ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1478 - (show annotations)
Tue May 27 13:18:31 2014 UTC (5 years, 2 months ago) by ph10
File size: 265906 byte(s)
Fix empty-matching possessive zero-repeat groups bug.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.36 xx-xxx-2014
5 ------------------------
7 1. Got rid of some compiler warnings in the C++ modules that were shown up by
8 -Wmissing-field-initializers and -Wunused-parameter.
10 2. The tests for quantifiers being too big (greater than 65535) were being
11 applied after reading the number, and stupidly assuming that integer
12 overflow would give a negative number. The tests are now applied as the
13 numbers are read.
15 3. Tidy code in pcre_exec.c where two branches that used to be different are
16 now the same.
18 4. The JIT compiler did not generate match limit checks for certain
19 bracketed expressions with quantifiers. This may lead to exponential
20 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
21 issue should be resolved now.
23 5. Fixed an issue, which occures when nested alternatives are optimized
24 with table jumps.
26 6. Inserted two casts and changed some ints to size_t in the light of some
27 reported 64-bit compiler warnings (Bugzilla 1477).
29 7. Fixed a bug concerned with zero-minimum possessive groups that could match
30 an empty string, which sometimes were behaving incorrectly in the
31 interpreter (though correctly in the JIT matcher). This pcretest input is
32 an example:
34 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
37 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
38 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
39 for an empty string was breaking the inner loop and carrying on at a lower
40 level, when possessive repeated groups should always return to a higher
41 level as they have no backtrack points in them. The empty string test now
42 occurs at the outer level.
45 Version 8.35 04-April-2014
46 --------------------------
48 1. A new flag is set, when property checks are present in an XCLASS.
49 When this flag is not set, PCRE can perform certain optimizations
50 such as studying these XCLASS-es.
52 2. The auto-possessification of character sets were improved: a normal
53 and an extended character set can be compared now. Furthermore
54 the JIT compiler optimizes more character set checks.
56 3. Got rid of some compiler warnings for potentially uninitialized variables
57 that show up only when compiled with -O2.
59 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
60 of a match later then the end of the match. The pcretest program was not
61 handling the case sensibly - it was outputting from the start to the next
62 binary zero. It now reports this situation in a message, and outputs the
63 text from the end to the start.
65 5. Fast forward search is improved in JIT. Instead of the first three
66 characters, any three characters with fixed position can be searched.
67 Search order: first, last, middle.
69 6. Improve character range checks in JIT. Characters are read by an inprecise
70 function now, which returns with an unknown value if the character code is
71 above a certain treshold (e.g: 256). The only limitation is that the value
72 must be bigger than the treshold as well. This function is useful, when
73 the characters above the treshold are handled in the same way.
75 7. The macros whose names start with RAWUCHAR are placeholders for a future
76 mode in which only the bottom 21 bits of 32-bit data items are used. To
77 make this more memorable for those maintaining the code, the names have
78 been changed to start with UCHAR21, and an extensive comment has been added
79 to their definition.
81 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
82 to the export list in Makefile.am (they were accidentally omitted from the
83 8.34 tarball).
85 9. The informational output from pcretest used the phrase "starting byte set"
86 which is inappropriate for the 16-bit and 32-bit libraries. As the output
87 for "first char" and "need char" really means "non-UTF-char", I've changed
88 "byte" to "char", and slightly reworded the output. The documentation about
89 these values has also been (I hope) clarified.
91 10. Another JIT related optimization: use table jumps for selecting the correct
92 backtracking path, when more than four alternatives are present inside a
93 bracket.
95 11. Empty match is not possible, when the minimum length is greater than zero,
96 and there is no \K in the pattern. JIT should avoid empty match checks in
97 such cases.
99 12. In a caseless character class with UCP support, when a character with more
100 than one alternative case was not the first character of a range, not all
101 the alternative cases were added to the class. For example, s and \x{17f}
102 are both alternative cases for S: the class [RST] was handled correctly,
103 but [R-T] was not.
105 13. The configure.ac file always checked for pthread support when JIT was
106 enabled. This is not used in Windows, so I have put this test inside a
107 check for the presence of windows.h (which was already tested for).
109 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
110 The algorithm provides a way to skip certain starting offsets, and usually
111 faster than linear prefix searches.
113 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
114 as for 'fr_FR' and 'french'. For some reason, however, it then used the
115 Windows-specific input and output files, which have 'french' screwed in.
116 So this could never have worked. One of the problems with locales is that
117 they aren't always the same. I have now updated RunTest so that it checks
118 the output of the locale test (test 3) against three different output
119 files, and it allows the test to pass if any one of them matches. With luck
120 this should make the test pass on some versions of Solaris where it was
121 failing. Because of the uncertainty, the script did not used to stop if
122 test 3 failed; it now does. If further versions of a French locale ever
123 come to light, they can now easily be added.
125 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
126 there was a message during ./configure, but it did not stop. This now
127 provokes an error. The invalid example in README has been corrected.
128 If a value less than the minimum is given, the minimum value has always
129 been used, but now a warning is given.
131 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
132 was a bug in the test system, which is now fixed. Also, the list of various
133 configurations that are tested for each release did not have one with both
134 16/32 bits and --enable-bar-anycrlf. It now does.
136 18. pcretest was missing "-C bsr" for displaying the \R default setting.
138 19. Little endian PowerPC systems are supported now by the JIT compiler.
140 20. The fast forward newline mechanism could enter to an infinite loop on
141 certain invalid UTF-8 input. Although we don't support these cases
142 this issue can be fixed by a performance optimization.
144 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
145 not take account if existing stack usage. There is now a new global
146 variable called pcre_stack_guard that can be set to point to an external
147 function to check stack availability. It is called at the start of
148 processing every parenthesized group.
150 22. A typo in the code meant that in ungreedy mode the max/min qualifier
151 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
152 not match "ab".
154 23. When UTF was disabled, the JIT program reported some incorrect compile
155 errors. These messages are silenced now.
157 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
158 compiler.
160 25. Change all the temporary files used in RunGrepTest to be different to those
161 used by RunTest so that the tests can be run simultaneously, for example by
162 "make -j check".
165 Version 8.34 15-December-2013
166 -----------------------------
168 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
169 executable memory. Patch inspired by Carsten Klein.
171 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
172 this macro is never tested and has no effect, because the work to support
173 coverage involves only compiling and linking options and special targets in
174 the Makefile. The comment in config.h implied that defining the macro would
175 enable coverage support, which is totally false. There was also support for
176 setting this macro in the CMake files (my fault, I just copied it from
177 configure). SUPPORT_GCOV has now been removed.
179 3. Make a small performance improvement in strlen16() and strlen32() in
180 pcretest.
182 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
183 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
184 these cases). There was also one in pcretest.c.
186 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
188 6. In UTF mode, the code for checking whether a group could match an empty
189 string (which is used for indefinitely repeated groups to allow for
190 breaking an infinite loop) was broken when the group contained a repeated
191 negated single-character class with a character that occupied more than one
192 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
193 UTF-8 mode). The effect was undefined: the group might or might not be
194 deemed as matching an empty string, or the program might have crashed.
196 7. The code for checking whether a group could match an empty string was not
197 recognizing that \h, \H, \v, \V, and \R must match a character.
199 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
200 an empty string. If it can, pcretest shows this in its information output.
202 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
203 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
204 matched by pcre_exec() without using JIT:
206 (a) If the rest of the pattern did not match after a maximal run of
207 grapheme clusters, the code for backing up to try with fewer of them
208 did not always back up over a full grapheme when characters that do not
209 have the modifier quality were involved, e.g. Hangul syllables.
211 (b) If the match point in a subject started with modifier character, and
212 there was no match, the code could incorrectly back up beyond the match
213 point, and potentially beyond the first character in the subject,
214 leading to a segfault or an incorrect match result.
216 10. A conditional group with an assertion condition could lead to PCRE
217 recording an incorrect first data item for a match if no other first data
218 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
219 first data item, and therefore matched "ca" after "c" instead of at the
220 start.
222 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
223 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
225 12. The source of pcregrep now includes z/OS-specific code so that it can be
226 compiled for z/OS as part of the special z/OS distribution.
228 13. Added the -T and -TM options to pcretest.
230 14. The code in pcre_compile.c for creating the table of named capturing groups
231 has been refactored. Instead of creating the table dynamically during the
232 actual compiling pass, the information is remembered during the pre-compile
233 pass (on the stack unless there are more than 20 named groups, in which
234 case malloc() is used) and the whole table is created before the actual
235 compile happens. This has simplified the code (it is now nearly 150 lines
236 shorter) and prepared the way for better handling of references to groups
237 with duplicate names.
239 15. A back reference to a named subpattern when there is more than one of the
240 same name now checks them in the order in which they appear in the pattern.
241 The first one that is set is used for the reference. Previously only the
242 first one was inspected. This change makes PCRE more compatible with Perl.
244 16. Unicode character properties were updated from Unicode 6.3.0.
246 17. The compile-time code for auto-possessification has been refactored, based
247 on a patch by Zoltan Herczeg. It now happens after instead of during
248 compilation. The code is cleaner, and more cases are handled. The option
249 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
250 options in pcretest are provided to set it. It can also be set by
251 (*NO_AUTO_POSSESS) at the start of a pattern.
253 18. The character VT has been added to the default ("C" locale) set of
254 characters that match \s and are generally treated as white space,
255 following this same change in Perl 5.18. There is now no difference between
256 "Perl space" and "POSIX space". Whether VT is treated as white space in
257 other locales depends on the locale.
259 19. The code for checking named groups as conditions, either for being set or
260 for being recursed, has been refactored (this is related to 14 and 15
261 above). Processing unduplicated named groups should now be as fast at
262 numerical groups, and processing duplicated groups should be faster than
263 before.
265 20. Two patches to the CMake build system, by Alexander Barkov:
267 (1) Replace the "source" command by "." in CMakeLists.txt because
268 "source" is a bash-ism.
270 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
271 without these the CMake build does not work on Solaris.
273 21. Perl has changed its handling of \8 and \9. If there is no previously
274 encountered capturing group of those numbers, they are treated as the
275 literal characters 8 and 9 instead of a binary zero followed by the
276 literals. PCRE now does the same.
278 22. Following Perl, added \o{} to specify codepoints in octal, making it
279 possible to specify values greater than 0777 and also making them
280 unambiguous.
282 23. Perl now gives an error for missing closing braces after \x{... instead of
283 treating the string as literal. PCRE now does the same.
285 24. RunTest used to grumble if an inappropriate test was selected explicitly,
286 but just skip it when running all tests. This make it awkward to run ranges
287 of tests when one of them was inappropriate. Now it just skips any
288 inappropriate tests, as it always did when running all tests.
290 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
291 character types such as \d or \w, too many callouts were inserted, and the
292 data that they returned was rubbish.
294 26. In UCP mode, \s was not matching two of the characters that Perl matches,
295 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
296 were matched by \h. The code has now been refactored so that the lists of
297 the horizontal and vertical whitespace characters used for \h and \v (which
298 are defined only in one place) are now also used for \s.
300 27. Add JIT support for the 64 bit TileGX architecture.
301 Patch by Jiong Wang (Tilera Corporation).
303 28. Possessive quantifiers for classes (both explicit and automatically
304 generated) now use special opcodes instead of wrapping in ONCE brackets.
306 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
307 (because it's meaningless), this was not happening when PCRE_CASELESS was
308 set. Not wrong, but inefficient.
310 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
311 properties for \w, \d, etc) is present in a test regex. Otherwise if the
312 test contains no characters greater than 255, Perl doesn't realise it
313 should be using Unicode semantics.
315 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
316 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
317 does in Unicode mode.
319 32. Added the "forbid" facility to pcretest so that putting tests into the
320 wrong test files can sometimes be quickly detected.
322 33. There is now a limit (default 250) on the depth of nesting of parentheses.
323 This limit is imposed to control the amount of system stack used at compile
324 time. It can be changed at build time by --with-parens-nest-limit=xxx or
325 the equivalent in CMake.
327 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
328 errors. Perl warns for these when in warning mode, but PCRE has no facility
329 for giving warnings.
331 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
332 However, this was not working for (?!) because it is optimized to (*FAIL),
333 for which PCRE does not allow quantifiers. The optimization is now disabled
334 when a quantifier follows (?!). I can't see any use for this, but it makes
335 things uniform.
337 36. Perl no longer allows group names to start with digits, so I have made this
338 change also in PCRE. It simplifies the code a bit.
340 37. In extended mode, Perl ignores spaces before a + that indicates a
341 possessive quantifier. PCRE allowed a space before the quantifier, but not
342 before the possessive +. It now does.
344 38. The use of \K (reset reported match start) within a repeated possessive
345 group such as (a\Kb)*+ was not working.
347 40. Document that the same character tables must be used at compile time and
348 run time, and that the facility to pass tables to pcre_exec() and
349 pcre_dfa_exec() is for use only with saved/restored patterns.
351 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
352 features for Builds with MSVC:
354 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
355 builds). The .rc files can be used to set FileDescription and many other
356 attributes.
358 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
359 This allows higher-level build scripts which want .pdb files to avoid
360 hard-coding the exact files needed."
362 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
363 mean "start of word" and "end of word", respectively, as a transition aid.
365 43. A minimizing repeat of a class containing codepoints greater than 255 in
366 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
367 compiled to use the heap for recursion.
369 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
370 is configured.
373 Version 8.33 28-May-2013
374 ------------------------
376 1. Added 'U' to some constants that are compared to unsigned integers, to
377 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
378 variables that are added to signed variables, to ensure the result is
379 signed and can be negated.
381 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
382 CMake config files.
384 3. Revise the creation of config.h.generic so that all boolean macros are
385 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
386 overriding via -D on the command line possible.
388 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
389 to unsigned int is reported to make a quite noticeable speed difference in
390 a specific Windows environment. Testing on Linux did also appear to show
391 some benefit (and it is clearly not harmful). Also fixed the definition of
392 Xop which should be unsigned.
394 5. Related to (4), changing the definition of the intermediate variable cc
395 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
396 improvements.
398 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
399 unnecessary spaces.
401 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
402 and later.
404 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
406 9. Optimizing fast_forward_start_bits in JIT.
408 10. Adding support for callouts in JIT, and fixing some issues revealed
409 during this work. Namely:
411 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
413 (b) Minimum length was not checked before the matching is started.
415 11. The value of capture_last that is passed to callouts was incorrect in some
416 cases when there was a capture on one path that was subsequently abandoned
417 after a backtrack. Also, the capture_last value is now reset after a
418 recursion, since all captures are also reset in this case.
420 12. The interpreter no longer returns the "too many substrings" error in the
421 case when an overflowing capture is in a branch that is subsequently
422 abandoned after a backtrack.
424 13. In the pathological case when an offset vector of size 2 is used, pcretest
425 now prints out the matched string after a yield of 0 or 1.
427 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
428 Only supported by the JIT compiler at the moment.
430 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
432 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
433 offset of the starting point of the matching process, provided the offsets
434 vector is large enough.
436 17. The \A escape now records a lookbehind value of 1, though its execution
437 does not actually inspect the previous character. This is to ensure that,
438 in partial multi-segment matching, at least one character from the old
439 segment is retained when a new segment is processed. Otherwise, if there
440 are no lookbehinds in the pattern, \A might match incorrectly at the start
441 of a new segment.
443 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
445 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
446 gives some modest performance improvement in 8-bit mode.
448 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
449 be expressed in certain programming languages using Universal Character
450 Names.
452 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
453 which points out that "non characters" are not "characters that may not
454 appear in Unicode strings" but rather "characters that are reserved for
455 internal use and have only local meaning".
457 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
458 there was a conditional group that depended on an assertion, if the
459 assertion was false, the callout that immediately followed the alternation
460 in the condition was skipped when pcre_exec() was used for matching.
462 23. Allow an explicit callout to be inserted before an assertion that is the
463 condition for a conditional group, for compatibility with automatic
464 callouts, which always insert a callout at this point.
466 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
467 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
469 25. (*PRUNE) is now supported by the JIT compiler.
471 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
473 27. Fix the case where there are two or more SKIPs with arguments that may be
474 ignored.
476 28. (*SKIP) is now supported by the JIT compiler.
478 29. (*THEN) is now supported by the JIT compiler.
480 30. Update RunTest with additional test selector options.
482 31. The way PCRE handles backtracking verbs has been changed in two ways.
484 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
485 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
486 backtracking. In some cases this makes it more Perl-compatible, but Perl's
487 rather obscure rules do not always do the same thing.
489 (2) Previously, backtracking verbs were confined within assertions. This is
490 no longer the case for positive assertions, except for (*ACCEPT). Again,
491 this sometimes improves Perl compatibility, and sometimes does not.
493 32. A number of tests that were in test 2 because Perl did things differently
494 have been moved to test 1, because either Perl or PCRE has changed, and
495 these tests are now compatible.
497 32. Backtracking control verbs are now handled in the same way in JIT and
498 interpreter.
500 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
501 contained a forward subroutine reference caused a compile error.
503 34. Auto-detect and optimize limited repetitions in JIT.
505 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
506 blocking (*UTF) etc.
508 36. In the interpreter, maximizing pattern repetitions for characters and
509 character types now use tail recursion, which reduces stack usage.
511 37. The value of the max lookbehind was not correctly preserved if a compiled
512 and saved regex was reloaded on a host of different endianness.
514 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
515 of the compiled pattern block, expand the flags field from 16 to 32 bits
516 because it was almost full.
518 39. Try madvise first before posix_madvise.
520 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
521 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
522 original change locked out all matches of empty strings. This has been
523 changed so that one match of an empty string per line is recognized.
524 Subsequent searches on the same line (for colouring or for --only-matching,
525 for example) do not recognize empty strings.
527 41. Applied a user patch to fix a number of spelling mistakes in comments.
529 42. Data lines longer than 65536 caused pcretest to crash.
531 43. Clarified the data type for length and startoffset arguments for pcre_exec
532 and pcre_dfa_exec in the function-specific man pages, where they were
533 explicitly stated to be in bytes, never having been updated. I also added
534 some clarification to the pcreapi man page.
536 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
537 a segmentation fault.
540 Version 8.32 30-November-2012
541 -----------------------------
543 1. Improved JIT compiler optimizations for first character search and single
544 character iterators.
546 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
547 Patch by Daniel Richard G.
549 3. Single character iterator optimizations in the JIT compiler.
551 4. Improved JIT compiler optimizations for character ranges.
553 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
554 Reported by Giuseppe D'Angelo.
556 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
557 of a line, was being set incorrectly in cases where .* appeared inside
558 atomic brackets at the start of a pattern, or where there was a subsequent
559 *PRUNE or *SKIP.
561 7. Improved instruction cache flush for POWER/PowerPC.
562 Patch by Daniel Richard G.
564 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
565 grep:
567 (a) There is now no limit to the number of patterns to be matched.
569 (b) An error is given if a pattern is too long.
571 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
572 are now supported.
574 (d) --exclude-from and --include-from (multiple use) have been added.
576 (e) Exclusions and inclusions now apply to all files and directories, not
577 just to those obtained from scanning a directory recursively.
579 (f) Multiple uses of -f and --file-list are now supported.
581 (g) In a Windows environment, the default for -d has been changed from
582 "read" (the GNU grep default) to "skip", because otherwise the presence
583 of a directory in the file list provokes an error.
585 (h) The documentation has been revised and clarified in places.
587 9. Improve the matching speed of capturing brackets.
589 10. Changed the meaning of \X so that it now matches a Unicode extended
590 grapheme cluster.
592 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
593 out POSIX threads when JIT support is configured.
595 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
597 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
598 structure could go wrong in environments where size_t is not the same size
599 as int.
601 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
602 set.
604 15. The EBCDIC support had decayed; later updates to the code had included
605 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
606 general tidy up of EBCDIC-related issues, and the documentation was also
607 not quite right. There is now a test that can be run on ASCII systems to
608 check some of the EBCDIC-related things (but is it not a full test).
610 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
611 in a small tidy to the code.
613 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
615 18. If the --only-matching (-o) option in pcregrep is specified multiple
616 times, each one causes appropriate output. For example, -o1 -o2 outputs the
617 substrings matched by the 1st and 2nd capturing parentheses. A separating
618 string can be specified by --om-separator (default empty).
620 19. Improving the first n character searches.
622 20. Turn case lists for horizontal and vertical white space into macros so that
623 they are defined only once.
625 21. This set of changes together give more compatible Unicode case-folding
626 behaviour for characters that have more than one other case when UCP
627 support is available.
629 (a) The Unicode property table now has offsets into a new table of sets of
630 three or more characters that are case-equivalent. The MultiStage2.py
631 script that generates these tables (the pcre_ucd.c file) now scans
632 CaseFolding.txt instead of UnicodeData.txt for character case
633 information.
635 (b) The code for adding characters or ranges of characters to a character
636 class has been abstracted into a generalized function that also handles
637 case-independence. In UTF-mode with UCP support, this uses the new data
638 to handle characters with more than one other case.
640 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
641 whose other case is greater than 256 are now correctly matched
642 caselessly. Previously, the high codepoint matched the low one, but not
643 vice versa.
645 (d) The processing of \h, \H, \v, and \ in character classes now makes use
646 of the new class addition function, using character lists defined as
647 macros alongside the case definitions of 20 above.
649 (e) Caseless back references now work with characters that have more than
650 one other case.
652 (f) General caseless matching of characters with more than one other case
653 is supported.
655 22. Unicode character properties were updated from Unicode 6.2.0
657 23. Improved CMake support under Windows. Patch by Daniel Richard G.
659 24. Add support for 32-bit character strings, and UTF-32
661 25. Major JIT compiler update (code refactoring and bugfixing).
662 Experimental Sparc 32 support is added.
664 26. Applied a modified version of Daniel Richard G's patch to create
665 pcre.h.generic and config.h.generic by "make" instead of in the
666 PrepareRelease script.
668 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
669 pcre_compile.c when checking for a zero character.
671 28. Introducing a native interface for JIT. Through this interface, the compiled
672 machine code can be directly executed. The purpose of this interface is to
673 provide fast pattern matching, so several sanity checks are not performed.
674 However, feature tests are still performed. The new interface provides
675 1.4x speedup compared to the old one.
677 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
678 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
679 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
681 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
682 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
683 the "old" RFC 2279). Instead, it ended up passing a negative length to
684 pcre_exec().
686 31. Add support for GCC's visibility feature to hide internal functions.
688 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
689 "unknown -C option" after outputting 0 or 1.
691 33. There is now support for generating a code coverage report for the test
692 suite in environments where gcc is the compiler and lcov is installed. This
693 is mainly for the benefit of the developers.
695 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
696 unaddressable using valgrind annotations, allowing valgrind to detect
697 invalid memory accesses. This is mainly for the benefit of the developers.
699 25. (*UTF) can now be used to start a pattern in any of the three libraries.
701 26. Give configure error if --enable-cpp but no C++ compiler found.
704 Version 8.31 06-July-2012
705 -------------------------
707 1. Fixing a wrong JIT test case and some compiler warnings.
709 2. Removed a bashism from the RunTest script.
711 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
712 to unsigned type, result still unsigned" that was given by an MS compiler
713 on encountering the code "-sizeof(xxx)".
715 4. Partial matching support is added to the JIT compiler.
717 5. Fixed several bugs concerned with partial matching of items that consist
718 of more than one character:
720 (a) /^(..)\1/ did not partially match "aba" because checking references was
721 done on an "all or nothing" basis. This also applied to repeated
722 references.
724 (b) \R did not give a hard partial match if \r was found at the end of the
725 subject.
727 (c) \X did not give a hard partial match after matching one or more
728 characters at the end of the subject.
730 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
731 a partial match for the string "\r".
733 (e) When newline was set to CRLF, the metacharacter "." did not recognize
734 a partial match for a CR character at the end of the subject string.
736 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
737 running pcretest, the text "(JIT)" added to the output whenever JIT is
738 actually used to run the match.
740 7. Individual JIT compile options can be set in pcretest by following -s+[+]
741 or /S+[+] with a digit between 1 and 7.
743 8. OP_NOT now supports any UTF character not just single-byte ones.
745 9. (*MARK) control verb is now supported by the JIT compiler.
747 10. The command "./RunTest list" lists the available tests without actually
748 running any of them. (Because I keep forgetting what they all are.)
752 12. Applied a (slightly modified) user-supplied patch that improves performance
753 when the heap is used for recursion (compiled with --disable-stack-for-
754 recursion). Instead of malloc and free for each heap frame each time a
755 logical recursion happens, frames are retained on a chain and re-used where
756 possible. This sometimes gives as much as 30% improvement.
758 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
759 call.
761 14. As documented, (*COMMIT) is now confined to within a positive assertion.
763 15. It is now possible to link pcretest with libedit as an alternative to
764 libreadline.
766 16. (*COMMIT) control verb is now supported by the JIT compiler.
768 17. The Unicode data tables have been updated to Unicode 6.1.0.
770 18. Added --file-list option to pcregrep.
772 19. Added binary file support to pcregrep, including the -a, --binary-files,
773 -I, and --text options.
775 20. The madvise function is renamed for posix_madvise for QNX compatibility
776 reasons. Fixed by Giuseppe D'Angelo.
778 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
780 22. Changed the option for creating symbolic links for 16-bit man pages from
781 -s to -sf so that re-installing does not cause issues.
783 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
785 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
786 matches in certain environments (the workspace was not being correctly
787 retained). Also added to pcre_dfa_exec() a simple plausibility check on
788 some of the workspace data at the beginning of a restart.
790 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
791 was not doing so when it should - probably a typo introduced by SVN 528
792 (change 8.10/14).
794 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
795 \w+ when the character tables indicated that \x{c4} was a word character.
796 There were several related cases, all because the tests for doing a table
797 lookup were testing for characters less than 127 instead of 255.
799 27. If a pattern contains capturing parentheses that are not used in a match,
800 their slots in the ovector are set to -1. For those that are higher than
801 any matched groups, this happens at the end of processing. In the case when
802 there were back references that the ovector was too small to contain
803 (causing temporary malloc'd memory to be used during matching), and the
804 highest capturing number was not used, memory off the end of the ovector
805 was incorrectly being set to -1. (It was using the size of the temporary
806 memory instead of the true size.)
808 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
809 ovector size, it uses memory at the end of the block that it has got.
811 29. Check for an overlong MARK name and give an error at compile time. The
812 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
814 30. JIT compiler update.
816 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
817 Rill for the patch.
819 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
821 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
823 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
824 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
825 was enabled.
827 35. Improve JIT code generation for greedy plus quantifier.
829 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
830 1 to "aa" instead of to an empty string. The bug affected repeated groups
831 that could potentially match an empty string.
833 37. Optimizing single character iterators in JIT.
835 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
836 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
837 codepoints that are too big for the mode are faulted, and in a UTF mode,
838 disallowed codepoints are also faulted.
840 39. If PCRE was compiled with UTF support, in three places in the DFA
841 matcher there was code that should only have been obeyed in UTF mode, but
842 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
843 processing when bytes with values greater than 127 were present. In 16-bit
844 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
845 both cases the values are those that cannot be the first data item in a UTF
846 character. The three items that might have provoked this were recursions,
847 possessively repeated groups, and atomic groups.
849 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
850 and pcregrep, because some OS require shared objects to be explicitly
851 passed to ld, causing the link step to fail if they are not.
853 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
854 mode, patterns that started with \h* or \R* might be incorrectly matched.
857 Version 8.30 04-February-2012
858 -----------------------------
860 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
861 name is defined in ctype.h.
863 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
864 only in quite long subpatterns.
866 3. Removed the function pcre_info(), which has been obsolete and deprecated
867 since it was replaced by pcre_fullinfo() in February 2000.
869 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
870 match a (*MARK), and the match failed at the start of the subject, a
871 reference to memory before the start of the subject could occur. This bug
872 was introduced by fix 17 of release 8.21.
874 5. A reference to an unset group with zero minimum repetition was giving
875 totally wrong answers (in non-JavaScript-compatibility mode). For example,
876 /(another)?(\1?)test/ matched against "hello world test". This bug was
877 introduced in release 8.13.
879 6. Add support for 16-bit character strings (a large amount of work involving
880 many changes and refactorings).
882 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
883 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
884 from a file.
886 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
887 rounding is not applied in this particular case).
889 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
890 if they appear, or are escaped, in patterns.
892 10. Get rid of a number of -Wunused-but-set-variable warnings.
894 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
895 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
896 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
897 also returns the mark "x". This bug applied to capturing parentheses,
898 non-capturing parentheses, and atomic parentheses. It also applied to some
899 assertions.
901 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
902 information out of configure.ac instead of relying on pcre.h.generic, which
903 is not stored in the repository.
905 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
906 -lreadline.
908 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
910 15. Applied Graycode's patch to put the top-level frame on the stack rather
911 than the heap when not using the stack for recursion. This gives a
912 performance improvement in many cases when recursion is not deep.
914 16. Experimental code added to "pcretest -C" to output the stack frame size.
917 Version 8.21 12-Dec-2011
918 ------------------------
920 1. Updating the JIT compiler.
922 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
923 are added as well.
925 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
926 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
927 calling _pcre_jit_exec. Some extra comments are added.
929 4. (*MARK) settings inside atomic groups that do not contain any capturing
930 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
931 was introduced by change 18 for 8.20.
933 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
934 ECMA-262 standard.
936 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
937 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
938 This bug was probably introduced by change 9 of 8.13.
940 7. While fixing 6 above, I noticed that a number of other items were being
941 incorrectly rejected as "not fixed length". This arose partly because newer
942 opcodes had not been added to the fixed-length checking code. I have (a)
943 corrected the bug and added tests for these items, and (b) arranged for an
944 error to occur if an unknown opcode is encountered while checking for fixed
945 length instead of just assuming "not fixed length". The items that were
946 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
947 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
948 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
950 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
951 being incorrectly compiled and would have given unpredicatble results.
953 9. A possessively repeated subpattern with minimum repeat count greater than
954 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
955 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
956 the first (A) could occur when it should not.
958 10. Add a cast and remove a redundant test from the code.
960 11. JIT should use pcre_malloc/pcre_free for allocation.
962 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
963 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
964 is anything other than /usr, -L is still shown).
966 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
968 14. Perl does not support \N without a following name in a [] class; PCRE now
969 also gives an error.
971 15. If a forward reference was repeated with an upper limit of around 2000,
972 it caused the error "internal error: overran compiling workspace". The
973 maximum number of forward references (including repeats) was limited by the
974 internal workspace, and dependent on the LINK_SIZE. The code has been
975 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
976 the default depends on LINK_SIZE. There is a new upper limit (for safety)
977 of around 200,000 forward references. While doing this, I also speeded up
978 the filling in of repeated forward references.
980 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
981 incorrectly expecting the subject to contain another "a" after the start.
983 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
984 in the match, the SKIP should be ignored. This was not happening; instead
985 the SKIP was being treated as NOMATCH. For patterns such as
986 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
987 tested.
989 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
990 now much more compatible with Perl, in particular in cases where the result
991 is a non-match for a non-anchored pattern. For example, if
992 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
993 "m", where previously it did not return a name. A side effect of this
994 change is that for partial matches, the last encountered mark name is
995 returned, as for non matches. A number of tests that were previously not
996 Perl-compatible have been moved into the Perl-compatible test files. The
997 refactoring has had the pleasing side effect of removing one argument from
998 the match() function, thus reducing its stack requirements.
1000 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1001 subsequent uses of /S (without +) incorrectly behaved like /S+.
1003 21. Retrieve executable code size support for the JIT compiler and fixing
1004 some warnings.
1006 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1007 not work when the shorter character appeared right at the end of the
1008 subject string.
1010 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1011 systems.
1013 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1014 output it when the /M option is used in pcretest.
1016 25. The CheckMan script was not being included in the distribution. Also, added
1017 an explicit "perl" to run Perl scripts from the PrepareRelease script
1018 because this is reportedly needed in Windows.
1020 26. If study data was being save in a file and studying had not found a set of
1021 "starts with" bytes for the pattern, the data written to the file (though
1022 never used) was taken from uninitialized memory and so caused valgrind to
1023 complain.
1025 27. Updated RunTest.bat as provided by Sheri Pierce.
1027 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1029 29. Computation of memory usage for the table of capturing group names was
1030 giving an unnecessarily large value.
1033 Version 8.20 21-Oct-2011
1034 ------------------------
1036 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1037 a POSIX class. After further experiments with Perl, which convinced me that
1038 Perl has bugs and confusions, a closing square bracket is no longer allowed
1039 in a POSIX name. This bug also affected patterns with classes that started
1040 with full stops.
1042 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1043 captured substring, but while checking the failing first alternative,
1044 substring 1 is temporarily captured. If the output vector supplied to
1045 pcre_exec() was not big enough for this capture, the yield of the function
1046 was still zero ("insufficient space for captured substrings"). This cannot
1047 be totally fixed without adding another stack variable, which seems a lot
1048 of expense for a edge case. However, I have improved the situation in cases
1049 such as /(a)(b)x|abc/ matched against "abc", where the return code
1050 indicates that fewer than the maximum number of slots in the ovector have
1051 been set.
1053 3. Related to (2) above: when there are more back references in a pattern than
1054 slots in the output vector, pcre_exec() uses temporary memory during
1055 matching, and copies in the captures as far as possible afterwards. It was
1056 using the entire output vector, but this conflicts with the specification
1057 that only 2/3 is used for passing back captured substrings. Now it uses
1058 only the first 2/3, for compatibility. This is, of course, another edge
1059 case.
1061 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1062 main code base, and can be used by building with --enable-jit. When this is
1063 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1064 runtime --no-jit option is given.
1066 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1067 ovector, the return from the function was zero, implying that there were
1068 other matches that did not fit. The correct "exactly full" value is now
1069 returned.
1071 6. If a subpattern that was called recursively or as a subroutine contained
1072 (*PRUNE) or any other control that caused it to give a non-standard return,
1073 invalid errors such as "Error -26 (nested recursion at the same subject
1074 position)" or even infinite loops could occur.
1076 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1077 computing the minimum length on reaching *ACCEPT, and so ended up with the
1078 wrong value of 1 rather than 0. Further investigation indicates that
1079 computing a minimum subject length in the presence of *ACCEPT is difficult
1080 (think back references, subroutine calls), and so I have changed the code
1081 so that no minimum is registered for a pattern that contains *ACCEPT.
1083 8. If (*THEN) was present in the first (true) branch of a conditional group,
1084 it was not handled as intended. [But see 16 below.]
1086 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1087 Sheri Pierce.
1089 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1090 the first byte in a match must be "a".
1092 11. Change 17 for 8.13 increased the recursion depth for patterns like
1093 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1094 pattern contains any instances of (*THEN). If it does not, the old
1095 optimizations are restored. It would be nice to do this on a per-group
1096 basis, but at the moment that is not feasible.
1098 12. In some environments, the output of pcretest -C is CRLF terminated. This
1099 broke RunTest's code that checks for the link size. A single white space
1100 character after the value is now allowed for.
1102 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1103 For "fr", it uses the Windows-specific input and output files.
1105 14. If (*THEN) appeared in a group that was called recursively or as a
1106 subroutine, it did not work as intended. [But see next item.]
1108 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1109 pattern fragments (but not containing any | characters). If A and B are
1110 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1111 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1112 D. In other words, Perl considers parentheses that do not contain any |
1113 characters to be part of a surrounding alternative, whereas PCRE was
1114 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1115 differently. PCRE now behaves in the same way as Perl, except in the case
1116 of subroutine/recursion calls such as (?1) which have in any case always
1117 been different (but PCRE had them first :-).
1119 16. Related to 15 above: Perl does not treat the | in a conditional group as
1120 creating alternatives. Such a group is treated in the same way as an
1121 ordinary group without any | characters when processing (*THEN). PCRE has
1122 been changed to match Perl's behaviour.
1124 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1125 RunGrepTest script failed.
1127 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1128 inevitable for groups that contain captures, but it can lead to a lot of
1129 stack use in large patterns. The old behaviour has been restored for atomic
1130 groups that do not contain any capturing parentheses.
1132 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1133 suppress the check for a minimum subject length at run time. (If it was
1134 given to pcre_exec() or pcre_dfa_exec() it did work.)
1136 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1137 fail to work when decoding hex characters in data strings in EBCDIC
1138 environments.
1140 21. It appears that in at least one Mac OS environment, the isxdigit() function
1141 is implemented as a macro that evaluates to its argument more than once,
1142 contravening the C 90 Standard (I haven't checked a later standard). There
1143 was an instance in pcretest which caused it to go wrong when processing
1144 \x{...} escapes in subject strings. The has been rewritten to avoid using
1145 things like p++ in the argument of isxdigit().
1148 Version 8.13 16-Aug-2011
1149 ------------------------
1151 1. The Unicode data tables have been updated to Unicode 6.0.0.
1153 2. Two minor typos in pcre_internal.h have been fixed.
1155 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1156 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1157 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1159 4. There were a number of related bugs in the code for matching backrefences
1160 caselessly in UTF-8 mode when codes for the characters concerned were
1161 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1162 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1163 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1164 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1165 2-byte code at the end of the subject (it thought there wasn't enough data
1166 left).
1168 5. Comprehensive information about what went wrong is now returned by
1169 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1170 as the output vector has at least 2 elements. The offset of the start of
1171 the failing character and a reason code are placed in the vector.
1173 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1174 now returned is for the first byte of the failing character, instead of the
1175 last byte inspected. This is an incompatible change, but I hope it is small
1176 enough not to be a problem. It makes the returned offset consistent with
1177 pcre_exec() and pcre_dfa_exec().
1179 7. pcretest now gives a text phrase as well as the error number when
1180 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1181 failure, the offset and reason code are output.
1183 8. When \R was used with a maximizing quantifier it failed to skip backwards
1184 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1185 back over a single character (\n). This seems wrong (because it treated the
1186 two characters as a single entity when going forwards), conflicts with the
1187 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1188 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1189 has been changed.
1191 9. Some internal refactoring has changed the processing so that the handling
1192 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1193 time (the PCRE_DOTALL option was changed this way some time ago: version
1194 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1195 which was always a bit of a fudge. It also means that there is one less
1196 argument for the match() function, which reduces its stack requirements
1197 slightly. This change also fixes an incompatibility with Perl: the pattern
1198 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1200 10. More internal refactoring has drastically reduced the number of recursive
1201 calls to match() for possessively repeated groups such as (abc)++ when
1202 using pcre_exec().
1204 11. While implementing 10, a number of bugs in the handling of groups were
1205 discovered and fixed:
1207 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1208 (a|)*(?1) gave a compile-time internal error.
1209 ((a|)+)+ did not notice that the outer group could match an empty string.
1210 (^a|^)+ was not marked as anchored.
1211 (.*a|.*)+ was not marked as matching at start or after a newline.
1213 12. Yet more internal refactoring has removed another argument from the match()
1214 function. Special calls to this function are now indicated by setting a
1215 value in a variable in the "match data" data block.
1217 13. Be more explicit in pcre_study() instead of relying on "default" for
1218 opcodes that mean there is no starting character; this means that when new
1219 ones are added and accidentally left out of pcre_study(), testing should
1220 pick them up.
1222 14. The -s option of pcretest has been documented for ages as being an old
1223 synonym of -m (show memory usage). I have changed it to mean "force study
1224 for every regex", that is, assume /S for every regex. This is similar to -i
1225 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1226 using it. It makes it easier to run collections of tests with and without
1227 study enabled, and thereby test pcre_study() more easily. All the standard
1228 tests are now run with and without -s (but some patterns can be marked as
1229 "never study" - see 20 below).
1231 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1232 restoration of the capturing data to the outer values was not happening
1233 correctly.
1235 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1236 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1237 pattern had matched an empty string, and so incorrectly returned a no
1238 match.
1240 17. There was optimizing code for the last branch of non-capturing parentheses,
1241 and also for the obeyed branch of a conditional subexpression, which used
1242 tail recursion to cut down on stack usage. Unfortunately, now that there is
1243 the possibility of (*THEN) occurring in these branches, tail recursion is
1244 no longer possible because the return has to be checked for (*THEN). These
1245 two optimizations have therefore been removed. [But see 8.20/11 above.]
1247 18. If a pattern containing \R was studied, it was assumed that \R always
1248 matched two bytes, thus causing the minimum subject length to be
1249 incorrectly computed because \R can also match just one byte.
1251 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1252 was incorrectly computed.
1254 20. If /S is present twice on a test pattern in pcretest input, it now
1255 *disables* studying, thereby overriding the use of -s on the command line
1256 (see 14 above). This is necessary for one or two tests to keep the output
1257 identical in both cases.
1259 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1260 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1262 22. When an atomic group that contained a capturing parenthesis was
1263 successfully matched, but the branch in which it appeared failed, the
1264 capturing was not being forgotten if a higher numbered group was later
1265 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1266 group 1 to "a", when in fact it should be unset. This applied to multi-
1267 branched capturing and non-capturing groups, repeated or not, and also to
1268 positive assertions (capturing in negative assertions does not happen
1269 in PCRE) and also to nested atomic groups.
1271 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1272 subject after a captured substring, to make it easier to tell which of a
1273 number of identical substrings has been captured.
1275 24. The way atomic groups are processed by pcre_exec() has been changed so that
1276 if they are repeated, backtracking one repetition now resets captured
1277 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1278 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1279 "aaa". Previously, it would have been "a". As part of this code
1280 refactoring, the way recursive calls are handled has also been changed.
1282 25. If an assertion condition captured any substrings, they were not passed
1283 back unless some other capturing happened later. For example, if
1284 (?(?=(a))a) was matched against "a", no capturing was returned.
1286 26. When studying a pattern that contained subroutine calls or assertions,
1287 the code for finding the minimum length of a possible match was handling
1288 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1289 group 1 called group 2 while simultaneously a separate group 2 called group
1290 1). A stack overflow occurred in this case. I have fixed this by limiting
1291 the recursion depth to 10.
1293 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1294 Fortmann. This supports explicit test numbers on the command line, and has
1295 argument validation and error reporting.
1297 28. An instance of \X with an unlimited repeat could fail if at any point the
1298 first character it looked at was a mark character.
1300 29. Some minor code refactoring concerning Unicode properties and scripts
1301 should reduce the stack requirement of match() slightly.
1303 30. Added the '=' option to pcretest to check the setting of unused capturing
1304 slots at the end of the pattern, which are documented as being -1, but are
1305 not included in the return count.
1307 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1308 compiled something random. Now it gives a compile-time error (as does
1309 Perl).
1311 32. A *MARK encountered during the processing of a positive assertion is now
1312 recorded and passed back (compatible with Perl).
1314 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1315 had alternative anchored branches, the search for a second match in a line
1316 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1317 matched the line "0102" twice. The same bug affected patterns that started
1318 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1319 twice.
1321 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1322 does, and because of capturing effects, quantifying parenthesized
1323 assertions may at times be useful. Quantifiers are now allowed for
1324 parenthesized assertions.
1326 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1328 36. \g was being checked for fancy things in a character class, when it should
1329 just be a literal "g".
1331 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1332 appearance of a nested POSIX class supersedes an apparent external class.
1333 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1334 unescaped square brackets may also appear as part of class names. For
1335 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1336 more like Perl. (But see 8.20/1 above.)
1338 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1339 was because it thought it was \N{name}, which is not supported).
1341 39. Add minix to OS list not supporting the -S option in pcretest.
1343 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1344 cannot analyze patterns in sufficient detail to catch mutual recursions
1345 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1346 subgroup is called recursively as a subpattern for a second time at the
1347 same position in the subject string. In previous releases this might have
1348 been caught by the recursion limit, or it might have run out of stack.
1350 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1351 happen only once. PCRE was, however incorrectly giving a compile time error
1352 "recursive call could loop indefinitely" because it cannot analyze the
1353 pattern in sufficient detail. The compile time test no longer happens when
1354 PCRE is compiling a conditional subpattern, but actual runaway loops are
1355 now caught at runtime (see 40 above).
1357 42. It seems that Perl allows any characters other than a closing parenthesis
1358 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1359 has been changed to be the same.
1361 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1362 as not to get warnings when autogen.sh is called. Also changed
1363 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1365 44. To help people who use pcregrep to scan files containing exceedingly long
1366 lines, the following changes have been made:
1368 (a) The default value of the buffer size parameter has been increased from
1369 8K to 20K. (The actual buffer used is three times this size.)
1371 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1372 PCRE is built.
1374 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1375 to be set at run time.
1377 (d) Numerical values in pcregrep options can be followed by K or M, for
1378 example --buffer-size=50K.
1380 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1381 given and the return code is set to 2.
1383 45. Add a pointer to the latest mark to the callout data block.
1385 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1386 partial match of an empty string instead of no match. This was specific to
1387 the use of ".".
1389 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1390 complete match instead of a partial match. This bug was dependent on both
1391 the PCRE_UTF8 and PCRE_DOTALL options being set.
1393 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1394 starting byte set, because \b was not being ignored.
1397 Version 8.12 15-Jan-2011
1398 ------------------------
1400 1. Fixed some typos in the markup of the man pages, and wrote a script that
1401 checks for such things as part of the documentation building process.
1403 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1404 --match-limit and --recursion-limit options (added for 8.11). In
1405 particular, this made one of the standard tests fail. (The integer value
1406 went into the wrong half of a long int.)
1408 3. If the --colour option was given to pcregrep with -v (invert match), it
1409 did strange things, either producing crazy output, or crashing. It should,
1410 of course, ignore a request for colour when reporting lines that do not
1411 match.
1413 4. Another pcregrep bug caused similar problems if --colour was specified with
1414 -M (multiline) and the pattern match finished with a line ending.
1416 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1417 matched in multiline mode, the following line was shown as part of the
1418 match. This seems wrong, so I have changed it.
1420 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1421 the check for further matches in the same line (so they could be coloured)
1422 to overrun the end of the current line. If another match was found, it was
1423 incorrectly shown (and then shown again when found in the next line).
1425 7. If pcregrep was compiled under Windows, there was a reference to the
1426 function pcregrep_exit() before it was defined. I am assuming this was
1427 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1428 reported by a user. I've moved the definition above the reference.
1431 Version 8.11 10-Dec-2010
1432 ------------------------
1434 1. (*THEN) was not working properly if there were untried alternatives prior
1435 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1436 backtracked to try for "b" instead of moving to the next alternative branch
1437 at the same level (in this case, to look for "c"). The Perl documentation
1438 is clear that when (*THEN) is backtracked onto, it goes to the "next
1439 alternative in the innermost enclosing group".
1441 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1442 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1443 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1444 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1445 (*THEN).
1447 3. If \s appeared in a character class, it removed the VT character from
1448 the class, even if it had been included by some previous item, for example
1449 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1450 of \s, but is part of the POSIX "space" class.)
1452 4. A partial match never returns an empty string (because you can always
1453 match an empty string at the end of the subject); however the checking for
1454 an empty string was starting at the "start of match" point. This has been
1455 changed to the "earliest inspected character" point, because the returned
1456 data for a partial match starts at this character. This means that, for
1457 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1458 (previously it gave "no match").
1460 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1461 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1462 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1463 has an implication that the given string is incomplete (because a partial
1464 match is preferred over a full match). For this reason, these items now
1465 give a partial match in this situation. [Aside: previously, the one case
1466 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1467 match rather than a full match, which was wrong by the old rules, but is
1468 now correct.]
1470 6. There was a bug in the handling of #-introduced comments, recognized when
1471 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1472 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1473 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1474 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1475 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1476 places in pcre_compile().
1478 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1479 comments when looking ahead for named forward references to subpatterns,
1480 the only newline sequence it recognized was NL. It now handles newlines
1481 according to the set newline convention.
1483 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1484 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1485 cater for a lack of strerror(). These oversights have been fixed.
1487 9. Added --match-limit and --recursion-limit to pcregrep.
1489 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1491 11. When the -o option was used, pcregrep was setting a return code of 1, even
1492 when matches were found, and --line-buffered was not being honoured.
1494 12. Added an optional parentheses number to the -o and --only-matching options
1495 of pcregrep.
1497 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1498 can match an empty string. The code to do it in pcretest and pcredemo
1499 needed fixing:
1501 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1502 only one byte after an empty string match just before CRLF (this case
1503 just got forgotten; "any" and "anycrlf" were OK).
1505 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1506 mode when an empty string match preceded an ASCII character followed by
1507 a non-ASCII character. (The code for advancing by one character rather
1508 than one byte was nonsense.)
1510 (c) The pcredemo.c sample program did not have any code at all to handle
1511 the cases when CRLF is a valid newline sequence.
1513 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1514 as a starting offset was within the subject string. There is now a new
1515 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1516 negative or greater than the length of the string. In order to test this,
1517 pcretest is extended to allow the setting of negative starting offsets.
1519 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1520 starting offset points to the beginning of a UTF-8 character was
1521 unnecessarily clumsy. I tidied it up.
1523 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1524 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1526 17. Nobody had reported that the --include_dir option, which was added in
1527 release 7.7 should have been called --include-dir (hyphen, not underscore)
1528 for compatibility with GNU grep. I have changed it to --include-dir, but
1529 left --include_dir as an undocumented synonym, and the same for
1530 --exclude-dir, though that is not available in GNU grep, at least as of
1531 release 2.5.4.
1533 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1534 characters from a string of bytes) have been redefined so as not to use
1535 loops, in order to improve performance in some environments. At the same
1536 time, I abstracted some of the common code into auxiliary macros to save
1537 repetition (this should not affect the compiled code).
1539 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1540 compile-time error is now given if \c is not followed by an ASCII
1541 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1542 different, and any byte value is allowed.)
1544 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1545 START_OPTIMIZE option, which is now allowed at compile time - but just
1546 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1547 to pcregrep and other applications that have no direct access to PCRE
1548 options. The new /Y option in pcretest sets this option when calling
1549 pcre_compile().
1551 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1552 back references. Groups containing recursive back references were forced to
1553 be atomic by that change, but in the case of named groups, the amount of
1554 memory required was incorrectly computed, leading to "Failed: internal
1555 error: code overflow". This has been fixed.
1557 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1558 pcretest.c, to avoid build problems in some Borland environments.
1561 Version 8.10 25-Jun-2010
1562 ------------------------
1564 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1565 THEN.
1567 2. (*ACCEPT) was not working when inside an atomic group.
1569 3. Inside a character class, \B is treated as a literal by default, but
1570 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1571 causes the error). The code is unchanged, but I tidied the documentation.
1573 4. Inside a character class, PCRE always treated \R and \X as literals,
1574 whereas Perl faults them if its -w option is set. I have changed PCRE so
1575 that it faults them when PCRE_EXTRA is set.
1577 5. Added support for \N, which always matches any character other than
1578 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1580 6. When compiling pcregrep with newer versions of gcc which may have
1581 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1582 declared with attribute warn_unused_result" were given. Just casting the
1583 result to (void) does not stop the warnings; a more elaborate fudge is
1584 needed. I've used a macro to implement this.
1586 7. Minor change to pcretest.c to avoid a compiler warning.
1588 8. Added four artifical Unicode properties to help with an option to make
1589 \s etc use properties (see next item). The new properties are: Xan
1590 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1592 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1593 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1594 this option. Modified pcretest to add /W to test this facility. Added
1595 REG_UCP to make it available via the POSIX interface.
1597 10. Added --line-buffered to pcregrep.
1599 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1600 studied, and the match started with a letter with a code point greater than
1601 127 whose first byte was different to the first byte of the other case of
1602 the letter, the other case of this starting letter was not recognized
1603 (#976).
1605 12. If a pattern that was studied started with a repeated Unicode property
1606 test, for example, \p{Nd}+, there was the theoretical possibility of
1607 setting up an incorrect bitmap of starting bytes, but fortunately it could
1608 not have actually happened in practice until change 8 above was made (it
1609 added property types that matched character-matching opcodes).
1611 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1612 possible starting bytes for non-anchored patterns.
1614 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1615 \R, and also a number of cases that involve Unicode properties, both
1616 explicit and implicit when PCRE_UCP is set.
1618 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1619 input, it could crash or give wrong results if characters with values
1620 greater than 0xc0 were present in the subject string. (Detail: it assumed
1621 UTF-8 input when processing these items.)
1623 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1624 size_t is 64-bit (#991).
1626 17. Added a check for running out of memory when PCRE is compiled with
1627 --disable-stack-for-recursion (#990).
1629 18. If the last data line in a file for pcretest does not have a newline on
1630 the end, a newline was missing in the output.
1632 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1633 less than 128) in its various bitmaps. However, there is a facility for
1634 generating tables according to the current locale when PCRE is compiled. It
1635 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1636 characters, are recognized by isspace() and therefore were getting set in
1637 these tables, and indeed these tables seem to approximate to ISO 8859. This
1638 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1639 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1640 which of course cannot start UTF-8 characters. I have changed the code so
1641 that only real ASCII characters (less than 128) and the correct starting
1642 bytes for UTF-8 encodings are set for characters greater than 127 when in
1643 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1644 altogether.)
1646 20. Added the /T option to pcretest so as to be able to run tests with non-
1647 standard character tables, thus making it possible to include the tests
1648 used for 19 above in the standard set of tests.
1650 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1651 reference to a subpattern the other side of a comment that contains an
1652 opening parenthesis caused either an internal compiling error, or a
1653 reference to the wrong subpattern.
1656 Version 8.02 19-Mar-2010
1657 ------------------------
1659 1. The Unicode data tables have been updated to Unicode 5.2.0.
1661 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1662 configured.
1664 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1665 original author of that file, following a query about its status.
1667 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1668 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1670 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1671 quantifier applied to a forward-referencing subroutine call, could compile
1672 incorrect code or give the error "internal error: previously-checked
1673 referenced subpattern not found".
1675 6. Both MS Visual Studio and Symbian OS have problems with initializing
1676 variables to point to external functions. For these systems, therefore,
1677 pcre_malloc etc. are now initialized to local functions that call the
1678 relevant global functions.
1680 7. There were two entries missing in the vectors called coptable and poptable
1681 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1682 I've fixed the data, and added a kludgy way of testing at compile time that
1683 the lengths are correct (equal to the number of opcodes).
1685 8. Following on from 7, I added a similar kludge to check the length of the
1686 eint vector in pcreposix.c.
1688 9. Error texts for pcre_compile() are held as one long string to avoid too
1689 much relocation at load time. To find a text, the string is searched,
1690 counting zeros. There was no check for running off the end of the string,
1691 which could happen if a new error number was added without updating the
1692 string.
1694 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1696 11. \K was not working if it appeared in an atomic group or in a group that
1697 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1698 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1699 the assertion is positive, but not if it is negative.
1701 12. Change 11 fortuitously reduced the size of the stack frame used in the
1702 "match()" function of pcre_exec.c by one pointer. Forthcoming
1703 implementation of support for (*MARK) will need an extra pointer on the
1704 stack; I have reserved it now, so that the stack frame size does not
1705 decrease.
1707 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1708 item in branch that calls a recursion is a subroutine call - as in the
1709 second branch in the above example - was incorrectly given the compile-
1710 time error "recursive call could loop indefinitely" because pcre_compile()
1711 was not correctly checking the subroutine for matching a non-empty string.
1713 14. The checks for overrunning compiling workspace could trigger after an
1714 overrun had occurred. This is a "should never occur" error, but it can be
1715 triggered by pathological patterns such as hundreds of nested parentheses.
1716 The checks now trigger 100 bytes before the end of the workspace.
1718 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1721 Version 8.01 19-Jan-2010
1722 ------------------------
1724 1. If a pattern contained a conditional subpattern with only one branch (in
1725 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1726 computed the wrong minimum data length (which is of course zero for such
1727 subpatterns). This could cause incorrect "no match" results.
1729 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1730 the pattern is reset in the first branch, pcre_compile() failed with
1731 "internal error: code overflow at offset...". This happened only when
1732 the reset was to the original external option setting. (An optimization
1733 abstracts leading options settings into an external setting, which was the
1734 cause of this.)
1736 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1737 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1738 assertion pattern did not match (meaning that the assertion was true), it
1739 was incorrectly treated as false if the SKIP had been reached during the
1740 matching. This also applied to assertions used as conditions.
1742 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1743 assertion subpattern, including such a pattern used as a condition,
1744 unpredictable results occurred, instead of the error return
1747 5. The C++ GlobalReplace function was not working like Perl for the special
1748 situation when an empty string is matched. It now does the fancy magic
1749 stuff that is necessary.
1751 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1752 removed. (These were left over from very, very early versions of PCRE.)
1754 7. Some cosmetic changes to the code to make life easier when compiling it
1755 as part of something else:
1757 (a) Change DEBUG to PCRE_DEBUG.
1759 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1760 called "current" as "current_branch", to prevent a collision with the
1761 Linux macro when compiled as a kernel module.
1763 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1764 prevent a collision with the Linux macro when compiled as a kernel
1765 module.
1767 8. In pcre_compile() there are some checks for integer overflows that used to
1768 cast potentially large values to (double). This has been changed to that
1769 when building, a check for int64_t is made, and if it is found, it is used
1770 instead, thus avoiding the use of floating point arithmetic. (There is no
1771 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1772 double.
1774 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1775 2005 (difference between two addresses compared to an unsigned value).
1777 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1778 custom one, because of the following reported problem in Windows:
1780 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1781 under Win32.
1782 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1783 therefore missing the function definition.
1784 - The compiler thus generates a "C" signature for the test function.
1785 - The linker fails to find the "C" function.
1786 - PCRE fails to configure if asked to do so against libbz2.
1788 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1789 messages were output:
1791 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1792 rerunning libtoolize, to keep the correct libtool macros in-tree.
1793 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1795 I have done both of these things.
1797 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1798 most of the time, it *can* run out if it is given a pattern that contains a
1799 runaway infinite recursion. I updated the discussion in the pcrestack man
1800 page.
1802 13. Now that we have gone to the x.xx style of version numbers, the minor
1803 version may start with zero. Using 08 or 09 is a bad idea because users
1804 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1805 interpreted as invalid octal numbers. I've updated the previous comment in
1806 configure.ac, and also added a check that gives an error if 08 or 09 are
1807 used.
1809 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1810 causing partial matching to fail when the end of the subject matched \W
1811 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1813 15. There were some discrepancies between the declarations in pcre_internal.h
1814 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1815 their definitions. The declarations used "const uschar *" and the
1816 definitions used USPTR. Even though USPTR is normally defined as "const
1817 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1818 reported that: "This difference in casting confuses some C++ compilers, for
1819 example, SunCC recognizes above declarations as different functions and
1820 generates broken code for hbpcre." I have changed the declarations to use
1821 USPTR.
1823 16. GNU libtool is named differently on some systems. The autogen.sh script now
1824 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1825 (FreeBSD).
1827 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1828 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1829 comment: "Figure out how to create a longlong from a string: strtoll and
1830 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1831 instance, but it only takes 2 args instead of 3!"
1833 18. A subtle bug concerned with back references has been fixed by a change of
1834 specification, with a corresponding code fix. A pattern such as
1835 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1836 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1837 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1838 same bug. Such groups have to be quantified to be useful, or contained
1839 inside another quantified group. (If there's no repetition, the reference
1840 can never match.) The problem arises because, having left the group and
1841 moved on to the rest of the pattern, a later failure that backtracks into
1842 the group uses the captured value from the final iteration of the group
1843 rather than the correct earlier one. I have fixed this in PCRE by forcing
1844 any group that contains a reference to itself to be an atomic group; that
1845 is, there cannot be any backtracking into it once it has completed. This is
1846 similar to recursive and subroutine calls.
1849 Version 8.00 19-Oct-09
1850 ----------------------
1852 1. The table for translating pcre_compile() error codes into POSIX error codes
1853 was out-of-date, and there was no check on the pcre_compile() error code
1854 being within the table. This could lead to an OK return being given in
1855 error.
1857 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1858 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1859 in a Windows environment.
1861 3. The pcregrep --count option prints the count for each file even when it is
1862 zero, as does GNU grep. However, pcregrep was also printing all files when
1863 --files-with-matches was added. Now, when both options are given, it prints
1864 counts only for those files that have at least one match. (GNU grep just
1865 prints the file name in this circumstance, but including the count seems
1866 more useful - otherwise, why use --count?) Also ensured that the
1867 combination -clh just lists non-zero counts, with no names.
1869 4. The long form of the pcregrep -F option was incorrectly implemented as
1870 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1871 but it seems right to fix it, and I didn't think it was worth preserving
1872 the old behaviour.
1874 5. The command line items --regex=pattern and --regexp=pattern were not
1875 recognized by pcregrep, which required --regex pattern or --regexp pattern
1876 (with a space rather than an '='). The man page documented the '=' forms,
1877 which are compatible with GNU grep; these now work.
1879 6. No libpcreposix.pc file was created for pkg-config; there was just
1880 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1882 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1883 when UCP support is not needed, by modifying the Python script that
1884 generates it from Unicode data files. This should not matter if the module
1885 is correctly used as a library, but I received one complaint about 50K of
1886 unwanted data. My guess is that the person linked everything into his
1887 program rather than using a library. Anyway, it does no harm.
1889 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1890 was a minimum greater than 1 for a wide character in a possessive
1891 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1892 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1893 character. Chaos in the form of incorrect output or a compiling loop could
1894 result.
1896 9. The restrictions on what a pattern can contain when partial matching is
1897 requested for pcre_exec() have been removed. All patterns can now be
1898 partially matched by this function. In addition, if there are at least two
1899 slots in the offset vector, the offset of the earliest inspected character
1900 for the match and the offset of the end of the subject are set in them when
1901 PCRE_ERROR_PARTIAL is returned.
1903 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1904 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1905 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1906 and may be more useful for multi-segment matching.
1908 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1909 used to be given if ever the end of the subject was reached; now it is
1910 given only if matching could not proceed because another character was
1911 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1912 string "Z", which now yields "no match" instead of "partial match". In the
1913 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1914 final character ended with (*FAIL).
1916 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1917 if the pattern had a "must contain" character that was already found in the
1918 earlier partial match, unless partial matching was again requested. For
1919 example, with the pattern /dog.(body)?/, the "must contain" character is
1920 "g". If the first part-match was for the string "dog", restarting with
1921 "sbody" failed. This bug has been fixed.
1923 13. The string returned by pcre_dfa_exec() after a partial match has been
1924 changed so that it starts at the first inspected character rather than the
1925 first character of the match. This makes a difference only if the pattern
1926 starts with a lookbehind assertion or \b or \B (\K is not supported by
1927 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1928 matching functions compatible, and I think it's the right thing to do.
1930 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1931 so that the demonstration program is easily available in environments where
1932 PCRE has not been installed from source.
1934 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1935 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1936 library.
1938 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1939 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1940 is not the first non-POSIX option to be added. Clearly some people find
1941 these options useful.
1943 17. If a caller to the POSIX matching function regexec() passes a non-zero
1944 value for nmatch with a NULL value for pmatch, the value of
1945 nmatch is forced to zero.
1947 18. RunGrepTest did not have a test for the availability of the -u option of
1948 the diff command, as RunTest does. It now checks in the same way as
1949 RunTest, and also checks for the -b option.
1951 19. If an odd number of negated classes containing just a single character
1952 interposed, within parentheses, between a forward reference to a named
1953 subpattern and the definition of the subpattern, compilation crashed with
1954 an internal error, complaining that it could not find the referenced
1955 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1956 [The bug was that it was starting one character too far in when skipping
1957 over the character class, thus treating the ] as data rather than
1958 terminating the class. This meant it could skip too much.]
1960 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1961 /g option in pcretest when the pattern contains \K, which makes it possible
1962 to have an empty string match not at the start, even when the pattern is
1963 anchored. Updated pcretest and pcredemo to use this option.
1965 21. If the maximum number of capturing subpatterns in a recursion was greater
1966 than the maximum at the outer level, the higher number was returned, but
1967 with unset values at the outer level. The correct (outer level) value is
1968 now given.
1970 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1971 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1972 make it do so. The string so far is captured, making this feature
1973 compatible with Perl.
1975 23. The tests have been re-organized, adding tests 11 and 12, to make it
1976 possible to check the Perl 5.10 features against Perl 5.10.
1978 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1979 pattern matches a fixed length string. PCRE did not allow this; now it
1980 does. Neither allows recursion.
1982 25. I finally figured out how to implement a request to provide the minimum
1983 length of subject string that was needed in order to match a given pattern.
1984 (It was back references and recursion that I had previously got hung up
1985 on.) This code has now been added to pcre_study(); it finds a lower bound
1986 to the length of subject needed. It is not necessarily the greatest lower
1987 bound, but using it to avoid searching strings that are too short does give
1988 some useful speed-ups. The value is available to calling programs via
1989 pcre_fullinfo().
1991 26. While implementing 25, I discovered to my embarrassment that pcretest had
1992 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1993 study optimizations had never been tested with that matching function.
1994 Oops. What is worse, even when it was passed study data, there was a bug in
1995 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1996 were also very few tests of studied patterns with pcre_dfa_exec().
1998 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1999 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2000 on the other side of the coin, they are no longer allowed to have different
2001 names, because these cannot be distinguished in PCRE, and this has caused
2002 confusion. (This is a difference from Perl.)
2004 28. When duplicate subpattern names are present (necessarily with different
2005 numbers, as required by 27 above), and a test is made by name in a
2006 conditional pattern, either for a subpattern having been matched, or for
2007 recursion in such a pattern, all the associated numbered subpatterns are
2008 tested, and the overall condition is true if the condition is true for any
2009 one of them. This is the way Perl works, and is also more like the way
2010 testing by number works.
2013 Version 7.9 11-Apr-09
2014 ---------------------
2016 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2017 (pcretest), all targets were linked against these libraries. This included
2018 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2019 libraries. This caused unwanted dependencies to be created. This problem
2020 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2021 pcretest is linked with readline.
2023 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2024 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2025 moved outside it again, because FALSE and TRUE are already defined in AIX,
2026 but BOOL is not.
2028 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2029 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2031 4. The pcregrep documentation said spaces were inserted as well as colons (or
2032 hyphens) following file names and line numbers when outputting matching
2033 lines. This is not true; no spaces are inserted. I have also clarified the
2034 wording for the --colour (or --color) option.
2036 5. In pcregrep, when --colour was used with -o, the list of matching strings
2037 was not coloured; this is different to GNU grep, so I have changed it to be
2038 the same.
2040 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2041 each matching line was coloured. Now it goes on to look for further matches
2042 of any of the test patterns, which is the same behaviour as GNU grep.
2044 7. A pattern that could match an empty string could cause pcregrep to loop; it
2045 doesn't make sense to accept an empty string match in pcregrep, so I have
2046 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2047 seems to be how GNU grep behaves. [But see later change 40 for release
2048 8.33.]
2050 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2051 start or after a newline", because the conditional assertion was not being
2052 correctly handled. The rule now is that both the assertion and what follows
2053 in the first alternative must satisfy the test.
2055 9. If auto-callout was enabled in a pattern with a conditional group whose
2056 condition was an assertion, PCRE could crash during matching, both with
2057 pcre_exec() and pcre_dfa_exec().
2059 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2060 used for matching.
2062 11. Unicode property support in character classes was not working for
2063 characters (bytes) greater than 127 when not in UTF-8 mode.
2065 12. Added the -M command line option to pcretest.
2067 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2069 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2071 16. Added comments and documentation about mis-use of no_arg in the C++
2072 wrapper.
2074 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2075 from Martin Jerabek that uses macro names for all relevant character and
2076 string constants.
2078 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2079 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2080 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2081 these, but not everybody uses configure.
2083 19. A conditional group that had only one branch was not being correctly
2084 recognized as an item that could match an empty string. This meant that an
2085 enclosing group might also not be so recognized, causing infinite looping
2086 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2087 with the subject "ab", where knowledge that the repeated group can match
2088 nothing is needed in order to break the loop.
2090 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2091 exec(), but without supplying a callout function, matching went wrong.
2093 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2094 leak if the size of the offset vector was greater than 30. When the vector
2095 is smaller, the saved offsets during recursion go onto a local stack
2096 vector, but for larger vectors malloc() is used. It was failing to free
2097 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2098 error, in fact).
2100 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2101 heapframe that is used only when UTF-8 support is enabled. This caused no
2102 problem, but was untidy.
2104 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2105 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2106 included within another project.
2108 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2109 slightly modified by me:
2111 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2112 not building pcregrep.
2114 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2115 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2117 25. Forward references, both numeric and by name, in patterns that made use of
2118 duplicate group numbers, could behave incorrectly or give incorrect errors,
2119 because when scanning forward to find the reference group, PCRE was not
2120 taking into account the duplicate group numbers. A pattern such as
2121 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2123 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2124 the feature of a custom pointer more persuasive (as requested by a user).
2126 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2127 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2128 pre-defined.
2130 28. Added support for (*UTF8) at the start of a pattern.
2132 29. Arrange for flags added by the "release type" setting in CMake to be shown
2133 in the configuration summary.
2136 Version 7.8 05-Sep-08
2137 ---------------------
2139 1. Replaced UCP searching code with optimized version as implemented for Ad
2140 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2141 stage table and inline lookup instead of a function, giving speed ups of 2
2142 to 5 times on some simple patterns that I tested. Permission was given to
2143 distribute the MultiStage2.py script that generates the tables (it's not in
2144 the tarball, but is in the Subversion repository).
2146 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2147 scripts.
2149 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2150 a group with a zero qualifier. The result of the study could be incorrect,
2151 or the function might crash, depending on the pattern.
2153 4. Caseless matching was not working for non-ASCII characters in back
2154 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2155 It now works when Unicode Property Support is available.
2157 5. In pcretest, an escape such as \x{de} in the data was always generating
2158 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2159 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2160 truncation.
2162 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2164 7. Added two (int) casts to pcregrep when printing the difference of two
2165 pointers, in case they are 64-bit values.
2167 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2168 test 2 if it fails.
2170 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2171 and a #define of that name to empty if it is not externally set. This is to
2172 allow users of MSVC to set it if necessary.
2174 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2175 the convenience functions in the pcre_get.c source file.
2177 11. An option change at the start of a pattern that had top-level alternatives
2178 could cause overwriting and/or a crash. This command provoked a crash in
2179 some environments:
2181 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2183 This potential security problem was recorded as CVE-2008-2371.
2185 12. For a pattern where the match had to start at the beginning or immediately
2186 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2187 pcre_dfa_exec() could read past the end of the passed subject if there was
2188 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2189 pcretest so that it places the subject at the end of its malloc-ed buffer.
2191 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2192 exec() might read past the end of the data buffer in UTF-8 mode.
2194 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2195 the data contained the byte 0x85 as part of a UTF-8 character within its
2196 first line. This applied both to normal and DFA matching.
2198 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2199 /^[^d]*?$/8 failed to match "abc".
2201 16. Added a missing copyright notice to pcrecpp_internal.h.
2203 17. Make it more clear in the documentation that values returned from
2204 pcre_exec() in ovector are byte offsets, not character counts.
2206 18. Tidied a few places to stop certain compilers from issuing warnings.
2208 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2209 supplied by Stefan Weber. I made a further small update for 7.8 because
2210 there is a change of source arrangements: the pcre_searchfuncs.c module is
2211 replaced by pcre_ucd.c.
2214 Version 7.7 07-May-08
2215 ---------------------
2217 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2218 a string to a long long, pretend we don't even have a long long." This is
2219 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2221 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2222 pre-7.6 versions, which defined a global no_arg variable instead of putting
2223 it in the RE class. (See also #8 below.)
2225 3. Remove a line of dead code, identified by coverity and reported by Nuno
2226 Lopes.
2228 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2230 (1) The include/exclude patterns were being applied to the whole pathnames
2231 of files, instead of just to the final components.
2233 (2) If there was more than one level of directory, the subdirectories were
2234 skipped unless they satisfied the include/exclude conditions. This is
2235 inconsistent with GNU grep (and could even be seen as contrary to the
2236 pcregrep specification - which I improved to make it absolutely clear).
2237 The action now is always to scan all levels of directory, and just
2238 apply the include/exclude patterns to regular files.
2240 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2241 --exclude_dir in the tests to avoid scanning .svn directories.
2243 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2244 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2245 doesn't support NULs in patterns.
2247 7. Added some missing "const"s to declarations of static tables in
2248 pcre_compile.c and pcre_dfa_exec.c.
2250 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2251 caused by fix #2 above. (Subsequently also a second patch to fix the
2252 first patch. And a third patch - this was a messy problem.)
2254 9. Applied Craig's patch to remove the use of push_back().
2256 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2257 matching function regexec().
2259 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2260 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2261 references. PCRE supports relative numbers with this syntax (I don't think
2262 Oniguruma does).
2264 12. Previously, a group with a zero repeat such as (...){0} was completely
2265 omitted from the compiled regex. However, this means that if the group
2266 was called as a subroutine from elsewhere in the pattern, things went wrong
2267 (an internal error was given). Such groups are now left in the compiled
2268 pattern, with a new opcode that causes them to be skipped at execution
2269 time.
2271 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2272 to the way PCRE behaves:
2274 (a) A lone ] character is dis-allowed (Perl treats it as data).
2276 (b) A back reference to an unmatched subpattern matches an empty string
2277 (Perl fails the current match path).
2279 (c) A data ] in a character class must be notated as \] because if the
2280 first data character in a class is ], it defines an empty class. (In
2281 Perl it is not possible to have an empty class.) The empty class []
2282 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2283 The negative empty class [^] matches any one character, independently
2284 of the DOTALL setting.
2286 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2287 non-existent subpattern following a character class starting with ']' and
2288 containing () gave an internal compiling error instead of "reference to
2289 non-existent subpattern". Fortunately, when the pattern did exist, the
2290 compiled code was correct. (When scanning forwards to check for the
2291 existence of the subpattern, it was treating the data ']' as terminating
2292 the class, so got the count wrong. When actually compiling, the reference
2293 was subsequently set up correctly.)
2295 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2296 it was being rejected as not supported by pcre_dfa_exec(), even though
2297 other assertions are supported. I have made pcre_dfa_exec() support
2298 (*FAIL).
2300 16. The implementation of 13c above involved the invention of a new opcode,
2301 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2302 cannot be changed at match time, I realized I could make a small
2303 improvement to matching performance by compiling OP_ALLANY instead of
2304 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2305 on the OP_ANY path.
2307 17. Compiling pcretest on Windows with readline support failed without the
2308 following two fixes: (1) Make the unistd.h include conditional on
2309 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2311 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2312 ncurses library to be included for pcretest when ReadLine support is
2313 requested, but also to allow for it to be overridden. This patch came from
2314 Daniel Bergström.
2316 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2317 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2318 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2319 spotting this.
2322 Version 7.6 28-Jan-08
2323 ---------------------
2325 1. A character class containing a very large number of characters with
2326 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2327 overflow.
2329 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2330 HAVE_LONG_LONG is not defined.
2332 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2333 bring them up to date and include new features. This patch includes:
2335 - Fixed PH's badly added libz and libbz2 support.
2336 - Fixed a problem with static linking.
2337 - Added pcredemo. [But later removed - see 7 below.]
2338 - Fixed dftables problem and added an option.
2339 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2341 - Added readline support for pcretest.
2342 - Added an listing of the option settings after cmake has run.
2344 4. A user submitted a patch to Makefile that makes it easy to create
2345 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2346 Makefile.am that cause it to include this special target, without
2347 affecting anything else. Note that the same mingw target plus all
2348 the other distribution libraries and programs are now supported
2349 when configuring with CMake (see 6 below) instead of with
2350 Configure/Make.
2352 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2353 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2354 exported in the Windows port". It has not yet been confirmed that the patch
2355 solves the problem, but it does no harm.
2357 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2358 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2359 with CMake, and also correct the comment about stack recursion.
2361 7. Remove the automatic building of pcredemo from the ./configure system and
2362 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2363 of a program that users should build themselves after PCRE is installed, so
2364 building it automatically is not really right. What is more, it gave
2365 trouble in some build environments.
2367 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2370 Version 7.5 10-Jan-08
2371 ---------------------
2373 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2374 values in parens when parsing an RE using the C++ wrapper."
2376 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2377 Characters greater than 255 were excluded from the class instead of being
2378 included.
2380 3. The same bug as (2) above applied to negated POSIX classes such as
2381 [:^space:].
2383 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2384 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2385 I have changed it.
2387 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2388 first named subpattern) and a construct such as (?&a) would reference the
2389 first named subpattern whose name started with "a" (in other words, the
2390 length check was missing). Both these problems are fixed. "Subpattern name
2391 expected" is now given for (?&) (a zero-length name), and this patch also
2392 makes it give the same error for \k'' (previously it complained that that
2393 was a reference to a non-existent subpattern).
2395 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2396 this is right because (?- can be followed by option settings as well as by
2397 digits. I have, however, made the messages clearer.
2399 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2400 than the number used in the conditional) now cause a compile-time error.
2401 This is actually not compatible with Perl, which accepts such patterns, but
2402 treats the conditional as always being FALSE (as PCRE used to), but it
2403 seems to me that giving a diagnostic is better.
2405 8. Change "alphameric" to the more common word "alphanumeric" in comments
2406 and messages.
2408 9. Fix two occurrences of "backslash" in comments that should have been
2409 "backspace".
2411 10. Remove two redundant lines of code that can never be obeyed (their function
2412 was moved elsewhere).
2414 11. The program that makes PCRE's Unicode character property table had a bug
2415 which caused it to generate incorrect table entries for sequences of
2416 characters that have the same character type, but are in different scripts.
2417 It amalgamated them into a single range, with the script of the first of
2418 them. In other words, some characters were in the wrong script. There were
2419 thirteen such cases, affecting characters in the following ranges:
2421 U+002b0 - U+002c1
2422 U+0060c - U+0060d
2423 U+0061e - U+00612
2424 U+0064b - U+0065e
2425 U+0074d - U+0076d
2426 U+01800 - U+01805
2427 U+01d00 - U+01d77
2428 U+01d9b - U+01dbf
2429 U+0200b - U+0200f
2430 U+030fc - U+030fe
2431 U+03260 - U+0327f
2432 U+0fb46 - U+0fbb1
2433 U+10450 - U+1049d
2435 12. The -o option (show only the matching part of a line) for pcregrep was not
2436 compatible with GNU grep in that, if there was more than one match in a
2437 line, it showed only the first of them. It now behaves in the same way as
2438 GNU grep.
2440 13. If the -o and -v options were combined for pcregrep, it printed a blank
2441 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2442 does the same. The return code can be used to tell if there were any
2443 non-matching lines.
2445 14. Added --file-offsets and --line-offsets to pcregrep.
2447 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2448 infinitely looping recursion. The bug was that positive lookaheads were not
2449 being skipped when checking for a possible empty match (negative lookaheads
2450 and both kinds of lookbehind were skipped).
2452 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2453 inclusion of <windows.h> to before rather than after the definition of
2454 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2456 17. Specifying a possessive quantifier with a specific limit for a Unicode
2457 character property caused pcre_compile() to compile bad code, which led at
2458 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2459 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2460 caused the error; without that there was no problem.
2462 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2464 19. Added --enable-pcretest-libreadline.
2466 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2467 RE::GlobalReplace(). As a result, the number of replacements returned was
2468 double what it should be. I removed one of the increments, but Craig sent a
2469 later patch that removed the other one (the right fix) and added unit tests
2470 that check the return values (which was not done before).
2472 21. Several CMake things:
2474 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2475 the names libpcre and libpcreposix, not just pcre and pcreposix.
2477 (2) The above change means that pcretest and pcregrep are now correctly
2478 linked with the newly-built libraries, not previously installed ones.
2482 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2483 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2484 UTF-8 newline character). The key issue is that the pattern starts .*;
2485 this means that the match must be either at the beginning, or after a
2486 newline. The bug was in the code for advancing after a failed match and
2487 checking that the new position followed a newline. It was not taking
2488 account of UTF-8 characters correctly.
2490 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2491 character classes. PCRE was not treating the sequence [:...:] as a
2492 character class unless the ... were all letters. Perl, however, seems to
2493 allow any characters between [: and :], though of course it rejects as
2494 unknown any "names" that contain non-letters, because all the known class
2495 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2496 for example, whereas PCRE did not - it did not recognize a POSIX character
2497 class. This seemed a bit dangerous, so the code has been changed to be
2498 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2499 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2500 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2501 Perl does, and where it didn't before.
2503 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2504 Windows environments %n is disabled by default.
2507 Version 7.4 21-Sep-07
2508 ---------------------
2510 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2511 means that a class such as [\s] counted as "explicit reference to CR or
2512 LF". That isn't really right - the whole point of the change was to try to
2513 help when there was an actual mention of one of the two characters. So now
2514 the change happens only if \r or \n (or a literal CR or LF) character is
2515 encountered.
2517 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2518 of both had grown to the point where there were only 3 bits left.
2519 Fortunately, there was spare space in the data structure, and so I have
2520 moved the internal flags into a new 16-bit field to free up more option
2521 bits.
2523 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2524 but did not set the internal JCHANGED flag - either of these is enough to
2525 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2526 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2527 start sets both bits.
2529 4. Added options (at build time, compile time, exec time) to change \R from
2530 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2532 5. doc/pcresyntax.html was missing from the distribution.
2534 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2535 compatibility, even though it is no longer used.
2537 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2538 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2539 windows.h file is present (where different names are used). [This was
2540 reversed later after testing - see 16 below.]
2542 8. Changed all #include <config.h> to #include "config.h". There were also
2543 some further <pcre.h> cases that I changed to "pcre.h".
2545 9. When pcregrep was used with the --colour option, it missed the line ending
2546 sequence off the lines that it output.
2548 10. It was pointed out to me that arrays of string pointers cause lots of
2549 relocations when a shared library is dynamically loaded. A technique of
2550 using a single long string with a table of offsets can drastically reduce
2551 these. I have refactored PCRE in four places to do this. The result is
2552 dramatic:
2554 Originally: 290
2555 After changing UCP table: 187
2556 After changing error message table: 43
2557 After changing table of "verbs" 36
2558 After changing table of Posix names 22
2560 Thanks to the folks working on Gregex for glib for this insight.
2562 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2563 unicode-properties was also set.
2565 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2567 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2568 checked only for CRLF.
2570 14. Added casts to pcretest.c to avoid compiler warnings.
2572 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2574 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2575 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2576 entirely. This removes changes made in 7 above.
2578 17. The CMake files have been updated, and there is now more information about
2579 building with CMake in the NON-UNIX-USE document.
2582 Version 7.3 28-Aug-07
2583 ---------------------
2585 1. In the rejigging of the build system that eventually resulted in 7.1, the
2586 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2587 brackets there is not right, since it causes compilers to look for an
2588 installed pcre.h, not the version that is in the source that is being
2589 compiled (which of course may be different). I have changed it back to:
2591 #include "pcre.h"
2593 I have a vague recollection that the change was concerned with compiling in
2594 different directories, but in the new build system, that is taken care of
2595 by the VPATH setting the Makefile.
2597 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2598 when the subject happened to end in the byte 0x85 (e.g. if the last
2599 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2600 characters but of course it shouldn't be taken as a newline when it is part
2601 of another character. The bug was that, for an unlimited repeat of . in
2602 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2603 characters when looking for a newline.
2605 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2607 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2608 in debug output.
2610 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2611 long printing in the pcrecpp unittest when running under MinGW.
2613 6. ESC_K was left out of the EBCDIC table.
2615 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2616 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2617 limit also applies to "virtual nesting" when a pattern is recursive, and in
2618 this case 1000 isn't so big. I have been able to remove this limit at the
2619 expense of backing off one optimization in certain circumstances. Normally,
2620 when pcre_exec() would call its internal match() function recursively and
2621 immediately return the result unconditionally, it uses a "tail recursion"
2622 feature to save stack. However, when a subpattern that can match an empty
2623 string has an unlimited repetition quantifier, it no longer makes this
2624 optimization. That gives it a stack frame in which to save the data for
2625 checking that an empty string has been matched. Previously this was taken
2626 from the 1000-entry workspace that had been reserved. So now there is no
2627 explicit limit, but more stack is used.
2629 8. Applied Daniel's patches to solve problems with the import/export magic
2630 syntax that is required for Windows, and which was going wrong for the
2631 pcreposix and pcrecpp parts of the library. These were overlooked when this
2632 problem was solved for the main library.
2634 9. There were some crude static tests to avoid integer overflow when computing
2635 the size of patterns that contain repeated groups with explicit upper
2636 limits. As the maximum quantifier is 65535, the maximum group length was
2637 set at 30,000 so that the product of these two numbers did not overflow a
2638 32-bit integer. However, it turns out that people want to use groups that
2639 are longer than 30,000 bytes (though not repeat them that many times).
2640 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2641 made it possible to implement the integer overflow checks in a much more
2642 dynamic way, which I have now done. The artificial limitation on group
2643 length has been removed - we now have only the limit on the total length of
2644 the compiled pattern, which depends on the LINK_SIZE setting.
2646 10. Fixed a bug in the documentation for get/copy named substring when
2647 duplicate names are permitted. If none of the named substrings are set, the
2648 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2649 empty string.
2651 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2652 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2653 because the ] is interpreted as the first data character and the
2654 terminating ] is not found. PCRE has been made compatible with Perl in this
2655 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2656 cause memory overwriting.
2658 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2659 string has been matched (to stop an infinite loop). It was not recognizing
2660 a conditional subpattern that could match an empty string if that
2661 subpattern was within another subpattern. For example, it looped when
2662 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2663 condition was not nested. This bug has been fixed.
2665 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2666 past the start of the subject in the presence of bytes with the top bit
2667 set, for example "\x8aBCD".
2669 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2670 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2672 14. Optimized (?!) to (*FAIL).
2674 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2675 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2676 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2677 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2678 does: it's just the validity check that is more restrictive.
2680 16. Inserted checks for integer overflows during escape sequence (backslash)
2681 processing, and also fixed erroneous offset values for syntax errors during
2682 backslash processing.
2684 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2685 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2687 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2688 caused an overrun.
2690 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2691 something other than just ASCII characters) inside a group that had an
2692 unlimited repeat caused a loop at compile time (while checking to see
2693 whether the group could match an empty string).
2695 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2696 [\P{Any}] did so. (Error in the code for printing property names.)
2698 21. An orphan \E inside a character class could cause a crash.
2700 22. A repeated capturing bracket such as (A)? could cause a wild memory
2701 reference during compilation.
2703 23. There are several functions in pcre_compile() that scan along a compiled
2704 expression for various reasons (e.g. to see if it's fixed length for look
2705 behind). There were bugs in these functions when a repeated \p or \P was
2706 present in the pattern. These operators have additional parameters compared
2707 with \d, etc, and these were not being taken into account when moving along
2708 the compiled data. Specifically:
2710 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2711 length.
2713 (b) An item such as \pL+ within a repeated group could cause crashes or
2714 loops.
2716 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2717 "reference to non-existent subpattern" error.
2719 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2721 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2722 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2724 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2726 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2727 character were causing crashes (broken optimization).
2729 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2730 \p or \P) caused a compile-time loop.
2732 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2733 break. For example, the unstudied pattern [\r\n]A does not match the string
2734 "\r\nA" because change 7.0/46 below moves the current point on by two
2735 characters after failing to match at the start. However, the pattern \nA
2736 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2737 the same is true. There doesn't seem any very clean way out of this, but
2738 what I have chosen to do makes the common cases work: PCRE now takes note
2739 of whether there can be an explicit match for \r or \n anywhere in the
2740 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2741 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2742 pattern has explicit CR or LF references.
2744 29. Added (*CR) etc for changing newline setting at start of pattern.
2747 Version 7.2 19-Jun-07
2748 ---------------------
2750 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2751 which is apparently normally available under Windows.
2753 2. Re-jig the pcregrep tests with different newline settings in an attempt
2754 to make them independent of the local environment's newline setting.
2756 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2758 4. Some of the "internals" tests were previously cut out when the link size
2759 was not 2, because the output contained actual offsets. The recent new
2760 "Z" feature of pcretest means that these can be cut out, making the tests
2761 usable with all link sizes.
2763 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2764 stack recursion. This gives a massive performance boost under BSD, but just
2765 a small improvement under Linux. However, it saves one field in the frame
2766 in all cases.
2768 6. Added more features from the forthcoming Perl 5.10:
2770 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2771 recursion call. It refers to the nth most recently opened parentheses.
2773 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2774 to be opened parentheses.
2776 (c) Conditions that refer to capturing parentheses can be specified
2777 relatively, for example, (?(-2)... or (?(+3)...
2779 (d) \K resets the start of the current match so that everything before
2780 is not part of it.
2782 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2784 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2785 reference syntax.
2787 (g) (?| introduces a group in which the numbering of parentheses in each
2788 alternative starts with the same number.
2790 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2792 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2795 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2796 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2797 for detecting groups that can match an empty string.
2799 9. A pattern with a very large number of alternatives (more than several
2800 hundred) was running out of internal workspace during the pre-compile
2801 phase, where pcre_compile() figures out how much memory will be needed. A
2802 bit of new cunning has reduced the workspace needed for groups with
2803 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2804 workspace instead of running out of the 4096 that are available.
2806 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2808 11. Applied patch from Google to remove an optimization that didn't quite work.
2809 The report of the bug said:
2811 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2812 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2813 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2815 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2816 it matched the wrong number of bytes.
2819 Version 7.1 24-Apr-07
2820 ---------------------
2822 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2823 that is more "standard", making use of automake and other Autotools. There
2824 is some re-arrangement of the files and adjustment of comments consequent
2825 on this.
2827 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2828 for recursive directory scanning broke on some systems because the files
2829 are not scanned in any specific order and on different systems the order
2830 was different. A call to "sort" has been inserted into RunGrepTest for the
2831 approprate test as a short-term fix. In the longer term there may be an
2832 alternative.
2834 3. I had an email from Eric Raymond about problems translating some of PCRE's
2835 man pages to HTML (despite the fact that I distribute HTML pages, some
2836 people do their own conversions for various reasons). The problems
2837 concerned the use of low-level troff macros .br and .in. I have therefore
2838 removed all such uses from the man pages (some were redundant, some could
2839 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2840 HTML has been updated to handle .nf/.fi and to complain if it encounters
2841 .br or .in.
2843 4. Updated comments in configure.ac that get placed in config.h.in and also
2844 arranged for config.h to be included in the distribution, with the name
2845 config.h.generic, for the benefit of those who have to compile without
2846 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2848 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2849 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2850 makevp.bat for the current PCRE, using the additional files
2851 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2853 6. A Windows user reported a minor discrepancy with test 2, which turned out
2854 to be caused by a trailing space on an input line that had got lost in his
2855 copy. The trailing space was an accident, so I've just removed it.
2857 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2858 that is needed.
2860 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2861 as "const" (a) because they are and (b) because it helps the PHP
2862 maintainers who have recently made a script to detect big data structures
2863 in the php code that should be moved to the .rodata section. I remembered
2864 to update Builducptable as well, so it won't revert if ucptable.h is ever
2865 re-created.
2867 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2868 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2869 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2870 support is not required. This saves 1.5-2K of code, which is important in
2871 some applications.
2873 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2874 so as not to refer to the tables, even though these functions will never be
2875 called when UTF-8 support is disabled. Otherwise there are problems with a
2876 shared library.
2878 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2880 (a) It was defining its arguments as char * instead of void *.
2882 (b) It was assuming that all moves were upwards in memory; this was true
2883 a long time ago when I wrote it, but is no longer the case.
2885 The emulated memove() is provided for those environments that have neither
2886 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2887 is clearly not the case, as these two bugs were recently reported.
2889 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2890 and Detrail to create the HTML documentation, the .txt form of the man
2891 pages, and it removes trailing spaces from listed files. It also creates
2892 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2893 case, it wraps all the #defines with #ifndefs. This script should be run
2894 before "make dist".
2896 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2897 with Unicode property support.
2899 (a) For a maximizing quantifier, if the two different cases of the
2900 character were of different lengths in their UTF-8 codings (there are
2901 some cases like this - I found 11), and the matching function had to
2902 back up over a mixture of the two cases, it incorrectly assumed they
2903 were both the same length.
2905 (b) When PCRE was configured to use the heap rather than the stack for
2906 recursion during matching, it was not correctly preserving the data for
2907 the other case of a UTF-8 character when checking ahead for a match
2908 while processing a minimizing repeat. If the check also involved
2909 matching a wide character, but failed, corruption could cause an
2910 erroneous result when trying to check for a repeat of the original
2911 character.
2913 13. Some tidying changes to the testing mechanism:
2915 (a) The RunTest script now detects the internal link size and whether there
2916 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2917 values substituted by "configure". (The RunGrepTest script already did
2918 this for UTF-8.) The configure.ac script no longer substitutes the
2919 relevant variables.
2921 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2922 with length and offset values. This means that the output is different
2923 for different internal link sizes. Test 2 is skipped for link sizes
2924 other than 2 because of this, bypassing the problem. Unfortunately,
2925 there was also a test in test 3 (the locale tests) that used /B and
2926 failed for link sizes other than 2. Rather than cut the whole test out,
2927 I have added a new /Z option to pcretest that replaces the length and
2928 offset values with spaces. This is now used to make test 3 independent
2929 of link size. (Test 2 will be tidied up later.)
2931 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2932 segmentation fault instead of returning the appropriate error message.
2934 15. In multiline mode when the newline sequence was set to "any", the pattern
2935 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2936 This doesn't seem right; it now treats the CRLF combination as the line
2937 ending, and so does not match in that case. It's only a pattern such as ^$
2938 that would hit this one: something like ^ABC$ would have failed after \r
2939 and then tried again after \r\n.
2941 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2942 in an attempt to make files that differ only in their line terminators
2943 compare equal. This works on Linux.
2945 17. Under certain error circumstances pcregrep might try to free random memory
2946 as it exited. This is now fixed, thanks to valgrind.
2948 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2949 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2950 was because its rules for how to advance for /g after matching an empty
2951 string at the end of a line did not allow for this case. They now check for
2952 it specially.
2954 20. pcretest is supposed to handle patterns and data of any length, by
2955 extending its buffers when necessary. It was getting this wrong when the
2956 buffer for a data line had to be extended.
2958 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2959 CRLF as a newline sequence.
2961 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2962 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2963 I have nevertheless tidied it up.
2965 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2967 24. Added a man page for pcre-config.
2970 Version 7.0 19-Dec-06
2971 ---------------------
2973 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2974 moving to gcc 4.1.1.
2976 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2977 sys/time.h, which is documented as needed for this function. It doesn't
2978 seem to matter on Linux, but it showed up on some releases of OS X.
2980 3. It seems that there are systems where bytes whose values are greater than
2981 127 match isprint() in the "C" locale. The "C" locale should be the
2982 default when a C program starts up. In most systems, only ASCII printing
2983 characters match isprint(). This difference caused the output from pcretest
2984 to vary, making some of the tests fail. I have changed pcretest so that:
2986 (a) When it is outputting text in the compiled version of a pattern, bytes
2987 other than 32-126 are always shown as hex escapes.
2989 (b) When it is outputting text that is a matched part of a subject string,
2990 it does the same, unless a different locale has been set for the match
2991 (using the /L modifier). In this case, it uses isprint() to decide.
2993 4. Fixed a major bug that caused incorrect computation of the amount of memory
2994 required for a compiled pattern when options that changed within the
2995 pattern affected the logic of the preliminary scan that determines the
2996 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2997 that the computed length was too small. The symptoms of this bug were
2998 either the PCRE error "internal error: code overflow" from pcre_compile(),
2999 or a glibc crash with a message such as "pcretest: free(): invalid next
3000 size (fast)". Examples of patterns that provoked this bug (shown in
3001 pcretest format) are:
3003 /(?-x: )/x
3004 /(?x)(?-x: \s*#\s*)/
3005 /((?i)[\x{c0}])/8
3006 /(?i:[\x{c0}])/8
3008 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3009 is now done differently.
3011 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3012 wrapper classes; (b) implement a new function in the C++ scanner that is
3013 more efficient than the old way of doing things because it avoids levels of
3014 recursion in the regex matching; (c) add a paragraph to the documentation
3015 for the FullMatch() function.
3017 6. The escape sequence \n was being treated as whatever was defined as
3018 "newline". Not only was this contrary to the documentation, which states
3019 that \n is character 10 (hex 0A), but it also went horribly wrong when
3020 "newline" was defined as CRLF. This has been fixed.
3022 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3023 was being set to -1 for the "end of line" case (supposedly a value that no
3024 character can have). Though this value is never used (the check for end of
3025 line is "zero bytes in current character"), it caused compiler complaints.
3026 I've changed it to 0xffffffff.
3028 8. In pcre_version.c, the version string was being built by a sequence of
3029 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3030 string (as it is for production releases) called a macro with an empty
3031 argument. The C standard says the result of this is undefined. The gcc
3032 compiler treats it as an empty string (which was what was wanted) but it is
3033 reported that Visual C gives an error. The source has been hacked around to
3034 avoid this problem.
3036 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3037 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3038 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3039 of them did).
3041 10. Originally, pcretest opened its input and output without "b"; then I was
3042 told that "b" was needed in some environments, so it was added for release
3043 5.0 to both the input and output. (It makes no difference on Unix-like
3044 systems.) Later I was told that it is wrong for the input on Windows. I've
3045 now abstracted the modes into two macros, to make it easier to fiddle with
3046 them, and removed "b" from the input mode under Windows.
3048 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3050 12. Added -help and --help to pcretest as an official way of being reminded
3051 of the options.
3053 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3054 and pcrecpp.cc because they annoy compilers at high warning levels.
3056 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3058 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3059 scripts are not C programs :-) and which was not noticed because it works
3060 on Linux.
3062 16. pcretest is supposed to handle any length of pattern and data line (as one
3063 line or as a continued sequence of lines) by extending its input buffer if
3064 necessary. This feature was broken for very long pattern lines, leading to
3065 a string of junk being passed to pcre_compile() if the pattern was longer
3066 than about 50K.
3068 17. I have done a major re-factoring of the way pcre_compile() computes the
3069 amount of memory needed for a compiled pattern. Previously, there was code
3070 that made a preliminary scan of the pattern in order to do this. That was
3071 OK when PCRE was new, but as the facilities have expanded, it has become
3072 harder and harder to keep it in step with the real compile phase, and there
3073 have been a number of bugs (see for example, 4 above). I have now found a
3074 cunning way of running the real compile function in a "fake" mode that
3075 enables it to compute how much memory it would need, while actually only
3076 ever using a few hundred bytes of working memory and without too many
3077 tests of the mode. This should make future maintenance and development
3078 easier. A side effect of this work is that the limit of 200 on the nesting
3079 depth of parentheses has been removed (though this was never a serious
3080 limitation, I suspect). However, there is a downside: pcre_compile() now
3081 runs more slowly than before (30% or more, depending on the pattern). I
3082 hope this isn't a big issue. There is no effect on runtime performance.
3084 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3085 newline (only possible for the last line of a file) and it was a
3086 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3088 19. Added additional timing features to pcretest. (1) The -tm option now times
3089 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3090 separate command line item, by a number that specifies the number of
3091 repeats to use when timing. The default is 50000; this gives better
3092 precision, but takes uncomfortably long for very large patterns.
3094 20. Extended pcre_study() to be more clever in cases where a branch of a
3095 subpattern has no definite first character. For example, (a*|b*)[cd] would
3096 previously give no result from pcre_study(). Now it recognizes that the
3097 first character must be a, b, c, or d.
3099 21. There was an incorrect error "recursive call could loop indefinitely" if
3100 a subpattern (or the entire pattern) that was being tested for matching an
3101 empty string contained only one non-empty item after a nested subpattern.
3102 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3103 incorrectly, because the \d was being skipped in the check.
3105 22. The pcretest program now has a new pattern option /B and a command line
3106 option -b, which is equivalent to adding /B to every pattern. This causes
3107 it to show the compiled bytecode, without the additional information that
3108 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3109 is the same as /B/I).
3111 23. A new optimization is now able automatically to treat some sequences such
3112 as a*b as a*+b. More specifically, if something simple (such as a character
3113 or a simple class like \d) has an unlimited quantifier, and is followed by
3114 something that cannot possibly match the quantified thing, the quantifier
3115 is automatically "possessified".
3117 24. A recursive reference to a subpattern whose number was greater than 39
3118 went wrong under certain circumstances in UTF-8 mode. This bug could also
3119 have affected the operation of pcre_study().
3121 25. Realized that a little bit of performance could be had by replacing
3122 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3124 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3126 27. Possessive quantifiers such as a++ were previously implemented by turning
3127 them into atomic groups such as ($>a+). Now they have their own opcodes,
3128 which improves performance. This includes the automatically created ones
3129 from 23 above.
3131 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3132 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3133 the first matched character to be a colon. This applied both to named and
3134 numbered groups.
3136 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3138 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3139 building PCRE on Windows easier, so I have included it in the distribution.
3141 31. There is now a check in pcretest against a ridiculously large number being
3142 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3143 loop, the loop is abandoned.
3145 32. Forward references to subpatterns in conditions such as (?(2)...) where
3146 subpattern 2 is defined later cause pcre_compile() to search forwards in
3147 the pattern for the relevant set of parentheses. This search went wrong
3148 when there were unescaped parentheses in a character class, parentheses
3149 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3151 33. "Subroutine" calls and backreferences were previously restricted to
3152 referencing subpatterns earlier in the regex. This restriction has now
3153 been removed.
3155 34. Added a number of extra features that are going to be in Perl 5.10. On the
3156 whole, these are just syntactic alternatives for features that PCRE had
3157 previously implemented using the Python syntax or my own invention. The
3158 other formats are all retained for compatibility.
3160 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3161 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3162 also .NET compatible.
3164 (b) A recursion or subroutine call to a named group can now be defined as
3165 (?&name) as well as (?P>name).
3167 (c) A backreference to a named group can now be defined as \k<name> or
3168 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3169 5.10, are also .NET compatible.
3171 (d) A conditional reference to a named group can now use the syntax
3172 (?(<name>) or (?('name') as well as (?(name).
3174 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3175 groups (named and numbered) that are never evaluated inline, but can be
3176 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3177 is always false. There may be only one alternative in such a group.
3179 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3180 as the simple (?(R). The condition is true only if the most recent
3181 recursion is that of the given number or name. It does not search out
3182 through the entire recursion stack.
3184 (g) The escape \gN or \g{N} has been added, where N is a positive or
3185 negative number, specifying an absolute or relative reference.
3187 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3188 some "unreachable code" warnings.
3190 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3191 things, this adds five new scripts.
3193 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3194 There were also incompatibilities regarding the handling of \Q..\E inside
3195 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3196 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3198 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3199 matches an empty string, and forcibly breaks the loop. There were bugs in
3200 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3201 against aaaa the result was just "a" rather than "aaaa", for example. Two
3202 separate and independent bugs (that affected different cases) have been
3203 fixed.
3205 39. Refactored the code to abolish the use of different opcodes for small
3206 capturing bracket numbers. This is a tidy that I avoided doing when I
3207 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3208 The new approach is not only tidier, it makes it possible to reduce the
3209 memory needed to fix the previous bug (38).
3211 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3212 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3213 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3214 mode.
3216 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3217 report.
3219 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3220 copy construction and assignment in the C++ wrapper.
3222 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3223 couple of bugs that could have given wrong results in the "--newline=crlf"
3224 case.
3226 44. Added a number of casts and did some reorganization of signed/unsigned int
3227 variables following suggestions from Dair Grant. Also renamed the variable
3228 "this" as "item" because it is a C++ keyword.
3230 45. Arranged for dftables to add
3232 #include "pcre_internal.h"
3234 to pcre_chartables.c because without it, gcc 4.x may remove the array
3235 definition from the final binary if PCRE is built into a static library and
3236 dead code stripping is activated.
3238 46. For an unanchored pattern, if a match attempt fails at the start of a
3239 newline sequence, and the newline setting is CRLF or ANY, and the next two
3240 characters are CRLF, advance by two characters instead of one.
3243 Version 6.7 04-Jul-06
3244 ---------------------
3246 1. In order to handle tests when input lines are enormously long, pcretest has
3247 been re-factored so that it automatically extends its buffers when
3248 necessary. The code is crude, but this _is_ just a test program. The
3249 default size has been increased from 32K to 50K.
3251 2. The code in pcre_study() was using the value of the re argument before
3252 testing it for NULL. (Of course, in any sensible call of the function, it
3253 won't be NULL.)
3255 3. The memmove() emulation function in pcre_internal.h, which is used on
3256 systems that lack both memmove() and bcopy() - that is, hardly ever -
3257 was missing a "static" storage class specifier.
3259 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3260 containing an extended class (one that cannot be represented by a bitmap
3261 because it contains high-valued characters or Unicode property items, e.g.
3262 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3263 pattern, but PCRE should not loop if you do not (it no longer does).
3264 [Detail: two cases were found: (a) a repeated subpattern containing an
3265 extended class; (b) a recursive reference to a subpattern that followed a
3266 previous extended class. It wasn't skipping over the extended class
3267 correctly when UTF-8 mode was not set.]
3269 5. A negated single-character class was not being recognized as fixed-length
3270 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3271 compile error "lookbehind assertion is not fixed length".
3273 6. The RunPerlTest auxiliary script was showing an unexpected difference
3274 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3275 write a Perl script that can interpret lines of an input file either as
3276 byte characters or as UTF-8, which is what "perltest" was being required to
3277 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3278 can't do is switch easily at run time between having the "use utf8;" pragma
3279 or not. In the end, I fudged it by using the RunPerlTest script to insert
3280 "use utf8;" explicitly for the UTF-8 tests.
3282 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3283 the end of the subject string, contrary to the documentation and to what
3284 Perl does. This was true of both matching functions. Now it matches only at
3285 the start of the subject and immediately after *internal* newlines.
3287 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3288 a pointer to an int instead of a pointer to an unsigned long int. This
3289 caused problems on 64-bit systems.
3291 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3292 instance of the 'standard' template library not being so standard".
3294 10. There was no check on the number of named subpatterns nor the maximum
3295 length of a subpattern name. The product of these values is used to compute
3296 the size of the memory block for a compiled pattern. By supplying a very
3297 long subpattern name and a large number of named subpatterns, the size
3298 computation could be caused to overflow. This is now prevented by limiting
3299 the length of names to 32 characters, and the number of named subpatterns
3300 to 10,000.
3302 11. Subpatterns that are repeated with specific counts have to be replicated in
3303 the compiled pattern. The size of memory for this was computed from the
3304 length of the subpattern and the repeat count. The latter is limited to
3305 65535, but there was no limit on the former, meaning that integer overflow
3306 could in principle occur. The compiled length of a repeated subpattern is
3307 now limited to 30,000 bytes in order to prevent this.
3309 12. Added the optional facility to have named substrings with the same name.
3311 13. Added the ability to use a named substring as a condition, using the
3312 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3313 are numbers (not recommended). Forward references are permitted.
3315 14. Added forward references in named backreferences (if you see what I mean).
3317 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3318 pattern could run off the end of the subject. For example, the pattern
3319 "(?s)(.{1,5})"8 did this with the subject "ab".
3321 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3322 PCRE_CASELESS was set when matching characters that were quantified with ?
3323 or *.
3325 17. A character class other than a single negated character that had a minimum
3326 but no maximum quantifier - for example [ab]{6,} - was not handled
3327 correctly by pce_dfa_exec(). It would match only one character.
3329 18. A valid (though odd) pattern that looked like a POSIX character
3330 class but used an invalid character after [ (for example [[,abc,]]) caused
3331 pcre_compile() to give the error "Failed: internal error: code overflow" or
3332 in some cases to crash with a glibc free() error. This could even happen if
3333 the pattern terminated after [[ but there just happened to be a sequence of
3334 letters, a binary zero, and a closing ] in the memory that followed.
3336 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3337 over the years. Originally (before any Unicode support), just the bottom 8
3338 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3339 output from "man perlunicode" includes this:
3341 The regular expression compiler produces polymorphic opcodes. That
3342 is, the pattern adapts to the data and automatically switches to
3343 the Unicode character scheme when presented with Unicode data--or
3344 instead uses a traditional byte scheme when presented with byte
3345 data.
3347 Sadly, a wide octal escape does not cause a switch, and in a string with
3348 no other multibyte characters, these octal escapes are treated as before.
3349 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3350 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3351 Unicode string.
3353 I have not perpetrated such confusion in PCRE. Up till now, it took just
3354 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3355 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3356 translate to the appropriate multibyte character.
3358 29. Applied some refactoring to reduce the number of warnings from Microsoft
3359 and Borland compilers. This has included removing the fudge introduced
3360 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3361 a warning about an unused variable.
3363 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3364 characters since release 4.0, because Perl (from release 5.004) does not.
3365 [Or at least, is documented not to: some releases seem to be in conflict
3366 with the documentation.] However, when a pattern was studied with
3367 pcre_study() and all its branches started with \s, PCRE still included VT
3368 as a possible starting character. Of course, this did no harm; it just
3369 caused an unnecessary match attempt.
3371 22. Removed a now-redundant internal flag bit that recorded the fact that case
3372 dependency changed within the pattern. This was once needed for "required
3373 byte" processing, but is no longer used. This recovers a now-scarce options
3374 bit. Also moved the least significant internal flag bit to the most-
3375 significant bit of the word, which was not previously used (hangover from
3376 the days when it was an int rather than a uint) to free up another bit for
3377 the future.
3379 23. Added support for CRLF line endings as well as CR and LF. As well as the
3380 default being selectable at build time, it can now be changed at runtime
3381 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3382 specify that it is scanning data with non-default line endings.
3384 24. Changed the definition of CXXLINK to make it agree with the definition of
3385 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3387 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3388 recursions. This makes a big different to stack usage for some patterns.
3390 26. If a subpattern containing a named recursion or subroutine reference such
3391 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3392 the space required for the compiled pattern went wrong and gave too small a
3393 value. Depending on the environment, this could lead to "Failed: internal
3394 error: code overflow at offset 49" or "glibc detected double free or
3395 corruption" errors.
3397 27. Applied patches from Google (a) to support the new newline modes and (b) to
3398 advance over multibyte UTF-8 characters in GlobalReplace.
3400 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3401 difference for some implementation of PCRE in some Windows version.
3403 29. Added some extra testing facilities to pcretest:
3405 \q<number> in a data line sets the "match limit" value
3406 \Q<number> in a data line sets the "match recursion limt" value
3407 -S <number> sets the stack size, where <number> is in megabytes
3409 The -S option isn't available for Windows.
3412 Version 6.6 06-Feb-06
3413 ---------------------
3415 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3416 in pcreposix.h. I have copied the definition from pcre.h.
3418 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3419 because pcre.h is no longer a built file.
3421 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3422 not normally included in the compiled code.
3425 Version 6.5 01-Feb-06
3426 ---------------------
3428 1. When using the partial match feature with pcre_dfa_exec(), it was not
3429 anchoring the second and subsequent partial matches at the new starting
3430 point. This could lead to incorrect results. For example, with the pattern
3431 /1234/, partially matching against "123" and then "a4" gave a match.
3433 2. Changes to pcregrep:
3435 (a) All non-match returns from pcre_exec() were being treated as failures
3436 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3437 error message is output. Some extra information is given for the
3439 probably the only errors that are likely to be caused by users (by
3440 specifying a regex that has nested indefinite repeats, for instance).
3441 If there are more than 20 of these errors, pcregrep is abandoned.
3443 (b) A binary zero was treated as data while matching, but terminated the
3444 output line if it was written out. This has been fixed: binary zeroes
3445 are now no different to any other data bytes.
3447 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3448 used to set a locale for matching. The --locale=xxxx long option has
3449 been added (no short equivalent) to specify a locale explicitly on the
3450 pcregrep command, overriding the environment variables.
3452 (d) When -B was used with -n, some line numbers in the output were one less
3453 than they should have been.
3455 (e) Added the -o (--only-matching) option.
3457 (f) If -A or -C was used with -c (count only), some lines of context were
3458 accidentally printed for the final match.
3460 (g) Added the -H (--with-filename) option.
3462 (h) The combination of options -rh failed to suppress file names for files
3463 that were found from directory arguments.
3465 (i) Added the -D (--devices) and -d (--directories) options.
3467 (j) Added the -F (--fixed-strings) option.
3469 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3471 (l) Added the --colo(u)r option.
3473 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3474 is not present by default.
3476 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3477 items such as (?R) or (?1), when the recursion could match a number of
3478 alternatives. If it matched one of the alternatives, but subsequently,
3479 outside the recursion, there was a failure, the code tried to back up into
3480 the recursion. However, because of the way PCRE is implemented, this is not
3481 possible, and the result was an incorrect result from the match.
3483 In order to prevent this happening, the specification of recursion has
3484 been changed so that all such subpatterns are automatically treated as
3485 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3487 4. I had overlooked the fact that, in some locales, there are characters for
3488 which isalpha() is true but neither isupper() nor islower() are true. In
3489 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3490 and ordfeminine) are like this. This affected the treatment of \w and \W
3491 when they appeared in character classes, but not when they appeared outside
3492 a character class. The bit map for "word" characters is now created
3493 separately from the results of isalnum() instead of just taking it from the
3494 upper, lower, and digit maps. (Plus the underscore character, of course.)
3496 5. The above bug also affected the handling of POSIX character classes such as
3497 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3498 permanent tables. Instead, the bit maps for such a class were previously
3499 created as the appropriate unions of the upper, lower, and digit bitmaps.
3500 Now they are created by subtraction from the [[:word:]] class, which has
3501 its own bitmap.
3503 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3504 It is created by subtracting the vertical space characters (\x09, \x0a,
3505 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3506 subtraction was done in the overall bitmap for a character class, meaning
3507 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3508 be recognized. This bug has been fixed.
3510 7. Patches from the folks at Google:
3512 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3513 real life, but is still worth protecting against".
3515 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3516 regular expressions".
3518 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3519 have it.
3521 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3522 "configure" and the latter not, in order to fix a problem somebody had
3523 with compiling the Arg class on HP-UX.
3525 (e) Improve the error-handling of the C++ wrapper a little bit.
3527 (f) New tests for checking recursion limiting.
3529 8. The pcre_memmove() function, which is used only if the environment does not
3530 have a standard memmove() function (and is therefore rarely compiled),
3531 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3532 returning a result (though PCRE never actually uses the result).
3534 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3535 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3536 returned instead of calling malloc() with an overflowing number that would
3537 most likely cause subsequent chaos.
3539 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3541 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3542 with this option is matched, the nmatch and pmatch options of regexec() are
3543 ignored.
3545 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3546 provided in case anyone wants to the the POSIX interface with UTF-8
3547 strings.
3549 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3550 C++ linking (needed for some HP-UX environments).
3552 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3553 (unused parameter) and in the pcre_printint() function (omitted "default"
3554 switch label when the default is to do nothing).
3556 15. Added some code to make it possible, when PCRE is compiled as a C++
3557 library, to replace subject pointers for pcre_exec() with a smart pointer
3558 class, thus making it possible to process discontinuous strings.
3560 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3561 much the same function. They were added by different people who were trying
3562 to make PCRE easy to compile on non-Unix systems. It has been suggested
3563 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3564 for compiling on Windows systems. I have therefore replaced it with
3565 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3566 defaults to "extern" for C or "extern C" for C++, which works fine on
3567 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3568 SCOPE with something explicit in config.h. In addition:
3570 (a) pcreposix.h still had just "extern" instead of either of these macros;
3571 I have replaced it with PCRE_DATA_SCOPE.
3573 (b) Functions such as _pcre_xclass(), which are internal to the library,
3574 but external in the C sense, all had PCRE_EXPORT in their definitions.
3575 This is apparently wrong for the Windows case, so I have removed it.
3576 (It makes no difference on Unix-like systems.)
3578 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3579 of recursive calls to match(). This is different to MATCH_LIMIT because
3580 that limits the total number of calls to match(), not all of which increase
3581 the depth of recursion. Limiting the recursion depth limits the amount of
3582 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3583 when PCRE is compiled, and changed at run time. A patch from Google adds
3584 this functionality to the C++ interface.
3586 18. Changes to the handling of Unicode character properties:
3588 (a) Updated the table to Unicode 4.1.0.
3590 (b) Recognize characters that are not in the table as "Cn" (undefined).
3592 (c) I revised the way the table is implemented to a much improved format
3593 which includes recognition of ranges. It now supports the ranges that
3594 are defined in UnicodeData.txt, and it also amalgamates other
3595 characters into ranges. This has reduced the number of entries in the
3596 table from around 16,000 to around 3,000, thus reducing its size
3597 considerably. I realized I did not need to use a tree structure after
3598 all - a binary chop search is just as efficient. Having reduced the
3599 number of entries, I extended their size from 6 bytes to 8 bytes to
3600 allow for more data.
3602 (d) Added support for Unicode script names via properties such as \p{Han}.
3604 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3605 matching that character.
3607 20. When matching a repeated Unicode property with a minimum greater than zero,
3608 (for example \pL{2,}), PCRE could look past the end of the subject if it
3609 reached it while seeking the minimum number of characters. This could
3610 happen only if some of the characters were more than one byte long, because
3611 there is a check for at least the minimum number of bytes.
3613 21. Refactored the implementation of \p and \P so as to be more general, to
3614 allow for more different types of property in future. This has changed the
3615 compiled form incompatibly. Anybody with saved compiled patterns that use
3616 \p or \P will have to recompile them.
3618 22. Added "Any" and "L&" to the supported property types.
3620 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3621 but give a compile time error if the value is greater than 0xff.
3623 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3624 accidentally not being installed or uninstalled.
3626 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3627 made were to insert the current release number. This seemed silly, because
3628 it made things harder for people building PCRE on systems that don't run
3629 "configure". I have turned pcre.h into a distributed file, no longer built
3630 by "configure", with the version identification directly included. There is
3631 no longer a pcre.h.in file.
3633 However, this change necessitated a change to the pcre-config script as
3634 well. It is built from pcre-config.in, and one of the substitutions was the
3635 release number. I have updated configure.ac so that ./configure now finds
3636 the release number by grepping pcre.h.
3638 26. Added the ability to run the tests under valgrind.
3641 Version 6.4 05-Sep-05
3642 ---------------------
3644 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3645 "--" to be printed when multiple files were scanned, even when none of the
3646 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3647 consider it to be a bug, and have restored the previous behaviour.
3649 2. A couple of code tidies to get rid of compiler warnings.
3651 3. The pcretest program used to cheat by referring to symbols in the library
3652 whose names begin with _pcre_. These are internal symbols that are not
3653 really supposed to be visible externally, and in some environments it is
3654 possible to suppress them. The cheating is now confined to including
3655 certain files from the library's source, which is a bit cleaner.
3657 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3658 file's purpose clearer.
3660 5. Reorganized pcre_ucp_findchar().
3663 Version 6.3 15-Aug-05
3664 ---------------------
3666 1. The file libpcre.pc.in did not have general read permission in the tarball.
3668 2. There were some problems when building without C++ support:
3670 (a) If C++ support was not built, "make install" and "make test" still
3671 tried to test it.
3673 (b) There were problems when the value of CXX was explicitly set. Some
3674 changes have been made to try to fix these, and ...
3676 (c) --disable-cpp can now be used to explicitly disable C++ support.
3678 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3679 backslash in a target when C++ was disabled. This confuses some
3680 versions of "make", apparently. Using an intermediate variable solves
3681 this. (Same for CPP_LOBJ.)
3683 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3684 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3685 necessary on certain architectures.
3687 4. Added a setting of -export-symbols-regex to the link command to remove
3688 those symbols that are exported in the C sense, but actually are local
3689 within the library, and not documented. Their names all begin with
3690 "_pcre_". This is not a perfect job, because (a) we have to except some
3691 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3692 available (and never for static libraries). I have made a note to try to
3693 find a way round (a) in the future.
3696 Version 6.2 01-Aug-05
3697 ---------------------
3699 1. There was no test for integer overflow of quantifier values. A construction
3700 such as {1111111111111111} would give undefined results. What is worse, if
3701 a minimum quantifier for a parenthesized subpattern overflowed and became
3702 negative, the calculation of the memory size went wrong. This could have
3703 led to memory overwriting.
3705 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3707 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3708 operating environments where this matters.
3710 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3711 PCRE options from within the C++ wrapper.
3713 5. Named capturing subpatterns were not being correctly counted when a pattern
3714 was compiled. This caused two problems: (a) If there were more than 100
3715 such subpatterns, the calculation of the memory needed for the whole
3716 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3717 back references of the form \12, where the number was greater than 9, were
3718 not recognized as back references, even though there were sufficient
3719 previous subpatterns.
3721 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3722 versions of gcc, e.g. 2.95.4.
3725 Version 6.1 21-Jun-05
3726 ---------------------
3728 1. There was one reference to the variable "posix" in pcretest.c that was not
3729 surrounded by "#if !defined NOPOSIX".
3731 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3732 the cross-check on the old pcre_info() function, for the benefit of the
3733 cut-down version of PCRE that is currently imported into Exim.
3735 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3736 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3737 patterns that start (?i) but I don't think that matters. The use of (?i) is
3738 just an example; this all applies to the other options as well.
3740 4. Since libtool seems to echo the compile commands it is issuing, the output
3741 from "make" can be reduced a bit by putting "@" in front of each libtool
3742 compile command.
3744 5. Patch from the folks at Google for configure.in to be a bit more thorough
3745 in checking for a suitable C++ installation before trying to compile the
3746 C++ stuff. This should fix a reported problem when a compiler was present,
3747 but no suitable headers.
3749 6. The man pages all had just "PCRE" as their title. I have changed them to
3750 be the relevant file name. I have also arranged that these names are
3751 retained in the file doc/pcre.txt, which is a concatenation in text format
3752 of all the man pages except the little individual ones for each function.
3754 7. The NON-UNIX-USE file had not been updated for the different set of source
3755 files that come with release 6. I also added a few comments about the C++
3756 wrapper.
3759 Version 6.0 07-Jun-05
3760 ---------------------
3762 1. Some minor internal re-organization to help with my DFA experiments.
3764 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3765 didn't matter for the library itself when fully configured, but did matter
3766 when compiling without UCP support, or within Exim, where the ucp files are
3767 not imported.
3769 3. Refactoring of the library code to split up the various functions into
3770 different source modules. The addition of the new DFA matching code (see
3771 below) to a single monolithic source would have made it really too
3772 unwieldy, quite apart from causing all the code to be include in a
3773 statically linked application, when only some functions are used. This is
3774 relevant even without the DFA addition now that patterns can be compiled in
3775 one application and matched in another.
3777 The downside of splitting up is that there have to be some external
3778 functions and data tables that are used internally in different modules of
3779 the library but which are not part of the API. These have all had their
3780 names changed to start with "_pcre_" so that they are unlikely to clash
3781 with other external names.
3783 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3784 a different (DFA) algorithm. Although it is slower than the original
3785 function, it does have some advantages for certain types of matching
3786 problem.
3788 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3789 including restarting after a partial match.
3791 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3792 defined when compiling for Windows was sent to me. I have put it into the
3793 code, though I have no means of testing or verifying it.
3795 7. Added the pcre_refcount() auxiliary function.
3797 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3798 match before or at the first newline in the subject string. In pcretest,
3799 the /f option on a pattern can be used to set this.
3801 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3802 would behave wrongly. This has been present in PCRE since release 4.0.
3804 10. A number of changes to the pcregrep command:
3806 (a) Refactored how -x works; insert ^(...)$ instead of setting
3807 PCRE_ANCHORED and checking the length, in preparation for adding
3808 something similar for -w.
3810 (b) Added the -w (match as a word) option.
3812 (c) Refactored the way lines are read and buffered so as to have more
3813 than one at a time available.
3815 (d) Implemented a pcregrep test script.
3817 (e) Added the -M (multiline match) option. This allows patterns to match
3818 over several lines of the subject. The buffering ensures that at least
3819 8K, or the rest of the document (whichever is the shorter) is available
3820 for matching (and similarly the previous 8K for lookbehind assertions).
3822 (f) Changed the --help output so that it now says
3824 -w, --word-regex(p)
3826 instead of two lines, one with "regex" and the other with "regexp"
3827 because that confused at least one person since the short forms are the
3828 same. (This required a bit of code, as the output is generated
3829 automatically from a table. It wasn't just a text change.)
3831 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3832 option but starts with a hyphen. Could be a pattern or a path name
3833 starting with a hyphen, for instance.
3835 (h) "-" can be given as a file name to represent stdin.
3837 (i) When file names are being printed, "(standard input)" is used for
3838 the standard input, for compatibility with GNU grep. Previously
3839 "<stdin>" was used.
3841 (j) The option --label=xxx can be used to supply a name to be used for
3842 stdin when file names are being printed. There is no short form.
3844 (k) Re-factored the options decoding logic because we are going to add
3845 two more options that take data. Such options can now be given in four
3846 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3848 (l) Added the -A, -B, and -C options for requesting that lines of context
3849 around matches be printed.
3851 (m) Added the -L option to print the names of files that do not contain
3852 any matching lines, that is, the complement of -l.
3854 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3855 continue to scan other files.
3857 (o) The -s option was incorrectly implemented. For compatibility with other
3858 greps, it now suppresses the error message for a non-existent or non-
3859 accessible file (but not the return code). There is a new option called
3860 -q that suppresses the output of matching lines, which was what -s was
3861 previously doing.
3863 (p) Added --include and --exclude options to specify files for inclusion
3864 and exclusion when recursing.
3866 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3867 Hopefully, it now does.
3869 12. Missing cast in pcre_study().
3871 13. Added an "uninstall" target to the makefile.
3873 14. Replaced "extern" in the function prototypes in Makefile.in with
3874 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3875 world, but is set differently for Windows.
3877 15. Added a second compiling function called pcre_compile2(). The only
3878 difference is that it has an extra argument, which is a pointer to an
3879 integer error code. When there is a compile-time failure, this is set
3880 non-zero, in addition to the error test pointer being set to point to an
3881 error message. The new argument may be NULL if no error number is required
3882 (but then you may as well call pcre_compile(), which is now just a
3883 wrapper). This facility is provided because some applications need a
3884 numeric error indication, but it has also enabled me to tidy up the way
3885 compile-time errors are handled in the POSIX wrapper.
3887 16. Added VPATH=.libs to the makefile; this should help when building with one
3888 prefix path and installing with another. (Or so I'm told by someone who
3889 knows more about this stuff than I do.)
3891 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3892 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3893 match everything, including newlines. This is not POSIX-compatible, but
3894 somebody wanted the feature. From pcretest it can be activated by using
3895 both the P and the s flags.
3897 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3899 19. libpcre.pc was being incorrectly installed as executable.
3901 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3902 it now also looks for '\r' so that it will work unmodified on Windows.
3904 21. Added Google's contributed C++ wrapper to the distribution.
3906 22. Added some untidy missing memory free() calls in pcretest, to keep
3907 Electric Fence happy when testing.
3911 Version 5.0 13-Sep-04
3912 ---------------------
3914 1. Internal change: literal characters are no longer packed up into items
3915 containing multiple characters in a single byte-string. Each character
3916 is now matched using a separate opcode. However, there may be more than one
3917 byte in the character in UTF-8 mode.
3919 2. The pcre_callout_block structure has two new fields: pattern_position and
3920 next_item_length. These contain the offset in the pattern to the next match
3921 item, and its length, respectively.
3923 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3924 insertion of callouts before each pattern item. Added the /C option to
3925 pcretest to make use of this.
3927 4. On the advice of a Windows user, the lines
3929 #if defined(_WIN32) || defined(WIN32)
3930 _setmode( _fileno( stdout ), 0x8000 );
3931 #endif /* defined(_WIN32) || defined(WIN32) */
3933 have been added to the source of pcretest. This apparently does useful
3934 magic in relation to line terminators.
3936 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3937 for the benefit of those environments where the "b" makes a difference.
3939 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3940 to know about it. I have put a hack into configure.in that adds in code
3941 to set GCC=yes if CC=icc. This seems to end up at a point in the
3942 generated configure script that is early enough to affect the setting of
3943 compiler options, which is what is needed, but I have no means of testing
3944 whether it really works. (The user who reported this had patched the
3945 generated configure script, which of course I cannot do.)
3947 LATER: After change 22 below (new libtool files), the configure script
3948 seems to know about icc (and also ecc). Therefore, I have commented out
3949 this hack in configure.in.
3951 7. Added support for pkg-config (2 patches were sent in).
3953 8. Negated POSIX character classes that used a combination of internal tables
3954 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3955 [[:^ascii]]. Typically, they would match almost any characters. The other
3956 POSIX classes were not broken in this way.
3958 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3959 to find the match, as PCRE was deluded into thinking that the match had to
3960 start at the start point or following a newline. The same bug applied to
3961 patterns with negative forward assertions or any backward assertions
3962 preceding ".*" at the start, unless the pattern required a fixed first
3963 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3965 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3966 starting at the last subject character, bytes beyond the end of the subject
3967 string were read.
3969 11. Renamed the variable "class" as "classbits" to make life easier for C++
3970 users. (Previously there was a macro definition, but it apparently wasn't
3971 enough.)
3973 12. Added the new field "tables" to the extra data so that tables can be passed
3974 in at exec time, or the internal tables can be re-selected. This allows
3975 a compiled regex to be saved and re-used at a later time by a different
3976 program that might have everything at different addresses.
3978 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3979 -R library as well as a -L library.
3981 14. The debugging options of pcretest (-d on the command line or D on a
3982 pattern) showed incorrect output for anything following an extended class
3983 that contained multibyte characters and which was followed by a quantifier.
3985 15. Added optional support for general category Unicode character properties
3986 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3987 support. It adds about 90K to the size of the library. The meanings of the
3988 inbuilt class escapes such as \d and \s have NOT been changed.
3990 16. Updated pcredemo.c to include calls to free() to release the memory for the
3991 compiled pattern.
3993 17. The generated file chartables.c was being created in the source directory
3994 instead of in the building directory. This caused the build to fail if the
3995 source directory was different from the building directory, and was
3996 read-only.
3998 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3999 file. No doubt somebody will tell me if they don't make sense... Also added
4000 Dan Mooney's comments about building on OpenVMS.
4002 19. Added support for partial matching via the PCRE_PARTIAL option for
4003 pcre_exec() and the \P data escape in pcretest.
4005 20. Extended pcretest with 3 new pattern features:
4007 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4008 write the compiled pattern to the file whose name is "rest-of-line".
4009 This is a straight binary dump of the data, with the saved pointer to
4010 the character tables forced to be NULL. The study data, if any, is
4011 written too. After writing, pcretest reads a new pattern.
4013 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4014 compiled pattern from the given file. There must not be any
4015 occurrences of "<" in the file name (pretty unlikely); if there are,
4016 pcretest will instead treat the initial "<" as a pattern delimiter.
4017 After reading in the pattern, pcretest goes on to read data lines as
4018 usual.
4020 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4021 and 16-bit fields in a compiled pattern, to simulate a pattern that
4022 was compiled on a host of opposite endianness.
4024 21. The pcre-exec() function can now cope with patterns that were compiled on
4025 hosts of opposite endianness, with this restriction:
4027 As for any compiled expression that is saved and used later, the tables
4028 pointer field cannot be preserved; the extra_data field in the arguments
4029 to pcre_exec() should be used to pass in a tables address if a value
4030 other than the default internal tables were used at compile time.
4032 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4033 now diagnosed as an error. Previously, most of the time, a negative number
4034 would have been treated as zero, but if in addition "ovector" was passed as
4035 NULL, a crash could occur.
4037 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4038 new versions from the libtool 1.5 distribution (the last one is a copy of
4039 a file called libtool.m4). This seems to have fixed the need to patch
4040 "configure" to support Darwin 1.3 (which I used to do). However, I still
4041 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4042 workstation).
4044 24. Changed the PCRE licence to be the more standard "BSD" licence.
4047 Version 4.5 01-Dec-03
4048 ---------------------
4050 1. There has been some re-arrangement of the code for the match() function so
4051 that it can be compiled in a version that does not call itself recursively.
4052 Instead, it keeps those local variables that need separate instances for
4053 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4054 needs to "recurse". Keeping track of where control must go is done by means
4055 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4056 hide most of the details from the main code, and operates only if
4057 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4058 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4059 operating.
4061 To make it easier for callers to provide specially tailored get/free
4062 functions for this usage, two new functions, pcre_stack_malloc, and
4063 pcre_stack_free, are used. They are always called in strict stacking order,
4064 and the size of block requested is always the same.
4066 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4067 PCRE has been compiled to use the stack or the heap for recursion. The
4068 -C option of pcretest uses this to show which version is compiled.
4070 A new data escape \S, is added to pcretest; it causes the amounts of store
4071 obtained and freed by both kinds of malloc/free at match time to be added
4072 to the output.
4074 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4075 what's available on my current Linux desktop machine.
4077 3. When matching a UTF-8 string, the test for a valid string at the start has
4078 been extended. If start_offset is not zero, PCRE now checks that it points
4079 to a byte that is the start of a UTF-8 character. If not, it returns
4080 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4081 this is necessary because there may be backward assertions in the pattern.
4082 When matching the same subject several times, it may save resources to use
4083 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4085 4. The code for checking the validity of UTF-8 strings has been tightened so
4086 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4087 containing "overlong sequences".
4089 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4090 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4091 should have been "&", but it just so happened that all the cases this let
4092 through by mistake were picked up later in the function.
4094 6. I had used a variable called "isblank" - this is a C99 function, causing
4095 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4097 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4098 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4099 so the version is shown; (c) stop "make test" echoing "./RunTest".
4101 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4103 9. The prototype for memmove() for systems that don't have it was using
4104 size_t, but the inclusion of the header that defines size_t was later. I've
4105 moved the #includes for the C headers earlier to avoid this.
4107 10. Added some adjustments to the code to make it easier to compiler on certain
4108 special systems:
4110 (a) Some "const" qualifiers were missing.
4111 (b) Added the macro EXPORT before all exported functions; by default this
4112 is defined to be empty.
4113 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4114 that it reads its output file name as an argument instead of writing
4115 to the standard output and assuming this can be redirected.
4117 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4118 class containing characters with values greater than 255, PCRE compilation
4119 went into a loop.
4121 12. A recursive reference to a subpattern that was within another subpattern
4122 that had a minimum quantifier of zero caused PCRE to crash. For example,
4123 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4124 recursion. If the recursively-called subpattern itself had a zero repeat,
4125 that was OK.
4127 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4128 buffer into which it was copied (for escape processing) was still set at
4129 1024, so long lines caused crashes.
4131 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4132 "internal error: code overflow...". This applied to any character class
4133 that was followed by a possessive quantifier.
4135 15. Modified the Makefile to add libpcre.la as a prerequisite for
4136 libpcreposix.la because I was told this is needed for a parallel build to
4137 work.
4139 16. If a pattern that contained .* following optional items at the start was
4140 studied, the wrong optimizing data was generated, leading to matching
4141 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4142 matching string must start with a or b or c. The correct conclusion for
4143 this pattern is that a match can start with any character.
4146 Version 4.4 13-Aug-03
4147 ---------------------
4149 1. In UTF-8 mode, a character class containing characters with values between
4150 127 and 255 was not handled correctly if the compiled pattern was studied.
4151 In fixing this, I have also improved the studying algorithm for such
4152 classes (slightly).
4154 2. Three internal functions had redundant arguments passed to them. Removal
4155 might give a very teeny performance improvement.
4157 3. Documentation bug: the value of the capture_top field in a callout is *one
4158 more than* the number of the hightest numbered captured substring.
4160 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4161 in incorrectly linking with a previously installed version. They now link
4162 explicitly with libpcre.la.
4164 5. configure.in no longer needs to recognize Cygwin specially.
4166 6. A problem in pcre.in for Windows platforms is fixed.
4168 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4169 pcretest, it used to include the size of the study block as part of its
4170 output. Unfortunately, the structure contains a field that has a different
4171 size on different hardware architectures. This meant that the tests that
4172 showed this size failed. As the block is currently always of a fixed size,
4173 this information isn't actually particularly useful in pcretest output, so
4174 I have just removed it.
4176 8. Three pre-processor statements accidentally did not start in column 1.
4177 Sadly, there are *still* compilers around that complain, even though
4178 standard C has not required this for well over a decade. Sigh.
4180 9. In pcretest, the code for checking callouts passed small integers in the
4181 callout_data field, which is a void * field. However, some picky compilers
4182 complained about the casts involved for this on 64-bit systems. Now
4183 pcretest passes the address of the small integer instead, which should get
4184 rid of the warnings.
4186 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4187 both compile and run time, and gives an error if an invalid UTF-8 sequence
4188 is found. There is a option for disabling this check in cases where the
4189 string is known to be correct and/or the maximum performance is wanted.
4191 11. In response to a bug report, I changed one line in Makefile.in from
4193 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4194 to
4195 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4197 to look similar to other lines, but I have no way of telling whether this
4198 is the right thing to do, as I do not use Windows. No doubt I'll get told
4199 if it's wrong...
4202 Version 4.3 21-May-03
4203 ---------------------
4205 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4206 Makefile.
4208 2. Some refactoring to improve the quality of the code:
4210 (i) The utf8_table... variables are now declared "const".
4212 (ii) The code for \cx, which used the "case flipping" table to upper case
4213 lower case letters, now just substracts 32. This is ASCII-specific,
4214 but the whole concept of \cx is ASCII-specific, so it seems
4215 reasonable.
4217 (iii) PCRE was using its character types table to recognize decimal and
4218 hexadecimal digits in the pattern. This is silly, because it handles
4219 only 0-9, a-f, and A-F, but the character types table is locale-
4220 specific, which means strange things might happen. A private
4221 table is now used for this - though it costs 256 bytes, a table is
4222 much faster than multiple explicit tests. Of course, the standard
4223 character types table is still used for matching digits in subject
4224 strings against \d.
4226 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4227 ending in _t are). So I've renamed it as ESC_tee.
4229 3. The first argument for regexec() in the POSIX wrapper should have been
4230 defined as "const".
4232 4. Changed pcretest to use malloc() for its buffers so that they can be
4233 Electric Fenced for debugging.
4235 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4236 to read one or more bytes before the start of the subject string. Often this
4237 had no effect on PCRE's behaviour, but in some circumstances it could
4238 provoke a segmentation fault.
4240 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4241 to try to read one or more bytes before the start of the subject string.
4243 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4244 UTF-8 support could misbehave in various ways if the subject string
4245 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4246 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4247 back over UTF-8 characters.)
4250 Version 4.2 14-Apr-03
4251 ---------------------
4253 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4255 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4256 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4257 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4258 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4259 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4261 Note: automatic setting of the BUILD variables is not yet working
4262 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4263 compile-time but not at link-time
4264 [LINK]: use for linking executables only
4265 make different versions for Windows and non-Windows
4266 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4267 libraries
4268 [LINK_FOR_BUILD]: new variable
4269 [OBJEXT]: use throughout
4270 [EXEEXT]: use throughout
4271 <winshared>: new target
4272 <wininstall>: new target
4273 <dftables.o>: use native compiler
4274 <dftables>: use native linker
4275 <install>: handle Windows platform correctly
4276 <clean>: ditto
4277 <check>: ditto
4278 copy DLL to top builddir before testing
4280 As part of these changes, -no-undefined was removed again. This was reported
4281 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4282 in any case.
4284 3. Some tidies to get rid of compiler warnings:
4286 . In the match_data structure, match_limit was an unsigned long int, whereas
4287 match_call_count was an int. I've made them both unsigned long ints.
4289 . In pcretest the fact that a const uschar * doesn't automatically cast to
4290 a void * provoked a warning.
4292 . Turning on some more compiler warnings threw up some "shadow" variables
4293 and a few more missing casts.
4295 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4296 option, a class that contained a single character with a value between 128
4297 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4299 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4300 option, a class that contained several characters, but with at least one
4301 whose value was between 128 and 255 caused PCRE to crash.
4304 Version 4.1 12-Mar-03
4305 ---------------------
4307 1. Compiling with gcc -pedantic found a couple of places where casts were
4308 needed, and a string in dftables.c that was longer than standard compilers are
4309 required to support.
4311 2. Compiling with Sun's compiler found a few more places where the code could
4312 be tidied up in order to avoid warnings.
4314 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4315 first of these names is deprecated in the latest Autoconf in favour of the name
4316 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4317 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4318 analogy I have changed it to CFLAGS_FOR_BUILD.
4320 4. Added -no-undefined to the linking command in the Makefile, because this is
4321 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4322 linking step for the pcreposix library.
4324 5. PCRE was failing to diagnose the case of two named groups with the same
4325 name.
4327 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4328 literal character that is needed in the subject for a match, and scans along to
4329 ensure that it is present before embarking on the full matching process. This
4330 saves time in cases of nested unlimited repeats that are never going to match.
4331 Problem: the scan can take a lot of time if the subject is very long (e.g.
4332 megabytes), thus penalizing straightforward matches. It is now done only if the
4333 amount of subject to be scanned is less than 1000 bytes.
4335 7. A lesser problem with the same optimization is that it was recording the
4336 first character of an anchored pattern as "needed", thus provoking a search
4337 right along the subject, even when the first match of the pattern was going to
4338 fail. The "needed" character is now not set for anchored patterns, unless it
4339 follows something in the pattern that is of non-fixed length. Thus, it still
4340 fulfils its original purpose of finding quick non-matches in cases of nested
4341 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4344 Version 4.0 17-Feb-03
4345 ---------------------
4347 1. If a comment in an extended regex that started immediately after a meta-item
4348 extended to the end of string, PCRE compiled incorrect data. This could lead to
4349 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4351 2. Moved to autoconf 2.53 and libtool 1.4.2.
4353 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4354 the special perltest8 script is no longer needed - all the tests can be run
4355 from a single perltest script.
4357 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4358 by \s. It has now been removed in PCRE. This means it isn't recognized as
4359 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4360 class [:space:] *does* include VT, thereby creating a mess.
4362 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4363 space and tab.
4365 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4366 its new features into the main test script, reducing the number of scripts.
4368 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4369 were backward compatible, and made the (?i) apply to the whole pattern, as if
4370 /i were given. Now it behaves more logically, and applies the option setting
4371 only to what follows. PCRE has been changed to follow suit. However, if it
4372 finds options settings right at the start of the pattern, it extracts them into
4373 the global options, as before. Thus, they show up in the info data.
4375 8. Added support for the \Q...\E escape sequence. Characters in between are
4376 treated as literals. This is slightly different from Perl in that $ and @ are
4377 also handled as literals inside the quotes. In Perl, they will cause variable
4378 interpolation. Note the following examples:
4380 Pattern PCRE matches Perl matches
4382 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4383 \Qabc\$xyz\E abc\$xyz abc\$xyz
4384 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4386 For compatibility with Perl, \Q...\E sequences are recognized inside character
4387 classes as well as outside them.
4389 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4390 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4391 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4392 signed/unsigned warnings.
4394 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4395 option for pcretest, so I've replaced it by a simple function that does just
4396 that job.
4398 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4399 "pcregrep -".
4401 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4402 Java package. This provides some syntactic sugar for simple cases of what my
4403 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4404 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4405 item, you can use this simplified notation. Note that only makes sense with
4406 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4407 greediness, whatever the setting of the PCRE_UNGREEDY option.
4409 13. A change of greediness default within a pattern was not taking effect at
4410 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4411 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4412 was abstracted outside.
4414 14. PCRE now supports the \G assertion. It is true when the current matching
4415 position is at the start point of the match. This differs from \A when the
4416 starting offset is non-zero. Used with the /g option of pcretest (or similar
4417 code), it works in the same way as it does for Perl's /g option. If all
4418 alternatives of a regex begin with \G, the expression is anchored to the start
4419 match position, and the "anchored" flag is set in the compiled expression.
4421 15. Some bugs concerning the handling of certain option changes within patterns
4422 have been fixed. These applied to options other than (?ims). For example,
4423 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4424 been the other way round. Some of this was related to change 7 above.
4426 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4427 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4428 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4429 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4431 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4432 mode. Unlike ".", it always matches newline, whatever the setting of
4433 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4434 assertions. Perl allows it, but it doesn't (in general) work because it can't
4435 calculate the length of the lookbehind. At least, that's the case for Perl
4436 5.8.0 - I've been told they are going to document that it doesn't work in
4437 future.
4439 18. Added an error diagnosis for escapes that PCRE does not support: these are
4440 \L, \l, \N, \P, \p, \U, \u, and \X.
4442 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4443 reading past the end of the pattern in cases such as /[abcd/.
4445 20. PCRE was getting more memory than necessary for patterns with classes that
4446 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4448 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4449 compiling PCRE for use with Virtual Pascal.
4451 22. Small fix to the Makefile to make it work properly if the build is done
4452 outside the source tree.
4454 23. Added a new extension: a condition to go with recursion. If a conditional
4455 subpattern starts with (?(R) the "true" branch is used if recursion has
4456 happened, whereas the "false" branch is used only at the top level.
4458 24. When there was a very long string of literal characters (over 255 bytes
4459 without UTF support, over 250 bytes with UTF support), the computation of how
4460 much memory was required could be incorrect, leading to segfaults or other
4461 strange effects.
4463 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4464 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4465 there was a subsequent back reference to those brackets. This meant that, for
4466 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4467 possible to check for precisely this case. All we can do is abandon the
4468 optimization if .* occurs inside capturing brackets when there are any back
4469 references whatsoever. (See below for a better fix that came later.)
4471 26. The handling of the optimization for finding the first character of a
4472 non-anchored pattern, and for finding a character that is required later in the
4473 match were failing in some cases. This didn't break the matching; it just
4474 failed to optimize when it could. The way this is done has been re-implemented.
4476 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4478 28. Added a new feature that provides some of the functionality that Perl
4479 provides with (?{...}). The facility is termed a "callout". The way it is done
4480 in PCRE is for the caller to provide an optional function, by setting
4481 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4482 global variable. By default it is unset, which disables all calling out. To get
4483 the function called, the regex must include (?C) at appropriate points. This
4484 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4485 This provides a means of identifying different callout points. When PCRE
4486 reaches such a point in the regex, if pcre_callout has been set, the external
4487 function is called. It is provided with data in a structure called
4488 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4489 matching continues; if it returns a non-zero value, the match at the current
4490 point fails. However, backtracking will occur if possible. [This was changed
4491 later and other features added - see item 49 below.]
4493 29. pcretest is upgraded to test the callout functionality. It provides a
4494 callout function that displays information. By default, it shows the start of
4495 the match and the current position in the text. There are some new data escapes
4496 to vary what happens:
4498 \C+ in addition, show current contents of captured substrings
4499 \C- do not supply a callout function
4500 \C!n return 1 when callout number n is reached
4501 \C!n!m return 1 when callout number n is reached for the mth time
4503 30. If pcregrep was called with the -l option and just a single file name, it
4504 output "<stdin>" if a match was found, instead of the file name.
4506 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4507 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4508 pcre_exec(). This saves a malloc/free per call. The default value of
4509 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4510 when configuring.
4512 32. The default maximum size of a compiled pattern is 64K. There have been a
4513 few cases of people hitting this limit. The code now uses macros to handle the
4514 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4515 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4516 configuring. Tests 2 and 5 work only with 2-byte links because they output
4517 debugging information about compiled patterns.
4519 33. Internal code re-arrangements:
4521 (a) Moved the debugging function for printing out a compiled regex into
4522 its own source file (printint.c) and used #include to pull it into
4523 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4524 separate copies.
4526 (b) Defined the list of op-code names for debugging as a macro in
4527 internal.h so that it is next to the definition of the opcodes.
4529 (c) Defined a table of op-code lengths for simpler skipping along compiled
4530 code. This is again a macro in internal.h so that it is next to the
4531 definition of the opcodes.
4533 34. Added support for recursive calls to individual subpatterns, along the
4534 lines of Robin Houston's patch (but implemented somewhat differently).
4536 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4537 allow it to read and process whole directories in Win32. This code was
4538 contributed by Lionel Fourquaux; it has not been tested by me.
4540 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4541 used to name a group. Names consist of alphanumerics and underscores, and must
4542 be unique. Back references use the syntax (?P=name) and recursive calls use
4543 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4544 numbers. The function pcre_fullinfo() can be used after compilation to extract
4545 a name/number map. There are three relevant calls:
4547 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4548 PCRE_INFO_NAMECOUNT yields the number of entries
4549 PCRE_INFO_NAMETABLE yields a pointer to the map.
4551 The map is a vector of fixed-size entries. The size of each entry depends on
4552 the length of the longest name used. The first two bytes of each entry are the
4553 group number, most significant byte first. There follows the corresponding
4554 name, zero terminated. The names are in alphabetical order.
4556 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4557 case instead of 255. Making it the same both with and without UTF-8 support
4558 means that the same test output works with both.
4560 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4561 calling malloc() with a zero argument.
4563 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4564 optimization. I've improved things by keeping a bitmap of backreferences with
4565 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4566 fact referenced, the optimization can be applied. It is unlikely that a
4567 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4568 the match to follow \n) will appear inside brackets with a number greater than
4569 31, but if it does, any back reference > 31 suppresses the optimization.
4571 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4572 of disabling numbered capturing parentheses. Any opening parenthesis that is
4573 not followed by ? behaves as if it were followed by ?: but named parentheses
4574 can still be used for capturing (and they will acquire numbers in the usual
4575 way).
4577 41. Redesigned the return codes from the match() function into yes/no/error so
4578 that errors can be passed back from deep inside the nested calls. A malloc
4579 failure while inside a recursive subpattern call now causes the
4580 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4582 42. It is now possible to set a limit on the number of times the match()
4583 function is called in a call to pcre_exec(). This facility makes it possible to
4584 limit the amount of recursion and backtracking, though not in a directly
4585 obvious way, because the match() function is used in a number of different
4586 circumstances. The count starts from zero for each position in the subject
4587 string (for non-anchored patterns). The default limit is, for compatibility, a
4588 large number, namely 10 000 000. You can change this in two ways:
4590 (a) When configuring PCRE before making, you can use --with-match-limit=n
4591 to set a default value for the compiled library.
4593 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4594 a different value is set. See 45 below.
4596 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4598 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4599 of things that can be changed at compile time. The first argument specifies
4600 what is wanted and the second points to where the information is to be placed.
4601 The current list of available information is:
4605 The output is an integer that is set to one if UTF-8 support is available;
4606 otherwise it is set to zero.
4610 The output is an integer that it set to the value of the code that is used for
4611 newline. It is either LF (10) or CR (13).
4615 The output is an integer that contains the number of bytes used for internal
4616 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4620 The output is an integer that contains the threshold above which the POSIX
4621 interface uses malloc() for output vectors. See item 31 above.
4625 The output is an unsigned integer that contains the default limit of the number
4626 of match() calls in a pcre_exec() execution. See 42 above.
4628 44. pcretest has been upgraded by the addition of the -C option. This causes it
4629 to extract all the available output from the new pcre_config() function, and to
4630 output it. The program then exits immediately.
4632 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4633 order to support additional features. One way would have been to define
4634 pcre_exec2() (for example) with extra arguments, but this would not have been
4635 extensible, and would also have required all calls to the original function to
4636 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4637 is used for passing in "extra" data from pcre_study().
4639 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4640 contains the following fields:
4642 flags a bitmap indicating which of the following fields are set
4643 study_data opaque data from pcre_study()
4644 match_limit a way of specifying a limit on match() calls for a specific
4645 call to pcre_exec()
4646 callout_data data for callouts (see 49 below)
4648 The flag bits are also defined in pcre.h, and are
4654 The pcre_study() function now returns one of these new pcre_extra blocks, with
4655 the actual study data pointed to by the study_data field, and the
4656 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4657 before. That is, this change is entirely upwards-compatible and requires no
4658 change to existing code.
4660 If you want to pass in additional data to pcre_exec(), you can either place it
4661 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4662 block.
4664 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4665 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4666 times with different match limits, until it finds the minimum value needed for
4667 pcre_exec() to complete. The value is then output. This can be instructive; for
4668 most simple matches the number is quite small, but for pathological cases it
4669 gets very large very quickly.
4671 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4672 returns the size of the data block pointed to by the study_data field in a
4673 pcre_extra block, that is, the value that was passed as the argument to
4674 pcre_malloc() when PCRE was getting memory in which to place the information
4675 created by pcre_study(). The fourth argument should point to a size_t variable.
4676 pcretest has been extended so that this information is shown after a successful
4677 pcre_study() call when information about the compiled regex is being displayed.
4679 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4680 because what follows is always an absolute path. (Later: it turns out that this
4681 is more than cosmetic for MinGW, because it doesn't like empty path
4682 components.)
4684 49. Some changes have been made to the callout feature (see 28 above):
4686 (i) A callout function now has three choices for what it returns:
4688 0 => success, carry on matching
4689 > 0 => failure at this point, but backtrack if possible
4690 < 0 => serious error, return this value from pcre_exec()
4692 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4693 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4694 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4695 use by callout functions. It will never be used by PCRE itself.
4697 (ii) The pcre_extra structure (see 45 above) has a void * field called
4698 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4699 pcre_callout_block structure has a field of the same name. The contents of
4700 the field passed in the pcre_extra structure are passed to the callout
4701 function in the corresponding field in the callout block. This makes it
4702 easier to use the same callout-containing regex from multiple threads. For
4703 testing, the pcretest program has a new data escape
4705 \C*n pass the number n (may be negative) as callout_data
4707 If the callout function in pcretest receives a non-zero value as
4708 callout_data, it returns that value.
4710 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4711 there were some redundant $(CFLAGS) in commands that are now specified as
4712 $(LINK), which already includes $(CFLAGS).
4714 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4715 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4716 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4717 one-byte characters throughout. Note that case-insensitive matching applies
4718 only to characters whose values are less than 256. PCRE doesn't support the
4719 notion of cases for higher-valued characters.
4721 (i) A character class whose characters are all within 0-255 is handled as
4722 a bit map, and the map is inverted for negative classes. Previously, a
4723 character > 255 always failed to match such a class; however it should
4724 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4726 (ii) A negated character class with a single character < 255 is coded as
4727 "not this character" (OP_NOT). This wasn't working properly when the test
4728 character was multibyte, either singly or repeated.
4730 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4731 mode, for example: \x{100}{2,3}.
4733 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4734 singly or repeated) now correctly test multibyte characters. However,
4735 PCRE doesn't recognize any characters with values greater than 255 as
4736 digits, spaces, or word characters. Such characters always match \D, \S,
4737 and \W, and never match \d, \s, or \w.
4739 (v) Classes may now contain characters and character ranges with values
4740 greater than 255. For example: [ab\x{100}-\x{400}].
4742 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4743 PCRE in UTF-8 mode.
4745 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4746 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4747 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4748 value.)
4750 53. The single man page has become too large. I have therefore split it up into
4751 a number of separate man pages. These also give rise to individual HTML pages;
4752 these are now put in a separate directory, and there is an index.html page that
4753 lists them all. Some hyperlinking between the pages has been installed.
4755 54. Added convenience functions for handling named capturing parentheses.
4757 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4758 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4759 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4760 are faulted.
4762 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4763 calling configure. These values are used when compiling the dftables.c program
4764 which is run to generate the source of the default character tables. They
4765 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4766 you will need to set these values.
4768 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4771 Version 3.9 02-Jan-02
4772 ---------------------
4774 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4776 2. If --disable-static was given, the building process failed when trying to
4777 build pcretest and pcregrep. (For some reason it was using libtool to compile
4778 them, which is not right, as they aren't part of the library.)
4781 Version 3.8 18-Dec-01
4782 ---------------------
4784 1. The experimental UTF-8 code was completely screwed up. It was packing the
4785 bytes in the wrong order. How dumb can you get?
4788 Version 3.7 29-Oct-01
4789 ---------------------
4791 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4792 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4793 this didn't happen under Solaris 8, where I normally test things.
4795 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4796 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4797 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4798 this fix an hour or so after the initial 3.7 release.)
4801 Version 3.6 23-Oct-01
4802 ---------------------
4804 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4805 offsets passed as NULL with zero offset count.
4807 2. The config.guess and config.sub files had not been updated when I moved to
4808 the latest autoconf.
4811 Version 3.5 15-Aug-01
4812 ---------------------
4814 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4815 had been forgotten.
4817 2. By using declared but undefined structures, we can avoid using "void"
4818 definitions in pcre.h while keeping the internal definitions of the structures
4819 private.
4821 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4822 user point of view, this means that both static and shared libraries are built
4823 by default, but this can be individually controlled. More of the work of
4824 handling this static/shared cases is now inside libtool instead of PCRE's make
4825 file.
4827 4. The pcretest utility is now installed along with pcregrep because it is
4828 useful for users (to test regexs) and by doing this, it automatically gets
4829 relinked by libtool. The documentation has been turned into a man page, so
4830 there are now .1, .txt, and .html versions in /doc.
4832 5. Upgrades to pcregrep:
4833 (i) Added long-form option names like gnu grep.
4834 (ii) Added --help to list all options with an explanatory phrase.
4835 (iii) Added -r, --recursive to recurse into sub-directories.
4836 (iv) Added -f, --file to read patterns from a file.
4838 6. pcre_exec() was referring to its "code" argument before testing that
4839 argument for NULL (and giving an error if it was NULL).
4841 7. Upgraded Makefile.in to allow for compiling in a different directory from
4842 the source directory.
4844 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4845 options bits, the pointer it was passed was to an int instead of to an unsigned
4846 long int. This mattered only on 64-bit systems.
4848 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4849 generated) instead of pcre.in, which it its source. Also made the same change
4850 in several of the .c files.
4852 10. A new release of gcc defines printf() as a macro, which broke pcretest
4853 because it had an ifdef in the middle of a string argument for printf(). Fixed
4854 by using separate calls to printf().
4856 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4857 script, to force use of CR or LF instead of \n in the source. On non-Unix
4858 systems, the value can be set in config.h.
4860 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4861 absolute limit. Changed the text of the error message to make this clear, and
4862 likewise updated the man page.
4864 13. The limit of 99 on the number of capturing subpatterns has been removed.
4865 The new limit is 65535, which I hope will not be a "real" limit.
4868 Version 3.4 22-Aug-00
4869 ---------------------
4871 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4873 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4876 Version 3.3 01-Aug-00
4877 ---------------------
4879 1. If an octal character was given, but the value was greater than \377, it
4880 was not getting masked to the least significant bits, as documented. This could
4881 lead to crashes in some systems.
4883 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4884 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4886 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4887 These just pass their arguments on to (pcre_free)(), but they are provided
4888 because some uses of PCRE bind it to non-C systems that can call its functions,
4889 but cannot call free() or pcre_free() directly.
4891 4. Add "make test" as a synonym for "make check". Corrected some comments in
4892 the Makefile.
4894 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4895 Makefile.
4897 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4898 command called pgrep for grepping around the active processes.
4900 7. Added the beginnings of support for UTF-8 character strings.
4902 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4903 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4904 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4905 out for the ar command.)
4908 Version 3.2 12-May-00
4909 ---------------------
4911 This is purely a bug fixing release.
4913 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4914 of ZA. This was just one example of several cases that could provoke this bug,
4915 which was introduced by change 9 of version 2.00. The code for breaking
4916 infinite loops after an iteration that matches an empty string was't working
4917 correctly.
4919 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4920 when matched against abbab (for example). After matching an empty string, it
4921 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4922 caused it to match further down the string than it should.
4924 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4925 was there because it doesn't seem to be needed, and it causes trouble on some
4926 systems, as it is not a Standard C header. It has been removed.
4928 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4929 were reported on the Macintosh. The changes were from
4931 while ((c = *(++ptr)) != 0 && c != '\n');
4932 to
4933 while ((c = *(++ptr)) != 0 && c != '\n') ;
4935 Totally extraordinary, but if that's what it takes...
4937 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4938 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4939 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4940 assumes the way PCRE uses memmove() (always moving upwards).
4942 6. PCRE is being used in one environment where strchr() is not available. There
4943 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4944 faster code anyway.
4947 Version 3.1 09-Feb-00
4948 ---------------------
4950 The only change in this release is the fixing of some bugs in Makefile.in for
4951 the "install" target:
4953 (1) It was failing to install pcreposix.h.
4955 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4958 Version 3.0 01-Feb-00
4959 ---------------------
4961 1. Add support for the /+ modifier to perltest (to output $` like it does in
4962 pcretest).
4964 2. Add support for the /g modifier to perltest.
4966 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4967 matches null strings.
4969 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4970 pattern. Perl treats empty patterns specially - it reuses the most recent
4971 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4972 effect.
4974 5. The POSIX interface was broken in that it was just handing over the POSIX
4975 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4976 required a bigger vector, with some working space on the end. This means that
4977 the POSIX wrapper now has to get and free some memory, and copy the results.
4979 6. Added some simple autoconf support, placing the test data and the
4980 documentation in separate directories, re-organizing some of the
4981 information files, and making it build pcre-config (a GNU standard). Also added
4982 libtool support for building PCRE as a shared library, which is now the
4983 default.
4985 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4986 09 are not valid octal constants. Single digits will be used for minor values
4987 less than 10.
4989 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4990 existing programs that set these in the POSIX interface can use PCRE without
4991 modification.
4993 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4994 return all that pcre_info() returns, plus additional data. The pcre_info()
4995 function is retained for compatibility, but is considered to be obsolete.
4997 10. Added experimental recursion feature (?R) to handle one common case that
4998 Perl 5.6 will be able to do with (?p{...}).
5000 11. Added support for POSIX character classes like [:alpha:], which Perl is
5001 adopting.
5004 Version 2.08 31-Aug-99
5005 ----------------------
5007 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5008 trying to match at the startoffset position, but instead was moving forward to
5009 the next newline as if a previous match had failed.
5011 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5012 and could get into a loop if a null string was matched other than at the start
5013 of the subject.
5015 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5016 be distinguished at compile time, and for completeness also added PCRE_DATE.
5018 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5019 in GnuWin32 environments.
5022 Version 2.07 29-Jul-99
5023 ----------------------
5025 1. The documentation is now supplied in plain text form and HTML as well as in
5026 the form of man page sources.
5028 2. C++ compilers don't like assigning (void *) values to other pointer types.
5029 In particular this affects malloc(). Although there is no problem in Standard
5030 C, I've put in casts to keep C++ compilers happy.
5032 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5033 should be (const char *).
5035 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5036 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5037 However, I haven't made this a standard facility. The documentation doesn't
5038 mention it, and the Makefile doesn't support it.
5040 5. The Makefile now contains an "install" target, with editable destinations at
5041 the top of the file. The pcretest program is not installed.
5043 6. pgrep -V now gives the PCRE version number and date.
5045 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5046 causing the entire string to be ignored, instead of just the last character.
5048 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5049 non-matching string, it can take a very, very long time, even for strings of
5050 quite modest length, because of the nested recursion. PCRE now does better in
5051 some of these cases. It does this by remembering the last required literal
5052 character in the pattern, and pre-searching the subject to ensure it is present
5053 before running the real match. In other words, it applies a heuristic to detect
5054 some types of certain failure quickly, and in the above example, if presented
5055 with a string that has no trailing " it gives "no match" very quickly.
5057 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5058 other alternatives are tried instead.
5061 Version 2.06 09-Jun-99
5062 ----------------------
5064 1. Change pcretest's output for amount of store used to show just the code
5065 space, because the remainder (the data block) varies in size between 32-bit and
5066 64-bit systems.
5068 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5069 start matching at. This allows lookbehinds to work when searching for multiple
5070 occurrences in a string.
5072 3. Added additional options to pcretest for testing multiple occurrences:
5074 /+ outputs the rest of the string that follows a match
5075 /g loops for multiple occurrences, using the new startoffset argument
5076 /G loops for multiple occurrences by passing an incremented pointer
5078 4. PCRE wasn't doing the "first character" optimization for patterns starting
5079 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5080 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5081 the letter 'x'. On long subject strings, this gives a significant speed-up.
5084 Version 2.05 21-Apr-99
5085 ----------------------
5087 1. Changed the type of magic_number from int to long int so that it works
5088 properly on 16-bit systems.
5090 2. Fixed a bug which caused patterns starting with .* not to work correctly
5091 when the subject string contained newline characters. PCRE was assuming
5092 anchoring for such patterns in all cases, which is not correct because .* will
5093 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5094 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5095 must be retried after every newline in the subject.
5098 Version 2.04 18-Feb-99
5099 ----------------------
5101 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5102 computation of the store needed to hold the pattern was incorrect (too large).
5103 If such patterns were nested a few deep, this could multiply and become a real
5104 problem.
5106 2. Added /M option to pcretest to show the memory requirement of a specific
5107 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5109 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5110 compiled in such a way that the backtracking after subsequent failure was
5111 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5112 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5115 Version 2.03 02-Feb-99
5116 ----------------------
5118 1. Fixed typo and small mistake in man page.
5120 2. Added 4th condition (GPL supersedes if conflict) and created separate
5121 LICENCE file containing the conditions.
5123 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5124 Perl, that is the internal \ allows the delimiter to be included in the
5125 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5126 the final delimiter, add \ to the end of the pattern (to test the error).
5128 4. Added the convenience functions for extracting substrings after a successful
5129 match. Updated pcretest to make it able to test these functions.
5132 Version 2.02 14-Jan-99
5133 ----------------------
5135 1. Initialized the working variables associated with each extraction so that
5136 their saving and restoring doesn't refer to uninitialized store.
5138 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5139 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5140 fix the problem.
5142 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5143 calls, and wasn't printing the correct value for compiling calls. Increased the
5144 default value of LOOPREPEAT, and the number of significant figures in the
5145 times.
5147 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5149 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5150 a building problem on Windows NT with a FAT file system.
5153 Version 2.01 21-Oct-98
5154 ----------------------
5156 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5157 to character tables built by pcre_maketables() in the current locale. If NULL
5158 is passed, the default tables are used.
5161 Version 2.00 24-Sep-98
5162 ----------------------
5164 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5165 it any more.
5167 2. Allow quantification of (?>) groups, and make it work correctly.
5169 3. The first character computation wasn't working for (?>) groups.
5171 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5172 end of the subject) and add 5.005's \z, which really does match only at the
5173 very end of the subject.
5175 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5177 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5178 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5179 localized options. All options to pcre_study() were also removed.
5181 7. Add other new features from 5.005:
5183 $(?<= positive lookbehind
5184 $(?<! negative lookbehind
5185 (?imsx-imsx) added the unsetting capability
5186 such a setting is global if at outer level; local otherwise
5187 (?imsx-imsx:) non-capturing groups with option setting
5188 (?(cond)re|re) conditional pattern matching
5190 A backreference to itself in a repeated group matches the previous
5191 captured string.
5193 8. General tidying up of studying (both automatic and via "study")
5194 consequential on the addition of new assertions.
5196 9. As in 5.005, unlimited repeated groups that could match an empty substring
5197 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5198 runtime if any iteration does actually match an empty substring.
5200 10. Include the RunTest script in the distribution.
5202 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5203 discrepancies, some of which were old and were also with respect to 5.004. They
5204 have now been fixed.
5207 Version 1.09 28-Apr-98
5208 ----------------------
5210 1. A negated single character class followed by a quantifier with a minimum
5211 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5212 program crashes, or just wrong answers. This did not apply to negated classes
5213 containing more than one character, or to minima other than one.
5216 Version 1.08 27-Mar-98
5217 ----------------------
5219 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5221 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5222 latter must appear before anything that relies on it in the pattern.
5225 Version 1.07 16-Feb-98
5226 ----------------------
5228 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5229 repeat of a potentially empty string).
5232 Version 1.06 23-Jan-98
5233 ----------------------
5235 1. Added Markus Oberhumer's little patches for C++.
5237 2. Literal strings longer than 255 characters were broken.
5240 Version 1.05 23-Dec-97
5241 ----------------------
5243 1. Negated character classes containing more than one character were failing if
5244 PCRE_CASELESS was set at run time.
5247 Version 1.04 19-Dec-97
5248 ----------------------
5250 1. Corrected the man page, where some "const" qualifiers had been omitted.
5252 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5253 input syntax.
5255 3. Fixed memory leak which occurred when a regex with back references was
5256 matched with an offsets vector that wasn't big enough. The temporary memory
5257 that is used in this case wasn't being freed if the match failed.
5259 4. Tidied pcretest to ensure it frees memory that it gets.
5261 5. Temporary memory was being obtained in the case where the passed offsets
5262 vector was exactly big enough.
5264 6. Corrected definition of offsetof() from change 5 below.
5266 7. I had screwed up change 6 below and broken the rules for the use of
5267 setjmp(). Now fixed.
5270 Version 1.03 18-Dec-97
5271 ----------------------
5273 1. A erroneous regex with a missing opening parenthesis was correctly
5274 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5275 on some systems.
5277 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5278 it was reported that one broken compiler failed on the former because "code" is
5279 also an independent variable.
5281 3. The erroneous regex a[]b caused an array overrun reference.
5283 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5284 fail on data ending with that character. (It was going on too far, and checking
5285 the next character, typically a binary zero.) This was specific to the
5286 optimized code for single-character negative classes.
5288 5. Added a contributed patch from the TIN world which does the following:
5290 + Add an undef for memmove, in case the the system defines a macro for it.
5292 + Add a definition of offsetof(), in case there isn't one. (I don't know
5293 the reason behind this - offsetof() is part of the ANSI standard - but
5294 it does no harm).
5296 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5297 most of the places where whitespace preceded '#'. I have given up and
5298 allowed the remaining 2 cases to be at the margin.
5300 + Rename some variables in pcre to eliminate shadowing. This seems very
5301 pedantic, but does no harm, of course.
5303 6. Moved the call to setjmp() into its own function, to get rid of warnings
5304 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5306 7. Constructs such as \d{8,} were compiling into the equivalent of
5307 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5308 outcome, but in this particular case used more store than had been allocated,
5309 which caused the bug to be discovered because it threw up an internal error.
5311 8. The debugging code in both pcre and pcretest for outputting the compiled
5312 form of a regex was going wrong in the case of back references followed by
5313 curly-bracketed repeats.
5316 Version 1.02 12-Dec-97
5317 ----------------------
5319 1. Typos in pcre.3 and comments in the source fixed.
5321 2. Applied a contributed patch to get rid of places where it used to remove
5322 'const' from variables, and fixed some signed/unsigned and uninitialized
5323 variable warnings.
5325 3. Added the "runtest" target to Makefile.
5327 4. Set default compiler flag to -O2 rather than just -O.
5330 Version 1.01 19-Nov-97
5331 ----------------------
5333 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5334 like /([ab]*)*/, that is, for classes with more than one character in them.
5336 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5337 as /((?>a*))*/ (a PCRE_EXTRA facility).
5340 Version 1.00 18-Nov-97
5341 ----------------------
5343 1. Added compile-time macros to support systems such as SunOS4 which don't have
5344 memmove() or strerror() but have other things that can be used instead.
5346 2. Arranged that "make clean" removes the executables.
5349 Version 0.99 27-Oct-97
5350 ----------------------
5352 1. Fixed bug in code for optimizing classes with only one character. It was
5353 initializing a 32-byte map regardless, which could cause it to run off the end
5354 of the memory it had got.
5356 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5359 Version 0.98 22-Oct-97
5360 ----------------------
5362 1. Fixed bug in code for handling temporary memory usage when there are more
5363 back references than supplied space in the ovector. This could cause segfaults.
5366 Version 0.97 21-Oct-97
5367 ----------------------
5369 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5371 2. Optimized negated single characters not to use a bit map.
5373 3. Brought error texts together as macro definitions; clarified some of them;
5374 fixed one that was wrong - it said "range out of order" when it meant "invalid
5375 escape sequence".
5377 4. Changed some char * arguments to const char *.
5379 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5381 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5382 pcretest.
5385 Version 0.96 16-Oct-97
5386 ----------------------
5388 1. Added a simple "pgrep" utility to the distribution.
5390 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5391 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5392 where "ddd" means "one or more decimal digits".
5394 3. Fixed serious bug. If a pattern had a back reference, but the call to
5395 pcre_exec() didn't supply a large enough ovector to record the related
5396 identifying subpattern, the match always failed. PCRE now remembers the number
5397 of the largest back reference, and gets some temporary memory in which to save
5398 the offsets during matching if necessary, in order to ensure that
5399 backreferences always work.
5401 4. Increased the compatibility with Perl in a number of ways:
5403 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5404 to request this handling. The option can be set at compile or exec time.
5406 (b) $ matches before a terminating newline by default; an option
5407 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5408 mode). The option can be set at compile or exec time.
5410 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5411 the same as Perl's. If the decimal number it represents is less than 10
5412 or there aren't that many previous left capturing parentheses, an octal
5413 escape is read. Inside a character class, it's always an octal escape,
5414 even if it is a single digit.
5416 (d) An escaped but undefined alphabetic character is taken as a literal,
5417 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5418 escapes.
5420 (e) {0} is now permitted. (The previous item is removed from the compiled
5421 pattern).
5423 5. Changed all the names of code files so that the basic parts are no longer
5424 than 10 characters, and abolished the teeny "globals.c" file.
5426 6. Changed the handling of character classes; they are now done with a 32-byte
5427 bit map always.
5429 7. Added the -d and /D options to pcretest to make it possible to look at the
5430 internals of compilation without having to recompile pcre.
5433 Version 0.95 23-Sep-97
5434 ----------------------
5436 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5437 \x20 at the start of a run of normal characters. These were being treated as
5438 real characters, instead of the source characters being re-checked.
5441 Version 0.94 18-Sep-97
5442 ----------------------
5444 1. The functions are now thread-safe, with the caveat that the global variables
5445 containing pointers to malloc() and free() or alternative functions are the
5446 same for all threads.
5448 2. Get pcre_study() to generate a bitmap of initial characters for non-
5449 anchored patterns when this is possible, and use it if passed to pcre_exec().
5452 Version 0.93 15-Sep-97
5453 ----------------------
5455 1. /(b)|(:+)/ was computing an incorrect first character.
5457 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5458 but not actually doing anything yet.
5460 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5461 as Perl does (e.g. [-az] or [az-]).
5463 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5464 all possible positions.
5466 5. Split up into different modules to avoid including unneeded functions in a
5467 compiled binary. However, compile and exec are still in one module. The "study"
5468 function is split off.
5470 6. The character tables are now in a separate module whose source is generated
5471 by an auxiliary program - but can then be edited by hand if required. There are
5472 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5473 toupper() in the code.
5475 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5476 make them global. Abolish the function for setting them, as the caller can now
5477 set them directly.
5480 Version 0.92 11-Sep-97
5481 ----------------------
5483 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5484 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5486 2. Caseless matching was not working in character classes if the characters in
5487 the pattern were in upper case.
5489 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5491 4. Make PCRE_ANCHORED public and accept as a compile option.
5493 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5494 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5495 pass them.
5497 6. Give an error if bad option bits passed at compile or run time.
5499 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5500 pcretest to cause it to pass that flag.
5502 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5503 options, and the first character, if set.
5505 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5508 Version 0.91 10-Sep-97
5509 ----------------------
5511 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5512 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5514 2. PCRE was looping on encountering an indefinitely repeated back reference to
5515 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5516 Perl does - treats the match as successful.
5518 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5