ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1441 - (show annotations)
Sat Jan 11 21:56:18 2014 UTC (6 years, 1 month ago) by zherczeg
File size: 261442 byte(s)
Error occurred while calculating annotation data.
Fix wrong ChangeLog merge
1 ChangeLog for PCRE
2 ------------------
4 Version 8.35-RC1 xx-xxxx-201x
5 -----------------------------
7 1. A new flag is set, when property checks are present in an XCLASS.
8 When this flag is not set, PCRE can perform certain optimizations
9 such as studying these XCLASS-es.
11 2. The auto-possessification of character sets were improved: a normal
12 and an extended character set can be compared now. Furthermore
13 the JIT compiler optimizes more character set checks.
15 3. Got rid of some compiler warnings for potentially uninitialized variables
16 that show up only when compiled with -O2.
18 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
19 of a match later then the end of the match. The pcretest program was not
20 handling the case sensibly - it was outputting from the start to the next
21 binary zero. It now reports this situation in a message, and outputs the
22 text from the end to the start.
24 5. Fast forward search is improved in JIT. Instead of the first three
25 characters, any three characters with fixed position can be searched.
26 Search order: first, last, middle.
28 6. Improve character range checks in JIT. Characters are read by an inprecise
29 function now, which returns with an unknown value if the character code is
30 above a certain treshold (e.g: 256). The only limitation is that the value
31 must be bigger than the treshold as well. This function is useful, when
32 the characters above the treshold are handled in the same way.
34 7. The macros whose names start with RAWUCHAR are placeholders for a future
35 mode in which only the bottom 21 bits of 32-bit data items are used. To
36 make this more memorable for those maintaining the code, the names have
37 been changed to start with UCHAR21, and an extensive comment has been added
38 to their definition.
40 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
41 to the export list in Makefile.am (they were accidentally omitted from the
42 8.34 tarball).
44 9. The informational output from pcretest used the phrase "starting byte set"
45 which is inappropriate for the 16-bit and 32-bit libraries. As the output
46 for "first char" and "need char" really means "non-UTF-char", I've changed
47 "byte" to "char", and slightly reworded the output. The documentation about
48 these values has also been (I hope) clarified.
50 10. Another JIT related optimization: use table jumps for selecting the correct
51 backtracking path, when more than four alternatives are present inside a
52 bracket.
54 11. Empty match is not possible, when the minimum length is greater than zero,
55 and there is no \K in the pattern. JIT should avoid empty match checks in
56 such cases.
58 12. In a caseless character class with UCP support, when a character with more
59 than one alternative case was not the first character of a range, not all
60 the alternative cases were added to the class. For example, s and \x{17f}
61 are both alternative cases for S: the class [RST] was handled correctly,
62 but [R-T] was not.
64 13. The configure.ac file always checked for pthread support when JIT was
65 enabled. This is not used in Windows, so I have put this test inside a
66 check for the presence of windows.h (which was already tested for).
68 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
69 The algorithm provides a way to skip certain starting offsets, and usually
70 faster than linear prefix searches.
73 Version 8.34 15-December-2013
74 -----------------------------
76 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
77 executable memory. Patch inspired by Carsten Klein.
79 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
80 this macro is never tested and has no effect, because the work to support
81 coverage involves only compiling and linking options and special targets in
82 the Makefile. The comment in config.h implied that defining the macro would
83 enable coverage support, which is totally false. There was also support for
84 setting this macro in the CMake files (my fault, I just copied it from
85 configure). SUPPORT_GCOV has now been removed.
87 3. Make a small performance improvement in strlen16() and strlen32() in
88 pcretest.
90 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
91 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
92 these cases). There was also one in pcretest.c.
94 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
96 6. In UTF mode, the code for checking whether a group could match an empty
97 string (which is used for indefinitely repeated groups to allow for
98 breaking an infinite loop) was broken when the group contained a repeated
99 negated single-character class with a character that occupied more than one
100 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
101 UTF-8 mode). The effect was undefined: the group might or might not be
102 deemed as matching an empty string, or the program might have crashed.
104 7. The code for checking whether a group could match an empty string was not
105 recognizing that \h, \H, \v, \V, and \R must match a character.
107 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
108 an empty string. If it can, pcretest shows this in its information output.
110 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
111 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
112 matched by pcre_exec() without using JIT:
114 (a) If the rest of the pattern did not match after a maximal run of
115 grapheme clusters, the code for backing up to try with fewer of them
116 did not always back up over a full grapheme when characters that do not
117 have the modifier quality were involved, e.g. Hangul syllables.
119 (b) If the match point in a subject started with modifier character, and
120 there was no match, the code could incorrectly back up beyond the match
121 point, and potentially beyond the first character in the subject,
122 leading to a segfault or an incorrect match result.
124 10. A conditional group with an assertion condition could lead to PCRE
125 recording an incorrect first data item for a match if no other first data
126 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
127 first data item, and therefore matched "ca" after "c" instead of at the
128 start.
130 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
131 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
133 12. The source of pcregrep now includes z/OS-specific code so that it can be
134 compiled for z/OS as part of the special z/OS distribution.
136 13. Added the -T and -TM options to pcretest.
138 14. The code in pcre_compile.c for creating the table of named capturing groups
139 has been refactored. Instead of creating the table dynamically during the
140 actual compiling pass, the information is remembered during the pre-compile
141 pass (on the stack unless there are more than 20 named groups, in which
142 case malloc() is used) and the whole table is created before the actual
143 compile happens. This has simplified the code (it is now nearly 150 lines
144 shorter) and prepared the way for better handling of references to groups
145 with duplicate names.
147 15. A back reference to a named subpattern when there is more than one of the
148 same name now checks them in the order in which they appear in the pattern.
149 The first one that is set is used for the reference. Previously only the
150 first one was inspected. This change makes PCRE more compatible with Perl.
152 16. Unicode character properties were updated from Unicode 6.3.0.
154 17. The compile-time code for auto-possessification has been refactored, based
155 on a patch by Zoltan Herczeg. It now happens after instead of during
156 compilation. The code is cleaner, and more cases are handled. The option
157 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
158 options in pcretest are provided to set it. It can also be set by
159 (*NO_AUTO_POSSESS) at the start of a pattern.
161 18. The character VT has been added to the default ("C" locale) set of
162 characters that match \s and are generally treated as white space,
163 following this same change in Perl 5.18. There is now no difference between
164 "Perl space" and "POSIX space". Whether VT is treated as white space in
165 other locales depends on the locale.
167 19. The code for checking named groups as conditions, either for being set or
168 for being recursed, has been refactored (this is related to 14 and 15
169 above). Processing unduplicated named groups should now be as fast at
170 numerical groups, and processing duplicated groups should be faster than
171 before.
173 20. Two patches to the CMake build system, by Alexander Barkov:
175 (1) Replace the "source" command by "." in CMakeLists.txt because
176 "source" is a bash-ism.
178 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
179 without these the CMake build does not work on Solaris.
181 21. Perl has changed its handling of \8 and \9. If there is no previously
182 encountered capturing group of those numbers, they are treated as the
183 literal characters 8 and 9 instead of a binary zero followed by the
184 literals. PCRE now does the same.
186 22. Following Perl, added \o{} to specify codepoints in octal, making it
187 possible to specify values greater than 0777 and also making them
188 unambiguous.
190 23. Perl now gives an error for missing closing braces after \x{... instead of
191 treating the string as literal. PCRE now does the same.
193 24. RunTest used to grumble if an inappropriate test was selected explicitly,
194 but just skip it when running all tests. This make it awkward to run ranges
195 of tests when one of them was inappropriate. Now it just skips any
196 inappropriate tests, as it always did when running all tests.
198 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
199 character types such as \d or \w, too many callouts were inserted, and the
200 data that they returned was rubbish.
202 26. In UCP mode, \s was not matching two of the characters that Perl matches,
203 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
204 were matched by \h. The code has now been refactored so that the lists of
205 the horizontal and vertical whitespace characters used for \h and \v (which
206 are defined only in one place) are now also used for \s.
208 27. Add JIT support for the 64 bit TileGX architecture.
209 Patch by Jiong Wang (Tilera Corporation).
211 28. Possessive quantifiers for classes (both explicit and automatically
212 generated) now use special opcodes instead of wrapping in ONCE brackets.
214 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
215 (because it's meaningless), this was not happening when PCRE_CASELESS was
216 set. Not wrong, but inefficient.
218 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
219 properties for \w, \d, etc) is present in a test regex. Otherwise if the
220 test contains no characters greater than 255, Perl doesn't realise it
221 should be using Unicode semantics.
223 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
224 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
225 does in Unicode mode.
227 32. Added the "forbid" facility to pcretest so that putting tests into the
228 wrong test files can sometimes be quickly detected.
230 33. There is now a limit (default 250) on the depth of nesting of parentheses.
231 This limit is imposed to control the amount of system stack used at compile
232 time. It can be changed at build time by --with-parens-nest-limit=xxx or
233 the equivalent in CMake.
235 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
236 errors. Perl warns for these when in warning mode, but PCRE has no facility
237 for giving warnings.
239 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
240 However, this was not working for (?!) because it is optimized to (*FAIL),
241 for which PCRE does not allow quantifiers. The optimization is now disabled
242 when a quantifier follows (?!). I can't see any use for this, but it makes
243 things uniform.
245 36. Perl no longer allows group names to start with digits, so I have made this
246 change also in PCRE. It simplifies the code a bit.
248 37. In extended mode, Perl ignores spaces before a + that indicates a
249 possessive quantifier. PCRE allowed a space before the quantifier, but not
250 before the possessive +. It now does.
252 38. The use of \K (reset reported match start) within a repeated possessive
253 group such as (a\Kb)*+ was not working.
255 40. Document that the same character tables must be used at compile time and
256 run time, and that the facility to pass tables to pcre_exec() and
257 pcre_dfa_exec() is for use only with saved/restored patterns.
259 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
260 features for Builds with MSVC:
262 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
263 builds). The .rc files can be used to set FileDescription and many other
264 attributes.
266 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
267 This allows higher-level build scripts which want .pdb files to avoid
268 hard-coding the exact files needed."
270 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
271 mean "start of word" and "end of word", respectively, as a transition aid.
273 43. A minimizing repeat of a class containing codepoints greater than 255 in
274 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
275 compiled to use the heap for recursion.
277 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
278 is configured.
281 Version 8.33 28-May-2013
282 ------------------------
284 1. Added 'U' to some constants that are compared to unsigned integers, to
285 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
286 variables that are added to signed variables, to ensure the result is
287 signed and can be negated.
289 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
290 CMake config files.
292 3. Revise the creation of config.h.generic so that all boolean macros are
293 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
294 overriding via -D on the command line possible.
296 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
297 to unsigned int is reported to make a quite noticeable speed difference in
298 a specific Windows environment. Testing on Linux did also appear to show
299 some benefit (and it is clearly not harmful). Also fixed the definition of
300 Xop which should be unsigned.
302 5. Related to (4), changing the definition of the intermediate variable cc
303 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
304 improvements.
306 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
307 unnecessary spaces.
309 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
310 and later.
312 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
314 9. Optimizing fast_forward_start_bits in JIT.
316 10. Adding support for callouts in JIT, and fixing some issues revealed
317 during this work. Namely:
319 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
321 (b) Minimum length was not checked before the matching is started.
323 11. The value of capture_last that is passed to callouts was incorrect in some
324 cases when there was a capture on one path that was subsequently abandoned
325 after a backtrack. Also, the capture_last value is now reset after a
326 recursion, since all captures are also reset in this case.
328 12. The interpreter no longer returns the "too many substrings" error in the
329 case when an overflowing capture is in a branch that is subsequently
330 abandoned after a backtrack.
332 13. In the pathological case when an offset vector of size 2 is used, pcretest
333 now prints out the matched string after a yield of 0 or 1.
335 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
336 Only supported by the JIT compiler at the moment.
338 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
340 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
341 offset of the starting point of the matching process, provided the offsets
342 vector is large enough.
344 17. The \A escape now records a lookbehind value of 1, though its execution
345 does not actually inspect the previous character. This is to ensure that,
346 in partial multi-segment matching, at least one character from the old
347 segment is retained when a new segment is processed. Otherwise, if there
348 are no lookbehinds in the pattern, \A might match incorrectly at the start
349 of a new segment.
351 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
353 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
354 gives some modest performance improvement in 8-bit mode.
356 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
357 be expressed in certain programming languages using Universal Character
358 Names.
360 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
361 which points out that "non characters" are not "characters that may not
362 appear in Unicode strings" but rather "characters that are reserved for
363 internal use and have only local meaning".
365 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
366 there was a conditional group that depended on an assertion, if the
367 assertion was false, the callout that immediately followed the alternation
368 in the condition was skipped when pcre_exec() was used for matching.
370 23. Allow an explicit callout to be inserted before an assertion that is the
371 condition for a conditional group, for compatibility with automatic
372 callouts, which always insert a callout at this point.
374 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
375 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
377 25. (*PRUNE) is now supported by the JIT compiler.
379 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
381 27. Fix the case where there are two or more SKIPs with arguments that may be
382 ignored.
384 28. (*SKIP) is now supported by the JIT compiler.
386 29. (*THEN) is now supported by the JIT compiler.
388 30. Update RunTest with additional test selector options.
390 31. The way PCRE handles backtracking verbs has been changed in two ways.
392 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
393 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
394 backtracking. In some cases this makes it more Perl-compatible, but Perl's
395 rather obscure rules do not always do the same thing.
397 (2) Previously, backtracking verbs were confined within assertions. This is
398 no longer the case for positive assertions, except for (*ACCEPT). Again,
399 this sometimes improves Perl compatibility, and sometimes does not.
401 32. A number of tests that were in test 2 because Perl did things differently
402 have been moved to test 1, because either Perl or PCRE has changed, and
403 these tests are now compatible.
405 32. Backtracking control verbs are now handled in the same way in JIT and
406 interpreter.
408 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
409 contained a forward subroutine reference caused a compile error.
411 34. Auto-detect and optimize limited repetitions in JIT.
413 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
414 blocking (*UTF) etc.
416 36. In the interpreter, maximizing pattern repetitions for characters and
417 character types now use tail recursion, which reduces stack usage.
419 37. The value of the max lookbehind was not correctly preserved if a compiled
420 and saved regex was reloaded on a host of different endianness.
422 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
423 of the compiled pattern block, expand the flags field from 16 to 32 bits
424 because it was almost full.
426 39. Try madvise first before posix_madvise.
428 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
429 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
430 original change locked out all matches of empty strings. This has been
431 changed so that one match of an empty string per line is recognized.
432 Subsequent searches on the same line (for colouring or for --only-matching,
433 for example) do not recognize empty strings.
435 41. Applied a user patch to fix a number of spelling mistakes in comments.
437 42. Data lines longer than 65536 caused pcretest to crash.
439 43. Clarified the data type for length and startoffset arguments for pcre_exec
440 and pcre_dfa_exec in the function-specific man pages, where they were
441 explicitly stated to be in bytes, never having been updated. I also added
442 some clarification to the pcreapi man page.
444 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
445 a segmentation fault.
448 Version 8.32 30-November-2012
449 -----------------------------
451 1. Improved JIT compiler optimizations for first character search and single
452 character iterators.
454 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
455 Patch by Daniel Richard G.
457 3. Single character iterator optimizations in the JIT compiler.
459 4. Improved JIT compiler optimizations for character ranges.
461 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
462 Reported by Giuseppe D'Angelo.
464 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
465 of a line, was being set incorrectly in cases where .* appeared inside
466 atomic brackets at the start of a pattern, or where there was a subsequent
467 *PRUNE or *SKIP.
469 7. Improved instruction cache flush for POWER/PowerPC.
470 Patch by Daniel Richard G.
472 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
473 grep:
475 (a) There is now no limit to the number of patterns to be matched.
477 (b) An error is given if a pattern is too long.
479 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
480 are now supported.
482 (d) --exclude-from and --include-from (multiple use) have been added.
484 (e) Exclusions and inclusions now apply to all files and directories, not
485 just to those obtained from scanning a directory recursively.
487 (f) Multiple uses of -f and --file-list are now supported.
489 (g) In a Windows environment, the default for -d has been changed from
490 "read" (the GNU grep default) to "skip", because otherwise the presence
491 of a directory in the file list provokes an error.
493 (h) The documentation has been revised and clarified in places.
495 9. Improve the matching speed of capturing brackets.
497 10. Changed the meaning of \X so that it now matches a Unicode extended
498 grapheme cluster.
500 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
501 out POSIX threads when JIT support is configured.
503 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
505 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
506 structure could go wrong in environments where size_t is not the same size
507 as int.
509 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
510 set.
512 15. The EBCDIC support had decayed; later updates to the code had included
513 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
514 general tidy up of EBCDIC-related issues, and the documentation was also
515 not quite right. There is now a test that can be run on ASCII systems to
516 check some of the EBCDIC-related things (but is it not a full test).
518 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
519 in a small tidy to the code.
521 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
523 18. If the --only-matching (-o) option in pcregrep is specified multiple
524 times, each one causes appropriate output. For example, -o1 -o2 outputs the
525 substrings matched by the 1st and 2nd capturing parentheses. A separating
526 string can be specified by --om-separator (default empty).
528 19. Improving the first n character searches.
530 20. Turn case lists for horizontal and vertical white space into macros so that
531 they are defined only once.
533 21. This set of changes together give more compatible Unicode case-folding
534 behaviour for characters that have more than one other case when UCP
535 support is available.
537 (a) The Unicode property table now has offsets into a new table of sets of
538 three or more characters that are case-equivalent. The MultiStage2.py
539 script that generates these tables (the pcre_ucd.c file) now scans
540 CaseFolding.txt instead of UnicodeData.txt for character case
541 information.
543 (b) The code for adding characters or ranges of characters to a character
544 class has been abstracted into a generalized function that also handles
545 case-independence. In UTF-mode with UCP support, this uses the new data
546 to handle characters with more than one other case.
548 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
549 whose other case is greater than 256 are now correctly matched
550 caselessly. Previously, the high codepoint matched the low one, but not
551 vice versa.
553 (d) The processing of \h, \H, \v, and \ in character classes now makes use
554 of the new class addition function, using character lists defined as
555 macros alongside the case definitions of 20 above.
557 (e) Caseless back references now work with characters that have more than
558 one other case.
560 (f) General caseless matching of characters with more than one other case
561 is supported.
563 22. Unicode character properties were updated from Unicode 6.2.0
565 23. Improved CMake support under Windows. Patch by Daniel Richard G.
567 24. Add support for 32-bit character strings, and UTF-32
569 25. Major JIT compiler update (code refactoring and bugfixing).
570 Experimental Sparc 32 support is added.
572 26. Applied a modified version of Daniel Richard G's patch to create
573 pcre.h.generic and config.h.generic by "make" instead of in the
574 PrepareRelease script.
576 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
577 pcre_compile.c when checking for a zero character.
579 28. Introducing a native interface for JIT. Through this interface, the compiled
580 machine code can be directly executed. The purpose of this interface is to
581 provide fast pattern matching, so several sanity checks are not performed.
582 However, feature tests are still performed. The new interface provides
583 1.4x speedup compared to the old one.
585 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
586 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
587 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
589 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
590 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
591 the "old" RFC 2279). Instead, it ended up passing a negative length to
592 pcre_exec().
594 31. Add support for GCC's visibility feature to hide internal functions.
596 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
597 "unknown -C option" after outputting 0 or 1.
599 33. There is now support for generating a code coverage report for the test
600 suite in environments where gcc is the compiler and lcov is installed. This
601 is mainly for the benefit of the developers.
603 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
604 unaddressable using valgrind annotations, allowing valgrind to detect
605 invalid memory accesses. This is mainly for the benefit of the developers.
607 25. (*UTF) can now be used to start a pattern in any of the three libraries.
609 26. Give configure error if --enable-cpp but no C++ compiler found.
612 Version 8.31 06-July-2012
613 -------------------------
615 1. Fixing a wrong JIT test case and some compiler warnings.
617 2. Removed a bashism from the RunTest script.
619 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
620 to unsigned type, result still unsigned" that was given by an MS compiler
621 on encountering the code "-sizeof(xxx)".
623 4. Partial matching support is added to the JIT compiler.
625 5. Fixed several bugs concerned with partial matching of items that consist
626 of more than one character:
628 (a) /^(..)\1/ did not partially match "aba" because checking references was
629 done on an "all or nothing" basis. This also applied to repeated
630 references.
632 (b) \R did not give a hard partial match if \r was found at the end of the
633 subject.
635 (c) \X did not give a hard partial match after matching one or more
636 characters at the end of the subject.
638 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
639 a partial match for the string "\r".
641 (e) When newline was set to CRLF, the metacharacter "." did not recognize
642 a partial match for a CR character at the end of the subject string.
644 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
645 running pcretest, the text "(JIT)" added to the output whenever JIT is
646 actually used to run the match.
648 7. Individual JIT compile options can be set in pcretest by following -s+[+]
649 or /S+[+] with a digit between 1 and 7.
651 8. OP_NOT now supports any UTF character not just single-byte ones.
653 9. (*MARK) control verb is now supported by the JIT compiler.
655 10. The command "./RunTest list" lists the available tests without actually
656 running any of them. (Because I keep forgetting what they all are.)
660 12. Applied a (slightly modified) user-supplied patch that improves performance
661 when the heap is used for recursion (compiled with --disable-stack-for-
662 recursion). Instead of malloc and free for each heap frame each time a
663 logical recursion happens, frames are retained on a chain and re-used where
664 possible. This sometimes gives as much as 30% improvement.
666 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
667 call.
669 14. As documented, (*COMMIT) is now confined to within a positive assertion.
671 15. It is now possible to link pcretest with libedit as an alternative to
672 libreadline.
674 16. (*COMMIT) control verb is now supported by the JIT compiler.
676 17. The Unicode data tables have been updated to Unicode 6.1.0.
678 18. Added --file-list option to pcregrep.
680 19. Added binary file support to pcregrep, including the -a, --binary-files,
681 -I, and --text options.
683 20. The madvise function is renamed for posix_madvise for QNX compatibility
684 reasons. Fixed by Giuseppe D'Angelo.
686 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
688 22. Changed the option for creating symbolic links for 16-bit man pages from
689 -s to -sf so that re-installing does not cause issues.
691 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
693 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
694 matches in certain environments (the workspace was not being correctly
695 retained). Also added to pcre_dfa_exec() a simple plausibility check on
696 some of the workspace data at the beginning of a restart.
698 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
699 was not doing so when it should - probably a typo introduced by SVN 528
700 (change 8.10/14).
702 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
703 \w+ when the character tables indicated that \x{c4} was a word character.
704 There were several related cases, all because the tests for doing a table
705 lookup were testing for characters less than 127 instead of 255.
707 27. If a pattern contains capturing parentheses that are not used in a match,
708 their slots in the ovector are set to -1. For those that are higher than
709 any matched groups, this happens at the end of processing. In the case when
710 there were back references that the ovector was too small to contain
711 (causing temporary malloc'd memory to be used during matching), and the
712 highest capturing number was not used, memory off the end of the ovector
713 was incorrectly being set to -1. (It was using the size of the temporary
714 memory instead of the true size.)
716 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
717 ovector size, it uses memory at the end of the block that it has got.
719 29. Check for an overlong MARK name and give an error at compile time. The
720 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
722 30. JIT compiler update.
724 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
725 Rill for the patch.
727 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
729 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
731 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
732 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
733 was enabled.
735 35. Improve JIT code generation for greedy plus quantifier.
737 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
738 1 to "aa" instead of to an empty string. The bug affected repeated groups
739 that could potentially match an empty string.
741 37. Optimizing single character iterators in JIT.
743 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
744 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
745 codepoints that are too big for the mode are faulted, and in a UTF mode,
746 disallowed codepoints are also faulted.
748 39. If PCRE was compiled with UTF support, in three places in the DFA
749 matcher there was code that should only have been obeyed in UTF mode, but
750 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
751 processing when bytes with values greater than 127 were present. In 16-bit
752 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
753 both cases the values are those that cannot be the first data item in a UTF
754 character. The three items that might have provoked this were recursions,
755 possessively repeated groups, and atomic groups.
757 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
758 and pcregrep, because some OS require shared objects to be explicitly
759 passed to ld, causing the link step to fail if they are not.
761 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
762 mode, patterns that started with \h* or \R* might be incorrectly matched.
765 Version 8.30 04-February-2012
766 -----------------------------
768 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
769 name is defined in ctype.h.
771 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
772 only in quite long subpatterns.
774 3. Removed the function pcre_info(), which has been obsolete and deprecated
775 since it was replaced by pcre_fullinfo() in February 2000.
777 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
778 match a (*MARK), and the match failed at the start of the subject, a
779 reference to memory before the start of the subject could occur. This bug
780 was introduced by fix 17 of release 8.21.
782 5. A reference to an unset group with zero minimum repetition was giving
783 totally wrong answers (in non-JavaScript-compatibility mode). For example,
784 /(another)?(\1?)test/ matched against "hello world test". This bug was
785 introduced in release 8.13.
787 6. Add support for 16-bit character strings (a large amount of work involving
788 many changes and refactorings).
790 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
791 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
792 from a file.
794 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
795 rounding is not applied in this particular case).
797 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
798 if they appear, or are escaped, in patterns.
800 10. Get rid of a number of -Wunused-but-set-variable warnings.
802 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
803 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
804 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
805 also returns the mark "x". This bug applied to capturing parentheses,
806 non-capturing parentheses, and atomic parentheses. It also applied to some
807 assertions.
809 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
810 information out of configure.ac instead of relying on pcre.h.generic, which
811 is not stored in the repository.
813 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
814 -lreadline.
816 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
818 15. Applied Graycode's patch to put the top-level frame on the stack rather
819 than the heap when not using the stack for recursion. This gives a
820 performance improvement in many cases when recursion is not deep.
822 16. Experimental code added to "pcretest -C" to output the stack frame size.
825 Version 8.21 12-Dec-2011
826 ------------------------
828 1. Updating the JIT compiler.
830 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
831 are added as well.
833 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
834 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
835 calling _pcre_jit_exec. Some extra comments are added.
837 4. (*MARK) settings inside atomic groups that do not contain any capturing
838 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
839 was introduced by change 18 for 8.20.
841 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
842 ECMA-262 standard.
844 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
845 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
846 This bug was probably introduced by change 9 of 8.13.
848 7. While fixing 6 above, I noticed that a number of other items were being
849 incorrectly rejected as "not fixed length". This arose partly because newer
850 opcodes had not been added to the fixed-length checking code. I have (a)
851 corrected the bug and added tests for these items, and (b) arranged for an
852 error to occur if an unknown opcode is encountered while checking for fixed
853 length instead of just assuming "not fixed length". The items that were
854 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
855 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
856 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
858 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
859 being incorrectly compiled and would have given unpredicatble results.
861 9. A possessively repeated subpattern with minimum repeat count greater than
862 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
863 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
864 the first (A) could occur when it should not.
866 10. Add a cast and remove a redundant test from the code.
868 11. JIT should use pcre_malloc/pcre_free for allocation.
870 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
871 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
872 is anything other than /usr, -L is still shown).
874 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
876 14. Perl does not support \N without a following name in a [] class; PCRE now
877 also gives an error.
879 15. If a forward reference was repeated with an upper limit of around 2000,
880 it caused the error "internal error: overran compiling workspace". The
881 maximum number of forward references (including repeats) was limited by the
882 internal workspace, and dependent on the LINK_SIZE. The code has been
883 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
884 the default depends on LINK_SIZE. There is a new upper limit (for safety)
885 of around 200,000 forward references. While doing this, I also speeded up
886 the filling in of repeated forward references.
888 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
889 incorrectly expecting the subject to contain another "a" after the start.
891 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
892 in the match, the SKIP should be ignored. This was not happening; instead
893 the SKIP was being treated as NOMATCH. For patterns such as
894 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
895 tested.
897 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
898 now much more compatible with Perl, in particular in cases where the result
899 is a non-match for a non-anchored pattern. For example, if
900 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
901 "m", where previously it did not return a name. A side effect of this
902 change is that for partial matches, the last encountered mark name is
903 returned, as for non matches. A number of tests that were previously not
904 Perl-compatible have been moved into the Perl-compatible test files. The
905 refactoring has had the pleasing side effect of removing one argument from
906 the match() function, thus reducing its stack requirements.
908 19. If the /S+ option was used in pcretest to study a pattern using JIT,
909 subsequent uses of /S (without +) incorrectly behaved like /S+.
911 21. Retrieve executable code size support for the JIT compiler and fixing
912 some warnings.
914 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
915 not work when the shorter character appeared right at the end of the
916 subject string.
918 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
919 systems.
921 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
922 output it when the /M option is used in pcretest.
924 25. The CheckMan script was not being included in the distribution. Also, added
925 an explicit "perl" to run Perl scripts from the PrepareRelease script
926 because this is reportedly needed in Windows.
928 26. If study data was being save in a file and studying had not found a set of
929 "starts with" bytes for the pattern, the data written to the file (though
930 never used) was taken from uninitialized memory and so caused valgrind to
931 complain.
933 27. Updated RunTest.bat as provided by Sheri Pierce.
935 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
937 29. Computation of memory usage for the table of capturing group names was
938 giving an unnecessarily large value.
941 Version 8.20 21-Oct-2011
942 ------------------------
944 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
945 a POSIX class. After further experiments with Perl, which convinced me that
946 Perl has bugs and confusions, a closing square bracket is no longer allowed
947 in a POSIX name. This bug also affected patterns with classes that started
948 with full stops.
950 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
951 captured substring, but while checking the failing first alternative,
952 substring 1 is temporarily captured. If the output vector supplied to
953 pcre_exec() was not big enough for this capture, the yield of the function
954 was still zero ("insufficient space for captured substrings"). This cannot
955 be totally fixed without adding another stack variable, which seems a lot
956 of expense for a edge case. However, I have improved the situation in cases
957 such as /(a)(b)x|abc/ matched against "abc", where the return code
958 indicates that fewer than the maximum number of slots in the ovector have
959 been set.
961 3. Related to (2) above: when there are more back references in a pattern than
962 slots in the output vector, pcre_exec() uses temporary memory during
963 matching, and copies in the captures as far as possible afterwards. It was
964 using the entire output vector, but this conflicts with the specification
965 that only 2/3 is used for passing back captured substrings. Now it uses
966 only the first 2/3, for compatibility. This is, of course, another edge
967 case.
969 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
970 main code base, and can be used by building with --enable-jit. When this is
971 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
972 runtime --no-jit option is given.
974 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
975 ovector, the return from the function was zero, implying that there were
976 other matches that did not fit. The correct "exactly full" value is now
977 returned.
979 6. If a subpattern that was called recursively or as a subroutine contained
980 (*PRUNE) or any other control that caused it to give a non-standard return,
981 invalid errors such as "Error -26 (nested recursion at the same subject
982 position)" or even infinite loops could occur.
984 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
985 computing the minimum length on reaching *ACCEPT, and so ended up with the
986 wrong value of 1 rather than 0. Further investigation indicates that
987 computing a minimum subject length in the presence of *ACCEPT is difficult
988 (think back references, subroutine calls), and so I have changed the code
989 so that no minimum is registered for a pattern that contains *ACCEPT.
991 8. If (*THEN) was present in the first (true) branch of a conditional group,
992 it was not handled as intended. [But see 16 below.]
994 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
995 Sheri Pierce.
997 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
998 the first byte in a match must be "a".
1000 11. Change 17 for 8.13 increased the recursion depth for patterns like
1001 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1002 pattern contains any instances of (*THEN). If it does not, the old
1003 optimizations are restored. It would be nice to do this on a per-group
1004 basis, but at the moment that is not feasible.
1006 12. In some environments, the output of pcretest -C is CRLF terminated. This
1007 broke RunTest's code that checks for the link size. A single white space
1008 character after the value is now allowed for.
1010 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1011 For "fr", it uses the Windows-specific input and output files.
1013 14. If (*THEN) appeared in a group that was called recursively or as a
1014 subroutine, it did not work as intended. [But see next item.]
1016 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1017 pattern fragments (but not containing any | characters). If A and B are
1018 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1019 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1020 D. In other words, Perl considers parentheses that do not contain any |
1021 characters to be part of a surrounding alternative, whereas PCRE was
1022 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1023 differently. PCRE now behaves in the same way as Perl, except in the case
1024 of subroutine/recursion calls such as (?1) which have in any case always
1025 been different (but PCRE had them first :-).
1027 16. Related to 15 above: Perl does not treat the | in a conditional group as
1028 creating alternatives. Such a group is treated in the same way as an
1029 ordinary group without any | characters when processing (*THEN). PCRE has
1030 been changed to match Perl's behaviour.
1032 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1033 RunGrepTest script failed.
1035 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1036 inevitable for groups that contain captures, but it can lead to a lot of
1037 stack use in large patterns. The old behaviour has been restored for atomic
1038 groups that do not contain any capturing parentheses.
1040 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1041 suppress the check for a minimum subject length at run time. (If it was
1042 given to pcre_exec() or pcre_dfa_exec() it did work.)
1044 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1045 fail to work when decoding hex characters in data strings in EBCDIC
1046 environments.
1048 21. It appears that in at least one Mac OS environment, the isxdigit() function
1049 is implemented as a macro that evaluates to its argument more than once,
1050 contravening the C 90 Standard (I haven't checked a later standard). There
1051 was an instance in pcretest which caused it to go wrong when processing
1052 \x{...} escapes in subject strings. The has been rewritten to avoid using
1053 things like p++ in the argument of isxdigit().
1056 Version 8.13 16-Aug-2011
1057 ------------------------
1059 1. The Unicode data tables have been updated to Unicode 6.0.0.
1061 2. Two minor typos in pcre_internal.h have been fixed.
1063 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1064 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1065 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1067 4. There were a number of related bugs in the code for matching backrefences
1068 caselessly in UTF-8 mode when codes for the characters concerned were
1069 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1070 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1071 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1072 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1073 2-byte code at the end of the subject (it thought there wasn't enough data
1074 left).
1076 5. Comprehensive information about what went wrong is now returned by
1077 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1078 as the output vector has at least 2 elements. The offset of the start of
1079 the failing character and a reason code are placed in the vector.
1081 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1082 now returned is for the first byte of the failing character, instead of the
1083 last byte inspected. This is an incompatible change, but I hope it is small
1084 enough not to be a problem. It makes the returned offset consistent with
1085 pcre_exec() and pcre_dfa_exec().
1087 7. pcretest now gives a text phrase as well as the error number when
1088 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1089 failure, the offset and reason code are output.
1091 8. When \R was used with a maximizing quantifier it failed to skip backwards
1092 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1093 back over a single character (\n). This seems wrong (because it treated the
1094 two characters as a single entity when going forwards), conflicts with the
1095 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1096 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1097 has been changed.
1099 9. Some internal refactoring has changed the processing so that the handling
1100 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1101 time (the PCRE_DOTALL option was changed this way some time ago: version
1102 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1103 which was always a bit of a fudge. It also means that there is one less
1104 argument for the match() function, which reduces its stack requirements
1105 slightly. This change also fixes an incompatibility with Perl: the pattern
1106 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1108 10. More internal refactoring has drastically reduced the number of recursive
1109 calls to match() for possessively repeated groups such as (abc)++ when
1110 using pcre_exec().
1112 11. While implementing 10, a number of bugs in the handling of groups were
1113 discovered and fixed:
1115 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1116 (a|)*(?1) gave a compile-time internal error.
1117 ((a|)+)+ did not notice that the outer group could match an empty string.
1118 (^a|^)+ was not marked as anchored.
1119 (.*a|.*)+ was not marked as matching at start or after a newline.
1121 12. Yet more internal refactoring has removed another argument from the match()
1122 function. Special calls to this function are now indicated by setting a
1123 value in a variable in the "match data" data block.
1125 13. Be more explicit in pcre_study() instead of relying on "default" for
1126 opcodes that mean there is no starting character; this means that when new
1127 ones are added and accidentally left out of pcre_study(), testing should
1128 pick them up.
1130 14. The -s option of pcretest has been documented for ages as being an old
1131 synonym of -m (show memory usage). I have changed it to mean "force study
1132 for every regex", that is, assume /S for every regex. This is similar to -i
1133 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1134 using it. It makes it easier to run collections of tests with and without
1135 study enabled, and thereby test pcre_study() more easily. All the standard
1136 tests are now run with and without -s (but some patterns can be marked as
1137 "never study" - see 20 below).
1139 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1140 restoration of the capturing data to the outer values was not happening
1141 correctly.
1143 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1144 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1145 pattern had matched an empty string, and so incorrectly returned a no
1146 match.
1148 17. There was optimizing code for the last branch of non-capturing parentheses,
1149 and also for the obeyed branch of a conditional subexpression, which used
1150 tail recursion to cut down on stack usage. Unfortunately, now that there is
1151 the possibility of (*THEN) occurring in these branches, tail recursion is
1152 no longer possible because the return has to be checked for (*THEN). These
1153 two optimizations have therefore been removed. [But see 8.20/11 above.]
1155 18. If a pattern containing \R was studied, it was assumed that \R always
1156 matched two bytes, thus causing the minimum subject length to be
1157 incorrectly computed because \R can also match just one byte.
1159 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1160 was incorrectly computed.
1162 20. If /S is present twice on a test pattern in pcretest input, it now
1163 *disables* studying, thereby overriding the use of -s on the command line
1164 (see 14 above). This is necessary for one or two tests to keep the output
1165 identical in both cases.
1167 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1168 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1170 22. When an atomic group that contained a capturing parenthesis was
1171 successfully matched, but the branch in which it appeared failed, the
1172 capturing was not being forgotten if a higher numbered group was later
1173 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1174 group 1 to "a", when in fact it should be unset. This applied to multi-
1175 branched capturing and non-capturing groups, repeated or not, and also to
1176 positive assertions (capturing in negative assertions does not happen
1177 in PCRE) and also to nested atomic groups.
1179 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1180 subject after a captured substring, to make it easier to tell which of a
1181 number of identical substrings has been captured.
1183 24. The way atomic groups are processed by pcre_exec() has been changed so that
1184 if they are repeated, backtracking one repetition now resets captured
1185 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1186 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1187 "aaa". Previously, it would have been "a". As part of this code
1188 refactoring, the way recursive calls are handled has also been changed.
1190 25. If an assertion condition captured any substrings, they were not passed
1191 back unless some other capturing happened later. For example, if
1192 (?(?=(a))a) was matched against "a", no capturing was returned.
1194 26. When studying a pattern that contained subroutine calls or assertions,
1195 the code for finding the minimum length of a possible match was handling
1196 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1197 group 1 called group 2 while simultaneously a separate group 2 called group
1198 1). A stack overflow occurred in this case. I have fixed this by limiting
1199 the recursion depth to 10.
1201 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1202 Fortmann. This supports explicit test numbers on the command line, and has
1203 argument validation and error reporting.
1205 28. An instance of \X with an unlimited repeat could fail if at any point the
1206 first character it looked at was a mark character.
1208 29. Some minor code refactoring concerning Unicode properties and scripts
1209 should reduce the stack requirement of match() slightly.
1211 30. Added the '=' option to pcretest to check the setting of unused capturing
1212 slots at the end of the pattern, which are documented as being -1, but are
1213 not included in the return count.
1215 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1216 compiled something random. Now it gives a compile-time error (as does
1217 Perl).
1219 32. A *MARK encountered during the processing of a positive assertion is now
1220 recorded and passed back (compatible with Perl).
1222 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1223 had alternative anchored branches, the search for a second match in a line
1224 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1225 matched the line "0102" twice. The same bug affected patterns that started
1226 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1227 twice.
1229 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1230 does, and because of capturing effects, quantifying parenthesized
1231 assertions may at times be useful. Quantifiers are now allowed for
1232 parenthesized assertions.
1234 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1236 36. \g was being checked for fancy things in a character class, when it should
1237 just be a literal "g".
1239 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1240 appearance of a nested POSIX class supersedes an apparent external class.
1241 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1242 unescaped square brackets may also appear as part of class names. For
1243 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1244 more like Perl. (But see 8.20/1 above.)
1246 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1247 was because it thought it was \N{name}, which is not supported).
1249 39. Add minix to OS list not supporting the -S option in pcretest.
1251 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1252 cannot analyze patterns in sufficient detail to catch mutual recursions
1253 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1254 subgroup is called recursively as a subpattern for a second time at the
1255 same position in the subject string. In previous releases this might have
1256 been caught by the recursion limit, or it might have run out of stack.
1258 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1259 happen only once. PCRE was, however incorrectly giving a compile time error
1260 "recursive call could loop indefinitely" because it cannot analyze the
1261 pattern in sufficient detail. The compile time test no longer happens when
1262 PCRE is compiling a conditional subpattern, but actual runaway loops are
1263 now caught at runtime (see 40 above).
1265 42. It seems that Perl allows any characters other than a closing parenthesis
1266 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1267 has been changed to be the same.
1269 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1270 as not to get warnings when autogen.sh is called. Also changed
1271 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1273 44. To help people who use pcregrep to scan files containing exceedingly long
1274 lines, the following changes have been made:
1276 (a) The default value of the buffer size parameter has been increased from
1277 8K to 20K. (The actual buffer used is three times this size.)
1279 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1280 PCRE is built.
1282 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1283 to be set at run time.
1285 (d) Numerical values in pcregrep options can be followed by K or M, for
1286 example --buffer-size=50K.
1288 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1289 given and the return code is set to 2.
1291 45. Add a pointer to the latest mark to the callout data block.
1293 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1294 partial match of an empty string instead of no match. This was specific to
1295 the use of ".".
1297 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1298 complete match instead of a partial match. This bug was dependent on both
1299 the PCRE_UTF8 and PCRE_DOTALL options being set.
1301 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1302 starting byte set, because \b was not being ignored.
1305 Version 8.12 15-Jan-2011
1306 ------------------------
1308 1. Fixed some typos in the markup of the man pages, and wrote a script that
1309 checks for such things as part of the documentation building process.
1311 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1312 --match-limit and --recursion-limit options (added for 8.11). In
1313 particular, this made one of the standard tests fail. (The integer value
1314 went into the wrong half of a long int.)
1316 3. If the --colour option was given to pcregrep with -v (invert match), it
1317 did strange things, either producing crazy output, or crashing. It should,
1318 of course, ignore a request for colour when reporting lines that do not
1319 match.
1321 4. Another pcregrep bug caused similar problems if --colour was specified with
1322 -M (multiline) and the pattern match finished with a line ending.
1324 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1325 matched in multiline mode, the following line was shown as part of the
1326 match. This seems wrong, so I have changed it.
1328 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1329 the check for further matches in the same line (so they could be coloured)
1330 to overrun the end of the current line. If another match was found, it was
1331 incorrectly shown (and then shown again when found in the next line).
1333 7. If pcregrep was compiled under Windows, there was a reference to the
1334 function pcregrep_exit() before it was defined. I am assuming this was
1335 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1336 reported by a user. I've moved the definition above the reference.
1339 Version 8.11 10-Dec-2010
1340 ------------------------
1342 1. (*THEN) was not working properly if there were untried alternatives prior
1343 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1344 backtracked to try for "b" instead of moving to the next alternative branch
1345 at the same level (in this case, to look for "c"). The Perl documentation
1346 is clear that when (*THEN) is backtracked onto, it goes to the "next
1347 alternative in the innermost enclosing group".
1349 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1350 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1351 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1352 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1353 (*THEN).
1355 3. If \s appeared in a character class, it removed the VT character from
1356 the class, even if it had been included by some previous item, for example
1357 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1358 of \s, but is part of the POSIX "space" class.)
1360 4. A partial match never returns an empty string (because you can always
1361 match an empty string at the end of the subject); however the checking for
1362 an empty string was starting at the "start of match" point. This has been
1363 changed to the "earliest inspected character" point, because the returned
1364 data for a partial match starts at this character. This means that, for
1365 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1366 (previously it gave "no match").
1368 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1369 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1370 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1371 has an implication that the given string is incomplete (because a partial
1372 match is preferred over a full match). For this reason, these items now
1373 give a partial match in this situation. [Aside: previously, the one case
1374 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1375 match rather than a full match, which was wrong by the old rules, but is
1376 now correct.]
1378 6. There was a bug in the handling of #-introduced comments, recognized when
1379 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1380 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1381 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1382 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1383 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1384 places in pcre_compile().
1386 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1387 comments when looking ahead for named forward references to subpatterns,
1388 the only newline sequence it recognized was NL. It now handles newlines
1389 according to the set newline convention.
1391 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1392 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1393 cater for a lack of strerror(). These oversights have been fixed.
1395 9. Added --match-limit and --recursion-limit to pcregrep.
1397 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1399 11. When the -o option was used, pcregrep was setting a return code of 1, even
1400 when matches were found, and --line-buffered was not being honoured.
1402 12. Added an optional parentheses number to the -o and --only-matching options
1403 of pcregrep.
1405 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1406 can match an empty string. The code to do it in pcretest and pcredemo
1407 needed fixing:
1409 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1410 only one byte after an empty string match just before CRLF (this case
1411 just got forgotten; "any" and "anycrlf" were OK).
1413 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1414 mode when an empty string match preceded an ASCII character followed by
1415 a non-ASCII character. (The code for advancing by one character rather
1416 than one byte was nonsense.)
1418 (c) The pcredemo.c sample program did not have any code at all to handle
1419 the cases when CRLF is a valid newline sequence.
1421 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1422 as a starting offset was within the subject string. There is now a new
1423 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1424 negative or greater than the length of the string. In order to test this,
1425 pcretest is extended to allow the setting of negative starting offsets.
1427 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1428 starting offset points to the beginning of a UTF-8 character was
1429 unnecessarily clumsy. I tidied it up.
1431 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1432 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1434 17. Nobody had reported that the --include_dir option, which was added in
1435 release 7.7 should have been called --include-dir (hyphen, not underscore)
1436 for compatibility with GNU grep. I have changed it to --include-dir, but
1437 left --include_dir as an undocumented synonym, and the same for
1438 --exclude-dir, though that is not available in GNU grep, at least as of
1439 release 2.5.4.
1441 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1442 characters from a string of bytes) have been redefined so as not to use
1443 loops, in order to improve performance in some environments. At the same
1444 time, I abstracted some of the common code into auxiliary macros to save
1445 repetition (this should not affect the compiled code).
1447 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1448 compile-time error is now given if \c is not followed by an ASCII
1449 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1450 different, and any byte value is allowed.)
1452 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1453 START_OPTIMIZE option, which is now allowed at compile time - but just
1454 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1455 to pcregrep and other applications that have no direct access to PCRE
1456 options. The new /Y option in pcretest sets this option when calling
1457 pcre_compile().
1459 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1460 back references. Groups containing recursive back references were forced to
1461 be atomic by that change, but in the case of named groups, the amount of
1462 memory required was incorrectly computed, leading to "Failed: internal
1463 error: code overflow". This has been fixed.
1465 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1466 pcretest.c, to avoid build problems in some Borland environments.
1469 Version 8.10 25-Jun-2010
1470 ------------------------
1472 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1473 THEN.
1475 2. (*ACCEPT) was not working when inside an atomic group.
1477 3. Inside a character class, \B is treated as a literal by default, but
1478 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1479 causes the error). The code is unchanged, but I tidied the documentation.
1481 4. Inside a character class, PCRE always treated \R and \X as literals,
1482 whereas Perl faults them if its -w option is set. I have changed PCRE so
1483 that it faults them when PCRE_EXTRA is set.
1485 5. Added support for \N, which always matches any character other than
1486 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1488 6. When compiling pcregrep with newer versions of gcc which may have
1489 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1490 declared with attribute warn_unused_result" were given. Just casting the
1491 result to (void) does not stop the warnings; a more elaborate fudge is
1492 needed. I've used a macro to implement this.
1494 7. Minor change to pcretest.c to avoid a compiler warning.
1496 8. Added four artifical Unicode properties to help with an option to make
1497 \s etc use properties (see next item). The new properties are: Xan
1498 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1500 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1501 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1502 this option. Modified pcretest to add /W to test this facility. Added
1503 REG_UCP to make it available via the POSIX interface.
1505 10. Added --line-buffered to pcregrep.
1507 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1508 studied, and the match started with a letter with a code point greater than
1509 127 whose first byte was different to the first byte of the other case of
1510 the letter, the other case of this starting letter was not recognized
1511 (#976).
1513 12. If a pattern that was studied started with a repeated Unicode property
1514 test, for example, \p{Nd}+, there was the theoretical possibility of
1515 setting up an incorrect bitmap of starting bytes, but fortunately it could
1516 not have actually happened in practice until change 8 above was made (it
1517 added property types that matched character-matching opcodes).
1519 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1520 possible starting bytes for non-anchored patterns.
1522 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1523 \R, and also a number of cases that involve Unicode properties, both
1524 explicit and implicit when PCRE_UCP is set.
1526 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1527 input, it could crash or give wrong results if characters with values
1528 greater than 0xc0 were present in the subject string. (Detail: it assumed
1529 UTF-8 input when processing these items.)
1531 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1532 size_t is 64-bit (#991).
1534 17. Added a check for running out of memory when PCRE is compiled with
1535 --disable-stack-for-recursion (#990).
1537 18. If the last data line in a file for pcretest does not have a newline on
1538 the end, a newline was missing in the output.
1540 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1541 less than 128) in its various bitmaps. However, there is a facility for
1542 generating tables according to the current locale when PCRE is compiled. It
1543 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1544 characters, are recognized by isspace() and therefore were getting set in
1545 these tables, and indeed these tables seem to approximate to ISO 8859. This
1546 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1547 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1548 which of course cannot start UTF-8 characters. I have changed the code so
1549 that only real ASCII characters (less than 128) and the correct starting
1550 bytes for UTF-8 encodings are set for characters greater than 127 when in
1551 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1552 altogether.)
1554 20. Added the /T option to pcretest so as to be able to run tests with non-
1555 standard character tables, thus making it possible to include the tests
1556 used for 19 above in the standard set of tests.
1558 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1559 reference to a subpattern the other side of a comment that contains an
1560 opening parenthesis caused either an internal compiling error, or a
1561 reference to the wrong subpattern.
1564 Version 8.02 19-Mar-2010
1565 ------------------------
1567 1. The Unicode data tables have been updated to Unicode 5.2.0.
1569 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1570 configured.
1572 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1573 original author of that file, following a query about its status.
1575 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1576 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1578 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1579 quantifier applied to a forward-referencing subroutine call, could compile
1580 incorrect code or give the error "internal error: previously-checked
1581 referenced subpattern not found".
1583 6. Both MS Visual Studio and Symbian OS have problems with initializing
1584 variables to point to external functions. For these systems, therefore,
1585 pcre_malloc etc. are now initialized to local functions that call the
1586 relevant global functions.
1588 7. There were two entries missing in the vectors called coptable and poptable
1589 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1590 I've fixed the data, and added a kludgy way of testing at compile time that
1591 the lengths are correct (equal to the number of opcodes).
1593 8. Following on from 7, I added a similar kludge to check the length of the
1594 eint vector in pcreposix.c.
1596 9. Error texts for pcre_compile() are held as one long string to avoid too
1597 much relocation at load time. To find a text, the string is searched,
1598 counting zeros. There was no check for running off the end of the string,
1599 which could happen if a new error number was added without updating the
1600 string.
1602 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1604 11. \K was not working if it appeared in an atomic group or in a group that
1605 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1606 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1607 the assertion is positive, but not if it is negative.
1609 12. Change 11 fortuitously reduced the size of the stack frame used in the
1610 "match()" function of pcre_exec.c by one pointer. Forthcoming
1611 implementation of support for (*MARK) will need an extra pointer on the
1612 stack; I have reserved it now, so that the stack frame size does not
1613 decrease.
1615 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1616 item in branch that calls a recursion is a subroutine call - as in the
1617 second branch in the above example - was incorrectly given the compile-
1618 time error "recursive call could loop indefinitely" because pcre_compile()
1619 was not correctly checking the subroutine for matching a non-empty string.
1621 14. The checks for overrunning compiling workspace could trigger after an
1622 overrun had occurred. This is a "should never occur" error, but it can be
1623 triggered by pathological patterns such as hundreds of nested parentheses.
1624 The checks now trigger 100 bytes before the end of the workspace.
1626 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1629 Version 8.01 19-Jan-2010
1630 ------------------------
1632 1. If a pattern contained a conditional subpattern with only one branch (in
1633 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1634 computed the wrong minimum data length (which is of course zero for such
1635 subpatterns). This could cause incorrect "no match" results.
1637 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1638 the pattern is reset in the first branch, pcre_compile() failed with
1639 "internal error: code overflow at offset...". This happened only when
1640 the reset was to the original external option setting. (An optimization
1641 abstracts leading options settings into an external setting, which was the
1642 cause of this.)
1644 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1645 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1646 assertion pattern did not match (meaning that the assertion was true), it
1647 was incorrectly treated as false if the SKIP had been reached during the
1648 matching. This also applied to assertions used as conditions.
1650 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1651 assertion subpattern, including such a pattern used as a condition,
1652 unpredictable results occurred, instead of the error return
1655 5. The C++ GlobalReplace function was not working like Perl for the special
1656 situation when an empty string is matched. It now does the fancy magic
1657 stuff that is necessary.
1659 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1660 removed. (These were left over from very, very early versions of PCRE.)
1662 7. Some cosmetic changes to the code to make life easier when compiling it
1663 as part of something else:
1665 (a) Change DEBUG to PCRE_DEBUG.
1667 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1668 called "current" as "current_branch", to prevent a collision with the
1669 Linux macro when compiled as a kernel module.
1671 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1672 prevent a collision with the Linux macro when compiled as a kernel
1673 module.
1675 8. In pcre_compile() there are some checks for integer overflows that used to
1676 cast potentially large values to (double). This has been changed to that
1677 when building, a check for int64_t is made, and if it is found, it is used
1678 instead, thus avoiding the use of floating point arithmetic. (There is no
1679 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1680 double.
1682 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1683 2005 (difference between two addresses compared to an unsigned value).
1685 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1686 custom one, because of the following reported problem in Windows:
1688 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1689 under Win32.
1690 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1691 therefore missing the function definition.
1692 - The compiler thus generates a "C" signature for the test function.
1693 - The linker fails to find the "C" function.
1694 - PCRE fails to configure if asked to do so against libbz2.
1696 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1697 messages were output:
1699 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1700 rerunning libtoolize, to keep the correct libtool macros in-tree.
1701 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1703 I have done both of these things.
1705 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1706 most of the time, it *can* run out if it is given a pattern that contains a
1707 runaway infinite recursion. I updated the discussion in the pcrestack man
1708 page.
1710 13. Now that we have gone to the x.xx style of version numbers, the minor
1711 version may start with zero. Using 08 or 09 is a bad idea because users
1712 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1713 interpreted as invalid octal numbers. I've updated the previous comment in
1714 configure.ac, and also added a check that gives an error if 08 or 09 are
1715 used.
1717 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1718 causing partial matching to fail when the end of the subject matched \W
1719 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1721 15. There were some discrepancies between the declarations in pcre_internal.h
1722 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1723 their definitions. The declarations used "const uschar *" and the
1724 definitions used USPTR. Even though USPTR is normally defined as "const
1725 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1726 reported that: "This difference in casting confuses some C++ compilers, for
1727 example, SunCC recognizes above declarations as different functions and
1728 generates broken code for hbpcre." I have changed the declarations to use
1729 USPTR.
1731 16. GNU libtool is named differently on some systems. The autogen.sh script now
1732 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1733 (FreeBSD).
1735 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1736 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1737 comment: "Figure out how to create a longlong from a string: strtoll and
1738 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1739 instance, but it only takes 2 args instead of 3!"
1741 18. A subtle bug concerned with back references has been fixed by a change of
1742 specification, with a corresponding code fix. A pattern such as
1743 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1744 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1745 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1746 same bug. Such groups have to be quantified to be useful, or contained
1747 inside another quantified group. (If there's no repetition, the reference
1748 can never match.) The problem arises because, having left the group and
1749 moved on to the rest of the pattern, a later failure that backtracks into
1750 the group uses the captured value from the final iteration of the group
1751 rather than the correct earlier one. I have fixed this in PCRE by forcing
1752 any group that contains a reference to itself to be an atomic group; that
1753 is, there cannot be any backtracking into it once it has completed. This is
1754 similar to recursive and subroutine calls.
1757 Version 8.00 19-Oct-09
1758 ----------------------
1760 1. The table for translating pcre_compile() error codes into POSIX error codes
1761 was out-of-date, and there was no check on the pcre_compile() error code
1762 being within the table. This could lead to an OK return being given in
1763 error.
1765 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1766 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1767 in a Windows environment.
1769 3. The pcregrep --count option prints the count for each file even when it is
1770 zero, as does GNU grep. However, pcregrep was also printing all files when
1771 --files-with-matches was added. Now, when both options are given, it prints
1772 counts only for those files that have at least one match. (GNU grep just
1773 prints the file name in this circumstance, but including the count seems
1774 more useful - otherwise, why use --count?) Also ensured that the
1775 combination -clh just lists non-zero counts, with no names.
1777 4. The long form of the pcregrep -F option was incorrectly implemented as
1778 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1779 but it seems right to fix it, and I didn't think it was worth preserving
1780 the old behaviour.
1782 5. The command line items --regex=pattern and --regexp=pattern were not
1783 recognized by pcregrep, which required --regex pattern or --regexp pattern
1784 (with a space rather than an '='). The man page documented the '=' forms,
1785 which are compatible with GNU grep; these now work.
1787 6. No libpcreposix.pc file was created for pkg-config; there was just
1788 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1790 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1791 when UCP support is not needed, by modifying the Python script that
1792 generates it from Unicode data files. This should not matter if the module
1793 is correctly used as a library, but I received one complaint about 50K of
1794 unwanted data. My guess is that the person linked everything into his
1795 program rather than using a library. Anyway, it does no harm.
1797 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1798 was a minimum greater than 1 for a wide character in a possessive
1799 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1800 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1801 character. Chaos in the form of incorrect output or a compiling loop could
1802 result.
1804 9. The restrictions on what a pattern can contain when partial matching is
1805 requested for pcre_exec() have been removed. All patterns can now be
1806 partially matched by this function. In addition, if there are at least two
1807 slots in the offset vector, the offset of the earliest inspected character
1808 for the match and the offset of the end of the subject are set in them when
1809 PCRE_ERROR_PARTIAL is returned.
1811 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1812 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1813 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1814 and may be more useful for multi-segment matching.
1816 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1817 used to be given if ever the end of the subject was reached; now it is
1818 given only if matching could not proceed because another character was
1819 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1820 string "Z", which now yields "no match" instead of "partial match". In the
1821 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1822 final character ended with (*FAIL).
1824 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1825 if the pattern had a "must contain" character that was already found in the
1826 earlier partial match, unless partial matching was again requested. For
1827 example, with the pattern /dog.(body)?/, the "must contain" character is
1828 "g". If the first part-match was for the string "dog", restarting with
1829 "sbody" failed. This bug has been fixed.
1831 13. The string returned by pcre_dfa_exec() after a partial match has been
1832 changed so that it starts at the first inspected character rather than the
1833 first character of the match. This makes a difference only if the pattern
1834 starts with a lookbehind assertion or \b or \B (\K is not supported by
1835 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1836 matching functions compatible, and I think it's the right thing to do.
1838 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1839 so that the demonstration program is easily available in environments where
1840 PCRE has not been installed from source.
1842 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1843 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1844 library.
1846 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1847 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1848 is not the first non-POSIX option to be added. Clearly some people find
1849 these options useful.
1851 17. If a caller to the POSIX matching function regexec() passes a non-zero
1852 value for nmatch with a NULL value for pmatch, the value of
1853 nmatch is forced to zero.
1855 18. RunGrepTest did not have a test for the availability of the -u option of
1856 the diff command, as RunTest does. It now checks in the same way as
1857 RunTest, and also checks for the -b option.
1859 19. If an odd number of negated classes containing just a single character
1860 interposed, within parentheses, between a forward reference to a named
1861 subpattern and the definition of the subpattern, compilation crashed with
1862 an internal error, complaining that it could not find the referenced
1863 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1864 [The bug was that it was starting one character too far in when skipping
1865 over the character class, thus treating the ] as data rather than
1866 terminating the class. This meant it could skip too much.]
1868 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1869 /g option in pcretest when the pattern contains \K, which makes it possible
1870 to have an empty string match not at the start, even when the pattern is
1871 anchored. Updated pcretest and pcredemo to use this option.
1873 21. If the maximum number of capturing subpatterns in a recursion was greater
1874 than the maximum at the outer level, the higher number was returned, but
1875 with unset values at the outer level. The correct (outer level) value is
1876 now given.
1878 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1879 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1880 make it do so. The string so far is captured, making this feature
1881 compatible with Perl.
1883 23. The tests have been re-organized, adding tests 11 and 12, to make it
1884 possible to check the Perl 5.10 features against Perl 5.10.
1886 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1887 pattern matches a fixed length string. PCRE did not allow this; now it
1888 does. Neither allows recursion.
1890 25. I finally figured out how to implement a request to provide the minimum
1891 length of subject string that was needed in order to match a given pattern.
1892 (It was back references and recursion that I had previously got hung up
1893 on.) This code has now been added to pcre_study(); it finds a lower bound
1894 to the length of subject needed. It is not necessarily the greatest lower
1895 bound, but using it to avoid searching strings that are too short does give
1896 some useful speed-ups. The value is available to calling programs via
1897 pcre_fullinfo().
1899 26. While implementing 25, I discovered to my embarrassment that pcretest had
1900 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1901 study optimizations had never been tested with that matching function.
1902 Oops. What is worse, even when it was passed study data, there was a bug in
1903 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1904 were also very few tests of studied patterns with pcre_dfa_exec().
1906 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1907 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1908 on the other side of the coin, they are no longer allowed to have different
1909 names, because these cannot be distinguished in PCRE, and this has caused
1910 confusion. (This is a difference from Perl.)
1912 28. When duplicate subpattern names are present (necessarily with different
1913 numbers, as required by 27 above), and a test is made by name in a
1914 conditional pattern, either for a subpattern having been matched, or for
1915 recursion in such a pattern, all the associated numbered subpatterns are
1916 tested, and the overall condition is true if the condition is true for any
1917 one of them. This is the way Perl works, and is also more like the way
1918 testing by number works.
1921 Version 7.9 11-Apr-09
1922 ---------------------
1924 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1925 (pcretest), all targets were linked against these libraries. This included
1926 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1927 libraries. This caused unwanted dependencies to be created. This problem
1928 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1929 pcretest is linked with readline.
1931 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1932 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1933 moved outside it again, because FALSE and TRUE are already defined in AIX,
1934 but BOOL is not.
1936 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1937 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1939 4. The pcregrep documentation said spaces were inserted as well as colons (or
1940 hyphens) following file names and line numbers when outputting matching
1941 lines. This is not true; no spaces are inserted. I have also clarified the
1942 wording for the --colour (or --color) option.
1944 5. In pcregrep, when --colour was used with -o, the list of matching strings
1945 was not coloured; this is different to GNU grep, so I have changed it to be
1946 the same.
1948 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1949 each matching line was coloured. Now it goes on to look for further matches
1950 of any of the test patterns, which is the same behaviour as GNU grep.
1952 7. A pattern that could match an empty string could cause pcregrep to loop; it
1953 doesn't make sense to accept an empty string match in pcregrep, so I have
1954 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1955 seems to be how GNU grep behaves. [But see later change 40 for release
1956 8.33.]
1958 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1959 start or after a newline", because the conditional assertion was not being
1960 correctly handled. The rule now is that both the assertion and what follows
1961 in the first alternative must satisfy the test.
1963 9. If auto-callout was enabled in a pattern with a conditional group whose
1964 condition was an assertion, PCRE could crash during matching, both with
1965 pcre_exec() and pcre_dfa_exec().
1967 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1968 used for matching.
1970 11. Unicode property support in character classes was not working for
1971 characters (bytes) greater than 127 when not in UTF-8 mode.
1973 12. Added the -M command line option to pcretest.
1975 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1977 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1979 16. Added comments and documentation about mis-use of no_arg in the C++
1980 wrapper.
1982 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1983 from Martin Jerabek that uses macro names for all relevant character and
1984 string constants.
1986 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1987 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1988 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1989 these, but not everybody uses configure.
1991 19. A conditional group that had only one branch was not being correctly
1992 recognized as an item that could match an empty string. This meant that an
1993 enclosing group might also not be so recognized, causing infinite looping
1994 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1995 with the subject "ab", where knowledge that the repeated group can match
1996 nothing is needed in order to break the loop.
1998 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1999 exec(), but without supplying a callout function, matching went wrong.
2001 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2002 leak if the size of the offset vector was greater than 30. When the vector
2003 is smaller, the saved offsets during recursion go onto a local stack
2004 vector, but for larger vectors malloc() is used. It was failing to free
2005 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2006 error, in fact).
2008 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2009 heapframe that is used only when UTF-8 support is enabled. This caused no
2010 problem, but was untidy.
2012 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2013 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2014 included within another project.
2016 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2017 slightly modified by me:
2019 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2020 not building pcregrep.
2022 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2023 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2025 25. Forward references, both numeric and by name, in patterns that made use of
2026 duplicate group numbers, could behave incorrectly or give incorrect errors,
2027 because when scanning forward to find the reference group, PCRE was not
2028 taking into account the duplicate group numbers. A pattern such as
2029 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2031 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2032 the feature of a custom pointer more persuasive (as requested by a user).
2034 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2035 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2036 pre-defined.
2038 28. Added support for (*UTF8) at the start of a pattern.
2040 29. Arrange for flags added by the "release type" setting in CMake to be shown
2041 in the configuration summary.
2044 Version 7.8 05-Sep-08
2045 ---------------------
2047 1. Replaced UCP searching code with optimized version as implemented for Ad
2048 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2049 stage table and inline lookup instead of a function, giving speed ups of 2
2050 to 5 times on some simple patterns that I tested. Permission was given to
2051 distribute the MultiStage2.py script that generates the tables (it's not in
2052 the tarball, but is in the Subversion repository).
2054 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2055 scripts.
2057 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2058 a group with a zero qualifier. The result of the study could be incorrect,
2059 or the function might crash, depending on the pattern.
2061 4. Caseless matching was not working for non-ASCII characters in back
2062 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2063 It now works when Unicode Property Support is available.
2065 5. In pcretest, an escape such as \x{de} in the data was always generating
2066 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2067 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2068 truncation.
2070 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2072 7. Added two (int) casts to pcregrep when printing the difference of two
2073 pointers, in case they are 64-bit values.
2075 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2076 test 2 if it fails.
2078 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2079 and a #define of that name to empty if it is not externally set. This is to
2080 allow users of MSVC to set it if necessary.
2082 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2083 the convenience functions in the pcre_get.c source file.
2085 11. An option change at the start of a pattern that had top-level alternatives
2086 could cause overwriting and/or a crash. This command provoked a crash in
2087 some environments:
2089 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2091 This potential security problem was recorded as CVE-2008-2371.
2093 12. For a pattern where the match had to start at the beginning or immediately
2094 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2095 pcre_dfa_exec() could read past the end of the passed subject if there was
2096 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2097 pcretest so that it places the subject at the end of its malloc-ed buffer.
2099 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2100 exec() might read past the end of the data buffer in UTF-8 mode.
2102 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2103 the data contained the byte 0x85 as part of a UTF-8 character within its
2104 first line. This applied both to normal and DFA matching.
2106 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2107 /^[^d]*?$/8 failed to match "abc".
2109 16. Added a missing copyright notice to pcrecpp_internal.h.
2111 17. Make it more clear in the documentation that values returned from
2112 pcre_exec() in ovector are byte offsets, not character counts.
2114 18. Tidied a few places to stop certain compilers from issuing warnings.
2116 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2117 supplied by Stefan Weber. I made a further small update for 7.8 because
2118 there is a change of source arrangements: the pcre_searchfuncs.c module is
2119 replaced by pcre_ucd.c.
2122 Version 7.7 07-May-08
2123 ---------------------
2125 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2126 a string to a long long, pretend we don't even have a long long." This is
2127 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2129 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2130 pre-7.6 versions, which defined a global no_arg variable instead of putting
2131 it in the RE class. (See also #8 below.)
2133 3. Remove a line of dead code, identified by coverity and reported by Nuno
2134 Lopes.
2136 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2138 (1) The include/exclude patterns were being applied to the whole pathnames
2139 of files, instead of just to the final components.
2141 (2) If there was more than one level of directory, the subdirectories were
2142 skipped unless they satisfied the include/exclude conditions. This is
2143 inconsistent with GNU grep (and could even be seen as contrary to the
2144 pcregrep specification - which I improved to make it absolutely clear).
2145 The action now is always to scan all levels of directory, and just
2146 apply the include/exclude patterns to regular files.
2148 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2149 --exclude_dir in the tests to avoid scanning .svn directories.
2151 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2152 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2153 doesn't support NULs in patterns.
2155 7. Added some missing "const"s to declarations of static tables in
2156 pcre_compile.c and pcre_dfa_exec.c.
2158 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2159 caused by fix #2 above. (Subsequently also a second patch to fix the
2160 first patch. And a third patch - this was a messy problem.)
2162 9. Applied Craig's patch to remove the use of push_back().
2164 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2165 matching function regexec().
2167 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2168 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2169 references. PCRE supports relative numbers with this syntax (I don't think
2170 Oniguruma does).
2172 12. Previously, a group with a zero repeat such as (...){0} was completely
2173 omitted from the compiled regex. However, this means that if the group
2174 was called as a subroutine from elsewhere in the pattern, things went wrong
2175 (an internal error was given). Such groups are now left in the compiled
2176 pattern, with a new opcode that causes them to be skipped at execution
2177 time.
2179 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2180 to the way PCRE behaves:
2182 (a) A lone ] character is dis-allowed (Perl treats it as data).
2184 (b) A back reference to an unmatched subpattern matches an empty string
2185 (Perl fails the current match path).
2187 (c) A data ] in a character class must be notated as \] because if the
2188 first data character in a class is ], it defines an empty class. (In
2189 Perl it is not possible to have an empty class.) The empty class []
2190 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2191 The negative empty class [^] matches any one character, independently
2192 of the DOTALL setting.
2194 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2195 non-existent subpattern following a character class starting with ']' and
2196 containing () gave an internal compiling error instead of "reference to
2197 non-existent subpattern". Fortunately, when the pattern did exist, the
2198 compiled code was correct. (When scanning forwards to check for the
2199 existence of the subpattern, it was treating the data ']' as terminating
2200 the class, so got the count wrong. When actually compiling, the reference
2201 was subsequently set up correctly.)
2203 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2204 it was being rejected as not supported by pcre_dfa_exec(), even though
2205 other assertions are supported. I have made pcre_dfa_exec() support
2206 (*FAIL).
2208 16. The implementation of 13c above involved the invention of a new opcode,
2209 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2210 cannot be changed at match time, I realized I could make a small
2211 improvement to matching performance by compiling OP_ALLANY instead of
2212 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2213 on the OP_ANY path.
2215 17. Compiling pcretest on Windows with readline support failed without the
2216 following two fixes: (1) Make the unistd.h include conditional on
2217 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2219 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2220 ncurses library to be included for pcretest when ReadLine support is
2221 requested, but also to allow for it to be overridden. This patch came from
2222 Daniel Bergström.
2224 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2225 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2226 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2227 spotting this.
2230 Version 7.6 28-Jan-08
2231 ---------------------
2233 1. A character class containing a very large number of characters with
2234 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2235 overflow.
2237 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2238 HAVE_LONG_LONG is not defined.
2240 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2241 bring them up to date and include new features. This patch includes:
2243 - Fixed PH's badly added libz and libbz2 support.
2244 - Fixed a problem with static linking.
2245 - Added pcredemo. [But later removed - see 7 below.]
2246 - Fixed dftables problem and added an option.
2247 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2249 - Added readline support for pcretest.
2250 - Added an listing of the option settings after cmake has run.
2252 4. A user submitted a patch to Makefile that makes it easy to create
2253 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2254 Makefile.am that cause it to include this special target, without
2255 affecting anything else. Note that the same mingw target plus all
2256 the other distribution libraries and programs are now supported
2257 when configuring with CMake (see 6 below) instead of with
2258 Configure/Make.
2260 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2261 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2262 exported in the Windows port". It has not yet been confirmed that the patch
2263 solves the problem, but it does no harm.
2265 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2266 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2267 with CMake, and also correct the comment about stack recursion.
2269 7. Remove the automatic building of pcredemo from the ./configure system and
2270 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2271 of a program that users should build themselves after PCRE is installed, so
2272 building it automatically is not really right. What is more, it gave
2273 trouble in some build environments.
2275 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2278 Version 7.5 10-Jan-08
2279 ---------------------
2281 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2282 values in parens when parsing an RE using the C++ wrapper."
2284 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2285 Characters greater than 255 were excluded from the class instead of being
2286 included.
2288 3. The same bug as (2) above applied to negated POSIX classes such as
2289 [:^space:].
2291 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2292 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2293 I have changed it.
2295 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2296 first named subpattern) and a construct such as (?&a) would reference the
2297 first named subpattern whose name started with "a" (in other words, the
2298 length check was missing). Both these problems are fixed. "Subpattern name
2299 expected" is now given for (?&) (a zero-length name), and this patch also
2300 makes it give the same error for \k'' (previously it complained that that
2301 was a reference to a non-existent subpattern).
2303 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2304 this is right because (?- can be followed by option settings as well as by
2305 digits. I have, however, made the messages clearer.
2307 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2308 than the number used in the conditional) now cause a compile-time error.
2309 This is actually not compatible with Perl, which accepts such patterns, but
2310 treats the conditional as always being FALSE (as PCRE used to), but it
2311 seems to me that giving a diagnostic is better.
2313 8. Change "alphameric" to the more common word "alphanumeric" in comments
2314 and messages.
2316 9. Fix two occurrences of "backslash" in comments that should have been
2317 "backspace".
2319 10. Remove two redundant lines of code that can never be obeyed (their function
2320 was moved elsewhere).
2322 11. The program that makes PCRE's Unicode character property table had a bug
2323 which caused it to generate incorrect table entries for sequences of
2324 characters that have the same character type, but are in different scripts.
2325 It amalgamated them into a single range, with the script of the first of
2326 them. In other words, some characters were in the wrong script. There were
2327 thirteen such cases, affecting characters in the following ranges:
2329 U+002b0 - U+002c1
2330 U+0060c - U+0060d
2331 U+0061e - U+00612
2332 U+0064b - U+0065e
2333 U+0074d - U+0076d
2334 U+01800 - U+01805
2335 U+01d00 - U+01d77
2336 U+01d9b - U+01dbf
2337 U+0200b - U+0200f
2338 U+030fc - U+030fe
2339 U+03260 - U+0327f
2340 U+0fb46 - U+0fbb1
2341 U+10450 - U+1049d
2343 12. The -o option (show only the matching part of a line) for pcregrep was not
2344 compatible with GNU grep in that, if there was more than one match in a
2345 line, it showed only the first of them. It now behaves in the same way as
2346 GNU grep.
2348 13. If the -o and -v options were combined for pcregrep, it printed a blank
2349 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2350 does the same. The return code can be used to tell if there were any
2351 non-matching lines.
2353 14. Added --file-offsets and --line-offsets to pcregrep.
2355 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2356 infinitely looping recursion. The bug was that positive lookaheads were not
2357 being skipped when checking for a possible empty match (negative lookaheads
2358 and both kinds of lookbehind were skipped).
2360 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2361 inclusion of <windows.h> to before rather than after the definition of
2362 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2364 17. Specifying a possessive quantifier with a specific limit for a Unicode
2365 character property caused pcre_compile() to compile bad code, which led at
2366 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2367 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2368 caused the error; without that there was no problem.
2370 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2372 19. Added --enable-pcretest-libreadline.
2374 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2375 RE::GlobalReplace(). As a result, the number of replacements returned was
2376 double what it should be. I removed one of the increments, but Craig sent a
2377 later patch that removed the other one (the right fix) and added unit tests
2378 that check the return values (which was not done before).
2380 21. Several CMake things:
2382 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2383 the names libpcre and libpcreposix, not just pcre and pcreposix.
2385 (2) The above change means that pcretest and pcregrep are now correctly
2386 linked with the newly-built libraries, not previously installed ones.
2390 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2391 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2392 UTF-8 newline character). The key issue is that the pattern starts .*;
2393 this means that the match must be either at the beginning, or after a
2394 newline. The bug was in the code for advancing after a failed match and
2395 checking that the new position followed a newline. It was not taking
2396 account of UTF-8 characters correctly.
2398 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2399 character classes. PCRE was not treating the sequence [:...:] as a
2400 character class unless the ... were all letters. Perl, however, seems to
2401 allow any characters between [: and :], though of course it rejects as
2402 unknown any "names" that contain non-letters, because all the known class
2403 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2404 for example, whereas PCRE did not - it did not recognize a POSIX character
2405 class. This seemed a bit dangerous, so the code has been changed to be
2406 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2407 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2408 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2409 Perl does, and where it didn't before.
2411 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2412 Windows environments %n is disabled by default.
2415 Version 7.4 21-Sep-07
2416 ---------------------
2418 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2419 means that a class such as [\s] counted as "explicit reference to CR or
2420 LF". That isn't really right - the whole point of the change was to try to
2421 help when there was an actual mention of one of the two characters. So now
2422 the change happens only if \r or \n (or a literal CR or LF) character is
2423 encountered.
2425 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2426 of both had grown to the point where there were only 3 bits left.
2427 Fortunately, there was spare space in the data structure, and so I have
2428 moved the internal flags into a new 16-bit field to free up more option
2429 bits.
2431 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2432 but did not set the internal JCHANGED flag - either of these is enough to
2433 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2434 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2435 start sets both bits.
2437 4. Added options (at build time, compile time, exec time) to change \R from
2438 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2440 5. doc/pcresyntax.html was missing from the distribution.
2442 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2443 compatibility, even though it is no longer used.
2445 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2446 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2447 windows.h file is present (where different names are used). [This was
2448 reversed later after testing - see 16 below.]
2450 8. Changed all #include <config.h> to #include "config.h". There were also
2451 some further <pcre.h> cases that I changed to "pcre.h".
2453 9. When pcregrep was used with the --colour option, it missed the line ending
2454 sequence off the lines that it output.
2456 10. It was pointed out to me that arrays of string pointers cause lots of
2457 relocations when a shared library is dynamically loaded. A technique of
2458 using a single long string with a table of offsets can drastically reduce
2459 these. I have refactored PCRE in four places to do this. The result is
2460 dramatic:
2462 Originally: 290
2463 After changing UCP table: 187
2464 After changing error message table: 43
2465 After changing table of "verbs" 36
2466 After changing table of Posix names 22
2468 Thanks to the folks working on Gregex for glib for this insight.
2470 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2471 unicode-properties was also set.
2473 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2475 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2476 checked only for CRLF.
2478 14. Added casts to pcretest.c to avoid compiler warnings.
2480 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2482 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2483 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2484 entirely. This removes changes made in 7 above.
2486 17. The CMake files have been updated, and there is now more information about
2487 building with CMake in the NON-UNIX-USE document.
2490 Version 7.3 28-Aug-07
2491 ---------------------
2493 1. In the rejigging of the build system that eventually resulted in 7.1, the
2494 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2495 brackets there is not right, since it causes compilers to look for an
2496 installed pcre.h, not the version that is in the source that is being
2497 compiled (which of course may be different). I have changed it back to:
2499 #include "pcre.h"
2501 I have a vague recollection that the change was concerned with compiling in
2502 different directories, but in the new build system, that is taken care of
2503 by the VPATH setting the Makefile.
2505 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2506 when the subject happened to end in the byte 0x85 (e.g. if the last
2507 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2508 characters but of course it shouldn't be taken as a newline when it is part
2509 of another character. The bug was that, for an unlimited repeat of . in
2510 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2511 characters when looking for a newline.
2513 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2515 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2516 in debug output.
2518 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2519 long printing in the pcrecpp unittest when running under MinGW.
2521 6. ESC_K was left out of the EBCDIC table.
2523 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2524 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2525 limit also applies to "virtual nesting" when a pattern is recursive, and in
2526 this case 1000 isn't so big. I have been able to remove this limit at the
2527 expense of backing off one optimization in certain circumstances. Normally,
2528 when pcre_exec() would call its internal match() function recursively and
2529 immediately return the result unconditionally, it uses a "tail recursion"
2530 feature to save stack. However, when a subpattern that can match an empty
2531 string has an unlimited repetition quantifier, it no longer makes this
2532 optimization. That gives it a stack frame in which to save the data for
2533 checking that an empty string has been matched. Previously this was taken
2534 from the 1000-entry workspace that had been reserved. So now there is no
2535 explicit limit, but more stack is used.
2537 8. Applied Daniel's patches to solve problems with the import/export magic
2538 syntax that is required for Windows, and which was going wrong for the
2539 pcreposix and pcrecpp parts of the library. These were overlooked when this
2540 problem was solved for the main library.
2542 9. There were some crude static tests to avoid integer overflow when computing
2543 the size of patterns that contain repeated groups with explicit upper
2544 limits. As the maximum quantifier is 65535, the maximum group length was
2545 set at 30,000 so that the product of these two numbers did not overflow a
2546 32-bit integer. However, it turns out that people want to use groups that
2547 are longer than 30,000 bytes (though not repeat them that many times).
2548 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2549 made it possible to implement the integer overflow checks in a much more
2550 dynamic way, which I have now done. The artificial limitation on group
2551 length has been removed - we now have only the limit on the total length of
2552 the compiled pattern, which depends on the LINK_SIZE setting.
2554 10. Fixed a bug in the documentation for get/copy named substring when
2555 duplicate names are permitted. If none of the named substrings are set, the
2556 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2557 empty string.
2559 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2560 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2561 because the ] is interpreted as the first data character and the
2562 terminating ] is not found. PCRE has been made compatible with Perl in this
2563 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2564 cause memory overwriting.
2566 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2567 string has been matched (to stop an infinite loop). It was not recognizing
2568 a conditional subpattern that could match an empty string if that
2569 subpattern was within another subpattern. For example, it looped when
2570 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2571 condition was not nested. This bug has been fixed.
2573 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2574 past the start of the subject in the presence of bytes with the top bit
2575 set, for example "\x8aBCD".
2577 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2578 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2580 14. Optimized (?!) to (*FAIL).
2582 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2583 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2584 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2585 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2586 does: it's just the validity check that is more restrictive.
2588 16. Inserted checks for integer overflows during escape sequence (backslash)
2589 processing, and also fixed erroneous offset values for syntax errors during
2590 backslash processing.
2592 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2593 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2595 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2596 caused an overrun.
2598 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2599 something other than just ASCII characters) inside a group that had an
2600 unlimited repeat caused a loop at compile time (while checking to see
2601 whether the group could match an empty string).
2603 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2604 [\P{Any}] did so. (Error in the code for printing property names.)
2606 21. An orphan \E inside a character class could cause a crash.
2608 22. A repeated capturing bracket such as (A)? could cause a wild memory
2609 reference during compilation.
2611 23. There are several functions in pcre_compile() that scan along a compiled
2612 expression for various reasons (e.g. to see if it's fixed length for look
2613 behind). There were bugs in these functions when a repeated \p or \P was
2614 present in the pattern. These operators have additional parameters compared
2615 with \d, etc, and these were not being taken into account when moving along
2616 the compiled data. Specifically:
2618 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2619 length.
2621 (b) An item such as \pL+ within a repeated group could cause crashes or
2622 loops.
2624 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2625 "reference to non-existent subpattern" error.
2627 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2629 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2630 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2632 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2634 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2635 character were causing crashes (broken optimization).
2637 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2638 \p or \P) caused a compile-time loop.
2640 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2641 break. For example, the unstudied pattern [\r\n]A does not match the string
2642 "\r\nA" because change 7.0/46 below moves the current point on by two
2643 characters after failing to match at the start. However, the pattern \nA
2644 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2645 the same is true. There doesn't seem any very clean way out of this, but
2646 what I have chosen to do makes the common cases work: PCRE now takes note
2647 of whether there can be an explicit match for \r or \n anywhere in the
2648 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2649 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2650 pattern has explicit CR or LF references.
2652 29. Added (*CR) etc for changing newline setting at start of pattern.
2655 Version 7.2 19-Jun-07
2656 ---------------------
2658 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2659 which is apparently normally available under Windows.
2661 2. Re-jig the pcregrep tests with different newline settings in an attempt
2662 to make them independent of the local environment's newline setting.
2664 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2666 4. Some of the "internals" tests were previously cut out when the link size
2667 was not 2, because the output contained actual offsets. The recent new
2668 "Z" feature of pcretest means that these can be cut out, making the tests
2669 usable with all link sizes.
2671 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2672 stack recursion. This gives a massive performance boost under BSD, but just
2673 a small improvement under Linux. However, it saves one field in the frame
2674 in all cases.
2676 6. Added more features from the forthcoming Perl 5.10:
2678 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2679 recursion call. It refers to the nth most recently opened parentheses.
2681 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2682 to be opened parentheses.
2684 (c) Conditions that refer to capturing parentheses can be specified
2685 relatively, for example, (?(-2)... or (?(+3)...
2687 (d) \K resets the start of the current match so that everything before
2688 is not part of it.
2690 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2692 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2693 reference syntax.
2695 (g) (?| introduces a group in which the numbering of parentheses in each
2696 alternative starts with the same number.
2698 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2700 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2703 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2704 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2705 for detecting groups that can match an empty string.
2707 9. A pattern with a very large number of alternatives (more than several
2708 hundred) was running out of internal workspace during the pre-compile
2709 phase, where pcre_compile() figures out how much memory will be needed. A
2710 bit of new cunning has reduced the workspace needed for groups with
2711 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2712 workspace instead of running out of the 4096 that are available.
2714 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2716 11. Applied patch from Google to remove an optimization that didn't quite work.
2717 The report of the bug said:
2719 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2720 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2721 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2723 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2724 it matched the wrong number of bytes.
2727 Version 7.1 24-Apr-07
2728 ---------------------
2730 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2731 that is more "standard", making use of automake and other Autotools. There
2732 is some re-arrangement of the files and adjustment of comments consequent
2733 on this.
2735 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2736 for recursive directory scanning broke on some systems because the files
2737 are not scanned in any specific order and on different systems the order
2738 was different. A call to "sort" has been inserted into RunGrepTest for the
2739 approprate test as a short-term fix. In the longer term there may be an
2740 alternative.
2742 3. I had an email from Eric Raymond about problems translating some of PCRE's
2743 man pages to HTML (despite the fact that I distribute HTML pages, some
2744 people do their own conversions for various reasons). The problems
2745 concerned the use of low-level troff macros .br and .in. I have therefore
2746 removed all such uses from the man pages (some were redundant, some could
2747 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2748 HTML has been updated to handle .nf/.fi and to complain if it encounters
2749 .br or .in.
2751 4. Updated comments in configure.ac that get placed in config.h.in and also
2752 arranged for config.h to be included in the distribution, with the name
2753 config.h.generic, for the benefit of those who have to compile without
2754 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2756 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2757 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2758 makevp.bat for the current PCRE, using the additional files
2759 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2761 6. A Windows user reported a minor discrepancy with test 2, which turned out
2762 to be caused by a trailing space on an input line that had got lost in his
2763 copy. The trailing space was an accident, so I've just removed it.
2765 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2766 that is needed.
2768 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2769 as "const" (a) because they are and (b) because it helps the PHP
2770 maintainers who have recently made a script to detect big data structures
2771 in the php code that should be moved to the .rodata section. I remembered
2772 to update Builducptable as well, so it won't revert if ucptable.h is ever
2773 re-created.
2775 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2776 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2777 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2778 support is not required. This saves 1.5-2K of code, which is important in
2779 some applications.
2781 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2782 so as not to refer to the tables, even though these functions will never be
2783 called when UTF-8 support is disabled. Otherwise there are problems with a
2784 shared library.
2786 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2788 (a) It was defining its arguments as char * instead of void *.
2790 (b) It was assuming that all moves were upwards in memory; this was true
2791 a long time ago when I wrote it, but is no longer the case.
2793 The emulated memove() is provided for those environments that have neither
2794 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2795 is clearly not the case, as these two bugs were recently reported.
2797 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2798 and Detrail to create the HTML documentation, the .txt form of the man
2799 pages, and it removes trailing spaces from listed files. It also creates
2800 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2801 case, it wraps all the #defines with #ifndefs. This script should be run
2802 before "make dist".
2804 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2805 with Unicode property support.
2807 (a) For a maximizing quantifier, if the two different cases of the
2808 character were of different lengths in their UTF-8 codings (there are
2809 some cases like this - I found 11), and the matching function had to
2810 back up over a mixture of the two cases, it incorrectly assumed they
2811 were both the same length.
2813 (b) When PCRE was configured to use the heap rather than the stack for
2814 recursion during matching, it was not correctly preserving the data for
2815 the other case of a UTF-8 character when checking ahead for a match
2816 while processing a minimizing repeat. If the check also involved
2817 matching a wide character, but failed, corruption could cause an
2818 erroneous result when trying to check for a repeat of the original
2819 character.
2821 13. Some tidying changes to the testing mechanism:
2823 (a) The RunTest script now detects the internal link size and whether there
2824 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2825 values substituted by "configure". (The RunGrepTest script already did
2826 this for UTF-8.) The configure.ac script no longer substitutes the
2827 relevant variables.
2829 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2830 with length and offset values. This means that the output is different
2831 for different internal link sizes. Test 2 is skipped for link sizes
2832 other than 2 because of this, bypassing the problem. Unfortunately,
2833 there was also a test in test 3 (the locale tests) that used /B and
2834 failed for link sizes other than 2. Rather than cut the whole test out,
2835 I have added a new /Z option to pcretest that replaces the length and
2836 offset values with spaces. This is now used to make test 3 independent
2837 of link size. (Test 2 will be tidied up later.)
2839 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2840 segmentation fault instead of returning the appropriate error message.
2842 15. In multiline mode when the newline sequence was set to "any", the pattern
2843 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2844 This doesn't seem right; it now treats the CRLF combination as the line
2845 ending, and so does not match in that case. It's only a pattern such as ^$
2846 that would hit this one: something like ^ABC$ would have failed after \r
2847 and then tried again after \r\n.
2849 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2850 in an attempt to make files that differ only in their line terminators
2851 compare equal. This works on Linux.
2853 17. Under certain error circumstances pcregrep might try to free random memory
2854 as it exited. This is now fixed, thanks to valgrind.
2856 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2857 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2858 was because its rules for how to advance for /g after matching an empty
2859 string at the end of a line did not allow for this case. They now check for
2860 it specially.
2862 20. pcretest is supposed to handle patterns and data of any length, by
2863 extending its buffers when necessary. It was getting this wrong when the
2864 buffer for a data line had to be extended.
2866 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2867 CRLF as a newline sequence.
2869 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2870 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2871 I have nevertheless tidied it up.
2873 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2875 24. Added a man page for pcre-config.
2878 Version 7.0 19-Dec-06
2879 ---------------------
2881 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2882 moving to gcc 4.1.1.
2884 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2885 sys/time.h, which is documented as needed for this function. It doesn't
2886 seem to matter on Linux, but it showed up on some releases of OS X.
2888 3. It seems that there are systems where bytes whose values are greater than
2889 127 match isprint() in the "C" locale. The "C" locale should be the
2890 default when a C program starts up. In most systems, only ASCII printing
2891 characters match isprint(). This difference caused the output from pcretest
2892 to vary, making some of the tests fail. I have changed pcretest so that:
2894 (a) When it is outputting text in the compiled version of a pattern, bytes
2895 other than 32-126 are always shown as hex escapes.
2897 (b) When it is outputting text that is a matched part of a subject string,
2898 it does the same, unless a different locale has been set for the match
2899 (using the /L modifier). In this case, it uses isprint() to decide.
2901 4. Fixed a major bug that caused incorrect computation of the amount of memory
2902 required for a compiled pattern when options that changed within the
2903 pattern affected the logic of the preliminary scan that determines the
2904 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2905 that the computed length was too small. The symptoms of this bug were
2906 either the PCRE error "internal error: code overflow" from pcre_compile(),
2907 or a glibc crash with a message such as "pcretest: free(): invalid next
2908 size (fast)". Examples of patterns that provoked this bug (shown in
2909 pcretest format) are:
2911 /(?-x: )/x
2912 /(?x)(?-x: \s*#\s*)/
2913 /((?i)[\x{c0}])/8
2914 /(?i:[\x{c0}])/8
2916 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2917 is now done differently.
2919 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2920 wrapper classes; (b) implement a new function in the C++ scanner that is
2921 more efficient than the old way of doing things because it avoids levels of
2922 recursion in the regex matching; (c) add a paragraph to the documentation
2923 for the FullMatch() function.
2925 6. The escape sequence \n was being treated as whatever was defined as
2926 "newline". Not only was this contrary to the documentation, which states
2927 that \n is character 10 (hex 0A), but it also went horribly wrong when
2928 "newline" was defined as CRLF. This has been fixed.
2930 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2931 was being set to -1 for the "end of line" case (supposedly a value that no
2932 character can have). Though this value is never used (the check for end of
2933 line is "zero bytes in current character"), it caused compiler complaints.
2934 I've changed it to 0xffffffff.
2936 8. In pcre_version.c, the version string was being built by a sequence of
2937 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2938 string (as it is for production releases) called a macro with an empty
2939 argument. The C standard says the result of this is undefined. The gcc
2940 compiler treats it as an empty string (which was what was wanted) but it is
2941 reported that Visual C gives an error. The source has been hacked around to
2942 avoid this problem.
2944 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2945 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2946 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2947 of them did).
2949 10. Originally, pcretest opened its input and output without "b"; then I was
2950 told that "b" was needed in some environments, so it was added for release
2951 5.0 to both the input and output. (It makes no difference on Unix-like
2952 systems.) Later I was told that it is wrong for the input on Windows. I've
2953 now abstracted the modes into two macros, to make it easier to fiddle with
2954 them, and removed "b" from the input mode under Windows.
2956 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2958 12. Added -help and --help to pcretest as an official way of being reminded
2959 of the options.
2961 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2962 and pcrecpp.cc because they annoy compilers at high warning levels.
2964 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2966 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2967 scripts are not C programs :-) and which was not noticed because it works
2968 on Linux.
2970 16. pcretest is supposed to handle any length of pattern and data line (as one
2971 line or as a continued sequence of lines) by extending its input buffer if
2972 necessary. This feature was broken for very long pattern lines, leading to
2973 a string of junk being passed to pcre_compile() if the pattern was longer
2974 than about 50K.
2976 17. I have done a major re-factoring of the way pcre_compile() computes the
2977 amount of memory needed for a compiled pattern. Previously, there was code
2978 that made a preliminary scan of the pattern in order to do this. That was
2979 OK when PCRE was new, but as the facilities have expanded, it has become
2980 harder and harder to keep it in step with the real compile phase, and there
2981 have been a number of bugs (see for example, 4 above). I have now found a
2982 cunning way of running the real compile function in a "fake" mode that
2983 enables it to compute how much memory it would need, while actually only
2984 ever using a few hundred bytes of working memory and without too many
2985 tests of the mode. This should make future maintenance and development
2986 easier. A side effect of this work is that the limit of 200 on the nesting
2987 depth of parentheses has been removed (though this was never a serious
2988 limitation, I suspect). However, there is a downside: pcre_compile() now
2989 runs more slowly than before (30% or more, depending on the pattern). I
2990 hope this isn't a big issue. There is no effect on runtime performance.
2992 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2993 newline (only possible for the last line of a file) and it was a
2994 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2996 19. Added additional timing features to pcretest. (1) The -tm option now times
2997 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2998 separate command line item, by a number that specifies the number of
2999 repeats to use when timing. The default is 50000; this gives better
3000 precision, but takes uncomfortably long for very large patterns.
3002 20. Extended pcre_study() to be more clever in cases where a branch of a
3003 subpattern has no definite first character. For example, (a*|b*)[cd] would
3004 previously give no result from pcre_study(). Now it recognizes that the
3005 first character must be a, b, c, or d.
3007 21. There was an incorrect error "recursive call could loop indefinitely" if
3008 a subpattern (or the entire pattern) that was being tested for matching an
3009 empty string contained only one non-empty item after a nested subpattern.
3010 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3011 incorrectly, because the \d was being skipped in the check.
3013 22. The pcretest program now has a new pattern option /B and a command line
3014 option -b, which is equivalent to adding /B to every pattern. This causes
3015 it to show the compiled bytecode, without the additional information that
3016 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3017 is the same as /B/I).
3019 23. A new optimization is now able automatically to treat some sequences such
3020 as a*b as a*+b. More specifically, if something simple (such as a character
3021 or a simple class like \d) has an unlimited quantifier, and is followed by
3022 something that cannot possibly match the quantified thing, the quantifier
3023 is automatically "possessified".
3025 24. A recursive reference to a subpattern whose number was greater than 39
3026 went wrong under certain circumstances in UTF-8 mode. This bug could also
3027 have affected the operation of pcre_study().
3029 25. Realized that a little bit of performance could be had by replacing
3030 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3032 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3034 27. Possessive quantifiers such as a++ were previously implemented by turning
3035 them into atomic groups such as ($>a+). Now they have their own opcodes,
3036 which improves performance. This includes the automatically created ones
3037 from 23 above.
3039 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3040 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3041 the first matched character to be a colon. This applied both to named and
3042 numbered groups.
3044 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3046 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3047 building PCRE on Windows easier, so I have included it in the distribution.
3049 31. There is now a check in pcretest against a ridiculously large number being
3050 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3051 loop, the loop is abandoned.
3053 32. Forward references to subpatterns in conditions such as (?(2)...) where
3054 subpattern 2 is defined later cause pcre_compile() to search forwards in
3055 the pattern for the relevant set of parentheses. This search went wrong
3056 when there were unescaped parentheses in a character class, parentheses
3057 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3059 33. "Subroutine" calls and backreferences were previously restricted to
3060 referencing subpatterns earlier in the regex. This restriction has now
3061 been removed.
3063 34. Added a number of extra features that are going to be in Perl 5.10. On the
3064 whole, these are just syntactic alternatives for features that PCRE had
3065 previously implemented using the Python syntax or my own invention. The
3066 other formats are all retained for compatibility.
3068 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3069 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3070 also .NET compatible.
3072 (b) A recursion or subroutine call to a named group can now be defined as
3073 (?&name) as well as (?P>name).
3075 (c) A backreference to a named group can now be defined as \k<name> or
3076 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3077 5.10, are also .NET compatible.
3079 (d) A conditional reference to a named group can now use the syntax
3080 (?(<name>) or (?('name') as well as (?(name).
3082 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3083 groups (named and numbered) that are never evaluated inline, but can be
3084 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3085 is always false. There may be only one alternative in such a group.
3087 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3088 as the simple (?(R). The condition is true only if the most recent
3089 recursion is that of the given number or name. It does not search out
3090 through the entire recursion stack.
3092 (g) The escape \gN or \g{N} has been added, where N is a positive or
3093 negative number, specifying an absolute or relative reference.
3095 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3096 some "unreachable code" warnings.
3098 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3099 things, this adds five new scripts.
3101 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3102 There were also incompatibilities regarding the handling of \Q..\E inside
3103 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3104 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3106 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3107 matches an empty string, and forcibly breaks the loop. There were bugs in
3108 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3109 against aaaa the result was just "a" rather than "aaaa", for example. Two
3110 separate and independent bugs (that affected different cases) have been
3111 fixed.
3113 39. Refactored the code to abolish the use of different opcodes for small
3114 capturing bracket numbers. This is a tidy that I avoided doing when I
3115 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3116 The new approach is not only tidier, it makes it possible to reduce the
3117 memory needed to fix the previous bug (38).
3119 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3120 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3121 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3122 mode.
3124 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3125 report.
3127 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3128 copy construction and assignment in the C++ wrapper.
3130 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3131 couple of bugs that could have given wrong results in the "--newline=crlf"
3132 case.
3134 44. Added a number of casts and did some reorganization of signed/unsigned int
3135 variables following suggestions from Dair Grant. Also renamed the variable
3136 "this" as "item" because it is a C++ keyword.
3138 45. Arranged for dftables to add
3140 #include "pcre_internal.h"
3142 to pcre_chartables.c because without it, gcc 4.x may remove the array
3143 definition from the final binary if PCRE is built into a static library and
3144 dead code stripping is activated.
3146 46. For an unanchored pattern, if a match attempt fails at the start of a
3147 newline sequence, and the newline setting is CRLF or ANY, and the next two
3148 characters are CRLF, advance by two characters instead of one.
3151 Version 6.7 04-Jul-06
3152 ---------------------
3154 1. In order to handle tests when input lines are enormously long, pcretest has
3155 been re-factored so that it automatically extends its buffers when
3156 necessary. The code is crude, but this _is_ just a test program. The
3157 default size has been increased from 32K to 50K.
3159 2. The code in pcre_study() was using the value of the re argument before
3160 testing it for NULL. (Of course, in any sensible call of the function, it
3161 won't be NULL.)
3163 3. The memmove() emulation function in pcre_internal.h, which is used on
3164 systems that lack both memmove() and bcopy() - that is, hardly ever -
3165 was missing a "static" storage class specifier.
3167 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3168 containing an extended class (one that cannot be represented by a bitmap
3169 because it contains high-valued characters or Unicode property items, e.g.
3170 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3171 pattern, but PCRE should not loop if you do not (it no longer does).
3172 [Detail: two cases were found: (a) a repeated subpattern containing an
3173 extended class; (b) a recursive reference to a subpattern that followed a
3174 previous extended class. It wasn't skipping over the extended class
3175 correctly when UTF-8 mode was not set.]
3177 5. A negated single-character class was not being recognized as fixed-length
3178 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3179 compile error "lookbehind assertion is not fixed length".
3181 6. The RunPerlTest auxiliary script was showing an unexpected difference
3182 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3183 write a Perl script that can interpret lines of an input file either as
3184 byte characters or as UTF-8, which is what "perltest" was being required to
3185 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3186 can't do is switch easily at run time between having the "use utf8;" pragma
3187 or not. In the end, I fudged it by using the RunPerlTest script to insert
3188 "use utf8;" explicitly for the UTF-8 tests.
3190 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3191 the end of the subject string, contrary to the documentation and to what
3192 Perl does. This was true of both matching functions. Now it matches only at
3193 the start of the subject and immediately after *internal* newlines.
3195 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3196 a pointer to an int instead of a pointer to an unsigned long int. This
3197 caused problems on 64-bit systems.
3199 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3200 instance of the 'standard' template library not being so standard".
3202 10. There was no check on the number of named subpatterns nor the maximum
3203 length of a subpattern name. The product of these values is used to compute
3204 the size of the memory block for a compiled pattern. By supplying a very
3205 long subpattern name and a large number of named subpatterns, the size
3206 computation could be caused to overflow. This is now prevented by limiting
3207 the length of names to 32 characters, and the number of named subpatterns
3208 to 10,000.
3210 11. Subpatterns that are repeated with specific counts have to be replicated in
3211 the compiled pattern. The size of memory for this was computed from the
3212 length of the subpattern and the repeat count. The latter is limited to
3213 65535, but there was no limit on the former, meaning that integer overflow
3214 could in principle occur. The compiled length of a repeated subpattern is
3215 now limited to 30,000 bytes in order to prevent this.
3217 12. Added the optional facility to have named substrings with the same name.
3219 13. Added the ability to use a named substring as a condition, using the
3220 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3221 are numbers (not recommended). Forward references are permitted.
3223 14. Added forward references in named backreferences (if you see what I mean).
3225 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3226 pattern could run off the end of the subject. For example, the pattern
3227 "(?s)(.{1,5})"8 did this with the subject "ab".
3229 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3230 PCRE_CASELESS was set when matching characters that were quantified with ?
3231 or *.
3233 17. A character class other than a single negated character that had a minimum
3234 but no maximum quantifier - for example [ab]{6,} - was not handled
3235 correctly by pce_dfa_exec(). It would match only one character.
3237 18. A valid (though odd) pattern that looked like a POSIX character
3238 class but used an invalid character after [ (for example [[,abc,]]) caused
3239 pcre_compile() to give the error "Failed: internal error: code overflow" or
3240 in some cases to crash with a glibc free() error. This could even happen if
3241 the pattern terminated after [[ but there just happened to be a sequence of
3242 letters, a binary zero, and a closing ] in the memory that followed.
3244 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3245 over the years. Originally (before any Unicode support), just the bottom 8
3246 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3247 output from "man perlunicode" includes this:
3249 The regular expression compiler produces polymorphic opcodes. That
3250 is, the pattern adapts to the data and automatically switches to
3251 the Unicode character scheme when presented with Unicode data--or
3252 instead uses a traditional byte scheme when presented with byte
3253 data.
3255 Sadly, a wide octal escape does not cause a switch, and in a string with
3256 no other multibyte characters, these octal escapes are treated as before.
3257 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3258 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3259 Unicode string.
3261 I have not perpetrated such confusion in PCRE. Up till now, it took just
3262 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3263 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3264 translate to the appropriate multibyte character.
3266 29. Applied some refactoring to reduce the number of warnings from Microsoft
3267 and Borland compilers. This has included removing the fudge introduced
3268 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3269 a warning about an unused variable.
3271 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3272 characters since release 4.0, because Perl (from release 5.004) does not.
3273 [Or at least, is documented not to: some releases seem to be in conflict
3274 with the documentation.] However, when a pattern was studied with
3275 pcre_study() and all its branches started with \s, PCRE still included VT
3276 as a possible starting character. Of course, this did no harm; it just
3277 caused an unnecessary match attempt.
3279 22. Removed a now-redundant internal flag bit that recorded the fact that case
3280 dependency changed within the pattern. This was once needed for "required
3281 byte" processing, but is no longer used. This recovers a now-scarce options
3282 bit. Also moved the least significant internal flag bit to the most-
3283 significant bit of the word, which was not previously used (hangover from
3284 the days when it was an int rather than a uint) to free up another bit for
3285 the future.
3287 23. Added support for CRLF line endings as well as CR and LF. As well as the
3288 default being selectable at build time, it can now be changed at runtime
3289 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3290 specify that it is scanning data with non-default line endings.
3292 24. Changed the definition of CXXLINK to make it agree with the definition of
3293 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3295 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3296 recursions. This makes a big different to stack usage for some patterns.
3298 26. If a subpattern containing a named recursion or subroutine reference such
3299 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3300 the space required for the compiled pattern went wrong and gave too small a
3301 value. Depending on the environment, this could lead to "Failed: internal
3302 error: code overflow at offset 49" or "glibc detected double free or
3303 corruption" errors.
3305 27. Applied patches from Google (a) to support the new newline modes and (b) to
3306 advance over multibyte UTF-8 characters in GlobalReplace.
3308 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3309 difference for some implementation of PCRE in some Windows version.
3311 29. Added some extra testing facilities to pcretest:
3313 \q<number> in a data line sets the "match limit" value
3314 \Q<number> in a data line sets the "match recursion limt" value
3315 -S <number> sets the stack size, where <number> is in megabytes
3317 The -S option isn't available for Windows.
3320 Version 6.6 06-Feb-06
3321 ---------------------
3323 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3324 in pcreposix.h. I have copied the definition from pcre.h.
3326 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3327 because pcre.h is no longer a built file.
3329 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3330 not normally included in the compiled code.
3333 Version 6.5 01-Feb-06
3334 ---------------------
3336 1. When using the partial match feature with pcre_dfa_exec(), it was not
3337 anchoring the second and subsequent partial matches at the new starting
3338 point. This could lead to incorrect results. For example, with the pattern
3339 /1234/, partially matching against "123" and then "a4" gave a match.
3341 2. Changes to pcregrep:
3343 (a) All non-match returns from pcre_exec() were being treated as failures
3344 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3345 error message is output. Some extra information is given for the
3347 probably the only errors that are likely to be caused by users (by
3348 specifying a regex that has nested indefinite repeats, for instance).
3349 If there are more than 20 of these errors, pcregrep is abandoned.
3351 (b) A binary zero was treated as data while matching, but terminated the
3352 output line if it was written out. This has been fixed: binary zeroes
3353 are now no different to any other data bytes.
3355 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3356 used to set a locale for matching. The --locale=xxxx long option has
3357 been added (no short equivalent) to specify a locale explicitly on the
3358 pcregrep command, overriding the environment variables.
3360 (d) When -B was used with -n, some line numbers in the output were one less
3361 than they should have been.
3363 (e) Added the -o (--only-matching) option.
3365 (f) If -A or -C was used with -c (count only), some lines of context were
3366 accidentally printed for the final match.
3368 (g) Added the -H (--with-filename) option.
3370 (h) The combination of options -rh failed to suppress file names for files
3371 that were found from directory arguments.
3373 (i) Added the -D (--devices) and -d (--directories) options.
3375 (j) Added the -F (--fixed-strings) option.
3377 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3379 (l) Added the --colo(u)r option.
3381 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3382 is not present by default.
3384 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3385 items such as (?R) or (?1), when the recursion could match a number of
3386 alternatives. If it matched one of the alternatives, but subsequently,
3387 outside the recursion, there was a failure, the code tried to back up into
3388 the recursion. However, because of the way PCRE is implemented, this is not
3389 possible, and the result was an incorrect result from the match.
3391 In order to prevent this happening, the specification of recursion has
3392 been changed so that all such subpatterns are automatically treated as
3393 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3395 4. I had overlooked the fact that, in some locales, there are characters for
3396 which isalpha() is true but neither isupper() nor islower() are true. In
3397 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3398 and ordfeminine) are like this. This affected the treatment of \w and \W
3399 when they appeared in character classes, but not when they appeared outside
3400 a character class. The bit map for "word" characters is now created
3401 separately from the results of isalnum() instead of just taking it from the
3402 upper, lower, and digit maps. (Plus the underscore character, of course.)
3404 5. The above bug also affected the handling of POSIX character classes such as
3405 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3406 permanent tables. Instead, the bit maps for such a class were previously
3407 created as the appropriate unions of the upper, lower, and digit bitmaps.
3408 Now they are created by subtraction from the [[:word:]] class, which has
3409 its own bitmap.
3411 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3412 It is created by subtracting the vertical space characters (\x09, \x0a,
3413 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3414 subtraction was done in the overall bitmap for a character class, meaning
3415 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3416 be recognized. This bug has been fixed.
3418 7. Patches from the folks at Google:
3420 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3421 real life, but is still worth protecting against".
3423 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3424 regular expressions".
3426 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3427 have it.
3429 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3430 "configure" and the latter not, in order to fix a problem somebody had
3431 with compiling the Arg class on HP-UX.
3433 (e) Improve the error-handling of the C++ wrapper a little bit.
3435 (f) New tests for checking recursion limiting.
3437 8. The pcre_memmove() function, which is used only if the environment does not
3438 have a standard memmove() function (and is therefore rarely compiled),
3439 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3440 returning a result (though PCRE never actually uses the result).
3442 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3443 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3444 returned instead of calling malloc() with an overflowing number that would
3445 most likely cause subsequent chaos.
3447 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3449 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3450 with this option is matched, the nmatch and pmatch options of regexec() are
3451 ignored.
3453 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3454 provided in case anyone wants to the the POSIX interface with UTF-8
3455 strings.
3457 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3458 C++ linking (needed for some HP-UX environments).
3460 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3461 (unused parameter) and in the pcre_printint() function (omitted "default"
3462 switch label when the default is to do nothing).
3464 15. Added some code to make it possible, when PCRE is compiled as a C++
3465 library, to replace subject pointers for pcre_exec() with a smart pointer
3466 class, thus making it possible to process discontinuous strings.
3468 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3469 much the same function. They were added by different people who were trying
3470 to make PCRE easy to compile on non-Unix systems. It has been suggested
3471 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3472 for compiling on Windows systems. I have therefore replaced it with
3473 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3474 defaults to "extern" for C or "extern C" for C++, which works fine on
3475 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3476 SCOPE with something explicit in config.h. In addition:
3478 (a) pcreposix.h still had just "extern" instead of either of these macros;
3479 I have replaced it with PCRE_DATA_SCOPE.
3481 (b) Functions such as _pcre_xclass(), which are internal to the library,
3482 but external in the C sense, all had PCRE_EXPORT in their definitions.
3483 This is apparently wrong for the Windows case, so I have removed it.
3484 (It makes no difference on Unix-like systems.)
3486 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3487 of recursive calls to match(). This is different to MATCH_LIMIT because
3488 that limits the total number of calls to match(), not all of which increase
3489 the depth of recursion. Limiting the recursion depth limits the amount of
3490 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3491 when PCRE is compiled, and changed at run time. A patch from Google adds
3492 this functionality to the C++ interface.
3494 18. Changes to the handling of Unicode character properties:
3496 (a) Updated the table to Unicode 4.1.0.
3498 (b) Recognize characters that are not in the table as "Cn" (undefined).
3500 (c) I revised the way the table is implemented to a much improved format
3501 which includes recognition of ranges. It now supports the ranges that
3502 are defined in UnicodeData.txt, and it also amalgamates other
3503 characters into ranges. This has reduced the number of entries in the
3504 table from around 16,000 to around 3,000, thus reducing its size
3505 considerably. I realized I did not need to use a tree structure after
3506 all - a binary chop search is just as efficient. Having reduced the
3507 number of entries, I extended their size from 6 bytes to 8 bytes to
3508 allow for more data.
3510 (d) Added support for Unicode script names via properties such as \p{Han}.
3512 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3513 matching that character.
3515 20. When matching a repeated Unicode property with a minimum greater than zero,
3516 (for example \pL{2,}), PCRE could look past the end of the subject if it
3517 reached it while seeking the minimum number of characters. This could
3518 happen only if some of the characters were more than one byte long, because
3519 there is a check for at least the minimum number of bytes.
3521 21. Refactored the implementation of \p and \P so as to be more general, to
3522 allow for more different types of property in future. This has changed the
3523 compiled form incompatibly. Anybody with saved compiled patterns that use
3524 \p or \P will have to recompile them.
3526 22. Added "Any" and "L&" to the supported property types.
3528 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3529 but give a compile time error if the value is greater than 0xff.
3531 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3532 accidentally not being installed or uninstalled.
3534 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3535 made were to insert the current release number. This seemed silly, because
3536 it made things harder for people building PCRE on systems that don't run
3537 "configure". I have turned pcre.h into a distributed file, no longer built
3538 by "configure", with the version identification directly included. There is
3539 no longer a pcre.h.in file.
3541 However, this change necessitated a change to the pcre-config script as
3542 well. It is built from pcre-config.in, and one of the substitutions was the
3543 release number. I have updated configure.ac so that ./configure now finds
3544 the release number by grepping pcre.h.
3546 26. Added the ability to run the tests under valgrind.
3549 Version 6.4 05-Sep-05
3550 ---------------------
3552 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3553 "--" to be printed when multiple files were scanned, even when none of the
3554 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3555 consider it to be a bug, and have restored the previous behaviour.
3557 2. A couple of code tidies to get rid of compiler warnings.
3559 3. The pcretest program used to cheat by referring to symbols in the library
3560 whose names begin with _pcre_. These are internal symbols that are not
3561 really supposed to be visible externally, and in some environments it is
3562 possible to suppress them. The cheating is now confined to including
3563 certain files from the library's source, which is a bit cleaner.
3565 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3566 file's purpose clearer.
3568 5. Reorganized pcre_ucp_findchar().
3571 Version 6.3 15-Aug-05
3572 ---------------------
3574 1. The file libpcre.pc.in did not have general read permission in the tarball.
3576 2. There were some problems when building without C++ support:
3578 (a) If C++ support was not built, "make install" and "make test" still
3579 tried to test it.
3581 (b) There were problems when the value of CXX was explicitly set. Some
3582 changes have been made to try to fix these, and ...
3584 (c) --disable-cpp can now be used to explicitly disable C++ support.
3586 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3587 backslash in a target when C++ was disabled. This confuses some
3588 versions of "make", apparently. Using an intermediate variable solves
3589 this. (Same for CPP_LOBJ.)
3591 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3592 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3593 necessary on certain architectures.
3595 4. Added a setting of -export-symbols-regex to the link command to remove
3596 those symbols that are exported in the C sense, but actually are local
3597 within the library, and not documented. Their names all begin with
3598 "_pcre_". This is not a perfect job, because (a) we have to except some
3599 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3600 available (and never for static libraries). I have made a note to try to
3601 find a way round (a) in the future.
3604 Version 6.2 01-Aug-05
3605 ---------------------
3607 1. There was no test for integer overflow of quantifier values. A construction
3608 such as {1111111111111111} would give undefined results. What is worse, if
3609 a minimum quantifier for a parenthesized subpattern overflowed and became
3610 negative, the calculation of the memory size went wrong. This could have
3611 led to memory overwriting.
3613 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3615 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3616 operating environments where this matters.
3618 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3619 PCRE options from within the C++ wrapper.
3621 5. Named capturing subpatterns were not being correctly counted when a pattern
3622 was compiled. This caused two problems: (a) If there were more than 100
3623 such subpatterns, the calculation of the memory needed for the whole
3624 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3625 back references of the form \12, where the number was greater than 9, were
3626 not recognized as back references, even though there were sufficient
3627 previous subpatterns.
3629 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3630 versions of gcc, e.g. 2.95.4.
3633 Version 6.1 21-Jun-05
3634 ---------------------
3636 1. There was one reference to the variable "posix" in pcretest.c that was not
3637 surrounded by "#if !defined NOPOSIX".
3639 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3640 the cross-check on the old pcre_info() function, for the benefit of the
3641 cut-down version of PCRE that is currently imported into Exim.
3643 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3644 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3645 patterns that start (?i) but I don't think that matters. The use of (?i) is
3646 just an example; this all applies to the other options as well.
3648 4. Since libtool seems to echo the compile commands it is issuing, the output
3649 from "make" can be reduced a bit by putting "@" in front of each libtool
3650 compile command.
3652 5. Patch from the folks at Google for configure.in to be a bit more thorough
3653 in checking for a suitable C++ installation before trying to compile the
3654 C++ stuff. This should fix a reported problem when a compiler was present,
3655 but no suitable headers.
3657 6. The man pages all had just "PCRE" as their title. I have changed them to
3658 be the relevant file name. I have also arranged that these names are
3659 retained in the file doc/pcre.txt, which is a concatenation in text format
3660 of all the man pages except the little individual ones for each function.
3662 7. The NON-UNIX-USE file had not been updated for the different set of source
3663 files that come with release 6. I also added a few comments about the C++
3664 wrapper.
3667 Version 6.0 07-Jun-05
3668 ---------------------
3670 1. Some minor internal re-organization to help with my DFA experiments.
3672 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3673 didn't matter for the library itself when fully configured, but did matter
3674 when compiling without UCP support, or within Exim, where the ucp files are
3675 not imported.
3677 3. Refactoring of the library code to split up the various functions into
3678 different source modules. The addition of the new DFA matching code (see
3679 below) to a single monolithic source would have made it really too
3680 unwieldy, quite apart from causing all the code to be include in a
3681 statically linked application, when only some functions are used. This is
3682 relevant even without the DFA addition now that patterns can be compiled in
3683 one application and matched in another.
3685 The downside of splitting up is that there have to be some external
3686 functions and data tables that are used internally in different modules of
3687 the library but which are not part of the API. These have all had their
3688 names changed to start with "_pcre_" so that they are unlikely to clash
3689 with other external names.
3691 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3692 a different (DFA) algorithm. Although it is slower than the original
3693 function, it does have some advantages for certain types of matching
3694 problem.
3696 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3697 including restarting after a partial match.
3699 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3700 defined when compiling for Windows was sent to me. I have put it into the
3701 code, though I have no means of testing or verifying it.
3703 7. Added the pcre_refcount() auxiliary function.
3705 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3706 match before or at the first newline in the subject string. In pcretest,
3707 the /f option on a pattern can be used to set this.
3709 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3710 would behave wrongly. This has been present in PCRE since release 4.0.
3712 10. A number of changes to the pcregrep command:
3714 (a) Refactored how -x works; insert ^(...)$ instead of setting
3715 PCRE_ANCHORED and checking the length, in preparation for adding
3716 something similar for -w.
3718 (b) Added the -w (match as a word) option.
3720 (c) Refactored the way lines are read and buffered so as to have more
3721 than one at a time available.
3723 (d) Implemented a pcregrep test script.
3725 (e) Added the -M (multiline match) option. This allows patterns to match
3726 over several lines of the subject. The buffering ensures that at least
3727 8K, or the rest of the document (whichever is the shorter) is available
3728 for matching (and similarly the previous 8K for lookbehind assertions).
3730 (f) Changed the --help output so that it now says
3732 -w, --word-regex(p)
3734 instead of two lines, one with "regex" and the other with "regexp"
3735 because that confused at least one person since the short forms are the
3736 same. (This required a bit of code, as the output is generated
3737 automatically from a table. It wasn't just a text change.)
3739 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3740 option but starts with a hyphen. Could be a pattern or a path name
3741 starting with a hyphen, for instance.
3743 (h) "-" can be given as a file name to represent stdin.
3745 (i) When file names are being printed, "(standard input)" is used for
3746 the standard input, for compatibility with GNU grep. Previously
3747 "<stdin>" was used.
3749 (j) The option --label=xxx can be used to supply a name to be used for
3750 stdin when file names are being printed. There is no short form.
3752 (k) Re-factored the options decoding logic because we are going to add
3753 two more options that take data. Such options can now be given in four
3754 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3756 (l) Added the -A, -B, and -C options for requesting that lines of context
3757 around matches be printed.
3759 (m) Added the -L option to print the names of files that do not contain
3760 any matching lines, that is, the complement of -l.
3762 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3763 continue to scan other files.
3765 (o) The -s option was incorrectly implemented. For compatibility with other
3766 greps, it now suppresses the error message for a non-existent or non-
3767 accessible file (but not the return code). There is a new option called
3768 -q that suppresses the output of matching lines, which was what -s was
3769 previously doing.
3771 (p) Added --include and --exclude options to specify files for inclusion
3772 and exclusion when recursing.
3774 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3775 Hopefully, it now does.
3777 12. Missing cast in pcre_study().
3779 13. Added an "uninstall" target to the makefile.
3781 14. Replaced "extern" in the function prototypes in Makefile.in with
3782 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3783 world, but is set differently for Windows.
3785 15. Added a second compiling function called pcre_compile2(). The only
3786 difference is that it has an extra argument, which is a pointer to an
3787 integer error code. When there is a compile-time failure, this is set
3788 non-zero, in addition to the error test pointer being set to point to an
3789 error message. The new argument may be NULL if no error number is required
3790 (but then you may as well call pcre_compile(), which is now just a
3791 wrapper). This facility is provided because some applications need a
3792 numeric error indication, but it has also enabled me to tidy up the way
3793 compile-time errors are handled in the POSIX wrapper.
3795 16. Added VPATH=.libs to the makefile; this should help when building with one
3796 prefix path and installing with another. (Or so I'm told by someone who
3797 knows more about this stuff than I do.)
3799 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3800 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3801 match everything, including newlines. This is not POSIX-compatible, but
3802 somebody wanted the feature. From pcretest it can be activated by using
3803 both the P and the s flags.
3805 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3807 19. libpcre.pc was being incorrectly installed as executable.
3809 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3810 it now also looks for '\r' so that it will work unmodified on Windows.
3812 21. Added Google's contributed C++ wrapper to the distribution.
3814 22. Added some untidy missing memory free() calls in pcretest, to keep
3815 Electric Fence happy when testing.
3819 Version 5.0 13-Sep-04
3820 ---------------------
3822 1. Internal change: literal characters are no longer packed up into items
3823 containing multiple characters in a single byte-string. Each character
3824 is now matched using a separate opcode. However, there may be more than one
3825 byte in the character in UTF-8 mode.
3827 2. The pcre_callout_block structure has two new fields: pattern_position and
3828 next_item_length. These contain the offset in the pattern to the next match
3829 item, and its length, respectively.
3831 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3832 insertion of callouts before each pattern item. Added the /C option to
3833 pcretest to make use of this.
3835 4. On the advice of a Windows user, the lines
3837 #if defined(_WIN32) || defined(WIN32)
3838 _setmode( _fileno( stdout ), 0x8000 );
3839 #endif /* defined(_WIN32) || defined(WIN32) */
3841 have been added to the source of pcretest. This apparently does useful
3842 magic in relation to line terminators.
3844 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3845 for the benefit of those environments where the "b" makes a difference.
3847 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3848 to know about it. I have put a hack into configure.in that adds in code
3849 to set GCC=yes if CC=icc. This seems to end up at a point in the
3850 generated configure script that is early enough to affect the setting of
3851 compiler options, which is what is needed, but I have no means of testing
3852 whether it really works. (The user who reported this had patched the
3853 generated configure script, which of course I cannot do.)
3855 LATER: After change 22 below (new libtool files), the configure script
3856 seems to know about icc (and also ecc). Therefore, I have commented out
3857 this hack in configure.in.
3859 7. Added support for pkg-config (2 patches were sent in).
3861 8. Negated POSIX character classes that used a combination of internal tables
3862 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3863 [[:^ascii]]. Typically, they would match almost any characters. The other
3864 POSIX classes were not broken in this way.
3866 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3867 to find the match, as PCRE was deluded into thinking that the match had to
3868 start at the start point or following a newline. The same bug applied to
3869 patterns with negative forward assertions or any backward assertions
3870 preceding ".*" at the start, unless the pattern required a fixed first
3871 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3873 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3874 starting at the last subject character, bytes beyond the end of the subject
3875 string were read.
3877 11. Renamed the variable "class" as "classbits" to make life easier for C++
3878 users. (Previously there was a macro definition, but it apparently wasn't
3879 enough.)
3881 12. Added the new field "tables" to the extra data so that tables can be passed
3882 in at exec time, or the internal tables can be re-selected. This allows
3883 a compiled regex to be saved and re-used at a later time by a different
3884 program that might have everything at different addresses.
3886 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3887 -R library as well as a -L library.
3889 14. The debugging options of pcretest (-d on the command line or D on a
3890 pattern) showed incorrect output for anything following an extended class
3891 that contained multibyte characters and which was followed by a quantifier.
3893 15. Added optional support for general category Unicode character properties
3894 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3895 support. It adds about 90K to the size of the library. The meanings of the
3896 inbuilt class escapes such as \d and \s have NOT been changed.
3898 16. Updated pcredemo.c to include calls to free() to release the memory for the
3899 compiled pattern.
3901 17. The generated file chartables.c was being created in the source directory
3902 instead of in the building directory. This caused the build to fail if the
3903 source directory was different from the building directory, and was
3904 read-only.
3906 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3907 file. No doubt somebody will tell me if they don't make sense... Also added
3908 Dan Mooney's comments about building on OpenVMS.
3910 19. Added support for partial matching via the PCRE_PARTIAL option for
3911 pcre_exec() and the \P data escape in pcretest.
3913 20. Extended pcretest with 3 new pattern features:
3915 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3916 write the compiled pattern to the file whose name is "rest-of-line".
3917 This is a straight binary dump of the data, with the saved pointer to
3918 the character tables forced to be NULL. The study data, if any, is
3919 written too. After writing, pcretest reads a new pattern.
3921 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3922 compiled pattern from the given file. There must not be any
3923 occurrences of "<" in the file name (pretty unlikely); if there are,
3924 pcretest will instead treat the initial "<" as a pattern delimiter.
3925 After reading in the pattern, pcretest goes on to read data lines as
3926 usual.
3928 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3929 and 16-bit fields in a compiled pattern, to simulate a pattern that
3930 was compiled on a host of opposite endianness.
3932 21. The pcre-exec() function can now cope with patterns that were compiled on
3933 hosts of opposite endianness, with this restriction:
3935 As for any compiled expression that is saved and used later, the tables
3936 pointer field cannot be preserved; the extra_data field in the arguments
3937 to pcre_exec() should be used to pass in a tables address if a value
3938 other than the default internal tables were used at compile time.
3940 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3941 now diagnosed as an error. Previously, most of the time, a negative number
3942 would have been treated as zero, but if in addition "ovector" was passed as
3943 NULL, a crash could occur.
3945 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3946 new versions from the libtool 1.5 distribution (the last one is a copy of
3947 a file called libtool.m4). This seems to have fixed the need to patch
3948 "configure" to support Darwin 1.3 (which I used to do). However, I still
3949 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3950 workstation).
3952 24. Changed the PCRE licence to be the more standard "BSD" licence.
3955 Version 4.5 01-Dec-03
3956 ---------------------
3958 1. There has been some re-arrangement of the code for the match() function so
3959 that it can be compiled in a version that does not call itself recursively.
3960 Instead, it keeps those local variables that need separate instances for
3961 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3962 needs to "recurse". Keeping track of where control must go is done by means
3963 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3964 hide most of the details from the main code, and operates only if
3965 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3966 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3967 operating.
3969 To make it easier for callers to provide specially tailored get/free
3970 functions for this usage, two new functions, pcre_stack_malloc, and
3971 pcre_stack_free, are used. They are always called in strict stacking order,
3972 and the size of block requested is always the same.
3974 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3975 PCRE has been compiled to use the stack or the heap for recursion. The
3976 -C option of pcretest uses this to show which version is compiled.
3978 A new data escape \S, is added to pcretest; it causes the amounts of store
3979 obtained and freed by both kinds of malloc/free at match time to be added
3980 to the output.
3982 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3983 what's available on my current Linux desktop machine.
3985 3. When matching a UTF-8 string, the test for a valid string at the start has
3986 been extended. If start_offset is not zero, PCRE now checks that it points
3987 to a byte that is the start of a UTF-8 character. If not, it returns
3988 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3989 this is necessary because there may be backward assertions in the pattern.
3990 When matching the same subject several times, it may save resources to use
3991 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3993 4. The code for checking the validity of UTF-8 strings has been tightened so
3994 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3995 containing "overlong sequences".
3997 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3998 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3999 should have been "&", but it just so happened that all the cases this let
4000 through by mistake were picked up later in the function.
4002 6. I had used a variable called "isblank" - this is a C99 function, causing
4003 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4005 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4006 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4007 so the version is shown; (c) stop "make test" echoing "./RunTest".
4009 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4011 9. The prototype for memmove() for systems that don't have it was using
4012 size_t, but the inclusion of the header that defines size_t was later. I've
4013 moved the #includes for the C headers earlier to avoid this.
4015 10. Added some adjustments to the code to make it easier to compiler on certain
4016 special systems:
4018 (a) Some "const" qualifiers were missing.
4019 (b) Added the macro EXPORT before all exported functions; by default this
4020 is defined to be empty.
4021 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4022 that it reads its output file name as an argument instead of writing
4023 to the standard output and assuming this can be redirected.
4025 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4026 class containing characters with values greater than 255, PCRE compilation
4027 went into a loop.
4029 12. A recursive reference to a subpattern that was within another subpattern
4030 that had a minimum quantifier of zero caused PCRE to crash. For example,
4031 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4032 recursion. If the recursively-called subpattern itself had a zero repeat,
4033 that was OK.
4035 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4036 buffer into which it was copied (for escape processing) was still set at
4037 1024, so long lines caused crashes.
4039 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4040 "internal error: code overflow...". This applied to any character class
4041 that was followed by a possessive quantifier.
4043 15. Modified the Makefile to add libpcre.la as a prerequisite for
4044 libpcreposix.la because I was told this is needed for a parallel build to
4045 work.
4047 16. If a pattern that contained .* following optional items at the start was
4048 studied, the wrong optimizing data was generated, leading to matching
4049 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4050 matching string must start with a or b or c. The correct conclusion for
4051 this pattern is that a match can start with any character.
4054 Version 4.4 13-Aug-03
4055 ---------------------
4057 1. In UTF-8 mode, a character class containing characters with values between
4058 127 and 255 was not handled correctly if the compiled pattern was studied.
4059 In fixing this, I have also improved the studying algorithm for such
4060 classes (slightly).
4062 2. Three internal functions had redundant arguments passed to them. Removal
4063 might give a very teeny performance improvement.
4065 3. Documentation bug: the value of the capture_top field in a callout is *one
4066 more than* the number of the hightest numbered captured substring.
4068 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4069 in incorrectly linking with a previously installed version. They now link
4070 explicitly with libpcre.la.
4072 5. configure.in no longer needs to recognize Cygwin specially.
4074 6. A problem in pcre.in for Windows platforms is fixed.
4076 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4077 pcretest, it used to include the size of the study block as part of its
4078 output. Unfortunately, the structure contains a field that has a different
4079 size on different hardware architectures. This meant that the tests that
4080 showed this size failed. As the block is currently always of a fixed size,
4081 this information isn't actually particularly useful in pcretest output, so
4082 I have just removed it.
4084 8. Three pre-processor statements accidentally did not start in column 1.
4085 Sadly, there are *still* compilers around that complain, even though
4086 standard C has not required this for well over a decade. Sigh.
4088 9. In pcretest, the code for checking callouts passed small integers in the
4089 callout_data field, which is a void * field. However, some picky compilers
4090 complained about the casts involved for this on 64-bit systems. Now
4091 pcretest passes the address of the small integer instead, which should get
4092 rid of the warnings.
4094 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4095 both compile and run time, and gives an error if an invalid UTF-8 sequence
4096 is found. There is a option for disabling this check in cases where the
4097 string is known to be correct and/or the maximum performance is wanted.
4099 11. In response to a bug report, I changed one line in Makefile.in from
4101 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4102 to
4103 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4105 to look similar to other lines, but I have no way of telling whether this
4106 is the right thing to do, as I do not use Windows. No doubt I'll get told
4107 if it's wrong...
4110 Version 4.3 21-May-03
4111 ---------------------
4113 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4114 Makefile.
4116 2. Some refactoring to improve the quality of the code:
4118 (i) The utf8_table... variables are now declared "const".
4120 (ii) The code for \cx, which used the "case flipping" table to upper case
4121 lower case letters, now just substracts 32. This is ASCII-specific,
4122 but the whole concept of \cx is ASCII-specific, so it seems
4123 reasonable.
4125 (iii) PCRE was using its character types table to recognize decimal and
4126 hexadecimal digits in the pattern. This is silly, because it handles
4127 only 0-9, a-f, and A-F, but the character types table is locale-
4128 specific, which means strange things might happen. A private
4129 table is now used for this - though it costs 256 bytes, a table is
4130 much faster than multiple explicit tests. Of course, the standard
4131 character types table is still used for matching digits in subject
4132 strings against \d.
4134 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4135 ending in _t are). So I've renamed it as ESC_tee.
4137 3. The first argument for regexec() in the POSIX wrapper should have been
4138 defined as "const".
4140 4. Changed pcretest to use malloc() for its buffers so that they can be
4141 Electric Fenced for debugging.
4143 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4144 to read one or more bytes before the start of the subject string. Often this
4145 had no effect on PCRE's behaviour, but in some circumstances it could
4146 provoke a segmentation fault.
4148 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4149 to try to read one or more bytes before the start of the subject string.
4151 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4152 UTF-8 support could misbehave in various ways if the subject string
4153 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4154 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4155 back over UTF-8 characters.)
4158 Version 4.2 14-Apr-03
4159 ---------------------
4161 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4163 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4164 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4165 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4166 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4167 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4169 Note: automatic setting of the BUILD variables is not yet working
4170 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4171 compile-time but not at link-time
4172 [LINK]: use for linking executables only
4173 make different versions for Windows and non-Windows
4174 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4175 libraries
4176 [LINK_FOR_BUILD]: new variable
4177 [OBJEXT]: use throughout
4178 [EXEEXT]: use throughout
4179 <winshared>: new target
4180 <wininstall>: new target
4181 <dftables.o>: use native compiler
4182 <dftables>: use native linker
4183 <install>: handle Windows platform correctly
4184 <clean>: ditto
4185 <check>: ditto
4186 copy DLL to top builddir before testing
4188 As part of these changes, -no-undefined was removed again. This was reported
4189 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4190 in any case.
4192 3. Some tidies to get rid of compiler warnings:
4194 . In the match_data structure, match_limit was an unsigned long int, whereas
4195 match_call_count was an int. I've made them both unsigned long ints.
4197 . In pcretest the fact that a const uschar * doesn't automatically cast to
4198 a void * provoked a warning.
4200 . Turning on some more compiler warnings threw up some "shadow" variables
4201 and a few more missing casts.
4203 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4204 option, a class that contained a single character with a value between 128
4205 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4207 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4208 option, a class that contained several characters, but with at least one
4209 whose value was between 128 and 255 caused PCRE to crash.
4212 Version 4.1 12-Mar-03
4213 ---------------------
4215 1. Compiling with gcc -pedantic found a couple of places where casts were
4216 needed, and a string in dftables.c that was longer than standard compilers are
4217 required to support.
4219 2. Compiling with Sun's compiler found a few more places where the code could
4220 be tidied up in order to avoid warnings.
4222 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4223 first of these names is deprecated in the latest Autoconf in favour of the name
4224 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4225 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4226 analogy I have changed it to CFLAGS_FOR_BUILD.
4228 4. Added -no-undefined to the linking command in the Makefile, because this is
4229 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4230 linking step for the pcreposix library.
4232 5. PCRE was failing to diagnose the case of two named groups with the same
4233 name.
4235 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4236 literal character that is needed in the subject for a match, and scans along to
4237 ensure that it is present before embarking on the full matching process. This
4238 saves time in cases of nested unlimited repeats that are never going to match.
4239 Problem: the scan can take a lot of time if the subject is very long (e.g.
4240 megabytes), thus penalizing straightforward matches. It is now done only if the
4241 amount of subject to be scanned is less than 1000 bytes.
4243 7. A lesser problem with the same optimization is that it was recording the
4244 first character of an anchored pattern as "needed", thus provoking a search
4245 right along the subject, even when the first match of the pattern was going to
4246 fail. The "needed" character is now not set for anchored patterns, unless it
4247 follows something in the pattern that is of non-fixed length. Thus, it still
4248 fulfils its original purpose of finding quick non-matches in cases of nested
4249 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4252 Version 4.0 17-Feb-03
4253 ---------------------
4255 1. If a comment in an extended regex that started immediately after a meta-item
4256 extended to the end of string, PCRE compiled incorrect data. This could lead to
4257 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4259 2. Moved to autoconf 2.53 and libtool 1.4.2.
4261 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4262 the special perltest8 script is no longer needed - all the tests can be run
4263 from a single perltest script.
4265 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4266 by \s. It has now been removed in PCRE. This means it isn't recognized as
4267 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4268 class [:space:] *does* include VT, thereby creating a mess.
4270 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4271 space and tab.
4273 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4274 its new features into the main test script, reducing the number of scripts.
4276 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4277 were backward compatible, and made the (?i) apply to the whole pattern, as if
4278 /i were given. Now it behaves more logically, and applies the option setting
4279 only to what follows. PCRE has been changed to follow suit. However, if it
4280 finds options settings right at the start of the pattern, it extracts them into
4281 the global options, as before. Thus, they show up in the info data.
4283 8. Added support for the \Q...\E escape sequence. Characters in between are
4284 treated as literals. This is slightly different from Perl in that $ and @ are
4285 also handled as literals inside the quotes. In Perl, they will cause variable
4286 interpolation. Note the following examples:
4288 Pattern PCRE matches Perl matches
4290 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4291 \Qabc\$xyz\E abc\$xyz abc\$xyz
4292 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4294 For compatibility with Perl, \Q...\E sequences are recognized inside character
4295 classes as well as outside them.
4297 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4298 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4299 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4300 signed/unsigned warnings.
4302 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4303 option for pcretest, so I've replaced it by a simple function that does just
4304 that job.
4306 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4307 "pcregrep -".
4309 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4310 Java package. This provides some syntactic sugar for simple cases of what my
4311 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4312 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4313 item, you can use this simplified notation. Note that only makes sense with
4314 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4315 greediness, whatever the setting of the PCRE_UNGREEDY option.
4317 13. A change of greediness default within a pattern was not taking effect at
4318 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4319 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4320 was abstracted outside.
4322 14. PCRE now supports the \G assertion. It is true when the current matching
4323 position is at the start point of the match. This differs from \A when the
4324 starting offset is non-zero. Used with the /g option of pcretest (or similar
4325 code), it works in the same way as it does for Perl's /g option. If all
4326 alternatives of a regex begin with \G, the expression is anchored to the start
4327 match position, and the "anchored" flag is set in the compiled expression.
4329 15. Some bugs concerning the handling of certain option changes within patterns
4330 have been fixed. These applied to options other than (?ims). For example,
4331 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4332 been the other way round. Some of this was related to change 7 above.
4334 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4335 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4336 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4337 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4339 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4340 mode. Unlike ".", it always matches newline, whatever the setting of
4341 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4342 assertions. Perl allows it, but it doesn't (in general) work because it can't
4343 calculate the length of the lookbehind. At least, that's the case for Perl
4344 5.8.0 - I've been told they are going to document that it doesn't work in
4345 future.
4347 18. Added an error diagnosis for escapes that PCRE does not support: these are
4348 \L, \l, \N, \P, \p, \U, \u, and \X.
4350 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4351 reading past the end of the pattern in cases such as /[abcd/.
4353 20. PCRE was getting more memory than necessary for patterns with classes that
4354 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4356 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4357 compiling PCRE for use with Virtual Pascal.
4359 22. Small fix to the Makefile to make it work properly if the build is done
4360 outside the source tree.
4362 23. Added a new extension: a condition to go with recursion. If a conditional
4363 subpattern starts with (?(R) the "true" branch is used if recursion has
4364 happened, whereas the "false" branch is used only at the top level.
4366 24. When there was a very long string of literal characters (over 255 bytes
4367 without UTF support, over 250 bytes with UTF support), the computation of how
4368 much memory was required could be incorrect, leading to segfaults or other
4369 strange effects.
4371 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4372 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4373 there was a subsequent back reference to those brackets. This meant that, for
4374 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4375 possible to check for precisely this case. All we can do is abandon the
4376 optimization if .* occurs inside capturing brackets when there are any back
4377 references whatsoever. (See below for a better fix that came later.)
4379 26. The handling of the optimization for finding the first character of a
4380 non-anchored pattern, and for finding a character that is required later in the
4381 match were failing in some cases. This didn't break the matching; it just
4382 failed to optimize when it could. The way this is done has been re-implemented.
4384 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4386 28. Added a new feature that provides some of the functionality that Perl
4387 provides with (?{...}). The facility is termed a "callout". The way it is done
4388 in PCRE is for the caller to provide an optional function, by setting
4389 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4390 global variable. By default it is unset, which disables all calling out. To get
4391 the function called, the regex must include (?C) at appropriate points. This
4392 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4393 This provides a means of identifying different callout points. When PCRE
4394 reaches such a point in the regex, if pcre_callout has been set, the external
4395 function is called. It is provided with data in a structure called
4396 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4397 matching continues; if it returns a non-zero value, the match at the current
4398 point fails. However, backtracking will occur if possible. [This was changed
4399 later and other features added - see item 49 below.]
4401 29. pcretest is upgraded to test the callout functionality. It provides a
4402 callout function that displays information. By default, it shows the start of
4403 the match and the current position in the text. There are some new data escapes
4404 to vary what happens:
4406 \C+ in addition, show current contents of captured substrings
4407 \C- do not supply a callout function
4408 \C!n return 1 when callout number n is reached
4409 \C!n!m return 1 when callout number n is reached for the mth time
4411 30. If pcregrep was called with the -l option and just a single file name, it
4412 output "<stdin>" if a match was found, instead of the file name.
4414 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4415 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4416 pcre_exec(). This saves a malloc/free per call. The default value of
4417 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4418 when configuring.
4420 32. The default maximum size of a compiled pattern is 64K. There have been a
4421 few cases of people hitting this limit. The code now uses macros to handle the
4422 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4423 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4424 configuring. Tests 2 and 5 work only with 2-byte links because they output
4425 debugging information about compiled patterns.
4427 33. Internal code re-arrangements:
4429 (a) Moved the debugging function for printing out a compiled regex into
4430 its own source file (printint.c) and used #include to pull it into
4431 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4432 separate copies.
4434 (b) Defined the list of op-code names for debugging as a macro in
4435 internal.h so that it is next to the definition of the opcodes.
4437 (c) Defined a table of op-code lengths for simpler skipping along compiled
4438 code. This is again a macro in internal.h so that it is next to the
4439 definition of the opcodes.
4441 34. Added support for recursive calls to individual subpatterns, along the
4442 lines of Robin Houston's patch (but implemented somewhat differently).
4444 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4445 allow it to read and process whole directories in Win32. This code was
4446 contributed by Lionel Fourquaux; it has not been tested by me.
4448 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4449 used to name a group. Names consist of alphanumerics and underscores, and must
4450 be unique. Back references use the syntax (?P=name) and recursive calls use
4451 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4452 numbers. The function pcre_fullinfo() can be used after compilation to extract
4453 a name/number map. There are three relevant calls:
4455 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4456 PCRE_INFO_NAMECOUNT yields the number of entries
4457 PCRE_INFO_NAMETABLE yields a pointer to the map.
4459 The map is a vector of fixed-size entries. The size of each entry depends on
4460 the length of the longest name used. The first two bytes of each entry are the
4461 group number, most significant byte first. There follows the corresponding
4462 name, zero terminated. The names are in alphabetical order.
4464 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4465 case instead of 255. Making it the same both with and without UTF-8 support
4466 means that the same test output works with both.
4468 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4469 calling malloc() with a zero argument.
4471 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4472 optimization. I've improved things by keeping a bitmap of backreferences with
4473 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4474 fact referenced, the optimization can be applied. It is unlikely that a
4475 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4476 the match to follow \n) will appear inside brackets with a number greater than
4477 31, but if it does, any back reference > 31 suppresses the optimization.
4479 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4480 of disabling numbered capturing parentheses. Any opening parenthesis that is
4481 not followed by ? behaves as if it were followed by ?: but named parentheses
4482 can still be used for capturing (and they will acquire numbers in the usual
4483 way).
4485 41. Redesigned the return codes from the match() function into yes/no/error so
4486 that errors can be passed back from deep inside the nested calls. A malloc
4487 failure while inside a recursive subpattern call now causes the
4488 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4490 42. It is now possible to set a limit on the number of times the match()
4491 function is called in a call to pcre_exec(). This facility makes it possible to
4492 limit the amount of recursion and backtracking, though not in a directly
4493 obvious way, because the match() function is used in a number of different
4494 circumstances. The count starts from zero for each position in the subject
4495 string (for non-anchored patterns). The default limit is, for compatibility, a
4496 large number, namely 10 000 000. You can change this in two ways:
4498 (a) When configuring PCRE before making, you can use --with-match-limit=n
4499 to set a default value for the compiled library.
4501 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4502 a different value is set. See 45 below.
4504 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4506 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4507 of things that can be changed at compile time. The first argument specifies
4508 what is wanted and the second points to where the information is to be placed.
4509 The current list of available information is:
4513 The output is an integer that is set to one if UTF-8 support is available;
4514 otherwise it is set to zero.
4518 The output is an integer that it set to the value of the code that is used for
4519 newline. It is either LF (10) or CR (13).
4523 The output is an integer that contains the number of bytes used for internal
4524 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4528 The output is an integer that contains the threshold above which the POSIX
4529 interface uses malloc() for output vectors. See item 31 above.
4533 The output is an unsigned integer that contains the default limit of the number
4534 of match() calls in a pcre_exec() execution. See 42 above.
4536 44. pcretest has been upgraded by the addition of the -C option. This causes it
4537 to extract all the available output from the new pcre_config() function, and to
4538 output it. The program then exits immediately.
4540 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4541 order to support additional features. One way would have been to define
4542 pcre_exec2() (for example) with extra arguments, but this would not have been
4543 extensible, and would also have required all calls to the original function to
4544 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4545 is used for passing in "extra" data from pcre_study().
4547 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4548 contains the following fields:
4550 flags a bitmap indicating which of the following fields are set
4551 study_data opaque data from pcre_study()
4552 match_limit a way of specifying a limit on match() calls for a specific
4553 call to pcre_exec()
4554 callout_data data for callouts (see 49 below)
4556 The flag bits are also defined in pcre.h, and are
4562 The pcre_study() function now returns one of these new pcre_extra blocks, with
4563 the actual study data pointed to by the study_data field, and the
4564 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4565 before. That is, this change is entirely upwards-compatible and requires no
4566 change to existing code.
4568 If you want to pass in additional data to pcre_exec(), you can either place it
4569 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4570 block.
4572 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4573 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4574 times with different match limits, until it finds the minimum value needed for
4575 pcre_exec() to complete. The value is then output. This can be instructive; for
4576 most simple matches the number is quite small, but for pathological cases it
4577 gets very large very quickly.
4579 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4580 returns the size of the data block pointed to by the study_data field in a
4581 pcre_extra block, that is, the value that was passed as the argument to
4582 pcre_malloc() when PCRE was getting memory in which to place the information
4583 created by pcre_study(). The fourth argument should point to a size_t variable.
4584 pcretest has been extended so that this information is shown after a successful
4585 pcre_study() call when information about the compiled regex is being displayed.
4587 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4588 because what follows is always an absolute path. (Later: it turns out that this
4589 is more than cosmetic for MinGW, because it doesn't like empty path
4590 components.)
4592 49. Some changes have been made to the callout feature (see 28 above):
4594 (i) A callout function now has three choices for what it returns:
4596 0 => success, carry on matching
4597 > 0 => failure at this point, but backtrack if possible
4598 < 0 => serious error, return this value from pcre_exec()
4600 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4601 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4602 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4603 use by callout functions. It will never be used by PCRE itself.
4605 (ii) The pcre_extra structure (see 45 above) has a void * field called
4606 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4607 pcre_callout_block structure has a field of the same name. The contents of
4608 the field passed in the pcre_extra structure are passed to the callout
4609 function in the corresponding field in the callout block. This makes it
4610 easier to use the same callout-containing regex from multiple threads. For
4611 testing, the pcretest program has a new data escape
4613 \C*n pass the number n (may be negative) as callout_data
4615 If the callout function in pcretest receives a non-zero value as
4616 callout_data, it returns that value.
4618 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4619 there were some redundant $(CFLAGS) in commands that are now specified as
4620 $(LINK), which already includes $(CFLAGS).
4622 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4623 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4624 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4625 one-byte characters throughout. Note that case-insensitive matching applies
4626 only to characters whose values are less than 256. PCRE doesn't support the
4627 notion of cases for higher-valued characters.
4629 (i) A character class whose characters are all within 0-255 is handled as
4630 a bit map, and the map is inverted for negative classes. Previously, a
4631 character > 255 always failed to match such a class; however it should
4632 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4634 (ii) A negated character class with a single character < 255 is coded as
4635 "not this character" (OP_NOT). This wasn't working properly when the test
4636 character was multibyte, either singly or repeated.
4638 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4639 mode, for example: \x{100}{2,3}.
4641 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4642 singly or repeated) now correctly test multibyte characters. However,
4643 PCRE doesn't recognize any characters with values greater than 255 as
4644 digits, spaces, or word characters. Such characters always match \D, \S,
4645 and \W, and never match \d, \s, or \w.
4647 (v) Classes may now contain characters and character ranges with values
4648 greater than 255. For example: [ab\x{100}-\x{400}].
4650 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4651 PCRE in UTF-8 mode.
4653 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4654 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4655 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4656 value.)
4658 53. The single man page has become too large. I have therefore split it up into
4659 a number of separate man pages. These also give rise to individual HTML pages;
4660 these are now put in a separate directory, and there is an index.html page that
4661 lists them all. Some hyperlinking between the pages has been installed.
4663 54. Added convenience functions for handling named capturing parentheses.
4665 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4666 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4667 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4668 are faulted.
4670 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4671 calling configure. These values are used when compiling the dftables.c program
4672 which is run to generate the source of the default character tables. They
4673 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4674 you will need to set these values.
4676 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4679 Version 3.9 02-Jan-02
4680 ---------------------
4682 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4684 2. If --disable-static was given, the building process failed when trying to
4685 build pcretest and pcregrep. (For some reason it was using libtool to compile
4686 them, which is not right, as they aren't part of the library.)
4689 Version 3.8 18-Dec-01
4690 ---------------------
4692 1. The experimental UTF-8 code was completely screwed up. It was packing the
4693 bytes in the wrong order. How dumb can you get?
4696 Version 3.7 29-Oct-01
4697 ---------------------
4699 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4700 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4701 this didn't happen under Solaris 8, where I normally test things.
4703 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4704 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4705 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4706 this fix an hour or so after the initial 3.7 release.)
4709 Version 3.6 23-Oct-01
4710 ---------------------
4712 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4713 offsets passed as NULL with zero offset count.
4715 2. The config.guess and config.sub files had not been updated when I moved to
4716 the latest autoconf.
4719 Version 3.5 15-Aug-01
4720 ---------------------
4722 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4723 had been forgotten.
4725 2. By using declared but undefined structures, we can avoid using "void"
4726 definitions in pcre.h while keeping the internal definitions of the structures
4727 private.
4729 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4730 user point of view, this means that both static and shared libraries are built
4731 by default, but this can be individually controlled. More of the work of
4732 handling this static/shared cases is now inside libtool instead of PCRE's make
4733 file.
4735 4. The pcretest utility is now installed along with pcregrep because it is
4736 useful for users (to test regexs) and by doing this, it automatically gets
4737 relinked by libtool. The documentation has been turned into a man page, so
4738 there are now .1, .txt, and .html versions in /doc.
4740 5. Upgrades to pcregrep:
4741 (i) Added long-form option names like gnu grep.
4742 (ii) Added --help to list all options with an explanatory phrase.
4743 (iii) Added -r, --recursive to recurse into sub-directories.
4744 (iv) Added -f, --file to read patterns from a file.
4746 6. pcre_exec() was referring to its "code" argument before testing that
4747 argument for NULL (and giving an error if it was NULL).
4749 7. Upgraded Makefile.in to allow for compiling in a different directory from
4750 the source directory.
4752 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4753 options bits, the pointer it was passed was to an int instead of to an unsigned
4754 long int. This mattered only on 64-bit systems.
4756 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4757 generated) instead of pcre.in, which it its source. Also made the same change
4758 in several of the .c files.
4760 10. A new release of gcc defines printf() as a macro, which broke pcretest
4761 because it had an ifdef in the middle of a string argument for printf(). Fixed
4762 by using separate calls to printf().
4764 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4765 script, to force use of CR or LF instead of \n in the source. On non-Unix
4766 systems, the value can be set in config.h.
4768 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4769 absolute limit. Changed the text of the error message to make this clear, and
4770 likewise updated the man page.
4772 13. The limit of 99 on the number of capturing subpatterns has been removed.
4773 The new limit is 65535, which I hope will not be a "real" limit.
4776 Version 3.4 22-Aug-00
4777 ---------------------
4779 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4781 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4784 Version 3.3 01-Aug-00
4785 ---------------------
4787 1. If an octal character was given, but the value was greater than \377, it
4788 was not getting masked to the least significant bits, as documented. This could
4789 lead to crashes in some systems.
4791 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4792 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4794 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4795 These just pass their arguments on to (pcre_free)(), but they are provided
4796 because some uses of PCRE bind it to non-C systems that can call its functions,
4797 but cannot call free() or pcre_free() directly.
4799 4. Add "make test" as a synonym for "make check". Corrected some comments in
4800 the Makefile.
4802 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4803 Makefile.
4805 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4806 command called pgrep for grepping around the active processes.
4808 7. Added the beginnings of support for UTF-8 character strings.
4810 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4811 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4812 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4813 out for the ar command.)
4816 Version 3.2 12-May-00
4817 ---------------------
4819 This is purely a bug fixing release.
4821 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4822 of ZA. This was just one example of several cases that could provoke this bug,
4823 which was introduced by change 9 of version 2.00. The code for breaking
4824 infinite loops after an iteration that matches an empty string was't working
4825 correctly.
4827 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4828 when matched against abbab (for example). After matching an empty string, it
4829 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4830 caused it to match further down the string than it should.
4832 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4833 was there because it doesn't seem to be needed, and it causes trouble on some
4834 systems, as it is not a Standard C header. It has been removed.
4836 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4837 were reported on the Macintosh. The changes were from
4839 while ((c = *(++ptr)) != 0 && c != '\n');
4840 to
4841 while ((c = *(++ptr)) != 0 && c != '\n') ;
4843 Totally extraordinary, but if that's what it takes...
4845 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4846 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4847 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4848 assumes the way PCRE uses memmove() (always moving upwards).
4850 6. PCRE is being used in one environment where strchr() is not available. There
4851 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4852 faster code anyway.
4855 Version 3.1 09-Feb-00
4856 ---------------------
4858 The only change in this release is the fixing of some bugs in Makefile.in for
4859 the "install" target:
4861 (1) It was failing to install pcreposix.h.
4863 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4866 Version 3.0 01-Feb-00
4867 ---------------------
4869 1. Add support for the /+ modifier to perltest (to output $` like it does in
4870 pcretest).
4872 2. Add support for the /g modifier to perltest.
4874 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4875 matches null strings.
4877 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4878 pattern. Perl treats empty patterns specially - it reuses the most recent
4879 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4880 effect.
4882 5. The POSIX interface was broken in that it was just handing over the POSIX
4883 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4884 required a bigger vector, with some working space on the end. This means that
4885 the POSIX wrapper now has to get and free some memory, and copy the results.
4887 6. Added some simple autoconf support, placing the test data and the
4888 documentation in separate directories, re-organizing some of the
4889 information files, and making it build pcre-config (a GNU standard). Also added
4890 libtool support for building PCRE as a shared library, which is now the
4891 default.
4893 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4894 09 are not valid octal constants. Single digits will be used for minor values
4895 less than 10.
4897 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4898 existing programs that set these in the POSIX interface can use PCRE without
4899 modification.
4901 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4902 return all that pcre_info() returns, plus additional data. The pcre_info()
4903 function is retained for compatibility, but is considered to be obsolete.
4905 10. Added experimental recursion feature (?R) to handle one common case that
4906 Perl 5.6 will be able to do with (?p{...}).
4908 11. Added support for POSIX character classes like [:alpha:], which Perl is
4909 adopting.
4912 Version 2.08 31-Aug-99
4913 ----------------------
4915 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4916 trying to match at the startoffset position, but instead was moving forward to
4917 the next newline as if a previous match had failed.
4919 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4920 and could get into a loop if a null string was matched other than at the start
4921 of the subject.
4923 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4924 be distinguished at compile time, and for completeness also added PCRE_DATE.
4926 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4927 in GnuWin32 environments.
4930 Version 2.07 29-Jul-99
4931 ----------------------
4933 1. The documentation is now supplied in plain text form and HTML as well as in
4934 the form of man page sources.
4936 2. C++ compilers don't like assigning (void *) values to other pointer types.
4937 In particular this affects malloc(). Although there is no problem in Standard
4938 C, I've put in casts to keep C++ compilers happy.
4940 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4941 should be (const char *).
4943 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4944 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4945 However, I haven't made this a standard facility. The documentation doesn't
4946 mention it, and the Makefile doesn't support it.
4948 5. The Makefile now contains an "install" target, with editable destinations at
4949 the top of the file. The pcretest program is not installed.
4951 6. pgrep -V now gives the PCRE version number and date.
4953 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4954 causing the entire string to be ignored, instead of just the last character.
4956 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4957 non-matching string, it can take a very, very long time, even for strings of
4958 quite modest length, because of the nested recursion. PCRE now does better in
4959 some of these cases. It does this by remembering the last required literal
4960 character in the pattern, and pre-searching the subject to ensure it is present
4961 before running the real match. In other words, it applies a heuristic to detect
4962 some types of certain failure quickly, and in the above example, if presented
4963 with a string that has no trailing " it gives "no match" very quickly.
4965 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4966 other alternatives are tried instead.
4969 Version 2.06 09-Jun-99
4970 ----------------------
4972 1. Change pcretest's output for amount of store used to show just the code
4973 space, because the remainder (the data block) varies in size between 32-bit and
4974 64-bit systems.
4976 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4977 start matching at. This allows lookbehinds to work when searching for multiple
4978 occurrences in a string.
4980 3. Added additional options to pcretest for testing multiple occurrences:
4982 /+ outputs the rest of the string that follows a match
4983 /g loops for multiple occurrences, using the new startoffset argument
4984 /G loops for multiple occurrences by passing an incremented pointer
4986 4. PCRE wasn't doing the "first character" optimization for patterns starting
4987 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4988 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4989 the letter 'x'. On long subject strings, this gives a significant speed-up.
4992 Version 2.05 21-Apr-99
4993 ----------------------
4995 1. Changed the type of magic_number from int to long int so that it works
4996 properly on 16-bit systems.
4998 2. Fixed a bug which caused patterns starting with .* not to work correctly
4999 when the subject string contained newline characters. PCRE was assuming
5000 anchoring for such patterns in all cases, which is not correct because .* will
5001 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5002 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5003 must be retried after every newline in the subject.
5006 Version 2.04 18-Feb-99
5007 ----------------------
5009 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5010 computation of the store needed to hold the pattern was incorrect (too large).
5011 If such patterns were nested a few deep, this could multiply and become a real
5012 problem.
5014 2. Added /M option to pcretest to show the memory requirement of a specific
5015 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5017 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5018 compiled in such a way that the backtracking after subsequent failure was
5019 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5020 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5023 Version 2.03 02-Feb-99
5024 ----------------------
5026 1. Fixed typo and small mistake in man page.
5028 2. Added 4th condition (GPL supersedes if conflict) and created separate
5029 LICENCE file containing the conditions.
5031 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5032 Perl, that is the internal \ allows the delimiter to be included in the
5033 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5034 the final delimiter, add \ to the end of the pattern (to test the error).
5036 4. Added the convenience functions for extracting substrings after a successful
5037 match. Updated pcretest to make it able to test these functions.
5040 Version 2.02 14-Jan-99
5041 ----------------------
5043 1. Initialized the working variables associated with each extraction so that
5044 their saving and restoring doesn't refer to uninitialized store.
5046 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5047 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5048 fix the problem.
5050 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5051 calls, and wasn't printing the correct value for compiling calls. Increased the
5052 default value of LOOPREPEAT, and the number of significant figures in the
5053 times.
5055 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5057 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5058 a building problem on Windows NT with a FAT file system.
5061 Version 2.01 21-Oct-98
5062 ----------------------
5064 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5065 to character tables built by pcre_maketables() in the current locale. If NULL
5066 is passed, the default tables are used.
5069 Version 2.00 24-Sep-98
5070 ----------------------
5072 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5073 it any more.
5075 2. Allow quantification of (?>) groups, and make it work correctly.
5077 3. The first character computation wasn't working for (?>) groups.
5079 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5080 end of the subject) and add 5.005's \z, which really does match only at the
5081 very end of the subject.
5083 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5085 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5086 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5087 localized options. All options to pcre_study() were also removed.
5089 7. Add other new features from 5.005:
5091 $(?<= positive lookbehind
5092 $(?<! negative lookbehind
5093 (?imsx-imsx) added the unsetting capability
5094 such a setting is global if at outer level; local otherwise
5095 (?imsx-imsx:) non-capturing groups with option setting
5096 (?(cond)re|re) conditional pattern matching
5098 A backreference to itself in a repeated group matches the previous
5099 captured string.
5101 8. General tidying up of studying (both automatic and via "study")
5102 consequential on the addition of new assertions.
5104 9. As in 5.005, unlimited repeated groups that could match an empty substring
5105 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5106 runtime if any iteration does actually match an empty substring.
5108 10. Include the RunTest script in the distribution.
5110 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5111 discrepancies, some of which were old and were also with respect to 5.004. They
5112 have now been fixed.
5115 Version 1.09 28-Apr-98
5116 ----------------------
5118 1. A negated single character class followed by a quantifier with a minimum
5119 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5120 program crashes, or just wrong answers. This did not apply to negated classes
5121 containing more than one character, or to minima other than one.
5124 Version 1.08 27-Mar-98
5125 ----------------------
5127 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5129 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5130 latter must appear before anything that relies on it in the pattern.
5133 Version 1.07 16-Feb-98
5134 ----------------------
5136 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5137 repeat of a potentially empty string).
5140 Version 1.06 23-Jan-98
5141 ----------------------
5143 1. Added Markus Oberhumer's little patches for C++.
5145 2. Literal strings longer than 255 characters were broken.
5148 Version 1.05 23-Dec-97
5149 ----------------------
5151 1. Negated character classes containing more than one character were failing if
5152 PCRE_CASELESS was set at run time.
5155 Version 1.04 19-Dec-97
5156 ----------------------
5158 1. Corrected the man page, where some "const" qualifiers had been omitted.
5160 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5161 input syntax.
5163 3. Fixed memory leak which occurred when a regex with back references was
5164 matched with an offsets vector that wasn't big enough. The temporary memory
5165 that is used in this case wasn't being freed if the match failed.
5167 4. Tidied pcretest to ensure it frees memory that it gets.
5169 5. Temporary memory was being obtained in the case where the passed offsets
5170 vector was exactly big enough.
5172 6. Corrected definition of offsetof() from change 5 below.
5174 7. I had screwed up change 6 below and broken the rules for the use of
5175 setjmp(). Now fixed.
5178 Version 1.03 18-Dec-97
5179 ----------------------
5181 1. A erroneous regex with a missing opening parenthesis was correctly
5182 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5183 on some systems.
5185 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5186 it was reported that one broken compiler failed on the former because "code" is
5187 also an independent variable.
5189 3. The erroneous regex a[]b caused an array overrun reference.
5191 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5192 fail on data ending with that character. (It was going on too far, and checking
5193 the next character, typically a binary zero.) This was specific to the
5194 optimized code for single-character negative classes.
5196 5. Added a contributed patch from the TIN world which does the following:
5198 + Add an undef for memmove, in case the the system defines a macro for it.
5200 + Add a definition of offsetof(), in case there isn't one. (I don't know
5201 the reason behind this - offsetof() is part of the ANSI standard - but
5202 it does no harm).
5204 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5205 most of the places where whitespace preceded '#'. I have given up and
5206 allowed the remaining 2 cases to be at the margin.
5208 + Rename some variables in pcre to eliminate shadowing. This seems very
5209 pedantic, but does no harm, of course.
5211 6. Moved the call to setjmp() into its own function, to get rid of warnings
5212 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5214 7. Constructs such as \d{8,} were compiling into the equivalent of
5215 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5216 outcome, but in this particular case used more store than had been allocated,
5217 which caused the bug to be discovered because it threw up an internal error.
5219 8. The debugging code in both pcre and pcretest for outputting the compiled
5220 form of a regex was going wrong in the case of back references followed by
5221 curly-bracketed repeats.
5224 Version 1.02 12-Dec-97
5225 ----------------------
5227 1. Typos in pcre.3 and comments in the source fixed.
5229 2. Applied a contributed patch to get rid of places where it used to remove
5230 'const' from variables, and fixed some signed/unsigned and uninitialized
5231 variable warnings.
5233 3. Added the "runtest" target to Makefile.
5235 4. Set default compiler flag to -O2 rather than just -O.
5238 Version 1.01 19-Nov-97
5239 ----------------------
5241 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5242 like /([ab]*)*/, that is, for classes with more than one character in them.
5244 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5245 as /((?>a*))*/ (a PCRE_EXTRA facility).
5248 Version 1.00 18-Nov-97
5249 ----------------------
5251 1. Added compile-time macros to support systems such as SunOS4 which don't have
5252 memmove() or strerror() but have other things that can be used instead.
5254 2. Arranged that "make clean" removes the executables.
5257 Version 0.99 27-Oct-97
5258 ----------------------
5260 1. Fixed bug in code for optimizing classes with only one character. It was
5261 initializing a 32-byte map regardless, which could cause it to run off the end
5262 of the memory it had got.
5264 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5267 Version 0.98 22-Oct-97
5268 ----------------------
5270 1. Fixed bug in code for handling temporary memory usage when there are more
5271 back references than supplied space in the ovector. This could cause segfaults.
5274 Version 0.97 21-Oct-97
5275 ----------------------
5277 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5279 2. Optimized negated single characters not to use a bit map.
5281 3. Brought error texts together as macro definitions; clarified some of them;
5282 fixed one that was wrong - it said "range out of order" when it meant "invalid
5283 escape sequence".
5285 4. Changed some char * arguments to const char *.
5287 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5289 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5290 pcretest.
5293 Version 0.96 16-Oct-97
5294 ----------------------
5296 1. Added a simple "pgrep" utility to the distribution.
5298 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5299 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5300 where "ddd" means "one or more decimal digits".
5302 3. Fixed serious bug. If a pattern had a back reference, but the call to
5303 pcre_exec() didn't supply a large enough ovector to record the related
5304 identifying subpattern, the match always failed. PCRE now remembers the number
5305 of the largest back reference, and gets some temporary memory in which to save
5306 the offsets during matching if necessary, in order to ensure that
5307 backreferences always work.
5309 4. Increased the compatibility with Perl in a number of ways:
5311 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5312 to request this handling. The option can be set at compile or exec time.
5314 (b) $ matches before a terminating newline by default; an option
5315 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5316 mode). The option can be set at compile or exec time.
5318 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5319 the same as Perl's. If the decimal number it represents is less than 10
5320 or there aren't that many previous left capturing parentheses, an octal
5321 escape is read. Inside a character class, it's always an octal escape,
5322 even if it is a single digit.
5324 (d) An escaped but undefined alphabetic character is taken as a literal,
5325 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5326 escapes.
5328 (e) {0} is now permitted. (The previous item is removed from the compiled
5329 pattern).
5331 5. Changed all the names of code files so that the basic parts are no longer
5332 than 10 characters, and abolished the teeny "globals.c" file.
5334 6. Changed the handling of character classes; they are now done with a 32-byte
5335 bit map always.
5337 7. Added the -d and /D options to pcretest to make it possible to look at the
5338 internals of compilation without having to recompile pcre.
5341 Version 0.95 23-Sep-97
5342 ----------------------
5344 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5345 \x20 at the start of a run of normal characters. These were being treated as
5346 real characters, instead of the source characters being re-checked.
5349 Version 0.94 18-Sep-97
5350 ----------------------
5352 1. The functions are now thread-safe, with the caveat that the global variables
5353 containing pointers to malloc() and free() or alternative functions are the
5354 same for all threads.
5356 2. Get pcre_study() to generate a bitmap of initial characters for non-
5357 anchored patterns when this is possible, and use it if passed to pcre_exec().
5360 Version 0.93 15-Sep-97
5361 ----------------------
5363 1. /(b)|(:+)/ was computing an incorrect first character.
5365 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5366 but not actually doing anything yet.
5368 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5369 as Perl does (e.g. [-az] or [az-]).
5371 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5372 all possible positions.
5374 5. Split up into different modules to avoid including unneeded functions in a
5375 compiled binary. However, compile and exec are still in one module. The "study"
5376 function is split off.
5378 6. The character tables are now in a separate module whose source is generated
5379 by an auxiliary program - but can then be edited by hand if required. There are
5380 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5381 toupper() in the code.
5383 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5384 make them global. Abolish the function for setting them, as the caller can now
5385 set them directly.
5388 Version 0.92 11-Sep-97
5389 ----------------------
5391 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5392 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5394 2. Caseless matching was not working in character classes if the characters in
5395 the pattern were in upper case.
5397 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5399 4. Make PCRE_ANCHORED public and accept as a compile option.
5401 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5402 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5403 pass them.
5405 6. Give an error if bad option bits passed at compile or run time.
5407 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5408 pcretest to cause it to pass that flag.
5410 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5411 options, and the first character, if set.
5413 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5416 Version 0.91 10-Sep-97
5417 ----------------------
5419 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5420 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5422 2. PCRE was looping on encountering an indefinitely repeated back reference to
5423 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5424 Perl does - treats the match as successful.
5426 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5