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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1433 - (show annotations)
Fri Jan 3 15:15:00 2014 UTC (5 years, 9 months ago) by ph10
File size: 260329 byte(s)
Reword pcretest messages and clarify "first char" meaning.
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
51 Version 8.34 15-December-2013
52 -----------------------------
53
54 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
55 executable memory. Patch inspired by Carsten Klein.
56
57 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
58 this macro is never tested and has no effect, because the work to support
59 coverage involves only compiling and linking options and special targets in
60 the Makefile. The comment in config.h implied that defining the macro would
61 enable coverage support, which is totally false. There was also support for
62 setting this macro in the CMake files (my fault, I just copied it from
63 configure). SUPPORT_GCOV has now been removed.
64
65 3. Make a small performance improvement in strlen16() and strlen32() in
66 pcretest.
67
68 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
69 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
70 these cases). There was also one in pcretest.c.
71
72 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
73
74 6. In UTF mode, the code for checking whether a group could match an empty
75 string (which is used for indefinitely repeated groups to allow for
76 breaking an infinite loop) was broken when the group contained a repeated
77 negated single-character class with a character that occupied more than one
78 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
79 UTF-8 mode). The effect was undefined: the group might or might not be
80 deemed as matching an empty string, or the program might have crashed.
81
82 7. The code for checking whether a group could match an empty string was not
83 recognizing that \h, \H, \v, \V, and \R must match a character.
84
85 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
86 an empty string. If it can, pcretest shows this in its information output.
87
88 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
89 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
90 matched by pcre_exec() without using JIT:
91
92 (a) If the rest of the pattern did not match after a maximal run of
93 grapheme clusters, the code for backing up to try with fewer of them
94 did not always back up over a full grapheme when characters that do not
95 have the modifier quality were involved, e.g. Hangul syllables.
96
97 (b) If the match point in a subject started with modifier character, and
98 there was no match, the code could incorrectly back up beyond the match
99 point, and potentially beyond the first character in the subject,
100 leading to a segfault or an incorrect match result.
101
102 10. A conditional group with an assertion condition could lead to PCRE
103 recording an incorrect first data item for a match if no other first data
104 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
105 first data item, and therefore matched "ca" after "c" instead of at the
106 start.
107
108 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
109 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
110
111 12. The source of pcregrep now includes z/OS-specific code so that it can be
112 compiled for z/OS as part of the special z/OS distribution.
113
114 13. Added the -T and -TM options to pcretest.
115
116 14. The code in pcre_compile.c for creating the table of named capturing groups
117 has been refactored. Instead of creating the table dynamically during the
118 actual compiling pass, the information is remembered during the pre-compile
119 pass (on the stack unless there are more than 20 named groups, in which
120 case malloc() is used) and the whole table is created before the actual
121 compile happens. This has simplified the code (it is now nearly 150 lines
122 shorter) and prepared the way for better handling of references to groups
123 with duplicate names.
124
125 15. A back reference to a named subpattern when there is more than one of the
126 same name now checks them in the order in which they appear in the pattern.
127 The first one that is set is used for the reference. Previously only the
128 first one was inspected. This change makes PCRE more compatible with Perl.
129
130 16. Unicode character properties were updated from Unicode 6.3.0.
131
132 17. The compile-time code for auto-possessification has been refactored, based
133 on a patch by Zoltan Herczeg. It now happens after instead of during
134 compilation. The code is cleaner, and more cases are handled. The option
135 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
136 options in pcretest are provided to set it. It can also be set by
137 (*NO_AUTO_POSSESS) at the start of a pattern.
138
139 18. The character VT has been added to the default ("C" locale) set of
140 characters that match \s and are generally treated as white space,
141 following this same change in Perl 5.18. There is now no difference between
142 "Perl space" and "POSIX space". Whether VT is treated as white space in
143 other locales depends on the locale.
144
145 19. The code for checking named groups as conditions, either for being set or
146 for being recursed, has been refactored (this is related to 14 and 15
147 above). Processing unduplicated named groups should now be as fast at
148 numerical groups, and processing duplicated groups should be faster than
149 before.
150
151 20. Two patches to the CMake build system, by Alexander Barkov:
152
153 (1) Replace the "source" command by "." in CMakeLists.txt because
154 "source" is a bash-ism.
155
156 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
157 without these the CMake build does not work on Solaris.
158
159 21. Perl has changed its handling of \8 and \9. If there is no previously
160 encountered capturing group of those numbers, they are treated as the
161 literal characters 8 and 9 instead of a binary zero followed by the
162 literals. PCRE now does the same.
163
164 22. Following Perl, added \o{} to specify codepoints in octal, making it
165 possible to specify values greater than 0777 and also making them
166 unambiguous.
167
168 23. Perl now gives an error for missing closing braces after \x{... instead of
169 treating the string as literal. PCRE now does the same.
170
171 24. RunTest used to grumble if an inappropriate test was selected explicitly,
172 but just skip it when running all tests. This make it awkward to run ranges
173 of tests when one of them was inappropriate. Now it just skips any
174 inappropriate tests, as it always did when running all tests.
175
176 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
177 character types such as \d or \w, too many callouts were inserted, and the
178 data that they returned was rubbish.
179
180 26. In UCP mode, \s was not matching two of the characters that Perl matches,
181 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
182 were matched by \h. The code has now been refactored so that the lists of
183 the horizontal and vertical whitespace characters used for \h and \v (which
184 are defined only in one place) are now also used for \s.
185
186 27. Add JIT support for the 64 bit TileGX architecture.
187 Patch by Jiong Wang (Tilera Corporation).
188
189 28. Possessive quantifiers for classes (both explicit and automatically
190 generated) now use special opcodes instead of wrapping in ONCE brackets.
191
192 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
193 (because it's meaningless), this was not happening when PCRE_CASELESS was
194 set. Not wrong, but inefficient.
195
196 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
197 properties for \w, \d, etc) is present in a test regex. Otherwise if the
198 test contains no characters greater than 255, Perl doesn't realise it
199 should be using Unicode semantics.
200
201 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
202 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
203 does in Unicode mode.
204
205 32. Added the "forbid" facility to pcretest so that putting tests into the
206 wrong test files can sometimes be quickly detected.
207
208 33. There is now a limit (default 250) on the depth of nesting of parentheses.
209 This limit is imposed to control the amount of system stack used at compile
210 time. It can be changed at build time by --with-parens-nest-limit=xxx or
211 the equivalent in CMake.
212
213 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
214 errors. Perl warns for these when in warning mode, but PCRE has no facility
215 for giving warnings.
216
217 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
218 However, this was not working for (?!) because it is optimized to (*FAIL),
219 for which PCRE does not allow quantifiers. The optimization is now disabled
220 when a quantifier follows (?!). I can't see any use for this, but it makes
221 things uniform.
222
223 36. Perl no longer allows group names to start with digits, so I have made this
224 change also in PCRE. It simplifies the code a bit.
225
226 37. In extended mode, Perl ignores spaces before a + that indicates a
227 possessive quantifier. PCRE allowed a space before the quantifier, but not
228 before the possessive +. It now does.
229
230 38. The use of \K (reset reported match start) within a repeated possessive
231 group such as (a\Kb)*+ was not working.
232
233 40. Document that the same character tables must be used at compile time and
234 run time, and that the facility to pass tables to pcre_exec() and
235 pcre_dfa_exec() is for use only with saved/restored patterns.
236
237 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
238 features for Builds with MSVC:
239
240 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
241 builds). The .rc files can be used to set FileDescription and many other
242 attributes.
243
244 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
245 This allows higher-level build scripts which want .pdb files to avoid
246 hard-coding the exact files needed."
247
248 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
249 mean "start of word" and "end of word", respectively, as a transition aid.
250
251 43. A minimizing repeat of a class containing codepoints greater than 255 in
252 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
253 compiled to use the heap for recursion.
254
255 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
256 is configured.
257
258
259 Version 8.33 28-May-2013
260 ------------------------
261
262 1. Added 'U' to some constants that are compared to unsigned integers, to
263 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
264 variables that are added to signed variables, to ensure the result is
265 signed and can be negated.
266
267 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
268 CMake config files.
269
270 3. Revise the creation of config.h.generic so that all boolean macros are
271 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
272 overriding via -D on the command line possible.
273
274 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
275 to unsigned int is reported to make a quite noticeable speed difference in
276 a specific Windows environment. Testing on Linux did also appear to show
277 some benefit (and it is clearly not harmful). Also fixed the definition of
278 Xop which should be unsigned.
279
280 5. Related to (4), changing the definition of the intermediate variable cc
281 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
282 improvements.
283
284 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
285 unnecessary spaces.
286
287 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
288 and later.
289
290 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
291
292 9. Optimizing fast_forward_start_bits in JIT.
293
294 10. Adding support for callouts in JIT, and fixing some issues revealed
295 during this work. Namely:
296
297 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
298
299 (b) Minimum length was not checked before the matching is started.
300
301 11. The value of capture_last that is passed to callouts was incorrect in some
302 cases when there was a capture on one path that was subsequently abandoned
303 after a backtrack. Also, the capture_last value is now reset after a
304 recursion, since all captures are also reset in this case.
305
306 12. The interpreter no longer returns the "too many substrings" error in the
307 case when an overflowing capture is in a branch that is subsequently
308 abandoned after a backtrack.
309
310 13. In the pathological case when an offset vector of size 2 is used, pcretest
311 now prints out the matched string after a yield of 0 or 1.
312
313 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
314 Only supported by the JIT compiler at the moment.
315
316 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
317
318 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
319 offset of the starting point of the matching process, provided the offsets
320 vector is large enough.
321
322 17. The \A escape now records a lookbehind value of 1, though its execution
323 does not actually inspect the previous character. This is to ensure that,
324 in partial multi-segment matching, at least one character from the old
325 segment is retained when a new segment is processed. Otherwise, if there
326 are no lookbehinds in the pattern, \A might match incorrectly at the start
327 of a new segment.
328
329 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
330
331 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
332 gives some modest performance improvement in 8-bit mode.
333
334 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
335 be expressed in certain programming languages using Universal Character
336 Names.
337
338 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
339 which points out that "non characters" are not "characters that may not
340 appear in Unicode strings" but rather "characters that are reserved for
341 internal use and have only local meaning".
342
343 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
344 there was a conditional group that depended on an assertion, if the
345 assertion was false, the callout that immediately followed the alternation
346 in the condition was skipped when pcre_exec() was used for matching.
347
348 23. Allow an explicit callout to be inserted before an assertion that is the
349 condition for a conditional group, for compatibility with automatic
350 callouts, which always insert a callout at this point.
351
352 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
353 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
354
355 25. (*PRUNE) is now supported by the JIT compiler.
356
357 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
358
359 27. Fix the case where there are two or more SKIPs with arguments that may be
360 ignored.
361
362 28. (*SKIP) is now supported by the JIT compiler.
363
364 29. (*THEN) is now supported by the JIT compiler.
365
366 30. Update RunTest with additional test selector options.
367
368 31. The way PCRE handles backtracking verbs has been changed in two ways.
369
370 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
371 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
372 backtracking. In some cases this makes it more Perl-compatible, but Perl's
373 rather obscure rules do not always do the same thing.
374
375 (2) Previously, backtracking verbs were confined within assertions. This is
376 no longer the case for positive assertions, except for (*ACCEPT). Again,
377 this sometimes improves Perl compatibility, and sometimes does not.
378
379 32. A number of tests that were in test 2 because Perl did things differently
380 have been moved to test 1, because either Perl or PCRE has changed, and
381 these tests are now compatible.
382
383 32. Backtracking control verbs are now handled in the same way in JIT and
384 interpreter.
385
386 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
387 contained a forward subroutine reference caused a compile error.
388
389 34. Auto-detect and optimize limited repetitions in JIT.
390
391 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
392 blocking (*UTF) etc.
393
394 36. In the interpreter, maximizing pattern repetitions for characters and
395 character types now use tail recursion, which reduces stack usage.
396
397 37. The value of the max lookbehind was not correctly preserved if a compiled
398 and saved regex was reloaded on a host of different endianness.
399
400 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
401 of the compiled pattern block, expand the flags field from 16 to 32 bits
402 because it was almost full.
403
404 39. Try madvise first before posix_madvise.
405
406 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
407 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
408 original change locked out all matches of empty strings. This has been
409 changed so that one match of an empty string per line is recognized.
410 Subsequent searches on the same line (for colouring or for --only-matching,
411 for example) do not recognize empty strings.
412
413 41. Applied a user patch to fix a number of spelling mistakes in comments.
414
415 42. Data lines longer than 65536 caused pcretest to crash.
416
417 43. Clarified the data type for length and startoffset arguments for pcre_exec
418 and pcre_dfa_exec in the function-specific man pages, where they were
419 explicitly stated to be in bytes, never having been updated. I also added
420 some clarification to the pcreapi man page.
421
422 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
423 a segmentation fault.
424
425
426 Version 8.32 30-November-2012
427 -----------------------------
428
429 1. Improved JIT compiler optimizations for first character search and single
430 character iterators.
431
432 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
433 Patch by Daniel Richard G.
434
435 3. Single character iterator optimizations in the JIT compiler.
436
437 4. Improved JIT compiler optimizations for character ranges.
438
439 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
440 Reported by Giuseppe D'Angelo.
441
442 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
443 of a line, was being set incorrectly in cases where .* appeared inside
444 atomic brackets at the start of a pattern, or where there was a subsequent
445 *PRUNE or *SKIP.
446
447 7. Improved instruction cache flush for POWER/PowerPC.
448 Patch by Daniel Richard G.
449
450 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
451 grep:
452
453 (a) There is now no limit to the number of patterns to be matched.
454
455 (b) An error is given if a pattern is too long.
456
457 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
458 are now supported.
459
460 (d) --exclude-from and --include-from (multiple use) have been added.
461
462 (e) Exclusions and inclusions now apply to all files and directories, not
463 just to those obtained from scanning a directory recursively.
464
465 (f) Multiple uses of -f and --file-list are now supported.
466
467 (g) In a Windows environment, the default for -d has been changed from
468 "read" (the GNU grep default) to "skip", because otherwise the presence
469 of a directory in the file list provokes an error.
470
471 (h) The documentation has been revised and clarified in places.
472
473 9. Improve the matching speed of capturing brackets.
474
475 10. Changed the meaning of \X so that it now matches a Unicode extended
476 grapheme cluster.
477
478 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
479 out POSIX threads when JIT support is configured.
480
481 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
482
483 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
484 structure could go wrong in environments where size_t is not the same size
485 as int.
486
487 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
488 set.
489
490 15. The EBCDIC support had decayed; later updates to the code had included
491 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
492 general tidy up of EBCDIC-related issues, and the documentation was also
493 not quite right. There is now a test that can be run on ASCII systems to
494 check some of the EBCDIC-related things (but is it not a full test).
495
496 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
497 in a small tidy to the code.
498
499 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
500
501 18. If the --only-matching (-o) option in pcregrep is specified multiple
502 times, each one causes appropriate output. For example, -o1 -o2 outputs the
503 substrings matched by the 1st and 2nd capturing parentheses. A separating
504 string can be specified by --om-separator (default empty).
505
506 19. Improving the first n character searches.
507
508 20. Turn case lists for horizontal and vertical white space into macros so that
509 they are defined only once.
510
511 21. This set of changes together give more compatible Unicode case-folding
512 behaviour for characters that have more than one other case when UCP
513 support is available.
514
515 (a) The Unicode property table now has offsets into a new table of sets of
516 three or more characters that are case-equivalent. The MultiStage2.py
517 script that generates these tables (the pcre_ucd.c file) now scans
518 CaseFolding.txt instead of UnicodeData.txt for character case
519 information.
520
521 (b) The code for adding characters or ranges of characters to a character
522 class has been abstracted into a generalized function that also handles
523 case-independence. In UTF-mode with UCP support, this uses the new data
524 to handle characters with more than one other case.
525
526 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
527 whose other case is greater than 256 are now correctly matched
528 caselessly. Previously, the high codepoint matched the low one, but not
529 vice versa.
530
531 (d) The processing of \h, \H, \v, and \ in character classes now makes use
532 of the new class addition function, using character lists defined as
533 macros alongside the case definitions of 20 above.
534
535 (e) Caseless back references now work with characters that have more than
536 one other case.
537
538 (f) General caseless matching of characters with more than one other case
539 is supported.
540
541 22. Unicode character properties were updated from Unicode 6.2.0
542
543 23. Improved CMake support under Windows. Patch by Daniel Richard G.
544
545 24. Add support for 32-bit character strings, and UTF-32
546
547 25. Major JIT compiler update (code refactoring and bugfixing).
548 Experimental Sparc 32 support is added.
549
550 26. Applied a modified version of Daniel Richard G's patch to create
551 pcre.h.generic and config.h.generic by "make" instead of in the
552 PrepareRelease script.
553
554 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
555 pcre_compile.c when checking for a zero character.
556
557 28. Introducing a native interface for JIT. Through this interface, the compiled
558 machine code can be directly executed. The purpose of this interface is to
559 provide fast pattern matching, so several sanity checks are not performed.
560 However, feature tests are still performed. The new interface provides
561 1.4x speedup compared to the old one.
562
563 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
564 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
565 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
566
567 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
568 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
569 the "old" RFC 2279). Instead, it ended up passing a negative length to
570 pcre_exec().
571
572 31. Add support for GCC's visibility feature to hide internal functions.
573
574 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
575 "unknown -C option" after outputting 0 or 1.
576
577 33. There is now support for generating a code coverage report for the test
578 suite in environments where gcc is the compiler and lcov is installed. This
579 is mainly for the benefit of the developers.
580
581 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
582 unaddressable using valgrind annotations, allowing valgrind to detect
583 invalid memory accesses. This is mainly for the benefit of the developers.
584
585 25. (*UTF) can now be used to start a pattern in any of the three libraries.
586
587 26. Give configure error if --enable-cpp but no C++ compiler found.
588
589
590 Version 8.31 06-July-2012
591 -------------------------
592
593 1. Fixing a wrong JIT test case and some compiler warnings.
594
595 2. Removed a bashism from the RunTest script.
596
597 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
598 to unsigned type, result still unsigned" that was given by an MS compiler
599 on encountering the code "-sizeof(xxx)".
600
601 4. Partial matching support is added to the JIT compiler.
602
603 5. Fixed several bugs concerned with partial matching of items that consist
604 of more than one character:
605
606 (a) /^(..)\1/ did not partially match "aba" because checking references was
607 done on an "all or nothing" basis. This also applied to repeated
608 references.
609
610 (b) \R did not give a hard partial match if \r was found at the end of the
611 subject.
612
613 (c) \X did not give a hard partial match after matching one or more
614 characters at the end of the subject.
615
616 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
617 a partial match for the string "\r".
618
619 (e) When newline was set to CRLF, the metacharacter "." did not recognize
620 a partial match for a CR character at the end of the subject string.
621
622 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
623 running pcretest, the text "(JIT)" added to the output whenever JIT is
624 actually used to run the match.
625
626 7. Individual JIT compile options can be set in pcretest by following -s+[+]
627 or /S+[+] with a digit between 1 and 7.
628
629 8. OP_NOT now supports any UTF character not just single-byte ones.
630
631 9. (*MARK) control verb is now supported by the JIT compiler.
632
633 10. The command "./RunTest list" lists the available tests without actually
634 running any of them. (Because I keep forgetting what they all are.)
635
636 11. Add PCRE_INFO_MAXLOOKBEHIND.
637
638 12. Applied a (slightly modified) user-supplied patch that improves performance
639 when the heap is used for recursion (compiled with --disable-stack-for-
640 recursion). Instead of malloc and free for each heap frame each time a
641 logical recursion happens, frames are retained on a chain and re-used where
642 possible. This sometimes gives as much as 30% improvement.
643
644 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
645 call.
646
647 14. As documented, (*COMMIT) is now confined to within a positive assertion.
648
649 15. It is now possible to link pcretest with libedit as an alternative to
650 libreadline.
651
652 16. (*COMMIT) control verb is now supported by the JIT compiler.
653
654 17. The Unicode data tables have been updated to Unicode 6.1.0.
655
656 18. Added --file-list option to pcregrep.
657
658 19. Added binary file support to pcregrep, including the -a, --binary-files,
659 -I, and --text options.
660
661 20. The madvise function is renamed for posix_madvise for QNX compatibility
662 reasons. Fixed by Giuseppe D'Angelo.
663
664 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
665
666 22. Changed the option for creating symbolic links for 16-bit man pages from
667 -s to -sf so that re-installing does not cause issues.
668
669 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
670
671 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
672 matches in certain environments (the workspace was not being correctly
673 retained). Also added to pcre_dfa_exec() a simple plausibility check on
674 some of the workspace data at the beginning of a restart.
675
676 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
677 was not doing so when it should - probably a typo introduced by SVN 528
678 (change 8.10/14).
679
680 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
681 \w+ when the character tables indicated that \x{c4} was a word character.
682 There were several related cases, all because the tests for doing a table
683 lookup were testing for characters less than 127 instead of 255.
684
685 27. If a pattern contains capturing parentheses that are not used in a match,
686 their slots in the ovector are set to -1. For those that are higher than
687 any matched groups, this happens at the end of processing. In the case when
688 there were back references that the ovector was too small to contain
689 (causing temporary malloc'd memory to be used during matching), and the
690 highest capturing number was not used, memory off the end of the ovector
691 was incorrectly being set to -1. (It was using the size of the temporary
692 memory instead of the true size.)
693
694 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
695 ovector size, it uses memory at the end of the block that it has got.
696
697 29. Check for an overlong MARK name and give an error at compile time. The
698 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
699
700 30. JIT compiler update.
701
702 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
703 Rill for the patch.
704
705 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
706
707 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
708
709 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
710 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
711 was enabled.
712
713 35. Improve JIT code generation for greedy plus quantifier.
714
715 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
716 1 to "aa" instead of to an empty string. The bug affected repeated groups
717 that could potentially match an empty string.
718
719 37. Optimizing single character iterators in JIT.
720
721 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
722 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
723 codepoints that are too big for the mode are faulted, and in a UTF mode,
724 disallowed codepoints are also faulted.
725
726 39. If PCRE was compiled with UTF support, in three places in the DFA
727 matcher there was code that should only have been obeyed in UTF mode, but
728 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
729 processing when bytes with values greater than 127 were present. In 16-bit
730 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
731 both cases the values are those that cannot be the first data item in a UTF
732 character. The three items that might have provoked this were recursions,
733 possessively repeated groups, and atomic groups.
734
735 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
736 and pcregrep, because some OS require shared objects to be explicitly
737 passed to ld, causing the link step to fail if they are not.
738
739 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
740 mode, patterns that started with \h* or \R* might be incorrectly matched.
741
742
743 Version 8.30 04-February-2012
744 -----------------------------
745
746 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
747 name is defined in ctype.h.
748
749 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
750 only in quite long subpatterns.
751
752 3. Removed the function pcre_info(), which has been obsolete and deprecated
753 since it was replaced by pcre_fullinfo() in February 2000.
754
755 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
756 match a (*MARK), and the match failed at the start of the subject, a
757 reference to memory before the start of the subject could occur. This bug
758 was introduced by fix 17 of release 8.21.
759
760 5. A reference to an unset group with zero minimum repetition was giving
761 totally wrong answers (in non-JavaScript-compatibility mode). For example,
762 /(another)?(\1?)test/ matched against "hello world test". This bug was
763 introduced in release 8.13.
764
765 6. Add support for 16-bit character strings (a large amount of work involving
766 many changes and refactorings).
767
768 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
769 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
770 from a file.
771
772 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
773 rounding is not applied in this particular case).
774
775 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
776 if they appear, or are escaped, in patterns.
777
778 10. Get rid of a number of -Wunused-but-set-variable warnings.
779
780 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
781 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
782 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
783 also returns the mark "x". This bug applied to capturing parentheses,
784 non-capturing parentheses, and atomic parentheses. It also applied to some
785 assertions.
786
787 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
788 information out of configure.ac instead of relying on pcre.h.generic, which
789 is not stored in the repository.
790
791 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
792 -lreadline.
793
794 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
795
796 15. Applied Graycode's patch to put the top-level frame on the stack rather
797 than the heap when not using the stack for recursion. This gives a
798 performance improvement in many cases when recursion is not deep.
799
800 16. Experimental code added to "pcretest -C" to output the stack frame size.
801
802
803 Version 8.21 12-Dec-2011
804 ------------------------
805
806 1. Updating the JIT compiler.
807
808 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
809 are added as well.
810
811 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
812 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
813 calling _pcre_jit_exec. Some extra comments are added.
814
815 4. (*MARK) settings inside atomic groups that do not contain any capturing
816 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
817 was introduced by change 18 for 8.20.
818
819 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
820 ECMA-262 standard.
821
822 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
823 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
824 This bug was probably introduced by change 9 of 8.13.
825
826 7. While fixing 6 above, I noticed that a number of other items were being
827 incorrectly rejected as "not fixed length". This arose partly because newer
828 opcodes had not been added to the fixed-length checking code. I have (a)
829 corrected the bug and added tests for these items, and (b) arranged for an
830 error to occur if an unknown opcode is encountered while checking for fixed
831 length instead of just assuming "not fixed length". The items that were
832 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
833 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
834 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
835
836 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
837 being incorrectly compiled and would have given unpredicatble results.
838
839 9. A possessively repeated subpattern with minimum repeat count greater than
840 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
841 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
842 the first (A) could occur when it should not.
843
844 10. Add a cast and remove a redundant test from the code.
845
846 11. JIT should use pcre_malloc/pcre_free for allocation.
847
848 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
849 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
850 is anything other than /usr, -L is still shown).
851
852 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
853
854 14. Perl does not support \N without a following name in a [] class; PCRE now
855 also gives an error.
856
857 15. If a forward reference was repeated with an upper limit of around 2000,
858 it caused the error "internal error: overran compiling workspace". The
859 maximum number of forward references (including repeats) was limited by the
860 internal workspace, and dependent on the LINK_SIZE. The code has been
861 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
862 the default depends on LINK_SIZE. There is a new upper limit (for safety)
863 of around 200,000 forward references. While doing this, I also speeded up
864 the filling in of repeated forward references.
865
866 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
867 incorrectly expecting the subject to contain another "a" after the start.
868
869 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
870 in the match, the SKIP should be ignored. This was not happening; instead
871 the SKIP was being treated as NOMATCH. For patterns such as
872 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
873 tested.
874
875 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
876 now much more compatible with Perl, in particular in cases where the result
877 is a non-match for a non-anchored pattern. For example, if
878 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
879 "m", where previously it did not return a name. A side effect of this
880 change is that for partial matches, the last encountered mark name is
881 returned, as for non matches. A number of tests that were previously not
882 Perl-compatible have been moved into the Perl-compatible test files. The
883 refactoring has had the pleasing side effect of removing one argument from
884 the match() function, thus reducing its stack requirements.
885
886 19. If the /S+ option was used in pcretest to study a pattern using JIT,
887 subsequent uses of /S (without +) incorrectly behaved like /S+.
888
889 21. Retrieve executable code size support for the JIT compiler and fixing
890 some warnings.
891
892 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
893 not work when the shorter character appeared right at the end of the
894 subject string.
895
896 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
897 systems.
898
899 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
900 output it when the /M option is used in pcretest.
901
902 25. The CheckMan script was not being included in the distribution. Also, added
903 an explicit "perl" to run Perl scripts from the PrepareRelease script
904 because this is reportedly needed in Windows.
905
906 26. If study data was being save in a file and studying had not found a set of
907 "starts with" bytes for the pattern, the data written to the file (though
908 never used) was taken from uninitialized memory and so caused valgrind to
909 complain.
910
911 27. Updated RunTest.bat as provided by Sheri Pierce.
912
913 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
914
915 29. Computation of memory usage for the table of capturing group names was
916 giving an unnecessarily large value.
917
918
919 Version 8.20 21-Oct-2011
920 ------------------------
921
922 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
923 a POSIX class. After further experiments with Perl, which convinced me that
924 Perl has bugs and confusions, a closing square bracket is no longer allowed
925 in a POSIX name. This bug also affected patterns with classes that started
926 with full stops.
927
928 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
929 captured substring, but while checking the failing first alternative,
930 substring 1 is temporarily captured. If the output vector supplied to
931 pcre_exec() was not big enough for this capture, the yield of the function
932 was still zero ("insufficient space for captured substrings"). This cannot
933 be totally fixed without adding another stack variable, which seems a lot
934 of expense for a edge case. However, I have improved the situation in cases
935 such as /(a)(b)x|abc/ matched against "abc", where the return code
936 indicates that fewer than the maximum number of slots in the ovector have
937 been set.
938
939 3. Related to (2) above: when there are more back references in a pattern than
940 slots in the output vector, pcre_exec() uses temporary memory during
941 matching, and copies in the captures as far as possible afterwards. It was
942 using the entire output vector, but this conflicts with the specification
943 that only 2/3 is used for passing back captured substrings. Now it uses
944 only the first 2/3, for compatibility. This is, of course, another edge
945 case.
946
947 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
948 main code base, and can be used by building with --enable-jit. When this is
949 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
950 runtime --no-jit option is given.
951
952 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
953 ovector, the return from the function was zero, implying that there were
954 other matches that did not fit. The correct "exactly full" value is now
955 returned.
956
957 6. If a subpattern that was called recursively or as a subroutine contained
958 (*PRUNE) or any other control that caused it to give a non-standard return,
959 invalid errors such as "Error -26 (nested recursion at the same subject
960 position)" or even infinite loops could occur.
961
962 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
963 computing the minimum length on reaching *ACCEPT, and so ended up with the
964 wrong value of 1 rather than 0. Further investigation indicates that
965 computing a minimum subject length in the presence of *ACCEPT is difficult
966 (think back references, subroutine calls), and so I have changed the code
967 so that no minimum is registered for a pattern that contains *ACCEPT.
968
969 8. If (*THEN) was present in the first (true) branch of a conditional group,
970 it was not handled as intended. [But see 16 below.]
971
972 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
973 Sheri Pierce.
974
975 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
976 the first byte in a match must be "a".
977
978 11. Change 17 for 8.13 increased the recursion depth for patterns like
979 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
980 pattern contains any instances of (*THEN). If it does not, the old
981 optimizations are restored. It would be nice to do this on a per-group
982 basis, but at the moment that is not feasible.
983
984 12. In some environments, the output of pcretest -C is CRLF terminated. This
985 broke RunTest's code that checks for the link size. A single white space
986 character after the value is now allowed for.
987
988 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
989 For "fr", it uses the Windows-specific input and output files.
990
991 14. If (*THEN) appeared in a group that was called recursively or as a
992 subroutine, it did not work as intended. [But see next item.]
993
994 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
995 pattern fragments (but not containing any | characters). If A and B are
996 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
997 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
998 D. In other words, Perl considers parentheses that do not contain any |
999 characters to be part of a surrounding alternative, whereas PCRE was
1000 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1001 differently. PCRE now behaves in the same way as Perl, except in the case
1002 of subroutine/recursion calls such as (?1) which have in any case always
1003 been different (but PCRE had them first :-).
1004
1005 16. Related to 15 above: Perl does not treat the | in a conditional group as
1006 creating alternatives. Such a group is treated in the same way as an
1007 ordinary group without any | characters when processing (*THEN). PCRE has
1008 been changed to match Perl's behaviour.
1009
1010 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1011 RunGrepTest script failed.
1012
1013 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1014 inevitable for groups that contain captures, but it can lead to a lot of
1015 stack use in large patterns. The old behaviour has been restored for atomic
1016 groups that do not contain any capturing parentheses.
1017
1018 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1019 suppress the check for a minimum subject length at run time. (If it was
1020 given to pcre_exec() or pcre_dfa_exec() it did work.)
1021
1022 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1023 fail to work when decoding hex characters in data strings in EBCDIC
1024 environments.
1025
1026 21. It appears that in at least one Mac OS environment, the isxdigit() function
1027 is implemented as a macro that evaluates to its argument more than once,
1028 contravening the C 90 Standard (I haven't checked a later standard). There
1029 was an instance in pcretest which caused it to go wrong when processing
1030 \x{...} escapes in subject strings. The has been rewritten to avoid using
1031 things like p++ in the argument of isxdigit().
1032
1033
1034 Version 8.13 16-Aug-2011
1035 ------------------------
1036
1037 1. The Unicode data tables have been updated to Unicode 6.0.0.
1038
1039 2. Two minor typos in pcre_internal.h have been fixed.
1040
1041 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1042 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1043 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1044
1045 4. There were a number of related bugs in the code for matching backrefences
1046 caselessly in UTF-8 mode when codes for the characters concerned were
1047 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1048 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1049 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1050 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1051 2-byte code at the end of the subject (it thought there wasn't enough data
1052 left).
1053
1054 5. Comprehensive information about what went wrong is now returned by
1055 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1056 as the output vector has at least 2 elements. The offset of the start of
1057 the failing character and a reason code are placed in the vector.
1058
1059 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1060 now returned is for the first byte of the failing character, instead of the
1061 last byte inspected. This is an incompatible change, but I hope it is small
1062 enough not to be a problem. It makes the returned offset consistent with
1063 pcre_exec() and pcre_dfa_exec().
1064
1065 7. pcretest now gives a text phrase as well as the error number when
1066 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1067 failure, the offset and reason code are output.
1068
1069 8. When \R was used with a maximizing quantifier it failed to skip backwards
1070 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1071 back over a single character (\n). This seems wrong (because it treated the
1072 two characters as a single entity when going forwards), conflicts with the
1073 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1074 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1075 has been changed.
1076
1077 9. Some internal refactoring has changed the processing so that the handling
1078 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1079 time (the PCRE_DOTALL option was changed this way some time ago: version
1080 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1081 which was always a bit of a fudge. It also means that there is one less
1082 argument for the match() function, which reduces its stack requirements
1083 slightly. This change also fixes an incompatibility with Perl: the pattern
1084 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1085
1086 10. More internal refactoring has drastically reduced the number of recursive
1087 calls to match() for possessively repeated groups such as (abc)++ when
1088 using pcre_exec().
1089
1090 11. While implementing 10, a number of bugs in the handling of groups were
1091 discovered and fixed:
1092
1093 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1094 (a|)*(?1) gave a compile-time internal error.
1095 ((a|)+)+ did not notice that the outer group could match an empty string.
1096 (^a|^)+ was not marked as anchored.
1097 (.*a|.*)+ was not marked as matching at start or after a newline.
1098
1099 12. Yet more internal refactoring has removed another argument from the match()
1100 function. Special calls to this function are now indicated by setting a
1101 value in a variable in the "match data" data block.
1102
1103 13. Be more explicit in pcre_study() instead of relying on "default" for
1104 opcodes that mean there is no starting character; this means that when new
1105 ones are added and accidentally left out of pcre_study(), testing should
1106 pick them up.
1107
1108 14. The -s option of pcretest has been documented for ages as being an old
1109 synonym of -m (show memory usage). I have changed it to mean "force study
1110 for every regex", that is, assume /S for every regex. This is similar to -i
1111 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1112 using it. It makes it easier to run collections of tests with and without
1113 study enabled, and thereby test pcre_study() more easily. All the standard
1114 tests are now run with and without -s (but some patterns can be marked as
1115 "never study" - see 20 below).
1116
1117 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1118 restoration of the capturing data to the outer values was not happening
1119 correctly.
1120
1121 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1122 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1123 pattern had matched an empty string, and so incorrectly returned a no
1124 match.
1125
1126 17. There was optimizing code for the last branch of non-capturing parentheses,
1127 and also for the obeyed branch of a conditional subexpression, which used
1128 tail recursion to cut down on stack usage. Unfortunately, now that there is
1129 the possibility of (*THEN) occurring in these branches, tail recursion is
1130 no longer possible because the return has to be checked for (*THEN). These
1131 two optimizations have therefore been removed. [But see 8.20/11 above.]
1132
1133 18. If a pattern containing \R was studied, it was assumed that \R always
1134 matched two bytes, thus causing the minimum subject length to be
1135 incorrectly computed because \R can also match just one byte.
1136
1137 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1138 was incorrectly computed.
1139
1140 20. If /S is present twice on a test pattern in pcretest input, it now
1141 *disables* studying, thereby overriding the use of -s on the command line
1142 (see 14 above). This is necessary for one or two tests to keep the output
1143 identical in both cases.
1144
1145 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1146 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1147
1148 22. When an atomic group that contained a capturing parenthesis was
1149 successfully matched, but the branch in which it appeared failed, the
1150 capturing was not being forgotten if a higher numbered group was later
1151 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1152 group 1 to "a", when in fact it should be unset. This applied to multi-
1153 branched capturing and non-capturing groups, repeated or not, and also to
1154 positive assertions (capturing in negative assertions does not happen
1155 in PCRE) and also to nested atomic groups.
1156
1157 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1158 subject after a captured substring, to make it easier to tell which of a
1159 number of identical substrings has been captured.
1160
1161 24. The way atomic groups are processed by pcre_exec() has been changed so that
1162 if they are repeated, backtracking one repetition now resets captured
1163 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1164 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1165 "aaa". Previously, it would have been "a". As part of this code
1166 refactoring, the way recursive calls are handled has also been changed.
1167
1168 25. If an assertion condition captured any substrings, they were not passed
1169 back unless some other capturing happened later. For example, if
1170 (?(?=(a))a) was matched against "a", no capturing was returned.
1171
1172 26. When studying a pattern that contained subroutine calls or assertions,
1173 the code for finding the minimum length of a possible match was handling
1174 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1175 group 1 called group 2 while simultaneously a separate group 2 called group
1176 1). A stack overflow occurred in this case. I have fixed this by limiting
1177 the recursion depth to 10.
1178
1179 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1180 Fortmann. This supports explicit test numbers on the command line, and has
1181 argument validation and error reporting.
1182
1183 28. An instance of \X with an unlimited repeat could fail if at any point the
1184 first character it looked at was a mark character.
1185
1186 29. Some minor code refactoring concerning Unicode properties and scripts
1187 should reduce the stack requirement of match() slightly.
1188
1189 30. Added the '=' option to pcretest to check the setting of unused capturing
1190 slots at the end of the pattern, which are documented as being -1, but are
1191 not included in the return count.
1192
1193 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1194 compiled something random. Now it gives a compile-time error (as does
1195 Perl).
1196
1197 32. A *MARK encountered during the processing of a positive assertion is now
1198 recorded and passed back (compatible with Perl).
1199
1200 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1201 had alternative anchored branches, the search for a second match in a line
1202 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1203 matched the line "0102" twice. The same bug affected patterns that started
1204 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1205 twice.
1206
1207 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1208 does, and because of capturing effects, quantifying parenthesized
1209 assertions may at times be useful. Quantifiers are now allowed for
1210 parenthesized assertions.
1211
1212 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1213
1214 36. \g was being checked for fancy things in a character class, when it should
1215 just be a literal "g".
1216
1217 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1218 appearance of a nested POSIX class supersedes an apparent external class.
1219 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1220 unescaped square brackets may also appear as part of class names. For
1221 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1222 more like Perl. (But see 8.20/1 above.)
1223
1224 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1225 was because it thought it was \N{name}, which is not supported).
1226
1227 39. Add minix to OS list not supporting the -S option in pcretest.
1228
1229 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1230 cannot analyze patterns in sufficient detail to catch mutual recursions
1231 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1232 subgroup is called recursively as a subpattern for a second time at the
1233 same position in the subject string. In previous releases this might have
1234 been caught by the recursion limit, or it might have run out of stack.
1235
1236 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1237 happen only once. PCRE was, however incorrectly giving a compile time error
1238 "recursive call could loop indefinitely" because it cannot analyze the
1239 pattern in sufficient detail. The compile time test no longer happens when
1240 PCRE is compiling a conditional subpattern, but actual runaway loops are
1241 now caught at runtime (see 40 above).
1242
1243 42. It seems that Perl allows any characters other than a closing parenthesis
1244 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1245 has been changed to be the same.
1246
1247 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1248 as not to get warnings when autogen.sh is called. Also changed
1249 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1250
1251 44. To help people who use pcregrep to scan files containing exceedingly long
1252 lines, the following changes have been made:
1253
1254 (a) The default value of the buffer size parameter has been increased from
1255 8K to 20K. (The actual buffer used is three times this size.)
1256
1257 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1258 PCRE is built.
1259
1260 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1261 to be set at run time.
1262
1263 (d) Numerical values in pcregrep options can be followed by K or M, for
1264 example --buffer-size=50K.
1265
1266 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1267 given and the return code is set to 2.
1268
1269 45. Add a pointer to the latest mark to the callout data block.
1270
1271 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1272 partial match of an empty string instead of no match. This was specific to
1273 the use of ".".
1274
1275 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1276 complete match instead of a partial match. This bug was dependent on both
1277 the PCRE_UTF8 and PCRE_DOTALL options being set.
1278
1279 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1280 starting byte set, because \b was not being ignored.
1281
1282
1283 Version 8.12 15-Jan-2011
1284 ------------------------
1285
1286 1. Fixed some typos in the markup of the man pages, and wrote a script that
1287 checks for such things as part of the documentation building process.
1288
1289 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1290 --match-limit and --recursion-limit options (added for 8.11). In
1291 particular, this made one of the standard tests fail. (The integer value
1292 went into the wrong half of a long int.)
1293
1294 3. If the --colour option was given to pcregrep with -v (invert match), it
1295 did strange things, either producing crazy output, or crashing. It should,
1296 of course, ignore a request for colour when reporting lines that do not
1297 match.
1298
1299 4. Another pcregrep bug caused similar problems if --colour was specified with
1300 -M (multiline) and the pattern match finished with a line ending.
1301
1302 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1303 matched in multiline mode, the following line was shown as part of the
1304 match. This seems wrong, so I have changed it.
1305
1306 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1307 the check for further matches in the same line (so they could be coloured)
1308 to overrun the end of the current line. If another match was found, it was
1309 incorrectly shown (and then shown again when found in the next line).
1310
1311 7. If pcregrep was compiled under Windows, there was a reference to the
1312 function pcregrep_exit() before it was defined. I am assuming this was
1313 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1314 reported by a user. I've moved the definition above the reference.
1315
1316
1317 Version 8.11 10-Dec-2010
1318 ------------------------
1319
1320 1. (*THEN) was not working properly if there were untried alternatives prior
1321 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1322 backtracked to try for "b" instead of moving to the next alternative branch
1323 at the same level (in this case, to look for "c"). The Perl documentation
1324 is clear that when (*THEN) is backtracked onto, it goes to the "next
1325 alternative in the innermost enclosing group".
1326
1327 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1328 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1329 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1330 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1331 (*THEN).
1332
1333 3. If \s appeared in a character class, it removed the VT character from
1334 the class, even if it had been included by some previous item, for example
1335 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1336 of \s, but is part of the POSIX "space" class.)
1337
1338 4. A partial match never returns an empty string (because you can always
1339 match an empty string at the end of the subject); however the checking for
1340 an empty string was starting at the "start of match" point. This has been
1341 changed to the "earliest inspected character" point, because the returned
1342 data for a partial match starts at this character. This means that, for
1343 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1344 (previously it gave "no match").
1345
1346 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1347 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1348 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1349 has an implication that the given string is incomplete (because a partial
1350 match is preferred over a full match). For this reason, these items now
1351 give a partial match in this situation. [Aside: previously, the one case
1352 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1353 match rather than a full match, which was wrong by the old rules, but is
1354 now correct.]
1355
1356 6. There was a bug in the handling of #-introduced comments, recognized when
1357 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1358 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1359 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1360 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1361 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1362 places in pcre_compile().
1363
1364 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1365 comments when looking ahead for named forward references to subpatterns,
1366 the only newline sequence it recognized was NL. It now handles newlines
1367 according to the set newline convention.
1368
1369 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1370 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1371 cater for a lack of strerror(). These oversights have been fixed.
1372
1373 9. Added --match-limit and --recursion-limit to pcregrep.
1374
1375 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1376
1377 11. When the -o option was used, pcregrep was setting a return code of 1, even
1378 when matches were found, and --line-buffered was not being honoured.
1379
1380 12. Added an optional parentheses number to the -o and --only-matching options
1381 of pcregrep.
1382
1383 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1384 can match an empty string. The code to do it in pcretest and pcredemo
1385 needed fixing:
1386
1387 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1388 only one byte after an empty string match just before CRLF (this case
1389 just got forgotten; "any" and "anycrlf" were OK).
1390
1391 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1392 mode when an empty string match preceded an ASCII character followed by
1393 a non-ASCII character. (The code for advancing by one character rather
1394 than one byte was nonsense.)
1395
1396 (c) The pcredemo.c sample program did not have any code at all to handle
1397 the cases when CRLF is a valid newline sequence.
1398
1399 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1400 as a starting offset was within the subject string. There is now a new
1401 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1402 negative or greater than the length of the string. In order to test this,
1403 pcretest is extended to allow the setting of negative starting offsets.
1404
1405 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1406 starting offset points to the beginning of a UTF-8 character was
1407 unnecessarily clumsy. I tidied it up.
1408
1409 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1410 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1411
1412 17. Nobody had reported that the --include_dir option, which was added in
1413 release 7.7 should have been called --include-dir (hyphen, not underscore)
1414 for compatibility with GNU grep. I have changed it to --include-dir, but
1415 left --include_dir as an undocumented synonym, and the same for
1416 --exclude-dir, though that is not available in GNU grep, at least as of
1417 release 2.5.4.
1418
1419 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1420 characters from a string of bytes) have been redefined so as not to use
1421 loops, in order to improve performance in some environments. At the same
1422 time, I abstracted some of the common code into auxiliary macros to save
1423 repetition (this should not affect the compiled code).
1424
1425 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1426 compile-time error is now given if \c is not followed by an ASCII
1427 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1428 different, and any byte value is allowed.)
1429
1430 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1431 START_OPTIMIZE option, which is now allowed at compile time - but just
1432 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1433 to pcregrep and other applications that have no direct access to PCRE
1434 options. The new /Y option in pcretest sets this option when calling
1435 pcre_compile().
1436
1437 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1438 back references. Groups containing recursive back references were forced to
1439 be atomic by that change, but in the case of named groups, the amount of
1440 memory required was incorrectly computed, leading to "Failed: internal
1441 error: code overflow". This has been fixed.
1442
1443 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1444 pcretest.c, to avoid build problems in some Borland environments.
1445
1446
1447 Version 8.10 25-Jun-2010
1448 ------------------------
1449
1450 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1451 THEN.
1452
1453 2. (*ACCEPT) was not working when inside an atomic group.
1454
1455 3. Inside a character class, \B is treated as a literal by default, but
1456 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1457 causes the error). The code is unchanged, but I tidied the documentation.
1458
1459 4. Inside a character class, PCRE always treated \R and \X as literals,
1460 whereas Perl faults them if its -w option is set. I have changed PCRE so
1461 that it faults them when PCRE_EXTRA is set.
1462
1463 5. Added support for \N, which always matches any character other than
1464 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1465
1466 6. When compiling pcregrep with newer versions of gcc which may have
1467 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1468 declared with attribute warn_unused_result" were given. Just casting the
1469 result to (void) does not stop the warnings; a more elaborate fudge is
1470 needed. I've used a macro to implement this.
1471
1472 7. Minor change to pcretest.c to avoid a compiler warning.
1473
1474 8. Added four artifical Unicode properties to help with an option to make
1475 \s etc use properties (see next item). The new properties are: Xan
1476 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1477
1478 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1479 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1480 this option. Modified pcretest to add /W to test this facility. Added
1481 REG_UCP to make it available via the POSIX interface.
1482
1483 10. Added --line-buffered to pcregrep.
1484
1485 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1486 studied, and the match started with a letter with a code point greater than
1487 127 whose first byte was different to the first byte of the other case of
1488 the letter, the other case of this starting letter was not recognized
1489 (#976).
1490
1491 12. If a pattern that was studied started with a repeated Unicode property
1492 test, for example, \p{Nd}+, there was the theoretical possibility of
1493 setting up an incorrect bitmap of starting bytes, but fortunately it could
1494 not have actually happened in practice until change 8 above was made (it
1495 added property types that matched character-matching opcodes).
1496
1497 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1498 possible starting bytes for non-anchored patterns.
1499
1500 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1501 \R, and also a number of cases that involve Unicode properties, both
1502 explicit and implicit when PCRE_UCP is set.
1503
1504 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1505 input, it could crash or give wrong results if characters with values
1506 greater than 0xc0 were present in the subject string. (Detail: it assumed
1507 UTF-8 input when processing these items.)
1508
1509 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1510 size_t is 64-bit (#991).
1511
1512 17. Added a check for running out of memory when PCRE is compiled with
1513 --disable-stack-for-recursion (#990).
1514
1515 18. If the last data line in a file for pcretest does not have a newline on
1516 the end, a newline was missing in the output.
1517
1518 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1519 less than 128) in its various bitmaps. However, there is a facility for
1520 generating tables according to the current locale when PCRE is compiled. It
1521 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1522 characters, are recognized by isspace() and therefore were getting set in
1523 these tables, and indeed these tables seem to approximate to ISO 8859. This
1524 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1525 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1526 which of course cannot start UTF-8 characters. I have changed the code so
1527 that only real ASCII characters (less than 128) and the correct starting
1528 bytes for UTF-8 encodings are set for characters greater than 127 when in
1529 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1530 altogether.)
1531
1532 20. Added the /T option to pcretest so as to be able to run tests with non-
1533 standard character tables, thus making it possible to include the tests
1534 used for 19 above in the standard set of tests.
1535
1536 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1537 reference to a subpattern the other side of a comment that contains an
1538 opening parenthesis caused either an internal compiling error, or a
1539 reference to the wrong subpattern.
1540
1541
1542 Version 8.02 19-Mar-2010
1543 ------------------------
1544
1545 1. The Unicode data tables have been updated to Unicode 5.2.0.
1546
1547 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1548 configured.
1549
1550 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1551 original author of that file, following a query about its status.
1552
1553 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1554 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1555
1556 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1557 quantifier applied to a forward-referencing subroutine call, could compile
1558 incorrect code or give the error "internal error: previously-checked
1559 referenced subpattern not found".
1560
1561 6. Both MS Visual Studio and Symbian OS have problems with initializing
1562 variables to point to external functions. For these systems, therefore,
1563 pcre_malloc etc. are now initialized to local functions that call the
1564 relevant global functions.
1565
1566 7. There were two entries missing in the vectors called coptable and poptable
1567 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1568 I've fixed the data, and added a kludgy way of testing at compile time that
1569 the lengths are correct (equal to the number of opcodes).
1570
1571 8. Following on from 7, I added a similar kludge to check the length of the
1572 eint vector in pcreposix.c.
1573
1574 9. Error texts for pcre_compile() are held as one long string to avoid too
1575 much relocation at load time. To find a text, the string is searched,
1576 counting zeros. There was no check for running off the end of the string,
1577 which could happen if a new error number was added without updating the
1578 string.
1579
1580 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1581
1582 11. \K was not working if it appeared in an atomic group or in a group that
1583 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1584 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1585 the assertion is positive, but not if it is negative.
1586
1587 12. Change 11 fortuitously reduced the size of the stack frame used in the
1588 "match()" function of pcre_exec.c by one pointer. Forthcoming
1589 implementation of support for (*MARK) will need an extra pointer on the
1590 stack; I have reserved it now, so that the stack frame size does not
1591 decrease.
1592
1593 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1594 item in branch that calls a recursion is a subroutine call - as in the
1595 second branch in the above example - was incorrectly given the compile-
1596 time error "recursive call could loop indefinitely" because pcre_compile()
1597 was not correctly checking the subroutine for matching a non-empty string.
1598
1599 14. The checks for overrunning compiling workspace could trigger after an
1600 overrun had occurred. This is a "should never occur" error, but it can be
1601 triggered by pathological patterns such as hundreds of nested parentheses.
1602 The checks now trigger 100 bytes before the end of the workspace.
1603
1604 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1605
1606
1607 Version 8.01 19-Jan-2010
1608 ------------------------
1609
1610 1. If a pattern contained a conditional subpattern with only one branch (in
1611 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1612 computed the wrong minimum data length (which is of course zero for such
1613 subpatterns). This could cause incorrect "no match" results.
1614
1615 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1616 the pattern is reset in the first branch, pcre_compile() failed with
1617 "internal error: code overflow at offset...". This happened only when
1618 the reset was to the original external option setting. (An optimization
1619 abstracts leading options settings into an external setting, which was the
1620 cause of this.)
1621
1622 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1623 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1624 assertion pattern did not match (meaning that the assertion was true), it
1625 was incorrectly treated as false if the SKIP had been reached during the
1626 matching. This also applied to assertions used as conditions.
1627
1628 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1629 assertion subpattern, including such a pattern used as a condition,
1630 unpredictable results occurred, instead of the error return
1631 PCRE_ERROR_DFA_UITEM.
1632
1633 5. The C++ GlobalReplace function was not working like Perl for the special
1634 situation when an empty string is matched. It now does the fancy magic
1635 stuff that is necessary.
1636
1637 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1638 removed. (These were left over from very, very early versions of PCRE.)
1639
1640 7. Some cosmetic changes to the code to make life easier when compiling it
1641 as part of something else:
1642
1643 (a) Change DEBUG to PCRE_DEBUG.
1644
1645 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1646 called "current" as "current_branch", to prevent a collision with the
1647 Linux macro when compiled as a kernel module.
1648
1649 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1650 prevent a collision with the Linux macro when compiled as a kernel
1651 module.
1652
1653 8. In pcre_compile() there are some checks for integer overflows that used to
1654 cast potentially large values to (double). This has been changed to that
1655 when building, a check for int64_t is made, and if it is found, it is used
1656 instead, thus avoiding the use of floating point arithmetic. (There is no
1657 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1658 double.
1659
1660 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1661 2005 (difference between two addresses compared to an unsigned value).
1662
1663 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1664 custom one, because of the following reported problem in Windows:
1665
1666 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1667 under Win32.
1668 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1669 therefore missing the function definition.
1670 - The compiler thus generates a "C" signature for the test function.
1671 - The linker fails to find the "C" function.
1672 - PCRE fails to configure if asked to do so against libbz2.
1673
1674 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1675 messages were output:
1676
1677 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1678 rerunning libtoolize, to keep the correct libtool macros in-tree.
1679 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1680
1681 I have done both of these things.
1682
1683 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1684 most of the time, it *can* run out if it is given a pattern that contains a
1685 runaway infinite recursion. I updated the discussion in the pcrestack man
1686 page.
1687
1688 13. Now that we have gone to the x.xx style of version numbers, the minor
1689 version may start with zero. Using 08 or 09 is a bad idea because users
1690 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1691 interpreted as invalid octal numbers. I've updated the previous comment in
1692 configure.ac, and also added a check that gives an error if 08 or 09 are
1693 used.
1694
1695 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1696 causing partial matching to fail when the end of the subject matched \W
1697 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1698
1699 15. There were some discrepancies between the declarations in pcre_internal.h
1700 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1701 their definitions. The declarations used "const uschar *" and the
1702 definitions used USPTR. Even though USPTR is normally defined as "const
1703 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1704 reported that: "This difference in casting confuses some C++ compilers, for
1705 example, SunCC recognizes above declarations as different functions and
1706 generates broken code for hbpcre." I have changed the declarations to use
1707 USPTR.
1708
1709 16. GNU libtool is named differently on some systems. The autogen.sh script now
1710 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1711 (FreeBSD).
1712
1713 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1714 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1715 comment: "Figure out how to create a longlong from a string: strtoll and
1716 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1717 instance, but it only takes 2 args instead of 3!"
1718
1719 18. A subtle bug concerned with back references has been fixed by a change of
1720 specification, with a corresponding code fix. A pattern such as
1721 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1722 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1723 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1724 same bug. Such groups have to be quantified to be useful, or contained
1725 inside another quantified group. (If there's no repetition, the reference
1726 can never match.) The problem arises because, having left the group and
1727 moved on to the rest of the pattern, a later failure that backtracks into
1728 the group uses the captured value from the final iteration of the group
1729 rather than the correct earlier one. I have fixed this in PCRE by forcing
1730 any group that contains a reference to itself to be an atomic group; that
1731 is, there cannot be any backtracking into it once it has completed. This is
1732 similar to recursive and subroutine calls.
1733
1734
1735 Version 8.00 19-Oct-09
1736 ----------------------
1737
1738 1. The table for translating pcre_compile() error codes into POSIX error codes
1739 was out-of-date, and there was no check on the pcre_compile() error code
1740 being within the table. This could lead to an OK return being given in
1741 error.
1742
1743 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1744 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1745 in a Windows environment.
1746
1747 3. The pcregrep --count option prints the count for each file even when it is
1748 zero, as does GNU grep. However, pcregrep was also printing all files when
1749 --files-with-matches was added. Now, when both options are given, it prints
1750 counts only for those files that have at least one match. (GNU grep just
1751 prints the file name in this circumstance, but including the count seems
1752 more useful - otherwise, why use --count?) Also ensured that the
1753 combination -clh just lists non-zero counts, with no names.
1754
1755 4. The long form of the pcregrep -F option was incorrectly implemented as
1756 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1757 but it seems right to fix it, and I didn't think it was worth preserving
1758 the old behaviour.
1759
1760 5. The command line items --regex=pattern and --regexp=pattern were not
1761 recognized by pcregrep, which required --regex pattern or --regexp pattern
1762 (with a space rather than an '='). The man page documented the '=' forms,
1763 which are compatible with GNU grep; these now work.
1764
1765 6. No libpcreposix.pc file was created for pkg-config; there was just
1766 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1767
1768 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1769 when UCP support is not needed, by modifying the Python script that
1770 generates it from Unicode data files. This should not matter if the module
1771 is correctly used as a library, but I received one complaint about 50K of
1772 unwanted data. My guess is that the person linked everything into his
1773 program rather than using a library. Anyway, it does no harm.
1774
1775 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1776 was a minimum greater than 1 for a wide character in a possessive
1777 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1778 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1779 character. Chaos in the form of incorrect output or a compiling loop could
1780 result.
1781
1782 9. The restrictions on what a pattern can contain when partial matching is
1783 requested for pcre_exec() have been removed. All patterns can now be
1784 partially matched by this function. In addition, if there are at least two
1785 slots in the offset vector, the offset of the earliest inspected character
1786 for the match and the offset of the end of the subject are set in them when
1787 PCRE_ERROR_PARTIAL is returned.
1788
1789 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1790 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1791 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1792 and may be more useful for multi-segment matching.
1793
1794 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1795 used to be given if ever the end of the subject was reached; now it is
1796 given only if matching could not proceed because another character was
1797 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1798 string "Z", which now yields "no match" instead of "partial match". In the
1799 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1800 final character ended with (*FAIL).
1801
1802 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1803 if the pattern had a "must contain" character that was already found in the
1804 earlier partial match, unless partial matching was again requested. For
1805 example, with the pattern /dog.(body)?/, the "must contain" character is
1806 "g". If the first part-match was for the string "dog", restarting with
1807 "sbody" failed. This bug has been fixed.
1808
1809 13. The string returned by pcre_dfa_exec() after a partial match has been
1810 changed so that it starts at the first inspected character rather than the
1811 first character of the match. This makes a difference only if the pattern
1812 starts with a lookbehind assertion or \b or \B (\K is not supported by
1813 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1814 matching functions compatible, and I think it's the right thing to do.
1815
1816 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1817 so that the demonstration program is easily available in environments where
1818 PCRE has not been installed from source.
1819
1820 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1821 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1822 library.
1823
1824 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1825 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1826 is not the first non-POSIX option to be added. Clearly some people find
1827 these options useful.
1828
1829 17. If a caller to the POSIX matching function regexec() passes a non-zero
1830 value for nmatch with a NULL value for pmatch, the value of
1831 nmatch is forced to zero.
1832
1833 18. RunGrepTest did not have a test for the availability of the -u option of
1834 the diff command, as RunTest does. It now checks in the same way as
1835 RunTest, and also checks for the -b option.
1836
1837 19. If an odd number of negated classes containing just a single character
1838 interposed, within parentheses, between a forward reference to a named
1839 subpattern and the definition of the subpattern, compilation crashed with
1840 an internal error, complaining that it could not find the referenced
1841 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1842 [The bug was that it was starting one character too far in when skipping
1843 over the character class, thus treating the ] as data rather than
1844 terminating the class. This meant it could skip too much.]
1845
1846 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1847 /g option in pcretest when the pattern contains \K, which makes it possible
1848 to have an empty string match not at the start, even when the pattern is
1849 anchored. Updated pcretest and pcredemo to use this option.
1850
1851 21. If the maximum number of capturing subpatterns in a recursion was greater
1852 than the maximum at the outer level, the higher number was returned, but
1853 with unset values at the outer level. The correct (outer level) value is
1854 now given.
1855
1856 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1857 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1858 make it do so. The string so far is captured, making this feature
1859 compatible with Perl.
1860
1861 23. The tests have been re-organized, adding tests 11 and 12, to make it
1862 possible to check the Perl 5.10 features against Perl 5.10.
1863
1864 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1865 pattern matches a fixed length string. PCRE did not allow this; now it
1866 does. Neither allows recursion.
1867
1868 25. I finally figured out how to implement a request to provide the minimum
1869 length of subject string that was needed in order to match a given pattern.
1870 (It was back references and recursion that I had previously got hung up
1871 on.) This code has now been added to pcre_study(); it finds a lower bound
1872 to the length of subject needed. It is not necessarily the greatest lower
1873 bound, but using it to avoid searching strings that are too short does give
1874 some useful speed-ups. The value is available to calling programs via
1875 pcre_fullinfo().
1876
1877 26. While implementing 25, I discovered to my embarrassment that pcretest had
1878 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1879 study optimizations had never been tested with that matching function.
1880 Oops. What is worse, even when it was passed study data, there was a bug in
1881 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1882 were also very few tests of studied patterns with pcre_dfa_exec().
1883
1884 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1885 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1886 on the other side of the coin, they are no longer allowed to have different
1887 names, because these cannot be distinguished in PCRE, and this has caused
1888 confusion. (This is a difference from Perl.)
1889
1890 28. When duplicate subpattern names are present (necessarily with different
1891 numbers, as required by 27 above), and a test is made by name in a
1892 conditional pattern, either for a subpattern having been matched, or for
1893 recursion in such a pattern, all the associated numbered subpatterns are
1894 tested, and the overall condition is true if the condition is true for any
1895 one of them. This is the way Perl works, and is also more like the way
1896 testing by number works.
1897
1898
1899 Version 7.9 11-Apr-09
1900 ---------------------
1901
1902 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1903 (pcretest), all targets were linked against these libraries. This included
1904 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1905 libraries. This caused unwanted dependencies to be created. This problem
1906 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1907 pcretest is linked with readline.
1908
1909 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1910 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1911 moved outside it again, because FALSE and TRUE are already defined in AIX,
1912 but BOOL is not.
1913
1914 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1915 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1916
1917 4. The pcregrep documentation said spaces were inserted as well as colons (or
1918 hyphens) following file names and line numbers when outputting matching
1919 lines. This is not true; no spaces are inserted. I have also clarified the
1920 wording for the --colour (or --color) option.
1921
1922 5. In pcregrep, when --colour was used with -o, the list of matching strings
1923 was not coloured; this is different to GNU grep, so I have changed it to be
1924 the same.
1925
1926 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1927 each matching line was coloured. Now it goes on to look for further matches
1928 of any of the test patterns, which is the same behaviour as GNU grep.
1929
1930 7. A pattern that could match an empty string could cause pcregrep to loop; it
1931 doesn't make sense to accept an empty string match in pcregrep, so I have
1932 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1933 seems to be how GNU grep behaves. [But see later change 40 for release
1934 8.33.]
1935
1936 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1937 start or after a newline", because the conditional assertion was not being
1938 correctly handled. The rule now is that both the assertion and what follows
1939 in the first alternative must satisfy the test.
1940
1941 9. If auto-callout was enabled in a pattern with a conditional group whose
1942 condition was an assertion, PCRE could crash during matching, both with
1943 pcre_exec() and pcre_dfa_exec().
1944
1945 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1946 used for matching.
1947
1948 11. Unicode property support in character classes was not working for
1949 characters (bytes) greater than 127 when not in UTF-8 mode.
1950
1951 12. Added the -M command line option to pcretest.
1952
1953 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1954
1955 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1956
1957 16. Added comments and documentation about mis-use of no_arg in the C++
1958 wrapper.
1959
1960 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1961 from Martin Jerabek that uses macro names for all relevant character and
1962 string constants.
1963
1964 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1965 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1966 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1967 these, but not everybody uses configure.
1968
1969 19. A conditional group that had only one branch was not being correctly
1970 recognized as an item that could match an empty string. This meant that an
1971 enclosing group might also not be so recognized, causing infinite looping
1972 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1973 with the subject "ab", where knowledge that the repeated group can match
1974 nothing is needed in order to break the loop.
1975
1976 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1977 exec(), but without supplying a callout function, matching went wrong.
1978
1979 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1980 leak if the size of the offset vector was greater than 30. When the vector
1981 is smaller, the saved offsets during recursion go onto a local stack
1982 vector, but for larger vectors malloc() is used. It was failing to free
1983 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1984 error, in fact).
1985
1986 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1987 heapframe that is used only when UTF-8 support is enabled. This caused no
1988 problem, but was untidy.
1989
1990 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1991 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1992 included within another project.
1993
1994 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1995 slightly modified by me:
1996
1997 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1998 not building pcregrep.
1999
2000 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2001 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2002
2003 25. Forward references, both numeric and by name, in patterns that made use of
2004 duplicate group numbers, could behave incorrectly or give incorrect errors,
2005 because when scanning forward to find the reference group, PCRE was not
2006 taking into account the duplicate group numbers. A pattern such as
2007 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2008
2009 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2010 the feature of a custom pointer more persuasive (as requested by a user).
2011
2012 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2013 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2014 pre-defined.
2015
2016 28. Added support for (*UTF8) at the start of a pattern.
2017
2018 29. Arrange for flags added by the "release type" setting in CMake to be shown
2019 in the configuration summary.
2020
2021
2022 Version 7.8 05-Sep-08
2023 ---------------------
2024
2025 1. Replaced UCP searching code with optimized version as implemented for Ad
2026 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2027 stage table and inline lookup instead of a function, giving speed ups of 2
2028 to 5 times on some simple patterns that I tested. Permission was given to
2029 distribute the MultiStage2.py script that generates the tables (it's not in
2030 the tarball, but is in the Subversion repository).
2031
2032 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2033 scripts.
2034
2035 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2036 a group with a zero qualifier. The result of the study could be incorrect,
2037 or the function might crash, depending on the pattern.
2038
2039 4. Caseless matching was not working for non-ASCII characters in back
2040 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2041 It now works when Unicode Property Support is available.
2042
2043 5. In pcretest, an escape such as \x{de} in the data was always generating
2044 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2045 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2046 truncation.
2047
2048 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2049
2050 7. Added two (int) casts to pcregrep when printing the difference of two
2051 pointers, in case they are 64-bit values.
2052
2053 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2054 test 2 if it fails.
2055
2056 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2057 and a #define of that name to empty if it is not externally set. This is to
2058 allow users of MSVC to set it if necessary.
2059
2060 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2061 the convenience functions in the pcre_get.c source file.
2062
2063 11. An option change at the start of a pattern that had top-level alternatives
2064 could cause overwriting and/or a crash. This command provoked a crash in
2065 some environments:
2066
2067 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2068
2069 This potential security problem was recorded as CVE-2008-2371.
2070
2071 12. For a pattern where the match had to start at the beginning or immediately
2072 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2073 pcre_dfa_exec() could read past the end of the passed subject if there was
2074 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2075 pcretest so that it places the subject at the end of its malloc-ed buffer.
2076
2077 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2078 exec() might read past the end of the data buffer in UTF-8 mode.
2079
2080 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2081 the data contained the byte 0x85 as part of a UTF-8 character within its
2082 first line. This applied both to normal and DFA matching.
2083
2084 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2085 /^[^d]*?$/8 failed to match "abc".
2086
2087 16. Added a missing copyright notice to pcrecpp_internal.h.
2088
2089 17. Make it more clear in the documentation that values returned from
2090 pcre_exec() in ovector are byte offsets, not character counts.
2091
2092 18. Tidied a few places to stop certain compilers from issuing warnings.
2093
2094 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2095 supplied by Stefan Weber. I made a further small update for 7.8 because
2096 there is a change of source arrangements: the pcre_searchfuncs.c module is
2097 replaced by pcre_ucd.c.
2098
2099
2100 Version 7.7 07-May-08
2101 ---------------------
2102
2103 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2104 a string to a long long, pretend we don't even have a long long." This is
2105 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2106
2107 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2108 pre-7.6 versions, which defined a global no_arg variable instead of putting
2109 it in the RE class. (See also #8 below.)
2110
2111 3. Remove a line of dead code, identified by coverity and reported by Nuno
2112 Lopes.
2113
2114 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2115
2116 (1) The include/exclude patterns were being applied to the whole pathnames
2117 of files, instead of just to the final components.
2118
2119 (2) If there was more than one level of directory, the subdirectories were
2120 skipped unless they satisfied the include/exclude conditions. This is
2121 inconsistent with GNU grep (and could even be seen as contrary to the
2122 pcregrep specification - which I improved to make it absolutely clear).
2123 The action now is always to scan all levels of directory, and just
2124 apply the include/exclude patterns to regular files.
2125
2126 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2127 --exclude_dir in the tests to avoid scanning .svn directories.
2128
2129 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2130 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2131 doesn't support NULs in patterns.
2132
2133 7. Added some missing "const"s to declarations of static tables in
2134 pcre_compile.c and pcre_dfa_exec.c.
2135
2136 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2137 caused by fix #2 above. (Subsequently also a second patch to fix the
2138 first patch. And a third patch - this was a messy problem.)
2139
2140 9. Applied Craig's patch to remove the use of push_back().
2141
2142 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2143 matching function regexec().
2144
2145 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2146 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2147 references. PCRE supports relative numbers with this syntax (I don't think
2148 Oniguruma does).
2149
2150 12. Previously, a group with a zero repeat such as (...){0} was completely
2151 omitted from the compiled regex. However, this means that if the group
2152 was called as a subroutine from elsewhere in the pattern, things went wrong
2153 (an internal error was given). Such groups are now left in the compiled
2154 pattern, with a new opcode that causes them to be skipped at execution
2155 time.
2156
2157 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2158 to the way PCRE behaves:
2159
2160 (a) A lone ] character is dis-allowed (Perl treats it as data).
2161
2162 (b) A back reference to an unmatched subpattern matches an empty string
2163 (Perl fails the current match path).
2164
2165 (c) A data ] in a character class must be notated as \] because if the
2166 first data character in a class is ], it defines an empty class. (In
2167 Perl it is not possible to have an empty class.) The empty class []
2168 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2169 The negative empty class [^] matches any one character, independently
2170 of the DOTALL setting.
2171
2172 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2173 non-existent subpattern following a character class starting with ']' and
2174 containing () gave an internal compiling error instead of "reference to
2175 non-existent subpattern". Fortunately, when the pattern did exist, the
2176 compiled code was correct. (When scanning forwards to check for the
2177 existence of the subpattern, it was treating the data ']' as terminating
2178 the class, so got the count wrong. When actually compiling, the reference
2179 was subsequently set up correctly.)
2180
2181 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2182 it was being rejected as not supported by pcre_dfa_exec(), even though
2183 other assertions are supported. I have made pcre_dfa_exec() support
2184 (*FAIL).
2185
2186 16. The implementation of 13c above involved the invention of a new opcode,
2187 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2188 cannot be changed at match time, I realized I could make a small
2189 improvement to matching performance by compiling OP_ALLANY instead of
2190 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2191 on the OP_ANY path.
2192
2193 17. Compiling pcretest on Windows with readline support failed without the
2194 following two fixes: (1) Make the unistd.h include conditional on
2195 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2196
2197 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2198 ncurses library to be included for pcretest when ReadLine support is
2199 requested, but also to allow for it to be overridden. This patch came from
2200 Daniel Bergström.
2201
2202 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2203 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2204 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2205 spotting this.
2206
2207
2208 Version 7.6 28-Jan-08
2209 ---------------------
2210
2211 1. A character class containing a very large number of characters with
2212 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2213 overflow.
2214
2215 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2216 HAVE_LONG_LONG is not defined.
2217
2218 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2219 bring them up to date and include new features. This patch includes:
2220
2221 - Fixed PH's badly added libz and libbz2 support.
2222 - Fixed a problem with static linking.
2223 - Added pcredemo. [But later removed - see 7 below.]
2224 - Fixed dftables problem and added an option.
2225 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2226 HAVE_LONG_LONG.
2227 - Added readline support for pcretest.
2228 - Added an listing of the option settings after cmake has run.
2229
2230 4. A user submitted a patch to Makefile that makes it easy to create
2231 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2232 Makefile.am that cause it to include this special target, without
2233 affecting anything else. Note that the same mingw target plus all
2234 the other distribution libraries and programs are now supported
2235 when configuring with CMake (see 6 below) instead of with
2236 Configure/Make.
2237
2238 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2239 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2240 exported in the Windows port". It has not yet been confirmed that the patch
2241 solves the problem, but it does no harm.
2242
2243 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2244 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2245 with CMake, and also correct the comment about stack recursion.
2246
2247 7. Remove the automatic building of pcredemo from the ./configure system and
2248 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2249 of a program that users should build themselves after PCRE is installed, so
2250 building it automatically is not really right. What is more, it gave
2251 trouble in some build environments.
2252
2253 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2254
2255
2256 Version 7.5 10-Jan-08
2257 ---------------------
2258
2259 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2260 values in parens when parsing an RE using the C++ wrapper."
2261
2262 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2263 Characters greater than 255 were excluded from the class instead of being
2264 included.
2265
2266 3. The same bug as (2) above applied to negated POSIX classes such as
2267 [:^space:].
2268
2269 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2270 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2271 I have changed it.
2272
2273 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2274 first named subpattern) and a construct such as (?&a) would reference the
2275 first named subpattern whose name started with "a" (in other words, the
2276 length check was missing). Both these problems are fixed. "Subpattern name
2277 expected" is now given for (?&) (a zero-length name), and this patch also
2278 makes it give the same error for \k'' (previously it complained that that
2279 was a reference to a non-existent subpattern).
2280
2281 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2282 this is right because (?- can be followed by option settings as well as by
2283 digits. I have, however, made the messages clearer.
2284
2285 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2286 than the number used in the conditional) now cause a compile-time error.
2287 This is actually not compatible with Perl, which accepts such patterns, but
2288 treats the conditional as always being FALSE (as PCRE used to), but it
2289 seems to me that giving a diagnostic is better.
2290
2291 8. Change "alphameric" to the more common word "alphanumeric" in comments
2292 and messages.
2293
2294 9. Fix two occurrences of "backslash" in comments that should have been
2295 "backspace".
2296
2297 10. Remove two redundant lines of code that can never be obeyed (their function
2298 was moved elsewhere).
2299
2300 11. The program that makes PCRE's Unicode character property table had a bug
2301 which caused it to generate incorrect table entries for sequences of
2302 characters that have the same character type, but are in different scripts.
2303 It amalgamated them into a single range, with the script of the first of
2304 them. In other words, some characters were in the wrong script. There were
2305 thirteen such cases, affecting characters in the following ranges:
2306
2307 U+002b0 - U+002c1
2308 U+0060c - U+0060d
2309 U+0061e - U+00612
2310 U+0064b - U+0065e
2311 U+0074d - U+0076d
2312 U+01800 - U+01805
2313 U+01d00 - U+01d77
2314 U+01d9b - U+01dbf
2315 U+0200b - U+0200f
2316 U+030fc - U+030fe
2317 U+03260 - U+0327f
2318 U+0fb46 - U+0fbb1
2319 U+10450 - U+1049d
2320
2321 12. The -o option (show only the matching part of a line) for pcregrep was not
2322 compatible with GNU grep in that, if there was more than one match in a
2323 line, it showed only the first of them. It now behaves in the same way as
2324 GNU grep.
2325
2326 13. If the -o and -v options were combined for pcregrep, it printed a blank
2327 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2328 does the same. The return code can be used to tell if there were any
2329 non-matching lines.
2330
2331 14. Added --file-offsets and --line-offsets to pcregrep.
2332
2333 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2334 infinitely looping recursion. The bug was that positive lookaheads were not
2335 being skipped when checking for a possible empty match (negative lookaheads
2336 and both kinds of lookbehind were skipped).
2337
2338 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2339 inclusion of <windows.h> to before rather than after the definition of
2340 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2341
2342 17. Specifying a possessive quantifier with a specific limit for a Unicode
2343 character property caused pcre_compile() to compile bad code, which led at
2344 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2345 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2346 caused the error; without that there was no problem.
2347
2348 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2349
2350 19. Added --enable-pcretest-libreadline.
2351
2352 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2353 RE::GlobalReplace(). As a result, the number of replacements returned was
2354 double what it should be. I removed one of the increments, but Craig sent a
2355 later patch that removed the other one (the right fix) and added unit tests
2356 that check the return values (which was not done before).
2357
2358 21. Several CMake things:
2359
2360 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2361 the names libpcre and libpcreposix, not just pcre and pcreposix.
2362
2363 (2) The above change means that pcretest and pcregrep are now correctly
2364 linked with the newly-built libraries, not previously installed ones.
2365
2366 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2367
2368 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2369 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2370 UTF-8 newline character). The key issue is that the pattern starts .*;
2371 this means that the match must be either at the beginning, or after a
2372 newline. The bug was in the code for advancing after a failed match and
2373 checking that the new position followed a newline. It was not taking
2374 account of UTF-8 characters correctly.
2375
2376 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2377 character classes. PCRE was not treating the sequence [:...:] as a
2378 character class unless the ... were all letters. Perl, however, seems to
2379 allow any characters between [: and :], though of course it rejects as
2380 unknown any "names" that contain non-letters, because all the known class
2381 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2382 for example, whereas PCRE did not - it did not recognize a POSIX character
2383 class. This seemed a bit dangerous, so the code has been changed to be
2384 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2385 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2386 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2387 Perl does, and where it didn't before.
2388
2389 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2390 Windows environments %n is disabled by default.
2391
2392
2393 Version 7.4 21-Sep-07
2394 ---------------------
2395
2396 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2397 means that a class such as [\s] counted as "explicit reference to CR or
2398 LF". That isn't really right - the whole point of the change was to try to
2399 help when there was an actual mention of one of the two characters. So now
2400 the change happens only if \r or \n (or a literal CR or LF) character is
2401 encountered.
2402
2403 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2404 of both had grown to the point where there were only 3 bits left.
2405 Fortunately, there was spare space in the data structure, and so I have
2406 moved the internal flags into a new 16-bit field to free up more option
2407 bits.
2408
2409 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2410 but did not set the internal JCHANGED flag - either of these is enough to
2411 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2412 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2413 start sets both bits.
2414
2415 4. Added options (at build time, compile time, exec time) to change \R from
2416 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2417
2418 5. doc/pcresyntax.html was missing from the distribution.
2419
2420 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2421 compatibility, even though it is no longer used.
2422
2423 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2424 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2425 windows.h file is present (where different names are used). [This was
2426 reversed later after testing - see 16 below.]
2427
2428 8. Changed all #include <config.h> to #include "config.h". There were also
2429 some further <pcre.h> cases that I changed to "pcre.h".
2430
2431 9. When pcregrep was used with the --colour option, it missed the line ending
2432 sequence off the lines that it output.
2433
2434 10. It was pointed out to me that arrays of string pointers cause lots of
2435 relocations when a shared library is dynamically loaded. A technique of
2436 using a single long string with a table of offsets can drastically reduce
2437 these. I have refactored PCRE in four places to do this. The result is
2438 dramatic:
2439
2440 Originally: 290
2441 After changing UCP table: 187
2442 After changing error message table: 43
2443 After changing table of "verbs" 36
2444 After changing table of Posix names 22
2445
2446 Thanks to the folks working on Gregex for glib for this insight.
2447
2448 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2449 unicode-properties was also set.
2450
2451 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2452
2453 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2454 checked only for CRLF.
2455
2456 14. Added casts to pcretest.c to avoid compiler warnings.
2457
2458 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2459
2460 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2461 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2462 entirely. This removes changes made in 7 above.
2463
2464 17. The CMake files have been updated, and there is now more information about
2465 building with CMake in the NON-UNIX-USE document.
2466
2467
2468 Version 7.3 28-Aug-07
2469 ---------------------
2470
2471 1. In the rejigging of the build system that eventually resulted in 7.1, the
2472 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2473 brackets there is not right, since it causes compilers to look for an
2474 installed pcre.h, not the version that is in the source that is being
2475 compiled (which of course may be different). I have changed it back to:
2476
2477 #include "pcre.h"
2478
2479 I have a vague recollection that the change was concerned with compiling in
2480 different directories, but in the new build system, that is taken care of
2481 by the VPATH setting the Makefile.
2482
2483 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2484 when the subject happened to end in the byte 0x85 (e.g. if the last
2485 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2486 characters but of course it shouldn't be taken as a newline when it is part
2487 of another character. The bug was that, for an unlimited repeat of . in
2488 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2489 characters when looking for a newline.
2490
2491 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2492
2493 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2494 in debug output.
2495
2496 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2497 long printing in the pcrecpp unittest when running under MinGW.
2498
2499 6. ESC_K was left out of the EBCDIC table.
2500
2501 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2502 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2503 limit also applies to "virtual nesting" when a pattern is recursive, and in
2504 this case 1000 isn't so big. I have been able to remove this limit at the
2505 expense of backing off one optimization in certain circumstances. Normally,
2506 when pcre_exec() would call its internal match() function recursively and
2507 immediately return the result unconditionally, it uses a "tail recursion"
2508 feature to save stack. However, when a subpattern that can match an empty
2509 string has an unlimited repetition quantifier, it no longer makes this
2510 optimization. That gives it a stack frame in which to save the data for
2511 checking that an empty string has been matched. Previously this was taken
2512 from the 1000-entry workspace that had been reserved. So now there is no
2513 explicit limit, but more stack is used.
2514
2515 8. Applied Daniel's patches to solve problems with the import/export magic
2516 syntax that is required for Windows, and which was going wrong for the
2517 pcreposix and pcrecpp parts of the library. These were overlooked when this
2518 problem was solved for the main library.
2519
2520 9. There were some crude static tests to avoid integer overflow when computing
2521 the size of patterns that contain repeated groups with explicit upper
2522 limits. As the maximum quantifier is 65535, the maximum group length was
2523 set at 30,000 so that the product of these two numbers did not overflow a
2524 32-bit integer. However, it turns out that people want to use groups that
2525 are longer than 30,000 bytes (though not repeat them that many times).
2526 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2527 made it possible to implement the integer overflow checks in a much more
2528 dynamic way, which I have now done. The artificial limitation on group
2529 length has been removed - we now have only the limit on the total length of
2530 the compiled pattern, which depends on the LINK_SIZE setting.
2531
2532 10. Fixed a bug in the documentation for get/copy named substring when
2533 duplicate names are permitted. If none of the named substrings are set, the
2534 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2535 empty string.
2536
2537 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2538 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2539 because the ] is interpreted as the first data character and the
2540 terminating ] is not found. PCRE has been made compatible with Perl in this
2541 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2542 cause memory overwriting.
2543
2544 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2545 string has been matched (to stop an infinite loop). It was not recognizing
2546 a conditional subpattern that could match an empty string if that
2547 subpattern was within another subpattern. For example, it looped when
2548 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2549 condition was not nested. This bug has been fixed.
2550
2551 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2552 past the start of the subject in the presence of bytes with the top bit
2553 set, for example "\x8aBCD".
2554
2555 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2556 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2557
2558 14. Optimized (?!) to (*FAIL).
2559
2560 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2561 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2562 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2563 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2564 does: it's just the validity check that is more restrictive.
2565
2566 16. Inserted checks for integer overflows during escape sequence (backslash)
2567 processing, and also fixed erroneous offset values for syntax errors during
2568 backslash processing.
2569
2570 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2571 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2572
2573 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2574 caused an overrun.
2575
2576 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2577 something other than just ASCII characters) inside a group that had an
2578 unlimited repeat caused a loop at compile time (while checking to see
2579 whether the group could match an empty string).
2580
2581 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2582 [\P{Any}] did so. (Error in the code for printing property names.)
2583
2584 21. An orphan \E inside a character class could cause a crash.
2585
2586 22. A repeated capturing bracket such as (A)? could cause a wild memory
2587 reference during compilation.
2588
2589 23. There are several functions in pcre_compile() that scan along a compiled
2590 expression for various reasons (e.g. to see if it's fixed length for look
2591 behind). There were bugs in these functions when a repeated \p or \P was
2592 present in the pattern. These operators have additional parameters compared
2593 with \d, etc, and these were not being taken into account when moving along
2594 the compiled data. Specifically:
2595
2596 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2597 length.
2598
2599 (b) An item such as \pL+ within a repeated group could cause crashes or
2600 loops.
2601
2602 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2603 "reference to non-existent subpattern" error.
2604
2605 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2606
2607 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2608 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2609
2610 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2611
2612 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2613 character were causing crashes (broken optimization).
2614
2615 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2616 \p or \P) caused a compile-time loop.
2617
2618 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2619 break. For example, the unstudied pattern [\r\n]A does not match the string
2620 "\r\nA" because change 7.0/46 below moves the current point on by two
2621 characters after failing to match at the start. However, the pattern \nA
2622 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2623 the same is true. There doesn't seem any very clean way out of this, but
2624 what I have chosen to do makes the common cases work: PCRE now takes note
2625 of whether there can be an explicit match for \r or \n anywhere in the
2626 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2627 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2628 pattern has explicit CR or LF references.
2629
2630 29. Added (*CR) etc for changing newline setting at start of pattern.
2631
2632
2633 Version 7.2 19-Jun-07
2634 ---------------------
2635
2636 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2637 which is apparently normally available under Windows.
2638
2639 2. Re-jig the pcregrep tests with different newline settings in an attempt
2640 to make them independent of the local environment's newline setting.
2641
2642 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2643
2644 4. Some of the "internals" tests were previously cut out when the link size
2645 was not 2, because the output contained actual offsets. The recent new
2646 "Z" feature of pcretest means that these can be cut out, making the tests
2647 usable with all link sizes.
2648
2649 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2650 stack recursion. This gives a massive performance boost under BSD, but just
2651 a small improvement under Linux. However, it saves one field in the frame
2652 in all cases.
2653
2654 6. Added more features from the forthcoming Perl 5.10:
2655
2656 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2657 recursion call. It refers to the nth most recently opened parentheses.
2658
2659 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2660 to be opened parentheses.
2661
2662 (c) Conditions that refer to capturing parentheses can be specified
2663 relatively, for example, (?(-2)... or (?(+3)...
2664
2665 (d) \K resets the start of the current match so that everything before
2666 is not part of it.
2667
2668 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2669
2670 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2671 reference syntax.
2672
2673 (g) (?| introduces a group in which the numbering of parentheses in each
2674 alternative starts with the same number.
2675
2676 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2677
2678 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2679 PCRE_INFO_JCHANGED.
2680
2681 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2682 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2683 for detecting groups that can match an empty string.
2684
2685 9. A pattern with a very large number of alternatives (more than several
2686 hundred) was running out of internal workspace during the pre-compile
2687 phase, where pcre_compile() figures out how much memory will be needed. A
2688 bit of new cunning has reduced the workspace needed for groups with
2689 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2690 workspace instead of running out of the 4096 that are available.
2691
2692 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2693
2694 11. Applied patch from Google to remove an optimization that didn't quite work.
2695 The report of the bug said:
2696
2697 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2698 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2699 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2700
2701 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2702 it matched the wrong number of bytes.
2703
2704
2705 Version 7.1 24-Apr-07
2706 ---------------------
2707
2708 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2709 that is more "standard", making use of automake and other Autotools. There
2710 is some re-arrangement of the files and adjustment of comments consequent
2711 on this.
2712
2713 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2714 for recursive directory scanning broke on some systems because the files
2715 are not scanned in any specific order and on different systems the order
2716 was different. A call to "sort" has been inserted into RunGrepTest for the
2717 approprate test as a short-term fix. In the longer term there may be an
2718 alternative.
2719
2720 3. I had an email from Eric Raymond about problems translating some of PCRE's
2721 man pages to HTML (despite the fact that I distribute HTML pages, some
2722 people do their own conversions for various reasons). The problems
2723 concerned the use of low-level troff macros .br and .in. I have therefore
2724 removed all such uses from the man pages (some were redundant, some could
2725 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2726 HTML has been updated to handle .nf/.fi and to complain if it encounters
2727 .br or .in.
2728
2729 4. Updated comments in configure.ac that get placed in config.h.in and also
2730 arranged for config.h to be included in the distribution, with the name
2731 config.h.generic, for the benefit of those who have to compile without
2732 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2733
2734 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2735 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2736 makevp.bat for the current PCRE, using the additional files
2737 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2738
2739 6. A Windows user reported a minor discrepancy with test 2, which turned out
2740 to be caused by a trailing space on an input line that had got lost in his
2741 copy. The trailing space was an accident, so I've just removed it.
2742
2743 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2744 that is needed.
2745
2746 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2747 as "const" (a) because they are and (b) because it helps the PHP
2748 maintainers who have recently made a script to detect big data structures
2749 in the php code that should be moved to the .rodata section. I remembered
2750 to update Builducptable as well, so it won't revert if ucptable.h is ever
2751 re-created.
2752
2753 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2754 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2755 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2756 support is not required. This saves 1.5-2K of code, which is important in
2757 some applications.
2758
2759 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2760 so as not to refer to the tables, even though these functions will never be
2761 called when UTF-8 support is disabled. Otherwise there are problems with a
2762 shared library.
2763
2764 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2765
2766 (a) It was defining its arguments as char * instead of void *.
2767
2768 (b) It was assuming that all moves were upwards in memory; this was true
2769 a long time ago when I wrote it, but is no longer the case.
2770
2771 The emulated memove() is provided for those environments that have neither
2772 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2773 is clearly not the case, as these two bugs were recently reported.
2774
2775 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2776 and Detrail to create the HTML documentation, the .txt form of the man
2777 pages, and it removes trailing spaces from listed files. It also creates
2778 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2779 case, it wraps all the #defines with #ifndefs. This script should be run
2780 before "make dist".
2781
2782 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2783 with Unicode property support.
2784
2785 (a) For a maximizing quantifier, if the two different cases of the
2786 character were of different lengths in their UTF-8 codings (there are
2787 some cases like this - I found 11), and the matching function had to
2788 back up over a mixture of the two cases, it incorrectly assumed they
2789 were both the same length.
2790
2791 (b) When PCRE was configured to use the heap rather than the stack for
2792 recursion during matching, it was not correctly preserving the data for
2793 the other case of a UTF-8 character when checking ahead for a match
2794 while processing a minimizing repeat. If the check also involved
2795 matching a wide character, but failed, corruption could cause an
2796 erroneous result when trying to check for a repeat of the original
2797 character.
2798
2799 13. Some tidying changes to the testing mechanism:
2800
2801 (a) The RunTest script now detects the internal link size and whether there
2802 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2803 values substituted by "configure". (The RunGrepTest script already did
2804 this for UTF-8.) The configure.ac script no longer substitutes the
2805 relevant variables.
2806
2807 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2808 with length and offset values. This means that the output is different
2809 for different internal link sizes. Test 2 is skipped for link sizes
2810 other than 2 because of this, bypassing the problem. Unfortunately,
2811 there was also a test in test 3 (the locale tests) that used /B and
2812 failed for link sizes other than 2. Rather than cut the whole test out,
2813 I have added a new /Z option to pcretest that replaces the length and
2814 offset values with spaces. This is now used to make test 3 independent
2815 of link size. (Test 2 will be tidied up later.)
2816
2817 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2818 segmentation fault instead of returning the appropriate error message.
2819
2820 15. In multiline mode when the newline sequence was set to "any", the pattern
2821 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2822 This doesn't seem right; it now treats the CRLF combination as the line
2823 ending, and so does not match in that case. It's only a pattern such as ^$
2824 that would hit this one: something like ^ABC$ would have failed after \r
2825 and then tried again after \r\n.
2826
2827 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2828 in an attempt to make files that differ only in their line terminators
2829 compare equal. This works on Linux.
2830
2831 17. Under certain error circumstances pcregrep might try to free random memory
2832 as it exited. This is now fixed, thanks to valgrind.
2833
2834 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2835 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2836 was because its rules for how to advance for /g after matching an empty
2837 string at the end of a line did not allow for this case. They now check for
2838 it specially.
2839
2840 20. pcretest is supposed to handle patterns and data of any length, by
2841 extending its buffers when necessary. It was getting this wrong when the
2842 buffer for a data line had to be extended.
2843
2844 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2845 CRLF as a newline sequence.
2846
2847 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2848 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2849 I have nevertheless tidied it up.
2850
2851 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2852
2853 24. Added a man page for pcre-config.
2854
2855
2856 Version 7.0 19-Dec-06
2857 ---------------------
2858
2859 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2860 moving to gcc 4.1.1.
2861
2862 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2863 sys/time.h, which is documented as needed for this function. It doesn't
2864 seem to matter on Linux, but it showed up on some releases of OS X.
2865
2866 3. It seems that there are systems where bytes whose values are greater than
2867 127 match isprint() in the "C" locale. The "C" locale should be the
2868 default when a C program starts up. In most systems, only ASCII printing
2869 characters match isprint(). This difference caused the output from pcretest
2870 to vary, making some of the tests fail. I have changed pcretest so that:
2871
2872 (a) When it is outputting text in the compiled version of a pattern, bytes
2873 other than 32-126 are always shown as hex escapes.
2874
2875 (b) When it is outputting text that is a matched part of a subject string,
2876 it does the same, unless a different locale has been set for the match
2877 (using the /L modifier). In this case, it uses isprint() to decide.
2878
2879 4. Fixed a major bug that caused incorrect computation of the amount of memory
2880 required for a compiled pattern when options that changed within the
2881 pattern affected the logic of the preliminary scan that determines the
2882 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2883 that the computed length was too small. The symptoms of this bug were
2884 either the PCRE error "internal error: code overflow" from pcre_compile(),
2885 or a glibc crash with a message such as "pcretest: free(): invalid next
2886 size (fast)". Examples of patterns that provoked this bug (shown in
2887 pcretest format) are:
2888
2889 /(?-x: )/x
2890 /(?x)(?-x: \s*#\s*)/
2891 /((?i)[\x{c0}])/8
2892 /(?i:[\x{c0}])/8
2893
2894 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2895 is now done differently.
2896
2897 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2898 wrapper classes; (b) implement a new function in the C++ scanner that is
2899 more efficient than the old way of doing things because it avoids levels of
2900 recursion in the regex matching; (c) add a paragraph to the documentation
2901 for the FullMatch() function.
2902
2903 6. The escape sequence \n was being treated as whatever was defined as
2904 "newline". Not only was this contrary to the documentation, which states
2905 that \n is character 10 (hex 0A), but it also went horribly wrong when
2906 "newline" was defined as CRLF. This has been fixed.
2907
2908 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2909 was being set to -1 for the "end of line" case (supposedly a value that no
2910 character can have). Though this value is never used (the check for end of
2911 line is "zero bytes in current character"), it caused compiler complaints.
2912 I've changed it to 0xffffffff.
2913
2914 8. In pcre_version.c, the version string was being built by a sequence of
2915 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2916 string (as it is for production releases) called a macro with an empty
2917 argument. The C standard says the result of this is undefined. The gcc
2918 compiler treats it as an empty string (which was what was wanted) but it is
2919 reported that Visual C gives an error. The source has been hacked around to
2920 avoid this problem.
2921
2922 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2923 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2924 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2925 of them did).
2926
2927 10. Originally, pcretest opened its input and output without "b"; then I was
2928 told that "b" was needed in some environments, so it was added for release
2929 5.0 to both the input and output. (It makes no difference on Unix-like
2930 systems.) Later I was told that it is wrong for the input on Windows. I've
2931 now abstracted the modes into two macros, to make it easier to fiddle with
2932 them, and removed "b" from the input mode under Windows.
2933
2934 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2935
2936 12. Added -help and --help to pcretest as an official way of being reminded
2937 of the options.
2938
2939 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2940 and pcrecpp.cc because they annoy compilers at high warning levels.
2941
2942 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2943
2944 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2945 scripts are not C programs :-) and which was not noticed because it works
2946 on Linux.
2947
2948 16. pcretest is supposed to handle any length of pattern and data line (as one
2949 line or as a continued sequence of lines) by extending its input buffer if
2950 necessary. This feature was broken for very long pattern lines, leading to
2951 a string of junk being passed to pcre_compile() if the pattern was longer
2952 than about 50K.
2953
2954 17. I have done a major re-factoring of the way pcre_compile() computes the
2955 amount of memory needed for a compiled pattern. Previously, there was code
2956 that made a preliminary scan of the pattern in order to do this. That was
2957 OK when PCRE was new, but as the facilities have expanded, it has become
2958 harder and harder to keep it in step with the real compile phase, and there
2959 have been a number of bugs (see for example, 4 above). I have now found a
2960 cunning way of running the real compile function in a "fake" mode that
2961 enables it to compute how much memory it would need, while actually only
2962 ever using a few hundred bytes of working memory and without too many
2963 tests of the mode. This should make future maintenance and development
2964 easier. A side effect of this work is that the limit of 200 on the nesting
2965 depth of parentheses has been removed (though this was never a serious
2966 limitation, I suspect). However, there is a downside: pcre_compile() now
2967 runs more slowly than before (30% or more, depending on the pattern). I
2968 hope this isn't a big issue. There is no effect on runtime performance.
2969
2970 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2971 newline (only possible for the last line of a file) and it was a
2972 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2973
2974 19. Added additional timing features to pcretest. (1) The -tm option now times
2975 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2976 separate command line item, by a number that specifies the number of
2977 repeats to use when timing. The default is 50000; this gives better
2978 precision, but takes uncomfortably long for very large patterns.
2979
2980 20. Extended pcre_study() to be more clever in cases where a branch of a
2981 subpattern has no definite first character. For example, (a*|b*)[cd] would
2982 previously give no result from pcre_study(). Now it recognizes that the
2983 first character must be a, b, c, or d.
2984
2985 21. There was an incorrect error "recursive call could loop indefinitely" if
2986 a subpattern (or the entire pattern) that was being tested for matching an
2987 empty string contained only one non-empty item after a nested subpattern.
2988 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2989 incorrectly, because the \d was being skipped in the check.
2990
2991 22. The pcretest program now has a new pattern option /B and a command line
2992 option -b, which is equivalent to adding /B to every pattern. This causes
2993 it to show the compiled bytecode, without the additional information that
2994 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2995 is the same as /B/I).
2996
2997 23. A new optimization is now able automatically to treat some sequences such
2998 as a*b as a*+b. More specifically, if something simple (such as a character
2999 or a simple class like \d) has an unlimited quantifier, and is followed by
3000 something that cannot possibly match the quantified thing, the quantifier
3001 is automatically "possessified".
3002
3003 24. A recursive reference to a subpattern whose number was greater than 39
3004 went wrong under certain circumstances in UTF-8 mode. This bug could also
3005 have affected the operation of pcre_study().
3006
3007 25. Realized that a little bit of performance could be had by replacing
3008 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3009
3010 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3011
3012 27. Possessive quantifiers such as a++ were previously implemented by turning
3013 them into atomic groups such as ($>a+). Now they have their own opcodes,
3014 which improves performance. This includes the automatically created ones
3015 from 23 above.
3016
3017 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3018 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3019 the first matched character to be a colon. This applied both to named and
3020 numbered groups.
3021
3022 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3023
3024 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3025 building PCRE on Windows easier, so I have included it in the distribution.
3026
3027 31. There is now a check in pcretest against a ridiculously large number being
3028 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3029 loop, the loop is abandoned.
3030
3031 32. Forward references to subpatterns in conditions such as (?(2)...) where
3032 subpattern 2 is defined later cause pcre_compile() to search forwards in
3033 the pattern for the relevant set of parentheses. This search went wrong
3034 when there were unescaped parentheses in a character class, parentheses
3035 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3036
3037 33. "Subroutine" calls and backreferences were previously restricted to
3038 referencing subpatterns earlier in the regex. This restriction has now
3039 been removed.
3040
3041 34. Added a number of extra features that are going to be in Perl 5.10. On the
3042 whole, these are just syntactic alternatives for features that PCRE had
3043 previously implemented using the Python syntax or my own invention. The
3044 other formats are all retained for compatibility.
3045
3046 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3047 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3048 also .NET compatible.
3049
3050 (b) A recursion or subroutine call to a named group can now be defined as
3051 (?&name) as well as (?P>name).
3052
3053 (c) A backreference to a named group can now be defined as \k<name> or
3054 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3055 5.10, are also .NET compatible.
3056
3057 (d) A conditional reference to a named group can now use the syntax
3058 (?(<name>) or (?('name') as well as (?(name).
3059
3060 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3061 groups (named and numbered) that are never evaluated inline, but can be
3062 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3063 is always false. There may be only one alternative in such a group.
3064
3065 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3066 as the simple (?(R). The condition is true only if the most recent
3067 recursion is that of the given number or name. It does not search out
3068 through the entire recursion stack.
3069
3070 (g) The escape \gN or \g{N} has been added, where N is a positive or
3071 negative number, specifying an absolute or relative reference.
3072
3073 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3074 some "unreachable code" warnings.
3075
3076 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3077 things, this adds five new scripts.
3078
3079 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3080 There were also incompatibilities regarding the handling of \Q..\E inside
3081 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3082 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3083
3084 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3085 matches an empty string, and forcibly breaks the loop. There were bugs in
3086 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3087 against aaaa the result was just "a" rather than "aaaa", for example. Two
3088 separate and independent bugs (that affected different cases) have been
3089 fixed.
3090
3091 39. Refactored the code to abolish the use of different opcodes for small
3092 capturing bracket numbers. This is a tidy that I avoided doing when I
3093 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3094 The new approach is not only tidier, it makes it possible to reduce the
3095 memory needed to fix the previous bug (38).
3096
3097 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3098 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3099 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3100 mode.
3101
3102 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3103 report.
3104
3105 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3106 copy construction and assignment in the C++ wrapper.
3107
3108 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3109 couple of bugs that could have given wrong results in the "--newline=crlf"
3110 case.
3111
3112 44. Added a number of casts and did some reorganization of signed/unsigned int
3113 variables following suggestions from Dair Grant. Also renamed the variable
3114 "this" as "item" because it is a C++ keyword.
3115
3116 45. Arranged for dftables to add
3117
3118 #include "pcre_internal.h"
3119
3120 to pcre_chartables.c because without it, gcc 4.x may remove the array
3121 definition from the final binary if PCRE is built into a static library and
3122 dead code stripping is activated.
3123
3124 46. For an unanchored pattern, if a match attempt fails at the start of a
3125 newline sequence, and the newline setting is CRLF or ANY, and the next two
3126 characters are CRLF, advance by two characters instead of one.
3127
3128
3129 Version 6.7 04-Jul-06
3130 ---------------------
3131
3132 1. In order to handle tests when input lines are enormously long, pcretest has
3133 been re-factored so that it automatically extends its buffers when
3134 necessary. The code is crude, but this _is_ just a test program. The
3135 default size has been increased from 32K to 50K.
3136
3137 2. The code in pcre_study() was using the value of the re argument before
3138 testing it for NULL. (Of course, in any sensible call of the function, it
3139 won't be NULL.)
3140
3141 3. The memmove() emulation function in pcre_internal.h, which is used on
3142 systems that lack both memmove() and bcopy() - that is, hardly ever -
3143 was missing a "static" storage class specifier.
3144
3145 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3146 containing an extended class (one that cannot be represented by a bitmap
3147 because it contains high-valued characters or Unicode property items, e.g.
3148 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3149 pattern, but PCRE should not loop if you do not (it no longer does).
3150 [Detail: two cases were found: (a) a repeated subpattern containing an
3151 extended class; (b) a recursive reference to a subpattern that followed a
3152 previous extended class. It wasn't skipping over the extended class
3153 correctly when UTF-8 mode was not set.]
3154
3155 5. A negated single-character class was not being recognized as fixed-length
3156 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3157 compile error "lookbehind assertion is not fixed length".
3158
3159 6. The RunPerlTest auxiliary script was showing an unexpected difference
3160 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3161 write a Perl script that can interpret lines of an input file either as
3162 byte characters or as UTF-8, which is what "perltest" was being required to
3163 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3164 can't do is switch easily at run time between having the "use utf8;" pragma
3165 or not. In the end, I fudged it by using the RunPerlTest script to insert
3166 "use utf8;" explicitly for the UTF-8 tests.
3167
3168 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3169 the end of the subject string, contrary to the documentation and to what
3170 Perl does. This was true of both matching functions. Now it matches only at
3171 the start of the subject and immediately after *internal* newlines.
3172
3173 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3174 a pointer to an int instead of a pointer to an unsigned long int. This
3175 caused problems on 64-bit systems.
3176
3177 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3178 instance of the 'standard' template library not being so standard".
3179
3180 10. There was no check on the number of named subpatterns nor the maximum
3181 length of a subpattern name. The product of these values is used to compute
3182 the size of the memory block for a compiled pattern. By supplying a very
3183 long subpattern name and a large number of named subpatterns, the size
3184 computation could be caused to overflow. This is now prevented by limiting
3185 the length of names to 32 characters, and the number of named subpatterns
3186 to 10,000.
3187
3188 11. Subpatterns that are repeated with specific counts have to be replicated in
3189 the compiled pattern. The size of memory for this was computed from the
3190 length of the subpattern and the repeat count. The latter is limited to
3191 65535, but there was no limit on the former, meaning that integer overflow
3192 could in principle occur. The compiled length of a repeated subpattern is
3193 now limited to 30,000 bytes in order to prevent this.
3194
3195 12. Added the optional facility to have named substrings with the same name.
3196
3197 13. Added the ability to use a named substring as a condition, using the
3198 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3199 are numbers (not recommended). Forward references are permitted.
3200
3201 14. Added forward references in named backreferences (if you see what I mean).
3202
3203 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3204 pattern could run off the end of the subject. For example, the pattern
3205 "(?s)(.{1,5})"8 did this with the subject "ab".
3206
3207 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3208 PCRE_CASELESS was set when matching characters that were quantified with ?
3209 or *.
3210
3211 17. A character class other than a single negated character that had a minimum
3212 but no maximum quantifier - for example [ab]{6,} - was not handled
3213 correctly by pce_dfa_exec(). It would match only one character.
3214
3215 18. A valid (though odd) pattern that looked like a POSIX character
3216 class but used an invalid character after [ (for example [[,abc,]]) caused
3217 pcre_compile() to give the error "Failed: internal error: code overflow" or
3218 in some cases to crash with a glibc free() error. This could even happen if
3219 the pattern terminated after [[ but there just happened to be a sequence of
3220 letters, a binary zero, and a closing ] in the memory that followed.
3221
3222 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3223 over the years. Originally (before any Unicode support), just the bottom 8
3224 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3225 output from "man perlunicode" includes this:
3226
3227 The regular expression compiler produces polymorphic opcodes. That
3228 is, the pattern adapts to the data and automatically switches to
3229 the Unicode character scheme when presented with Unicode data--or
3230 instead uses a traditional byte scheme when presented with byte
3231 data.
3232
3233 Sadly, a wide octal escape does not cause a switch, and in a string with
3234 no other multibyte characters, these octal escapes are treated as before.
3235 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3236 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3237 Unicode string.
3238
3239 I have not perpetrated such confusion in PCRE. Up till now, it took just
3240 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3241 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3242 translate to the appropriate multibyte character.
3243
3244 29. Applied some refactoring to reduce the number of warnings from Microsoft
3245 and Borland compilers. This has included removing the fudge introduced
3246 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3247 a warning about an unused variable.
3248
3249 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3250 characters since release 4.0, because Perl (from release 5.004) does not.
3251 [Or at least, is documented not to: some releases seem to be in conflict
3252 with the documentation.] However, when a pattern was studied with
3253 pcre_study() and all its branches started with \s, PCRE still included VT
3254 as a possible starting character. Of course, this did no harm; it just
3255 caused an unnecessary match attempt.
3256
3257 22. Removed a now-redundant internal flag bit that recorded the fact that case
3258 dependency changed within the pattern. This was once needed for "required
3259 byte" processing, but is no longer used. This recovers a now-scarce options
3260 bit. Also moved the least significant internal flag bit to the most-
3261 significant bit of the word, which was not previously used (hangover from
3262 the days when it was an int rather than a uint) to free up another bit for
3263 the future.
3264
3265 23. Added support for CRLF line endings as well as CR and LF. As well as the
3266 default being selectable at build time, it can now be changed at runtime
3267 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3268 specify that it is scanning data with non-default line endings.
3269
3270 24. Changed the definition of CXXLINK to make it agree with the definition of
3271 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3272
3273 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3274 recursions. This makes a big different to stack usage for some patterns.
3275
3276 26. If a subpattern containing a named recursion or subroutine reference such
3277 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3278 the space required for the compiled pattern went wrong and gave too small a
3279 value. Depending on the environment, this could lead to "Failed: internal
3280 error: code overflow at offset 49" or "glibc detected double free or
3281 corruption" errors.
3282
3283 27. Applied patches from Google (a) to support the new newline modes and (b) to
3284 advance over multibyte UTF-8 characters in GlobalReplace.
3285
3286 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3287 difference for some implementation of PCRE in some Windows version.
3288
3289 29. Added some extra testing facilities to pcretest:
3290
3291 \q<number> in a data line sets the "match limit" value
3292 \Q<number> in a data line sets the "match recursion limt" value
3293 -S <number> sets the stack size, where <number> is in megabytes
3294
3295 The -S option isn't available for Windows.
3296
3297
3298 Version 6.6 06-Feb-06
3299 ---------------------
3300
3301 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3302 in pcreposix.h. I have copied the definition from pcre.h.
3303
3304 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3305 because pcre.h is no longer a built file.
3306
3307 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3308 not normally included in the compiled code.
3309
3310
3311 Version 6.5 01-Feb-06
3312 ---------------------
3313
3314 1. When using the partial match feature with pcre_dfa_exec(), it was not
3315 anchoring the second and subsequent partial matches at the new starting
3316 point. This could lead to incorrect results. For example, with the pattern
3317 /1234/, partially matching against "123" and then "a4" gave a match.
3318
3319 2. Changes to pcregrep:
3320
3321 (a) All non-match returns from pcre_exec() were being treated as failures
3322 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3323 error message is output. Some extra information is given for the
3324 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3325 probably the only errors that are likely to be caused by users (by
3326 specifying a regex that has nested indefinite repeats, for instance).
3327 If there are more than 20 of these errors, pcregrep is abandoned.
3328
3329 (b) A binary zero was treated as data while matching, but terminated the
3330 output line if it was written out. This has been fixed: binary zeroes
3331 are now no different to any other data bytes.
3332
3333 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3334 used to set a locale for matching. The --locale=xxxx long option has
3335 been added (no short equivalent) to specify a locale explicitly on the
3336 pcregrep command, overriding the environment variables.
3337
3338 (d) When -B was used with -n, some line numbers in the output were one less
3339 than they should have been.
3340
3341 (e) Added the -o (--only-matching) option.
3342
3343 (f) If -A or -C was used with -c (count only), some lines of context were
3344 accidentally printed for the final match.
3345
3346 (g) Added the -H (--with-filename) option.
3347
3348 (h) The combination of options -rh failed to suppress file names for files
3349 that were found from directory arguments.
3350
3351 (i) Added the -D (--devices) and -d (--directories) options.
3352
3353 (j) Added the -F (--fixed-strings) option.
3354
3355 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3356
3357 (l) Added the --colo(u)r option.
3358
3359 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3360 is not present by default.
3361
3362 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3363 items such as (?R) or (?1), when the recursion could match a number of
3364 alternatives. If it matched one of the alternatives, but subsequently,
3365 outside the recursion, there was a failure, the code tried to back up into
3366 the recursion. However, because of the way PCRE is implemented, this is not
3367 possible, and the result was an incorrect result from the match.
3368
3369 In order to prevent this happening, the specification of recursion has
3370 been changed so that all such subpatterns are automatically treated as
3371 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3372
3373 4. I had overlooked the fact that, in some locales, there are characters for
3374 which isalpha() is true but neither isupper() nor islower() are true. In
3375 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3376 and ordfeminine) are like this. This affected the treatment of \w and \W
3377 when they appeared in character classes, but not when they appeared outside
3378 a character class. The bit map for "word" characters is now created
3379 separately from the results of isalnum() instead of just taking it from the
3380 upper, lower, and digit maps. (Plus the underscore character, of course.)
3381
3382 5. The above bug also affected the handling of POSIX character classes such as
3383 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3384 permanent tables. Instead, the bit maps for such a class were previously
3385 created as the appropriate unions of the upper, lower, and digit bitmaps.
3386 Now they are created by subtraction from the [[:word:]] class, which has
3387 its own bitmap.
3388
3389 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3390 It is created by subtracting the vertical space characters (\x09, \x0a,
3391 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3392 subtraction was done in the overall bitmap for a character class, meaning
3393 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3394 be recognized. This bug has been fixed.
3395
3396 7. Patches from the folks at Google:
3397
3398 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3399 real life, but is still worth protecting against".
3400
3401 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3402 regular expressions".
3403
3404 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3405 have it.
3406
3407 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3408 "configure" and the latter not, in order to fix a problem somebody had
3409 with compiling the Arg class on HP-UX.
3410
3411 (e) Improve the error-handling of the C++ wrapper a little bit.
3412
3413 (f) New tests for checking recursion limiting.
3414
3415 8. The pcre_memmove() function, which is used only if the environment does not
3416 have a standard memmove() function (and is therefore rarely compiled),
3417 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3418 returning a result (though PCRE never actually uses the result).
3419
3420 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3421 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3422 returned instead of calling malloc() with an overflowing number that would
3423 most likely cause subsequent chaos.
3424
3425 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3426
3427 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3428 with this option is matched, the nmatch and pmatch options of regexec() are
3429 ignored.
3430
3431 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3432 provided in case anyone wants to the the POSIX interface with UTF-8
3433 strings.
3434
3435 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3436 C++ linking (needed for some HP-UX environments).
3437
3438 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3439 (unused parameter) and in the pcre_printint() function (omitted "default"
3440 switch label when the default is to do nothing).
3441
3442 15. Added some code to make it possible, when PCRE is compiled as a C++
3443 library, to replace subject pointers for pcre_exec() with a smart pointer
3444 class, thus making it possible to process discontinuous strings.
3445
3446 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3447 much the same function. They were added by different people who were trying
3448 to make PCRE easy to compile on non-Unix systems. It has been suggested
3449 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3450 for compiling on Windows systems. I have therefore replaced it with
3451 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3452 defaults to "extern" for C or "extern C" for C++, which works fine on
3453 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3454 SCOPE with something explicit in config.h. In addition:
3455
3456 (a) pcreposix.h still had just "extern" instead of either of these macros;
3457 I have replaced it with PCRE_DATA_SCOPE.
3458
3459 (b) Functions such as _pcre_xclass(), which are internal to the library,
3460 but external in the C sense, all had PCRE_EXPORT in their definitions.
3461 This is apparently wrong for the Windows case, so I have removed it.
3462 (It makes no difference on Unix-like systems.)
3463
3464 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3465 of recursive calls to match(). This is different to MATCH_LIMIT because
3466 that limits the total number of calls to match(), not all of which increase
3467 the depth of recursion. Limiting the recursion depth limits the amount of
3468 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3469 when PCRE is compiled, and changed at run time. A patch from Google adds
3470 this functionality to the C++ interface.
3471
3472 18. Changes to the handling of Unicode character properties:
3473
3474 (a) Updated the table to Unicode 4.1.0.
3475
3476 (b) Recognize characters that are not in the table as "Cn" (undefined).
3477
3478 (c) I revised the way the table is implemented to a much improved format
3479 which includes recognition of ranges. It now supports the ranges that
3480 are defined in UnicodeData.txt, and it also amalgamates other
3481 characters into ranges. This has reduced the number of entries in the
3482 table from around 16,000 to around 3,000, thus reducing its size
3483 considerably. I realized I did not need to use a tree structure after
3484 all - a binary chop search is just as efficient. Having reduced the
3485 number of entries, I extended their size from 6 bytes to 8 bytes to
3486 allow for more data.
3487
3488 (d) Added support for Unicode script names via properties such as \p{Han}.
3489
3490 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3491 matching that character.
3492
3493 20. When matching a repeated Unicode property with a minimum greater than zero,
3494 (for example \pL{2,}), PCRE could look past the end of the subject if it
3495 reached it while seeking the minimum number of characters. This could
3496 happen only if some of the characters were more than one byte long, because
3497 there is a check for at least the minimum number of bytes.
3498
3499 21. Refactored the implementation of \p and \P so as to be more general, to
3500 allow for more different types of property in future. This has changed the
3501 compiled form incompatibly. Anybody with saved compiled patterns that use
3502 \p or \P will have to recompile them.
3503
3504 22. Added "Any" and "L&" to the supported property types.
3505
3506 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3507 but give a compile time error if the value is greater than 0xff.
3508
3509 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3510 accidentally not being installed or uninstalled.
3511
3512 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3513 made were to insert the current release number. This seemed silly, because
3514 it made things harder for people building PCRE on systems that don't run
3515 "configure". I have turned pcre.h into a distributed file, no longer built
3516 by "configure", with the version identification directly included. There is
3517 no longer a pcre.h.in file.
3518
3519 However, this change necessitated a change to the pcre-config script as
3520 well. It is built from pcre-config.in, and one of the substitutions was the
3521 release number. I have updated configure.ac so that ./configure now finds
3522 the release number by grepping pcre.h.
3523
3524 26. Added the ability to run the tests under valgrind.
3525
3526
3527 Version 6.4 05-Sep-05
3528 ---------------------
3529
3530 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3531 "--" to be printed when multiple files were scanned, even when none of the
3532 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3533 consider it to be a bug, and have restored the previous behaviour.
3534
3535 2. A couple of code tidies to get rid of compiler warnings.
3536
3537 3. The pcretest program used to cheat by referring to symbols in the library
3538 whose names begin with _pcre_. These are internal symbols that are not
3539 really supposed to be visible externally, and in some environments it is
3540 possible to suppress them. The cheating is now confined to including
3541 certain files from the library's source, which is a bit cleaner.
3542
3543 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3544 file's purpose clearer.
3545
3546 5. Reorganized pcre_ucp_findchar().
3547
3548
3549 Version 6.3 15-Aug-05
3550 ---------------------
3551
3552 1. The file libpcre.pc.in did not have general read permission in the tarball.
3553
3554 2. There were some problems when building without C++ support:
3555
3556 (a) If C++ support was not built, "make install" and "make test" still
3557 tried to test it.
3558
3559 (b) There were problems when the value of CXX was explicitly set. Some
3560 changes have been made to try to fix these, and ...
3561
3562 (c) --disable-cpp can now be used to explicitly disable C++ support.
3563
3564 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3565 backslash in a target when C++ was disabled. This confuses some
3566 versions of "make", apparently. Using an intermediate variable solves
3567 this. (Same for CPP_LOBJ.)
3568
3569 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3570 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3571 necessary on certain architectures.
3572
3573 4. Added a setting of -export-symbols-regex to the link command to remove
3574 those symbols that are exported in the C sense, but actually are local
3575 within the library, and not documented. Their names all begin with
3576 "_pcre_". This is not a perfect job, because (a) we have to except some
3577 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3578 available (and never for static libraries). I have made a note to try to
3579 find a way round (a) in the future.
3580
3581
3582 Version 6.2 01-Aug-05
3583 ---------------------
3584
3585 1. There was no test for integer overflow of quantifier values. A construction
3586 such as {1111111111111111} would give undefined results. What is worse, if
3587 a minimum quantifier for a parenthesized subpattern overflowed and became
3588 negative, the calculation of the memory size went wrong. This could have
3589 led to memory overwriting.
3590
3591 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3592
3593 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3594 operating environments where this matters.
3595
3596 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3597 PCRE options from within the C++ wrapper.
3598
3599 5. Named capturing subpatterns were not being correctly counted when a pattern
3600 was compiled. This caused two problems: (a) If there were more than 100
3601 such subpatterns, the calculation of the memory needed for the whole
3602 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3603 back references of the form \12, where the number was greater than 9, were
3604 not recognized as back references, even though there were sufficient
3605 previous subpatterns.
3606
3607 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3608 versions of gcc, e.g. 2.95.4.
3609
3610
3611 Version 6.1 21-Jun-05
3612 ---------------------
3613
3614 1. There was one reference to the variable "posix" in pcretest.c that was not
3615 surrounded by "#if !defined NOPOSIX".
3616
3617 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3618 the cross-check on the old pcre_info() function, for the benefit of the
3619 cut-down version of PCRE that is currently imported into Exim.
3620
3621 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3622 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3623 patterns that start (?i) but I don't think that matters. The use of (?i) is
3624 just an example; this all applies to the other options as well.
3625
3626 4. Since libtool seems to echo the compile commands it is issuing, the output
3627 from "make" can be reduced a bit by putting "@" in front of each libtool
3628 compile command.
3629
3630 5. Patch from the folks at Google for configure.in to be a bit more thorough
3631 in checking for a suitable C++ installation before trying to compile the
3632 C++ stuff. This should fix a reported problem when a compiler was present,
3633 but no suitable headers.
3634
3635 6. The man pages all had just "PCRE" as their title. I have changed them to
3636 be the relevant file name. I have also arranged that these names are
3637 retained in the file doc/pcre.txt, which is a concatenation in text format
3638 of all the man pages except the little individual ones for each function.
3639
3640 7. The NON-UNIX-USE file had not been updated for the different set of source
3641 files that come with release 6. I also added a few comments about the C++
3642 wrapper.
3643
3644
3645 Version 6.0 07-Jun-05
3646 ---------------------
3647
3648 1. Some minor internal re-organization to help with my DFA experiments.
3649
3650 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3651 didn't matter for the library itself when fully configured, but did matter
3652 when compiling without UCP support, or within Exim, where the ucp files are
3653 not imported.
3654
3655 3. Refactoring of the library code to split up the various functions into
3656 different source modules. The addition of the new DFA matching code (see
3657 below) to a single monolithic source would have made it really too
3658 unwieldy, quite apart from causing all the code to be include in a
3659 statically linked application, when only some functions are used. This is
3660 relevant even without the DFA addition now that patterns can be compiled in
3661 one application and matched in another.
3662
3663 The downside of splitting up is that there have to be some external
3664 functions and data tables that are used internally in different modules of
3665 the library but which are not part of the API. These have all had their
3666 names changed to start with "_pcre_" so that they are unlikely to clash
3667 with other external names.
3668
3669 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3670 a different (DFA) algorithm. Although it is slower than the original
3671 function, it does have some advantages for certain types of matching
3672 problem.
3673
3674 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3675 including restarting after a partial match.
3676
3677 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3678 defined when compiling for Windows was sent to me. I have put it into the
3679 code, though I have no means of testing or verifying it.
3680
3681 7. Added the pcre_refcount() auxiliary function.
3682
3683 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3684 match before or at the first newline in the subject string. In pcretest,
3685 the /f option on a pattern can be used to set this.
3686
3687 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3688 would behave wrongly. This has been present in PCRE since release 4.0.
3689
3690 10. A number of changes to the pcregrep command:
3691
3692 (a) Refactored how -x works; insert ^(...)$ instead of setting
3693 PCRE_ANCHORED and checking the length, in preparation for adding
3694 something similar for -w.
3695
3696 (b) Added the -w (match as a word) option.
3697
3698 (c) Refactored the way lines are read and buffered so as to have more
3699 than one at a time available.
3700
3701 (d) Implemented a pcregrep test script.
3702
3703 (e) Added the -M (multiline match) option. This allows patterns to match
3704 over several lines of the subject. The buffering ensures that at least
3705 8K, or the rest of the document (whichever is the shorter) is available
3706 for matching (and similarly the previous 8K for lookbehind assertions).
3707
3708 (f) Changed the --help output so that it now says
3709
3710 -w, --word-regex(p)
3711
3712 instead of two lines, one with "regex" and the other with "regexp"
3713 because that confused at least one person since the short forms are the
3714 same. (This required a bit of code, as the output is generated
3715 automatically from a table. It wasn't just a text change.)
3716
3717 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3718 option but starts with a hyphen. Could be a pattern or a path name
3719 starting with a hyphen, for instance.
3720
3721 (h) "-" can be given as a file name to represent stdin.
3722
3723 (i) When file names are being printed, "(standard input)" is used for
3724 the standard input, for compatibility with GNU grep. Previously
3725 "<stdin>" was used.
3726
3727 (j) The option --label=xxx can be used to supply a name to be used for
3728 stdin when file names are being printed. There is no short form.
3729
3730 (k) Re-factored the options decoding logic because we are going to add
3731 two more options that take data. Such options can now be given in four
3732 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3733
3734 (l) Added the -A, -B, and -C options for requesting that lines of context
3735 around matches be printed.
3736
3737 (m) Added the -L option to print the names of files that do not contain
3738 any matching lines, that is, the complement of -l.
3739
3740 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3741 continue to scan other files.
3742
3743 (o) The -s option was incorrectly implemented. For compatibility with other
3744 greps, it now suppresses the error message for a non-existent or non-
3745 accessible file (but not the return code). There is a new option called
3746 -q that suppresses the output of matching lines, which was what -s was
3747 previously doing.
3748
3749 (p) Added --include and --exclude options to specify files for inclusion
3750 and exclusion when recursing.
3751
3752 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3753 Hopefully, it now does.
3754
3755 12. Missing cast in pcre_study().
3756
3757 13. Added an "uninstall" target to the makefile.
3758
3759 14. Replaced "extern" in the function prototypes in Makefile.in with
3760 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3761 world, but is set differently for Windows.
3762
3763 15. Added a second compiling function called pcre_compile2(). The only
3764 difference is that it has an extra argument, which is a pointer to an
3765 integer error code. When there is a compile-time failure, this is set
3766 non-zero, in addition to the error test pointer being set to point to an
3767 error message. The new argument may be NULL if no error number is required
3768 (but then you may as well call pcre_compile(), which is now just a
3769 wrapper). This facility is provided because some applications need a
3770 numeric error indication, but it has also enabled me to tidy up the way
3771 compile-time errors are handled in the POSIX wrapper.
3772
3773 16. Added VPATH=.libs to the makefile; this should help when building with one
3774 prefix path and installing with another. (Or so I'm told by someone who
3775 knows more about this stuff than I do.)
3776
3777 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3778 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3779 match everything, including newlines. This is not POSIX-compatible, but
3780 somebody wanted the feature. From pcretest it can be activated by using
3781 both the P and the s flags.
3782
3783 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3784
3785 19. libpcre.pc was being incorrectly installed as executable.
3786
3787 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3788 it now also looks for '\r' so that it will work unmodified on Windows.
3789
3790 21. Added Google's contributed C++ wrapper to the distribution.
3791
3792 22. Added some untidy missing memory free() calls in pcretest, to keep
3793 Electric Fence happy when testing.
3794
3795
3796
3797 Version 5.0 13-Sep-04
3798 ---------------------
3799
3800 1. Internal change: literal characters are no longer packed up into items
3801 containing multiple characters in a single byte-string. Each character
3802 is now matched using a separate opcode. However, there may be more than one
3803 byte in the character in UTF-8 mode.
3804
3805 2. The pcre_callout_block structure has two new fields: pattern_position and
3806 next_item_length. These contain the offset in the pattern to the next match
3807 item, and its length, respectively.
3808
3809 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3810 insertion of callouts before each pattern item. Added the /C option to
3811 pcretest to make use of this.
3812
3813 4. On the advice of a Windows user, the lines
3814
3815 #if defined(_WIN32) || defined(WIN32)
3816 _setmode( _fileno( stdout ), 0x8000 );
3817 #endif /* defined(_WIN32) || defined(WIN32) */
3818
3819 have been added to the source of pcretest. This apparently does useful
3820 magic in relation to line terminators.
3821
3822 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3823 for the benefit of those environments where the "b" makes a difference.
3824
3825 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3826 to know about it. I have put a hack into configure.in that adds in code
3827 to set GCC=yes if CC=icc. This seems to end up at a point in the
3828 generated configure script that is early enough to affect the setting of
3829 compiler options, which is what is needed, but I have no means of testing
3830 whether it really works. (The user who reported this had patched the
3831 generated configure script, which of course I cannot do.)
3832
3833 LATER: After change 22 below (new libtool files), the configure script
3834 seems to know about icc (and also ecc). Therefore, I have commented out
3835 this hack in configure.in.
3836
3837 7. Added support for pkg-config (2 patches were sent in).
3838
3839 8. Negated POSIX character classes that used a combination of internal tables
3840 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3841 [[:^ascii]]. Typically, they would match almost any characters. The other
3842 POSIX classes were not broken in this way.
3843
3844 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3845 to find the match, as PCRE was deluded into thinking that the match had to
3846 start at the start point or following a newline. The same bug applied to
3847 patterns with negative forward assertions or any backward assertions
3848 preceding ".*" at the start, unless the pattern required a fixed first
3849 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3850
3851 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3852 starting at the last subject character, bytes beyond the end of the subject
3853 string were read.
3854
3855 11. Renamed the variable "class" as "classbits" to make life easier for C++
3856 users. (Previously there was a macro definition, but it apparently wasn't
3857 enough.)
3858
3859 12. Added the new field "tables" to the extra data so that tables can be passed
3860 in at exec time, or the internal tables can be re-selected. This allows
3861 a compiled regex to be saved and re-used at a later time by a different
3862 program that might have everything at different addresses.
3863
3864 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3865 -R library as well as a -L library.
3866
3867 14. The debugging options of pcretest (-d on the command line or D on a
3868 pattern) showed incorrect output for anything following an extended class
3869 that contained multibyte characters and which was followed by a quantifier.
3870
3871 15. Added optional support for general category Unicode character properties
3872 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3873 support. It adds about 90K to the size of the library. The meanings of the
3874 inbuilt class escapes such as \d and \s have NOT been changed.
3875
3876 16. Updated pcredemo.c to include calls to free() to release the memory for the
3877 compiled pattern.
3878
3879 17. The generated file chartables.c was being created in the source directory
3880 instead of in the building directory. This caused the build to fail if the
3881 source directory was different from the building directory, and was
3882 read-only.
3883
3884 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3885 file. No doubt somebody will tell me if they don't make sense... Also added
3886 Dan Mooney's comments about building on OpenVMS.
3887
3888 19. Added support for partial matching via the PCRE_PARTIAL option for
3889 pcre_exec() and the \P data escape in pcretest.
3890
3891 20. Extended pcretest with 3 new pattern features:
3892
3893 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3894 write the compiled pattern to the file whose name is "rest-of-line".
3895 This is a straight binary dump of the data, with the saved pointer to
3896 the character tables forced to be NULL. The study data, if any, is
3897 written too. After writing, pcretest reads a new pattern.
3898
3899 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3900 compiled pattern from the given file. There must not be any
3901 occurrences of "<" in the file name (pretty unlikely); if there are,
3902 pcretest will instead treat the initial "<" as a pattern delimiter.
3903 After reading in the pattern, pcretest goes on to read data lines as
3904 usual.
3905
3906 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3907 and 16-bit fields in a compiled pattern, to simulate a pattern that
3908 was compiled on a host of opposite endianness.
3909
3910 21. The pcre-exec() function can now cope with patterns that were compiled on
3911 hosts of opposite endianness, with this restriction:
3912
3913 As for any compiled expression that is saved and used later, the tables
3914 pointer field cannot be preserved; the extra_data field in the arguments
3915 to pcre_exec() should be used to pass in a tables address if a value
3916 other than the default internal tables were used at compile time.
3917
3918 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3919 now diagnosed as an error. Previously, most of the time, a negative number
3920 would have been treated as zero, but if in addition "ovector" was passed as
3921 NULL, a crash could occur.
3922
3923 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3924 new versions from the libtool 1.5 distribution (the last one is a copy of
3925 a file called libtool.m4). This seems to have fixed the need to patch
3926 "configure" to support Darwin 1.3 (which I used to do). However, I still
3927 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3928 workstation).
3929
3930 24. Changed the PCRE licence to be the more standard "BSD" licence.
3931
3932
3933 Version 4.5 01-Dec-03
3934 ---------------------
3935
3936 1. There has been some re-arrangement of the code for the match() function so
3937 that it can be compiled in a version that does not call itself recursively.
3938 Instead, it keeps those local variables that need separate instances for
3939 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3940 needs to "recurse". Keeping track of where control must go is done by means
3941 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3942 hide most of the details from the main code, and operates only if
3943 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3944 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3945 operating.
3946
3947 To make it easier for callers to provide specially tailored get/free
3948 functions for this usage, two new functions, pcre_stack_malloc, and
3949 pcre_stack_free, are used. They are always called in strict stacking order,
3950 and the size of block requested is always the same.
3951
3952 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3953 PCRE has been compiled to use the stack or the heap for recursion. The
3954 -C option of pcretest uses this to show which version is compiled.
3955
3956 A new data escape \S, is added to pcretest; it causes the amounts of store
3957 obtained and freed by both kinds of malloc/free at match time to be added
3958 to the output.
3959
3960 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3961 what's available on my current Linux desktop machine.
3962
3963 3. When matching a UTF-8 string, the test for a valid string at the start has
3964 been extended. If start_offset is not zero, PCRE now checks that it points
3965 to a byte that is the start of a UTF-8 character. If not, it returns
3966 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3967 this is necessary because there may be backward assertions in the pattern.
3968 When matching the same subject several times, it may save resources to use
3969 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3970
3971 4. The code for checking the validity of UTF-8 strings has been tightened so
3972 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3973 containing "overlong sequences".
3974
3975 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3976 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3977 should have been "&", but it just so happened that all the cases this let
3978 through by mistake were picked up later in the function.
3979
3980 6. I had used a variable called "isblank" - this is a C99 function, causing
3981 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3982
3983 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3984 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3985 so the version is shown; (c) stop "make test" echoing "./RunTest".
3986
3987 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3988
3989 9. The prototype for memmove() for systems that don't have it was using
3990 size_t, but the inclusion of the header that defines size_t was later. I've
3991 moved the #includes for the C headers earlier to avoid this.
3992
3993 10. Added some adjustments to the code to make it easier to compiler on certain
3994 special systems:
3995
3996 (a) Some "const" qualifiers were missing.
3997 (b) Added the macro EXPORT before all exported functions; by default this
3998 is defined to be empty.
3999 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4000 that it reads its output file name as an argument instead of writing
4001 to the standard output and assuming this can be redirected.
4002
4003 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4004 class containing characters with values greater than 255, PCRE compilation
4005 went into a loop.
4006
4007 12. A recursive reference to a subpattern that was within another subpattern
4008 that had a minimum quantifier of zero caused PCRE to crash. For example,
4009 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4010 recursion. If the recursively-called subpattern itself had a zero repeat,
4011 that was OK.
4012
4013 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4014 buffer into which it was copied (for escape processing) was still set at
4015 1024, so long lines caused crashes.
4016
4017 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4018 "internal error: code overflow...". This applied to any character class
4019 that was followed by a possessive quantifier.
4020
4021 15. Modified the Makefile to add libpcre.la as a prerequisite for
4022 libpcreposix.la because I was told this is needed for a parallel build to
4023 work.
4024
4025 16. If a pattern that contained .* following optional items at the start was
4026 studied, the wrong optimizing data was generated, leading to matching
4027 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4028 matching string must start with a or b or c. The correct conclusion for
4029 this pattern is that a match can start with any character.
4030
4031
4032 Version 4.4 13-Aug-03
4033 ---------------------
4034
4035 1. In UTF-8 mode, a character class containing characters with values between
4036 127 and 255 was not handled correctly if the compiled pattern was studied.
4037 In fixing this, I have also improved the studying algorithm for such
4038 classes (slightly).
4039
4040 2. Three internal functions had redundant arguments passed to them. Removal
4041 might give a very teeny performance improvement.
4042
4043 3. Documentation bug: the value of the capture_top field in a callout is *one
4044 more than* the number of the hightest numbered captured substring.
4045
4046 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4047 in incorrectly linking with a previously installed version. They now link
4048 explicitly with libpcre.la.
4049
4050 5. configure.in no longer needs to recognize Cygwin specially.
4051
4052 6. A problem in pcre.in for Windows platforms is fixed.
4053
4054 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4055 pcretest, it used to include the size of the study block as part of its
4056 output. Unfortunately, the structure contains a field that has a different
4057 size on different hardware architectures. This meant that the tests that
4058 showed this size failed. As the block is currently always of a fixed size,
4059 this information isn't actually particularly useful in pcretest output, so
4060 I have just removed it.
4061
4062 8. Three pre-processor statements accidentally did not start in column 1.
4063 Sadly, there are *still* compilers around that complain, even though
4064 standard C has not required this for well over a decade. Sigh.
4065
4066 9. In pcretest, the code for checking callouts passed small integers in the
4067 callout_data field, which is a void * field. However, some picky compilers
4068 complained about the casts involved for this on 64-bit systems. Now
4069 pcretest passes the address of the small integer instead, which should get
4070 rid of the warnings.
4071
4072 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4073 both compile and run time, and gives an error if an invalid UTF-8 sequence
4074 is found. There is a option for disabling this check in cases where the
4075 string is known to be correct and/or the maximum performance is wanted.
4076
4077 11. In response to a bug report, I changed one line in Makefile.in from
4078
4079 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4080 to
4081 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4082
4083 to look similar to other lines, but I have no way of telling whether this
4084 is the right thing to do, as I do not use Windows. No doubt I'll get told
4085 if it's wrong...
4086
4087
4088 Version 4.3 21-May-03
4089 ---------------------
4090
4091 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4092 Makefile.
4093
4094 2. Some refactoring to improve the quality of the code:
4095
4096 (i) The utf8_table... variables are now declared "const".
4097
4098 (ii) The code for \cx, which used the "case flipping" table to upper case
4099 lower case letters, now just substracts 32. This is ASCII-specific,
4100 but the whole concept of \cx is ASCII-specific, so it seems
4101 reasonable.
4102
4103 (iii) PCRE was using its character types table to recognize decimal and
4104 hexadecimal digits in the pattern. This is silly, because it handles
4105 only 0-9, a-f, and A-F, but the character types table is locale-
4106 specific, which means strange things might happen. A private
4107 table is now used for this - though it costs 256 bytes, a table is
4108 much faster than multiple explicit tests. Of course, the standard
4109 character types table is still used for matching digits in subject
4110 strings against \d.
4111
4112 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4113 ending in _t are). So I've renamed it as ESC_tee.
4114
4115 3. The first argument for regexec() in the POSIX wrapper should have been
4116 defined as "const".
4117
4118 4. Changed pcretest to use malloc() for its buffers so that they can be
4119 Electric Fenced for debugging.
4120
4121 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4122 to read one or more bytes before the start of the subject string. Often this
4123 had no effect on PCRE's behaviour, but in some circumstances it could
4124 provoke a segmentation fault.
4125
4126 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4127 to try to read one or more bytes before the start of the subject string.
4128
4129 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4130 UTF-8 support could misbehave in various ways if the subject string
4131 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4132 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4133 back over UTF-8 characters.)
4134
4135
4136 Version 4.2 14-Apr-03
4137 ---------------------
4138
4139 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4140
4141 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4142 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4143 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4144 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4145 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4146 and BUILD_EXEEXT
4147 Note: automatic setting of the BUILD variables is not yet working
4148 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4149 compile-time but not at link-time
4150 [LINK]: use for linking executables only
4151 make different versions for Windows and non-Windows
4152 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4153 libraries
4154 [LINK_FOR_BUILD]: new variable
4155 [OBJEXT]: use throughout
4156 [EXEEXT]: use throughout
4157 <winshared>: new target
4158 <wininstall>: new target
4159 <dftables.o>: use native compiler
4160 <dftables>: use native linker
4161 <install>: handle Windows platform correctly
4162 <clean>: ditto
4163 <check>: ditto
4164 copy DLL to top builddir before testing
4165
4166 As part of these changes, -no-undefined was removed again. This was reported
4167 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4168 in any case.
4169
4170 3. Some tidies to get rid of compiler warnings:
4171
4172 . In the match_data structure, match_limit was an unsigned long int, whereas
4173 match_call_count was an int. I've made them both unsigned long ints.
4174
4175 . In pcretest the fact that a const uschar * doesn't automatically cast to
4176 a void * provoked a warning.
4177
4178 . Turning on some more compiler warnings threw up some "shadow" variables
4179 and a few more missing casts.
4180
4181 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4182 option, a class that contained a single character with a value between 128
4183 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4184
4185 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4186 option, a class that contained several characters, but with at least one
4187 whose value was between 128 and 255 caused PCRE to crash.
4188
4189
4190 Version 4.1 12-Mar-03
4191 ---------------------
4192
4193 1. Compiling with gcc -pedantic found a couple of places where casts were
4194 needed, and a string in dftables.c that was longer than standard compilers are
4195 required to support.
4196
4197 2. Compiling with Sun's compiler found a few more places where the code could
4198 be tidied up in order to avoid warnings.
4199
4200 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4201 first of these names is deprecated in the latest Autoconf in favour of the name
4202 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4203 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4204 analogy I have changed it to CFLAGS_FOR_BUILD.
4205
4206 4. Added -no-undefined to the linking command in the Makefile, because this is
4207 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4208 linking step for the pcreposix library.
4209
4210 5. PCRE was failing to diagnose the case of two named groups with the same
4211 name.
4212
4213 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4214 literal character that is needed in the subject for a match, and scans along to
4215 ensure that it is present before embarking on the full matching process. This
4216 saves time in cases of nested unlimited repeats that are never going to match.
4217 Problem: the scan can take a lot of time if the subject is very long (e.g.
4218 megabytes), thus penalizing straightforward matches. It is now done only if the
4219 amount of subject to be scanned is less than 1000 bytes.
4220
4221 7. A lesser problem with the same optimization is that it was recording the
4222 first character of an anchored pattern as "needed", thus provoking a search
4223 right along the subject, even when the first match of the pattern was going to
4224 fail. The "needed" character is now not set for anchored patterns, unless it
4225 follows something in the pattern that is of non-fixed length. Thus, it still
4226 fulfils its original purpose of finding quick non-matches in cases of nested
4227 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4228
4229
4230 Version 4.0 17-Feb-03
4231 ---------------------
4232
4233 1. If a comment in an extended regex that started immediately after a meta-item
4234 extended to the end of string, PCRE compiled incorrect data. This could lead to
4235 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4236
4237 2. Moved to autoconf 2.53 and libtool 1.4.2.
4238
4239 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4240 the special perltest8 script is no longer needed - all the tests can be run
4241 from a single perltest script.
4242
4243 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4244 by \s. It has now been removed in PCRE. This means it isn't recognized as
4245 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4246 class [:space:] *does* include VT, thereby creating a mess.
4247
4248 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4249 space and tab.
4250
4251 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4252 its new features into the main test script, reducing the number of scripts.
4253
4254 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4255 were backward compatible, and made the (?i) apply to the whole pattern, as if
4256 /i were given. Now it behaves more logically, and applies the option setting
4257 only to what follows. PCRE has been changed to follow suit. However, if it
4258 finds options settings right at the start of the pattern, it extracts them into
4259 the global options, as before. Thus, they show up in the info data.
4260
4261 8. Added support for the \Q...\E escape sequence. Characters in between are
4262 treated as literals. This is slightly different from Perl in that $ and @ are
4263 also handled as literals inside the quotes. In Perl, they will cause variable
4264 interpolation. Note the following examples:
4265
4266 Pattern PCRE matches Perl matches
4267
4268 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4269 \Qabc\$xyz\E abc\$xyz abc\$xyz
4270 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4271
4272 For compatibility with Perl, \Q...\E sequences are recognized inside character
4273 classes as well as outside them.
4274
4275 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4276 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4277 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4278 signed/unsigned warnings.
4279
4280 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4281 option for pcretest, so I've replaced it by a simple function that does just
4282 that job.
4283
4284 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4285 "pcregrep -".
4286
4287 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4288 Java package. This provides some syntactic sugar for simple cases of what my
4289 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4290 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4291 item, you can use this simplified notation. Note that only makes sense with
4292 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4293 greediness, whatever the setting of the PCRE_UNGREEDY option.
4294
4295 13. A change of greediness default within a pattern was not taking effect at
4296 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4297 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4298 was abstracted outside.
4299
4300 14. PCRE now supports the \G assertion. It is true when the current matching
4301 position is at the start point of the match. This differs from \A when the
4302 starting offset is non-zero. Used with the /g option of pcretest (or similar
4303 code), it works in the same way as it does for Perl's /g option. If all
4304 alternatives of a regex begin with \G, the expression is anchored to the start
4305 match position, and the "anchored" flag is set in the compiled expression.
4306
4307 15. Some bugs concerning the handling of certain option changes within patterns
4308 have been fixed. These applied to options other than (?ims). For example,
4309 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4310 been the other way round. Some of this was related to change 7 above.
4311
4312 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4313 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4314 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4315 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4316
4317 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4318 mode. Unlike ".", it always matches newline, whatever the setting of
4319 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4320 assertions. Perl allows it, but it doesn't (in general) work because it can't
4321 calculate the length of the lookbehind. At least, that's the case for Perl
4322 5.8.0 - I've been told they are going to document that it doesn't work in
4323 future.
4324
4325 18. Added an error diagnosis for escapes that PCRE does not support: these are
4326 \L, \l, \N, \P, \p, \U, \u, and \X.
4327
4328 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4329 reading past the end of the pattern in cases such as /[abcd/.
4330
4331 20. PCRE was getting more memory than necessary for patterns with classes that
4332 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4333
4334 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4335 compiling PCRE for use with Virtual Pascal.
4336
4337 22. Small fix to the Makefile to make it work properly if the build is done
4338 outside the source tree.
4339
4340 23. Added a new extension: a condition to go with recursion. If a conditional
4341 subpattern starts with (?(R) the "true" branch is used if recursion has
4342 happened, whereas the "false" branch is used only at the top level.
4343
4344 24. When there was a very long string of literal characters (over 255 bytes
4345 without UTF support, over 250 bytes with UTF support), the computation of how
4346 much memory was required could be incorrect, leading to segfaults or other
4347 strange effects.
4348
4349 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4350 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4351 there was a subsequent back reference to those brackets. This meant that, for
4352 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4353 possible to check for precisely this case. All we can do is abandon the
4354 optimization if .* occurs inside capturing brackets when there are any back
4355 references whatsoever. (See below for a better fix that came later.)
4356
4357 26. The handling of the optimization for finding the first character of a
4358 non-anchored pattern, and for finding a character that is required later in the
4359 match were failing in some cases. This didn't break the matching; it just
4360 failed to optimize when it could. The way this is done has been re-implemented.
4361
4362 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4363
4364 28. Added a new feature that provides some of the functionality that Perl
4365 provides with (?{...}). The facility is termed a "callout". The way it is done
4366 in PCRE is for the caller to provide an optional function, by setting
4367 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4368 global variable. By default it is unset, which disables all calling out. To get
4369 the function called, the regex must include (?C) at appropriate points. This
4370 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4371 This provides a means of identifying different callout points. When PCRE
4372 reaches such a point in the regex, if pcre_callout has been set, the external
4373 function is called. It is provided with data in a structure called
4374 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4375 matching continues; if it returns a non-zero value, the match at the current
4376 point fails. However, backtracking will occur if possible. [This was changed
4377 later and other features added - see item 49 below.]
4378
4379 29. pcretest is upgraded to test the callout functionality. It provides a
4380 callout function that displays information. By default, it shows the start of
4381 the match and the current position in the text. There are some new data escapes
4382 to vary what happens:
4383
4384 \C+ in addition, show current contents of captured substrings
4385 \C- do not supply a callout function
4386 \C!n return 1 when callout number n is reached
4387 \C!n!m return 1 when callout number n is reached for the mth time
4388
4389 30. If pcregrep was called with the -l option and just a single file name, it
4390 output "<stdin>" if a match was found, instead of the file name.
4391
4392 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4393 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4394 pcre_exec(). This saves a malloc/free per call. The default value of
4395 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4396 when configuring.
4397
4398 32. The default maximum size of a compiled pattern is 64K. There have been a
4399 few cases of people hitting this limit. The code now uses macros to handle the
4400 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4401 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4402 configuring. Tests 2 and 5 work only with 2-byte links because they output
4403 debugging information about compiled patterns.
4404
4405 33. Internal code re-arrangements:
4406
4407 (a) Moved the debugging function for printing out a compiled regex into
4408 its own source file (printint.c) and used #include to pull it into
4409 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4410 separate copies.
4411
4412 (b) Defined the list of op-code names for debugging as a macro in
4413 internal.h so that it is next to the definition of the opcodes.
4414
4415 (c) Defined a table of op-code lengths for simpler skipping along compiled
4416 code. This is again a macro in internal.h so that it is next to the
4417 definition of the opcodes.
4418
4419 34. Added support for recursive calls to individual subpatterns, along the
4420 lines of Robin Houston's patch (but implemented somewhat differently).
4421
4422 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4423 allow it to read and process whole directories in Win32. This code was
4424 contributed by Lionel Fourquaux; it has not been tested by me.
4425
4426 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4427 used to name a group. Names consist of alphanumerics and underscores, and must
4428 be unique. Back references use the syntax (?P=name) and recursive calls use
4429 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4430 numbers. The function pcre_fullinfo() can be used after compilation to extract
4431 a name/number map. There are three relevant calls:
4432
4433 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4434 PCRE_INFO_NAMECOUNT yields the number of entries
4435 PCRE_INFO_NAMETABLE yields a pointer to the map.
4436
4437 The map is a vector of fixed-size entries. The size of each entry depends on
4438 the length of the longest name used. The first two bytes of each entry are the
4439 group number, most significant byte first. There follows the corresponding
4440 name, zero terminated. The names are in alphabetical order.
4441
4442 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4443 case instead of 255. Making it the same both with and without UTF-8 support
4444 means that the same test output works with both.
4445
4446 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4447 calling malloc() with a zero argument.
4448
4449 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4450 optimization. I've improved things by keeping a bitmap of backreferences with
4451 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4452 fact referenced, the optimization can be applied. It is unlikely that a
4453 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4454 the match to follow \n) will appear inside brackets with a number greater than
4455 31, but if it does, any back reference > 31 suppresses the optimization.
4456
4457 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4458 of disabling numbered capturing parentheses. Any opening parenthesis that is
4459 not followed by ? behaves as if it were followed by ?: but named parentheses
4460 can still be used for capturing (and they will acquire numbers in the usual
4461 way).
4462
4463 41. Redesigned the return codes from the match() function into yes/no/error so
4464 that errors can be passed back from deep inside the nested calls. A malloc
4465 failure while inside a recursive subpattern call now causes the
4466 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4467
4468 42. It is now possible to set a limit on the number of times the match()
4469 function is called in a call to pcre_exec(). This facility makes it possible to
4470 limit the amount of recursion and backtracking, though not in a directly
4471 obvious way, because the match() function is used in a number of different
4472 circumstances. The count starts from zero for each position in the subject
4473 string (for non-anchored patterns). The default limit is, for compatibility, a
4474 large number, namely 10 000 000. You can change this in two ways:
4475
4476 (a) When configuring PCRE before making, you can use --with-match-limit=n
4477 to set a default value for the compiled library.
4478
4479 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4480 a different value is set. See 45 below.
4481
4482 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4483
4484 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4485 of things that can be changed at compile time. The first argument specifies
4486 what is wanted and the second points to where the information is to be placed.
4487 The current list of available information is:
4488
4489 PCRE_CONFIG_UTF8
4490
4491 The output is an integer that is set to one if UTF-8 support is available;
4492 otherwise it is set to zero.
4493
4494 PCRE_CONFIG_NEWLINE
4495
4496 The output is an integer that it set to the value of the code that is used for
4497 newline. It is either LF (10) or CR (13).
4498
4499 PCRE_CONFIG_LINK_SIZE
4500
4501 The output is an integer that contains the number of bytes used for internal
4502 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4503
4504 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4505
4506 The output is an integer that contains the threshold above which the POSIX
4507 interface uses malloc() for output vectors. See item 31 above.
4508
4509 PCRE_CONFIG_MATCH_LIMIT
4510
4511 The output is an unsigned integer that contains the default limit of the number
4512 of match() calls in a pcre_exec() execution. See 42 above.
4513
4514 44. pcretest has been upgraded by the addition of the -C option. This causes it
4515 to extract all the available output from the new pcre_config() function, and to
4516 output it. The program then exits immediately.
4517
4518 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4519 order to support additional features. One way would have been to define
4520 pcre_exec2() (for example) with extra arguments, but this would not have been
4521 extensible, and would also have required all calls to the original function to
4522 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4523 is used for passing in "extra" data from pcre_study().
4524
4525 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4526 contains the following fields:
4527
4528 flags a bitmap indicating which of the following fields are set
4529 study_data opaque data from pcre_study()
4530 match_limit a way of specifying a limit on match() calls for a specific
4531 call to pcre_exec()
4532 callout_data data for callouts (see 49 below)
4533
4534 The flag bits are also defined in pcre.h, and are
4535
4536 PCRE_EXTRA_STUDY_DATA
4537 PCRE_EXTRA_MATCH_LIMIT
4538 PCRE_EXTRA_CALLOUT_DATA
4539
4540 The pcre_study() function now returns one of these new pcre_extra blocks, with
4541 the actual study data pointed to by the study_data field, and the
4542 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4543 before. That is, this change is entirely upwards-compatible and requires no
4544 change to existing code.
4545
4546 If you want to pass in additional data to pcre_exec(), you can either place it
4547 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4548 block.
4549
4550 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4551 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4552 times with different match limits, until it finds the minimum value needed for
4553 pcre_exec() to complete. The value is then output. This can be instructive; for
4554 most simple matches the number is quite small, but for pathological cases it
4555 gets very large very quickly.
4556
4557 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4558 returns the size of the data block pointed to by the study_data field in a
4559 pcre_extra block, that is, the value that was passed as the argument to
4560 pcre_malloc() when PCRE was getting memory in which to place the information
4561 created by pcre_study(). The fourth argument should point to a size_t variable.
4562 pcretest has been extended so that this information is shown after a successful
4563 pcre_study() call when information about the compiled regex is being displayed.
4564
4565 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4566 because what follows is always an absolute path. (Later: it turns out that this
4567 is more than cosmetic for MinGW, because it doesn't like empty path
4568 components.)
4569
4570 49. Some changes have been made to the callout feature (see 28 above):
4571
4572 (i) A callout function now has three choices for what it returns:
4573
4574 0 => success, carry on matching
4575 > 0 => failure at this point, but backtrack if possible
4576 < 0 => serious error, return this value from pcre_exec()
4577
4578 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4579 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4580 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4581 use by callout functions. It will never be used by PCRE itself.
4582
4583 (ii) The pcre_extra structure (see 45 above) has a void * field called
4584 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4585 pcre_callout_block structure has a field of the same name. The contents of
4586 the field passed in the pcre_extra structure are passed to the callout
4587 function in the corresponding field in the callout block. This makes it
4588 easier to use the same callout-containing regex from multiple threads. For
4589 testing, the pcretest program has a new data escape
4590
4591 \C*n pass the number n (may be negative) as callout_data
4592
4593 If the callout function in pcretest receives a non-zero value as
4594 callout_data, it returns that value.
4595
4596 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4597 there were some redundant $(CFLAGS) in commands that are now specified as
4598 $(LINK), which already includes $(CFLAGS).
4599
4600 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4601 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4602 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4603 one-byte characters throughout. Note that case-insensitive matching applies
4604 only to characters whose values are less than 256. PCRE doesn't support the
4605 notion of cases for higher-valued characters.
4606
4607 (i) A character class whose characters are all within 0-255 is handled as
4608 a bit map, and the map is inverted for negative classes. Previously, a
4609 character > 255 always failed to match such a class; however it should
4610 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4611
4612 (ii) A negated character class with a single character < 255 is coded as
4613 "not this character" (OP_NOT). This wasn't working properly when the test
4614 character was multibyte, either singly or repeated.
4615
4616 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4617 mode, for example: \x{100}{2,3}.
4618
4619 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4620 singly or repeated) now correctly test multibyte characters. However,
4621 PCRE doesn't recognize any characters with values greater than 255 as
4622 digits, spaces, or word characters. Such characters always match \D, \S,
4623 and \W, and never match \d, \s, or \w.
4624
4625 (v) Classes may now contain characters and character ranges with values
4626 greater than 255. For example: [ab\x{100}-\x{400}].
4627
4628 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4629 PCRE in UTF-8 mode.
4630
4631 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4632 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4633 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4634 value.)
4635
4636 53. The single man page has become too large. I have therefore split it up into
4637 a number of separate man pages. These also give rise to individual HTML pages;
4638 these are now put in a separate directory, and there is an index.html page that
4639 lists them all. Some hyperlinking between the pages has been installed.
4640
4641 54. Added convenience functions for handling named capturing parentheses.
4642
4643 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4644 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4645 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4646 are faulted.
4647
4648 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4649 calling configure. These values are used when compiling the dftables.c program
4650 which is run to generate the source of the default character tables. They
4651 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4652 you will need to set these values.
4653
4654 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4655
4656
4657 Version 3.9 02-Jan-02
4658 ---------------------
4659
4660 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4661
4662 2. If --disable-static was given, the building process failed when trying to
4663 build pcretest and pcregrep. (For some reason it was using libtool to compile
4664 them, which is not right, as they aren't part of the library.)
4665
4666
4667 Version 3.8 18-Dec-01
4668 ---------------------
4669
4670 1. The experimental UTF-8 code was completely screwed up. It was packing the
4671 bytes in the wrong order. How dumb can you get?
4672
4673
4674 Version 3.7 29-Oct-01
4675 ---------------------
4676
4677 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4678 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4679 this didn't happen under Solaris 8, where I normally test things.
4680
4681 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4682 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4683 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4684 this fix an hour or so after the initial 3.7 release.)
4685
4686
4687 Version 3.6 23-Oct-01
4688 ---------------------
4689
4690 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4691 offsets passed as NULL with zero offset count.
4692
4693 2. The config.guess and config.sub files had not been updated when I moved to
4694 the latest autoconf.
4695
4696
4697 Version 3.5 15-Aug-01
4698 ---------------------
4699
4700 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4701 had been forgotten.
4702
4703 2. By using declared but undefined structures, we can avoid using "void"
4704 definitions in pcre.h while keeping the internal definitions of the structures
4705 private.
4706
4707 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4708 user point of view, this means that both static and shared libraries are built
4709 by default, but this can be individually controlled. More of the work of
4710 handling this static/shared cases is now inside libtool instead of PCRE's make
4711 file.
4712
4713 4. The pcretest utility is now installed along with pcregrep because it is
4714 useful for users (to test regexs) and by doing this, it automatically gets
4715 relinked by libtool. The documentation has been turned into a man page, so
4716 there are now .1, .txt, and .html versions in /doc.
4717
4718 5. Upgrades to pcregrep:
4719 (i) Added long-form option names like gnu grep.
4720 (ii) Added --help to list all options with an explanatory phrase.
4721 (iii) Added -r, --recursive to recurse into sub-directories.
4722 (iv) Added -f, --file to read patterns from a file.
4723
4724 6. pcre_exec() was referring to its "code" argument before testing that
4725 argument for NULL (and giving an error if it was NULL).
4726
4727 7. Upgraded Makefile.in to allow for compiling in a different directory from
4728 the source directory.
4729
4730 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4731 options bits, the pointer it was passed was to an int instead of to an unsigned
4732 long int. This mattered only on 64-bit systems.
4733
4734 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4735 generated) instead of pcre.in, which it its source. Also made the same change
4736 in several of the .c files.
4737
4738 10. A new release of gcc defines printf() as a macro, which broke pcretest
4739 because it had an ifdef in the middle of a string argument for printf(). Fixed
4740 by using separate calls to printf().
4741
4742 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4743 script, to force use of CR or LF instead of \n in the source. On non-Unix
4744 systems, the value can be set in config.h.
4745
4746 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4747 absolute limit. Changed the text of the error message to make this clear, and
4748 likewise updated the man page.
4749
4750 13. The limit of 99 on the number of capturing subpatterns has been removed.
4751 The new limit is 65535, which I hope will not be a "real" limit.
4752
4753
4754 Version 3.4 22-Aug-00
4755 ---------------------
4756
4757 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4758
4759 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4760
4761
4762 Version 3.3 01-Aug-00
4763 ---------------------
4764
4765 1. If an octal character was given, but the value was greater than \377, it
4766 was not getting masked to the least significant bits, as documented. This could
4767 lead to crashes in some systems.
4768
4769 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4770 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4771
4772 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4773 These just pass their arguments on to (pcre_free)(), but they are provided
4774 because some uses of PCRE bind it to non-C systems that can call its functions,
4775 but cannot call free() or pcre_free() directly.
4776
4777 4. Add "make test" as a synonym for "make check". Corrected some comments in
4778 the Makefile.
4779
4780 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4781 Makefile.
4782
4783 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4784 command called pgrep for grepping around the active processes.
4785
4786 7. Added the beginnings of support for UTF-8 character strings.
4787
4788 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4789 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4790 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4791 out for the ar command.)
4792
4793
4794 Version 3.2 12-May-00
4795 ---------------------
4796
4797 This is purely a bug fixing release.
4798
4799 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4800 of ZA. This was just one example of several cases that could provoke this bug,
4801 which was introduced by change 9 of version 2.00. The code for breaking
4802 infinite loops after an iteration that matches an empty string was't working
4803 correctly.
4804
4805 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4806 when matched against abbab (for example). After matching an empty string, it
4807 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4808 caused it to match further down the string than it should.
4809
4810 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4811 was there because it doesn't seem to be needed, and it causes trouble on some
4812 systems, as it is not a Standard C header. It has been removed.
4813
4814 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4815 were reported on the Macintosh. The changes were from
4816
4817 while ((c = *(++ptr)) != 0 && c != '\n');
4818 to
4819 while ((c = *(++ptr)) != 0 && c != '\n') ;
4820
4821 Totally extraordinary, but if that's what it takes...
4822
4823 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4824 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4825 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4826 assumes the way PCRE uses memmove() (always moving upwards).
4827
4828 6. PCRE is being used in one environment where strchr() is not available. There
4829 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4830 faster code anyway.
4831
4832
4833 Version 3.1 09-Feb-00
4834 ---------------------
4835
4836 The only change in this release is the fixing of some bugs in Makefile.in for
4837 the "install" target:
4838
4839 (1) It was failing to install pcreposix.h.
4840
4841 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4842
4843
4844 Version 3.0 01-Feb-00
4845 ---------------------
4846
4847 1. Add support for the /+ modifier to perltest (to output $` like it does in
4848 pcretest).
4849
4850 2. Add support for the /g modifier to perltest.
4851
4852 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4853 matches null strings.
4854
4855 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4856 pattern. Perl treats empty patterns specially - it reuses the most recent
4857 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4858 effect.
4859
4860 5. The POSIX interface was broken in that it was just handing over the POSIX
4861 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4862 required a bigger vector, with some working space on the end. This means that
4863 the POSIX wrapper now has to get and free some memory, and copy the results.
4864
4865 6. Added some simple autoconf support, placing the test data and the
4866 documentation in separate directories, re-organizing some of the
4867 information files, and making it build pcre-config (a GNU standard). Also added
4868 libtool support for building PCRE as a shared library, which is now the
4869 default.
4870
4871 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4872 09 are not valid octal constants. Single digits will be used for minor values
4873 less than 10.
4874
4875 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4876 existing programs that set these in the POSIX interface can use PCRE without
4877 modification.
4878
4879 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4880 return all that pcre_info() returns, plus additional data. The pcre_info()
4881 function is retained for compatibility, but is considered to be obsolete.
4882
4883 10. Added experimental recursion feature (?R) to handle one common case that
4884 Perl 5.6 will be able to do with (?p{...}).
4885
4886 11. Added support for POSIX character classes like [:alpha:], which Perl is
4887 adopting.
4888
4889
4890 Version 2.08 31-Aug-99
4891 ----------------------
4892
4893 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4894 trying to match at the startoffset position, but instead was moving forward to
4895 the next newline as if a previous match had failed.
4896
4897 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4898 and could get into a loop if a null string was matched other than at the start
4899 of the subject.
4900
4901 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4902 be distinguished at compile time, and for completeness also added PCRE_DATE.
4903
4904 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4905 in GnuWin32 environments.
4906
4907
4908 Version 2.07 29-Jul-99
4909 ----------------------
4910
4911 1. The documentation is now supplied in plain text form and HTML as well as in
4912 the form of man page sources.
4913
4914 2. C++ compilers don't like assigning (void *) values to other pointer types.
4915 In particular this affects malloc(). Although there is no problem in Standard
4916 C, I've put in casts to keep C++ compilers happy.
4917
4918 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4919 should be (const char *).
4920
4921 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4922 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4923 However, I haven't made this a standard facility. The documentation doesn't
4924 mention it, and the Makefile doesn't support it.
4925
4926 5. The Makefile now contains an "install" target, with editable destinations at
4927 the top of the file. The pcretest program is not installed.
4928
4929 6. pgrep -V now gives the PCRE version number and date.
4930
4931 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4932 causing the entire string to be ignored, instead of just the last character.
4933
4934 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4935 non-matching string, it can take a very, very long time, even for strings of
4936 quite modest length, because of the nested recursion. PCRE now does better in
4937 some of these cases. It does this by remembering the last required literal
4938 character in the pattern, and pre-searching the subject to ensure it is present
4939 before running the real match. In other words, it applies a heuristic to detect
4940 some types of certain failure quickly, and in the above example, if presented
4941 with a string that has no trailing " it gives "no match" very quickly.
4942
4943 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4944 other alternatives are tried instead.
4945
4946
4947 Version 2.06 09-Jun-99
4948 ----------------------
4949
4950 1. Change pcretest's output for amount of store used to show just the code
4951 space, because the remainder (the data block) varies in size between 32-bit and
4952 64-bit systems.
4953
4954 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4955 start matching at. This allows lookbehinds to work when searching for multiple
4956 occurrences in a string.
4957
4958 3. Added additional options to pcretest for testing multiple occurrences:
4959
4960 /+ outputs the rest of the string that follows a match
4961 /g loops for multiple occurrences, using the new startoffset argument
4962 /G loops for multiple occurrences by passing an incremented pointer
4963
4964 4. PCRE wasn't doing the "first character" optimization for patterns starting
4965 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4966 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4967 the letter 'x'. On long subject strings, this gives a significant speed-up.
4968
4969
4970 Version 2.05 21-Apr-99
4971 ----------------------
4972
4973 1. Changed the type of magic_number from int to long int so that it works
4974 properly on 16-bit systems.
4975
4976 2. Fixed a bug which caused patterns starting with .* not to work correctly
4977 when the subject string contained newline characters. PCRE was assuming
4978 anchoring for such patterns in all cases, which is not correct because .* will
4979 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4980 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4981 must be retried after every newline in the subject.
4982
4983
4984 Version 2.04 18-Feb-99
4985 ----------------------
4986
4987 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4988 computation of the store needed to hold the pattern was incorrect (too large).
4989 If such patterns were nested a few deep, this could multiply and become a real
4990 problem.
4991
4992 2. Added /M option to pcretest to show the memory requirement of a specific
4993 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4994
4995 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4996 compiled in such a way that the backtracking after subsequent failure was
4997 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4998 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4999
5000
5001 Version 2.03 02-Feb-99
5002 ----------------------
5003
5004 1. Fixed typo and small mistake in man page.
5005
5006 2. Added 4th condition (GPL supersedes if conflict) and created separate
5007 LICENCE file containing the conditions.
5008
5009 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5010 Perl, that is the internal \ allows the delimiter to be included in the
5011 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5012 the final delimiter, add \ to the end of the pattern (to test the error).
5013
5014 4. Added the convenience functions for extracting substrings after a successful
5015 match. Updated pcretest to make it able to test these functions.
5016
5017
5018 Version 2.02 14-Jan-99
5019 ----------------------
5020
5021 1. Initialized the working variables associated with each extraction so that
5022 their saving and restoring doesn't refer to uninitialized store.
5023
5024 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5025 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5026 fix the problem.
5027
5028 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5029 calls, and wasn't printing the correct value for compiling calls. Increased the
5030 default value of LOOPREPEAT, and the number of significant figures in the
5031 times.
5032
5033 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5034
5035 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5036 a building problem on Windows NT with a FAT file system.
5037
5038
5039 Version 2.01 21-Oct-98
5040 ----------------------
5041
5042 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5043 to character tables built by pcre_maketables() in the current locale. If NULL
5044 is passed, the default tables are used.
5045
5046
5047 Version 2.00 24-Sep-98
5048 ----------------------
5049
5050 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5051 it any more.
5052
5053 2. Allow quantification of (?>) groups, and make it work correctly.
5054
5055 3. The first character computation wasn't working for (?>) groups.
5056
5057 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5058 end of the subject) and add 5.005's \z, which really does match only at the
5059 very end of the subject.
5060
5061 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5062
5063 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5064 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5065 localized options. All options to pcre_study() were also removed.
5066
5067 7. Add other new features from 5.005:
5068
5069 $(?<= positive lookbehind
5070 $(?<! negative lookbehind
5071 (?imsx-imsx) added the unsetting capability
5072 such a setting is global if at outer level; local otherwise
5073 (?imsx-imsx:) non-capturing groups with option setting
5074 (?(cond)re|re) conditional pattern matching
5075
5076 A backreference to itself in a repeated group matches the previous
5077 captured string.
5078
5079 8. General tidying up of studying (both automatic and via "study")
5080 consequential on the addition of new assertions.
5081
5082 9. As in 5.005, unlimited repeated groups that could match an empty substring
5083 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5084 runtime if any iteration does actually match an empty substring.
5085
5086 10. Include the RunTest script in the distribution.
5087
5088 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5089 discrepancies, some of which were old and were also with respect to 5.004. They
5090 have now been fixed.
5091
5092
5093 Version 1.09 28-Apr-98
5094 ----------------------
5095
5096 1. A negated single character class followed by a quantifier with a minimum
5097 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5098 program crashes, or just wrong answers. This did not apply to negated classes
5099 containing more than one character, or to minima other than one.
5100
5101
5102 Version 1.08 27-Mar-98
5103 ----------------------
5104
5105 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5106
5107 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5108 latter must appear before anything that relies on it in the pattern.
5109
5110
5111 Version 1.07 16-Feb-98
5112 ----------------------
5113
5114 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5115 repeat of a potentially empty string).
5116
5117
5118 Version 1.06 23-Jan-98
5119 ----------------------
5120
5121 1. Added Markus Oberhumer's little patches for C++.
5122
5123 2. Literal strings longer than 255 characters were broken.
5124
5125
5126 Version 1.05 23-Dec-97
5127 ----------------------
5128
5129 1. Negated character classes containing more than one character were failing if
5130 PCRE_CASELESS was set at run time.
5131
5132
5133 Version 1.04 19-Dec-97
5134 ----------------------
5135
5136 1. Corrected the man page, where some "const" qualifiers had been omitted.
5137
5138 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5139 input syntax.
5140
5141 3. Fixed memory leak which occurred when a regex with back references was
5142 matched with an offsets vector that wasn't big enough. The temporary memory
5143 that is used in this case wasn't being freed if the match failed.
5144
5145 4. Tidied pcretest to ensure it frees memory that it gets.
5146
5147 5. Temporary memory was being obtained in the case where the passed offsets
5148 vector was exactly big enough.
5149
5150 6. Corrected definition of offsetof() from change 5 below.
5151
5152 7. I had screwed up change 6 below and broken the rules for the use of
5153 setjmp(). Now fixed.
5154
5155
5156 Version 1.03 18-Dec-97
5157 ----------------------
5158
5159 1. A erroneous regex with a missing opening parenthesis was correctly
5160 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5161 on some systems.
5162
5163 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5164 it was reported that one broken compiler failed on the former because "code" is
5165 also an independent variable.
5166
5167 3. The erroneous regex a[]b caused an array overrun reference.
5168
5169 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5170 fail on data ending with that character. (It was going on too far, and checking
5171 the next character, typically a binary zero.) This was specific to the
5172 optimized code for single-character negative classes.
5173
5174 5. Added a contributed patch from the TIN world which does the following:
5175
5176 + Add an undef for memmove, in case the the system defines a macro for it.
5177
5178 + Add a definition of offsetof(), in case there isn't one. (I don't know
5179 the reason behind this - offsetof() is part of the ANSI standard - but
5180 it does no harm).
5181
5182 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5183 most of the places where whitespace preceded '#'. I have given up and
5184 allowed the remaining 2 cases to be at the margin.
5185
5186 + Rename some variables in pcre to eliminate shadowing. This seems very
5187 pedantic, but does no harm, of course.
5188
5189 6. Moved the call to setjmp() into its own function, to get rid of warnings
5190 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5191
5192 7. Constructs such as \d{8,} were compiling into the equivalent of
5193 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5194 outcome, but in this particular case used more store than had been allocated,
5195 which caused the bug to be discovered because it threw up an internal error.
5196
5197 8. The debugging code in both pcre and pcretest for outputting the compiled
5198 form of a regex was going wrong in the case of back references followed by
5199 curly-bracketed repeats.
5200
5201
5202 Version 1.02 12-Dec-97
5203 ----------------------
5204
5205 1. Typos in pcre.3 and comments in the source fixed.
5206
5207 2. Applied a contributed patch to get rid of places where it used to remove
5208 'const' from variables, and fixed some signed/unsigned and uninitialized
5209 variable warnings.
5210
5211 3. Added the "runtest" target to Makefile.
5212
5213 4. Set default compiler flag to -O2 rather than just -O.
5214
5215
5216 Version 1.01 19-Nov-97
5217 ----------------------
5218
5219 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5220 like /([ab]*)*/, that is, for classes with more than one character in them.
5221
5222 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5223 as /((?>a*))*/ (a PCRE_EXTRA facility).
5224
5225
5226 Version 1.00 18-Nov-97
5227 ----------------------
5228
5229 1. Added compile-time macros to support systems such as SunOS4 which don't have
5230 memmove() or strerror() but have other things that can be used instead.
5231
5232 2. Arranged that "make clean" removes the executables.
5233
5234
5235 Version 0.99 27-Oct-97
5236 ----------------------
5237
5238 1. Fixed bug in code for optimizing classes with only one character. It was
5239 initializing a 32-byte map regardless, which could cause it to run off the end
5240 of the memory it had got.
5241
5242 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5243
5244
5245 Version 0.98 22-Oct-97
5246 ----------------------
5247
5248 1. Fixed bug in code for handling temporary memory usage when there are more
5249 back references than supplied space in the ovector. This could cause segfaults.
5250
5251
5252 Version 0.97 21-Oct-97
5253 ----------------------
5254
5255 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5256
5257 2. Optimized negated single characters not to use a bit map.
5258
5259 3. Brought error texts together as macro definitions; clarified some of them;
5260 fixed one that was wrong - it said "range out of order" when it meant "invalid
5261 escape sequence".
5262
5263 4. Changed some char * arguments to const char *.
5264
5265 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5266
5267 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5268 pcretest.
5269
5270
5271 Version 0.96 16-Oct-97
5272 ----------------------
5273
5274 1. Added a simple "pgrep" utility to the distribution.
5275
5276 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5277 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5278 where "ddd" means "one or more decimal digits".
5279
5280 3. Fixed serious bug. If a pattern had a back reference, but the call to
5281 pcre_exec() didn't supply a large enough ovector to record the related
5282 identifying subpattern, the match always failed. PCRE now remembers the number
5283 of the largest back reference, and gets some temporary memory in which to save
5284 the offsets during matching if necessary, in order to ensure that
5285 backreferences always work.
5286
5287 4. Increased the compatibility with Perl in a number of ways:
5288
5289 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5290 to request this handling. The option can be set at compile or exec time.
5291
5292 (b) $ matches before a terminating newline by default; an option
5293 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5294 mode). The option can be set at compile or exec time.
5295
5296 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5297 the same as Perl's. If the decimal number it represents is less than 10
5298 or there aren't that many previous left capturing parentheses, an octal
5299 escape is read. Inside a character class, it's always an octal escape,
5300 even if it is a single digit.
5301
5302 (d) An escaped but undefined alphabetic character is taken as a literal,
5303 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5304 escapes.
5305
5306 (e) {0} is now permitted. (The previous item is removed from the compiled
5307 pattern).
5308
5309 5. Changed all the names of code files so that the basic parts are no longer
5310 than 10 characters, and abolished the teeny "globals.c" file.
5311
5312 6. Changed the handling of character classes; they are now done with a 32-byte
5313 bit map always.
5314
5315 7. Added the -d and /D options to pcretest to make it possible to look at the
5316 internals of compilation without having to recompile pcre.
5317
5318
5319 Version 0.95 23-Sep-97
5320 ----------------------
5321
5322 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5323 \x20 at the start of a run of normal characters. These were being treated as
5324 real characters, instead of the source characters being re-checked.
5325
5326
5327 Version 0.94 18-Sep-97
5328 ----------------------
5329
5330 1. The functions are now thread-safe, with the caveat that the global variables
5331 containing pointers to malloc() and free() or alternative functions are the
5332 same for all threads.
5333
5334 2. Get pcre_study() to generate a bitmap of initial characters for non-
5335 anchored patterns when this is possible, and use it if passed to pcre_exec().
5336
5337
5338 Version 0.93 15-Sep-97
5339 ----------------------
5340
5341 1. /(b)|(:+)/ was computing an incorrect first character.
5342
5343 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5344 but not actually doing anything yet.
5345
5346 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5347 as Perl does (e.g. [-az] or [az-]).
5348
5349 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5350 all possible positions.
5351
5352 5. Split up into different modules to avoid including unneeded functions in a
5353 compiled binary. However, compile and exec are still in one module. The "study"
5354 function is split off.
5355
5356 6. The character tables are now in a separate module whose source is generated
5357 by an auxiliary program - but can then be edited by hand if required. There are
5358 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5359 toupper() in the code.
5360
5361 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5362 make them global. Abolish the function for setting them, as the caller can now
5363 set them directly.
5364
5365
5366 Version 0.92 11-Sep-97
5367 ----------------------
5368
5369 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5370 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5371
5372 2. Caseless matching was not working in character classes if the characters in
5373 the pattern were in upper case.
5374
5375 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5376
5377 4. Make PCRE_ANCHORED public and accept as a compile option.
5378
5379 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5380 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5381 pass them.
5382
5383 6. Give an error if bad option bits passed at compile or run time.
5384
5385 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5386 pcretest to cause it to pass that flag.
5387
5388 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5389 options, and the first character, if set.
5390
5391 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5392
5393
5394 Version 0.91 10-Sep-97
5395 ----------------------
5396
5397 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5398 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5399
5400 2. PCRE was looping on encountering an indefinitely repeated back reference to
5401 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5402 Perl does - treats the match as successful.
5403
5404 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5