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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1451 - (show annotations)
Tue Jan 21 01:49:14 2014 UTC (5 years, 10 months ago) by zherczeg
File size: 263074 byte(s)
Error occurred while calculating annotation data.
Little endian PowerPC systems are supported now by the JIT compiler.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.35-RC1 xx-xxxx-201x
5 -----------------------------
6
7 1. A new flag is set, when property checks are present in an XCLASS.
8 When this flag is not set, PCRE can perform certain optimizations
9 such as studying these XCLASS-es.
10
11 2. The auto-possessification of character sets were improved: a normal
12 and an extended character set can be compared now. Furthermore
13 the JIT compiler optimizes more character set checks.
14
15 3. Got rid of some compiler warnings for potentially uninitialized variables
16 that show up only when compiled with -O2.
17
18 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
19 of a match later then the end of the match. The pcretest program was not
20 handling the case sensibly - it was outputting from the start to the next
21 binary zero. It now reports this situation in a message, and outputs the
22 text from the end to the start.
23
24 5. Fast forward search is improved in JIT. Instead of the first three
25 characters, any three characters with fixed position can be searched.
26 Search order: first, last, middle.
27
28 6. Improve character range checks in JIT. Characters are read by an inprecise
29 function now, which returns with an unknown value if the character code is
30 above a certain treshold (e.g: 256). The only limitation is that the value
31 must be bigger than the treshold as well. This function is useful, when
32 the characters above the treshold are handled in the same way.
33
34 7. The macros whose names start with RAWUCHAR are placeholders for a future
35 mode in which only the bottom 21 bits of 32-bit data items are used. To
36 make this more memorable for those maintaining the code, the names have
37 been changed to start with UCHAR21, and an extensive comment has been added
38 to their definition.
39
40 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
41 to the export list in Makefile.am (they were accidentally omitted from the
42 8.34 tarball).
43
44 9. The informational output from pcretest used the phrase "starting byte set"
45 which is inappropriate for the 16-bit and 32-bit libraries. As the output
46 for "first char" and "need char" really means "non-UTF-char", I've changed
47 "byte" to "char", and slightly reworded the output. The documentation about
48 these values has also been (I hope) clarified.
49
50 10. Another JIT related optimization: use table jumps for selecting the correct
51 backtracking path, when more than four alternatives are present inside a
52 bracket.
53
54 11. Empty match is not possible, when the minimum length is greater than zero,
55 and there is no \K in the pattern. JIT should avoid empty match checks in
56 such cases.
57
58 12. In a caseless character class with UCP support, when a character with more
59 than one alternative case was not the first character of a range, not all
60 the alternative cases were added to the class. For example, s and \x{17f}
61 are both alternative cases for S: the class [RST] was handled correctly,
62 but [R-T] was not.
63
64 13. The configure.ac file always checked for pthread support when JIT was
65 enabled. This is not used in Windows, so I have put this test inside a
66 check for the presence of windows.h (which was already tested for).
67
68 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
69 The algorithm provides a way to skip certain starting offsets, and usually
70 faster than linear prefix searches.
71
72 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
73 as for 'fr_FR' and 'french'. For some reason, however, it then used the
74 Windows-specific input and output files, which have 'french' screwed in.
75 So this could never have worked. One of the problems with locales is that
76 they aren't always the same. I have now updated RunTest so that it checks
77 the output of the locale test (test 3) against three different output
78 files, and it allows the test to pass if any one of them matches. With luck
79 this should make the test pass on some versions of Solaris where it was
80 failing. Because of the uncertainty, the script did not used to stop if
81 test 3 failed; it now does. If further versions of a French locale ever
82 come to light, they can now easily be added.
83
84 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
85 there was a message during ./configure, but it did not stop. This now
86 provokes an error. The invalid example in README has been corrected.
87 If a value less than the minimum is given, the minimum value has always
88 been used, but now a warning is given.
89
90 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
91 was a bug in the test system, which is now fixed. Also, the list of various
92 configurations that are tested for each release did not have one with both
93 16/32 bits and --enable-bar-anycrlf. It now does.
94
95 18. pcretest was missing "-C bsr" for displaying the \R default setting.
96
97 19. Little endian PowerPC systems are supported now by the JIT compiler.
98
99
100 Version 8.34 15-December-2013
101 -----------------------------
102
103 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
104 executable memory. Patch inspired by Carsten Klein.
105
106 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
107 this macro is never tested and has no effect, because the work to support
108 coverage involves only compiling and linking options and special targets in
109 the Makefile. The comment in config.h implied that defining the macro would
110 enable coverage support, which is totally false. There was also support for
111 setting this macro in the CMake files (my fault, I just copied it from
112 configure). SUPPORT_GCOV has now been removed.
113
114 3. Make a small performance improvement in strlen16() and strlen32() in
115 pcretest.
116
117 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
118 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
119 these cases). There was also one in pcretest.c.
120
121 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
122
123 6. In UTF mode, the code for checking whether a group could match an empty
124 string (which is used for indefinitely repeated groups to allow for
125 breaking an infinite loop) was broken when the group contained a repeated
126 negated single-character class with a character that occupied more than one
127 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
128 UTF-8 mode). The effect was undefined: the group might or might not be
129 deemed as matching an empty string, or the program might have crashed.
130
131 7. The code for checking whether a group could match an empty string was not
132 recognizing that \h, \H, \v, \V, and \R must match a character.
133
134 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
135 an empty string. If it can, pcretest shows this in its information output.
136
137 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
138 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
139 matched by pcre_exec() without using JIT:
140
141 (a) If the rest of the pattern did not match after a maximal run of
142 grapheme clusters, the code for backing up to try with fewer of them
143 did not always back up over a full grapheme when characters that do not
144 have the modifier quality were involved, e.g. Hangul syllables.
145
146 (b) If the match point in a subject started with modifier character, and
147 there was no match, the code could incorrectly back up beyond the match
148 point, and potentially beyond the first character in the subject,
149 leading to a segfault or an incorrect match result.
150
151 10. A conditional group with an assertion condition could lead to PCRE
152 recording an incorrect first data item for a match if no other first data
153 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
154 first data item, and therefore matched "ca" after "c" instead of at the
155 start.
156
157 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
158 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
159
160 12. The source of pcregrep now includes z/OS-specific code so that it can be
161 compiled for z/OS as part of the special z/OS distribution.
162
163 13. Added the -T and -TM options to pcretest.
164
165 14. The code in pcre_compile.c for creating the table of named capturing groups
166 has been refactored. Instead of creating the table dynamically during the
167 actual compiling pass, the information is remembered during the pre-compile
168 pass (on the stack unless there are more than 20 named groups, in which
169 case malloc() is used) and the whole table is created before the actual
170 compile happens. This has simplified the code (it is now nearly 150 lines
171 shorter) and prepared the way for better handling of references to groups
172 with duplicate names.
173
174 15. A back reference to a named subpattern when there is more than one of the
175 same name now checks them in the order in which they appear in the pattern.
176 The first one that is set is used for the reference. Previously only the
177 first one was inspected. This change makes PCRE more compatible with Perl.
178
179 16. Unicode character properties were updated from Unicode 6.3.0.
180
181 17. The compile-time code for auto-possessification has been refactored, based
182 on a patch by Zoltan Herczeg. It now happens after instead of during
183 compilation. The code is cleaner, and more cases are handled. The option
184 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
185 options in pcretest are provided to set it. It can also be set by
186 (*NO_AUTO_POSSESS) at the start of a pattern.
187
188 18. The character VT has been added to the default ("C" locale) set of
189 characters that match \s and are generally treated as white space,
190 following this same change in Perl 5.18. There is now no difference between
191 "Perl space" and "POSIX space". Whether VT is treated as white space in
192 other locales depends on the locale.
193
194 19. The code for checking named groups as conditions, either for being set or
195 for being recursed, has been refactored (this is related to 14 and 15
196 above). Processing unduplicated named groups should now be as fast at
197 numerical groups, and processing duplicated groups should be faster than
198 before.
199
200 20. Two patches to the CMake build system, by Alexander Barkov:
201
202 (1) Replace the "source" command by "." in CMakeLists.txt because
203 "source" is a bash-ism.
204
205 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
206 without these the CMake build does not work on Solaris.
207
208 21. Perl has changed its handling of \8 and \9. If there is no previously
209 encountered capturing group of those numbers, they are treated as the
210 literal characters 8 and 9 instead of a binary zero followed by the
211 literals. PCRE now does the same.
212
213 22. Following Perl, added \o{} to specify codepoints in octal, making it
214 possible to specify values greater than 0777 and also making them
215 unambiguous.
216
217 23. Perl now gives an error for missing closing braces after \x{... instead of
218 treating the string as literal. PCRE now does the same.
219
220 24. RunTest used to grumble if an inappropriate test was selected explicitly,
221 but just skip it when running all tests. This make it awkward to run ranges
222 of tests when one of them was inappropriate. Now it just skips any
223 inappropriate tests, as it always did when running all tests.
224
225 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
226 character types such as \d or \w, too many callouts were inserted, and the
227 data that they returned was rubbish.
228
229 26. In UCP mode, \s was not matching two of the characters that Perl matches,
230 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
231 were matched by \h. The code has now been refactored so that the lists of
232 the horizontal and vertical whitespace characters used for \h and \v (which
233 are defined only in one place) are now also used for \s.
234
235 27. Add JIT support for the 64 bit TileGX architecture.
236 Patch by Jiong Wang (Tilera Corporation).
237
238 28. Possessive quantifiers for classes (both explicit and automatically
239 generated) now use special opcodes instead of wrapping in ONCE brackets.
240
241 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
242 (because it's meaningless), this was not happening when PCRE_CASELESS was
243 set. Not wrong, but inefficient.
244
245 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
246 properties for \w, \d, etc) is present in a test regex. Otherwise if the
247 test contains no characters greater than 255, Perl doesn't realise it
248 should be using Unicode semantics.
249
250 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
251 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
252 does in Unicode mode.
253
254 32. Added the "forbid" facility to pcretest so that putting tests into the
255 wrong test files can sometimes be quickly detected.
256
257 33. There is now a limit (default 250) on the depth of nesting of parentheses.
258 This limit is imposed to control the amount of system stack used at compile
259 time. It can be changed at build time by --with-parens-nest-limit=xxx or
260 the equivalent in CMake.
261
262 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
263 errors. Perl warns for these when in warning mode, but PCRE has no facility
264 for giving warnings.
265
266 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
267 However, this was not working for (?!) because it is optimized to (*FAIL),
268 for which PCRE does not allow quantifiers. The optimization is now disabled
269 when a quantifier follows (?!). I can't see any use for this, but it makes
270 things uniform.
271
272 36. Perl no longer allows group names to start with digits, so I have made this
273 change also in PCRE. It simplifies the code a bit.
274
275 37. In extended mode, Perl ignores spaces before a + that indicates a
276 possessive quantifier. PCRE allowed a space before the quantifier, but not
277 before the possessive +. It now does.
278
279 38. The use of \K (reset reported match start) within a repeated possessive
280 group such as (a\Kb)*+ was not working.
281
282 40. Document that the same character tables must be used at compile time and
283 run time, and that the facility to pass tables to pcre_exec() and
284 pcre_dfa_exec() is for use only with saved/restored patterns.
285
286 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
287 features for Builds with MSVC:
288
289 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
290 builds). The .rc files can be used to set FileDescription and many other
291 attributes.
292
293 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
294 This allows higher-level build scripts which want .pdb files to avoid
295 hard-coding the exact files needed."
296
297 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
298 mean "start of word" and "end of word", respectively, as a transition aid.
299
300 43. A minimizing repeat of a class containing codepoints greater than 255 in
301 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
302 compiled to use the heap for recursion.
303
304 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
305 is configured.
306
307
308 Version 8.33 28-May-2013
309 ------------------------
310
311 1. Added 'U' to some constants that are compared to unsigned integers, to
312 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
313 variables that are added to signed variables, to ensure the result is
314 signed and can be negated.
315
316 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
317 CMake config files.
318
319 3. Revise the creation of config.h.generic so that all boolean macros are
320 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
321 overriding via -D on the command line possible.
322
323 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
324 to unsigned int is reported to make a quite noticeable speed difference in
325 a specific Windows environment. Testing on Linux did also appear to show
326 some benefit (and it is clearly not harmful). Also fixed the definition of
327 Xop which should be unsigned.
328
329 5. Related to (4), changing the definition of the intermediate variable cc
330 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
331 improvements.
332
333 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
334 unnecessary spaces.
335
336 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
337 and later.
338
339 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
340
341 9. Optimizing fast_forward_start_bits in JIT.
342
343 10. Adding support for callouts in JIT, and fixing some issues revealed
344 during this work. Namely:
345
346 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
347
348 (b) Minimum length was not checked before the matching is started.
349
350 11. The value of capture_last that is passed to callouts was incorrect in some
351 cases when there was a capture on one path that was subsequently abandoned
352 after a backtrack. Also, the capture_last value is now reset after a
353 recursion, since all captures are also reset in this case.
354
355 12. The interpreter no longer returns the "too many substrings" error in the
356 case when an overflowing capture is in a branch that is subsequently
357 abandoned after a backtrack.
358
359 13. In the pathological case when an offset vector of size 2 is used, pcretest
360 now prints out the matched string after a yield of 0 or 1.
361
362 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
363 Only supported by the JIT compiler at the moment.
364
365 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
366
367 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
368 offset of the starting point of the matching process, provided the offsets
369 vector is large enough.
370
371 17. The \A escape now records a lookbehind value of 1, though its execution
372 does not actually inspect the previous character. This is to ensure that,
373 in partial multi-segment matching, at least one character from the old
374 segment is retained when a new segment is processed. Otherwise, if there
375 are no lookbehinds in the pattern, \A might match incorrectly at the start
376 of a new segment.
377
378 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
379
380 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
381 gives some modest performance improvement in 8-bit mode.
382
383 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
384 be expressed in certain programming languages using Universal Character
385 Names.
386
387 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
388 which points out that "non characters" are not "characters that may not
389 appear in Unicode strings" but rather "characters that are reserved for
390 internal use and have only local meaning".
391
392 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
393 there was a conditional group that depended on an assertion, if the
394 assertion was false, the callout that immediately followed the alternation
395 in the condition was skipped when pcre_exec() was used for matching.
396
397 23. Allow an explicit callout to be inserted before an assertion that is the
398 condition for a conditional group, for compatibility with automatic
399 callouts, which always insert a callout at this point.
400
401 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
402 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
403
404 25. (*PRUNE) is now supported by the JIT compiler.
405
406 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
407
408 27. Fix the case where there are two or more SKIPs with arguments that may be
409 ignored.
410
411 28. (*SKIP) is now supported by the JIT compiler.
412
413 29. (*THEN) is now supported by the JIT compiler.
414
415 30. Update RunTest with additional test selector options.
416
417 31. The way PCRE handles backtracking verbs has been changed in two ways.
418
419 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
420 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
421 backtracking. In some cases this makes it more Perl-compatible, but Perl's
422 rather obscure rules do not always do the same thing.
423
424 (2) Previously, backtracking verbs were confined within assertions. This is
425 no longer the case for positive assertions, except for (*ACCEPT). Again,
426 this sometimes improves Perl compatibility, and sometimes does not.
427
428 32. A number of tests that were in test 2 because Perl did things differently
429 have been moved to test 1, because either Perl or PCRE has changed, and
430 these tests are now compatible.
431
432 32. Backtracking control verbs are now handled in the same way in JIT and
433 interpreter.
434
435 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
436 contained a forward subroutine reference caused a compile error.
437
438 34. Auto-detect and optimize limited repetitions in JIT.
439
440 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
441 blocking (*UTF) etc.
442
443 36. In the interpreter, maximizing pattern repetitions for characters and
444 character types now use tail recursion, which reduces stack usage.
445
446 37. The value of the max lookbehind was not correctly preserved if a compiled
447 and saved regex was reloaded on a host of different endianness.
448
449 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
450 of the compiled pattern block, expand the flags field from 16 to 32 bits
451 because it was almost full.
452
453 39. Try madvise first before posix_madvise.
454
455 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
456 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
457 original change locked out all matches of empty strings. This has been
458 changed so that one match of an empty string per line is recognized.
459 Subsequent searches on the same line (for colouring or for --only-matching,
460 for example) do not recognize empty strings.
461
462 41. Applied a user patch to fix a number of spelling mistakes in comments.
463
464 42. Data lines longer than 65536 caused pcretest to crash.
465
466 43. Clarified the data type for length and startoffset arguments for pcre_exec
467 and pcre_dfa_exec in the function-specific man pages, where they were
468 explicitly stated to be in bytes, never having been updated. I also added
469 some clarification to the pcreapi man page.
470
471 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
472 a segmentation fault.
473
474
475 Version 8.32 30-November-2012
476 -----------------------------
477
478 1. Improved JIT compiler optimizations for first character search and single
479 character iterators.
480
481 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
482 Patch by Daniel Richard G.
483
484 3. Single character iterator optimizations in the JIT compiler.
485
486 4. Improved JIT compiler optimizations for character ranges.
487
488 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
489 Reported by Giuseppe D'Angelo.
490
491 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
492 of a line, was being set incorrectly in cases where .* appeared inside
493 atomic brackets at the start of a pattern, or where there was a subsequent
494 *PRUNE or *SKIP.
495
496 7. Improved instruction cache flush for POWER/PowerPC.
497 Patch by Daniel Richard G.
498
499 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
500 grep:
501
502 (a) There is now no limit to the number of patterns to be matched.
503
504 (b) An error is given if a pattern is too long.
505
506 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
507 are now supported.
508
509 (d) --exclude-from and --include-from (multiple use) have been added.
510
511 (e) Exclusions and inclusions now apply to all files and directories, not
512 just to those obtained from scanning a directory recursively.
513
514 (f) Multiple uses of -f and --file-list are now supported.
515
516 (g) In a Windows environment, the default for -d has been changed from
517 "read" (the GNU grep default) to "skip", because otherwise the presence
518 of a directory in the file list provokes an error.
519
520 (h) The documentation has been revised and clarified in places.
521
522 9. Improve the matching speed of capturing brackets.
523
524 10. Changed the meaning of \X so that it now matches a Unicode extended
525 grapheme cluster.
526
527 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
528 out POSIX threads when JIT support is configured.
529
530 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
531
532 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
533 structure could go wrong in environments where size_t is not the same size
534 as int.
535
536 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
537 set.
538
539 15. The EBCDIC support had decayed; later updates to the code had included
540 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
541 general tidy up of EBCDIC-related issues, and the documentation was also
542 not quite right. There is now a test that can be run on ASCII systems to
543 check some of the EBCDIC-related things (but is it not a full test).
544
545 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
546 in a small tidy to the code.
547
548 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
549
550 18. If the --only-matching (-o) option in pcregrep is specified multiple
551 times, each one causes appropriate output. For example, -o1 -o2 outputs the
552 substrings matched by the 1st and 2nd capturing parentheses. A separating
553 string can be specified by --om-separator (default empty).
554
555 19. Improving the first n character searches.
556
557 20. Turn case lists for horizontal and vertical white space into macros so that
558 they are defined only once.
559
560 21. This set of changes together give more compatible Unicode case-folding
561 behaviour for characters that have more than one other case when UCP
562 support is available.
563
564 (a) The Unicode property table now has offsets into a new table of sets of
565 three or more characters that are case-equivalent. The MultiStage2.py
566 script that generates these tables (the pcre_ucd.c file) now scans
567 CaseFolding.txt instead of UnicodeData.txt for character case
568 information.
569
570 (b) The code for adding characters or ranges of characters to a character
571 class has been abstracted into a generalized function that also handles
572 case-independence. In UTF-mode with UCP support, this uses the new data
573 to handle characters with more than one other case.
574
575 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
576 whose other case is greater than 256 are now correctly matched
577 caselessly. Previously, the high codepoint matched the low one, but not
578 vice versa.
579
580 (d) The processing of \h, \H, \v, and \ in character classes now makes use
581 of the new class addition function, using character lists defined as
582 macros alongside the case definitions of 20 above.
583
584 (e) Caseless back references now work with characters that have more than
585 one other case.
586
587 (f) General caseless matching of characters with more than one other case
588 is supported.
589
590 22. Unicode character properties were updated from Unicode 6.2.0
591
592 23. Improved CMake support under Windows. Patch by Daniel Richard G.
593
594 24. Add support for 32-bit character strings, and UTF-32
595
596 25. Major JIT compiler update (code refactoring and bugfixing).
597 Experimental Sparc 32 support is added.
598
599 26. Applied a modified version of Daniel Richard G's patch to create
600 pcre.h.generic and config.h.generic by "make" instead of in the
601 PrepareRelease script.
602
603 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
604 pcre_compile.c when checking for a zero character.
605
606 28. Introducing a native interface for JIT. Through this interface, the compiled
607 machine code can be directly executed. The purpose of this interface is to
608 provide fast pattern matching, so several sanity checks are not performed.
609 However, feature tests are still performed. The new interface provides
610 1.4x speedup compared to the old one.
611
612 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
613 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
614 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
615
616 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
617 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
618 the "old" RFC 2279). Instead, it ended up passing a negative length to
619 pcre_exec().
620
621 31. Add support for GCC's visibility feature to hide internal functions.
622
623 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
624 "unknown -C option" after outputting 0 or 1.
625
626 33. There is now support for generating a code coverage report for the test
627 suite in environments where gcc is the compiler and lcov is installed. This
628 is mainly for the benefit of the developers.
629
630 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
631 unaddressable using valgrind annotations, allowing valgrind to detect
632 invalid memory accesses. This is mainly for the benefit of the developers.
633
634 25. (*UTF) can now be used to start a pattern in any of the three libraries.
635
636 26. Give configure error if --enable-cpp but no C++ compiler found.
637
638
639 Version 8.31 06-July-2012
640 -------------------------
641
642 1. Fixing a wrong JIT test case and some compiler warnings.
643
644 2. Removed a bashism from the RunTest script.
645
646 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
647 to unsigned type, result still unsigned" that was given by an MS compiler
648 on encountering the code "-sizeof(xxx)".
649
650 4. Partial matching support is added to the JIT compiler.
651
652 5. Fixed several bugs concerned with partial matching of items that consist
653 of more than one character:
654
655 (a) /^(..)\1/ did not partially match "aba" because checking references was
656 done on an "all or nothing" basis. This also applied to repeated
657 references.
658
659 (b) \R did not give a hard partial match if \r was found at the end of the
660 subject.
661
662 (c) \X did not give a hard partial match after matching one or more
663 characters at the end of the subject.
664
665 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
666 a partial match for the string "\r".
667
668 (e) When newline was set to CRLF, the metacharacter "." did not recognize
669 a partial match for a CR character at the end of the subject string.
670
671 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
672 running pcretest, the text "(JIT)" added to the output whenever JIT is
673 actually used to run the match.
674
675 7. Individual JIT compile options can be set in pcretest by following -s+[+]
676 or /S+[+] with a digit between 1 and 7.
677
678 8. OP_NOT now supports any UTF character not just single-byte ones.
679
680 9. (*MARK) control verb is now supported by the JIT compiler.
681
682 10. The command "./RunTest list" lists the available tests without actually
683 running any of them. (Because I keep forgetting what they all are.)
684
685 11. Add PCRE_INFO_MAXLOOKBEHIND.
686
687 12. Applied a (slightly modified) user-supplied patch that improves performance
688 when the heap is used for recursion (compiled with --disable-stack-for-
689 recursion). Instead of malloc and free for each heap frame each time a
690 logical recursion happens, frames are retained on a chain and re-used where
691 possible. This sometimes gives as much as 30% improvement.
692
693 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
694 call.
695
696 14. As documented, (*COMMIT) is now confined to within a positive assertion.
697
698 15. It is now possible to link pcretest with libedit as an alternative to
699 libreadline.
700
701 16. (*COMMIT) control verb is now supported by the JIT compiler.
702
703 17. The Unicode data tables have been updated to Unicode 6.1.0.
704
705 18. Added --file-list option to pcregrep.
706
707 19. Added binary file support to pcregrep, including the -a, --binary-files,
708 -I, and --text options.
709
710 20. The madvise function is renamed for posix_madvise for QNX compatibility
711 reasons. Fixed by Giuseppe D'Angelo.
712
713 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
714
715 22. Changed the option for creating symbolic links for 16-bit man pages from
716 -s to -sf so that re-installing does not cause issues.
717
718 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
719
720 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
721 matches in certain environments (the workspace was not being correctly
722 retained). Also added to pcre_dfa_exec() a simple plausibility check on
723 some of the workspace data at the beginning of a restart.
724
725 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
726 was not doing so when it should - probably a typo introduced by SVN 528
727 (change 8.10/14).
728
729 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
730 \w+ when the character tables indicated that \x{c4} was a word character.
731 There were several related cases, all because the tests for doing a table
732 lookup were testing for characters less than 127 instead of 255.
733
734 27. If a pattern contains capturing parentheses that are not used in a match,
735 their slots in the ovector are set to -1. For those that are higher than
736 any matched groups, this happens at the end of processing. In the case when
737 there were back references that the ovector was too small to contain
738 (causing temporary malloc'd memory to be used during matching), and the
739 highest capturing number was not used, memory off the end of the ovector
740 was incorrectly being set to -1. (It was using the size of the temporary
741 memory instead of the true size.)
742
743 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
744 ovector size, it uses memory at the end of the block that it has got.
745
746 29. Check for an overlong MARK name and give an error at compile time. The
747 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
748
749 30. JIT compiler update.
750
751 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
752 Rill for the patch.
753
754 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
755
756 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
757
758 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
759 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
760 was enabled.
761
762 35. Improve JIT code generation for greedy plus quantifier.
763
764 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
765 1 to "aa" instead of to an empty string. The bug affected repeated groups
766 that could potentially match an empty string.
767
768 37. Optimizing single character iterators in JIT.
769
770 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
771 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
772 codepoints that are too big for the mode are faulted, and in a UTF mode,
773 disallowed codepoints are also faulted.
774
775 39. If PCRE was compiled with UTF support, in three places in the DFA
776 matcher there was code that should only have been obeyed in UTF mode, but
777 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
778 processing when bytes with values greater than 127 were present. In 16-bit
779 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
780 both cases the values are those that cannot be the first data item in a UTF
781 character. The three items that might have provoked this were recursions,
782 possessively repeated groups, and atomic groups.
783
784 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
785 and pcregrep, because some OS require shared objects to be explicitly
786 passed to ld, causing the link step to fail if they are not.
787
788 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
789 mode, patterns that started with \h* or \R* might be incorrectly matched.
790
791
792 Version 8.30 04-February-2012
793 -----------------------------
794
795 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
796 name is defined in ctype.h.
797
798 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
799 only in quite long subpatterns.
800
801 3. Removed the function pcre_info(), which has been obsolete and deprecated
802 since it was replaced by pcre_fullinfo() in February 2000.
803
804 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
805 match a (*MARK), and the match failed at the start of the subject, a
806 reference to memory before the start of the subject could occur. This bug
807 was introduced by fix 17 of release 8.21.
808
809 5. A reference to an unset group with zero minimum repetition was giving
810 totally wrong answers (in non-JavaScript-compatibility mode). For example,
811 /(another)?(\1?)test/ matched against "hello world test". This bug was
812 introduced in release 8.13.
813
814 6. Add support for 16-bit character strings (a large amount of work involving
815 many changes and refactorings).
816
817 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
818 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
819 from a file.
820
821 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
822 rounding is not applied in this particular case).
823
824 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
825 if they appear, or are escaped, in patterns.
826
827 10. Get rid of a number of -Wunused-but-set-variable warnings.
828
829 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
830 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
831 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
832 also returns the mark "x". This bug applied to capturing parentheses,
833 non-capturing parentheses, and atomic parentheses. It also applied to some
834 assertions.
835
836 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
837 information out of configure.ac instead of relying on pcre.h.generic, which
838 is not stored in the repository.
839
840 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
841 -lreadline.
842
843 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
844
845 15. Applied Graycode's patch to put the top-level frame on the stack rather
846 than the heap when not using the stack for recursion. This gives a
847 performance improvement in many cases when recursion is not deep.
848
849 16. Experimental code added to "pcretest -C" to output the stack frame size.
850
851
852 Version 8.21 12-Dec-2011
853 ------------------------
854
855 1. Updating the JIT compiler.
856
857 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
858 are added as well.
859
860 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
861 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
862 calling _pcre_jit_exec. Some extra comments are added.
863
864 4. (*MARK) settings inside atomic groups that do not contain any capturing
865 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
866 was introduced by change 18 for 8.20.
867
868 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
869 ECMA-262 standard.
870
871 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
872 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
873 This bug was probably introduced by change 9 of 8.13.
874
875 7. While fixing 6 above, I noticed that a number of other items were being
876 incorrectly rejected as "not fixed length". This arose partly because newer
877 opcodes had not been added to the fixed-length checking code. I have (a)
878 corrected the bug and added tests for these items, and (b) arranged for an
879 error to occur if an unknown opcode is encountered while checking for fixed
880 length instead of just assuming "not fixed length". The items that were
881 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
882 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
883 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
884
885 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
886 being incorrectly compiled and would have given unpredicatble results.
887
888 9. A possessively repeated subpattern with minimum repeat count greater than
889 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
890 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
891 the first (A) could occur when it should not.
892
893 10. Add a cast and remove a redundant test from the code.
894
895 11. JIT should use pcre_malloc/pcre_free for allocation.
896
897 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
898 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
899 is anything other than /usr, -L is still shown).
900
901 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
902
903 14. Perl does not support \N without a following name in a [] class; PCRE now
904 also gives an error.
905
906 15. If a forward reference was repeated with an upper limit of around 2000,
907 it caused the error "internal error: overran compiling workspace". The
908 maximum number of forward references (including repeats) was limited by the
909 internal workspace, and dependent on the LINK_SIZE. The code has been
910 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
911 the default depends on LINK_SIZE. There is a new upper limit (for safety)
912 of around 200,000 forward references. While doing this, I also speeded up
913 the filling in of repeated forward references.
914
915 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
916 incorrectly expecting the subject to contain another "a" after the start.
917
918 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
919 in the match, the SKIP should be ignored. This was not happening; instead
920 the SKIP was being treated as NOMATCH. For patterns such as
921 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
922 tested.
923
924 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
925 now much more compatible with Perl, in particular in cases where the result
926 is a non-match for a non-anchored pattern. For example, if
927 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
928 "m", where previously it did not return a name. A side effect of this
929 change is that for partial matches, the last encountered mark name is
930 returned, as for non matches. A number of tests that were previously not
931 Perl-compatible have been moved into the Perl-compatible test files. The
932 refactoring has had the pleasing side effect of removing one argument from
933 the match() function, thus reducing its stack requirements.
934
935 19. If the /S+ option was used in pcretest to study a pattern using JIT,
936 subsequent uses of /S (without +) incorrectly behaved like /S+.
937
938 21. Retrieve executable code size support for the JIT compiler and fixing
939 some warnings.
940
941 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
942 not work when the shorter character appeared right at the end of the
943 subject string.
944
945 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
946 systems.
947
948 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
949 output it when the /M option is used in pcretest.
950
951 25. The CheckMan script was not being included in the distribution. Also, added
952 an explicit "perl" to run Perl scripts from the PrepareRelease script
953 because this is reportedly needed in Windows.
954
955 26. If study data was being save in a file and studying had not found a set of
956 "starts with" bytes for the pattern, the data written to the file (though
957 never used) was taken from uninitialized memory and so caused valgrind to
958 complain.
959
960 27. Updated RunTest.bat as provided by Sheri Pierce.
961
962 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
963
964 29. Computation of memory usage for the table of capturing group names was
965 giving an unnecessarily large value.
966
967
968 Version 8.20 21-Oct-2011
969 ------------------------
970
971 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
972 a POSIX class. After further experiments with Perl, which convinced me that
973 Perl has bugs and confusions, a closing square bracket is no longer allowed
974 in a POSIX name. This bug also affected patterns with classes that started
975 with full stops.
976
977 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
978 captured substring, but while checking the failing first alternative,
979 substring 1 is temporarily captured. If the output vector supplied to
980 pcre_exec() was not big enough for this capture, the yield of the function
981 was still zero ("insufficient space for captured substrings"). This cannot
982 be totally fixed without adding another stack variable, which seems a lot
983 of expense for a edge case. However, I have improved the situation in cases
984 such as /(a)(b)x|abc/ matched against "abc", where the return code
985 indicates that fewer than the maximum number of slots in the ovector have
986 been set.
987
988 3. Related to (2) above: when there are more back references in a pattern than
989 slots in the output vector, pcre_exec() uses temporary memory during
990 matching, and copies in the captures as far as possible afterwards. It was
991 using the entire output vector, but this conflicts with the specification
992 that only 2/3 is used for passing back captured substrings. Now it uses
993 only the first 2/3, for compatibility. This is, of course, another edge
994 case.
995
996 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
997 main code base, and can be used by building with --enable-jit. When this is
998 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
999 runtime --no-jit option is given.
1000
1001 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1002 ovector, the return from the function was zero, implying that there were
1003 other matches that did not fit. The correct "exactly full" value is now
1004 returned.
1005
1006 6. If a subpattern that was called recursively or as a subroutine contained
1007 (*PRUNE) or any other control that caused it to give a non-standard return,
1008 invalid errors such as "Error -26 (nested recursion at the same subject
1009 position)" or even infinite loops could occur.
1010
1011 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1012 computing the minimum length on reaching *ACCEPT, and so ended up with the
1013 wrong value of 1 rather than 0. Further investigation indicates that
1014 computing a minimum subject length in the presence of *ACCEPT is difficult
1015 (think back references, subroutine calls), and so I have changed the code
1016 so that no minimum is registered for a pattern that contains *ACCEPT.
1017
1018 8. If (*THEN) was present in the first (true) branch of a conditional group,
1019 it was not handled as intended. [But see 16 below.]
1020
1021 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1022 Sheri Pierce.
1023
1024 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1025 the first byte in a match must be "a".
1026
1027 11. Change 17 for 8.13 increased the recursion depth for patterns like
1028 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1029 pattern contains any instances of (*THEN). If it does not, the old
1030 optimizations are restored. It would be nice to do this on a per-group
1031 basis, but at the moment that is not feasible.
1032
1033 12. In some environments, the output of pcretest -C is CRLF terminated. This
1034 broke RunTest's code that checks for the link size. A single white space
1035 character after the value is now allowed for.
1036
1037 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1038 For "fr", it uses the Windows-specific input and output files.
1039
1040 14. If (*THEN) appeared in a group that was called recursively or as a
1041 subroutine, it did not work as intended. [But see next item.]
1042
1043 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1044 pattern fragments (but not containing any | characters). If A and B are
1045 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1046 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1047 D. In other words, Perl considers parentheses that do not contain any |
1048 characters to be part of a surrounding alternative, whereas PCRE was
1049 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1050 differently. PCRE now behaves in the same way as Perl, except in the case
1051 of subroutine/recursion calls such as (?1) which have in any case always
1052 been different (but PCRE had them first :-).
1053
1054 16. Related to 15 above: Perl does not treat the | in a conditional group as
1055 creating alternatives. Such a group is treated in the same way as an
1056 ordinary group without any | characters when processing (*THEN). PCRE has
1057 been changed to match Perl's behaviour.
1058
1059 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1060 RunGrepTest script failed.
1061
1062 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1063 inevitable for groups that contain captures, but it can lead to a lot of
1064 stack use in large patterns. The old behaviour has been restored for atomic
1065 groups that do not contain any capturing parentheses.
1066
1067 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1068 suppress the check for a minimum subject length at run time. (If it was
1069 given to pcre_exec() or pcre_dfa_exec() it did work.)
1070
1071 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1072 fail to work when decoding hex characters in data strings in EBCDIC
1073 environments.
1074
1075 21. It appears that in at least one Mac OS environment, the isxdigit() function
1076 is implemented as a macro that evaluates to its argument more than once,
1077 contravening the C 90 Standard (I haven't checked a later standard). There
1078 was an instance in pcretest which caused it to go wrong when processing
1079 \x{...} escapes in subject strings. The has been rewritten to avoid using
1080 things like p++ in the argument of isxdigit().
1081
1082
1083 Version 8.13 16-Aug-2011
1084 ------------------------
1085
1086 1. The Unicode data tables have been updated to Unicode 6.0.0.
1087
1088 2. Two minor typos in pcre_internal.h have been fixed.
1089
1090 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1091 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1092 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1093
1094 4. There were a number of related bugs in the code for matching backrefences
1095 caselessly in UTF-8 mode when codes for the characters concerned were
1096 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1097 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1098 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1099 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1100 2-byte code at the end of the subject (it thought there wasn't enough data
1101 left).
1102
1103 5. Comprehensive information about what went wrong is now returned by
1104 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1105 as the output vector has at least 2 elements. The offset of the start of
1106 the failing character and a reason code are placed in the vector.
1107
1108 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1109 now returned is for the first byte of the failing character, instead of the
1110 last byte inspected. This is an incompatible change, but I hope it is small
1111 enough not to be a problem. It makes the returned offset consistent with
1112 pcre_exec() and pcre_dfa_exec().
1113
1114 7. pcretest now gives a text phrase as well as the error number when
1115 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1116 failure, the offset and reason code are output.
1117
1118 8. When \R was used with a maximizing quantifier it failed to skip backwards
1119 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1120 back over a single character (\n). This seems wrong (because it treated the
1121 two characters as a single entity when going forwards), conflicts with the
1122 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1123 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1124 has been changed.
1125
1126 9. Some internal refactoring has changed the processing so that the handling
1127 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1128 time (the PCRE_DOTALL option was changed this way some time ago: version
1129 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1130 which was always a bit of a fudge. It also means that there is one less
1131 argument for the match() function, which reduces its stack requirements
1132 slightly. This change also fixes an incompatibility with Perl: the pattern
1133 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1134
1135 10. More internal refactoring has drastically reduced the number of recursive
1136 calls to match() for possessively repeated groups such as (abc)++ when
1137 using pcre_exec().
1138
1139 11. While implementing 10, a number of bugs in the handling of groups were
1140 discovered and fixed:
1141
1142 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1143 (a|)*(?1) gave a compile-time internal error.
1144 ((a|)+)+ did not notice that the outer group could match an empty string.
1145 (^a|^)+ was not marked as anchored.
1146 (.*a|.*)+ was not marked as matching at start or after a newline.
1147
1148 12. Yet more internal refactoring has removed another argument from the match()
1149 function. Special calls to this function are now indicated by setting a
1150 value in a variable in the "match data" data block.
1151
1152 13. Be more explicit in pcre_study() instead of relying on "default" for
1153 opcodes that mean there is no starting character; this means that when new
1154 ones are added and accidentally left out of pcre_study(), testing should
1155 pick them up.
1156
1157 14. The -s option of pcretest has been documented for ages as being an old
1158 synonym of -m (show memory usage). I have changed it to mean "force study
1159 for every regex", that is, assume /S for every regex. This is similar to -i
1160 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1161 using it. It makes it easier to run collections of tests with and without
1162 study enabled, and thereby test pcre_study() more easily. All the standard
1163 tests are now run with and without -s (but some patterns can be marked as
1164 "never study" - see 20 below).
1165
1166 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1167 restoration of the capturing data to the outer values was not happening
1168 correctly.
1169
1170 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1171 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1172 pattern had matched an empty string, and so incorrectly returned a no
1173 match.
1174
1175 17. There was optimizing code for the last branch of non-capturing parentheses,
1176 and also for the obeyed branch of a conditional subexpression, which used
1177 tail recursion to cut down on stack usage. Unfortunately, now that there is
1178 the possibility of (*THEN) occurring in these branches, tail recursion is
1179 no longer possible because the return has to be checked for (*THEN). These
1180 two optimizations have therefore been removed. [But see 8.20/11 above.]
1181
1182 18. If a pattern containing \R was studied, it was assumed that \R always
1183 matched two bytes, thus causing the minimum subject length to be
1184 incorrectly computed because \R can also match just one byte.
1185
1186 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1187 was incorrectly computed.
1188
1189 20. If /S is present twice on a test pattern in pcretest input, it now
1190 *disables* studying, thereby overriding the use of -s on the command line
1191 (see 14 above). This is necessary for one or two tests to keep the output
1192 identical in both cases.
1193
1194 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1195 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1196
1197 22. When an atomic group that contained a capturing parenthesis was
1198 successfully matched, but the branch in which it appeared failed, the
1199 capturing was not being forgotten if a higher numbered group was later
1200 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1201 group 1 to "a", when in fact it should be unset. This applied to multi-
1202 branched capturing and non-capturing groups, repeated or not, and also to
1203 positive assertions (capturing in negative assertions does not happen
1204 in PCRE) and also to nested atomic groups.
1205
1206 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1207 subject after a captured substring, to make it easier to tell which of a
1208 number of identical substrings has been captured.
1209
1210 24. The way atomic groups are processed by pcre_exec() has been changed so that
1211 if they are repeated, backtracking one repetition now resets captured
1212 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1213 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1214 "aaa". Previously, it would have been "a". As part of this code
1215 refactoring, the way recursive calls are handled has also been changed.
1216
1217 25. If an assertion condition captured any substrings, they were not passed
1218 back unless some other capturing happened later. For example, if
1219 (?(?=(a))a) was matched against "a", no capturing was returned.
1220
1221 26. When studying a pattern that contained subroutine calls or assertions,
1222 the code for finding the minimum length of a possible match was handling
1223 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1224 group 1 called group 2 while simultaneously a separate group 2 called group
1225 1). A stack overflow occurred in this case. I have fixed this by limiting
1226 the recursion depth to 10.
1227
1228 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1229 Fortmann. This supports explicit test numbers on the command line, and has
1230 argument validation and error reporting.
1231
1232 28. An instance of \X with an unlimited repeat could fail if at any point the
1233 first character it looked at was a mark character.
1234
1235 29. Some minor code refactoring concerning Unicode properties and scripts
1236 should reduce the stack requirement of match() slightly.
1237
1238 30. Added the '=' option to pcretest to check the setting of unused capturing
1239 slots at the end of the pattern, which are documented as being -1, but are
1240 not included in the return count.
1241
1242 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1243 compiled something random. Now it gives a compile-time error (as does
1244 Perl).
1245
1246 32. A *MARK encountered during the processing of a positive assertion is now
1247 recorded and passed back (compatible with Perl).
1248
1249 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1250 had alternative anchored branches, the search for a second match in a line
1251 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1252 matched the line "0102" twice. The same bug affected patterns that started
1253 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1254 twice.
1255
1256 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1257 does, and because of capturing effects, quantifying parenthesized
1258 assertions may at times be useful. Quantifiers are now allowed for
1259 parenthesized assertions.
1260
1261 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1262
1263 36. \g was being checked for fancy things in a character class, when it should
1264 just be a literal "g".
1265
1266 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1267 appearance of a nested POSIX class supersedes an apparent external class.
1268 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1269 unescaped square brackets may also appear as part of class names. For
1270 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1271 more like Perl. (But see 8.20/1 above.)
1272
1273 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1274 was because it thought it was \N{name}, which is not supported).
1275
1276 39. Add minix to OS list not supporting the -S option in pcretest.
1277
1278 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1279 cannot analyze patterns in sufficient detail to catch mutual recursions
1280 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1281 subgroup is called recursively as a subpattern for a second time at the
1282 same position in the subject string. In previous releases this might have
1283 been caught by the recursion limit, or it might have run out of stack.
1284
1285 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1286 happen only once. PCRE was, however incorrectly giving a compile time error
1287 "recursive call could loop indefinitely" because it cannot analyze the
1288 pattern in sufficient detail. The compile time test no longer happens when
1289 PCRE is compiling a conditional subpattern, but actual runaway loops are
1290 now caught at runtime (see 40 above).
1291
1292 42. It seems that Perl allows any characters other than a closing parenthesis
1293 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1294 has been changed to be the same.
1295
1296 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1297 as not to get warnings when autogen.sh is called. Also changed
1298 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1299
1300 44. To help people who use pcregrep to scan files containing exceedingly long
1301 lines, the following changes have been made:
1302
1303 (a) The default value of the buffer size parameter has been increased from
1304 8K to 20K. (The actual buffer used is three times this size.)
1305
1306 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1307 PCRE is built.
1308
1309 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1310 to be set at run time.
1311
1312 (d) Numerical values in pcregrep options can be followed by K or M, for
1313 example --buffer-size=50K.
1314
1315 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1316 given and the return code is set to 2.
1317
1318 45. Add a pointer to the latest mark to the callout data block.
1319
1320 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1321 partial match of an empty string instead of no match. This was specific to
1322 the use of ".".
1323
1324 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1325 complete match instead of a partial match. This bug was dependent on both
1326 the PCRE_UTF8 and PCRE_DOTALL options being set.
1327
1328 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1329 starting byte set, because \b was not being ignored.
1330
1331
1332 Version 8.12 15-Jan-2011
1333 ------------------------
1334
1335 1. Fixed some typos in the markup of the man pages, and wrote a script that
1336 checks for such things as part of the documentation building process.
1337
1338 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1339 --match-limit and --recursion-limit options (added for 8.11). In
1340 particular, this made one of the standard tests fail. (The integer value
1341 went into the wrong half of a long int.)
1342
1343 3. If the --colour option was given to pcregrep with -v (invert match), it
1344 did strange things, either producing crazy output, or crashing. It should,
1345 of course, ignore a request for colour when reporting lines that do not
1346 match.
1347
1348 4. Another pcregrep bug caused similar problems if --colour was specified with
1349 -M (multiline) and the pattern match finished with a line ending.
1350
1351 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1352 matched in multiline mode, the following line was shown as part of the
1353 match. This seems wrong, so I have changed it.
1354
1355 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1356 the check for further matches in the same line (so they could be coloured)
1357 to overrun the end of the current line. If another match was found, it was
1358 incorrectly shown (and then shown again when found in the next line).
1359
1360 7. If pcregrep was compiled under Windows, there was a reference to the
1361 function pcregrep_exit() before it was defined. I am assuming this was
1362 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1363 reported by a user. I've moved the definition above the reference.
1364
1365
1366 Version 8.11 10-Dec-2010
1367 ------------------------
1368
1369 1. (*THEN) was not working properly if there were untried alternatives prior
1370 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1371 backtracked to try for "b" instead of moving to the next alternative branch
1372 at the same level (in this case, to look for "c"). The Perl documentation
1373 is clear that when (*THEN) is backtracked onto, it goes to the "next
1374 alternative in the innermost enclosing group".
1375
1376 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1377 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1378 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1379 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1380 (*THEN).
1381
1382 3. If \s appeared in a character class, it removed the VT character from
1383 the class, even if it had been included by some previous item, for example
1384 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1385 of \s, but is part of the POSIX "space" class.)
1386
1387 4. A partial match never returns an empty string (because you can always
1388 match an empty string at the end of the subject); however the checking for
1389 an empty string was starting at the "start of match" point. This has been
1390 changed to the "earliest inspected character" point, because the returned
1391 data for a partial match starts at this character. This means that, for
1392 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1393 (previously it gave "no match").
1394
1395 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1396 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1397 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1398 has an implication that the given string is incomplete (because a partial
1399 match is preferred over a full match). For this reason, these items now
1400 give a partial match in this situation. [Aside: previously, the one case
1401 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1402 match rather than a full match, which was wrong by the old rules, but is
1403 now correct.]
1404
1405 6. There was a bug in the handling of #-introduced comments, recognized when
1406 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1407 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1408 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1409 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1410 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1411 places in pcre_compile().
1412
1413 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1414 comments when looking ahead for named forward references to subpatterns,
1415 the only newline sequence it recognized was NL. It now handles newlines
1416 according to the set newline convention.
1417
1418 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1419 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1420 cater for a lack of strerror(). These oversights have been fixed.
1421
1422 9. Added --match-limit and --recursion-limit to pcregrep.
1423
1424 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1425
1426 11. When the -o option was used, pcregrep was setting a return code of 1, even
1427 when matches were found, and --line-buffered was not being honoured.
1428
1429 12. Added an optional parentheses number to the -o and --only-matching options
1430 of pcregrep.
1431
1432 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1433 can match an empty string. The code to do it in pcretest and pcredemo
1434 needed fixing:
1435
1436 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1437 only one byte after an empty string match just before CRLF (this case
1438 just got forgotten; "any" and "anycrlf" were OK).
1439
1440 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1441 mode when an empty string match preceded an ASCII character followed by
1442 a non-ASCII character. (The code for advancing by one character rather
1443 than one byte was nonsense.)
1444
1445 (c) The pcredemo.c sample program did not have any code at all to handle
1446 the cases when CRLF is a valid newline sequence.
1447
1448 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1449 as a starting offset was within the subject string. There is now a new
1450 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1451 negative or greater than the length of the string. In order to test this,
1452 pcretest is extended to allow the setting of negative starting offsets.
1453
1454 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1455 starting offset points to the beginning of a UTF-8 character was
1456 unnecessarily clumsy. I tidied it up.
1457
1458 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1459 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1460
1461 17. Nobody had reported that the --include_dir option, which was added in
1462 release 7.7 should have been called --include-dir (hyphen, not underscore)
1463 for compatibility with GNU grep. I have changed it to --include-dir, but
1464 left --include_dir as an undocumented synonym, and the same for
1465 --exclude-dir, though that is not available in GNU grep, at least as of
1466 release 2.5.4.
1467
1468 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1469 characters from a string of bytes) have been redefined so as not to use
1470 loops, in order to improve performance in some environments. At the same
1471 time, I abstracted some of the common code into auxiliary macros to save
1472 repetition (this should not affect the compiled code).
1473
1474 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1475 compile-time error is now given if \c is not followed by an ASCII
1476 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1477 different, and any byte value is allowed.)
1478
1479 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1480 START_OPTIMIZE option, which is now allowed at compile time - but just
1481 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1482 to pcregrep and other applications that have no direct access to PCRE
1483 options. The new /Y option in pcretest sets this option when calling
1484 pcre_compile().
1485
1486 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1487 back references. Groups containing recursive back references were forced to
1488 be atomic by that change, but in the case of named groups, the amount of
1489 memory required was incorrectly computed, leading to "Failed: internal
1490 error: code overflow". This has been fixed.
1491
1492 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1493 pcretest.c, to avoid build problems in some Borland environments.
1494
1495
1496 Version 8.10 25-Jun-2010
1497 ------------------------
1498
1499 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1500 THEN.
1501
1502 2. (*ACCEPT) was not working when inside an atomic group.
1503
1504 3. Inside a character class, \B is treated as a literal by default, but
1505 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1506 causes the error). The code is unchanged, but I tidied the documentation.
1507
1508 4. Inside a character class, PCRE always treated \R and \X as literals,
1509 whereas Perl faults them if its -w option is set. I have changed PCRE so
1510 that it faults them when PCRE_EXTRA is set.
1511
1512 5. Added support for \N, which always matches any character other than
1513 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1514
1515 6. When compiling pcregrep with newer versions of gcc which may have
1516 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1517 declared with attribute warn_unused_result" were given. Just casting the
1518 result to (void) does not stop the warnings; a more elaborate fudge is
1519 needed. I've used a macro to implement this.
1520
1521 7. Minor change to pcretest.c to avoid a compiler warning.
1522
1523 8. Added four artifical Unicode properties to help with an option to make
1524 \s etc use properties (see next item). The new properties are: Xan
1525 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1526
1527 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1528 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1529 this option. Modified pcretest to add /W to test this facility. Added
1530 REG_UCP to make it available via the POSIX interface.
1531
1532 10. Added --line-buffered to pcregrep.
1533
1534 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1535 studied, and the match started with a letter with a code point greater than
1536 127 whose first byte was different to the first byte of the other case of
1537 the letter, the other case of this starting letter was not recognized
1538 (#976).
1539
1540 12. If a pattern that was studied started with a repeated Unicode property
1541 test, for example, \p{Nd}+, there was the theoretical possibility of
1542 setting up an incorrect bitmap of starting bytes, but fortunately it could
1543 not have actually happened in practice until change 8 above was made (it
1544 added property types that matched character-matching opcodes).
1545
1546 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1547 possible starting bytes for non-anchored patterns.
1548
1549 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1550 \R, and also a number of cases that involve Unicode properties, both
1551 explicit and implicit when PCRE_UCP is set.
1552
1553 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1554 input, it could crash or give wrong results if characters with values
1555 greater than 0xc0 were present in the subject string. (Detail: it assumed
1556 UTF-8 input when processing these items.)
1557
1558 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1559 size_t is 64-bit (#991).
1560
1561 17. Added a check for running out of memory when PCRE is compiled with
1562 --disable-stack-for-recursion (#990).
1563
1564 18. If the last data line in a file for pcretest does not have a newline on
1565 the end, a newline was missing in the output.
1566
1567 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1568 less than 128) in its various bitmaps. However, there is a facility for
1569 generating tables according to the current locale when PCRE is compiled. It
1570 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1571 characters, are recognized by isspace() and therefore were getting set in
1572 these tables, and indeed these tables seem to approximate to ISO 8859. This
1573 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1574 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1575 which of course cannot start UTF-8 characters. I have changed the code so
1576 that only real ASCII characters (less than 128) and the correct starting
1577 bytes for UTF-8 encodings are set for characters greater than 127 when in
1578 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1579 altogether.)
1580
1581 20. Added the /T option to pcretest so as to be able to run tests with non-
1582 standard character tables, thus making it possible to include the tests
1583 used for 19 above in the standard set of tests.
1584
1585 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1586 reference to a subpattern the other side of a comment that contains an
1587 opening parenthesis caused either an internal compiling error, or a
1588 reference to the wrong subpattern.
1589
1590
1591 Version 8.02 19-Mar-2010
1592 ------------------------
1593
1594 1. The Unicode data tables have been updated to Unicode 5.2.0.
1595
1596 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1597 configured.
1598
1599 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1600 original author of that file, following a query about its status.
1601
1602 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1603 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1604
1605 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1606 quantifier applied to a forward-referencing subroutine call, could compile
1607 incorrect code or give the error "internal error: previously-checked
1608 referenced subpattern not found".
1609
1610 6. Both MS Visual Studio and Symbian OS have problems with initializing
1611 variables to point to external functions. For these systems, therefore,
1612 pcre_malloc etc. are now initialized to local functions that call the
1613 relevant global functions.
1614
1615 7. There were two entries missing in the vectors called coptable and poptable
1616 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1617 I've fixed the data, and added a kludgy way of testing at compile time that
1618 the lengths are correct (equal to the number of opcodes).
1619
1620 8. Following on from 7, I added a similar kludge to check the length of the
1621 eint vector in pcreposix.c.
1622
1623 9. Error texts for pcre_compile() are held as one long string to avoid too
1624 much relocation at load time. To find a text, the string is searched,
1625 counting zeros. There was no check for running off the end of the string,
1626 which could happen if a new error number was added without updating the
1627 string.
1628
1629 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1630
1631 11. \K was not working if it appeared in an atomic group or in a group that
1632 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1633 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1634 the assertion is positive, but not if it is negative.
1635
1636 12. Change 11 fortuitously reduced the size of the stack frame used in the
1637 "match()" function of pcre_exec.c by one pointer. Forthcoming
1638 implementation of support for (*MARK) will need an extra pointer on the
1639 stack; I have reserved it now, so that the stack frame size does not
1640 decrease.
1641
1642 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1643 item in branch that calls a recursion is a subroutine call - as in the
1644 second branch in the above example - was incorrectly given the compile-
1645 time error "recursive call could loop indefinitely" because pcre_compile()
1646 was not correctly checking the subroutine for matching a non-empty string.
1647
1648 14. The checks for overrunning compiling workspace could trigger after an
1649 overrun had occurred. This is a "should never occur" error, but it can be
1650 triggered by pathological patterns such as hundreds of nested parentheses.
1651 The checks now trigger 100 bytes before the end of the workspace.
1652
1653 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1654
1655
1656 Version 8.01 19-Jan-2010
1657 ------------------------
1658
1659 1. If a pattern contained a conditional subpattern with only one branch (in
1660 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1661 computed the wrong minimum data length (which is of course zero for such
1662 subpatterns). This could cause incorrect "no match" results.
1663
1664 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1665 the pattern is reset in the first branch, pcre_compile() failed with
1666 "internal error: code overflow at offset...". This happened only when
1667 the reset was to the original external option setting. (An optimization
1668 abstracts leading options settings into an external setting, which was the
1669 cause of this.)
1670
1671 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1672 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1673 assertion pattern did not match (meaning that the assertion was true), it
1674 was incorrectly treated as false if the SKIP had been reached during the
1675 matching. This also applied to assertions used as conditions.
1676
1677 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1678 assertion subpattern, including such a pattern used as a condition,
1679 unpredictable results occurred, instead of the error return
1680 PCRE_ERROR_DFA_UITEM.
1681
1682 5. The C++ GlobalReplace function was not working like Perl for the special
1683 situation when an empty string is matched. It now does the fancy magic
1684 stuff that is necessary.
1685
1686 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1687 removed. (These were left over from very, very early versions of PCRE.)
1688
1689 7. Some cosmetic changes to the code to make life easier when compiling it
1690 as part of something else:
1691
1692 (a) Change DEBUG to PCRE_DEBUG.
1693
1694 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1695 called "current" as "current_branch", to prevent a collision with the
1696 Linux macro when compiled as a kernel module.
1697
1698 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1699 prevent a collision with the Linux macro when compiled as a kernel
1700 module.
1701
1702 8. In pcre_compile() there are some checks for integer overflows that used to
1703 cast potentially large values to (double). This has been changed to that
1704 when building, a check for int64_t is made, and if it is found, it is used
1705 instead, thus avoiding the use of floating point arithmetic. (There is no
1706 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1707 double.
1708
1709 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1710 2005 (difference between two addresses compared to an unsigned value).
1711
1712 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1713 custom one, because of the following reported problem in Windows:
1714
1715 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1716 under Win32.
1717 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1718 therefore missing the function definition.
1719 - The compiler thus generates a "C" signature for the test function.
1720 - The linker fails to find the "C" function.
1721 - PCRE fails to configure if asked to do so against libbz2.
1722
1723 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1724 messages were output:
1725
1726 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1727 rerunning libtoolize, to keep the correct libtool macros in-tree.
1728 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1729
1730 I have done both of these things.
1731
1732 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1733 most of the time, it *can* run out if it is given a pattern that contains a
1734 runaway infinite recursion. I updated the discussion in the pcrestack man
1735 page.
1736
1737 13. Now that we have gone to the x.xx style of version numbers, the minor
1738 version may start with zero. Using 08 or 09 is a bad idea because users
1739 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1740 interpreted as invalid octal numbers. I've updated the previous comment in
1741 configure.ac, and also added a check that gives an error if 08 or 09 are
1742 used.
1743
1744 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1745 causing partial matching to fail when the end of the subject matched \W
1746 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1747
1748 15. There were some discrepancies between the declarations in pcre_internal.h
1749 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1750 their definitions. The declarations used "const uschar *" and the
1751 definitions used USPTR. Even though USPTR is normally defined as "const
1752 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1753 reported that: "This difference in casting confuses some C++ compilers, for
1754 example, SunCC recognizes above declarations as different functions and
1755 generates broken code for hbpcre." I have changed the declarations to use
1756 USPTR.
1757
1758 16. GNU libtool is named differently on some systems. The autogen.sh script now
1759 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1760 (FreeBSD).
1761
1762 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1763 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1764 comment: "Figure out how to create a longlong from a string: strtoll and
1765 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1766 instance, but it only takes 2 args instead of 3!"
1767
1768 18. A subtle bug concerned with back references has been fixed by a change of
1769 specification, with a corresponding code fix. A pattern such as
1770 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1771 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1772 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1773 same bug. Such groups have to be quantified to be useful, or contained
1774 inside another quantified group. (If there's no repetition, the reference
1775 can never match.) The problem arises because, having left the group and
1776 moved on to the rest of the pattern, a later failure that backtracks into
1777 the group uses the captured value from the final iteration of the group
1778 rather than the correct earlier one. I have fixed this in PCRE by forcing
1779 any group that contains a reference to itself to be an atomic group; that
1780 is, there cannot be any backtracking into it once it has completed. This is
1781 similar to recursive and subroutine calls.
1782
1783
1784 Version 8.00 19-Oct-09
1785 ----------------------
1786
1787 1. The table for translating pcre_compile() error codes into POSIX error codes
1788 was out-of-date, and there was no check on the pcre_compile() error code
1789 being within the table. This could lead to an OK return being given in
1790 error.
1791
1792 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1793 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1794 in a Windows environment.
1795
1796 3. The pcregrep --count option prints the count for each file even when it is
1797 zero, as does GNU grep. However, pcregrep was also printing all files when
1798 --files-with-matches was added. Now, when both options are given, it prints
1799 counts only for those files that have at least one match. (GNU grep just
1800 prints the file name in this circumstance, but including the count seems
1801 more useful - otherwise, why use --count?) Also ensured that the
1802 combination -clh just lists non-zero counts, with no names.
1803
1804 4. The long form of the pcregrep -F option was incorrectly implemented as
1805 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1806 but it seems right to fix it, and I didn't think it was worth preserving
1807 the old behaviour.
1808
1809 5. The command line items --regex=pattern and --regexp=pattern were not
1810 recognized by pcregrep, which required --regex pattern or --regexp pattern
1811 (with a space rather than an '='). The man page documented the '=' forms,
1812 which are compatible with GNU grep; these now work.
1813
1814 6. No libpcreposix.pc file was created for pkg-config; there was just
1815 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1816
1817 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1818 when UCP support is not needed, by modifying the Python script that
1819 generates it from Unicode data files. This should not matter if the module
1820 is correctly used as a library, but I received one complaint about 50K of
1821 unwanted data. My guess is that the person linked everything into his
1822 program rather than using a library. Anyway, it does no harm.
1823
1824 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1825 was a minimum greater than 1 for a wide character in a possessive
1826 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1827 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1828 character. Chaos in the form of incorrect output or a compiling loop could
1829 result.
1830
1831 9. The restrictions on what a pattern can contain when partial matching is
1832 requested for pcre_exec() have been removed. All patterns can now be
1833 partially matched by this function. In addition, if there are at least two
1834 slots in the offset vector, the offset of the earliest inspected character
1835 for the match and the offset of the end of the subject are set in them when
1836 PCRE_ERROR_PARTIAL is returned.
1837
1838 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1839 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1840 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1841 and may be more useful for multi-segment matching.
1842
1843 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1844 used to be given if ever the end of the subject was reached; now it is
1845 given only if matching could not proceed because another character was
1846 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1847 string "Z", which now yields "no match" instead of "partial match". In the
1848 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1849 final character ended with (*FAIL).
1850
1851 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1852 if the pattern had a "must contain" character that was already found in the
1853 earlier partial match, unless partial matching was again requested. For
1854 example, with the pattern /dog.(body)?/, the "must contain" character is
1855 "g". If the first part-match was for the string "dog", restarting with
1856 "sbody" failed. This bug has been fixed.
1857
1858 13. The string returned by pcre_dfa_exec() after a partial match has been
1859 changed so that it starts at the first inspected character rather than the
1860 first character of the match. This makes a difference only if the pattern
1861 starts with a lookbehind assertion or \b or \B (\K is not supported by
1862 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1863 matching functions compatible, and I think it's the right thing to do.
1864
1865 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1866 so that the demonstration program is easily available in environments where
1867 PCRE has not been installed from source.
1868
1869 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1870 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1871 library.
1872
1873 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1874 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1875 is not the first non-POSIX option to be added. Clearly some people find
1876 these options useful.
1877
1878 17. If a caller to the POSIX matching function regexec() passes a non-zero
1879 value for nmatch with a NULL value for pmatch, the value of
1880 nmatch is forced to zero.
1881
1882 18. RunGrepTest did not have a test for the availability of the -u option of
1883 the diff command, as RunTest does. It now checks in the same way as
1884 RunTest, and also checks for the -b option.
1885
1886 19. If an odd number of negated classes containing just a single character
1887 interposed, within parentheses, between a forward reference to a named
1888 subpattern and the definition of the subpattern, compilation crashed with
1889 an internal error, complaining that it could not find the referenced
1890 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1891 [The bug was that it was starting one character too far in when skipping
1892 over the character class, thus treating the ] as data rather than
1893 terminating the class. This meant it could skip too much.]
1894
1895 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1896 /g option in pcretest when the pattern contains \K, which makes it possible
1897 to have an empty string match not at the start, even when the pattern is
1898 anchored. Updated pcretest and pcredemo to use this option.
1899
1900 21. If the maximum number of capturing subpatterns in a recursion was greater
1901 than the maximum at the outer level, the higher number was returned, but
1902 with unset values at the outer level. The correct (outer level) value is
1903 now given.
1904
1905 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1906 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1907 make it do so. The string so far is captured, making this feature
1908 compatible with Perl.
1909
1910 23. The tests have been re-organized, adding tests 11 and 12, to make it
1911 possible to check the Perl 5.10 features against Perl 5.10.
1912
1913 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1914 pattern matches a fixed length string. PCRE did not allow this; now it
1915 does. Neither allows recursion.
1916
1917 25. I finally figured out how to implement a request to provide the minimum
1918 length of subject string that was needed in order to match a given pattern.
1919 (It was back references and recursion that I had previously got hung up
1920 on.) This code has now been added to pcre_study(); it finds a lower bound
1921 to the length of subject needed. It is not necessarily the greatest lower
1922 bound, but using it to avoid searching strings that are too short does give
1923 some useful speed-ups. The value is available to calling programs via
1924 pcre_fullinfo().
1925
1926 26. While implementing 25, I discovered to my embarrassment that pcretest had
1927 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1928 study optimizations had never been tested with that matching function.
1929 Oops. What is worse, even when it was passed study data, there was a bug in
1930 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1931 were also very few tests of studied patterns with pcre_dfa_exec().
1932
1933 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1934 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1935 on the other side of the coin, they are no longer allowed to have different
1936 names, because these cannot be distinguished in PCRE, and this has caused
1937 confusion. (This is a difference from Perl.)
1938
1939 28. When duplicate subpattern names are present (necessarily with different
1940 numbers, as required by 27 above), and a test is made by name in a
1941 conditional pattern, either for a subpattern having been matched, or for
1942 recursion in such a pattern, all the associated numbered subpatterns are
1943 tested, and the overall condition is true if the condition is true for any
1944 one of them. This is the way Perl works, and is also more like the way
1945 testing by number works.
1946
1947
1948 Version 7.9 11-Apr-09
1949 ---------------------
1950
1951 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1952 (pcretest), all targets were linked against these libraries. This included
1953 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1954 libraries. This caused unwanted dependencies to be created. This problem
1955 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1956 pcretest is linked with readline.
1957
1958 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1959 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1960 moved outside it again, because FALSE and TRUE are already defined in AIX,
1961 but BOOL is not.
1962
1963 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1964 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1965
1966 4. The pcregrep documentation said spaces were inserted as well as colons (or
1967 hyphens) following file names and line numbers when outputting matching
1968 lines. This is not true; no spaces are inserted. I have also clarified the
1969 wording for the --colour (or --color) option.
1970
1971 5. In pcregrep, when --colour was used with -o, the list of matching strings
1972 was not coloured; this is different to GNU grep, so I have changed it to be
1973 the same.
1974
1975 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1976 each matching line was coloured. Now it goes on to look for further matches
1977 of any of the test patterns, which is the same behaviour as GNU grep.
1978
1979 7. A pattern that could match an empty string could cause pcregrep to loop; it
1980 doesn't make sense to accept an empty string match in pcregrep, so I have
1981 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1982 seems to be how GNU grep behaves. [But see later change 40 for release
1983 8.33.]
1984
1985 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1986 start or after a newline", because the conditional assertion was not being
1987 correctly handled. The rule now is that both the assertion and what follows
1988 in the first alternative must satisfy the test.
1989
1990 9. If auto-callout was enabled in a pattern with a conditional group whose
1991 condition was an assertion, PCRE could crash during matching, both with
1992 pcre_exec() and pcre_dfa_exec().
1993
1994 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1995 used for matching.
1996
1997 11. Unicode property support in character classes was not working for
1998 characters (bytes) greater than 127 when not in UTF-8 mode.
1999
2000 12. Added the -M command line option to pcretest.
2001
2002 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2003
2004 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2005
2006 16. Added comments and documentation about mis-use of no_arg in the C++
2007 wrapper.
2008
2009 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2010 from Martin Jerabek that uses macro names for all relevant character and
2011 string constants.
2012
2013 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2014 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2015 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2016 these, but not everybody uses configure.
2017
2018 19. A conditional group that had only one branch was not being correctly
2019 recognized as an item that could match an empty string. This meant that an
2020 enclosing group might also not be so recognized, causing infinite looping
2021 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2022 with the subject "ab", where knowledge that the repeated group can match
2023 nothing is needed in order to break the loop.
2024
2025 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2026 exec(), but without supplying a callout function, matching went wrong.
2027
2028 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2029 leak if the size of the offset vector was greater than 30. When the vector
2030 is smaller, the saved offsets during recursion go onto a local stack
2031 vector, but for larger vectors malloc() is used. It was failing to free
2032 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2033 error, in fact).
2034
2035 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2036 heapframe that is used only when UTF-8 support is enabled. This caused no
2037 problem, but was untidy.
2038
2039 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2040 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2041 included within another project.
2042
2043 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2044 slightly modified by me:
2045
2046 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2047 not building pcregrep.
2048
2049 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2050 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2051
2052 25. Forward references, both numeric and by name, in patterns that made use of
2053 duplicate group numbers, could behave incorrectly or give incorrect errors,
2054 because when scanning forward to find the reference group, PCRE was not
2055 taking into account the duplicate group numbers. A pattern such as
2056 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2057
2058 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2059 the feature of a custom pointer more persuasive (as requested by a user).
2060
2061 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2062 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2063 pre-defined.
2064
2065 28. Added support for (*UTF8) at the start of a pattern.
2066
2067 29. Arrange for flags added by the "release type" setting in CMake to be shown
2068 in the configuration summary.
2069
2070
2071 Version 7.8 05-Sep-08
2072 ---------------------
2073
2074 1. Replaced UCP searching code with optimized version as implemented for Ad
2075 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2076 stage table and inline lookup instead of a function, giving speed ups of 2
2077 to 5 times on some simple patterns that I tested. Permission was given to
2078 distribute the MultiStage2.py script that generates the tables (it's not in
2079 the tarball, but is in the Subversion repository).
2080
2081 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2082 scripts.
2083
2084 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2085 a group with a zero qualifier. The result of the study could be incorrect,
2086 or the function might crash, depending on the pattern.
2087
2088 4. Caseless matching was not working for non-ASCII characters in back
2089 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2090 It now works when Unicode Property Support is available.
2091
2092 5. In pcretest, an escape such as \x{de} in the data was always generating
2093 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2094 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2095 truncation.
2096
2097 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2098
2099 7. Added two (int) casts to pcregrep when printing the difference of two
2100 pointers, in case they are 64-bit values.
2101
2102 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2103 test 2 if it fails.
2104
2105 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2106 and a #define of that name to empty if it is not externally set. This is to
2107 allow users of MSVC to set it if necessary.
2108
2109 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2110 the convenience functions in the pcre_get.c source file.
2111
2112 11. An option change at the start of a pattern that had top-level alternatives
2113 could cause overwriting and/or a crash. This command provoked a crash in
2114 some environments:
2115
2116 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2117
2118 This potential security problem was recorded as CVE-2008-2371.
2119
2120 12. For a pattern where the match had to start at the beginning or immediately
2121 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2122 pcre_dfa_exec() could read past the end of the passed subject if there was
2123 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2124 pcretest so that it places the subject at the end of its malloc-ed buffer.
2125
2126 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2127 exec() might read past the end of the data buffer in UTF-8 mode.
2128
2129 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2130 the data contained the byte 0x85 as part of a UTF-8 character within its
2131 first line. This applied both to normal and DFA matching.
2132
2133 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2134 /^[^d]*?$/8 failed to match "abc".
2135
2136 16. Added a missing copyright notice to pcrecpp_internal.h.
2137
2138 17. Make it more clear in the documentation that values returned from
2139 pcre_exec() in ovector are byte offsets, not character counts.
2140
2141 18. Tidied a few places to stop certain compilers from issuing warnings.
2142
2143 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2144 supplied by Stefan Weber. I made a further small update for 7.8 because
2145 there is a change of source arrangements: the pcre_searchfuncs.c module is
2146 replaced by pcre_ucd.c.
2147
2148
2149 Version 7.7 07-May-08
2150 ---------------------
2151
2152 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2153 a string to a long long, pretend we don't even have a long long." This is
2154 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2155
2156 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2157 pre-7.6 versions, which defined a global no_arg variable instead of putting
2158 it in the RE class. (See also #8 below.)
2159
2160 3. Remove a line of dead code, identified by coverity and reported by Nuno
2161 Lopes.
2162
2163 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2164
2165 (1) The include/exclude patterns were being applied to the whole pathnames
2166 of files, instead of just to the final components.
2167
2168 (2) If there was more than one level of directory, the subdirectories were
2169 skipped unless they satisfied the include/exclude conditions. This is
2170 inconsistent with GNU grep (and could even be seen as contrary to the
2171 pcregrep specification - which I improved to make it absolutely clear).
2172 The action now is always to scan all levels of directory, and just
2173 apply the include/exclude patterns to regular files.
2174
2175 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2176 --exclude_dir in the tests to avoid scanning .svn directories.
2177
2178 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2179 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2180 doesn't support NULs in patterns.
2181
2182 7. Added some missing "const"s to declarations of static tables in
2183 pcre_compile.c and pcre_dfa_exec.c.
2184
2185 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2186 caused by fix #2 above. (Subsequently also a second patch to fix the
2187 first patch. And a third patch - this was a messy problem.)
2188
2189 9. Applied Craig's patch to remove the use of push_back().
2190
2191 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2192 matching function regexec().
2193
2194 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2195 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2196 references. PCRE supports relative numbers with this syntax (I don't think
2197 Oniguruma does).
2198
2199 12. Previously, a group with a zero repeat such as (...){0} was completely
2200 omitted from the compiled regex. However, this means that if the group
2201 was called as a subroutine from elsewhere in the pattern, things went wrong
2202 (an internal error was given). Such groups are now left in the compiled
2203 pattern, with a new opcode that causes them to be skipped at execution
2204 time.
2205
2206 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2207 to the way PCRE behaves:
2208
2209 (a) A lone ] character is dis-allowed (Perl treats it as data).
2210
2211 (b) A back reference to an unmatched subpattern matches an empty string
2212 (Perl fails the current match path).
2213
2214 (c) A data ] in a character class must be notated as \] because if the
2215 first data character in a class is ], it defines an empty class. (In
2216 Perl it is not possible to have an empty class.) The empty class []
2217 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2218 The negative empty class [^] matches any one character, independently
2219 of the DOTALL setting.
2220
2221 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2222 non-existent subpattern following a character class starting with ']' and
2223 containing () gave an internal compiling error instead of "reference to
2224 non-existent subpattern". Fortunately, when the pattern did exist, the
2225 compiled code was correct. (When scanning forwards to check for the
2226 existence of the subpattern, it was treating the data ']' as terminating
2227 the class, so got the count wrong. When actually compiling, the reference
2228 was subsequently set up correctly.)
2229
2230 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2231 it was being rejected as not supported by pcre_dfa_exec(), even though
2232 other assertions are supported. I have made pcre_dfa_exec() support
2233 (*FAIL).
2234
2235 16. The implementation of 13c above involved the invention of a new opcode,
2236 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2237 cannot be changed at match time, I realized I could make a small
2238 improvement to matching performance by compiling OP_ALLANY instead of
2239 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2240 on the OP_ANY path.
2241
2242 17. Compiling pcretest on Windows with readline support failed without the
2243 following two fixes: (1) Make the unistd.h include conditional on
2244 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2245
2246 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2247 ncurses library to be included for pcretest when ReadLine support is
2248 requested, but also to allow for it to be overridden. This patch came from
2249 Daniel Bergström.
2250
2251 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2252 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2253 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2254 spotting this.
2255
2256
2257 Version 7.6 28-Jan-08
2258 ---------------------
2259
2260 1. A character class containing a very large number of characters with
2261 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2262 overflow.
2263
2264 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2265 HAVE_LONG_LONG is not defined.
2266
2267 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2268 bring them up to date and include new features. This patch includes:
2269
2270 - Fixed PH's badly added libz and libbz2 support.
2271 - Fixed a problem with static linking.
2272 - Added pcredemo. [But later removed - see 7 below.]
2273 - Fixed dftables problem and added an option.
2274 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2275 HAVE_LONG_LONG.
2276 - Added readline support for pcretest.
2277 - Added an listing of the option settings after cmake has run.
2278
2279 4. A user submitted a patch to Makefile that makes it easy to create
2280 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2281 Makefile.am that cause it to include this special target, without
2282 affecting anything else. Note that the same mingw target plus all
2283 the other distribution libraries and programs are now supported
2284 when configuring with CMake (see 6 below) instead of with
2285 Configure/Make.
2286
2287 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2288 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2289 exported in the Windows port". It has not yet been confirmed that the patch
2290 solves the problem, but it does no harm.
2291
2292 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2293 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2294 with CMake, and also correct the comment about stack recursion.
2295
2296 7. Remove the automatic building of pcredemo from the ./configure system and
2297 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2298 of a program that users should build themselves after PCRE is installed, so
2299 building it automatically is not really right. What is more, it gave
2300 trouble in some build environments.
2301
2302 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2303
2304
2305 Version 7.5 10-Jan-08
2306 ---------------------
2307
2308 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2309 values in parens when parsing an RE using the C++ wrapper."
2310
2311 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2312 Characters greater than 255 were excluded from the class instead of being
2313 included.
2314
2315 3. The same bug as (2) above applied to negated POSIX classes such as
2316 [:^space:].
2317
2318 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2319 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2320 I have changed it.
2321
2322 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2323 first named subpattern) and a construct such as (?&a) would reference the
2324 first named subpattern whose name started with "a" (in other words, the
2325 length check was missing). Both these problems are fixed. "Subpattern name
2326 expected" is now given for (?&) (a zero-length name), and this patch also
2327 makes it give the same error for \k'' (previously it complained that that
2328 was a reference to a non-existent subpattern).
2329
2330 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2331 this is right because (?- can be followed by option settings as well as by
2332 digits. I have, however, made the messages clearer.
2333
2334 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2335 than the number used in the conditional) now cause a compile-time error.
2336 This is actually not compatible with Perl, which accepts such patterns, but
2337 treats the conditional as always being FALSE (as PCRE used to), but it
2338 seems to me that giving a diagnostic is better.
2339
2340 8. Change "alphameric" to the more common word "alphanumeric" in comments
2341 and messages.
2342
2343 9. Fix two occurrences of "backslash" in comments that should have been
2344 "backspace".
2345
2346 10. Remove two redundant lines of code that can never be obeyed (their function
2347 was moved elsewhere).
2348
2349 11. The program that makes PCRE's Unicode character property table had a bug
2350 which caused it to generate incorrect table entries for sequences of
2351 characters that have the same character type, but are in different scripts.
2352 It amalgamated them into a single range, with the script of the first of
2353 them. In other words, some characters were in the wrong script. There were
2354 thirteen such cases, affecting characters in the following ranges:
2355
2356 U+002b0 - U+002c1
2357 U+0060c - U+0060d
2358 U+0061e - U+00612
2359 U+0064b - U+0065e
2360 U+0074d - U+0076d
2361 U+01800 - U+01805
2362 U+01d00 - U+01d77
2363 U+01d9b - U+01dbf
2364 U+0200b - U+0200f
2365 U+030fc - U+030fe
2366 U+03260 - U+0327f
2367 U+0fb46 - U+0fbb1
2368 U+10450 - U+1049d
2369
2370 12. The -o option (show only the matching part of a line) for pcregrep was not
2371 compatible with GNU grep in that, if there was more than one match in a
2372 line, it showed only the first of them. It now behaves in the same way as
2373 GNU grep.
2374
2375 13. If the -o and -v options were combined for pcregrep, it printed a blank
2376 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2377 does the same. The return code can be used to tell if there were any
2378 non-matching lines.
2379
2380 14. Added --file-offsets and --line-offsets to pcregrep.
2381
2382 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2383 infinitely looping recursion. The bug was that positive lookaheads were not
2384 being skipped when checking for a possible empty match (negative lookaheads
2385 and both kinds of lookbehind were skipped).
2386
2387 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2388 inclusion of <windows.h> to before rather than after the definition of
2389 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2390
2391 17. Specifying a possessive quantifier with a specific limit for a Unicode
2392 character property caused pcre_compile() to compile bad code, which led at
2393 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2394 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2395 caused the error; without that there was no problem.
2396
2397 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2398
2399 19. Added --enable-pcretest-libreadline.
2400
2401 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2402 RE::GlobalReplace(). As a result, the number of replacements returned was
2403 double what it should be. I removed one of the increments, but Craig sent a
2404 later patch that removed the other one (the right fix) and added unit tests
2405 that check the return values (which was not done before).
2406
2407 21. Several CMake things:
2408
2409 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2410 the names libpcre and libpcreposix, not just pcre and pcreposix.
2411
2412 (2) The above change means that pcretest and pcregrep are now correctly
2413 linked with the newly-built libraries, not previously installed ones.
2414
2415 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2416
2417 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2418 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2419 UTF-8 newline character). The key issue is that the pattern starts .*;
2420 this means that the match must be either at the beginning, or after a
2421 newline. The bug was in the code for advancing after a failed match and
2422 checking that the new position followed a newline. It was not taking
2423 account of UTF-8 characters correctly.
2424
2425 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2426 character classes. PCRE was not treating the sequence [:...:] as a
2427 character class unless the ... were all letters. Perl, however, seems to
2428 allow any characters between [: and :], though of course it rejects as
2429 unknown any "names" that contain non-letters, because all the known class
2430 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2431 for example, whereas PCRE did not - it did not recognize a POSIX character
2432 class. This seemed a bit dangerous, so the code has been changed to be
2433 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2434 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2435 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2436 Perl does, and where it didn't before.
2437
2438 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2439 Windows environments %n is disabled by default.
2440
2441
2442 Version 7.4 21-Sep-07
2443 ---------------------
2444
2445 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2446 means that a class such as [\s] counted as "explicit reference to CR or
2447 LF". That isn't really right - the whole point of the change was to try to
2448 help when there was an actual mention of one of the two characters. So now
2449 the change happens only if \r or \n (or a literal CR or LF) character is
2450 encountered.
2451
2452 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2453 of both had grown to the point where there were only 3 bits left.
2454 Fortunately, there was spare space in the data structure, and so I have
2455 moved the internal flags into a new 16-bit field to free up more option
2456 bits.
2457
2458 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2459 but did not set the internal JCHANGED flag - either of these is enough to
2460 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2461 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2462 start sets both bits.
2463
2464 4. Added options (at build time, compile time, exec time) to change \R from
2465 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2466
2467 5. doc/pcresyntax.html was missing from the distribution.
2468
2469 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2470 compatibility, even though it is no longer used.
2471
2472 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2473 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2474 windows.h file is present (where different names are used). [This was
2475 reversed later after testing - see 16 below.]
2476
2477 8. Changed all #include <config.h> to #include "config.h". There were also
2478 some further <pcre.h> cases that I changed to "pcre.h".
2479
2480 9. When pcregrep was used with the --colour option, it missed the line ending
2481 sequence off the lines that it output.
2482
2483 10. It was pointed out to me that arrays of string pointers cause lots of
2484 relocations when a shared library is dynamically loaded. A technique of
2485 using a single long string with a table of offsets can drastically reduce
2486 these. I have refactored PCRE in four places to do this. The result is
2487 dramatic:
2488
2489 Originally: 290
2490 After changing UCP table: 187
2491 After changing error message table: 43
2492 After changing table of "verbs" 36
2493 After changing table of Posix names 22
2494
2495 Thanks to the folks working on Gregex for glib for this insight.
2496
2497 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2498 unicode-properties was also set.
2499
2500 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2501
2502 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2503 checked only for CRLF.
2504
2505 14. Added casts to pcretest.c to avoid compiler warnings.
2506
2507 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2508
2509 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2510 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2511 entirely. This removes changes made in 7 above.
2512
2513 17. The CMake files have been updated, and there is now more information about
2514 building with CMake in the NON-UNIX-USE document.
2515
2516
2517 Version 7.3 28-Aug-07
2518 ---------------------
2519
2520 1. In the rejigging of the build system that eventually resulted in 7.1, the
2521 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2522 brackets there is not right, since it causes compilers to look for an
2523 installed pcre.h, not the version that is in the source that is being
2524 compiled (which of course may be different). I have changed it back to:
2525
2526 #include "pcre.h"
2527
2528 I have a vague recollection that the change was concerned with compiling in
2529 different directories, but in the new build system, that is taken care of
2530 by the VPATH setting the Makefile.
2531
2532 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2533 when the subject happened to end in the byte 0x85 (e.g. if the last
2534 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2535 characters but of course it shouldn't be taken as a newline when it is part
2536 of another character. The bug was that, for an unlimited repeat of . in
2537 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2538 characters when looking for a newline.
2539
2540 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2541
2542 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2543 in debug output.
2544
2545 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2546 long printing in the pcrecpp unittest when running under MinGW.
2547
2548 6. ESC_K was left out of the EBCDIC table.
2549
2550 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2551 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2552 limit also applies to "virtual nesting" when a pattern is recursive, and in
2553 this case 1000 isn't so big. I have been able to remove this limit at the
2554 expense of backing off one optimization in certain circumstances. Normally,
2555 when pcre_exec() would call its internal match() function recursively and
2556 immediately return the result unconditionally, it uses a "tail recursion"
2557 feature to save stack. However, when a subpattern that can match an empty
2558 string has an unlimited repetition quantifier, it no longer makes this
2559 optimization. That gives it a stack frame in which to save the data for
2560 checking that an empty string has been matched. Previously this was taken
2561 from the 1000-entry workspace that had been reserved. So now there is no
2562 explicit limit, but more stack is used.
2563
2564 8. Applied Daniel's patches to solve problems with the import/export magic
2565 syntax that is required for Windows, and which was going wrong for the
2566 pcreposix and pcrecpp parts of the library. These were overlooked when this
2567 problem was solved for the main library.
2568
2569 9. There were some crude static tests to avoid integer overflow when computing
2570 the size of patterns that contain repeated groups with explicit upper
2571 limits. As the maximum quantifier is 65535, the maximum group length was
2572 set at 30,000 so that the product of these two numbers did not overflow a
2573 32-bit integer. However, it turns out that people want to use groups that
2574 are longer than 30,000 bytes (though not repeat them that many times).
2575 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2576 made it possible to implement the integer overflow checks in a much more
2577 dynamic way, which I have now done. The artificial limitation on group
2578 length has been removed - we now have only the limit on the total length of
2579 the compiled pattern, which depends on the LINK_SIZE setting.
2580
2581 10. Fixed a bug in the documentation for get/copy named substring when
2582 duplicate names are permitted. If none of the named substrings are set, the
2583 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2584 empty string.
2585
2586 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2587 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2588 because the ] is interpreted as the first data character and the
2589 terminating ] is not found. PCRE has been made compatible with Perl in this
2590 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2591 cause memory overwriting.
2592
2593 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2594 string has been matched (to stop an infinite loop). It was not recognizing
2595 a conditional subpattern that could match an empty string if that
2596 subpattern was within another subpattern. For example, it looped when
2597 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2598 condition was not nested. This bug has been fixed.
2599
2600 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2601 past the start of the subject in the presence of bytes with the top bit
2602 set, for example "\x8aBCD".
2603
2604 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2605 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2606
2607 14. Optimized (?!) to (*FAIL).
2608
2609 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2610 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2611 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2612 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2613 does: it's just the validity check that is more restrictive.
2614
2615 16. Inserted checks for integer overflows during escape sequence (backslash)
2616 processing, and also fixed erroneous offset values for syntax errors during
2617 backslash processing.
2618
2619 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2620 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2621
2622 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2623 caused an overrun.
2624
2625 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2626 something other than just ASCII characters) inside a group that had an
2627 unlimited repeat caused a loop at compile time (while checking to see
2628 whether the group could match an empty string).
2629
2630 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2631 [\P{Any}] did so. (Error in the code for printing property names.)
2632
2633 21. An orphan \E inside a character class could cause a crash.
2634
2635 22. A repeated capturing bracket such as (A)? could cause a wild memory
2636 reference during compilation.
2637
2638 23. There are several functions in pcre_compile() that scan along a compiled
2639 expression for various reasons (e.g. to see if it's fixed length for look
2640 behind). There were bugs in these functions when a repeated \p or \P was
2641 present in the pattern. These operators have additional parameters compared
2642 with \d, etc, and these were not being taken into account when moving along
2643 the compiled data. Specifically:
2644
2645 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2646 length.
2647
2648 (b) An item such as \pL+ within a repeated group could cause crashes or
2649 loops.
2650
2651 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2652 "reference to non-existent subpattern" error.
2653
2654 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2655
2656 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2657 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2658
2659 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2660
2661 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2662 character were causing crashes (broken optimization).
2663
2664 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2665 \p or \P) caused a compile-time loop.
2666
2667 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2668 break. For example, the unstudied pattern [\r\n]A does not match the string
2669 "\r\nA" because change 7.0/46 below moves the current point on by two
2670 characters after failing to match at the start. However, the pattern \nA
2671 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2672 the same is true. There doesn't seem any very clean way out of this, but
2673 what I have chosen to do makes the common cases work: PCRE now takes note
2674 of whether there can be an explicit match for \r or \n anywhere in the
2675 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2676 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2677 pattern has explicit CR or LF references.
2678
2679 29. Added (*CR) etc for changing newline setting at start of pattern.
2680
2681
2682 Version 7.2 19-Jun-07
2683 ---------------------
2684
2685 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2686 which is apparently normally available under Windows.
2687
2688 2. Re-jig the pcregrep tests with different newline settings in an attempt
2689 to make them independent of the local environment's newline setting.
2690
2691 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2692
2693 4. Some of the "internals" tests were previously cut out when the link size
2694 was not 2, because the output contained actual offsets. The recent new
2695 "Z" feature of pcretest means that these can be cut out, making the tests
2696 usable with all link sizes.
2697
2698 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2699 stack recursion. This gives a massive performance boost under BSD, but just
2700 a small improvement under Linux. However, it saves one field in the frame
2701 in all cases.
2702
2703 6. Added more features from the forthcoming Perl 5.10:
2704
2705 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2706 recursion call. It refers to the nth most recently opened parentheses.
2707
2708 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2709 to be opened parentheses.
2710
2711 (c) Conditions that refer to capturing parentheses can be specified
2712 relatively, for example, (?(-2)... or (?(+3)...
2713
2714 (d) \K resets the start of the current match so that everything before
2715 is not part of it.
2716
2717 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2718
2719 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2720 reference syntax.
2721
2722 (g) (?| introduces a group in which the numbering of parentheses in each
2723 alternative starts with the same number.
2724
2725 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2726
2727 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2728 PCRE_INFO_JCHANGED.
2729
2730 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2731 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2732 for detecting groups that can match an empty string.
2733
2734 9. A pattern with a very large number of alternatives (more than several
2735 hundred) was running out of internal workspace during the pre-compile
2736 phase, where pcre_compile() figures out how much memory will be needed. A
2737 bit of new cunning has reduced the workspace needed for groups with
2738 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2739 workspace instead of running out of the 4096 that are available.
2740
2741 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2742
2743 11. Applied patch from Google to remove an optimization that didn't quite work.
2744 The report of the bug said:
2745
2746 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2747 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2748 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2749
2750 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2751 it matched the wrong number of bytes.
2752
2753
2754 Version 7.1 24-Apr-07
2755 ---------------------
2756
2757 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2758 that is more "standard", making use of automake and other Autotools. There
2759 is some re-arrangement of the files and adjustment of comments consequent
2760 on this.
2761
2762 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2763 for recursive directory scanning broke on some systems because the files
2764 are not scanned in any specific order and on different systems the order
2765 was different. A call to "sort" has been inserted into RunGrepTest for the
2766 approprate test as a short-term fix. In the longer term there may be an
2767 alternative.
2768
2769 3. I had an email from Eric Raymond about problems translating some of PCRE's
2770 man pages to HTML (despite the fact that I distribute HTML pages, some
2771 people do their own conversions for various reasons). The problems
2772 concerned the use of low-level troff macros .br and .in. I have therefore
2773 removed all such uses from the man pages (some were redundant, some could
2774 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2775 HTML has been updated to handle .nf/.fi and to complain if it encounters
2776 .br or .in.
2777
2778 4. Updated comments in configure.ac that get placed in config.h.in and also
2779 arranged for config.h to be included in the distribution, with the name
2780 config.h.generic, for the benefit of those who have to compile without
2781 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2782
2783 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2784 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2785 makevp.bat for the current PCRE, using the additional files
2786 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2787
2788 6. A Windows user reported a minor discrepancy with test 2, which turned out
2789 to be caused by a trailing space on an input line that had got lost in his
2790 copy. The trailing space was an accident, so I've just removed it.
2791
2792 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2793 that is needed.
2794
2795 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2796 as "const" (a) because they are and (b) because it helps the PHP
2797 maintainers who have recently made a script to detect big data structures
2798 in the php code that should be moved to the .rodata section. I remembered
2799 to update Builducptable as well, so it won't revert if ucptable.h is ever
2800 re-created.
2801
2802 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2803 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2804 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2805 support is not required. This saves 1.5-2K of code, which is important in
2806 some applications.
2807
2808 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2809 so as not to refer to the tables, even though these functions will never be
2810 called when UTF-8 support is disabled. Otherwise there are problems with a
2811 shared library.
2812
2813 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2814
2815 (a) It was defining its arguments as char * instead of void *.
2816
2817 (b) It was assuming that all moves were upwards in memory; this was true
2818 a long time ago when I wrote it, but is no longer the case.
2819
2820 The emulated memove() is provided for those environments that have neither
2821 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2822 is clearly not the case, as these two bugs were recently reported.
2823
2824 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2825 and Detrail to create the HTML documentation, the .txt form of the man
2826 pages, and it removes trailing spaces from listed files. It also creates
2827 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2828 case, it wraps all the #defines with #ifndefs. This script should be run
2829 before "make dist".
2830
2831 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2832 with Unicode property support.
2833
2834 (a) For a maximizing quantifier, if the two different cases of the
2835 character were of different lengths in their UTF-8 codings (there are
2836 some cases like this - I found 11), and the matching function had to
2837 back up over a mixture of the two cases, it incorrectly assumed they
2838 were both the same length.
2839
2840 (b) When PCRE was configured to use the heap rather than the stack for
2841 recursion during matching, it was not correctly preserving the data for
2842 the other case of a UTF-8 character when checking ahead for a match
2843 while processing a minimizing repeat. If the check also involved
2844 matching a wide character, but failed, corruption could cause an
2845 erroneous result when trying to check for a repeat of the original
2846 character.
2847
2848 13. Some tidying changes to the testing mechanism:
2849
2850 (a) The RunTest script now detects the internal link size and whether there
2851 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2852 values substituted by "configure". (The RunGrepTest script already did
2853 this for UTF-8.) The configure.ac script no longer substitutes the
2854 relevant variables.
2855
2856 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2857 with length and offset values. This means that the output is different
2858 for different internal link sizes. Test 2 is skipped for link sizes
2859 other than 2 because of this, bypassing the problem. Unfortunately,
2860 there was also a test in test 3 (the locale tests) that used /B and
2861 failed for link sizes other than 2. Rather than cut the whole test out,
2862 I have added a new /Z option to pcretest that replaces the length and
2863 offset values with spaces. This is now used to make test 3 independent
2864 of link size. (Test 2 will be tidied up later.)
2865
2866 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2867 segmentation fault instead of returning the appropriate error message.
2868
2869 15. In multiline mode when the newline sequence was set to "any", the pattern
2870 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2871 This doesn't seem right; it now treats the CRLF combination as the line
2872 ending, and so does not match in that case. It's only a pattern such as ^$
2873 that would hit this one: something like ^ABC$ would have failed after \r
2874 and then tried again after \r\n.
2875
2876 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2877 in an attempt to make files that differ only in their line terminators
2878 compare equal. This works on Linux.
2879
2880 17. Under certain error circumstances pcregrep might try to free random memory
2881 as it exited. This is now fixed, thanks to valgrind.
2882
2883 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2884 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2885 was because its rules for how to advance for /g after matching an empty
2886 string at the end of a line did not allow for this case. They now check for
2887 it specially.
2888
2889 20. pcretest is supposed to handle patterns and data of any length, by
2890 extending its buffers when necessary. It was getting this wrong when the
2891 buffer for a data line had to be extended.
2892
2893 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2894 CRLF as a newline sequence.
2895
2896 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2897 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2898 I have nevertheless tidied it up.
2899
2900 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2901
2902 24. Added a man page for pcre-config.
2903
2904
2905 Version 7.0 19-Dec-06
2906 ---------------------
2907
2908 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2909 moving to gcc 4.1.1.
2910
2911 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2912 sys/time.h, which is documented as needed for this function. It doesn't
2913 seem to matter on Linux, but it showed up on some releases of OS X.
2914
2915 3. It seems that there are systems where bytes whose values are greater than
2916 127 match isprint() in the "C" locale. The "C" locale should be the
2917 default when a C program starts up. In most systems, only ASCII printing
2918 characters match isprint(). This difference caused the output from pcretest
2919 to vary, making some of the tests fail. I have changed pcretest so that:
2920
2921 (a) When it is outputting text in the compiled version of a pattern, bytes
2922 other than 32-126 are always shown as hex escapes.
2923
2924 (b) When it is outputting text that is a matched part of a subject string,
2925 it does the same, unless a different locale has been set for the match
2926 (using the /L modifier). In this case, it uses isprint() to decide.
2927
2928 4. Fixed a major bug that caused incorrect computation of the amount of memory
2929 required for a compiled pattern when options that changed within the
2930 pattern affected the logic of the preliminary scan that determines the
2931 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2932 that the computed length was too small. The symptoms of this bug were
2933 either the PCRE error "internal error: code overflow" from pcre_compile(),
2934 or a glibc crash with a message such as "pcretest: free(): invalid next
2935 size (fast)". Examples of patterns that provoked this bug (shown in
2936 pcretest format) are:
2937
2938 /(?-x: )/x
2939 /(?x)(?-x: \s*#\s*)/
2940 /((?i)[\x{c0}])/8
2941 /(?i:[\x{c0}])/8
2942
2943 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2944 is now done differently.
2945
2946 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2947 wrapper classes; (b) implement a new function in the C++ scanner that is
2948 more efficient than the old way of doing things because it avoids levels of
2949 recursion in the regex matching; (c) add a paragraph to the documentation
2950 for the FullMatch() function.
2951
2952 6. The escape sequence \n was being treated as whatever was defined as
2953 "newline". Not only was this contrary to the documentation, which states
2954 that \n is character 10 (hex 0A), but it also went horribly wrong when
2955 "newline" was defined as CRLF. This has been fixed.
2956
2957 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2958 was being set to -1 for the "end of line" case (supposedly a value that no
2959 character can have). Though this value is never used (the check for end of
2960 line is "zero bytes in current character"), it caused compiler complaints.
2961 I've changed it to 0xffffffff.
2962
2963 8. In pcre_version.c, the version string was being built by a sequence of
2964 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2965 string (as it is for production releases) called a macro with an empty
2966 argument. The C standard says the result of this is undefined. The gcc
2967 compiler treats it as an empty string (which was what was wanted) but it is
2968 reported that Visual C gives an error. The source has been hacked around to
2969 avoid this problem.
2970
2971 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2972 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2973 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2974 of them did).
2975
2976 10. Originally, pcretest opened its input and output without "b"; then I was
2977 told that "b" was needed in some environments, so it was added for release
2978 5.0 to both the input and output. (It makes no difference on Unix-like
2979 systems.) Later I was told that it is wrong for the input on Windows. I've
2980 now abstracted the modes into two macros, to make it easier to fiddle with
2981 them, and removed "b" from the input mode under Windows.
2982
2983 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2984
2985 12. Added -help and --help to pcretest as an official way of being reminded
2986 of the options.
2987
2988 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2989 and pcrecpp.cc because they annoy compilers at high warning levels.
2990
2991 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2992
2993 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2994 scripts are not C programs :-) and which was not noticed because it works
2995 on Linux.
2996
2997 16. pcretest is supposed to handle any length of pattern and data line (as one
2998 line or as a continued sequence of lines) by extending its input buffer if
2999 necessary. This feature was broken for very long pattern lines, leading to
3000 a string of junk being passed to pcre_compile() if the pattern was longer
3001 than about 50K.
3002
3003 17. I have done a major re-factoring of the way pcre_compile() computes the
3004 amount of memory needed for a compiled pattern. Previously, there was code
3005 that made a preliminary scan of the pattern in order to do this. That was
3006 OK when PCRE was new, but as the facilities have expanded, it has become
3007 harder and harder to keep it in step with the real compile phase, and there
3008 have been a number of bugs (see for example, 4 above). I have now found a
3009 cunning way of running the real compile function in a "fake" mode that
3010 enables it to compute how much memory it would need, while actually only
3011 ever using a few hundred bytes of working memory and without too many
3012 tests of the mode. This should make future maintenance and development
3013 easier. A side effect of this work is that the limit of 200 on the nesting
3014 depth of parentheses has been removed (though this was never a serious
3015 limitation, I suspect). However, there is a downside: pcre_compile() now
3016 runs more slowly than before (30% or more, depending on the pattern). I
3017 hope this isn't a big issue. There is no effect on runtime performance.
3018
3019 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3020 newline (only possible for the last line of a file) and it was a
3021 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3022
3023 19. Added additional timing features to pcretest. (1) The -tm option now times
3024 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3025 separate command line item, by a number that specifies the number of
3026 repeats to use when timing. The default is 50000; this gives better
3027 precision, but takes uncomfortably long for very large patterns.
3028
3029 20. Extended pcre_study() to be more clever in cases where a branch of a
3030 subpattern has no definite first character. For example, (a*|b*)[cd] would
3031 previously give no result from pcre_study(). Now it recognizes that the
3032 first character must be a, b, c, or d.
3033
3034 21. There was an incorrect error "recursive call could loop indefinitely" if
3035 a subpattern (or the entire pattern) that was being tested for matching an
3036 empty string contained only one non-empty item after a nested subpattern.
3037 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3038 incorrectly, because the \d was being skipped in the check.
3039
3040 22. The pcretest program now has a new pattern option /B and a command line
3041 option -b, which is equivalent to adding /B to every pattern. This causes
3042 it to show the compiled bytecode, without the additional information that
3043 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3044 is the same as /B/I).
3045
3046 23. A new optimization is now able automatically to treat some sequences such
3047 as a*b as a*+b. More specifically, if something simple (such as a character
3048 or a simple class like \d) has an unlimited quantifier, and is followed by
3049 something that cannot possibly match the quantified thing, the quantifier
3050 is automatically "possessified".
3051
3052 24. A recursive reference to a subpattern whose number was greater than 39
3053 went wrong under certain circumstances in UTF-8 mode. This bug could also
3054 have affected the operation of pcre_study().
3055
3056 25. Realized that a little bit of performance could be had by replacing
3057 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3058
3059 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3060
3061 27. Possessive quantifiers such as a++ were previously implemented by turning
3062 them into atomic groups such as ($>a+). Now they have their own opcodes,
3063 which improves performance. This includes the automatically created ones
3064 from 23 above.
3065
3066 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3067 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3068 the first matched character to be a colon. This applied both to named and
3069 numbered groups.
3070
3071 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3072
3073 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3074 building PCRE on Windows easier, so I have included it in the distribution.
3075
3076 31. There is now a check in pcretest against a ridiculously large number being
3077 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3078 loop, the loop is abandoned.
3079
3080 32. Forward references to subpatterns in conditions such as (?(2)...) where
3081 subpattern 2 is defined later cause pcre_compile() to search forwards in
3082 the pattern for the relevant set of parentheses. This search went wrong
3083 when there were unescaped parentheses in a character class, parentheses
3084 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3085
3086 33. "Subroutine" calls and backreferences were previously restricted to
3087 referencing subpatterns earlier in the regex. This restriction has now
3088 been removed.
3089
3090 34. Added a number of extra features that are going to be in Perl 5.10. On the
3091 whole, these are just syntactic alternatives for features that PCRE had
3092 previously implemented using the Python syntax or my own invention. The
3093 other formats are all retained for compatibility.
3094
3095 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3096 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3097 also .NET compatible.
3098
3099 (b) A recursion or subroutine call to a named group can now be defined as
3100 (?&name) as well as (?P>name).
3101
3102 (c) A backreference to a named group can now be defined as \k<name> or
3103 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3104 5.10, are also .NET compatible.
3105
3106 (d) A conditional reference to a named group can now use the syntax
3107 (?(<name>) or (?('name') as well as (?(name).
3108
3109 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3110 groups (named and numbered) that are never evaluated inline, but can be
3111 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3112 is always false. There may be only one alternative in such a group.
3113
3114 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3115 as the simple (?(R). The condition is true only if the most recent
3116 recursion is that of the given number or name. It does not search out
3117 through the entire recursion stack.
3118
3119 (g) The escape \gN or \g{N} has been added, where N is a positive or
3120 negative number, specifying an absolute or relative reference.
3121
3122 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3123 some "unreachable code" warnings.
3124
3125 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3126 things, this adds five new scripts.
3127
3128 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3129 There were also incompatibilities regarding the handling of \Q..\E inside
3130 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3131 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3132
3133 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3134 matches an empty string, and forcibly breaks the loop. There were bugs in
3135 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3136 against aaaa the result was just "a" rather than "aaaa", for example. Two
3137 separate and independent bugs (that affected different cases) have been
3138 fixed.
3139
3140 39. Refactored the code to abolish the use of different opcodes for small
3141 capturing bracket numbers. This is a tidy that I avoided doing when I
3142 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3143 The new approach is not only tidier, it makes it possible to reduce the
3144 memory needed to fix the previous bug (38).
3145
3146 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3147 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3148 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3149 mode.
3150
3151 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3152 report.
3153
3154 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3155 copy construction and assignment in the C++ wrapper.
3156
3157 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3158 couple of bugs that could have given wrong results in the "--newline=crlf"
3159 case.
3160
3161 44. Added a number of casts and did some reorganization of signed/unsigned int
3162 variables following suggestions from Dair Grant. Also renamed the variable
3163 "this" as "item" because it is a C++ keyword.
3164
3165 45. Arranged for dftables to add
3166
3167 #include "pcre_internal.h"
3168
3169 to pcre_chartables.c because without it, gcc 4.x may remove the array
3170 definition from the final binary if PCRE is built into a static library and
3171 dead code stripping is activated.
3172
3173 46. For an unanchored pattern, if a match attempt fails at the start of a
3174 newline sequence, and the newline setting is CRLF or ANY, and the next two
3175 characters are CRLF, advance by two characters instead of one.
3176
3177
3178 Version 6.7 04-Jul-06
3179 ---------------------
3180
3181 1. In order to handle tests when input lines are enormously long, pcretest has
3182 been re-factored so that it automatically extends its buffers when
3183 necessary. The code is crude, but this _is_ just a test program. The
3184 default size has been increased from 32K to 50K.
3185
3186 2. The code in pcre_study() was using the value of the re argument before
3187 testing it for NULL. (Of course, in any sensible call of the function, it
3188 won't be NULL.)
3189
3190 3. The memmove() emulation function in pcre_internal.h, which is used on
3191 systems that lack both memmove() and bcopy() - that is, hardly ever -
3192 was missing a "static" storage class specifier.
3193
3194 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3195 containing an extended class (one that cannot be represented by a bitmap
3196 because it contains high-valued characters or Unicode property items, e.g.
3197 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3198 pattern, but PCRE should not loop if you do not (it no longer does).
3199 [Detail: two cases were found: (a) a repeated subpattern containing an
3200 extended class; (b) a recursive reference to a subpattern that followed a
3201 previous extended class. It wasn't skipping over the extended class
3202 correctly when UTF-8 mode was not set.]
3203
3204 5. A negated single-character class was not being recognized as fixed-length
3205 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3206 compile error "lookbehind assertion is not fixed length".
3207
3208 6. The RunPerlTest auxiliary script was showing an unexpected difference
3209 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3210 write a Perl script that can interpret lines of an input file either as
3211 byte characters or as UTF-8, which is what "perltest" was being required to
3212 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3213 can't do is switch easily at run time between having the "use utf8;" pragma
3214 or not. In the end, I fudged it by using the RunPerlTest script to insert
3215 "use utf8;" explicitly for the UTF-8 tests.
3216
3217 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3218 the end of the subject string, contrary to the documentation and to what
3219 Perl does. This was true of both matching functions. Now it matches only at
3220 the start of the subject and immediately after *internal* newlines.
3221
3222 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3223 a pointer to an int instead of a pointer to an unsigned long int. This
3224 caused problems on 64-bit systems.
3225
3226 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3227 instance of the 'standard' template library not being so standard".
3228
3229 10. There was no check on the number of named subpatterns nor the maximum
3230 length of a subpattern name. The product of these values is used to compute
3231 the size of the memory block for a compiled pattern. By supplying a very
3232 long subpattern name and a large number of named subpatterns, the size
3233 computation could be caused to overflow. This is now prevented by limiting
3234 the length of names to 32 characters, and the number of named subpatterns
3235 to 10,000.
3236
3237 11. Subpatterns that are repeated with specific counts have to be replicated in
3238 the compiled pattern. The size of memory for this was computed from the
3239 length of the subpattern and the repeat count. The latter is limited to
3240 65535, but there was no limit on the former, meaning that integer overflow
3241 could in principle occur. The compiled length of a repeated subpattern is
3242 now limited to 30,000 bytes in order to prevent this.
3243
3244 12. Added the optional facility to have named substrings with the same name.
3245
3246 13. Added the ability to use a named substring as a condition, using the
3247 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3248 are numbers (not recommended). Forward references are permitted.
3249
3250 14. Added forward references in named backreferences (if you see what I mean).
3251
3252 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3253 pattern could run off the end of the subject. For example, the pattern
3254 "(?s)(.{1,5})"8 did this with the subject "ab".
3255
3256 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3257 PCRE_CASELESS was set when matching characters that were quantified with ?
3258 or *.
3259
3260 17. A character class other than a single negated character that had a minimum
3261 but no maximum quantifier - for example [ab]{6,} - was not handled
3262 correctly by pce_dfa_exec(). It would match only one character.
3263
3264 18. A valid (though odd) pattern that looked like a POSIX character
3265 class but used an invalid character after [ (for example [[,abc,]]) caused
3266 pcre_compile() to give the error "Failed: internal error: code overflow" or
3267 in some cases to crash with a glibc free() error. This could even happen if
3268 the pattern terminated after [[ but there just happened to be a sequence of
3269 letters, a binary zero, and a closing ] in the memory that followed.
3270
3271 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3272 over the years. Originally (before any Unicode support), just the bottom 8
3273 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3274 output from "man perlunicode" includes this:
3275
3276 The regular expression compiler produces polymorphic opcodes. That
3277 is, the pattern adapts to the data and automatically switches to
3278 the Unicode character scheme when presented with Unicode data--or
3279 instead uses a traditional byte scheme when presented with byte
3280 data.
3281
3282 Sadly, a wide octal escape does not cause a switch, and in a string with
3283 no other multibyte characters, these octal escapes are treated as before.
3284 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3285 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3286 Unicode string.
3287
3288 I have not perpetrated such confusion in PCRE. Up till now, it took just
3289 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3290 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3291 translate to the appropriate multibyte character.
3292
3293 29. Applied some refactoring to reduce the number of warnings from Microsoft
3294 and Borland compilers. This has included removing the fudge introduced
3295 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3296 a warning about an unused variable.
3297
3298 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3299 characters since release 4.0, because Perl (from release 5.004) does not.
3300 [Or at least, is documented not to: some releases seem to be in conflict
3301 with the documentation.] However, when a pattern was studied with
3302 pcre_study() and all its branches started with \s, PCRE still included VT
3303 as a possible starting character. Of course, this did no harm; it just
3304 caused an unnecessary match attempt.
3305
3306 22. Removed a now-redundant internal flag bit that recorded the fact that case
3307 dependency changed within the pattern. This was once needed for "required
3308 byte" processing, but is no longer used. This recovers a now-scarce options
3309 bit. Also moved the least significant internal flag bit to the most-
3310 significant bit of the word, which was not previously used (hangover from
3311 the days when it was an int rather than a uint) to free up another bit for
3312 the future.
3313
3314 23. Added support for CRLF line endings as well as CR and LF. As well as the
3315 default being selectable at build time, it can now be changed at runtime
3316 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3317 specify that it is scanning data with non-default line endings.
3318
3319 24. Changed the definition of CXXLINK to make it agree with the definition of
3320 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3321
3322 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3323 recursions. This makes a big different to stack usage for some patterns.
3324
3325 26. If a subpattern containing a named recursion or subroutine reference such
3326 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3327 the space required for the compiled pattern went wrong and gave too small a
3328 value. Depending on the environment, this could lead to "Failed: internal
3329 error: code overflow at offset 49" or "glibc detected double free or
3330 corruption" errors.
3331
3332 27. Applied patches from Google (a) to support the new newline modes and (b) to
3333 advance over multibyte UTF-8 characters in GlobalReplace.
3334
3335 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3336 difference for some implementation of PCRE in some Windows version.
3337
3338 29. Added some extra testing facilities to pcretest:
3339
3340 \q<number> in a data line sets the "match limit" value
3341 \Q<number> in a data line sets the "match recursion limt" value
3342 -S <number> sets the stack size, where <number> is in megabytes
3343
3344 The -S option isn't available for Windows.
3345
3346
3347 Version 6.6 06-Feb-06
3348 ---------------------
3349
3350 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3351 in pcreposix.h. I have copied the definition from pcre.h.
3352
3353 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3354 because pcre.h is no longer a built file.
3355
3356 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3357 not normally included in the compiled code.
3358
3359
3360 Version 6.5 01-Feb-06
3361 ---------------------
3362
3363 1. When using the partial match feature with pcre_dfa_exec(), it was not
3364 anchoring the second and subsequent partial matches at the new starting
3365 point. This could lead to incorrect results. For example, with the pattern
3366 /1234/, partially matching against "123" and then "a4" gave a match.
3367
3368 2. Changes to pcregrep:
3369
3370 (a) All non-match returns from pcre_exec() were being treated as failures
3371 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3372 error message is output. Some extra information is given for the
3373 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3374 probably the only errors that are likely to be caused by users (by
3375 specifying a regex that has nested indefinite repeats, for instance).
3376 If there are more than 20 of these errors, pcregrep is abandoned.
3377
3378 (b) A binary zero was treated as data while matching, but terminated the
3379 output line if it was written out. This has been fixed: binary zeroes
3380 are now no different to any other data bytes.
3381
3382 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3383 used to set a locale for matching. The --locale=xxxx long option has
3384 been added (no short equivalent) to specify a locale explicitly on the
3385 pcregrep command, overriding the environment variables.
3386
3387 (d) When -B was used with -n, some line numbers in the output were one less
3388 than they should have been.
3389
3390 (e) Added the -o (--only-matching) option.
3391
3392 (f) If -A or -C was used with -c (count only), some lines of context were
3393 accidentally printed for the final match.
3394
3395 (g) Added the -H (--with-filename) option.
3396
3397 (h) The combination of options -rh failed to suppress file names for files
3398 that were found from directory arguments.
3399
3400 (i) Added the -D (--devices) and -d (--directories) options.
3401
3402 (j) Added the -F (--fixed-strings) option.
3403
3404 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3405
3406 (l) Added the --colo(u)r option.
3407
3408 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3409 is not present by default.
3410
3411 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3412 items such as (?R) or (?1), when the recursion could match a number of
3413 alternatives. If it matched one of the alternatives, but subsequently,
3414 outside the recursion, there was a failure, the code tried to back up into
3415 the recursion. However, because of the way PCRE is implemented, this is not
3416 possible, and the result was an incorrect result from the match.
3417
3418 In order to prevent this happening, the specification of recursion has
3419 been changed so that all such subpatterns are automatically treated as
3420 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3421
3422 4. I had overlooked the fact that, in some locales, there are characters for
3423 which isalpha() is true but neither isupper() nor islower() are true. In
3424 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3425 and ordfeminine) are like this. This affected the treatment of \w and \W
3426 when they appeared in character classes, but not when they appeared outside
3427 a character class. The bit map for "word" characters is now created
3428 separately from the results of isalnum() instead of just taking it from the
3429 upper, lower, and digit maps. (Plus the underscore character, of course.)
3430
3431 5. The above bug also affected the handling of POSIX character classes such as
3432 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3433 permanent tables. Instead, the bit maps for such a class were previously
3434 created as the appropriate unions of the upper, lower, and digit bitmaps.
3435 Now they are created by subtraction from the [[:word:]] class, which has
3436 its own bitmap.
3437
3438 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3439 It is created by subtracting the vertical space characters (\x09, \x0a,
3440 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3441 subtraction was done in the overall bitmap for a character class, meaning
3442 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3443 be recognized. This bug has been fixed.
3444
3445 7. Patches from the folks at Google:
3446
3447 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3448 real life, but is still worth protecting against".
3449
3450 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3451 regular expressions".
3452
3453 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3454 have it.
3455
3456 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3457 "configure" and the latter not, in order to fix a problem somebody had
3458 with compiling the Arg class on HP-UX.
3459
3460 (e) Improve the error-handling of the C++ wrapper a little bit.
3461
3462 (f) New tests for checking recursion limiting.
3463
3464 8. The pcre_memmove() function, which is used only if the environment does not
3465 have a standard memmove() function (and is therefore rarely compiled),
3466 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3467 returning a result (though PCRE never actually uses the result).
3468
3469 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3470 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3471 returned instead of calling malloc() with an overflowing number that would
3472 most likely cause subsequent chaos.
3473
3474 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3475
3476 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3477 with this option is matched, the nmatch and pmatch options of regexec() are
3478 ignored.
3479
3480 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3481 provided in case anyone wants to the the POSIX interface with UTF-8
3482 strings.
3483
3484 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3485 C++ linking (needed for some HP-UX environments).
3486
3487 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3488 (unused parameter) and in the pcre_printint() function (omitted "default"
3489 switch label when the default is to do nothing).
3490
3491 15. Added some code to make it possible, when PCRE is compiled as a C++
3492 library, to replace subject pointers for pcre_exec() with a smart pointer
3493 class, thus making it possible to process discontinuous strings.
3494
3495 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3496 much the same function. They were added by different people who were trying
3497 to make PCRE easy to compile on non-Unix systems. It has been suggested
3498 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3499 for compiling on Windows systems. I have therefore replaced it with
3500 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3501 defaults to "extern" for C or "extern C" for C++, which works fine on
3502 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3503 SCOPE with something explicit in config.h. In addition:
3504
3505 (a) pcreposix.h still had just "extern" instead of either of these macros;
3506 I have replaced it with PCRE_DATA_SCOPE.
3507
3508 (b) Functions such as _pcre_xclass(), which are internal to the library,
3509 but external in the C sense, all had PCRE_EXPORT in their definitions.
3510 This is apparently wrong for the Windows case, so I have removed it.
3511 (It makes no difference on Unix-like systems.)
3512
3513 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3514 of recursive calls to match(). This is different to MATCH_LIMIT because
3515 that limits the total number of calls to match(), not all of which increase
3516 the depth of recursion. Limiting the recursion depth limits the amount of
3517 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3518 when PCRE is compiled, and changed at run time. A patch from Google adds
3519 this functionality to the C++ interface.
3520
3521 18. Changes to the handling of Unicode character properties:
3522
3523 (a) Updated the table to Unicode 4.1.0.
3524
3525 (b) Recognize characters that are not in the table as "Cn" (undefined).
3526
3527 (c) I revised the way the table is implemented to a much improved format
3528 which includes recognition of ranges. It now supports the ranges that
3529 are defined in UnicodeData.txt, and it also amalgamates other
3530 characters into ranges. This has reduced the number of entries in the
3531 table from around 16,000 to around 3,000, thus reducing its size
3532 considerably. I realized I did not need to use a tree structure after
3533 all - a binary chop search is just as efficient. Having reduced the
3534 number of entries, I extended their size from 6 bytes to 8 bytes to
3535 allow for more data.
3536
3537 (d) Added support for Unicode script names via properties such as \p{Han}.
3538
3539 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3540 matching that character.
3541
3542 20. When matching a repeated Unicode property with a minimum greater than zero,
3543 (for example \pL{2,}), PCRE could look past the end of the subject if it
3544 reached it while seeking the minimum number of characters. This could
3545 happen only if some of the characters were more than one byte long, because
3546 there is a check for at least the minimum number of bytes.
3547
3548 21. Refactored the implementation of \p and \P so as to be more general, to
3549 allow for more different types of property in future. This has changed the
3550 compiled form incompatibly. Anybody with saved compiled patterns that use
3551 \p or \P will have to recompile them.
3552
3553 22. Added "Any" and "L&" to the supported property types.
3554
3555 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3556 but give a compile time error if the value is greater than 0xff.
3557
3558 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3559 accidentally not being installed or uninstalled.
3560
3561 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3562 made were to insert the current release number. This seemed silly, because
3563 it made things harder for people building PCRE on systems that don't run
3564 "configure". I have turned pcre.h into a distributed file, no longer built
3565 by "configure", with the version identification directly included. There is
3566 no longer a pcre.h.in file.
3567
3568 However, this change necessitated a change to the pcre-config script as
3569 well. It is built from pcre-config.in, and one of the substitutions was the
3570 release number. I have updated configure.ac so that ./configure now finds
3571 the release number by grepping pcre.h.
3572
3573 26. Added the ability to run the tests under valgrind.
3574
3575
3576 Version 6.4 05-Sep-05
3577 ---------------------
3578
3579 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3580 "--" to be printed when multiple files were scanned, even when none of the
3581 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3582 consider it to be a bug, and have restored the previous behaviour.
3583
3584 2. A couple of code tidies to get rid of compiler warnings.
3585
3586 3. The pcretest program used to cheat by referring to symbols in the library
3587 whose names begin with _pcre_. These are internal symbols that are not
3588 really supposed to be visible externally, and in some environments it is
3589 possible to suppress them. The cheating is now confined to including
3590 certain files from the library's source, which is a bit cleaner.
3591
3592 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3593 file's purpose clearer.
3594
3595 5. Reorganized pcre_ucp_findchar().
3596
3597
3598 Version 6.3 15-Aug-05
3599 ---------------------
3600
3601 1. The file libpcre.pc.in did not have general read permission in the tarball.
3602
3603 2. There were some problems when building without C++ support:
3604
3605 (a) If C++ support was not built, "make install" and "make test" still
3606 tried to test it.
3607
3608 (b) There were problems when the value of CXX was explicitly set. Some
3609 changes have been made to try to fix these, and ...
3610
3611 (c) --disable-cpp can now be used to explicitly disable C++ support.
3612
3613 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3614 backslash in a target when C++ was disabled. This confuses some
3615 versions of "make", apparently. Using an intermediate variable solves
3616 this. (Same for CPP_LOBJ.)
3617
3618 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3619 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3620 necessary on certain architectures.
3621
3622 4. Added a setting of -export-symbols-regex to the link command to remove
3623 those symbols that are exported in the C sense, but actually are local
3624 within the library, and not documented. Their names all begin with
3625 "_pcre_". This is not a perfect job, because (a) we have to except some
3626 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3627 available (and never for static libraries). I have made a note to try to
3628 find a way round (a) in the future.
3629
3630
3631 Version 6.2 01-Aug-05
3632 ---------------------
3633
3634 1. There was no test for integer overflow of quantifier values. A construction
3635 such as {1111111111111111} would give undefined results. What is worse, if
3636 a minimum quantifier for a parenthesized subpattern overflowed and became
3637 negative, the calculation of the memory size went wrong. This could have
3638 led to memory overwriting.
3639
3640 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3641
3642 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3643 operating environments where this matters.
3644
3645 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3646 PCRE options from within the C++ wrapper.
3647
3648 5. Named capturing subpatterns were not being correctly counted when a pattern
3649 was compiled. This caused two problems: (a) If there were more than 100
3650 such subpatterns, the calculation of the memory needed for the whole
3651 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3652 back references of the form \12, where the number was greater than 9, were
3653 not recognized as back references, even though there were sufficient
3654 previous subpatterns.
3655
3656 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3657 versions of gcc, e.g. 2.95.4.
3658
3659
3660 Version 6.1 21-Jun-05
3661 ---------------------
3662
3663 1. There was one reference to the variable "posix" in pcretest.c that was not
3664 surrounded by "#if !defined NOPOSIX".
3665
3666 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3667 the cross-check on the old pcre_info() function, for the benefit of the
3668 cut-down version of PCRE that is currently imported into Exim.
3669
3670 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3671 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3672 patterns that start (?i) but I don't think that matters. The use of (?i) is
3673 just an example; this all applies to the other options as well.
3674
3675 4. Since libtool seems to echo the compile commands it is issuing, the output
3676 from "make" can be reduced a bit by putting "@" in front of each libtool
3677 compile command.
3678
3679 5. Patch from the folks at Google for configure.in to be a bit more thorough
3680 in checking for a suitable C++ installation before trying to compile the
3681 C++ stuff. This should fix a reported problem when a compiler was present,
3682 but no suitable headers.
3683
3684 6. The man pages all had just "PCRE" as their title. I have changed them to
3685 be the relevant file name. I have also arranged that these names are
3686 retained in the file doc/pcre.txt, which is a concatenation in text format
3687 of all the man pages except the little individual ones for each function.
3688
3689 7. The NON-UNIX-USE file had not been updated for the different set of source
3690 files that come with release 6. I also added a few comments about the C++
3691 wrapper.
3692
3693
3694 Version 6.0 07-Jun-05
3695 ---------------------
3696
3697 1. Some minor internal re-organization to help with my DFA experiments.
3698
3699 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3700 didn't matter for the library itself when fully configured, but did matter
3701 when compiling without UCP support, or within Exim, where the ucp files are
3702 not imported.
3703
3704 3. Refactoring of the library code to split up the various functions into
3705 different source modules. The addition of the new DFA matching code (see
3706 below) to a single monolithic source would have made it really too
3707 unwieldy, quite apart from causing all the code to be include in a
3708 statically linked application, when only some functions are used. This is
3709 relevant even without the DFA addition now that patterns can be compiled in
3710 one application and matched in another.
3711
3712 The downside of splitting up is that there have to be some external
3713 functions and data tables that are used internally in different modules of
3714 the library but which are not part of the API. These have all had their
3715 names changed to start with "_pcre_" so that they are unlikely to clash
3716 with other external names.
3717
3718 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3719 a different (DFA) algorithm. Although it is slower than the original
3720 function, it does have some advantages for certain types of matching
3721 problem.
3722
3723 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3724 including restarting after a partial match.
3725
3726 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3727 defined when compiling for Windows was sent to me. I have put it into the
3728 code, though I have no means of testing or verifying it.
3729
3730 7. Added the pcre_refcount() auxiliary function.
3731
3732 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3733 match before or at the first newline in the subject string. In pcretest,
3734 the /f option on a pattern can be used to set this.
3735
3736 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3737 would behave wrongly. This has been present in PCRE since release 4.0.
3738
3739 10. A number of changes to the pcregrep command:
3740
3741 (a) Refactored how -x works; insert ^(...)$ instead of setting
3742 PCRE_ANCHORED and checking the length, in preparation for adding
3743 something similar for -w.
3744
3745 (b) Added the -w (match as a word) option.
3746
3747 (c) Refactored the way lines are read and buffered so as to have more
3748 than one at a time available.
3749
3750 (d) Implemented a pcregrep test script.
3751
3752 (e) Added the -M (multiline match) option. This allows patterns to match
3753 over several lines of the subject. The buffering ensures that at least
3754 8K, or the rest of the document (whichever is the shorter) is available
3755 for matching (and similarly the previous 8K for lookbehind assertions).
3756
3757 (f) Changed the --help output so that it now says
3758
3759 -w, --word-regex(p)
3760
3761 instead of two lines, one with "regex" and the other with "regexp"
3762 because that confused at least one person since the short forms are the
3763 same. (This required a bit of code, as the output is generated
3764 automatically from a table. It wasn't just a text change.)
3765
3766 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3767 option but starts with a hyphen. Could be a pattern or a path name
3768 starting with a hyphen, for instance.
3769
3770 (h) "-" can be given as a file name to represent stdin.
3771
3772 (i) When file names are being printed, "(standard input)" is used for
3773 the standard input, for compatibility with GNU grep. Previously
3774 "<stdin>" was used.
3775
3776 (j) The option --label=xxx can be used to supply a name to be used for
3777 stdin when file names are being printed. There is no short form.
3778
3779 (k) Re-factored the options decoding logic because we are going to add
3780 two more options that take data. Such options can now be given in four
3781 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3782
3783 (l) Added the -A, -B, and -C options for requesting that lines of context
3784 around matches be printed.
3785
3786 (m) Added the -L option to print the names of files that do not contain
3787 any matching lines, that is, the complement of -l.
3788
3789 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3790 continue to scan other files.
3791
3792 (o) The -s option was incorrectly implemented. For compatibility with other
3793 greps, it now suppresses the error message for a non-existent or non-
3794 accessible file (but not the return code). There is a new option called
3795 -q that suppresses the output of matching lines, which was what -s was
3796 previously doing.
3797
3798 (p) Added --include and --exclude options to specify files for inclusion
3799 and exclusion when recursing.
3800
3801 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3802 Hopefully, it now does.
3803
3804 12. Missing cast in pcre_study().
3805
3806 13. Added an "uninstall" target to the makefile.
3807
3808 14. Replaced "extern" in the function prototypes in Makefile.in with
3809 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3810 world, but is set differently for Windows.
3811
3812 15. Added a second compiling function called pcre_compile2(). The only
3813 difference is that it has an extra argument, which is a pointer to an
3814 integer error code. When there is a compile-time failure, this is set
3815 non-zero, in addition to the error test pointer being set to point to an
3816 error message. The new argument may be NULL if no error number is required
3817 (but then you may as well call pcre_compile(), which is now just a
3818 wrapper). This facility is provided because some applications need a
3819 numeric error indication, but it has also enabled me to tidy up the way
3820 compile-time errors are handled in the POSIX wrapper.
3821
3822 16. Added VPATH=.libs to the makefile; this should help when building with one
3823 prefix path and installing with another. (Or so I'm told by someone who
3824 knows more about this stuff than I do.)
3825
3826 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3827 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3828 match everything, including newlines. This is not POSIX-compatible, but
3829 somebody wanted the feature. From pcretest it can be activated by using
3830 both the P and the s flags.
3831
3832 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3833
3834 19. libpcre.pc was being incorrectly installed as executable.
3835
3836 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3837 it now also looks for '\r' so that it will work unmodified on Windows.
3838
3839 21. Added Google's contributed C++ wrapper to the distribution.
3840
3841 22. Added some untidy missing memory free() calls in pcretest, to keep
3842 Electric Fence happy when testing.
3843
3844
3845
3846 Version 5.0 13-Sep-04
3847 ---------------------
3848
3849 1. Internal change: literal characters are no longer packed up into items
3850 containing multiple characters in a single byte-string. Each character
3851 is now matched using a separate opcode. However, there may be more than one
3852 byte in the character in UTF-8 mode.
3853
3854 2. The pcre_callout_block structure has two new fields: pattern_position and
3855 next_item_length. These contain the offset in the pattern to the next match
3856 item, and its length, respectively.
3857
3858 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3859 insertion of callouts before each pattern item. Added the /C option to
3860 pcretest to make use of this.
3861
3862 4. On the advice of a Windows user, the lines
3863
3864 #if defined(_WIN32) || defined(WIN32)
3865 _setmode( _fileno( stdout ), 0x8000 );
3866 #endif /* defined(_WIN32) || defined(WIN32) */
3867
3868 have been added to the source of pcretest. This apparently does useful
3869 magic in relation to line terminators.
3870
3871 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3872 for the benefit of those environments where the "b" makes a difference.
3873
3874 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3875 to know about it. I have put a hack into configure.in that adds in code
3876 to set GCC=yes if CC=icc. This seems to end up at a point in the
3877 generated configure script that is early enough to affect the setting of
3878 compiler options, which is what is needed, but I have no means of testing
3879 whether it really works. (The user who reported this had patched the
3880 generated configure script, which of course I cannot do.)
3881
3882 LATER: After change 22 below (new libtool files), the configure script
3883 seems to know about icc (and also ecc). Therefore, I have commented out
3884 this hack in configure.in.
3885
3886 7. Added support for pkg-config (2 patches were sent in).
3887
3888 8. Negated POSIX character classes that used a combination of internal tables
3889 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3890 [[:^ascii]]. Typically, they would match almost any characters. The other
3891 POSIX classes were not broken in this way.
3892
3893 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3894 to find the match, as PCRE was deluded into thinking that the match had to
3895 start at the start point or following a newline. The same bug applied to
3896 patterns with negative forward assertions or any backward assertions
3897 preceding ".*" at the start, unless the pattern required a fixed first
3898 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3899
3900 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3901 starting at the last subject character, bytes beyond the end of the subject
3902 string were read.
3903
3904 11. Renamed the variable "class" as "classbits" to make life easier for C++
3905 users. (Previously there was a macro definition, but it apparently wasn't
3906 enough.)
3907
3908 12. Added the new field "tables" to the extra data so that tables can be passed
3909 in at exec time, or the internal tables can be re-selected. This allows
3910 a compiled regex to be saved and re-used at a later time by a different
3911 program that might have everything at different addresses.
3912
3913 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3914 -R library as well as a -L library.
3915
3916 14. The debugging options of pcretest (-d on the command line or D on a
3917 pattern) showed incorrect output for anything following an extended class
3918 that contained multibyte characters and which was followed by a quantifier.
3919
3920 15. Added optional support for general category Unicode character properties
3921 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3922 support. It adds about 90K to the size of the library. The meanings of the
3923 inbuilt class escapes such as \d and \s have NOT been changed.
3924
3925 16. Updated pcredemo.c to include calls to free() to release the memory for the
3926 compiled pattern.
3927
3928 17. The generated file chartables.c was being created in the source directory
3929 instead of in the building directory. This caused the build to fail if the
3930 source directory was different from the building directory, and was
3931 read-only.
3932
3933 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3934 file. No doubt somebody will tell me if they don't make sense... Also added
3935 Dan Mooney's comments about building on OpenVMS.
3936
3937 19. Added support for partial matching via the PCRE_PARTIAL option for
3938 pcre_exec() and the \P data escape in pcretest.
3939
3940 20. Extended pcretest with 3 new pattern features:
3941
3942 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3943 write the compiled pattern to the file whose name is "rest-of-line".
3944 This is a straight binary dump of the data, with the saved pointer to
3945 the character tables forced to be NULL. The study data, if any, is
3946 written too. After writing, pcretest reads a new pattern.
3947
3948 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3949 compiled pattern from the given file. There must not be any
3950 occurrences of "<" in the file name (pretty unlikely); if there are,
3951 pcretest will instead treat the initial "<" as a pattern delimiter.
3952 After reading in the pattern, pcretest goes on to read data lines as
3953 usual.
3954
3955 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3956 and 16-bit fields in a compiled pattern, to simulate a pattern that
3957 was compiled on a host of opposite endianness.
3958
3959 21. The pcre-exec() function can now cope with patterns that were compiled on
3960 hosts of opposite endianness, with this restriction:
3961
3962 As for any compiled expression that is saved and used later, the tables
3963 pointer field cannot be preserved; the extra_data field in the arguments
3964 to pcre_exec() should be used to pass in a tables address if a value
3965 other than the default internal tables were used at compile time.
3966
3967 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3968 now diagnosed as an error. Previously, most of the time, a negative number
3969 would have been treated as zero, but if in addition "ovector" was passed as
3970 NULL, a crash could occur.
3971
3972 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3973 new versions from the libtool 1.5 distribution (the last one is a copy of
3974 a file called libtool.m4). This seems to have fixed the need to patch
3975 "configure" to support Darwin 1.3 (which I used to do). However, I still
3976 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3977 workstation).
3978
3979 24. Changed the PCRE licence to be the more standard "BSD" licence.
3980
3981
3982 Version 4.5 01-Dec-03
3983 ---------------------
3984
3985 1. There has been some re-arrangement of the code for the match() function so
3986 that it can be compiled in a version that does not call itself recursively.
3987 Instead, it keeps those local variables that need separate instances for
3988 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3989 needs to "recurse". Keeping track of where control must go is done by means
3990 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3991 hide most of the details from the main code, and operates only if
3992 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3993 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3994 operating.
3995
3996 To make it easier for callers to provide specially tailored get/free
3997 functions for this usage, two new functions, pcre_stack_malloc, and
3998 pcre_stack_free, are used. They are always called in strict stacking order,
3999 and the size of block requested is always the same.
4000
4001 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4002 PCRE has been compiled to use the stack or the heap for recursion. The
4003 -C option of pcretest uses this to show which version is compiled.
4004
4005 A new data escape \S, is added to pcretest; it causes the amounts of store
4006 obtained and freed by both kinds of malloc/free at match time to be added
4007 to the output.
4008
4009 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4010 what's available on my current Linux desktop machine.
4011
4012 3. When matching a UTF-8 string, the test for a valid string at the start has
4013 been extended. If start_offset is not zero, PCRE now checks that it points
4014 to a byte that is the start of a UTF-8 character. If not, it returns
4015 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4016 this is necessary because there may be backward assertions in the pattern.
4017 When matching the same subject several times, it may save resources to use
4018 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4019
4020 4. The code for checking the validity of UTF-8 strings has been tightened so
4021 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4022 containing "overlong sequences".
4023
4024 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4025 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4026 should have been "&", but it just so happened that all the cases this let
4027 through by mistake were picked up later in the function.
4028
4029 6. I had used a variable called "isblank" - this is a C99 function, causing
4030 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4031
4032 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4033 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4034 so the version is shown; (c) stop "make test" echoing "./RunTest".
4035
4036 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4037
4038 9. The prototype for memmove() for systems that don't have it was using
4039 size_t, but the inclusion of the header that defines size_t was later. I've
4040 moved the #includes for the C headers earlier to avoid this.
4041
4042 10. Added some adjustments to the code to make it easier to compiler on certain
4043 special systems:
4044
4045 (a) Some "const" qualifiers were missing.
4046 (b) Added the macro EXPORT before all exported functions; by default this
4047 is defined to be empty.
4048 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4049 that it reads its output file name as an argument instead of writing
4050 to the standard output and assuming this can be redirected.
4051
4052 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4053 class containing characters with values greater than 255, PCRE compilation
4054 went into a loop.
4055
4056 12. A recursive reference to a subpattern that was within another subpattern
4057 that had a minimum quantifier of zero caused PCRE to crash. For example,
4058 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4059 recursion. If the recursively-called subpattern itself had a zero repeat,
4060 that was OK.
4061
4062 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4063 buffer into which it was copied (for escape processing) was still set at
4064 1024, so long lines caused crashes.
4065
4066 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4067 "internal error: code overflow...". This applied to any character class
4068 that was followed by a possessive quantifier.
4069
4070 15. Modified the Makefile to add libpcre.la as a prerequisite for
4071 libpcreposix.la because I was told this is needed for a parallel build to
4072 work.
4073
4074 16. If a pattern that contained .* following optional items at the start was
4075 studied, the wrong optimizing data was generated, leading to matching
4076 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4077 matching string must start with a or b or c. The correct conclusion for
4078 this pattern is that a match can start with any character.
4079
4080
4081 Version 4.4 13-Aug-03
4082 ---------------------
4083
4084 1. In UTF-8 mode, a character class containing characters with values between
4085 127 and 255 was not handled correctly if the compiled pattern was studied.
4086 In fixing this, I have also improved the studying algorithm for such
4087 classes (slightly).
4088
4089 2. Three internal functions had redundant arguments passed to them. Removal
4090 might give a very teeny performance improvement.
4091
4092 3. Documentation bug: the value of the capture_top field in a callout is *one
4093 more than* the number of the hightest numbered captured substring.
4094
4095 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4096 in incorrectly linking with a previously installed version. They now link
4097 explicitly with libpcre.la.
4098
4099 5. configure.in no longer needs to recognize Cygwin specially.
4100
4101 6. A problem in pcre.in for Windows platforms is fixed.
4102
4103 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4104 pcretest, it used to include the size of the study block as part of its
4105 output. Unfortunately, the structure contains a field that has a different
4106 size on different hardware architectures. This meant that the tests that
4107 showed this size failed. As the block is currently always of a fixed size,
4108 this information isn't actually particularly useful in pcretest output, so
4109 I have just removed it.
4110
4111 8. Three pre-processor statements accidentally did not start in column 1.
4112 Sadly, there are *still* compilers around that complain, even though
4113 standard C has not required this for well over a decade. Sigh.
4114
4115 9. In pcretest, the code for checking callouts passed small integers in the
4116 callout_data field, which is a void * field. However, some picky compilers
4117 complained about the casts involved for this on 64-bit systems. Now
4118 pcretest passes the address of the small integer instead, which should get
4119 rid of the warnings.
4120
4121 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4122 both compile and run time, and gives an error if an invalid UTF-8 sequence
4123 is found. There is a option for disabling this check in cases where the
4124 string is known to be correct and/or the maximum performance is wanted.
4125
4126 11. In response to a bug report, I changed one line in Makefile.in from
4127
4128 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4129 to
4130 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4131
4132 to look similar to other lines, but I have no way of telling whether this
4133 is the right thing to do, as I do not use Windows. No doubt I'll get told
4134 if it's wrong...
4135
4136
4137 Version 4.3 21-May-03
4138 ---------------------
4139
4140 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4141 Makefile.
4142
4143 2. Some refactoring to improve the quality of the code:
4144
4145 (i) The utf8_table... variables are now declared "const".
4146
4147 (ii) The code for \cx, which used the "case flipping" table to upper case
4148 lower case letters, now just substracts 32. This is ASCII-specific,
4149 but the whole concept of \cx is ASCII-specific, so it seems
4150 reasonable.
4151
4152 (iii) PCRE was using its character types table to recognize decimal and
4153 hexadecimal digits in the pattern. This is silly, because it handles
4154 only 0-9, a-f, and A-F, but the character types table is locale-
4155 specific, which means strange things might happen. A private
4156 table is now used for this - though it costs 256 bytes, a table is
4157 much faster than multiple explicit tests. Of course, the standard
4158 character types table is still used for matching digits in subject
4159 strings against \d.
4160
4161 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4162 ending in _t are). So I've renamed it as ESC_tee.
4163
4164 3. The first argument for regexec() in the POSIX wrapper should have been
4165 defined as "const".
4166
4167 4. Changed pcretest to use malloc() for its buffers so that they can be
4168 Electric Fenced for debugging.
4169
4170 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4171 to read one or more bytes before the start of the subject string. Often this
4172 had no effect on PCRE's behaviour, but in some circumstances it could
4173 provoke a segmentation fault.
4174
4175 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4176 to try to read one or more bytes before the start of the subject string.
4177
4178 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4179 UTF-8 support could misbehave in various ways if the subject string
4180 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4181 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4182 back over UTF-8 characters.)
4183
4184
4185 Version 4.2 14-Apr-03
4186 ---------------------
4187
4188 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4189
4190 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4191 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4192 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4193 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4194 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4195 and BUILD_EXEEXT
4196 Note: automatic setting of the BUILD variables is not yet working
4197 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4198 compile-time but not at link-time
4199 [LINK]: use for linking executables only
4200 make different versions for Windows and non-Windows
4201 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4202 libraries
4203 [LINK_FOR_BUILD]: new variable
4204 [OBJEXT]: use throughout
4205 [EXEEXT]: use throughout
4206 <winshared>: new target
4207 <wininstall>: new target
4208 <dftables.o>: use native compiler
4209 <dftables>: use native linker
4210 <install>: handle Windows platform correctly
4211 <clean>: ditto
4212 <check>: ditto
4213 copy DLL to top builddir before testing
4214
4215 As part of these changes, -no-undefined was removed again. This was reported
4216 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4217 in any case.
4218
4219 3. Some tidies to get rid of compiler warnings:
4220
4221 . In the match_data structure, match_limit was an unsigned long int, whereas
4222 match_call_count was an int. I've made them both unsigned long ints.
4223
4224 . In pcretest the fact that a const uschar * doesn't automatically cast to
4225 a void * provoked a warning.
4226
4227 . Turning on some more compiler warnings threw up some "shadow" variables
4228 and a few more missing casts.
4229
4230 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4231 option, a class that contained a single character with a value between 128
4232 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4233
4234 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4235 option, a class that contained several characters, but with at least one
4236 whose value was between 128 and 255 caused PCRE to crash.
4237
4238
4239 Version 4.1 12-Mar-03
4240 ---------------------
4241
4242 1. Compiling with gcc -pedantic found a couple of places where casts were
4243 needed, and a string in dftables.c that was longer than standard compilers are
4244 required to support.
4245
4246 2. Compiling with Sun's compiler found a few more places where the code could
4247 be tidied up in order to avoid warnings.
4248
4249 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4250 first of these names is deprecated in the latest Autoconf in favour of the name
4251 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4252 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4253 analogy I have changed it to CFLAGS_FOR_BUILD.
4254
4255 4. Added -no-undefined to the linking command in the Makefile, because this is
4256 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4257 linking step for the pcreposix library.
4258
4259 5. PCRE was failing to diagnose the case of two named groups with the same
4260 name.
4261
4262 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4263 literal character that is needed in the subject for a match, and scans along to
4264 ensure that it is present before embarking on the full matching process. This
4265 saves time in cases of nested unlimited repeats that are never going to match.
4266 Problem: the scan can take a lot of time if the subject is very long (e.g.
4267 megabytes), thus penalizing straightforward matches. It is now done only if the
4268 amount of subject to be scanned is less than 1000 bytes.
4269
4270 7. A lesser problem with the same optimization is that it was recording the
4271 first character of an anchored pattern as "needed", thus provoking a search
4272 right along the subject, even when the first match of the pattern was going to
4273 fail. The "needed" character is now not set for anchored patterns, unless it
4274 follows something in the pattern that is of non-fixed length. Thus, it still
4275 fulfils its original purpose of finding quick non-matches in cases of nested
4276 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4277
4278
4279 Version 4.0 17-Feb-03
4280 ---------------------
4281
4282 1. If a comment in an extended regex that started immediately after a meta-item
4283 extended to the end of string, PCRE compiled incorrect data. This could lead to
4284 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4285
4286 2. Moved to autoconf 2.53 and libtool 1.4.2.
4287
4288 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4289 the special perltest8 script is no longer needed - all the tests can be run
4290 from a single perltest script.
4291
4292 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4293 by \s. It has now been removed in PCRE. This means it isn't recognized as
4294 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4295 class [:space:] *does* include VT, thereby creating a mess.
4296
4297 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4298 space and tab.
4299
4300 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4301 its new features into the main test script, reducing the number of scripts.
4302
4303 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4304 were backward compatible, and made the (?i) apply to the whole pattern, as if
4305 /i were given. Now it behaves more logically, and applies the option setting
4306 only to what follows. PCRE has been changed to follow suit. However, if it
4307 finds options settings right at the start of the pattern, it extracts them into
4308 the global options, as before. Thus, they show up in the info data.
4309
4310 8. Added support for the \Q...\E escape sequence. Characters in between are
4311 treated as literals. This is slightly different from Perl in that $ and @ are
4312 also handled as literals inside the quotes. In Perl, they will cause variable
4313 interpolation. Note the following examples:
4314
4315 Pattern PCRE matches Perl matches
4316
4317 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4318 \Qabc\$xyz\E abc\$xyz abc\$xyz
4319 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4320
4321 For compatibility with Perl, \Q...\E sequences are recognized inside character
4322 classes as well as outside them.
4323
4324 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4325 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4326 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4327 signed/unsigned warnings.
4328
4329 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4330 option for pcretest, so I've replaced it by a simple function that does just
4331 that job.
4332
4333 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4334 "pcregrep -".
4335
4336 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4337 Java package. This provides some syntactic sugar for simple cases of what my
4338 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4339 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4340 item, you can use this simplified notation. Note that only makes sense with
4341 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4342 greediness, whatever the setting of the PCRE_UNGREEDY option.
4343
4344 13. A change of greediness default within a pattern was not taking effect at
4345 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4346 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4347 was abstracted outside.
4348
4349 14. PCRE now supports the \G assertion. It is true when the current matching
4350 position is at the start point of the match. This differs from \A when the
4351 starting offset is non-zero. Used with the /g option of pcretest (or similar
4352 code), it works in the same way as it does for Perl's /g option. If all
4353 alternatives of a regex begin with \G, the expression is anchored to the start
4354 match position, and the "anchored" flag is set in the compiled expression.
4355
4356 15. Some bugs concerning the handling of certain option changes within patterns
4357 have been fixed. These applied to options other than (?ims). For example,
4358 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4359 been the other way round. Some of this was related to change 7 above.
4360
4361 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4362 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4363 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4364 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4365
4366 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4367 mode. Unlike ".", it always matches newline, whatever the setting of
4368 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4369 assertions. Perl allows it, but it doesn't (in general) work because it can't
4370 calculate the length of the lookbehind. At least, that's the case for Perl
4371 5.8.0 - I've been told they are going to document that it doesn't work in
4372 future.
4373
4374 18. Added an error diagnosis for escapes that PCRE does not support: these are
4375 \L, \l, \N, \P, \p, \U, \u, and \X.
4376
4377 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4378 reading past the end of the pattern in cases such as /[abcd/.
4379
4380 20. PCRE was getting more memory than necessary for patterns with classes that
4381 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4382
4383 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4384 compiling PCRE for use with Virtual Pascal.
4385
4386 22. Small fix to the Makefile to make it work properly if the build is done
4387 outside the source tree.
4388
4389 23. Added a new extension: a condition to go with recursion. If a conditional
4390 subpattern starts with (?(R) the "true" branch is used if recursion has
4391 happened, whereas the "false" branch is used only at the top level.
4392
4393 24. When there was a very long string of literal characters (over 255 bytes
4394 without UTF support, over 250 bytes with UTF support), the computation of how
4395 much memory was required could be incorrect, leading to segfaults or other
4396 strange effects.
4397
4398 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4399 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4400 there was a subsequent back reference to those brackets. This meant that, for
4401 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4402 possible to check for precisely this case. All we can do is abandon the
4403 optimization if .* occurs inside capturing brackets when there are any back
4404 references whatsoever. (See below for a better fix that came later.)
4405
4406 26. The handling of the optimization for finding the first character of a
4407 non-anchored pattern, and for finding a character that is required later in the
4408 match were failing in some cases. This didn't break the matching; it just
4409 failed to optimize when it could. The way this is done has been re-implemented.
4410
4411 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4412
4413 28. Added a new feature that provides some of the functionality that Perl
4414 provides with (?{...}). The facility is termed a "callout". The way it is done
4415 in PCRE is for the caller to provide an optional function, by setting
4416 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4417 global variable. By default it is unset, which disables all calling out. To get
4418 the function called, the regex must include (?C) at appropriate points. This
4419 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4420 This provides a means of identifying different callout points. When PCRE
4421 reaches such a point in the regex, if pcre_callout has been set, the external
4422 function is called. It is provided with data in a structure called
4423 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4424 matching continues; if it returns a non-zero value, the match at the current
4425 point fails. However, backtracking will occur if possible. [This was changed
4426 later and other features added - see item 49 below.]
4427
4428 29. pcretest is upgraded to test the callout functionality. It provides a
4429 callout function that displays information. By default, it shows the start of
4430 the match and the current position in the text. There are some new data escapes
4431 to vary what happens:
4432
4433 \C+ in addition, show current contents of captured substrings
4434 \C- do not supply a callout function
4435 \C!n return 1 when callout number n is reached
4436 \C!n!m return 1 when callout number n is reached for the mth time
4437
4438 30. If pcregrep was called with the -l option and just a single file name, it
4439 output "<stdin>" if a match was found, instead of the file name.
4440
4441 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4442 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4443 pcre_exec(). This saves a malloc/free per call. The default value of
4444 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4445 when configuring.
4446
4447 32. The default maximum size of a compiled pattern is 64K. There have been a
4448 few cases of people hitting this limit. The code now uses macros to handle the
4449 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4450 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4451 configuring. Tests 2 and 5 work only with 2-byte links because they output
4452 debugging information about compiled patterns.
4453
4454 33. Internal code re-arrangements:
4455
4456 (a) Moved the debugging function for printing out a compiled regex into
4457 its own source file (printint.c) and used #include to pull it into
4458 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4459 separate copies.
4460
4461 (b) Defined the list of op-code names for debugging as a macro in
4462 internal.h so that it is next to the definition of the opcodes.
4463
4464 (c) Defined a table of op-code lengths for simpler skipping along compiled
4465 code. This is again a macro in internal.h so that it is next to the
4466 definition of the opcodes.
4467
4468 34. Added support for recursive calls to individual subpatterns, along the
4469 lines of Robin Houston's patch (but implemented somewhat differently).
4470
4471 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4472 allow it to read and process whole directories in Win32. This code was
4473 contributed by Lionel Fourquaux; it has not been tested by me.
4474
4475 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4476 used to name a group. Names consist of alphanumerics and underscores, and must
4477 be unique. Back references use the syntax (?P=name) and recursive calls use
4478 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4479 numbers. The function pcre_fullinfo() can be used after compilation to extract
4480 a name/number map. There are three relevant calls:
4481
4482 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4483 PCRE_INFO_NAMECOUNT yields the number of entries
4484 PCRE_INFO_NAMETABLE yields a pointer to the map.
4485
4486 The map is a vector of fixed-size entries. The size of each entry depends on
4487 the length of the longest name used. The first two bytes of each entry are the
4488 group number, most significant byte first. There follows the corresponding
4489 name, zero terminated. The names are in alphabetical order.
4490
4491 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4492 case instead of 255. Making it the same both with and without UTF-8 support
4493 means that the same test output works with both.
4494
4495 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4496 calling malloc() with a zero argument.
4497
4498 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4499 optimization. I've improved things by keeping a bitmap of backreferences with
4500 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4501 fact referenced, the optimization can be applied. It is unlikely that a
4502 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4503 the match to follow \n) will appear inside brackets with a number greater than
4504 31, but if it does, any back reference > 31 suppresses the optimization.
4505
4506 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4507 of disabling numbered capturing parentheses. Any opening parenthesis that is
4508 not followed by ? behaves as if it were followed by ?: but named parentheses
4509 can still be used for capturing (and they will acquire numbers in the usual
4510 way).
4511
4512 41. Redesigned the return codes from the match() function into yes/no/error so
4513 that errors can be passed back from deep inside the nested calls. A malloc
4514 failure while inside a recursive subpattern call now causes the
4515 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4516
4517 42. It is now possible to set a limit on the number of times the match()
4518 function is called in a call to pcre_exec(). This facility makes it possible to
4519 limit the amount of recursion and backtracking, though not in a directly
4520 obvious way, because the match() function is used in a number of different
4521 circumstances. The count starts from zero for each position in the subject
4522 string (for non-anchored patterns). The default limit is, for compatibility, a
4523 large number, namely 10 000 000. You can change this in two ways:
4524
4525 (a) When configuring PCRE before making, you can use --with-match-limit=n
4526 to set a default value for the compiled library.
4527
4528 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4529 a different value is set. See 45 below.
4530
4531 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4532
4533 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4534 of things that can be changed at compile time. The first argument specifies
4535 what is wanted and the second points to where the information is to be placed.
4536 The current list of available information is:
4537
4538 PCRE_CONFIG_UTF8
4539
4540 The output is an integer that is set to one if UTF-8 support is available;
4541 otherwise it is set to zero.
4542
4543 PCRE_CONFIG_NEWLINE
4544
4545 The output is an integer that it set to the value of the code that is used for
4546 newline. It is either LF (10) or CR (13).
4547
4548 PCRE_CONFIG_LINK_SIZE
4549
4550 The output is an integer that contains the number of bytes used for internal
4551 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4552
4553 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4554
4555 The output is an integer that contains the threshold above which the POSIX
4556 interface uses malloc() for output vectors. See item 31 above.
4557
4558 PCRE_CONFIG_MATCH_LIMIT
4559
4560 The output is an unsigned integer that contains the default limit of the number
4561 of match() calls in a pcre_exec() execution. See 42 above.
4562
4563 44. pcretest has been upgraded by the addition of the -C option. This causes it
4564 to extract all the available output from the new pcre_config() function, and to
4565 output it. The program then exits immediately.
4566
4567 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4568 order to support additional features. One way would have been to define
4569 pcre_exec2() (for example) with extra arguments, but this would not have been
4570 extensible, and would also have required all calls to the original function to
4571 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4572 is used for passing in "extra" data from pcre_study().
4573
4574 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4575 contains the following fields:
4576
4577 flags a bitmap indicating which of the following fields are set
4578 study_data opaque data from pcre_study()
4579 match_limit a way of specifying a limit on match() calls for a specific
4580 call to pcre_exec()
4581 callout_data data for callouts (see 49 below)
4582
4583 The flag bits are also defined in pcre.h, and are
4584
4585 PCRE_EXTRA_STUDY_DATA
4586 PCRE_EXTRA_MATCH_LIMIT
4587 PCRE_EXTRA_CALLOUT_DATA
4588
4589 The pcre_study() function now returns one of these new pcre_extra blocks, with
4590 the actual study data pointed to by the study_data field, and the
4591 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4592 before. That is, this change is entirely upwards-compatible and requires no
4593 change to existing code.
4594
4595 If you want to pass in additional data to pcre_exec(), you can either place it
4596 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4597 block.
4598
4599 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4600 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4601 times with different match limits, until it finds the minimum value needed for
4602 pcre_exec() to complete. The value is then output. This can be instructive; for
4603 most simple matches the number is quite small, but for pathological cases it
4604 gets very large very quickly.
4605
4606 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4607 returns the size of the data block pointed to by the study_data field in a
4608 pcre_extra block, that is, the value that was passed as the argument to
4609 pcre_malloc() when PCRE was getting memory in which to place the information
4610 created by pcre_study(). The fourth argument should point to a size_t variable.
4611 pcretest has been extended so that this information is shown after a successful
4612 pcre_study() call when information about the compiled regex is being displayed.
4613
4614 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4615 because what follows is always an absolute path. (Later: it turns out that this
4616 is more than cosmetic for MinGW, because it doesn't like empty path
4617 components.)
4618
4619 49. Some changes have been made to the callout feature (see 28 above):
4620
4621 (i) A callout function now has three choices for what it returns:
4622
4623 0 => success, carry on matching
4624 > 0 => failure at this point, but backtrack if possible
4625 < 0 => serious error, return this value from pcre_exec()
4626
4627 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4628 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4629 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4630 use by callout functions. It will never be used by PCRE itself.
4631
4632 (ii) The pcre_extra structure (see 45 above) has a void * field called
4633 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4634 pcre_callout_block structure has a field of the same name. The contents of
4635 the field passed in the pcre_extra structure are passed to the callout
4636 function in the corresponding field in the callout block. This makes it
4637 easier to use the same callout-containing regex from multiple threads. For
4638 testing, the pcretest program has a new data escape
4639
4640 \C*n pass the number n (may be negative) as callout_data
4641
4642 If the callout function in pcretest receives a non-zero value as
4643 callout_data, it returns that value.
4644
4645 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4646 there were some redundant $(CFLAGS) in commands that are now specified as
4647 $(LINK), which already includes $(CFLAGS).
4648
4649 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4650 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4651 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4652 one-byte characters throughout. Note that case-insensitive matching applies
4653 only to characters whose values are less than 256. PCRE doesn't support the
4654 notion of cases for higher-valued characters.
4655
4656 (i) A character class whose characters are all within 0-255 is handled as
4657 a bit map, and the map is inverted for negative classes. Previously, a
4658 character > 255 always failed to match such a class; however it should
4659 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4660
4661 (ii) A negated character class with a single character < 255 is coded as
4662 "not this character" (OP_NOT). This wasn't working properly when the test
4663 character was multibyte, either singly or repeated.
4664
4665 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4666 mode, for example: \x{100}{2,3}.
4667
4668 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4669 singly or repeated) now correctly test multibyte characters. However,
4670 PCRE doesn't recognize any characters with values greater than 255 as
4671 digits, spaces, or word characters. Such characters always match \D, \S,
4672 and \W, and never match \d, \s, or \w.
4673
4674 (v) Classes may now contain characters and character ranges with values
4675 greater than 255. For example: [ab\x{100}-\x{400}].
4676
4677 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4678 PCRE in UTF-8 mode.
4679
4680 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4681 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4682 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4683 value.)
4684
4685 53. The single man page has become too large. I have therefore split it up into
4686 a number of separate man pages. These also give rise to individual HTML pages;
4687 these are now put in a separate directory, and there is an index.html page that
4688 lists them all. Some hyperlinking between the pages has been installed.
4689
4690 54. Added convenience functions for handling named capturing parentheses.
4691
4692 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4693 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4694 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4695 are faulted.
4696
4697 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4698 calling configure. These values are used when compiling the dftables.c program
4699 which is run to generate the source of the default character tables. They
4700 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4701 you will need to set these values.
4702
4703 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4704
4705
4706 Version 3.9 02-Jan-02
4707 ---------------------
4708
4709 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4710
4711 2. If --disable-static was given, the building process failed when trying to
4712 build pcretest and pcregrep. (For some reason it was using libtool to compile
4713 them, which is not right, as they aren't part of the library.)
4714
4715
4716 Version 3.8 18-Dec-01
4717 ---------------------
4718
4719 1. The experimental UTF-8 code was completely screwed up. It was packing the
4720 bytes in the wrong order. How dumb can you get?
4721
4722
4723 Version 3.7 29-Oct-01
4724 ---------------------
4725
4726 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4727 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4728 this didn't happen under Solaris 8, where I normally test things.
4729
4730 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4731 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4732 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4733 this fix an hour or so after the initial 3.7 release.)
4734
4735
4736 Version 3.6 23-Oct-01
4737 ---------------------
4738
4739 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4740 offsets passed as NULL with zero offset count.
4741
4742 2. The config.guess and config.sub files had not been updated when I moved to
4743 the latest autoconf.
4744
4745
4746 Version 3.5 15-Aug-01
4747 ---------------------
4748
4749 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4750 had been forgotten.
4751
4752 2. By using declared but undefined structures, we can avoid using "void"
4753 definitions in pcre.h while keeping the internal definitions of the structures
4754 private.
4755
4756 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4757 user point of view, this means that both static and shared libraries are built
4758 by default, but this can be individually controlled. More of the work of
4759 handling this static/shared cases is now inside libtool instead of PCRE's make
4760 file.
4761
4762 4. The pcretest utility is now installed along with pcregrep because it is
4763 useful for users (to test regexs) and by doing this, it automatically gets
4764 relinked by libtool. The documentation has been turned into a man page, so
4765 there are now .1, .txt, and .html versions in /doc.
4766
4767 5. Upgrades to pcregrep:
4768 (i) Added long-form option names like gnu grep.
4769 (ii) Added --help to list all options with an explanatory phrase.
4770 (iii) Added -r, --recursive to recurse into sub-directories.
4771 (iv) Added -f, --file to read patterns from a file.
4772
4773 6. pcre_exec() was referring to its "code" argument before testing that
4774 argument for NULL (and giving an error if it was NULL).
4775
4776 7. Upgraded Makefile.in to allow for compiling in a different directory from
4777 the source directory.
4778
4779 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4780 options bits, the pointer it was passed was to an int instead of to an unsigned
4781 long int. This mattered only on 64-bit systems.
4782
4783 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4784 generated) instead of pcre.in, which it its source. Also made the same change
4785 in several of the .c files.
4786
4787 10. A new release of gcc defines printf() as a macro, which broke pcretest
4788 because it had an ifdef in the middle of a string argument for printf(). Fixed
4789 by using separate calls to printf().
4790
4791 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4792 script, to force use of CR or LF instead of \n in the source. On non-Unix
4793 systems, the value can be set in config.h.
4794
4795 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4796 absolute limit. Changed the text of the error message to make this clear, and
4797 likewise updated the man page.
4798
4799 13. The limit of 99 on the number of capturing subpatterns has been removed.
4800 The new limit is 65535, which I hope will not be a "real" limit.
4801
4802
4803 Version 3.4 22-Aug-00
4804 ---------------------
4805
4806 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4807
4808 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4809
4810
4811 Version 3.3 01-Aug-00
4812 ---------------------
4813
4814 1. If an octal character was given, but the value was greater than \377, it
4815 was not getting masked to the least significant bits, as documented. This could
4816 lead to crashes in some systems.
4817
4818 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4819 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4820
4821 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4822 These just pass their arguments on to (pcre_free)(), but they are provided
4823 because some uses of PCRE bind it to non-C systems that can call its functions,
4824 but cannot call free() or pcre_free() directly.
4825
4826 4. Add "make test" as a synonym for "make check". Corrected some comments in
4827 the Makefile.
4828
4829 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4830 Makefile.
4831
4832 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4833 command called pgrep for grepping around the active processes.
4834
4835 7. Added the beginnings of support for UTF-8 character strings.
4836
4837 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4838 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4839 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4840 out for the ar command.)
4841
4842
4843 Version 3.2 12-May-00
4844 ---------------------
4845
4846 This is purely a bug fixing release.
4847
4848 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4849 of ZA. This was just one example of several cases that could provoke this bug,
4850 which was introduced by change 9 of version 2.00. The code for breaking
4851 infinite loops after an iteration that matches an empty string was't working
4852 correctly.
4853
4854 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4855 when matched against abbab (for example). After matching an empty string, it
4856 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4857 caused it to match further down the string than it should.
4858
4859 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4860 was there because it doesn't seem to be needed, and it causes trouble on some
4861 systems, as it is not a Standard C header. It has been removed.
4862
4863 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4864 were reported on the Macintosh. The changes were from
4865
4866 while ((c = *(++ptr)) != 0 && c != '\n');
4867 to
4868 while ((c = *(++ptr)) != 0 && c != '\n') ;
4869
4870 Totally extraordinary, but if that's what it takes...
4871
4872 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4873 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4874 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4875 assumes the way PCRE uses memmove() (always moving upwards).
4876
4877 6. PCRE is being used in one environment where strchr() is not available. There
4878 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4879 faster code anyway.
4880
4881
4882 Version 3.1 09-Feb-00
4883 ---------------------
4884
4885 The only change in this release is the fixing of some bugs in Makefile.in for
4886 the "install" target:
4887
4888 (1) It was failing to install pcreposix.h.
4889
4890 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4891
4892
4893 Version 3.0 01-Feb-00
4894 ---------------------
4895
4896 1. Add support for the /+ modifier to perltest (to output $` like it does in
4897 pcretest).
4898
4899 2. Add support for the /g modifier to perltest.
4900
4901 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4902 matches null strings.
4903
4904 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4905 pattern. Perl treats empty patterns specially - it reuses the most recent
4906 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4907 effect.
4908
4909 5. The POSIX interface was broken in that it was just handing over the POSIX
4910 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4911 required a bigger vector, with some working space on the end. This means that
4912 the POSIX wrapper now has to get and free some memory, and copy the results.
4913
4914 6. Added some simple autoconf support, placing the test data and the
4915 documentation in separate directories, re-organizing some of the
4916 information files, and making it build pcre-config (a GNU standard). Also added
4917 libtool support for building PCRE as a shared library, which is now the
4918 default.
4919
4920 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4921 09 are not valid octal constants. Single digits will be used for minor values
4922 less than 10.
4923
4924 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4925 existing programs that set these in the POSIX interface can use PCRE without
4926 modification.
4927
4928 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4929 return all that pcre_info() returns, plus additional data. The pcre_info()
4930 function is retained for compatibility, but is considered to be obsolete.
4931
4932 10. Added experimental recursion feature (?R) to handle one common case that
4933 Perl 5.6 will be able to do with (?p{...}).
4934
4935 11. Added support for POSIX character classes like [:alpha:], which Perl is
4936 adopting.
4937
4938
4939 Version 2.08 31-Aug-99
4940 ----------------------
4941
4942 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4943 trying to match at the startoffset position, but instead was moving forward to
4944 the next newline as if a previous match had failed.
4945
4946 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4947 and could get into a loop if a null string was matched other than at the start
4948 of the subject.
4949
4950 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4951 be distinguished at compile time, and for completeness also added PCRE_DATE.
4952
4953 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4954 in GnuWin32 environments.
4955
4956
4957 Version 2.07 29-Jul-99
4958 ----------------------
4959
4960 1. The documentation is now supplied in plain text form and HTML as well as in
4961 the form of man page sources.
4962
4963 2. C++ compilers don't like assigning (void *) values to other pointer types.
4964 In particular this affects malloc(). Although there is no problem in Standard
4965 C, I've put in casts to keep C++ compilers happy.
4966
4967 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4968 should be (const char *).
4969
4970 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4971 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4972 However, I haven't made this a standard facility. The documentation doesn't
4973 mention it, and the Makefile doesn't support it.
4974
4975 5. The Makefile now contains an "install" target, with editable destinations at
4976 the top of the file. The pcretest program is not installed.
4977
4978 6. pgrep -V now gives the PCRE version number and date.
4979
4980 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4981 causing the entire string to be ignored, instead of just the last character.
4982
4983 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4984 non-matching string, it can take a very, very long time, even for strings of
4985 quite modest length, because of the nested recursion. PCRE now does better in
4986 some of these cases. It does this by remembering the last required literal
4987 character in the pattern, and pre-searching the subject to ensure it is present
4988 before running the real match. In other words, it applies a heuristic to detect
4989 some types of certain failure quickly, and in the above example, if presented
4990 with a string that has no trailing " it gives "no match" very quickly.
4991
4992 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4993 other alternatives are tried instead.
4994
4995
4996 Version 2.06 09-Jun-99
4997 ----------------------
4998
4999 1. Change pcretest's output for amount of store used to show just the code
5000 space, because the remainder (the data block) varies in size between 32-bit and
5001 64-bit systems.
5002
5003 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5004 start matching at. This allows lookbehinds to work when searching for multiple
5005 occurrences in a string.
5006
5007 3. Added additional options to pcretest for testing multiple occurrences:
5008
5009 /+ outputs the rest of the string that follows a match
5010 /g loops for multiple occurrences, using the new startoffset argument
5011 /G loops for multiple occurrences by passing an incremented pointer
5012
5013 4. PCRE wasn't doing the "first character" optimization for patterns starting
5014 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5015 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5016 the letter 'x'. On long subject strings, this gives a significant speed-up.
5017
5018
5019 Version 2.05 21-Apr-99
5020 ----------------------
5021
5022 1. Changed the type of magic_number from int to long int so that it works
5023 properly on 16-bit systems.
5024
5025 2. Fixed a bug which caused patterns starting with .* not to work correctly
5026 when the subject string contained newline characters. PCRE was assuming
5027 anchoring for such patterns in all cases, which is not correct because .* will
5028 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5029 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5030 must be retried after every newline in the subject.
5031
5032
5033 Version 2.04 18-Feb-99
5034 ----------------------
5035
5036 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5037 computation of the store needed to hold the pattern was incorrect (too large).
5038 If such patterns were nested a few deep, this could multiply and become a real
5039 problem.
5040
5041 2. Added /M option to pcretest to show the memory requirement of a specific
5042 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5043
5044 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5045 compiled in such a way that the backtracking after subsequent failure was
5046 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5047 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5048
5049
5050 Version 2.03 02-Feb-99
5051 ----------------------
5052
5053 1. Fixed typo and small mistake in man page.
5054
5055 2. Added 4th condition (GPL supersedes if conflict) and created separate
5056 LICENCE file containing the conditions.
5057
5058 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5059 Perl, that is the internal \ allows the delimiter to be included in the
5060 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5061 the final delimiter, add \ to the end of the pattern (to test the error).
5062
5063 4. Added the convenience functions for extracting substrings after a successful
5064 match. Updated pcretest to make it able to test these functions.
5065
5066
5067 Version 2.02 14-Jan-99
5068 ----------------------
5069
5070 1. Initialized the working variables associated with each extraction so that
5071 their saving and restoring doesn't refer to uninitialized store.
5072
5073 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5074 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5075 fix the problem.
5076
5077 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5078 calls, and wasn't printing the correct value for compiling calls. Increased the
5079 default value of LOOPREPEAT, and the number of significant figures in the
5080 times.
5081
5082 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5083
5084 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5085 a building problem on Windows NT with a FAT file system.
5086
5087
5088 Version 2.01 21-Oct-98
5089 ----------------------
5090
5091 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5092 to character tables built by pcre_maketables() in the current locale. If NULL
5093 is passed, the default tables are used.
5094
5095
5096 Version 2.00 24-Sep-98
5097 ----------------------
5098
5099 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5100 it any more.
5101
5102 2. Allow quantification of (?>) groups, and make it work correctly.
5103
5104 3. The first character computation wasn't working for (?>) groups.
5105
5106 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5107 end of the subject) and add 5.005's \z, which really does match only at the
5108 very end of the subject.
5109
5110 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5111
5112 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5113 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5114 localized options. All options to pcre_study() were also removed.
5115
5116 7. Add other new features from 5.005:
5117
5118 $(?<= positive lookbehind
5119 $(?<! negative lookbehind
5120 (?imsx-imsx) added the unsetting capability
5121 such a setting is global if at outer level; local otherwise
5122 (?imsx-imsx:) non-capturing groups with option setting
5123 (?(cond)re|re) conditional pattern matching
5124
5125 A backreference to itself in a repeated group matches the previous
5126 captured string.
5127
5128 8. General tidying up of studying (both automatic and via "study")
5129 consequential on the addition of new assertions.
5130
5131 9. As in 5.005, unlimited repeated groups that could match an empty substring
5132 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5133 runtime if any iteration does actually match an empty substring.
5134
5135 10. Include the RunTest script in the distribution.
5136
5137 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5138 discrepancies, some of which were old and were also with respect to 5.004. They
5139 have now been fixed.
5140
5141
5142 Version 1.09 28-Apr-98
5143 ----------------------
5144
5145 1. A negated single character class followed by a quantifier with a minimum
5146 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5147 program crashes, or just wrong answers. This did not apply to negated classes
5148 containing more than one character, or to minima other than one.
5149
5150
5151 Version 1.08 27-Mar-98
5152 ----------------------
5153
5154 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5155
5156 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5157 latter must appear before anything that relies on it in the pattern.
5158
5159
5160 Version 1.07 16-Feb-98
5161 ----------------------
5162
5163 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5164 repeat of a potentially empty string).
5165
5166
5167 Version 1.06 23-Jan-98
5168 ----------------------
5169
5170 1. Added Markus Oberhumer's little patches for C++.
5171
5172 2. Literal strings longer than 255 characters were broken.
5173
5174
5175 Version 1.05 23-Dec-97
5176 ----------------------
5177
5178 1. Negated character classes containing more than one character were failing if
5179 PCRE_CASELESS was set at run time.
5180
5181
5182 Version 1.04 19-Dec-97
5183 ----------------------
5184
5185 1. Corrected the man page, where some "const" qualifiers had been omitted.
5186
5187 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5188 input syntax.
5189
5190 3. Fixed memory leak which occurred when a regex with back references was
5191 matched with an offsets vector that wasn't big enough. The temporary memory
5192 that is used in this case wasn't being freed if the match failed.
5193
5194 4. Tidied pcretest to ensure it frees memory that it gets.
5195
5196 5. Temporary memory was being obtained in the case where the passed offsets
5197 vector was exactly big enough.
5198
5199 6. Corrected definition of offsetof() from change 5 below.
5200
5201 7. I had screwed up change 6 below and broken the rules for the use of
5202 setjmp(). Now fixed.
5203
5204
5205 Version 1.03 18-Dec-97
5206 ----------------------
5207
5208 1. A erroneous regex with a missing opening parenthesis was correctly
5209 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5210 on some systems.
5211
5212 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5213 it was reported that one broken compiler failed on the former because "code" is
5214 also an independent variable.
5215
5216 3. The erroneous regex a[]b caused an array overrun reference.
5217
5218 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5219 fail on data ending with that character. (It was going on too far, and checking
5220 the next character, typically a binary zero.) This was specific to the
5221 optimized code for single-character negative classes.
5222
5223 5. Added a contributed patch from the TIN world which does the following:
5224
5225 + Add an undef for memmove, in case the the system defines a macro for it.
5226
5227 + Add a definition of offsetof(), in case there isn't one. (I don't know
5228 the reason behind this - offsetof() is part of the ANSI standard - but
5229 it does no harm).
5230
5231 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5232 most of the places where whitespace preceded '#'. I have given up and
5233 allowed the remaining 2 cases to be at the margin.
5234
5235 + Rename some variables in pcre to eliminate shadowing. This seems very
5236 pedantic, but does no harm, of course.
5237
5238 6. Moved the call to setjmp() into its own function, to get rid of warnings
5239 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5240
5241 7. Constructs such as \d{8,} were compiling into the equivalent of
5242 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5243 outcome, but in this particular case used more store than had been allocated,
5244 which caused the bug to be discovered because it threw up an internal error.
5245
5246 8. The debugging code in both pcre and pcretest for outputting the compiled
5247 form of a regex was going wrong in the case of back references followed by
5248 curly-bracketed repeats.
5249
5250
5251 Version 1.02 12-Dec-97
5252 ----------------------
5253
5254 1. Typos in pcre.3 and comments in the source fixed.
5255
5256 2. Applied a contributed patch to get rid of places where it used to remove
5257 'const' from variables, and fixed some signed/unsigned and uninitialized
5258 variable warnings.
5259
5260 3. Added the "runtest" target to Makefile.
5261
5262 4. Set default compiler flag to -O2 rather than just -O.
5263
5264
5265 Version 1.01 19-Nov-97
5266 ----------------------
5267
5268 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5269 like /([ab]*)*/, that is, for classes with more than one character in them.
5270
5271 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5272 as /((?>a*))*/ (a PCRE_EXTRA facility).
5273
5274
5275 Version 1.00 18-Nov-97
5276 ----------------------
5277
5278 1. Added compile-time macros to support systems such as SunOS4 which don't have
5279 memmove() or strerror() but have other things that can be used instead.
5280
5281 2. Arranged that "make clean" removes the executables.
5282
5283
5284 Version 0.99 27-Oct-97
5285 ----------------------
5286
5287 1. Fixed bug in code for optimizing classes with only one character. It was
5288 initializing a 32-byte map regardless, which could cause it to run off the end
5289 of the memory it had got.
5290
5291 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5292
5293
5294 Version 0.98 22-Oct-97
5295 ----------------------
5296
5297 1. Fixed bug in code for handling temporary memory usage when there are more
5298 back references than supplied space in the ovector. This could cause segfaults.
5299
5300
5301 Version 0.97 21-Oct-97
5302 ----------------------
5303
5304 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5305
5306 2. Optimized negated single characters not to use a bit map.
5307
5308 3. Brought error texts together as macro definitions; clarified some of them;
5309 fixed one that was wrong - it said "range out of order" when it meant "invalid
5310 escape sequence".
5311
5312 4. Changed some char * arguments to const char *.
5313
5314 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5315
5316 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5317 pcretest.
5318
5319
5320 Version 0.96 16-Oct-97
5321 ----------------------
5322
5323 1. Added a simple "pgrep" utility to the distribution.
5324
5325 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5326 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5327 where "ddd" means "one or more decimal digits".
5328
5329 3. Fixed serious bug. If a pattern had a back reference, but the call to
5330 pcre_exec() didn't supply a large enough ovector to record the related
5331 identifying subpattern, the match always failed. PCRE now remembers the number
5332 of the largest back reference, and gets some temporary memory in which to save
5333 the offsets during matching if necessary, in order to ensure that
5334 backreferences always work.
5335
5336 4. Increased the compatibility with Perl in a number of ways:
5337
5338 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5339 to request this handling. The option can be set at compile or exec time.
5340
5341 (b) $ matches before a terminating newline by default; an option
5342 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5343 mode). The option can be set at compile or exec time.
5344
5345 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5346 the same as Perl's. If the decimal number it represents is less than 10
5347 or there aren't that many previous left capturing parentheses, an octal
5348 escape is read. Inside a character class, it's always an octal escape,
5349 even if it is a single digit.
5350
5351 (d) An escaped but undefined alphabetic character is taken as a literal,
5352 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5353 escapes.
5354
5355 (e) {0} is now permitted. (The previous item is removed from the compiled
5356 pattern).
5357
5358 5. Changed all the names of code files so that the basic parts are no longer
5359 than 10 characters, and abolished the teeny "globals.c" file.
5360
5361 6. Changed the handling of character classes; they are now done with a 32-byte
5362 bit map always.
5363
5364 7. Added the -d and /D options to pcretest to make it possible to look at the
5365 internals of compilation without having to recompile pcre.
5366
5367
5368 Version 0.95 23-Sep-97
5369 ----------------------
5370
5371 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5372 \x20 at the start of a run of normal characters. These were being treated as
5373 real characters, instead of the source characters being re-checked.
5374
5375
5376 Version 0.94 18-Sep-97
5377 ----------------------
5378
5379 1. The functions are now thread-safe, with the caveat that the global variables
5380 containing pointers to malloc() and free() or alternative functions are the
5381 same for all threads.
5382
5383 2. Get pcre_study() to generate a bitmap of initial characters for non-
5384 anchored patterns when this is possible, and use it if passed to pcre_exec().
5385
5386
5387 Version 0.93 15-Sep-97
5388 ----------------------
5389
5390 1. /(b)|(:+)/ was computing an incorrect first character.
5391
5392 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5393 but not actually doing anything yet.
5394
5395 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5396 as Perl does (e.g. [-az] or [az-]).
5397
5398 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5399 all possible positions.
5400
5401 5. Split up into different modules to avoid including unneeded functions in a
5402 compiled binary. However, compile and exec are still in one module. The "study"
5403 function is split off.
5404
5405 6. The character tables are now in a separate module whose source is generated
5406 by an auxiliary program - but can then be edited by hand if required. There are
5407 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5408 toupper() in the code.
5409
5410 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5411 make them global. Abolish the function for setting them, as the caller can now
5412 set them directly.
5413
5414
5415 Version 0.92 11-Sep-97
5416 ----------------------
5417
5418 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5419 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5420
5421 2. Caseless matching was not working in character classes if the characters in
5422 the pattern were in upper case.
5423
5424 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5425
5426 4. Make PCRE_ANCHORED public and accept as a compile option.
5427
5428 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5429 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5430 pass them.
5431
5432 6. Give an error if bad option bits passed at compile or run time.
5433
5434 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5435 pcretest to cause it to pass that flag.
5436
5437 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5438 options, and the first character, if set.
5439
5440 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5441
5442
5443 Version 0.91 10-Sep-97
5444 ----------------------
5445
5446 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5447 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5448
5449 2. PCRE was looping on encountering an indefinitely repeated back reference to
5450 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5451 Perl does - treats the match as successful.
5452
5453 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5