabled me to tidy up the way
3888 compile-time errors are handled in the POSIX wrapper.
3889
3890 16. Added VPATH=.libs to the makefile; this should help when building with one
3891 prefix path and installing with another. (Or so I'm told by someone who
3892 knows more about this stuff than I do.)
3893
3894 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3895 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3896 match everything, including newlines. This is not POSIX-compatible, but
3897 somebody wanted the feature. From pcretest it can be activated by using
3898 both the P and the s flags.
3899
3900 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3901
3902 19. libpcre.pc was being incorrectly installed as executable.
3903
3904 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3905 it now also looks for '\r' so that it will work unmodified on Windows.
3906
3907 21. Added Google's contributed C++ wrapper to the distribution.
3908
3909 22. Added some untidy missing memory free() calls in pcretest, to keep
3910 Electric Fence happy when testing.
3911
3912
3913
3914 Version 5.0 13-Sep-04
3915 ---------------------
3916
3917 1. Internal change: literal characters are no longer packed up into items
3918 containing multiple characters in a single byte-string. Each character
3919 is now matched using a separate opcode. However, there may be more than one
3920 byte in the character in UTF-8 mode.
3921
3922 2. The pcre_callout_block structure has two new fields: pattern_position and
3923 next_item_length. These contain the offset in the pattern to the next match
3924 item, and its length, respectively.
3925
3926 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3927 insertion of callouts before each pattern item. Added the /C option to
3928 pcretest to make use of this.
3929
3930 4. On the advice of a Windows user, the lines
3931
3932 #if defined(_WIN32) || defined(WIN32)
3933 _setmode( _fileno( stdout ), 0x8000 );
3934 #endif /* defined(_WIN32) || defined(WIN32) */
3935
3936 have been added to the source of pcretest. This apparently does useful
3937 magic in relation to line terminators.
3938
3939 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3940 for the benefit of those environments where the "b" makes a difference.
3941
3942 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3943 to know about it. I have put a hack into configure.in that adds in code
3944 to set GCC=yes if CC=icc. This seems to end up at a point in the
3945 generated configure script that is early enough to affect the setting of
3946 compiler options, which is what is needed, but I have no means of testing
3947 whether it really works. (The user who reported this had patched the
3948 generated configure script, which of course I cannot do.)
3949
3950 LATER: After change 22 below (new libtool files), the configure script
3951 seems to know about icc (and also ecc). Therefore, I have commented out
3952 this hack in configure.in.
3953
3954 7. Added support for pkg-config (2 patches were sent in).
3955
3956 8. Negated POSIX character classes that used a combination of internal tables
3957 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3958 [[:^ascii]]. Typically, they would match almost any characters. The other
3959 POSIX classes were not broken in this way.
3960
3961 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3962 to find the match, as PCRE was deluded into thinking that the match had to
3963 start at the start point or following a newline. The same bug applied to
3964 patterns with negative forward assertions or any backward assertions
3965 preceding ".*" at the start, unless the pattern required a fixed first
3966 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3967
3968 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3969 starting at the last subject character, bytes beyond the end of the subject
3970 string were read.
3971
3972 11. Renamed the variable "class" as "classbits" to make life easier for C++
3973 users. (Previously there was a macro definition, but it apparently wasn't
3974 enough.)
3975
3976 12. Added the new field "tables" to the extra data so that tables can be passed
3977 in at exec time, or the internal tables can be re-selected. This allows
3978 a compiled regex to be saved and re-used at a later time by a different
3979 program that might have everything at different addresses.
3980
3981 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3982 -R library as well as a -L library.
3983
3984 14. The debugging options of pcretest (-d on the command line or D on a
3985 pattern) showed incorrect output for anything following an extended class
3986 that contained multibyte characters and which was followed by a quantifier.
3987
3988 15. Added optional support for general category Unicode character properties
3989 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3990 support. It adds about 90K to the size of the library. The meanings of the
3991 inbuilt class escapes such as \d and \s have NOT been changed.
3992
3993 16. Updated pcredemo.c to include calls to free() to release the memory for the
3994 compiled pattern.
3995
3996 17. The generated file chartables.c was being created in the source directory
3997 instead of in the building directory. This caused the build to fail if the
3998 source directory was different from the building directory, and was
3999 read-only.
4000
4001 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4002 file. No doubt somebody will tell me if they don't make sense... Also added
4003 Dan Mooney's comments about building on OpenVMS.
4004
4005 19. Added support for partial matching via the PCRE_PARTIAL option for
4006 pcre_exec() and the \P data escape in pcretest.
4007
4008 20. Extended pcretest with 3 new pattern features:
4009
4010 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4011 write the compiled pattern to the file whose name is "rest-of-line".
4012 This is a straight binary dump of the data, with the saved pointer to
4013 the character tables forced to be NULL. The study data, if any, is
4014 written too. After writing, pcretest reads a new pattern.
4015
4016 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4017 compiled pattern from the given file. There must not be any
4018 occurrences of "<" in the file name (pretty unlikely); if there are,
4019 pcretest will instead treat the initial "<" as a pattern delimiter.
4020 After reading in the pattern, pcretest goes on to read data lines as
4021 usual.
4022
4023 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4024 and 16-bit fields in a compiled pattern, to simulate a pattern that
4025 was compiled on a host of opposite endianness.
4026
4027 21. The pcre-exec() function can now cope with patterns that were compiled on
4028 hosts of opposite endianness, with this restriction:
4029
4030 As for any compiled expression that is saved and used later, the tables
4031 pointer field cannot be preserved; the extra_data field in the arguments
4032 to pcre_exec() should be used to pass in a tables address if a value
4033 other than the default internal tables were used at compile time.
4034
4035 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4036 now diagnosed as an error. Previously, most of the time, a negative number
4037 would have been treated as zero, but if in addition "ovector" was passed as
4038 NULL, a crash could occur.
4039
4040 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4041 new versions from the libtool 1.5 distribution (the last one is a copy of
4042 a file called libtool.m4). This seems to have fixed the need to patch
4043 "configure" to support Darwin 1.3 (which I used to do). However, I still
4044 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4045 workstation).
4046
4047 24. Changed the PCRE licence to be the more standard "BSD" licence.
4048
4049
4050 Version 4.5 01-Dec-03
4051 ---------------------
4052
4053 1. There has been some re-arrangement of the code for the match() function so
4054 that it can be compiled in a version that does not call itself recursively.
4055 Instead, it keeps those local variables that need separate instances for
4056 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4057 needs to "recurse". Keeping track of where control must go is done by means
4058 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4059 hide most of the details from the main code, and operates only if
4060 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4061 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4062 operating.
4063
4064 To make it easier for callers to provide specially tailored get/free
4065 functions for this usage, two new functions, pcre_stack_malloc, and
4066 pcre_stack_free, are used. They are always called in strict stacking order,
4067 and the size of block requested is always the same.
4068
4069 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4070 PCRE has been compiled to use the stack or the heap for recursion. The
4071 -C option of pcretest uses this to show which version is compiled.
4072
4073 A new data escape \S, is added to pcretest; it causes the amounts of store
4074 obtained and freed by both kinds of malloc/free at match time to be added
4075 to the output.
4076
4077 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4078 what's available on my current Linux desktop machine.
4079
4080 3. When matching a UTF-8 string, the test for a valid string at the start has
4081 been extended. If start_offset is not zero, PCRE now checks that it points
4082 to a byte that is the start of a UTF-8 character. If not, it returns
4083 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4084 this is necessary because there may be backward assertions in the pattern.
4085 When matching the same subject several times, it may save resources to use
4086 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4087
4088 4. The code for checking the validity of UTF-8 strings has been tightened so
4089 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4090 containing "overlong sequences".
4091
4092 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4093 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4094 should have been "&", but it just so happened that all the cases this let
4095 through by mistake were picked up later in the function.
4096
4097 6. I had used a variable called "isblank" - this is a C99 function, causing
4098 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4099
4100 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4101 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4102 so the version is shown; (c) stop "make test" echoing "./RunTest".
4103
4104 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4105
4106 9. The prototype for memmove() for systems that don't have it was using
4107 size_t, but the inclusion of the header that defines size_t was later. I've
4108 moved the #includes for the C headers earlier to avoid this.
4109
4110 10. Added some adjustments to the code to make it easier to compiler on certain
4111 special systems:
4112
4113 (a) Some "const" qualifiers were missing.
4114 (b) Added the macro EXPORT before all exported functions; by default this
4115 is defined to be empty.
4116 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4117 that it reads its output file name as an argument instead of writing
4118 to the standard output and assuming this can be redirected.
4119
4120 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4121 class containing characters with values greater than 255, PCRE compilation
4122 went into a loop.
4123
4124 12. A recursive reference to a subpattern that was within another subpattern
4125 that had a minimum quantifier of zero caused PCRE to crash. For example,
4126 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4127 recursion. If the recursively-called subpattern itself had a zero repeat,
4128 that was OK.
4129
4130 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4131 buffer into which it was copied (for escape processing) was still set at
4132 1024, so long lines caused crashes.
4133
4134 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4135 "internal error: code overflow...". This applied to any character class
4136 that was followed by a possessive quantifier.
4137
4138 15. Modified the Makefile to add libpcre.la as a prerequisite for
4139 libpcreposix.la because I was told this is needed for a parallel build to
4140 work.
4141
4142 16. If a pattern that contained .* following optional items at the start was
4143 studied, the wrong optimizing data was generated, leading to matching
4144 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4145 matching string must start with a or b or c. The correct conclusion for
4146 this pattern is that a match can start with any character.
4147
4148
4149 Version 4.4 13-Aug-03
4150 ---------------------
4151
4152 1. In UTF-8 mode, a character class containing characters with values between
4153 127 and 255 was not handled correctly if the compiled pattern was studied.
4154 In fixing this, I have also improved the studying algorithm for such
4155 classes (slightly).
4156
4157 2. Three internal functions had redundant arguments passed to them. Removal
4158 might give a very teeny performance improvement.
4159
4160 3. Documentation bug: the value of the capture_top field in a callout is *one
4161 more than* the number of the hightest numbered captured substring.
4162
4163 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4164 in incorrectly linking with a previously installed version. They now link
4165 explicitly with libpcre.la.
4166
4167 5. configure.in no longer needs to recognize Cygwin specially.
4168
4169 6. A problem in pcre.in for Windows platforms is fixed.
4170
4171 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4172 pcretest, it used to include the size of the study block as part of its
4173 output. Unfortunately, the structure contains a field that has a different
4174 size on different hardware architectures. This meant that the tests that
4175 showed this size failed. As the block is currently always of a fixed size,
4176 this information isn't actually particularly useful in pcretest output, so
4177 I have just removed it.
4178
4179 8. Three pre-processor statements accidentally did not start in column 1.
4180 Sadly, there are *still* compilers around that complain, even though
4181 standard C has not required this for well over a decade. Sigh.
4182
4183 9. In pcretest, the code for checking callouts passed small integers in the
4184 callout_data field, which is a void * field. However, some picky compilers
4185 complained about the casts involved for this on 64-bit systems. Now
4186 pcretest passes the address of the small integer instead, which should get
4187 rid of the warnings.
4188
4189 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4190 both compile and run time, and gives an error if an invalid UTF-8 sequence
4191 is found. There is a option for disabling this check in cases where the
4192 string is known to be correct and/or the maximum performance is wanted.
4193
4194 11. In response to a bug report, I changed one line in Makefile.in from
4195
4196 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4197 to
4198 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4199
4200 to look similar to other lines, but I have no way of telling whether this
4201 is the right thing to do, as I do not use Windows. No doubt I'll get told
4202 if it's wrong...
4203
4204
4205 Version 4.3 21-May-03
4206 ---------------------
4207
4208 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4209 Makefile.
4210
4211 2. Some refactoring to improve the quality of the code:
4212
4213 (i) The utf8_table... variables are now declared "const".
4214
4215 (ii) The code for \cx, which used the "case flipping" table to upper case
4216 lower case letters, now just substracts 32. This is ASCII-specific,
4217 but the whole concept of \cx is ASCII-specific, so it seems
4218 reasonable.
4219
4220 (iii) PCRE was using its character types table to recognize decimal and
4221 hexadecimal digits in the pattern. This is silly, because it handles
4222 only 0-9, a-f, and A-F, but the character types table is locale-
4223 specific, which means strange things might happen. A private
4224 table is now used for this - though it costs 256 bytes, a table is
4225 much faster than multiple explicit tests. Of course, the standard
4226 character types table is still used for matching digits in subject
4227 strings against \d.
4228
4229 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4230 ending in _t are). So I've renamed it as ESC_tee.
4231
4232 3. The first argument for regexec() in the POSIX wrapper should have been
4233 defined as "const".
4234
4235 4. Changed pcretest to use malloc() for its buffers so that they can be
4236 Electric Fenced for debugging.
4237
4238 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4239 to read one or more bytes before the start of the subject string. Often this
4240 had no effect on PCRE's behaviour, but in some circumstances it could
4241 provoke a segmentation fault.
4242
4243 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4244 to try to read one or more bytes before the start of the subject string.
4245
4246 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4247 UTF-8 support could misbehave in various ways if the subject string
4248 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4249 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4250 back over UTF-8 characters.)
4251
4252
4253 Version 4.2 14-Apr-03
4254 ---------------------
4255
4256 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4257
4258 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4259 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4260 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4261 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4262 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4263 and BUILD_EXEEXT
4264 Note: automatic setting of the BUILD variables is not yet working
4265 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4266 compile-time but not at link-time
4267 [LINK]: use for linking executables only
4268 make different versions for Windows and non-Windows
4269 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4270 libraries
4271 [LINK_FOR_BUILD]: new variable
4272 [OBJEXT]: use throughout
4273 [EXEEXT]: use throughout
4274 <winshared>: new target
4275 <wininstall>: new target
4276 <dftables.o>: use native compiler
4277 <dftables>: use native linker
4278 <install>: handle Windows platform correctly
4279 <clean>: ditto
4280 <check>: ditto
4281 copy DLL to top builddir before testing
4282
4283 As part of these changes, -no-undefined was removed again. This was reported
4284 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4285 in any case.
4286
4287 3. Some tidies to get rid of compiler warnings:
4288
4289 . In the match_data structure, match_limit was an unsigned long int, whereas
4290 match_call_count was an int. I've made them both unsigned long ints.
4291
4292 . In pcretest the fact that a const uschar * doesn't automatically cast to
4293 a void * provoked a warning.
4294
4295 . Turning on some more compiler warnings threw up some "shadow" variables
4296 and a few more missing casts.
4297
4298 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4299 option, a class that contained a single character with a value between 128
4300 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4301
4302 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4303 option, a class that contained several characters, but with at least one
4304 whose value was between 128 and 255 caused PCRE to crash.
4305
4306
4307 Version 4.1 12-Mar-03
4308 ---------------------
4309
4310 1. Compiling with gcc -pedantic found a couple of places where casts were
4311 needed, and a string in dftables.c that was longer than standard compilers are
4312 required to support.
4313
4314 2. Compiling with Sun's compiler found a few more places where the code could
4315 be tidied up in order to avoid warnings.
4316
4317 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4318 first of these names is deprecated in the latest Autoconf in favour of the name
4319 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4320 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4321 analogy I have changed it to CFLAGS_FOR_BUILD.
4322
4323 4. Added -no-undefined to the linking command in the Makefile, because this is
4324 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4325 linking step for the pcreposix library.
4326
4327 5. PCRE was failing to diagnose the case of two named groups with the same
4328 name.
4329
4330 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4331 literal character that is needed in the subject for a match, and scans along to
4332 ensure that it is present before embarking on the full matching process. This
4333 saves time in cases of nested unlimited repeats that are never going to match.
4334 Problem: the scan can take a lot of time if the subject is very long (e.g.
4335 megabytes), thus penalizing straightforward matches. It is now done only if the
4336 amount of subject to be scanned is less than 1000 bytes.
4337
4338 7. A lesser problem with the same optimization is that it was recording the
4339 first character of an anchored pattern as "needed", thus provoking a search
4340 right along the subject, even when the first match of the pattern was going to
4341 fail. The "needed" character is now not set for anchored patterns, unless it
4342 follows something in the pattern that is of non-fixed length. Thus, it still
4343 fulfils its original purpose of finding quick non-matches in cases of nested
4344 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4345
4346
4347 Version 4.0 17-Feb-03
4348 ---------------------
4349
4350 1. If a comment in an extended regex that started immediately after a meta-item
4351 extended to the end of string, PCRE compiled incorrect data. This could lead to
4352 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4353
4354 2. Moved to autoconf 2.53 and libtool 1.4.2.
4355
4356 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4357 the special perltest8 script is no longer needed - all the tests can be run
4358 from a single perltest script.
4359
4360 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4361 by \s. It has now been removed in PCRE. This means it isn't recognized as
4362 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4363 class [:space:] *does* include VT, thereby creating a mess.
4364
4365 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4366 space and tab.
4367
4368 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4369 its new features into the main test script, reducing the number of scripts.
4370
4371 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4372 were backward compatible, and made the (?i) apply to the whole pattern, as if
4373 /i were given. Now it behaves more logically, and applies the option setting
4374 only to what follows. PCRE has been changed to follow suit. However, if it
4375 finds options settings right at the start of the pattern, it extracts them into
4376 the global options, as before. Thus, they show up in the info data.
4377
4378 8. Added support for the \Q...\E escape sequence. Characters in between are
4379 treated as literals. This is slightly different from Perl in that $ and @ are
4380 also handled as literals inside the quotes. In Perl, they will cause variable
4381 interpolation. Note the following examples:
4382
4383 Pattern PCRE matches Perl matches
4384
4385 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4386 \Qabc\$xyz\E abc\$xyz abc\$xyz
4387 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4388
4389 For compatibility with Perl, \Q...\E sequences are recognized inside character
4390 classes as well as outside them.
4391
4392 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4393 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4394 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4395 signed/unsigned warnings.
4396
4397 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4398 option for pcretest, so I've replaced it by a simple function that does just
4399 that job.
4400
4401 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4402 "pcregrep -".
4403
4404 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4405 Java package. This provides some syntactic sugar for simple cases of what my
4406 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4407 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4408 item, you can use this simplified notation. Note that only makes sense with
4409 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4410 greediness, whatever the setting of the PCRE_UNGREEDY option.
4411
4412 13. A change of greediness default within a pattern was not taking effect at
4413 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4414 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4415 was abstracted outside.
4416
4417 14. PCRE now supports the \G assertion. It is true when the current matching
4418 position is at the start point of the match. This differs from \A when the
4419 starting offset is non-zero. Used with the /g option of pcretest (or similar
4420 code), it works in the same way as it does for Perl's /g option. If all
4421 alternatives of a regex begin with \G, the expression is anchored to the start
4422 match position, and the "anchored" flag is set in the compiled expression.
4423
4424 15. Some bugs concerning the handling of certain option changes within patterns
4425 have been fixed. These applied to options other than (?ims). For example,
4426 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4427 been the other way round. Some of this was related to change 7 above.
4428
4429 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4430 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4431 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4432 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4433
4434 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4435 mode. Unlike ".", it always matches newline, whatever the setting of
4436 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4437 assertions. Perl allows it, but it doesn't (in general) work because it can't
4438 calculate the length of the lookbehind. At least, that's the case for Perl
4439 5.8.0 - I've been told they are going to document that it doesn't work in
4440 future.
4441
4442 18. Added an error diagnosis for escapes that PCRE does not support: these are
4443 \L, \l, \N, \P, \p, \U, \u, and \X.
4444
4445 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4446 reading past the end of the pattern in cases such as /[abcd/.
4447
4448 20. PCRE was getting more memory than necessary for patterns with classes that
4449 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4450
4451 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4452 compiling PCRE for use with Virtual Pascal.
4453
4454 22. Small fix to the Makefile to make it work properly if the build is done
4455 outside the source tree.
4456
4457 23. Added a new extension: a condition to go with recursion. If a conditional
4458 subpattern starts with (?(R) the "true" branch is used if recursion has
4459 happened, whereas the "false" branch is used only at the top level.
4460
4461 24. When there was a very long string of literal characters (over 255 bytes
4462 without UTF support, over 250 bytes with UTF support), the computation of how
4463 much memory was required could be incorrect, leading to segfaults or other
4464 strange effects.
4465
4466 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4467 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4468 there was a subsequent back reference to those brackets. This meant that, for
4469 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4470 possible to check for precisely this case. All we can do is abandon the
4471 optimization if .* occurs inside capturing brackets when there are any back
4472 references whatsoever. (See below for a better fix that came later.)
4473
4474 26. The handling of the optimization for finding the first character of a
4475 non-anchored pattern, and for finding a character that is required later in the
4476 match were failing in some cases. This didn't break the matching; it just
4477 failed to optimize when it could. The way this is done has been re-implemented.
4478
4479 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4480
4481 28. Added a new feature that provides some of the functionality that Perl
4482 provides with (?{...}). The facility is termed a "callout". The way it is done
4483 in PCRE is for the caller to provide an optional function, by setting
4484 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4485 global variable. By default it is unset, which disables all calling out. To get
4486 the function called, the regex must include (?C) at appropriate points. This
4487 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4488 This provides a means of identifying different callout points. When PCRE
4489 reaches such a point in the regex, if pcre_callout has been set, the external
4490 function is called. It is provided with data in a structure called
4491 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4492 matching continues; if it returns a non-zero value, the match at the current
4493 point fails. However, backtracking will occur if possible. [This was changed
4494 later and other features added - see item 49 below.]
4495
4496 29. pcretest is upgraded to test the callout functionality. It provides a
4497 callout function that displays information. By default, it shows the start of
4498 the match and the current position in the text. There are some new data escapes
4499 to vary what happens:
4500
4501 \C+ in addition, show current contents of captured substrings
4502 \C- do not supply a callout function
4503 \C!n return 1 when callout number n is reached
4504 \C!n!m return 1 when callout number n is reached for the mth time
4505
4506 30. If pcregrep was called with the -l option and just a single file name, it
4507 output "<stdin>" if a match was found, instead of the file name.
4508
4509 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4510 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4511 pcre_exec(). This saves a malloc/free per call. The default value of
4512 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4513 when configuring.
4514
4515 32. The default maximum size of a compiled pattern is 64K. There have been a
4516 few cases of people hitting this limit. The code now uses macros to handle the
4517 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4518 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4519 configuring. Tests 2 and 5 work only with 2-byte links because they output
4520 debugging information about compiled patterns.
4521
4522 33. Internal code re-arrangements:
4523
4524 (a) Moved the debugging function for printing out a compiled regex into
4525 its own source file (printint.c) and used #include to pull it into
4526 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4527 separate copies.
4528
4529 (b) Defined the list of op-code names for debugging as a macro in
4530 internal.h so that it is next to the definition of the opcodes.
4531
4532 (c) Defined a table of op-code lengths for simpler skipping along compiled
4533 code. This is again a macro in internal.h so that it is next to the
4534 definition of the opcodes.
4535
4536 34. Added support for recursive calls to individual subpatterns, along the
4537 lines of Robin Houston's patch (but implemented somewhat differently).
4538
4539 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4540 allow it to read and process whole directories in Win32. This code was
4541 contributed by Lionel Fourquaux; it has not been tested by me.
4542
4543 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4544 used to name a group. Names consist of alphanumerics and underscores, and must
4545 be unique. Back references use the syntax (?P=name) and recursive calls use
4546 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4547 numbers. The function pcre_fullinfo() can be used after compilation to extract
4548 a name/number map. There are three relevant calls:
4549
4550 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4551 PCRE_INFO_NAMECOUNT yields the number of entries
4552 PCRE_INFO_NAMETABLE yields a pointer to the map.
4553
4554 The map is a vector of fixed-size entries. The size of each entry depends on
4555 the length of the longest name used. The first two bytes of each entry are the
4556 group number, most significant byte first. There follows the corresponding
4557 name, zero terminated. The names are in alphabetical order.
4558
4559 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4560 case instead of 255. Making it the same both with and without UTF-8 support
4561 means that the same test output works with both.
4562
4563 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4564 calling malloc() with a zero argument.
4565
4566 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4567 optimization. I've improved things by keeping a bitmap of backreferences with
4568 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4569 fact referenced, the optimization can be applied. It is unlikely that a
4570 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4571 the match to follow \n) will appear inside brackets with a number greater than
4572 31, but if it does, any back reference > 31 suppresses the optimization.
4573
4574 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4575 of disabling numbered capturing parentheses. Any opening parenthesis that is
4576 not followed by ? behaves as if it were followed by ?: but named parentheses
4577 can still be used for capturing (and they will acquire numbers in the usual
4578 way).
4579
4580 41. Redesigned the return codes from the match() function into yes/no/error so
4581 that errors can be passed back from deep inside the nested calls. A malloc
4582 failure while inside a recursive subpattern call now causes the
4583 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4584
4585 42. It is now possible to set a limit on the number of times the match()
4586 function is called in a call to pcre_exec(). This facility makes it possible to
4587 limit the amount of recursion and backtracking, though not in a directly
4588 obvious way, because the match() function is used in a number of different
4589 circumstances. The count starts from zero for each position in the subject
4590 string (for non-anchored patterns). The default limit is, for compatibility, a
4591 large number, namely 10 000 000. You can change this in two ways:
4592
4593 (a) When configuring PCRE before making, you can use --with-match-limit=n
4594 to set a default value for the compiled library.
4595
4596 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4597 a different value is set. See 45 below.
4598
4599 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4600
4601 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4602 of things that can be changed at compile time. The first argument specifies
4603 what is wanted and the second points to where the information is to be placed.
4604 The current list of available information is:
4605
4606 PCRE_CONFIG_UTF8
4607
4608 The output is an integer that is set to one if UTF-8 support is available;
4609 otherwise it is set to zero.
4610
4611 PCRE_CONFIG_NEWLINE
4612
4613 The output is an integer that it set to the value of the code that is used for
4614 newline. It is either LF (10) or CR (13).
4615
4616 PCRE_CONFIG_LINK_SIZE
4617
4618 The output is an integer that contains the number of bytes used for internal
4619 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4620
4621 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4622
4623 The output is an integer that contains the threshold above which the POSIX
4624 interface uses malloc() for output vectors. See item 31 above.
4625
4626 PCRE_CONFIG_MATCH_LIMIT
4627
4628 The output is an unsigned integer that contains the default limit of the number
4629 of match() calls in a pcre_exec() execution. See 42 above.
4630
4631 44. pcretest has been upgraded by the addition of the -C option. This causes it
4632 to extract all the available output from the new pcre_config() function, and to
4633 output it. The program then exits immediately.
4634
4635 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4636 order to support additional features. One way would have been to define
4637 pcre_exec2() (for example) with extra arguments, but this would not have been
4638 extensible, and would also have required all calls to the original function to
4639 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4640 is used for passing in "extra" data from pcre_study().
4641
4642 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4643 contains the following fields:
4644
4645 flags a bitmap indicating which of the following fields are set
4646 study_data opaque data from pcre_study()
4647 match_limit a way of specifying a limit on match() calls for a specific
4648 call to pcre_exec()
4649 callout_data data for callouts (see 49 below)
4650
4651 The flag bits are also defined in pcre.h, and are
4652
4653 PCRE_EXTRA_STUDY_DATA
4654 PCRE_EXTRA_MATCH_LIMIT
4655 PCRE_EXTRA_CALLOUT_DATA
4656
4657 The pcre_study() function now returns one of these new pcre_extra blocks, with
4658 the actual study data pointed to by the study_data field, and the
4659 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4660 before. That is, this change is entirely upwards-compatible and requires no
4661 change to existing code.
4662
4663 If you want to pass in additional data to pcre_exec(), you can either place it
4664 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4665 block.
4666
4667 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4668 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4669 times with different match limits, until it finds the minimum value needed for
4670 pcre_exec() to complete. The value is then output. This can be instructive; for
4671 most simple matches the number is quite small, but for pathological cases it
4672 gets very large very quickly.
4673
4674 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4675 returns the size of the data block pointed to by the study_data field in a
4676 pcre_extra block, that is, the value that was passed as the argument to
4677 pcre_malloc() when PCRE was getting memory in which to place the information
4678 created by pcre_study(). The fourth argument should point to a size_t variable.
4679 pcretest has been extended so that this information is shown after a successful
4680 pcre_study() call when information about the compiled regex is being displayed.
4681
4682 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4683 because what follows is always an absolute path. (Later: it turns out that this
4684 is more than cosmetic for MinGW, because it doesn't like empty path
4685 components.)
4686
4687 49. Some changes have been made to the callout feature (see 28 above):
4688
4689 (i) A callout function now has three choices for what it returns:
4690
4691 0 => success, carry on matching
4692 > 0 => failure at this point, but backtrack if possible
4693 < 0 => serious error, return this value from pcre_exec()
4694
4695 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4696 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4697 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4698 use by callout functions. It will never be used by PCRE itself.
4699
4700 (ii) The pcre_extra structure (see 45 above) has a void * field called
4701 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4702 pcre_callout_block structure has a field of the same name. The contents of
4703 the field passed in the pcre_extra structure are passed to the callout
4704 function in the corresponding field in the callout block. This makes it
4705 easier to use the same callout-containing regex from multiple threads. For
4706 testing, the pcretest program has a new data escape
4707
4708 \C*n pass the number n (may be negative) as callout_data
4709
4710 If the callout function in pcretest receives a non-zero value as
4711 callout_data, it returns that value.
4712
4713 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4714 there were some redundant $(CFLAGS) in commands that are now specified as
4715 $(LINK), which already includes $(CFLAGS).
4716
4717 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4718 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4719 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4720 one-byte characters throughout. Note that case-insensitive matching applies
4721 only to characters whose values are less than 256. PCRE doesn't support the
4722 notion of cases for higher-valued characters.
4723
4724 (i) A character class whose characters are all within 0-255 is handled as
4725 a bit map, and the map is inverted for negative classes. Previously, a
4726 character > 255 always failed to match such a class; however it should
4727 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4728
4729 (ii) A negated character class with a single character < 255 is coded as
4730 "not this character" (OP_NOT). This wasn't working properly when the test
4731 character was multibyte, either singly or repeated.
4732
4733 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4734 mode, for example: \x{100}{2,3}.
4735
4736 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4737 singly or repeated) now correctly test multibyte characters. However,
4738 PCRE doesn't recognize any characters with values greater than 255 as
4739 digits, spaces, or word characters. Such characters always match \D, \S,
4740 and \W, and never match \d, \s, or \w.
4741
4742 (v) Classes may now contain characters and character ranges with values
4743 greater than 255. For example: [ab\x{100}-\x{400}].
4744
4745 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4746 PCRE in UTF-8 mode.
4747
4748 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4749 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4750 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4751 value.)
4752
4753 53. The single man page has become too large. I have therefore split it up into
4754 a number of separate man pages. These also give rise to individual HTML pages;
4755 these are now put in a separate directory, and there is an index.html page that
4756 lists them all. Some hyperlinking between the pages has been installed.
4757
4758 54. Added convenience functions for handling named capturing parentheses.
4759
4760 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4761 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4762 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4763 are faulted.
4764
4765 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4766 calling configure. These values are used when compiling the dftables.c program
4767 which is run to generate the source of the default character tables. They
4768 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4769 you will need to set these values.
4770
4771 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4772
4773
4774 Version 3.9 02-Jan-02
4775 ---------------------
4776
4777 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4778
4779 2. If --disable-static was given, the building process failed when trying to
4780 build pcretest and pcregrep. (For some reason it was using libtool to compile
4781 them, which is not right, as they aren't part of the library.)
4782
4783
4784 Version 3.8 18-Dec-01
4785 ---------------------
4786
4787 1. The experimental UTF-8 code was completely screwed up. It was packing the
4788 bytes in the wrong order. How dumb can you get?
4789
4790
4791 Version 3.7 29-Oct-01
4792 ---------------------
4793
4794 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4795 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4796 this didn't happen under Solaris 8, where I normally test things.
4797
4798 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4799 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4800 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4801 this fix an hour or so after the initial 3.7 release.)
4802
4803
4804 Version 3.6 23-Oct-01
4805 ---------------------
4806
4807 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4808 offsets passed as NULL with zero offset count.
4809
4810 2. The config.guess and config.sub files had not been updated when I moved to
4811 the latest autoconf.
4812
4813
4814 Version 3.5 15-Aug-01
4815 ---------------------
4816
4817 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4818 had been forgotten.
4819
4820 2. By using declared but undefined structures, we can avoid using "void"
4821 definitions in pcre.h while keeping the internal definitions of the structures
4822 private.
4823
4824 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4825 user point of view, this means that both static and shared libraries are built
4826 by default, but this can be individually controlled. More of the work of
4827 handling this static/shared cases is now inside libtool instead of PCRE's make
4828 file.
4829
4830 4. The pcretest utility is now installed along with pcregrep because it is
4831 useful for users (to test regexs) and by doing this, it automatically gets
4832 relinked by libtool. The documentation has been turned into a man page, so
4833 there are now .1, .txt, and .html versions in /doc.
4834
4835 5. Upgrades to pcregrep:
4836 (i) Added long-form option names like gnu grep.
4837 (ii) Added --help to list all options with an explanatory phrase.
4838 (iii) Added -r, --recursive to recurse into sub-directories.
4839 (iv) Added -f, --file to read patterns from a file.
4840
4841 6. pcre_exec() was referring to its "code" argument before testing that
4842 argument for NULL (and giving an error if it was NULL).
4843
4844 7. Upgraded Makefile.in to allow for compiling in a different directory from
4845 the source directory.
4846
4847 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4848 options bits, the pointer it was passed was to an int instead of to an unsigned
4849 long int. This mattered only on 64-bit systems.
4850
4851 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4852 generated) instead of pcre.in, which it its source. Also made the same change
4853 in several of the .c files.
4854
4855 10. A new release of gcc defines printf() as a macro, which broke pcretest
4856 because it had an ifdef in the middle of a string argument for printf(). Fixed
4857 by using separate calls to printf().
4858
4859 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4860 script, to force use of CR or LF instead of \n in the source. On non-Unix
4861 systems, the value can be set in config.h.
4862
4863 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4864 absolute limit. Changed the text of the error message to make this clear, and
4865 likewise updated the man page.
4866
4867 13. The limit of 99 on the number of capturing subpatterns has been removed.
4868 The new limit is 65535, which I hope will not be a "real" limit.
4869
4870
4871 Version 3.4 22-Aug-00
4872 ---------------------
4873
4874 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4875
4876 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4877
4878
4879 Version 3.3 01-Aug-00
4880 ---------------------
4881
4882 1. If an octal character was given, but the value was greater than \377, it
4883 was not getting masked to the least significant bits, as documented. This could
4884 lead to crashes in some systems.
4885
4886 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4887 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4888
4889 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4890 These just pass their arguments on to (pcre_free)(), but they are provided
4891 because some uses of PCRE bind it to non-C systems that can call its functions,
4892 but cannot call free() or pcre_free() directly.
4893
4894 4. Add "make test" as a synonym for "make check". Corrected some comments in
4895 the Makefile.
4896
4897 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4898 Makefile.
4899
4900 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4901 command called pgrep for grepping around the active processes.
4902
4903 7. Added the beginnings of support for UTF-8 character strings.
4904
4905 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4906 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4907 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4908 out for the ar command.)
4909
4910
4911 Version 3.2 12-May-00
4912 ---------------------
4913
4914 This is purely a bug fixing release.
4915
4916 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4917 of ZA. This was just one example of several cases that could provoke this bug,
4918 which was introduced by change 9 of version 2.00. The code for breaking
4919 infinite loops after an iteration that matches an empty string was't working
4920 correctly.
4921
4922 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4923 when matched against abbab (for example). After matching an empty string, it
4924 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4925 caused it to match further down the string than it should.
4926
4927 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4928 was there because it doesn't seem to be needed, and it causes trouble on some
4929 systems, as it is not a Standard C header. It has been removed.
4930
4931 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4932 were reported on the Macintosh. The changes were from
4933
4934 while ((c = *(++ptr)) != 0 && c != '\n');
4935 to
4936 while ((c = *(++ptr)) != 0 && c != '\n') ;
4937
4938 Totally extraordinary, but if that's what it takes...
4939
4940 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4941 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4942 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4943 assumes the way PCRE uses memmove() (always moving upwards).
4944
4945 6. PCRE is being used in one environment where strchr() is not available. There
4946 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4947 faster code anyway.
4948
4949
4950 Version 3.1 09-Feb-00
4951 ---------------------
4952
4953 The only change in this release is the fixing of some bugs in Makefile.in for
4954 the "install" target:
4955
4956 (1) It was failing to install pcreposix.h.
4957
4958 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4959
4960
4961 Version 3.0 01-Feb-00
4962 ---------------------
4963
4964 1. Add support for the /+ modifier to perltest (to output $` like it does in
4965 pcretest).
4966
4967 2. Add support for the /g modifier to perltest.
4968
4969 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4970 matches null strings.
4971
4972 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4973 pattern. Perl treats empty patterns specially - it reuses the most recent
4974 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4975 effect.
4976
4977 5. The POSIX interface was broken in that it was just handing over the POSIX
4978 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4979 required a bigger vector, with some working space on the end. This means that
4980 the POSIX wrapper now has to get and free some memory, and copy the results.
4981
4982 6. Added some simple autoconf support, placing the test data and the
4983 documentation in separate directories, re-organizing some of the
4984 information files, and making it build pcre-config (a GNU standard). Also added
4985 libtool support for building PCRE as a shared library, which is now the
4986 default.
4987
4988 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4989 09 are not valid octal constants. Single digits will be used for minor values
4990 less than 10.
4991
4992 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4993 existing programs that set these in the POSIX interface can use PCRE without
4994 modification.
4995
4996 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4997 return all that pcre_info() returns, plus additional data. The pcre_info()
4998 function is retained for compatibility, but is considered to be obsolete.
4999
5000 10. Added experimental recursion feature (?R) to handle one common case that
5001 Perl 5.6 will be able to do with (?p{...}).
5002
5003 11. Added support for POSIX character classes like [:alpha:], which Perl is
5004 adopting.
5005
5006
5007 Version 2.08 31-Aug-99
5008 ----------------------
5009
5010 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5011 trying to match at the startoffset position, but instead was moving forward to
5012 the next newline as if a previous match had failed.
5013
5014 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5015 and could get into a loop if a null string was matched other than at the start
5016 of the subject.
5017
5018 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5019 be distinguished at compile time, and for completeness also added PCRE_DATE.
5020
5021 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5022 in GnuWin32 environments.
5023
5024
5025 Version 2.07 29-Jul-99
5026 ----------------------
5027
5028 1. The documentation is now supplied in plain text form and HTML as well as in
5029 the form of man page sources.
5030
5031 2. C++ compilers don't like assigning (void *) values to other pointer types.
5032 In particular this affects malloc(). Although there is no problem in Standard
5033 C, I've put in casts to keep C++ compilers happy.
5034
5035 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5036 should be (const char *).
5037
5038 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5039 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5040 However, I haven't made this a standard facility. The documentation doesn't
5041 mention it, and the Makefile doesn't support it.
5042
5043 5. The Makefile now contains an "install" target, with editable destinations at
5044 the top of the file. The pcretest program is not installed.
5045
5046 6. pgrep -V now gives the PCRE version number and date.
5047
5048 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5049 causing the entire string to be ignored, instead of just the last character.
5050
5051 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5052 non-matching string, it can take a very, very long time, even for strings of
5053 quite modest length, because of the nested recursion. PCRE now does better in
5054 some of these cases. It does this by remembering the last required literal
5055 character in the pattern, and pre-searching the subject to ensure it is present
5056 before running the real match. In other words, it applies a heuristic to detect
5057 some types of certain failure quickly, and in the above example, if presented
5058 with a string that has no trailing " it gives "no match" very quickly.
5059
5060 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5061 other alternatives are tried instead.
5062
5063
5064 Version 2.06 09-Jun-99
5065 ----------------------
5066
5067 1. Change pcretest's output for amount of store used to show just the code
5068 space, because the remainder (the data block) varies in size between 32-bit and
5069 64-bit systems.
5070
5071 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5072 start matching at. This allows lookbehinds to work when searching for multiple
5073 occurrences in a string.
5074
5075 3. Added additional options to pcretest for testing multiple occurrences:
5076
5077 /+ outputs the rest of the string that follows a match
5078 /g loops for multiple occurrences, using the new startoffset argument
5079 /G loops for multiple occurrences by passing an incremented pointer
5080
5081 4. PCRE wasn't doing the "first character" optimization for patterns starting
5082 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5083 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5084 the letter 'x'. On long subject strings, this gives a significant speed-up.
5085
5086
5087 Version 2.05 21-Apr-99
5088 ----------------------
5089
5090 1. Changed the type of magic_number from int to long int so that it works
5091 properly on 16-bit systems.
5092
5093 2. Fixed a bug which caused patterns starting with .* not to work correctly
5094 when the subject string contained newline characters. PCRE was assuming
5095 anchoring for such patterns in all cases, which is not correct because .* will
5096 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5097 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5098 must be retried after every newline in the subject.
5099
5100
5101 Version 2.04 18-Feb-99
5102 ----------------------
5103
5104 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5105 computation of the store needed to hold the pattern was incorrect (too large).
5106 If such patterns were nested a few deep, this could multiply and become a real
5107 problem.
5108
5109 2. Added /M option to pcretest to show the memory requirement of a specific
5110 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5111
5112 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5113 compiled in such a way that the backtracking after subsequent failure was
5114 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5115 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5116
5117
5118 Version 2.03 02-Feb-99
5119 ----------------------
5120
5121 1. Fixed typo and small mistake in man page.
5122
5123 2. Added 4th condition (GPL supersedes if conflict) and created separate
5124 LICENCE file containing the conditions.
5125
5126 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5127 Perl, that is the internal \ allows the delimiter to be included in the
5128 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5129 the final delimiter, add \ to the end of the pattern (to test the error).
5130
5131 4. Added the convenience functions for extracting substrings after a successful
5132 match. Updated pcretest to make it able to test these functions.
5133
5134
5135 Version 2.02 14-Jan-99
5136 ----------------------
5137
5138 1. Initialized the working variables associated with each extraction so that
5139 their saving and restoring doesn't refer to uninitialized store.
5140
5141 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5142 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5143 fix the problem.
5144
5145 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5146 calls, and wasn't printing the correct value for compiling calls. Increased the
5147 default value of LOOPREPEAT, and the number of significant figures in the
5148 times.
5149
5150 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5151
5152 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5153 a building problem on Windows NT with a FAT file system.
5154
5155
5156 Version 2.01 21-Oct-98
5157 ----------------------
5158
5159 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5160 to character tables built by pcre_maketables() in the current locale. If NULL
5161 is passed, the default tables are used.
5162
5163
5164 Version 2.00 24-Sep-98
5165 ----------------------
5166
5167 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5168 it any more.
5169
5170 2. Allow quantification of (?>) groups, and make it work correctly.
5171
5172 3. The first character computation wasn't working for (?>) groups.
5173
5174 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5175 end of the subject) and add 5.005's \z, which really does match only at the
5176 very end of the subject.
5177
5178 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5179
5180 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5181 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5182 localized options. All options to pcre_study() were also removed.
5183
5184 7. Add other new features from 5.005:
5185
5186 $(?<= positive lookbehind
5187 $(?<! negative lookbehind
5188 (?imsx-imsx) added the unsetting capability
5189 such a setting is global if at outer level; local otherwise
5190 (?imsx-imsx:) non-capturing groups with option setting
5191 (?(cond)re|re) conditional pattern matching
5192
5193 A backreference to itself in a repeated group matches the previous
5194 captured string.
5195
5196 8. General tidying up of studying (both automatic and via "study")
5197 consequential on the addition of new assertions.
5198
5199 9. As in 5.005, unlimited repeated groups that could match an empty substring
5200 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5201 runtime if any iteration does actually match an empty substring.
5202
5203 10. Include the RunTest script in the distribution.
5204
5205 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5206 discrepancies, some of which were old and were also with respect to 5.004. They
5207 have now been fixed.
5208
5209
5210 Version 1.09 28-Apr-98
5211 ----------------------
5212
5213 1. A negated single character class followed by a quantifier with a minimum
5214 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5215 program crashes, or just wrong answers. This did not apply to negated classes
5216 containing more than one character, or to minima other than one.
5217
5218
5219 Version 1.08 27-Mar-98
5220 ----------------------
5221
5222 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5223
5224 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5225 latter must appear before anything that relies on it in the pattern.
5226
5227
5228 Version 1.07 16-Feb-98
5229 ----------------------
5230
5231 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5232 repeat of a potentially empty string).
5233
5234
5235 Version 1.06 23-Jan-98
5236 ----------------------
5237
5238 1. Added Markus Oberhumer's little patches for C++.
5239
5240 2. Literal strings longer than 255 characters were broken.
5241
5242
5243 Version 1.05 23-Dec-97
5244 ----------------------
5245
5246 1. Negated character classes containing more than one character were failing if
5247 PCRE_CASELESS was set at run time.
5248
5249
5250 Version 1.04 19-Dec-97
5251 ----------------------
5252
5253 1. Corrected the man page, where some "const" qualifiers had been omitted.
5254
5255 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5256 input syntax.
5257
5258 3. Fixed memory leak which occurred when a regex with back references was
5259 matched with an offsets vector that wasn't big enough. The temporary memory
5260 that is used in this case wasn't being freed if the match failed.
5261
5262 4. Tidied pcretest to ensure it frees memory that it gets.
5263
5264 5. Temporary memory was being obtained in the case where the passed offsets
5265 vector was exactly big enough.
5266
5267 6. Corrected definition of offsetof() from change 5 below.
5268
5269 7. I had screwed up change 6 below and broken the rules for the use of
5270 setjmp(). Now fixed.
5271
5272
5273 Version 1.03 18-Dec-97
5274 ----------------------
5275
5276 1. A erroneous regex with a missing opening parenthesis was correctly
5277 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5278 on some systems.
5279
5280 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5281 it was reported that one broken compiler failed on the former because "code" is
5282 also an independent variable.
5283
5284 3. The erroneous regex a[]b caused an array overrun reference.
5285
5286 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5287 fail on data ending with that character. (It was going on too far, and checking
5288 the next character, typically a binary zero.) This was specific to the
5289 optimized code for single-character negative classes.
5290
5291 5. Added a contributed patch from the TIN world which does the following:
5292
5293 + Add an undef for memmove, in case the the system defines a macro for it.
5294
5295 + Add a definition of offsetof(), in case there isn't one. (I don't know
5296 the reason behind this - offsetof() is part of the ANSI standard - but
5297 it does no harm).
5298
5299 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5300 most of the places where whitespace preceded '#'. I have given up and
5301 allowed the remaining 2 cases to be at the margin.
5302
5303 + Rename some variables in pcre to eliminate shadowing. This seems very
5304 pedantic, but does no harm, of course.
5305
5306 6. Moved the call to setjmp() into its own function, to get rid of warnings
5307 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5308
5309 7. Constructs such as \d{8,} were compiling into the equivalent of
5310 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5311 outcome, but in this particular case used more store than had been allocated,
5312 which caused the bug to be discovered because it threw up an internal error.
5313
5314 8. The debugging code in both pcre and pcretest for outputting the compiled
5315 form of a regex was going wrong in the case of back references followed by
5316 curly-bracketed repeats.
5317
5318
5319 Version 1.02 12-Dec-97
5320 ----------------------
5321
5322 1. Typos in pcre.3 and comments in the source fixed.
5323
5324 2. Applied a contributed patch to get rid of places where it used to remove
5325 'const' from variables, and fixed some signed/unsigned and uninitialized
5326 variable warnings.
5327
5328 3. Added the "runtest" target to Makefile.
5329
5330 4. Set default compiler flag to -O2 rather than just -O.
5331
5332
5333 Version 1.01 19-Nov-97
5334 ----------------------
5335
5336 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5337 like /([ab]*)*/, that is, for classes with more than one character in them.
5338
5339 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5340 as /((?>a*))*/ (a PCRE_EXTRA facility).
5341
5342
5343 Version 1.00 18-Nov-97
5344 ----------------------
5345
5346 1. Added compile-time macros to support systems such as SunOS4 which don't have
5347 memmove() or strerror() but have other things that can be used instead.
5348
5349 2. Arranged that "make clean" removes the executables.
5350
5351
5352 Version 0.99 27-Oct-97
5353 ----------------------
5354
5355 1. Fixed bug in code for optimizing classes with only one character. It was
5356 initializing a 32-byte map regardless, which could cause it to run off the end
5357 of the memory it had got.
5358
5359 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5360
5361
5362 Version 0.98 22-Oct-97
5363 ----------------------
5364
5365 1. Fixed bug in code for handling temporary memory usage when there are more
5366 back references than supplied space in the ovector. This could cause segfaults.
5367
5368
5369 Version 0.97 21-Oct-97
5370 ----------------------
5371
5372 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5373
5374 2. Optimized negated single characters not to use a bit map.
5375
5376 3. Brought error texts together as macro definitions; clarified some of them;
5377 fixed one that was wrong - it said "range out of order" when it meant "invalid
5378 escape sequence".
5379
5380 4. Changed some char * arguments to const char *.
5381
5382 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5383
5384 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5385 pcretest.
5386
5387
5388 Version 0.96 16-Oct-97
5389 ----------------------
5390
5391 1. Added a simple "pgrep" utility to the distribution.
5392
5393 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5394 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5395 where "ddd" means "one or more decimal digits".
5396
5397 3. Fixed serious bug. If a pattern had a back reference, but the call to
5398 pcre_exec() didn't supply a large enough ovector to record the related
5399 identifying subpattern, the match always failed. PCRE now remembers the number
5400 of the largest back reference, and gets some temporary memory in which to save
5401 the offsets during matching if necessary, in order to ensure that
5402 backreferences always work.
5403
5404 4. Increased the compatibility with Perl in a number of ways:
5405
5406 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5407 to request this handling. The option can be set at compile or exec time.
5408
5409 (b) $ matches before a terminating newline by default; an option
5410 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5411 mode). The option can be set at compile or exec time.
5412
5413 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5414 the same as Perl's. If the decimal number it represents is less than 10
5415 or there aren't that many previous left capturing parentheses, an octal
5416 escape is read. Inside a character class, it's always an octal escape,
5417 even if it is a single digit.
5418
5419 (d) An escaped but undefined alphabetic character is taken as a literal,
5420 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5421 escapes.
5422
5423 (e) {0} is now permitted. (The previous item is removed from the compiled
5424 pattern).
5425
5426 5. Changed all the names of code files so that the basic parts are no longer
5427 than 10 characters, and abolished the teeny "globals.c" file.
5428
5429 6. Changed the handling of character classes; they are now done with a 32-byte
5430 bit map always.
5431
5432 7. Added the -d and /D options to pcretest to make it possible to look at the
5433 internals of compilation without having to recompile pcre.
5434
5435
5436 Version 0.95 23-Sep-97
5437 ----------------------
5438
5439 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5440 \x20 at the start of a run of normal characters. These were being treated as
5441 real characters, instead of the source characters being re-checked.
5442
5443
5444 Version 0.94 18-Sep-97
5445 ----------------------
5446
5447 1. The functions are now thread-safe, with the caveat that the global variables
5448 containing pointers to malloc() and free() or alternative functions are the
5449 same for all threads.
5450
5451 2. Get pcre_study() to generate a bitmap of initial characters for non-
5452 anchored patterns when this is possible, and use it if passed to pcre_exec().
5453
5454
5455 Version 0.93 15-Sep-97
5456 ----------------------
5457
5458 1. /(b)|(:+)/ was computing an incorrect first character.
5459
5460 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5461 but not actually doing anything yet.
5462
5463 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5464 as Perl does (e.g. [-az] or [az-]).
5465
5466 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5467 all possible positions.
5468
5469 5. Split up into different modules to avoid including unneeded functions in a
5470 compiled binary. However, compile and exec are still in one module. The "study"
5471 function is split off.
5472
5473 6. The character tables are now in a separate module whose source is generated
5474 by an auxiliary program - but can then be edited by hand if required. There are
5475 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5476 toupper() in the code.
5477
5478 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5479 make them global. Abolish the function for setting them, as the caller can now
5480 set them directly.
5481
5482
5483 Version 0.92 11-Sep-97
5484 ----------------------
5485
5486 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5487 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5488
5489 2. Caseless matching was not working in character classes if the characters in
5490 the pattern were in upper case.
5491
5492 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5493
5494 4. Make PCRE_ANCHORED public and accept as a compile option.
5495
5496 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5497 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5498 pass them.
5499
5500 6. Give an error if bad option bits passed at compile or run time.
5501
5502 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5503 pcretest to cause it to pass that flag.
5504
5505 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5506 options, and the first character, if set.
5507
5508 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5509
5510
5511 Version 0.91 10-Sep-97
5512 ----------------------
5513
5514 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5515 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5516
5517 2. PCRE was looping on encountering an indefinitely repeated back reference to
5518 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5519 Perl does - treats the match as successful.
5520
5521 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5