ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1404 - (show annotations)
Tue Nov 19 15:36:57 2013 UTC (7 years, 2 months ago) by ph10
File size: 256904 byte(s)
Source tidies for 8.34-RC1.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.34 19-November-2013
5 -----------------------------
7 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8 executable memory. Patch inspired by Carsten Klein.
10 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11 this macro is never tested and has no effect, because the work to support
12 coverage involves only compiling and linking options and special targets in
13 the Makefile. The comment in config.h implied that defining the macro would
14 enable coverage support, which is totally false. There was also support for
15 setting this macro in the CMake files (my fault, I just copied it from
16 configure). SUPPORT_GCOV has now been removed.
18 3. Make a small performance improvement in strlen16() and strlen32() in
19 pcretest.
21 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23 these cases). There was also one in pcretest.c.
25 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
27 6. In UTF mode, the code for checking whether a group could match an empty
28 string (which is used for indefinitely repeated groups to allow for
29 breaking an infinite loop) was broken when the group contained a repeated
30 negated single-character class with a character that occupied more than one
31 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32 UTF-8 mode). The effect was undefined: the group might or might not be
33 deemed as matching an empty string, or the program might have crashed.
35 7. The code for checking whether a group could match an empty string was not
36 recognizing that \h, \H, \v, \V, and \R must match a character.
38 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39 an empty string. If it can, pcretest shows this in its information output.
41 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
42 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43 matched by pcre_exec() without using JIT:
45 (a) If the rest of the pattern did not match after a maximal run of
46 grapheme clusters, the code for backing up to try with fewer of them
47 did not always back up over a full grapheme when characters that do not
48 have the modifier quality were involved, e.g. Hangul syllables.
50 (b) If the match point in a subject started with modifier character, and
51 there was no match, the code could incorrectly back up beyond the match
52 point, and potentially beyond the first character in the subject,
53 leading to a segfault or an incorrect match result.
55 10. A conditional group with an assertion condition could lead to PCRE
56 recording an incorrect first data item for a match if no other first data
57 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58 first data item, and therefore matched "ca" after "c" instead of at the
59 start.
61 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
64 12. The source of pcregrep now includes z/OS-specific code so that it can be
65 compiled for z/OS as part of the special z/OS distribution.
67 13. Added the -T and -TM options to pcretest.
69 14. The code in pcre_compile.c for creating the table of named capturing groups
70 has been refactored. Instead of creating the table dynamically during the
71 actual compiling pass, the information is remembered during the pre-compile
72 pass (on the stack unless there are more than 20 named groups, in which
73 case malloc() is used) and the whole table is created before the actual
74 compile happens. This has simplified the code (it is now nearly 150 lines
75 shorter) and prepared the way for better handling of references to groups
76 with duplicate names.
78 15. A back reference to a named subpattern when there is more than one of the
79 same name now checks them in the order in which they appear in the pattern.
80 The first one that is set is used for the reference. Previously only the
81 first one was inspected. This change makes PCRE more compatible with Perl.
83 16. Unicode character properties were updated from Unicode 6.3.0.
85 17. The compile-time code for auto-possessification has been refactored, based
86 on a patch by Zoltan Herczeg. It now happens after instead of during
87 compilation. The code is cleaner, and more cases are handled. The option
88 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
89 options in pcretest are provided to set it. It can also be set by
90 (*NO_AUTO_POSSESS) at the start of a pattern.
92 18. The character VT has been added to the set of characters that match \s and
93 are generally treated as white space, following this same change in Perl
94 5.18. There is now no difference between "Perl space" and "POSIX space".
96 19. The code for checking named groups as conditions, either for being set or
97 for being recursed, has been refactored (this is related to 14 and 15
98 above). Processing unduplicated named groups should now be as fast at
99 numerical groups, and processing duplicated groups should be faster than
100 before.
102 20. Two patches to the CMake build system, by Alexander Barkov:
104 (1) Replace the "source" command by "." in CMakeLists.txt because
105 "source" is a bash-ism.
107 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
108 without these the CMake build does not work on Solaris.
110 21. Perl has changed its handling of \8 and \9. If there is no previously
111 encountered capturing group of those numbers, they are treated as the
112 literal characters 8 and 9 instead of a binary zero followed by the
113 literals. PCRE now does the same.
115 22. Following Perl, added \o{} to specify codepoints in octal, making it
116 possible to specify values greater than 0777 and also making them
117 unambiguous.
119 23. Perl now gives an error for missing closing braces after \x{... instead of
120 treating the string as literal. PCRE now does the same.
122 24. RunTest used to grumble if an inappropriate test was selected explicitly,
123 but just skip it when running all tests. This make it awkward to run ranges
124 of tests when one of them was inappropriate. Now it just skips any
125 inappropriate tests, as it always did when running all tests.
127 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
128 character types such as \d or \w, too many callouts were inserted, and the
129 data that they returned was rubbish.
131 26. In UCP mode, \s was not matching two of the characters that Perl matches,
132 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
133 were matched by \h. The code has now been refactored so that the lists of
134 the horizontal and vertical whitespace characters used for \h and \v (which
135 are defined only in one place) are now also used for \s.
137 27. Add JIT support for the 64 bit TileGX architecture.
138 Patch by Jiong Wang (Tilera Corporation).
140 28. Possessive quantifiers for classes (both explicit and automatically
141 generated) now use special opcodes instead of wrapping in ONCE brackets.
143 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
144 (because it's meaningless), this was not happening when PCRE_CASELESS was
145 set. Not wrong, but inefficient.
147 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
148 properties for \w, \d, etc) is present in a test regex. Otherwise if the
149 test contains no characters greater than 255, Perl doesn't realise it
150 should be using Unicode semantics.
152 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
153 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
154 does in Unicode mode.
156 32. Added the "forbid" facility to pcretest so that putting tests into the
157 wrong test files can sometimes be quickly detected.
159 33. There is now a limit (default 250) on the depth of nesting of parentheses.
160 This limit is imposed to control the amount of system stack used at compile
161 time. It can be changed at build time by --with-parens-nest-limit=xxx or
162 the equivalent in CMake.
164 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
165 errors. Perl warns for these when in warning mode, but PCRE has no facility
166 for giving warnings.
168 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
169 However, this was not working for (?!) because it is optimized to (*FAIL),
170 for which PCRE does not allow quantifiers. The optimization is now disabled
171 when a quantifier follows (?!). I can't see any use for this, but it makes
172 things uniform.
174 36. Perl no longer allows group names to start with digits, so I have made this
175 change also in PCRE. It simplifies the code a bit.
177 37. In extended mode, Perl ignores spaces before a + that indicates a
178 possessive quantifier. PCRE allowed a space before the quantifier, but not
179 before the possessive +. It now does.
181 38. The use of \K (reset reported match start) within a repeated possessive
182 group such as (a\Kb)*+ was not working.
184 40. Document that the same character tables must be used at compile time and
185 run time, and that the facility to pass tables to pcre_exec() and
186 pcre_dfa_exec() is for use only with saved/restored patterns.
189 Version 8.33 28-May-2013
190 ------------------------
192 1. Added 'U' to some constants that are compared to unsigned integers, to
193 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
194 variables that are added to signed variables, to ensure the result is
195 signed and can be negated.
197 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
198 CMake config files.
200 3. Revise the creation of config.h.generic so that all boolean macros are
201 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
202 overriding via -D on the command line possible.
204 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
205 to unsigned int is reported to make a quite noticeable speed difference in
206 a specific Windows environment. Testing on Linux did also appear to show
207 some benefit (and it is clearly not harmful). Also fixed the definition of
208 Xop which should be unsigned.
210 5. Related to (4), changing the definition of the intermediate variable cc
211 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
212 improvements.
214 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
215 unnecessary spaces.
217 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
218 and later.
220 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
222 9. Optimizing fast_forward_start_bits in JIT.
224 10. Adding support for callouts in JIT, and fixing some issues revealed
225 during this work. Namely:
227 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
229 (b) Minimum length was not checked before the matching is started.
231 11. The value of capture_last that is passed to callouts was incorrect in some
232 cases when there was a capture on one path that was subsequently abandoned
233 after a backtrack. Also, the capture_last value is now reset after a
234 recursion, since all captures are also reset in this case.
236 12. The interpreter no longer returns the "too many substrings" error in the
237 case when an overflowing capture is in a branch that is subsequently
238 abandoned after a backtrack.
240 13. In the pathological case when an offset vector of size 2 is used, pcretest
241 now prints out the matched string after a yield of 0 or 1.
243 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
244 Only supported by the JIT compiler at the moment.
246 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
248 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
249 offset of the starting point of the matching process, provided the offsets
250 vector is large enough.
252 17. The \A escape now records a lookbehind value of 1, though its execution
253 does not actually inspect the previous character. This is to ensure that,
254 in partial multi-segment matching, at least one character from the old
255 segment is retained when a new segment is processed. Otherwise, if there
256 are no lookbehinds in the pattern, \A might match incorrectly at the start
257 of a new segment.
259 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
261 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
262 gives some modest performance improvement in 8-bit mode.
264 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
265 be expressed in certain programming languages using Universal Character
266 Names.
268 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
269 which points out that "non characters" are not "characters that may not
270 appear in Unicode strings" but rather "characters that are reserved for
271 internal use and have only local meaning".
273 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
274 there was a conditional group that depended on an assertion, if the
275 assertion was false, the callout that immediately followed the alternation
276 in the condition was skipped when pcre_exec() was used for matching.
278 23. Allow an explicit callout to be inserted before an assertion that is the
279 condition for a conditional group, for compatibility with automatic
280 callouts, which always insert a callout at this point.
282 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
283 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
285 25. (*PRUNE) is now supported by the JIT compiler.
287 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
289 27. Fix the case where there are two or more SKIPs with arguments that may be
290 ignored.
292 28. (*SKIP) is now supported by the JIT compiler.
294 29. (*THEN) is now supported by the JIT compiler.
296 30. Update RunTest with additional test selector options.
298 31. The way PCRE handles backtracking verbs has been changed in two ways.
300 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
301 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
302 backtracking. In some cases this makes it more Perl-compatible, but Perl's
303 rather obscure rules do not always do the same thing.
305 (2) Previously, backtracking verbs were confined within assertions. This is
306 no longer the case for positive assertions, except for (*ACCEPT). Again,
307 this sometimes improves Perl compatibility, and sometimes does not.
309 32. A number of tests that were in test 2 because Perl did things differently
310 have been moved to test 1, because either Perl or PCRE has changed, and
311 these tests are now compatible.
313 32. Backtracking control verbs are now handled in the same way in JIT and
314 interpreter.
316 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
317 contained a forward subroutine reference caused a compile error.
319 34. Auto-detect and optimize limited repetitions in JIT.
321 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
322 blocking (*UTF) etc.
324 36. In the interpreter, maximizing pattern repetitions for characters and
325 character types now use tail recursion, which reduces stack usage.
327 37. The value of the max lookbehind was not correctly preserved if a compiled
328 and saved regex was reloaded on a host of different endianness.
330 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
331 of the compiled pattern block, expand the flags field from 16 to 32 bits
332 because it was almost full.
334 39. Try madvise first before posix_madvise.
336 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
337 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
338 original change locked out all matches of empty strings. This has been
339 changed so that one match of an empty string per line is recognized.
340 Subsequent searches on the same line (for colouring or for --only-matching,
341 for example) do not recognize empty strings.
343 41. Applied a user patch to fix a number of spelling mistakes in comments.
345 42. Data lines longer than 65536 caused pcretest to crash.
347 43. Clarified the data type for length and startoffset arguments for pcre_exec
348 and pcre_dfa_exec in the function-specific man pages, where they were
349 explicitly stated to be in bytes, never having been updated. I also added
350 some clarification to the pcreapi man page.
352 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
353 a segmentation fault.
356 Version 8.32 30-November-2012
357 -----------------------------
359 1. Improved JIT compiler optimizations for first character search and single
360 character iterators.
362 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
363 Patch by Daniel Richard G.
365 3. Single character iterator optimizations in the JIT compiler.
367 4. Improved JIT compiler optimizations for character ranges.
369 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
370 Reported by Giuseppe D'Angelo.
372 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
373 of a line, was being set incorrectly in cases where .* appeared inside
374 atomic brackets at the start of a pattern, or where there was a subsequent
375 *PRUNE or *SKIP.
377 7. Improved instruction cache flush for POWER/PowerPC.
378 Patch by Daniel Richard G.
380 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
381 grep:
383 (a) There is now no limit to the number of patterns to be matched.
385 (b) An error is given if a pattern is too long.
387 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
388 are now supported.
390 (d) --exclude-from and --include-from (multiple use) have been added.
392 (e) Exclusions and inclusions now apply to all files and directories, not
393 just to those obtained from scanning a directory recursively.
395 (f) Multiple uses of -f and --file-list are now supported.
397 (g) In a Windows environment, the default for -d has been changed from
398 "read" (the GNU grep default) to "skip", because otherwise the presence
399 of a directory in the file list provokes an error.
401 (h) The documentation has been revised and clarified in places.
403 9. Improve the matching speed of capturing brackets.
405 10. Changed the meaning of \X so that it now matches a Unicode extended
406 grapheme cluster.
408 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
409 out POSIX threads when JIT support is configured.
411 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
413 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
414 structure could go wrong in environments where size_t is not the same size
415 as int.
417 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
418 set.
420 15. The EBCDIC support had decayed; later updates to the code had included
421 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
422 general tidy up of EBCDIC-related issues, and the documentation was also
423 not quite right. There is now a test that can be run on ASCII systems to
424 check some of the EBCDIC-related things (but is it not a full test).
426 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
427 in a small tidy to the code.
429 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
431 18. If the --only-matching (-o) option in pcregrep is specified multiple
432 times, each one causes appropriate output. For example, -o1 -o2 outputs the
433 substrings matched by the 1st and 2nd capturing parentheses. A separating
434 string can be specified by --om-separator (default empty).
436 19. Improving the first n character searches.
438 20. Turn case lists for horizontal and vertical white space into macros so that
439 they are defined only once.
441 21. This set of changes together give more compatible Unicode case-folding
442 behaviour for characters that have more than one other case when UCP
443 support is available.
445 (a) The Unicode property table now has offsets into a new table of sets of
446 three or more characters that are case-equivalent. The MultiStage2.py
447 script that generates these tables (the pcre_ucd.c file) now scans
448 CaseFolding.txt instead of UnicodeData.txt for character case
449 information.
451 (b) The code for adding characters or ranges of characters to a character
452 class has been abstracted into a generalized function that also handles
453 case-independence. In UTF-mode with UCP support, this uses the new data
454 to handle characters with more than one other case.
456 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
457 whose other case is greater than 256 are now correctly matched
458 caselessly. Previously, the high codepoint matched the low one, but not
459 vice versa.
461 (d) The processing of \h, \H, \v, and \ in character classes now makes use
462 of the new class addition function, using character lists defined as
463 macros alongside the case definitions of 20 above.
465 (e) Caseless back references now work with characters that have more than
466 one other case.
468 (f) General caseless matching of characters with more than one other case
469 is supported.
471 22. Unicode character properties were updated from Unicode 6.2.0
473 23. Improved CMake support under Windows. Patch by Daniel Richard G.
475 24. Add support for 32-bit character strings, and UTF-32
477 25. Major JIT compiler update (code refactoring and bugfixing).
478 Experimental Sparc 32 support is added.
480 26. Applied a modified version of Daniel Richard G's patch to create
481 pcre.h.generic and config.h.generic by "make" instead of in the
482 PrepareRelease script.
484 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
485 pcre_compile.c when checking for a zero character.
487 28. Introducing a native interface for JIT. Through this interface, the compiled
488 machine code can be directly executed. The purpose of this interface is to
489 provide fast pattern matching, so several sanity checks are not performed.
490 However, feature tests are still performed. The new interface provides
491 1.4x speedup compared to the old one.
493 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
494 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
495 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
497 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
498 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
499 the "old" RFC 2279). Instead, it ended up passing a negative length to
500 pcre_exec().
502 31. Add support for GCC's visibility feature to hide internal functions.
504 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
505 "unknown -C option" after outputting 0 or 1.
507 33. There is now support for generating a code coverage report for the test
508 suite in environments where gcc is the compiler and lcov is installed. This
509 is mainly for the benefit of the developers.
511 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
512 unaddressable using valgrind annotations, allowing valgrind to detect
513 invalid memory accesses. This is mainly for the benefit of the developers.
515 25. (*UTF) can now be used to start a pattern in any of the three libraries.
517 26. Give configure error if --enable-cpp but no C++ compiler found.
520 Version 8.31 06-July-2012
521 -------------------------
523 1. Fixing a wrong JIT test case and some compiler warnings.
525 2. Removed a bashism from the RunTest script.
527 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
528 to unsigned type, result still unsigned" that was given by an MS compiler
529 on encountering the code "-sizeof(xxx)".
531 4. Partial matching support is added to the JIT compiler.
533 5. Fixed several bugs concerned with partial matching of items that consist
534 of more than one character:
536 (a) /^(..)\1/ did not partially match "aba" because checking references was
537 done on an "all or nothing" basis. This also applied to repeated
538 references.
540 (b) \R did not give a hard partial match if \r was found at the end of the
541 subject.
543 (c) \X did not give a hard partial match after matching one or more
544 characters at the end of the subject.
546 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
547 a partial match for the string "\r".
549 (e) When newline was set to CRLF, the metacharacter "." did not recognize
550 a partial match for a CR character at the end of the subject string.
552 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
553 running pcretest, the text "(JIT)" added to the output whenever JIT is
554 actually used to run the match.
556 7. Individual JIT compile options can be set in pcretest by following -s+[+]
557 or /S+[+] with a digit between 1 and 7.
559 8. OP_NOT now supports any UTF character not just single-byte ones.
561 9. (*MARK) control verb is now supported by the JIT compiler.
563 10. The command "./RunTest list" lists the available tests without actually
564 running any of them. (Because I keep forgetting what they all are.)
568 12. Applied a (slightly modified) user-supplied patch that improves performance
569 when the heap is used for recursion (compiled with --disable-stack-for-
570 recursion). Instead of malloc and free for each heap frame each time a
571 logical recursion happens, frames are retained on a chain and re-used where
572 possible. This sometimes gives as much as 30% improvement.
574 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
575 call.
577 14. As documented, (*COMMIT) is now confined to within a positive assertion.
579 15. It is now possible to link pcretest with libedit as an alternative to
580 libreadline.
582 16. (*COMMIT) control verb is now supported by the JIT compiler.
584 17. The Unicode data tables have been updated to Unicode 6.1.0.
586 18. Added --file-list option to pcregrep.
588 19. Added binary file support to pcregrep, including the -a, --binary-files,
589 -I, and --text options.
591 20. The madvise function is renamed for posix_madvise for QNX compatibility
592 reasons. Fixed by Giuseppe D'Angelo.
594 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
596 22. Changed the option for creating symbolic links for 16-bit man pages from
597 -s to -sf so that re-installing does not cause issues.
599 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
601 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
602 matches in certain environments (the workspace was not being correctly
603 retained). Also added to pcre_dfa_exec() a simple plausibility check on
604 some of the workspace data at the beginning of a restart.
606 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
607 was not doing so when it should - probably a typo introduced by SVN 528
608 (change 8.10/14).
610 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
611 \w+ when the character tables indicated that \x{c4} was a word character.
612 There were several related cases, all because the tests for doing a table
613 lookup were testing for characters less than 127 instead of 255.
615 27. If a pattern contains capturing parentheses that are not used in a match,
616 their slots in the ovector are set to -1. For those that are higher than
617 any matched groups, this happens at the end of processing. In the case when
618 there were back references that the ovector was too small to contain
619 (causing temporary malloc'd memory to be used during matching), and the
620 highest capturing number was not used, memory off the end of the ovector
621 was incorrectly being set to -1. (It was using the size of the temporary
622 memory instead of the true size.)
624 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
625 ovector size, it uses memory at the end of the block that it has got.
627 29. Check for an overlong MARK name and give an error at compile time. The
628 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
630 30. JIT compiler update.
632 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
633 Rill for the patch.
635 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
637 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
639 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
640 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
641 was enabled.
643 35. Improve JIT code generation for greedy plus quantifier.
645 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
646 1 to "aa" instead of to an empty string. The bug affected repeated groups
647 that could potentially match an empty string.
649 37. Optimizing single character iterators in JIT.
651 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
652 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
653 codepoints that are too big for the mode are faulted, and in a UTF mode,
654 disallowed codepoints are also faulted.
656 39. If PCRE was compiled with UTF support, in three places in the DFA
657 matcher there was code that should only have been obeyed in UTF mode, but
658 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
659 processing when bytes with values greater than 127 were present. In 16-bit
660 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
661 both cases the values are those that cannot be the first data item in a UTF
662 character. The three items that might have provoked this were recursions,
663 possessively repeated groups, and atomic groups.
665 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
666 and pcregrep, because some OS require shared objects to be explicitly
667 passed to ld, causing the link step to fail if they are not.
669 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
670 mode, patterns that started with \h* or \R* might be incorrectly matched.
673 Version 8.30 04-February-2012
674 -----------------------------
676 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
677 name is defined in ctype.h.
679 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
680 only in quite long subpatterns.
682 3. Removed the function pcre_info(), which has been obsolete and deprecated
683 since it was replaced by pcre_fullinfo() in February 2000.
685 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
686 match a (*MARK), and the match failed at the start of the subject, a
687 reference to memory before the start of the subject could occur. This bug
688 was introduced by fix 17 of release 8.21.
690 5. A reference to an unset group with zero minimum repetition was giving
691 totally wrong answers (in non-JavaScript-compatibility mode). For example,
692 /(another)?(\1?)test/ matched against "hello world test". This bug was
693 introduced in release 8.13.
695 6. Add support for 16-bit character strings (a large amount of work involving
696 many changes and refactorings).
698 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
699 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
700 from a file.
702 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
703 rounding is not applied in this particular case).
705 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
706 if they appear, or are escaped, in patterns.
708 10. Get rid of a number of -Wunused-but-set-variable warnings.
710 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
711 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
712 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
713 also returns the mark "x". This bug applied to capturing parentheses,
714 non-capturing parentheses, and atomic parentheses. It also applied to some
715 assertions.
717 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
718 information out of configure.ac instead of relying on pcre.h.generic, which
719 is not stored in the repository.
721 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
722 -lreadline.
724 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
726 15. Applied Graycode's patch to put the top-level frame on the stack rather
727 than the heap when not using the stack for recursion. This gives a
728 performance improvement in many cases when recursion is not deep.
730 16. Experimental code added to "pcretest -C" to output the stack frame size.
733 Version 8.21 12-Dec-2011
734 ------------------------
736 1. Updating the JIT compiler.
738 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
739 are added as well.
741 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
742 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
743 calling _pcre_jit_exec. Some extra comments are added.
745 4. (*MARK) settings inside atomic groups that do not contain any capturing
746 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
747 was introduced by change 18 for 8.20.
749 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
750 ECMA-262 standard.
752 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
753 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
754 This bug was probably introduced by change 9 of 8.13.
756 7. While fixing 6 above, I noticed that a number of other items were being
757 incorrectly rejected as "not fixed length". This arose partly because newer
758 opcodes had not been added to the fixed-length checking code. I have (a)
759 corrected the bug and added tests for these items, and (b) arranged for an
760 error to occur if an unknown opcode is encountered while checking for fixed
761 length instead of just assuming "not fixed length". The items that were
762 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
763 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
764 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
766 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
767 being incorrectly compiled and would have given unpredicatble results.
769 9. A possessively repeated subpattern with minimum repeat count greater than
770 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
771 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
772 the first (A) could occur when it should not.
774 10. Add a cast and remove a redundant test from the code.
776 11. JIT should use pcre_malloc/pcre_free for allocation.
778 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
779 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
780 is anything other than /usr, -L is still shown).
782 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
784 14. Perl does not support \N without a following name in a [] class; PCRE now
785 also gives an error.
787 15. If a forward reference was repeated with an upper limit of around 2000,
788 it caused the error "internal error: overran compiling workspace". The
789 maximum number of forward references (including repeats) was limited by the
790 internal workspace, and dependent on the LINK_SIZE. The code has been
791 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
792 the default depends on LINK_SIZE. There is a new upper limit (for safety)
793 of around 200,000 forward references. While doing this, I also speeded up
794 the filling in of repeated forward references.
796 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
797 incorrectly expecting the subject to contain another "a" after the start.
799 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
800 in the match, the SKIP should be ignored. This was not happening; instead
801 the SKIP was being treated as NOMATCH. For patterns such as
802 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
803 tested.
805 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
806 now much more compatible with Perl, in particular in cases where the result
807 is a non-match for a non-anchored pattern. For example, if
808 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
809 "m", where previously it did not return a name. A side effect of this
810 change is that for partial matches, the last encountered mark name is
811 returned, as for non matches. A number of tests that were previously not
812 Perl-compatible have been moved into the Perl-compatible test files. The
813 refactoring has had the pleasing side effect of removing one argument from
814 the match() function, thus reducing its stack requirements.
816 19. If the /S+ option was used in pcretest to study a pattern using JIT,
817 subsequent uses of /S (without +) incorrectly behaved like /S+.
819 21. Retrieve executable code size support for the JIT compiler and fixing
820 some warnings.
822 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
823 not work when the shorter character appeared right at the end of the
824 subject string.
826 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
827 systems.
829 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
830 output it when the /M option is used in pcretest.
832 25. The CheckMan script was not being included in the distribution. Also, added
833 an explicit "perl" to run Perl scripts from the PrepareRelease script
834 because this is reportedly needed in Windows.
836 26. If study data was being save in a file and studying had not found a set of
837 "starts with" bytes for the pattern, the data written to the file (though
838 never used) was taken from uninitialized memory and so caused valgrind to
839 complain.
841 27. Updated RunTest.bat as provided by Sheri Pierce.
843 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
845 29. Computation of memory usage for the table of capturing group names was
846 giving an unnecessarily large value.
849 Version 8.20 21-Oct-2011
850 ------------------------
852 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
853 a POSIX class. After further experiments with Perl, which convinced me that
854 Perl has bugs and confusions, a closing square bracket is no longer allowed
855 in a POSIX name. This bug also affected patterns with classes that started
856 with full stops.
858 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
859 captured substring, but while checking the failing first alternative,
860 substring 1 is temporarily captured. If the output vector supplied to
861 pcre_exec() was not big enough for this capture, the yield of the function
862 was still zero ("insufficient space for captured substrings"). This cannot
863 be totally fixed without adding another stack variable, which seems a lot
864 of expense for a edge case. However, I have improved the situation in cases
865 such as /(a)(b)x|abc/ matched against "abc", where the return code
866 indicates that fewer than the maximum number of slots in the ovector have
867 been set.
869 3. Related to (2) above: when there are more back references in a pattern than
870 slots in the output vector, pcre_exec() uses temporary memory during
871 matching, and copies in the captures as far as possible afterwards. It was
872 using the entire output vector, but this conflicts with the specification
873 that only 2/3 is used for passing back captured substrings. Now it uses
874 only the first 2/3, for compatibility. This is, of course, another edge
875 case.
877 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
878 main code base, and can be used by building with --enable-jit. When this is
879 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
880 runtime --no-jit option is given.
882 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
883 ovector, the return from the function was zero, implying that there were
884 other matches that did not fit. The correct "exactly full" value is now
885 returned.
887 6. If a subpattern that was called recursively or as a subroutine contained
888 (*PRUNE) or any other control that caused it to give a non-standard return,
889 invalid errors such as "Error -26 (nested recursion at the same subject
890 position)" or even infinite loops could occur.
892 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
893 computing the minimum length on reaching *ACCEPT, and so ended up with the
894 wrong value of 1 rather than 0. Further investigation indicates that
895 computing a minimum subject length in the presence of *ACCEPT is difficult
896 (think back references, subroutine calls), and so I have changed the code
897 so that no minimum is registered for a pattern that contains *ACCEPT.
899 8. If (*THEN) was present in the first (true) branch of a conditional group,
900 it was not handled as intended. [But see 16 below.]
902 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
903 Sheri Pierce.
905 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
906 the first byte in a match must be "a".
908 11. Change 17 for 8.13 increased the recursion depth for patterns like
909 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
910 pattern contains any instances of (*THEN). If it does not, the old
911 optimizations are restored. It would be nice to do this on a per-group
912 basis, but at the moment that is not feasible.
914 12. In some environments, the output of pcretest -C is CRLF terminated. This
915 broke RunTest's code that checks for the link size. A single white space
916 character after the value is now allowed for.
918 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
919 For "fr", it uses the Windows-specific input and output files.
921 14. If (*THEN) appeared in a group that was called recursively or as a
922 subroutine, it did not work as intended. [But see next item.]
924 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
925 pattern fragments (but not containing any | characters). If A and B are
926 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
927 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
928 D. In other words, Perl considers parentheses that do not contain any |
929 characters to be part of a surrounding alternative, whereas PCRE was
930 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
931 differently. PCRE now behaves in the same way as Perl, except in the case
932 of subroutine/recursion calls such as (?1) which have in any case always
933 been different (but PCRE had them first :-).
935 16. Related to 15 above: Perl does not treat the | in a conditional group as
936 creating alternatives. Such a group is treated in the same way as an
937 ordinary group without any | characters when processing (*THEN). PCRE has
938 been changed to match Perl's behaviour.
940 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
941 RunGrepTest script failed.
943 18. Change 22 for version 13 caused atomic groups to use more stack. This is
944 inevitable for groups that contain captures, but it can lead to a lot of
945 stack use in large patterns. The old behaviour has been restored for atomic
946 groups that do not contain any capturing parentheses.
948 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
949 suppress the check for a minimum subject length at run time. (If it was
950 given to pcre_exec() or pcre_dfa_exec() it did work.)
952 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
953 fail to work when decoding hex characters in data strings in EBCDIC
954 environments.
956 21. It appears that in at least one Mac OS environment, the isxdigit() function
957 is implemented as a macro that evaluates to its argument more than once,
958 contravening the C 90 Standard (I haven't checked a later standard). There
959 was an instance in pcretest which caused it to go wrong when processing
960 \x{...} escapes in subject strings. The has been rewritten to avoid using
961 things like p++ in the argument of isxdigit().
964 Version 8.13 16-Aug-2011
965 ------------------------
967 1. The Unicode data tables have been updated to Unicode 6.0.0.
969 2. Two minor typos in pcre_internal.h have been fixed.
971 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
972 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
973 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
975 4. There were a number of related bugs in the code for matching backrefences
976 caselessly in UTF-8 mode when codes for the characters concerned were
977 different numbers of bytes. For example, U+023A and U+2C65 are an upper
978 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
979 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
980 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
981 2-byte code at the end of the subject (it thought there wasn't enough data
982 left).
984 5. Comprehensive information about what went wrong is now returned by
985 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
986 as the output vector has at least 2 elements. The offset of the start of
987 the failing character and a reason code are placed in the vector.
989 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
990 now returned is for the first byte of the failing character, instead of the
991 last byte inspected. This is an incompatible change, but I hope it is small
992 enough not to be a problem. It makes the returned offset consistent with
993 pcre_exec() and pcre_dfa_exec().
995 7. pcretest now gives a text phrase as well as the error number when
996 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
997 failure, the offset and reason code are output.
999 8. When \R was used with a maximizing quantifier it failed to skip backwards
1000 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1001 back over a single character (\n). This seems wrong (because it treated the
1002 two characters as a single entity when going forwards), conflicts with the
1003 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1004 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1005 has been changed.
1007 9. Some internal refactoring has changed the processing so that the handling
1008 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1009 time (the PCRE_DOTALL option was changed this way some time ago: version
1010 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1011 which was always a bit of a fudge. It also means that there is one less
1012 argument for the match() function, which reduces its stack requirements
1013 slightly. This change also fixes an incompatibility with Perl: the pattern
1014 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1016 10. More internal refactoring has drastically reduced the number of recursive
1017 calls to match() for possessively repeated groups such as (abc)++ when
1018 using pcre_exec().
1020 11. While implementing 10, a number of bugs in the handling of groups were
1021 discovered and fixed:
1023 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1024 (a|)*(?1) gave a compile-time internal error.
1025 ((a|)+)+ did not notice that the outer group could match an empty string.
1026 (^a|^)+ was not marked as anchored.
1027 (.*a|.*)+ was not marked as matching at start or after a newline.
1029 12. Yet more internal refactoring has removed another argument from the match()
1030 function. Special calls to this function are now indicated by setting a
1031 value in a variable in the "match data" data block.
1033 13. Be more explicit in pcre_study() instead of relying on "default" for
1034 opcodes that mean there is no starting character; this means that when new
1035 ones are added and accidentally left out of pcre_study(), testing should
1036 pick them up.
1038 14. The -s option of pcretest has been documented for ages as being an old
1039 synonym of -m (show memory usage). I have changed it to mean "force study
1040 for every regex", that is, assume /S for every regex. This is similar to -i
1041 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1042 using it. It makes it easier to run collections of tests with and without
1043 study enabled, and thereby test pcre_study() more easily. All the standard
1044 tests are now run with and without -s (but some patterns can be marked as
1045 "never study" - see 20 below).
1047 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1048 restoration of the capturing data to the outer values was not happening
1049 correctly.
1051 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1052 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1053 pattern had matched an empty string, and so incorrectly returned a no
1054 match.
1056 17. There was optimizing code for the last branch of non-capturing parentheses,
1057 and also for the obeyed branch of a conditional subexpression, which used
1058 tail recursion to cut down on stack usage. Unfortunately, now that there is
1059 the possibility of (*THEN) occurring in these branches, tail recursion is
1060 no longer possible because the return has to be checked for (*THEN). These
1061 two optimizations have therefore been removed. [But see 8.20/11 above.]
1063 18. If a pattern containing \R was studied, it was assumed that \R always
1064 matched two bytes, thus causing the minimum subject length to be
1065 incorrectly computed because \R can also match just one byte.
1067 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1068 was incorrectly computed.
1070 20. If /S is present twice on a test pattern in pcretest input, it now
1071 *disables* studying, thereby overriding the use of -s on the command line
1072 (see 14 above). This is necessary for one or two tests to keep the output
1073 identical in both cases.
1075 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1076 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1078 22. When an atomic group that contained a capturing parenthesis was
1079 successfully matched, but the branch in which it appeared failed, the
1080 capturing was not being forgotten if a higher numbered group was later
1081 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1082 group 1 to "a", when in fact it should be unset. This applied to multi-
1083 branched capturing and non-capturing groups, repeated or not, and also to
1084 positive assertions (capturing in negative assertions does not happen
1085 in PCRE) and also to nested atomic groups.
1087 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1088 subject after a captured substring, to make it easier to tell which of a
1089 number of identical substrings has been captured.
1091 24. The way atomic groups are processed by pcre_exec() has been changed so that
1092 if they are repeated, backtracking one repetition now resets captured
1093 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1094 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1095 "aaa". Previously, it would have been "a". As part of this code
1096 refactoring, the way recursive calls are handled has also been changed.
1098 25. If an assertion condition captured any substrings, they were not passed
1099 back unless some other capturing happened later. For example, if
1100 (?(?=(a))a) was matched against "a", no capturing was returned.
1102 26. When studying a pattern that contained subroutine calls or assertions,
1103 the code for finding the minimum length of a possible match was handling
1104 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1105 group 1 called group 2 while simultaneously a separate group 2 called group
1106 1). A stack overflow occurred in this case. I have fixed this by limiting
1107 the recursion depth to 10.
1109 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1110 Fortmann. This supports explicit test numbers on the command line, and has
1111 argument validation and error reporting.
1113 28. An instance of \X with an unlimited repeat could fail if at any point the
1114 first character it looked at was a mark character.
1116 29. Some minor code refactoring concerning Unicode properties and scripts
1117 should reduce the stack requirement of match() slightly.
1119 30. Added the '=' option to pcretest to check the setting of unused capturing
1120 slots at the end of the pattern, which are documented as being -1, but are
1121 not included in the return count.
1123 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1124 compiled something random. Now it gives a compile-time error (as does
1125 Perl).
1127 32. A *MARK encountered during the processing of a positive assertion is now
1128 recorded and passed back (compatible with Perl).
1130 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1131 had alternative anchored branches, the search for a second match in a line
1132 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1133 matched the line "0102" twice. The same bug affected patterns that started
1134 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1135 twice.
1137 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1138 does, and because of capturing effects, quantifying parenthesized
1139 assertions may at times be useful. Quantifiers are now allowed for
1140 parenthesized assertions.
1142 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1144 36. \g was being checked for fancy things in a character class, when it should
1145 just be a literal "g".
1147 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1148 appearance of a nested POSIX class supersedes an apparent external class.
1149 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1150 unescaped square brackets may also appear as part of class names. For
1151 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1152 more like Perl. (But see 8.20/1 above.)
1154 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1155 was because it thought it was \N{name}, which is not supported).
1157 39. Add minix to OS list not supporting the -S option in pcretest.
1159 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1160 cannot analyze patterns in sufficient detail to catch mutual recursions
1161 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1162 subgroup is called recursively as a subpattern for a second time at the
1163 same position in the subject string. In previous releases this might have
1164 been caught by the recursion limit, or it might have run out of stack.
1166 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1167 happen only once. PCRE was, however incorrectly giving a compile time error
1168 "recursive call could loop indefinitely" because it cannot analyze the
1169 pattern in sufficient detail. The compile time test no longer happens when
1170 PCRE is compiling a conditional subpattern, but actual runaway loops are
1171 now caught at runtime (see 40 above).
1173 42. It seems that Perl allows any characters other than a closing parenthesis
1174 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1175 has been changed to be the same.
1177 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1178 as not to get warnings when autogen.sh is called. Also changed
1179 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1181 44. To help people who use pcregrep to scan files containing exceedingly long
1182 lines, the following changes have been made:
1184 (a) The default value of the buffer size parameter has been increased from
1185 8K to 20K. (The actual buffer used is three times this size.)
1187 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1188 PCRE is built.
1190 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1191 to be set at run time.
1193 (d) Numerical values in pcregrep options can be followed by K or M, for
1194 example --buffer-size=50K.
1196 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1197 given and the return code is set to 2.
1199 45. Add a pointer to the latest mark to the callout data block.
1201 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1202 partial match of an empty string instead of no match. This was specific to
1203 the use of ".".
1205 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1206 complete match instead of a partial match. This bug was dependent on both
1207 the PCRE_UTF8 and PCRE_DOTALL options being set.
1209 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1210 starting byte set, because \b was not being ignored.
1213 Version 8.12 15-Jan-2011
1214 ------------------------
1216 1. Fixed some typos in the markup of the man pages, and wrote a script that
1217 checks for such things as part of the documentation building process.
1219 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1220 --match-limit and --recursion-limit options (added for 8.11). In
1221 particular, this made one of the standard tests fail. (The integer value
1222 went into the wrong half of a long int.)
1224 3. If the --colour option was given to pcregrep with -v (invert match), it
1225 did strange things, either producing crazy output, or crashing. It should,
1226 of course, ignore a request for colour when reporting lines that do not
1227 match.
1229 4. Another pcregrep bug caused similar problems if --colour was specified with
1230 -M (multiline) and the pattern match finished with a line ending.
1232 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1233 matched in multiline mode, the following line was shown as part of the
1234 match. This seems wrong, so I have changed it.
1236 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1237 the check for further matches in the same line (so they could be coloured)
1238 to overrun the end of the current line. If another match was found, it was
1239 incorrectly shown (and then shown again when found in the next line).
1241 7. If pcregrep was compiled under Windows, there was a reference to the
1242 function pcregrep_exit() before it was defined. I am assuming this was
1243 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1244 reported by a user. I've moved the definition above the reference.
1247 Version 8.11 10-Dec-2010
1248 ------------------------
1250 1. (*THEN) was not working properly if there were untried alternatives prior
1251 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1252 backtracked to try for "b" instead of moving to the next alternative branch
1253 at the same level (in this case, to look for "c"). The Perl documentation
1254 is clear that when (*THEN) is backtracked onto, it goes to the "next
1255 alternative in the innermost enclosing group".
1257 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1258 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1259 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1260 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1261 (*THEN).
1263 3. If \s appeared in a character class, it removed the VT character from
1264 the class, even if it had been included by some previous item, for example
1265 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1266 of \s, but is part of the POSIX "space" class.)
1268 4. A partial match never returns an empty string (because you can always
1269 match an empty string at the end of the subject); however the checking for
1270 an empty string was starting at the "start of match" point. This has been
1271 changed to the "earliest inspected character" point, because the returned
1272 data for a partial match starts at this character. This means that, for
1273 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1274 (previously it gave "no match").
1276 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1277 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1278 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1279 has an implication that the given string is incomplete (because a partial
1280 match is preferred over a full match). For this reason, these items now
1281 give a partial match in this situation. [Aside: previously, the one case
1282 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1283 match rather than a full match, which was wrong by the old rules, but is
1284 now correct.]
1286 6. There was a bug in the handling of #-introduced comments, recognized when
1287 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1288 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1289 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1290 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1291 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1292 places in pcre_compile().
1294 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1295 comments when looking ahead for named forward references to subpatterns,
1296 the only newline sequence it recognized was NL. It now handles newlines
1297 according to the set newline convention.
1299 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1300 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1301 cater for a lack of strerror(). These oversights have been fixed.
1303 9. Added --match-limit and --recursion-limit to pcregrep.
1305 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1307 11. When the -o option was used, pcregrep was setting a return code of 1, even
1308 when matches were found, and --line-buffered was not being honoured.
1310 12. Added an optional parentheses number to the -o and --only-matching options
1311 of pcregrep.
1313 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1314 can match an empty string. The code to do it in pcretest and pcredemo
1315 needed fixing:
1317 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1318 only one byte after an empty string match just before CRLF (this case
1319 just got forgotten; "any" and "anycrlf" were OK).
1321 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1322 mode when an empty string match preceded an ASCII character followed by
1323 a non-ASCII character. (The code for advancing by one character rather
1324 than one byte was nonsense.)
1326 (c) The pcredemo.c sample program did not have any code at all to handle
1327 the cases when CRLF is a valid newline sequence.
1329 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1330 as a starting offset was within the subject string. There is now a new
1331 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1332 negative or greater than the length of the string. In order to test this,
1333 pcretest is extended to allow the setting of negative starting offsets.
1335 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1336 starting offset points to the beginning of a UTF-8 character was
1337 unnecessarily clumsy. I tidied it up.
1339 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1340 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1342 17. Nobody had reported that the --include_dir option, which was added in
1343 release 7.7 should have been called --include-dir (hyphen, not underscore)
1344 for compatibility with GNU grep. I have changed it to --include-dir, but
1345 left --include_dir as an undocumented synonym, and the same for
1346 --exclude-dir, though that is not available in GNU grep, at least as of
1347 release 2.5.4.
1349 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1350 characters from a string of bytes) have been redefined so as not to use
1351 loops, in order to improve performance in some environments. At the same
1352 time, I abstracted some of the common code into auxiliary macros to save
1353 repetition (this should not affect the compiled code).
1355 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1356 compile-time error is now given if \c is not followed by an ASCII
1357 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1358 different, and any byte value is allowed.)
1360 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1361 START_OPTIMIZE option, which is now allowed at compile time - but just
1362 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1363 to pcregrep and other applications that have no direct access to PCRE
1364 options. The new /Y option in pcretest sets this option when calling
1365 pcre_compile().
1367 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1368 back references. Groups containing recursive back references were forced to
1369 be atomic by that change, but in the case of named groups, the amount of
1370 memory required was incorrectly computed, leading to "Failed: internal
1371 error: code overflow". This has been fixed.
1373 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1374 pcretest.c, to avoid build problems in some Borland environments.
1377 Version 8.10 25-Jun-2010
1378 ------------------------
1380 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1381 THEN.
1383 2. (*ACCEPT) was not working when inside an atomic group.
1385 3. Inside a character class, \B is treated as a literal by default, but
1386 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1387 causes the error). The code is unchanged, but I tidied the documentation.
1389 4. Inside a character class, PCRE always treated \R and \X as literals,
1390 whereas Perl faults them if its -w option is set. I have changed PCRE so
1391 that it faults them when PCRE_EXTRA is set.
1393 5. Added support for \N, which always matches any character other than
1394 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1396 6. When compiling pcregrep with newer versions of gcc which may have
1397 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1398 declared with attribute warn_unused_result" were given. Just casting the
1399 result to (void) does not stop the warnings; a more elaborate fudge is
1400 needed. I've used a macro to implement this.
1402 7. Minor change to pcretest.c to avoid a compiler warning.
1404 8. Added four artifical Unicode properties to help with an option to make
1405 \s etc use properties (see next item). The new properties are: Xan
1406 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1408 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1409 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1410 this option. Modified pcretest to add /W to test this facility. Added
1411 REG_UCP to make it available via the POSIX interface.
1413 10. Added --line-buffered to pcregrep.
1415 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1416 studied, and the match started with a letter with a code point greater than
1417 127 whose first byte was different to the first byte of the other case of
1418 the letter, the other case of this starting letter was not recognized
1419 (#976).
1421 12. If a pattern that was studied started with a repeated Unicode property
1422 test, for example, \p{Nd}+, there was the theoretical possibility of
1423 setting up an incorrect bitmap of starting bytes, but fortunately it could
1424 not have actually happened in practice until change 8 above was made (it
1425 added property types that matched character-matching opcodes).
1427 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1428 possible starting bytes for non-anchored patterns.
1430 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1431 \R, and also a number of cases that involve Unicode properties, both
1432 explicit and implicit when PCRE_UCP is set.
1434 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1435 input, it could crash or give wrong results if characters with values
1436 greater than 0xc0 were present in the subject string. (Detail: it assumed
1437 UTF-8 input when processing these items.)
1439 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1440 size_t is 64-bit (#991).
1442 17. Added a check for running out of memory when PCRE is compiled with
1443 --disable-stack-for-recursion (#990).
1445 18. If the last data line in a file for pcretest does not have a newline on
1446 the end, a newline was missing in the output.
1448 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1449 less than 128) in its various bitmaps. However, there is a facility for
1450 generating tables according to the current locale when PCRE is compiled. It
1451 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1452 characters, are recognized by isspace() and therefore were getting set in
1453 these tables, and indeed these tables seem to approximate to ISO 8859. This
1454 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1455 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1456 which of course cannot start UTF-8 characters. I have changed the code so
1457 that only real ASCII characters (less than 128) and the correct starting
1458 bytes for UTF-8 encodings are set for characters greater than 127 when in
1459 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1460 altogether.)
1462 20. Added the /T option to pcretest so as to be able to run tests with non-
1463 standard character tables, thus making it possible to include the tests
1464 used for 19 above in the standard set of tests.
1466 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1467 reference to a subpattern the other side of a comment that contains an
1468 opening parenthesis caused either an internal compiling error, or a
1469 reference to the wrong subpattern.
1472 Version 8.02 19-Mar-2010
1473 ------------------------
1475 1. The Unicode data tables have been updated to Unicode 5.2.0.
1477 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1478 configured.
1480 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1481 original author of that file, following a query about its status.
1483 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1484 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1486 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1487 quantifier applied to a forward-referencing subroutine call, could compile
1488 incorrect code or give the error "internal error: previously-checked
1489 referenced subpattern not found".
1491 6. Both MS Visual Studio and Symbian OS have problems with initializing
1492 variables to point to external functions. For these systems, therefore,
1493 pcre_malloc etc. are now initialized to local functions that call the
1494 relevant global functions.
1496 7. There were two entries missing in the vectors called coptable and poptable
1497 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1498 I've fixed the data, and added a kludgy way of testing at compile time that
1499 the lengths are correct (equal to the number of opcodes).
1501 8. Following on from 7, I added a similar kludge to check the length of the
1502 eint vector in pcreposix.c.
1504 9. Error texts for pcre_compile() are held as one long string to avoid too
1505 much relocation at load time. To find a text, the string is searched,
1506 counting zeros. There was no check for running off the end of the string,
1507 which could happen if a new error number was added without updating the
1508 string.
1510 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1512 11. \K was not working if it appeared in an atomic group or in a group that
1513 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1514 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1515 the assertion is positive, but not if it is negative.
1517 12. Change 11 fortuitously reduced the size of the stack frame used in the
1518 "match()" function of pcre_exec.c by one pointer. Forthcoming
1519 implementation of support for (*MARK) will need an extra pointer on the
1520 stack; I have reserved it now, so that the stack frame size does not
1521 decrease.
1523 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1524 item in branch that calls a recursion is a subroutine call - as in the
1525 second branch in the above example - was incorrectly given the compile-
1526 time error "recursive call could loop indefinitely" because pcre_compile()
1527 was not correctly checking the subroutine for matching a non-empty string.
1529 14. The checks for overrunning compiling workspace could trigger after an
1530 overrun had occurred. This is a "should never occur" error, but it can be
1531 triggered by pathological patterns such as hundreds of nested parentheses.
1532 The checks now trigger 100 bytes before the end of the workspace.
1534 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1537 Version 8.01 19-Jan-2010
1538 ------------------------
1540 1. If a pattern contained a conditional subpattern with only one branch (in
1541 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1542 computed the wrong minimum data length (which is of course zero for such
1543 subpatterns). This could cause incorrect "no match" results.
1545 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1546 the pattern is reset in the first branch, pcre_compile() failed with
1547 "internal error: code overflow at offset...". This happened only when
1548 the reset was to the original external option setting. (An optimization
1549 abstracts leading options settings into an external setting, which was the
1550 cause of this.)
1552 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1553 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1554 assertion pattern did not match (meaning that the assertion was true), it
1555 was incorrectly treated as false if the SKIP had been reached during the
1556 matching. This also applied to assertions used as conditions.
1558 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1559 assertion subpattern, including such a pattern used as a condition,
1560 unpredictable results occurred, instead of the error return
1563 5. The C++ GlobalReplace function was not working like Perl for the special
1564 situation when an empty string is matched. It now does the fancy magic
1565 stuff that is necessary.
1567 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1568 removed. (These were left over from very, very early versions of PCRE.)
1570 7. Some cosmetic changes to the code to make life easier when compiling it
1571 as part of something else:
1573 (a) Change DEBUG to PCRE_DEBUG.
1575 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1576 called "current" as "current_branch", to prevent a collision with the
1577 Linux macro when compiled as a kernel module.
1579 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1580 prevent a collision with the Linux macro when compiled as a kernel
1581 module.
1583 8. In pcre_compile() there are some checks for integer overflows that used to
1584 cast potentially large values to (double). This has been changed to that
1585 when building, a check for int64_t is made, and if it is found, it is used
1586 instead, thus avoiding the use of floating point arithmetic. (There is no
1587 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1588 double.
1590 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1591 2005 (difference between two addresses compared to an unsigned value).
1593 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1594 custom one, because of the following reported problem in Windows:
1596 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1597 under Win32.
1598 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1599 therefore missing the function definition.
1600 - The compiler thus generates a "C" signature for the test function.
1601 - The linker fails to find the "C" function.
1602 - PCRE fails to configure if asked to do so against libbz2.
1604 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1605 messages were output:
1607 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1608 rerunning libtoolize, to keep the correct libtool macros in-tree.
1609 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1611 I have done both of these things.
1613 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1614 most of the time, it *can* run out if it is given a pattern that contains a
1615 runaway infinite recursion. I updated the discussion in the pcrestack man
1616 page.
1618 13. Now that we have gone to the x.xx style of version numbers, the minor
1619 version may start with zero. Using 08 or 09 is a bad idea because users
1620 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1621 interpreted as invalid octal numbers. I've updated the previous comment in
1622 configure.ac, and also added a check that gives an error if 08 or 09 are
1623 used.
1625 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1626 causing partial matching to fail when the end of the subject matched \W
1627 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1629 15. There were some discrepancies between the declarations in pcre_internal.h
1630 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1631 their definitions. The declarations used "const uschar *" and the
1632 definitions used USPTR. Even though USPTR is normally defined as "const
1633 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1634 reported that: "This difference in casting confuses some C++ compilers, for
1635 example, SunCC recognizes above declarations as different functions and
1636 generates broken code for hbpcre." I have changed the declarations to use
1637 USPTR.
1639 16. GNU libtool is named differently on some systems. The autogen.sh script now
1640 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1641 (FreeBSD).
1643 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1644 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1645 comment: "Figure out how to create a longlong from a string: strtoll and
1646 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1647 instance, but it only takes 2 args instead of 3!"
1649 18. A subtle bug concerned with back references has been fixed by a change of
1650 specification, with a corresponding code fix. A pattern such as
1651 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1652 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1653 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1654 same bug. Such groups have to be quantified to be useful, or contained
1655 inside another quantified group. (If there's no repetition, the reference
1656 can never match.) The problem arises because, having left the group and
1657 moved on to the rest of the pattern, a later failure that backtracks into
1658 the group uses the captured value from the final iteration of the group
1659 rather than the correct earlier one. I have fixed this in PCRE by forcing
1660 any group that contains a reference to itself to be an atomic group; that
1661 is, there cannot be any backtracking into it once it has completed. This is
1662 similar to recursive and subroutine calls.
1665 Version 8.00 19-Oct-09
1666 ----------------------
1668 1. The table for translating pcre_compile() error codes into POSIX error codes
1669 was out-of-date, and there was no check on the pcre_compile() error code
1670 being within the table. This could lead to an OK return being given in
1671 error.
1673 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1674 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1675 in a Windows environment.
1677 3. The pcregrep --count option prints the count for each file even when it is
1678 zero, as does GNU grep. However, pcregrep was also printing all files when
1679 --files-with-matches was added. Now, when both options are given, it prints
1680 counts only for those files that have at least one match. (GNU grep just
1681 prints the file name in this circumstance, but including the count seems
1682 more useful - otherwise, why use --count?) Also ensured that the
1683 combination -clh just lists non-zero counts, with no names.
1685 4. The long form of the pcregrep -F option was incorrectly implemented as
1686 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1687 but it seems right to fix it, and I didn't think it was worth preserving
1688 the old behaviour.
1690 5. The command line items --regex=pattern and --regexp=pattern were not
1691 recognized by pcregrep, which required --regex pattern or --regexp pattern
1692 (with a space rather than an '='). The man page documented the '=' forms,
1693 which are compatible with GNU grep; these now work.
1695 6. No libpcreposix.pc file was created for pkg-config; there was just
1696 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1698 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1699 when UCP support is not needed, by modifying the Python script that
1700 generates it from Unicode data files. This should not matter if the module
1701 is correctly used as a library, but I received one complaint about 50K of
1702 unwanted data. My guess is that the person linked everything into his
1703 program rather than using a library. Anyway, it does no harm.
1705 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1706 was a minimum greater than 1 for a wide character in a possessive
1707 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1708 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1709 character. Chaos in the form of incorrect output or a compiling loop could
1710 result.
1712 9. The restrictions on what a pattern can contain when partial matching is
1713 requested for pcre_exec() have been removed. All patterns can now be
1714 partially matched by this function. In addition, if there are at least two
1715 slots in the offset vector, the offset of the earliest inspected character
1716 for the match and the offset of the end of the subject are set in them when
1717 PCRE_ERROR_PARTIAL is returned.
1719 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1720 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1721 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1722 and may be more useful for multi-segment matching.
1724 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1725 used to be given if ever the end of the subject was reached; now it is
1726 given only if matching could not proceed because another character was
1727 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1728 string "Z", which now yields "no match" instead of "partial match". In the
1729 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1730 final character ended with (*FAIL).
1732 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1733 if the pattern had a "must contain" character that was already found in the
1734 earlier partial match, unless partial matching was again requested. For
1735 example, with the pattern /dog.(body)?/, the "must contain" character is
1736 "g". If the first part-match was for the string "dog", restarting with
1737 "sbody" failed. This bug has been fixed.
1739 13. The string returned by pcre_dfa_exec() after a partial match has been
1740 changed so that it starts at the first inspected character rather than the
1741 first character of the match. This makes a difference only if the pattern
1742 starts with a lookbehind assertion or \b or \B (\K is not supported by
1743 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1744 matching functions compatible, and I think it's the right thing to do.
1746 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1747 so that the demonstration program is easily available in environments where
1748 PCRE has not been installed from source.
1750 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1751 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1752 library.
1754 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1755 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1756 is not the first non-POSIX option to be added. Clearly some people find
1757 these options useful.
1759 17. If a caller to the POSIX matching function regexec() passes a non-zero
1760 value for nmatch with a NULL value for pmatch, the value of
1761 nmatch is forced to zero.
1763 18. RunGrepTest did not have a test for the availability of the -u option of
1764 the diff command, as RunTest does. It now checks in the same way as
1765 RunTest, and also checks for the -b option.
1767 19. If an odd number of negated classes containing just a single character
1768 interposed, within parentheses, between a forward reference to a named
1769 subpattern and the definition of the subpattern, compilation crashed with
1770 an internal error, complaining that it could not find the referenced
1771 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1772 [The bug was that it was starting one character too far in when skipping
1773 over the character class, thus treating the ] as data rather than
1774 terminating the class. This meant it could skip too much.]
1776 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1777 /g option in pcretest when the pattern contains \K, which makes it possible
1778 to have an empty string match not at the start, even when the pattern is
1779 anchored. Updated pcretest and pcredemo to use this option.
1781 21. If the maximum number of capturing subpatterns in a recursion was greater
1782 than the maximum at the outer level, the higher number was returned, but
1783 with unset values at the outer level. The correct (outer level) value is
1784 now given.
1786 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1787 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1788 make it do so. The string so far is captured, making this feature
1789 compatible with Perl.
1791 23. The tests have been re-organized, adding tests 11 and 12, to make it
1792 possible to check the Perl 5.10 features against Perl 5.10.
1794 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1795 pattern matches a fixed length string. PCRE did not allow this; now it
1796 does. Neither allows recursion.
1798 25. I finally figured out how to implement a request to provide the minimum
1799 length of subject string that was needed in order to match a given pattern.
1800 (It was back references and recursion that I had previously got hung up
1801 on.) This code has now been added to pcre_study(); it finds a lower bound
1802 to the length of subject needed. It is not necessarily the greatest lower
1803 bound, but using it to avoid searching strings that are too short does give
1804 some useful speed-ups. The value is available to calling programs via
1805 pcre_fullinfo().
1807 26. While implementing 25, I discovered to my embarrassment that pcretest had
1808 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1809 study optimizations had never been tested with that matching function.
1810 Oops. What is worse, even when it was passed study data, there was a bug in
1811 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1812 were also very few tests of studied patterns with pcre_dfa_exec().
1814 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1815 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1816 on the other side of the coin, they are no longer allowed to have different
1817 names, because these cannot be distinguished in PCRE, and this has caused
1818 confusion. (This is a difference from Perl.)
1820 28. When duplicate subpattern names are present (necessarily with different
1821 numbers, as required by 27 above), and a test is made by name in a
1822 conditional pattern, either for a subpattern having been matched, or for
1823 recursion in such a pattern, all the associated numbered subpatterns are
1824 tested, and the overall condition is true if the condition is true for any
1825 one of them. This is the way Perl works, and is also more like the way
1826 testing by number works.
1829 Version 7.9 11-Apr-09
1830 ---------------------
1832 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1833 (pcretest), all targets were linked against these libraries. This included
1834 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1835 libraries. This caused unwanted dependencies to be created. This problem
1836 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1837 pcretest is linked with readline.
1839 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1840 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1841 moved outside it again, because FALSE and TRUE are already defined in AIX,
1842 but BOOL is not.
1844 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1845 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1847 4. The pcregrep documentation said spaces were inserted as well as colons (or
1848 hyphens) following file names and line numbers when outputting matching
1849 lines. This is not true; no spaces are inserted. I have also clarified the
1850 wording for the --colour (or --color) option.
1852 5. In pcregrep, when --colour was used with -o, the list of matching strings
1853 was not coloured; this is different to GNU grep, so I have changed it to be
1854 the same.
1856 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1857 each matching line was coloured. Now it goes on to look for further matches
1858 of any of the test patterns, which is the same behaviour as GNU grep.
1860 7. A pattern that could match an empty string could cause pcregrep to loop; it
1861 doesn't make sense to accept an empty string match in pcregrep, so I have
1862 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1863 seems to be how GNU grep behaves. [But see later change 40 for release
1864 8.33.]
1866 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1867 start or after a newline", because the conditional assertion was not being
1868 correctly handled. The rule now is that both the assertion and what follows
1869 in the first alternative must satisfy the test.
1871 9. If auto-callout was enabled in a pattern with a conditional group whose
1872 condition was an assertion, PCRE could crash during matching, both with
1873 pcre_exec() and pcre_dfa_exec().
1875 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1876 used for matching.
1878 11. Unicode property support in character classes was not working for
1879 characters (bytes) greater than 127 when not in UTF-8 mode.
1881 12. Added the -M command line option to pcretest.
1883 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1885 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1887 16. Added comments and documentation about mis-use of no_arg in the C++
1888 wrapper.
1890 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1891 from Martin Jerabek that uses macro names for all relevant character and
1892 string constants.
1894 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1895 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1896 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1897 these, but not everybody uses configure.
1899 19. A conditional group that had only one branch was not being correctly
1900 recognized as an item that could match an empty string. This meant that an
1901 enclosing group might also not be so recognized, causing infinite looping
1902 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1903 with the subject "ab", where knowledge that the repeated group can match
1904 nothing is needed in order to break the loop.
1906 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1907 exec(), but without supplying a callout function, matching went wrong.
1909 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1910 leak if the size of the offset vector was greater than 30. When the vector
1911 is smaller, the saved offsets during recursion go onto a local stack
1912 vector, but for larger vectors malloc() is used. It was failing to free
1913 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1914 error, in fact).
1916 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1917 heapframe that is used only when UTF-8 support is enabled. This caused no
1918 problem, but was untidy.
1920 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1921 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1922 included within another project.
1924 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1925 slightly modified by me:
1927 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1928 not building pcregrep.
1930 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1931 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1933 25. Forward references, both numeric and by name, in patterns that made use of
1934 duplicate group numbers, could behave incorrectly or give incorrect errors,
1935 because when scanning forward to find the reference group, PCRE was not
1936 taking into account the duplicate group numbers. A pattern such as
1937 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1939 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1940 the feature of a custom pointer more persuasive (as requested by a user).
1942 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1943 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1944 pre-defined.
1946 28. Added support for (*UTF8) at the start of a pattern.
1948 29. Arrange for flags added by the "release type" setting in CMake to be shown
1949 in the configuration summary.
1952 Version 7.8 05-Sep-08
1953 ---------------------
1955 1. Replaced UCP searching code with optimized version as implemented for Ad
1956 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1957 stage table and inline lookup instead of a function, giving speed ups of 2
1958 to 5 times on some simple patterns that I tested. Permission was given to
1959 distribute the MultiStage2.py script that generates the tables (it's not in
1960 the tarball, but is in the Subversion repository).
1962 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1963 scripts.
1965 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1966 a group with a zero qualifier. The result of the study could be incorrect,
1967 or the function might crash, depending on the pattern.
1969 4. Caseless matching was not working for non-ASCII characters in back
1970 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1971 It now works when Unicode Property Support is available.
1973 5. In pcretest, an escape such as \x{de} in the data was always generating
1974 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1975 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1976 truncation.
1978 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1980 7. Added two (int) casts to pcregrep when printing the difference of two
1981 pointers, in case they are 64-bit values.
1983 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1984 test 2 if it fails.
1986 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1987 and a #define of that name to empty if it is not externally set. This is to
1988 allow users of MSVC to set it if necessary.
1990 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1991 the convenience functions in the pcre_get.c source file.
1993 11. An option change at the start of a pattern that had top-level alternatives
1994 could cause overwriting and/or a crash. This command provoked a crash in
1995 some environments:
1997 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1999 This potential security problem was recorded as CVE-2008-2371.
2001 12. For a pattern where the match had to start at the beginning or immediately
2002 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2003 pcre_dfa_exec() could read past the end of the passed subject if there was
2004 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2005 pcretest so that it places the subject at the end of its malloc-ed buffer.
2007 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2008 exec() might read past the end of the data buffer in UTF-8 mode.
2010 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2011 the data contained the byte 0x85 as part of a UTF-8 character within its
2012 first line. This applied both to normal and DFA matching.
2014 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2015 /^[^d]*?$/8 failed to match "abc".
2017 16. Added a missing copyright notice to pcrecpp_internal.h.
2019 17. Make it more clear in the documentation that values returned from
2020 pcre_exec() in ovector are byte offsets, not character counts.
2022 18. Tidied a few places to stop certain compilers from issuing warnings.
2024 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2025 supplied by Stefan Weber. I made a further small update for 7.8 because
2026 there is a change of source arrangements: the pcre_searchfuncs.c module is
2027 replaced by pcre_ucd.c.
2030 Version 7.7 07-May-08
2031 ---------------------
2033 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2034 a string to a long long, pretend we don't even have a long long." This is
2035 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2037 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2038 pre-7.6 versions, which defined a global no_arg variable instead of putting
2039 it in the RE class. (See also #8 below.)
2041 3. Remove a line of dead code, identified by coverity and reported by Nuno
2042 Lopes.
2044 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2046 (1) The include/exclude patterns were being applied to the whole pathnames
2047 of files, instead of just to the final components.
2049 (2) If there was more than one level of directory, the subdirectories were
2050 skipped unless they satisfied the include/exclude conditions. This is
2051 inconsistent with GNU grep (and could even be seen as contrary to the
2052 pcregrep specification - which I improved to make it absolutely clear).
2053 The action now is always to scan all levels of directory, and just
2054 apply the include/exclude patterns to regular files.
2056 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2057 --exclude_dir in the tests to avoid scanning .svn directories.
2059 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2060 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2061 doesn't support NULs in patterns.
2063 7. Added some missing "const"s to declarations of static tables in
2064 pcre_compile.c and pcre_dfa_exec.c.
2066 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2067 caused by fix #2 above. (Subsequently also a second patch to fix the
2068 first patch. And a third patch - this was a messy problem.)
2070 9. Applied Craig's patch to remove the use of push_back().
2072 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2073 matching function regexec().
2075 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2076 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2077 references. PCRE supports relative numbers with this syntax (I don't think
2078 Oniguruma does).
2080 12. Previously, a group with a zero repeat such as (...){0} was completely
2081 omitted from the compiled regex. However, this means that if the group
2082 was called as a subroutine from elsewhere in the pattern, things went wrong
2083 (an internal error was given). Such groups are now left in the compiled
2084 pattern, with a new opcode that causes them to be skipped at execution
2085 time.
2087 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2088 to the way PCRE behaves:
2090 (a) A lone ] character is dis-allowed (Perl treats it as data).
2092 (b) A back reference to an unmatched subpattern matches an empty string
2093 (Perl fails the current match path).
2095 (c) A data ] in a character class must be notated as \] because if the
2096 first data character in a class is ], it defines an empty class. (In
2097 Perl it is not possible to have an empty class.) The empty class []
2098 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2099 The negative empty class [^] matches any one character, independently
2100 of the DOTALL setting.
2102 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2103 non-existent subpattern following a character class starting with ']' and
2104 containing () gave an internal compiling error instead of "reference to
2105 non-existent subpattern". Fortunately, when the pattern did exist, the
2106 compiled code was correct. (When scanning forwards to check for the
2107 existence of the subpattern, it was treating the data ']' as terminating
2108 the class, so got the count wrong. When actually compiling, the reference
2109 was subsequently set up correctly.)
2111 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2112 it was being rejected as not supported by pcre_dfa_exec(), even though
2113 other assertions are supported. I have made pcre_dfa_exec() support
2114 (*FAIL).
2116 16. The implementation of 13c above involved the invention of a new opcode,
2117 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2118 cannot be changed at match time, I realized I could make a small
2119 improvement to matching performance by compiling OP_ALLANY instead of
2120 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2121 on the OP_ANY path.
2123 17. Compiling pcretest on Windows with readline support failed without the
2124 following two fixes: (1) Make the unistd.h include conditional on
2125 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2127 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2128 ncurses library to be included for pcretest when ReadLine support is
2129 requested, but also to allow for it to be overridden. This patch came from
2130 Daniel Bergström.
2132 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2133 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2134 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2135 spotting this.
2138 Version 7.6 28-Jan-08
2139 ---------------------
2141 1. A character class containing a very large number of characters with
2142 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2143 overflow.
2145 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2146 HAVE_LONG_LONG is not defined.
2148 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2149 bring them up to date and include new features. This patch includes:
2151 - Fixed PH's badly added libz and libbz2 support.
2152 - Fixed a problem with static linking.
2153 - Added pcredemo. [But later removed - see 7 below.]
2154 - Fixed dftables problem and added an option.
2155 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2157 - Added readline support for pcretest.
2158 - Added an listing of the option settings after cmake has run.
2160 4. A user submitted a patch to Makefile that makes it easy to create
2161 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2162 Makefile.am that cause it to include this special target, without
2163 affecting anything else. Note that the same mingw target plus all
2164 the other distribution libraries and programs are now supported
2165 when configuring with CMake (see 6 below) instead of with
2166 Configure/Make.
2168 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2169 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2170 exported in the Windows port". It has not yet been confirmed that the patch
2171 solves the problem, but it does no harm.
2173 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2174 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2175 with CMake, and also correct the comment about stack recursion.
2177 7. Remove the automatic building of pcredemo from the ./configure system and
2178 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2179 of a program that users should build themselves after PCRE is installed, so
2180 building it automatically is not really right. What is more, it gave
2181 trouble in some build environments.
2183 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2186 Version 7.5 10-Jan-08
2187 ---------------------
2189 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2190 values in parens when parsing an RE using the C++ wrapper."
2192 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2193 Characters greater than 255 were excluded from the class instead of being
2194 included.
2196 3. The same bug as (2) above applied to negated POSIX classes such as
2197 [:^space:].
2199 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2200 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2201 I have changed it.
2203 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2204 first named subpattern) and a construct such as (?&a) would reference the
2205 first named subpattern whose name started with "a" (in other words, the
2206 length check was missing). Both these problems are fixed. "Subpattern name
2207 expected" is now given for (?&) (a zero-length name), and this patch also
2208 makes it give the same error for \k'' (previously it complained that that
2209 was a reference to a non-existent subpattern).
2211 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2212 this is right because (?- can be followed by option settings as well as by
2213 digits. I have, however, made the messages clearer.
2215 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2216 than the number used in the conditional) now cause a compile-time error.
2217 This is actually not compatible with Perl, which accepts such patterns, but
2218 treats the conditional as always being FALSE (as PCRE used to), but it
2219 seems to me that giving a diagnostic is better.
2221 8. Change "alphameric" to the more common word "alphanumeric" in comments
2222 and messages.
2224 9. Fix two occurrences of "backslash" in comments that should have been
2225 "backspace".
2227 10. Remove two redundant lines of code that can never be obeyed (their function
2228 was moved elsewhere).
2230 11. The program that makes PCRE's Unicode character property table had a bug
2231 which caused it to generate incorrect table entries for sequences of
2232 characters that have the same character type, but are in different scripts.
2233 It amalgamated them into a single range, with the script of the first of
2234 them. In other words, some characters were in the wrong script. There were
2235 thirteen such cases, affecting characters in the following ranges:
2237 U+002b0 - U+002c1
2238 U+0060c - U+0060d
2239 U+0061e - U+00612
2240 U+0064b - U+0065e
2241 U+0074d - U+0076d
2242 U+01800 - U+01805
2243 U+01d00 - U+01d77
2244 U+01d9b - U+01dbf
2245 U+0200b - U+0200f
2246 U+030fc - U+030fe
2247 U+03260 - U+0327f
2248 U+0fb46 - U+0fbb1
2249 U+10450 - U+1049d
2251 12. The -o option (show only the matching part of a line) for pcregrep was not
2252 compatible with GNU grep in that, if there was more than one match in a
2253 line, it showed only the first of them. It now behaves in the same way as
2254 GNU grep.
2256 13. If the -o and -v options were combined for pcregrep, it printed a blank
2257 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2258 does the same. The return code can be used to tell if there were any
2259 non-matching lines.
2261 14. Added --file-offsets and --line-offsets to pcregrep.
2263 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2264 infinitely looping recursion. The bug was that positive lookaheads were not
2265 being skipped when checking for a possible empty match (negative lookaheads
2266 and both kinds of lookbehind were skipped).
2268 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2269 inclusion of <windows.h> to before rather than after the definition of
2270 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2272 17. Specifying a possessive quantifier with a specific limit for a Unicode
2273 character property caused pcre_compile() to compile bad code, which led at
2274 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2275 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2276 caused the error; without that there was no problem.
2278 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2280 19. Added --enable-pcretest-libreadline.
2282 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2283 RE::GlobalReplace(). As a result, the number of replacements returned was
2284 double what it should be. I removed one of the increments, but Craig sent a
2285 later patch that removed the other one (the right fix) and added unit tests
2286 that check the return values (which was not done before).
2288 21. Several CMake things:
2290 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2291 the names libpcre and libpcreposix, not just pcre and pcreposix.
2293 (2) The above change means that pcretest and pcregrep are now correctly
2294 linked with the newly-built libraries, not previously installed ones.
2298 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2299 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2300 UTF-8 newline character). The key issue is that the pattern starts .*;
2301 this means that the match must be either at the beginning, or after a
2302 newline. The bug was in the code for advancing after a failed match and
2303 checking that the new position followed a newline. It was not taking
2304 account of UTF-8 characters correctly.
2306 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2307 character classes. PCRE was not treating the sequence [:...:] as a
2308 character class unless the ... were all letters. Perl, however, seems to
2309 allow any characters between [: and :], though of course it rejects as
2310 unknown any "names" that contain non-letters, because all the known class
2311 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2312 for example, whereas PCRE did not - it did not recognize a POSIX character
2313 class. This seemed a bit dangerous, so the code has been changed to be
2314 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2315 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2316 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2317 Perl does, and where it didn't before.
2319 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2320 Windows environments %n is disabled by default.
2323 Version 7.4 21-Sep-07
2324 ---------------------
2326 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2327 means that a class such as [\s] counted as "explicit reference to CR or
2328 LF". That isn't really right - the whole point of the change was to try to
2329 help when there was an actual mention of one of the two characters. So now
2330 the change happens only if \r or \n (or a literal CR or LF) character is
2331 encountered.
2333 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2334 of both had grown to the point where there were only 3 bits left.
2335 Fortunately, there was spare space in the data structure, and so I have
2336 moved the internal flags into a new 16-bit field to free up more option
2337 bits.
2339 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2340 but did not set the internal JCHANGED flag - either of these is enough to
2341 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2342 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2343 start sets both bits.
2345 4. Added options (at build time, compile time, exec time) to change \R from
2346 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2348 5. doc/pcresyntax.html was missing from the distribution.
2350 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2351 compatibility, even though it is no longer used.
2353 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2354 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2355 windows.h file is present (where different names are used). [This was
2356 reversed later after testing - see 16 below.]
2358 8. Changed all #include <config.h> to #include "config.h". There were also
2359 some further <pcre.h> cases that I changed to "pcre.h".
2361 9. When pcregrep was used with the --colour option, it missed the line ending
2362 sequence off the lines that it output.
2364 10. It was pointed out to me that arrays of string pointers cause lots of
2365 relocations when a shared library is dynamically loaded. A technique of
2366 using a single long string with a table of offsets can drastically reduce
2367 these. I have refactored PCRE in four places to do this. The result is
2368 dramatic:
2370 Originally: 290
2371 After changing UCP table: 187
2372 After changing error message table: 43
2373 After changing table of "verbs" 36
2374 After changing table of Posix names 22
2376 Thanks to the folks working on Gregex for glib for this insight.
2378 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2379 unicode-properties was also set.
2381 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2383 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2384 checked only for CRLF.
2386 14. Added casts to pcretest.c to avoid compiler warnings.
2388 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2390 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2391 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2392 entirely. This removes changes made in 7 above.
2394 17. The CMake files have been updated, and there is now more information about
2395 building with CMake in the NON-UNIX-USE document.
2398 Version 7.3 28-Aug-07
2399 ---------------------
2401 1. In the rejigging of the build system that eventually resulted in 7.1, the
2402 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2403 brackets there is not right, since it causes compilers to look for an
2404 installed pcre.h, not the version that is in the source that is being
2405 compiled (which of course may be different). I have changed it back to:
2407 #include "pcre.h"
2409 I have a vague recollection that the change was concerned with compiling in
2410 different directories, but in the new build system, that is taken care of
2411 by the VPATH setting the Makefile.
2413 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2414 when the subject happened to end in the byte 0x85 (e.g. if the last
2415 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2416 characters but of course it shouldn't be taken as a newline when it is part
2417 of another character. The bug was that, for an unlimited repeat of . in
2418 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2419 characters when looking for a newline.
2421 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2423 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2424 in debug output.
2426 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2427 long printing in the pcrecpp unittest when running under MinGW.
2429 6. ESC_K was left out of the EBCDIC table.
2431 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2432 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2433 limit also applies to "virtual nesting" when a pattern is recursive, and in
2434 this case 1000 isn't so big. I have been able to remove this limit at the
2435 expense of backing off one optimization in certain circumstances. Normally,
2436 when pcre_exec() would call its internal match() function recursively and
2437 immediately return the result unconditionally, it uses a "tail recursion"
2438 feature to save stack. However, when a subpattern that can match an empty
2439 string has an unlimited repetition quantifier, it no longer makes this
2440 optimization. That gives it a stack frame in which to save the data for
2441 checking that an empty string has been matched. Previously this was taken
2442 from the 1000-entry workspace that had been reserved. So now there is no
2443 explicit limit, but more stack is used.
2445 8. Applied Daniel's patches to solve problems with the import/export magic
2446 syntax that is required for Windows, and which was going wrong for the
2447 pcreposix and pcrecpp parts of the library. These were overlooked when this
2448 problem was solved for the main library.
2450 9. There were some crude static tests to avoid integer overflow when computing
2451 the size of patterns that contain repeated groups with explicit upper
2452 limits. As the maximum quantifier is 65535, the maximum group length was
2453 set at 30,000 so that the product of these two numbers did not overflow a
2454 32-bit integer. However, it turns out that people want to use groups that
2455 are longer than 30,000 bytes (though not repeat them that many times).
2456 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2457 made it possible to implement the integer overflow checks in a much more
2458 dynamic way, which I have now done. The artificial limitation on group
2459 length has been removed - we now have only the limit on the total length of
2460 the compiled pattern, which depends on the LINK_SIZE setting.
2462 10. Fixed a bug in the documentation for get/copy named substring when
2463 duplicate names are permitted. If none of the named substrings are set, the
2464 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2465 empty string.
2467 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2468 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2469 because the ] is interpreted as the first data character and the
2470 terminating ] is not found. PCRE has been made compatible with Perl in this
2471 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2472 cause memory overwriting.
2474 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2475 string has been matched (to stop an infinite loop). It was not recognizing
2476 a conditional subpattern that could match an empty string if that
2477 subpattern was within another subpattern. For example, it looped when
2478 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2479 condition was not nested. This bug has been fixed.
2481 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2482 past the start of the subject in the presence of bytes with the top bit
2483 set, for example "\x8aBCD".
2485 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2486 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2488 14. Optimized (?!) to (*FAIL).
2490 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2491 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2492 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2493 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2494 does: it's just the validity check that is more restrictive.
2496 16. Inserted checks for integer overflows during escape sequence (backslash)
2497 processing, and also fixed erroneous offset values for syntax errors during
2498 backslash processing.
2500 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2501 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2503 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2504 caused an overrun.
2506 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2507 something other than just ASCII characters) inside a group that had an
2508 unlimited repeat caused a loop at compile time (while checking to see
2509 whether the group could match an empty string).
2511 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2512 [\P{Any}] did so. (Error in the code for printing property names.)
2514 21. An orphan \E inside a character class could cause a crash.
2516 22. A repeated capturing bracket such as (A)? could cause a wild memory
2517 reference during compilation.
2519 23. There are several functions in pcre_compile() that scan along a compiled
2520 expression for various reasons (e.g. to see if it's fixed length for look
2521 behind). There were bugs in these functions when a repeated \p or \P was
2522 present in the pattern. These operators have additional parameters compared
2523 with \d, etc, and these were not being taken into account when moving along
2524 the compiled data. Specifically:
2526 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2527 length.
2529 (b) An item such as \pL+ within a repeated group could cause crashes or
2530 loops.
2532 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2533 "reference to non-existent subpattern" error.
2535 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2537 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2538 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2540 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2542 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2543 character were causing crashes (broken optimization).
2545 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2546 \p or \P) caused a compile-time loop.
2548 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2549 break. For example, the unstudied pattern [\r\n]A does not match the string
2550 "\r\nA" because change 7.0/46 below moves the current point on by two
2551 characters after failing to match at the start. However, the pattern \nA
2552 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2553 the same is true. There doesn't seem any very clean way out of this, but
2554 what I have chosen to do makes the common cases work: PCRE now takes note
2555 of whether there can be an explicit match for \r or \n anywhere in the
2556 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2557 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2558 pattern has explicit CR or LF references.
2560 29. Added (*CR) etc for changing newline setting at start of pattern.
2563 Version 7.2 19-Jun-07
2564 ---------------------
2566 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2567 which is apparently normally available under Windows.
2569 2. Re-jig the pcregrep tests with different newline settings in an attempt
2570 to make them independent of the local environment's newline setting.
2572 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2574 4. Some of the "internals" tests were previously cut out when the link size
2575 was not 2, because the output contained actual offsets. The recent new
2576 "Z" feature of pcretest means that these can be cut out, making the tests
2577 usable with all link sizes.
2579 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2580 stack recursion. This gives a massive performance boost under BSD, but just
2581 a small improvement under Linux. However, it saves one field in the frame
2582 in all cases.
2584 6. Added more features from the forthcoming Perl 5.10:
2586 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2587 recursion call. It refers to the nth most recently opened parentheses.
2589 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2590 to be opened parentheses.
2592 (c) Conditions that refer to capturing parentheses can be specified
2593 relatively, for example, (?(-2)... or (?(+3)...
2595 (d) \K resets the start of the current match so that everything before
2596 is not part of it.
2598 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2600 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2601 reference syntax.
2603 (g) (?| introduces a group in which the numbering of parentheses in each
2604 alternative starts with the same number.
2606 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2608 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2611 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2612 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2613 for detecting groups that can match an empty string.
2615 9. A pattern with a very large number of alternatives (more than several
2616 hundred) was running out of internal workspace during the pre-compile
2617 phase, where pcre_compile() figures out how much memory will be needed. A
2618 bit of new cunning has reduced the workspace needed for groups with
2619 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2620 workspace instead of running out of the 4096 that are available.
2622 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2624 11. Applied patch from Google to remove an optimization that didn't quite work.
2625 The report of the bug said:
2627 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2628 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2629 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2631 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2632 it matched the wrong number of bytes.
2635 Version 7.1 24-Apr-07
2636 ---------------------
2638 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2639 that is more "standard", making use of automake and other Autotools. There
2640 is some re-arrangement of the files and adjustment of comments consequent
2641 on this.
2643 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2644 for recursive directory scanning broke on some systems because the files
2645 are not scanned in any specific order and on different systems the order
2646 was different. A call to "sort" has been inserted into RunGrepTest for the
2647 approprate test as a short-term fix. In the longer term there may be an
2648 alternative.
2650 3. I had an email from Eric Raymond about problems translating some of PCRE's
2651 man pages to HTML (despite the fact that I distribute HTML pages, some
2652 people do their own conversions for various reasons). The problems
2653 concerned the use of low-level troff macros .br and .in. I have therefore
2654 removed all such uses from the man pages (some were redundant, some could
2655 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2656 HTML has been updated to handle .nf/.fi and to complain if it encounters
2657 .br or .in.
2659 4. Updated comments in configure.ac that get placed in config.h.in and also
2660 arranged for config.h to be included in the distribution, with the name
2661 config.h.generic, for the benefit of those who have to compile without
2662 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2664 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2665 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2666 makevp.bat for the current PCRE, using the additional files
2667 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2669 6. A Windows user reported a minor discrepancy with test 2, which turned out
2670 to be caused by a trailing space on an input line that had got lost in his
2671 copy. The trailing space was an accident, so I've just removed it.
2673 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2674 that is needed.
2676 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2677 as "const" (a) because they are and (b) because it helps the PHP
2678 maintainers who have recently made a script to detect big data structures
2679 in the php code that should be moved to the .rodata section. I remembered
2680 to update Builducptable as well, so it won't revert if ucptable.h is ever
2681 re-created.
2683 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2684 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2685 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2686 support is not required. This saves 1.5-2K of code, which is important in
2687 some applications.
2689 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2690 so as not to refer to the tables, even though these functions will never be
2691 called when UTF-8 support is disabled. Otherwise there are problems with a
2692 shared library.
2694 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2696 (a) It was defining its arguments as char * instead of void *.
2698 (b) It was assuming that all moves were upwards in memory; this was true
2699 a long time ago when I wrote it, but is no longer the case.
2701 The emulated memove() is provided for those environments that have neither
2702 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2703 is clearly not the case, as these two bugs were recently reported.
2705 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2706 and Detrail to create the HTML documentation, the .txt form of the man
2707 pages, and it removes trailing spaces from listed files. It also creates
2708 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2709 case, it wraps all the #defines with #ifndefs. This script should be run
2710 before "make dist".
2712 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2713 with Unicode property support.
2715 (a) For a maximizing quantifier, if the two different cases of the
2716 character were of different lengths in their UTF-8 codings (there are
2717 some cases like this - I found 11), and the matching function had to
2718 back up over a mixture of the two cases, it incorrectly assumed they
2719 were both the same length.
2721 (b) When PCRE was configured to use the heap rather than the stack for
2722 recursion during matching, it was not correctly preserving the data for
2723 the other case of a UTF-8 character when checking ahead for a match
2724 while processing a minimizing repeat. If the check also involved
2725 matching a wide character, but failed, corruption could cause an
2726 erroneous result when trying to check for a repeat of the original
2727 character.
2729 13. Some tidying changes to the testing mechanism:
2731 (a) The RunTest script now detects the internal link size and whether there
2732 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2733 values substituted by "configure". (The RunGrepTest script already did
2734 this for UTF-8.) The configure.ac script no longer substitutes the
2735 relevant variables.
2737 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2738 with length and offset values. This means that the output is different
2739 for different internal link sizes. Test 2 is skipped for link sizes
2740 other than 2 because of this, bypassing the problem. Unfortunately,
2741 there was also a test in test 3 (the locale tests) that used /B and
2742 failed for link sizes other than 2. Rather than cut the whole test out,
2743 I have added a new /Z option to pcretest that replaces the length and
2744 offset values with spaces. This is now used to make test 3 independent
2745 of link size. (Test 2 will be tidied up later.)
2747 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2748 segmentation fault instead of returning the appropriate error message.
2750 15. In multiline mode when the newline sequence was set to "any", the pattern
2751 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2752 This doesn't seem right; it now treats the CRLF combination as the line
2753 ending, and so does not match in that case. It's only a pattern such as ^$
2754 that would hit this one: something like ^ABC$ would have failed after \r
2755 and then tried again after \r\n.
2757 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2758 in an attempt to make files that differ only in their line terminators
2759 compare equal. This works on Linux.
2761 17. Under certain error circumstances pcregrep might try to free random memory
2762 as it exited. This is now fixed, thanks to valgrind.
2764 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2765 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2766 was because its rules for how to advance for /g after matching an empty
2767 string at the end of a line did not allow for this case. They now check for
2768 it specially.
2770 20. pcretest is supposed to handle patterns and data of any length, by
2771 extending its buffers when necessary. It was getting this wrong when the
2772 buffer for a data line had to be extended.
2774 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2775 CRLF as a newline sequence.
2777 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2778 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2779 I have nevertheless tidied it up.
2781 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2783 24. Added a man page for pcre-config.
2786 Version 7.0 19-Dec-06
2787 ---------------------
2789 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2790 moving to gcc 4.1.1.
2792 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2793 sys/time.h, which is documented as needed for this function. It doesn't
2794 seem to matter on Linux, but it showed up on some releases of OS X.
2796 3. It seems that there are systems where bytes whose values are greater than
2797 127 match isprint() in the "C" locale. The "C" locale should be the
2798 default when a C program starts up. In most systems, only ASCII printing
2799 characters match isprint(). This difference caused the output from pcretest
2800 to vary, making some of the tests fail. I have changed pcretest so that:
2802 (a) When it is outputting text in the compiled version of a pattern, bytes
2803 other than 32-126 are always shown as hex escapes.
2805 (b) When it is outputting text that is a matched part of a subject string,
2806 it does the same, unless a different locale has been set for the match
2807 (using the /L modifier). In this case, it uses isprint() to decide.
2809 4. Fixed a major bug that caused incorrect computation of the amount of memory
2810 required for a compiled pattern when options that changed within the
2811 pattern affected the logic of the preliminary scan that determines the
2812 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2813 that the computed length was too small. The symptoms of this bug were
2814 either the PCRE error "internal error: code overflow" from pcre_compile(),
2815 or a glibc crash with a message such as "pcretest: free(): invalid next
2816 size (fast)". Examples of patterns that provoked this bug (shown in
2817 pcretest format) are:
2819 /(?-x: )/x
2820 /(?x)(?-x: \s*#\s*)/
2821 /((?i)[\x{c0}])/8
2822 /(?i:[\x{c0}])/8
2824 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2825 is now done differently.
2827 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2828 wrapper classes; (b) implement a new function in the C++ scanner that is
2829 more efficient than the old way of doing things because it avoids levels of
2830 recursion in the regex matching; (c) add a paragraph to the documentation
2831 for the FullMatch() function.
2833 6. The escape sequence \n was being treated as whatever was defined as
2834 "newline". Not only was this contrary to the documentation, which states
2835 that \n is character 10 (hex 0A), but it also went horribly wrong when
2836 "newline" was defined as CRLF. This has been fixed.
2838 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2839 was being set to -1 for the "end of line" case (supposedly a value that no
2840 character can have). Though this value is never used (the check for end of
2841 line is "zero bytes in current character"), it caused compiler complaints.
2842 I've changed it to 0xffffffff.
2844 8. In pcre_version.c, the version string was being built by a sequence of
2845 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2846 string (as it is for production releases) called a macro with an empty
2847 argument. The C standard says the result of this is undefined. The gcc
2848 compiler treats it as an empty string (which was what was wanted) but it is
2849 reported that Visual C gives an error. The source has been hacked around to
2850 avoid this problem.
2852 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2853 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2854 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2855 of them did).
2857 10. Originally, pcretest opened its input and output without "b"; then I was
2858 told that "b" was needed in some environments, so it was added for release
2859 5.0 to both the input and output. (It makes no difference on Unix-like
2860 systems.) Later I was told that it is wrong for the input on Windows. I've
2861 now abstracted the modes into two macros, to make it easier to fiddle with
2862 them, and removed "b" from the input mode under Windows.
2864 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2866 12. Added -help and --help to pcretest as an official way of being reminded
2867 of the options.
2869 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2870 and pcrecpp.cc because they annoy compilers at high warning levels.
2872 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2874 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2875 scripts are not C programs :-) and which was not noticed because it works
2876 on Linux.
2878 16. pcretest is supposed to handle any length of pattern and data line (as one
2879 line or as a continued sequence of lines) by extending its input buffer if
2880 necessary. This feature was broken for very long pattern lines, leading to
2881 a string of junk being passed to pcre_compile() if the pattern was longer
2882 than about 50K.
2884 17. I have done a major re-factoring of the way pcre_compile() computes the
2885 amount of memory needed for a compiled pattern. Previously, there was code
2886 that made a preliminary scan of the pattern in order to do this. That was
2887 OK when PCRE was new, but as the facilities have expanded, it has become
2888 harder and harder to keep it in step with the real compile phase, and there
2889 have been a number of bugs (see for example, 4 above). I have now found a
2890 cunning way of running the real compile function in a "fake" mode that
2891 enables it to compute how much memory it would need, while actually only
2892 ever using a few hundred bytes of working memory and without too many
2893 tests of the mode. This should make future maintenance and development
2894 easier. A side effect of this work is that the limit of 200 on the nesting
2895 depth of parentheses has been removed (though this was never a serious
2896 limitation, I suspect). However, there is a downside: pcre_compile() now
2897 runs more slowly than before (30% or more, depending on the pattern). I
2898 hope this isn't a big issue. There is no effect on runtime performance.
2900 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2901 newline (only possible for the last line of a file) and it was a
2902 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2904 19. Added additional timing features to pcretest. (1) The -tm option now times
2905 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2906 separate command line item, by a number that specifies the number of
2907 repeats to use when timing. The default is 50000; this gives better
2908 precision, but takes uncomfortably long for very large patterns.
2910 20. Extended pcre_study() to be more clever in cases where a branch of a
2911 subpattern has no definite first character. For example, (a*|b*)[cd] would
2912 previously give no result from pcre_study(). Now it recognizes that the
2913 first character must be a, b, c, or d.
2915 21. There was an incorrect error "recursive call could loop indefinitely" if
2916 a subpattern (or the entire pattern) that was being tested for matching an
2917 empty string contained only one non-empty item after a nested subpattern.
2918 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2919 incorrectly, because the \d was being skipped in the check.
2921 22. The pcretest program now has a new pattern option /B and a command line
2922 option -b, which is equivalent to adding /B to every pattern. This causes
2923 it to show the compiled bytecode, without the additional information that
2924 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2925 is the same as /B/I).
2927 23. A new optimization is now able automatically to treat some sequences such
2928 as a*b as a*+b. More specifically, if something simple (such as a character
2929 or a simple class like \d) has an unlimited quantifier, and is followed by
2930 something that cannot possibly match the quantified thing, the quantifier
2931 is automatically "possessified".
2933 24. A recursive reference to a subpattern whose number was greater than 39
2934 went wrong under certain circumstances in UTF-8 mode. This bug could also
2935 have affected the operation of pcre_study().
2937 25. Realized that a little bit of performance could be had by replacing
2938 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2940 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2942 27. Possessive quantifiers such as a++ were previously implemented by turning
2943 them into atomic groups such as ($>a+). Now they have their own opcodes,
2944 which improves performance. This includes the automatically created ones
2945 from 23 above.
2947 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2948 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2949 the first matched character to be a colon. This applied both to named and
2950 numbered groups.
2952 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2954 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2955 building PCRE on Windows easier, so I have included it in the distribution.
2957 31. There is now a check in pcretest against a ridiculously large number being
2958 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2959 loop, the loop is abandoned.
2961 32. Forward references to subpatterns in conditions such as (?(2)...) where
2962 subpattern 2 is defined later cause pcre_compile() to search forwards in
2963 the pattern for the relevant set of parentheses. This search went wrong
2964 when there were unescaped parentheses in a character class, parentheses
2965 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2967 33. "Subroutine" calls and backreferences were previously restricted to
2968 referencing subpatterns earlier in the regex. This restriction has now
2969 been removed.
2971 34. Added a number of extra features that are going to be in Perl 5.10. On the
2972 whole, these are just syntactic alternatives for features that PCRE had
2973 previously implemented using the Python syntax or my own invention. The
2974 other formats are all retained for compatibility.
2976 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2977 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2978 also .NET compatible.
2980 (b) A recursion or subroutine call to a named group can now be defined as
2981 (?&name) as well as (?P>name).
2983 (c) A backreference to a named group can now be defined as \k<name> or
2984 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2985 5.10, are also .NET compatible.
2987 (d) A conditional reference to a named group can now use the syntax
2988 (?(<name>) or (?('name') as well as (?(name).
2990 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2991 groups (named and numbered) that are never evaluated inline, but can be
2992 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2993 is always false. There may be only one alternative in such a group.
2995 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2996 as the simple (?(R). The condition is true only if the most recent
2997 recursion is that of the given number or name. It does not search out
2998 through the entire recursion stack.
3000 (g) The escape \gN or \g{N} has been added, where N is a positive or
3001 negative number, specifying an absolute or relative reference.
3003 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3004 some "unreachable code" warnings.
3006 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3007 things, this adds five new scripts.
3009 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3010 There were also incompatibilities regarding the handling of \Q..\E inside
3011 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3012 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3014 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3015 matches an empty string, and forcibly breaks the loop. There were bugs in
3016 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3017 against aaaa the result was just "a" rather than "aaaa", for example. Two
3018 separate and independent bugs (that affected different cases) have been
3019 fixed.
3021 39. Refactored the code to abolish the use of different opcodes for small
3022 capturing bracket numbers. This is a tidy that I avoided doing when I
3023 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3024 The new approach is not only tidier, it makes it possible to reduce the
3025 memory needed to fix the previous bug (38).
3027 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3028 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3029 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3030 mode.
3032 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3033 report.
3035 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3036 copy construction and assignment in the C++ wrapper.
3038 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3039 couple of bugs that could have given wrong results in the "--newline=crlf"
3040 case.
3042 44. Added a number of casts and did some reorganization of signed/unsigned int
3043 variables following suggestions from Dair Grant. Also renamed the variable
3044 "this" as "item" because it is a C++ keyword.
3046 45. Arranged for dftables to add
3048 #include "pcre_internal.h"
3050 to pcre_chartables.c because without it, gcc 4.x may remove the array
3051 definition from the final binary if PCRE is built into a static library and
3052 dead code stripping is activated.
3054 46. For an unanchored pattern, if a match attempt fails at the start of a
3055 newline sequence, and the newline setting is CRLF or ANY, and the next two
3056 characters are CRLF, advance by two characters instead of one.
3059 Version 6.7 04-Jul-06
3060 ---------------------
3062 1. In order to handle tests when input lines are enormously long, pcretest has
3063 been re-factored so that it automatically extends its buffers when
3064 necessary. The code is crude, but this _is_ just a test program. The
3065 default size has been increased from 32K to 50K.
3067 2. The code in pcre_study() was using the value of the re argument before
3068 testing it for NULL. (Of course, in any sensible call of the function, it
3069 won't be NULL.)
3071 3. The memmove() emulation function in pcre_internal.h, which is used on
3072 systems that lack both memmove() and bcopy() - that is, hardly ever -
3073 was missing a "static" storage class specifier.
3075 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3076 containing an extended class (one that cannot be represented by a bitmap
3077 because it contains high-valued characters or Unicode property items, e.g.
3078 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3079 pattern, but PCRE should not loop if you do not (it no longer does).
3080 [Detail: two cases were found: (a) a repeated subpattern containing an
3081 extended class; (b) a recursive reference to a subpattern that followed a
3082 previous extended class. It wasn't skipping over the extended class
3083 correctly when UTF-8 mode was not set.]
3085 5. A negated single-character class was not being recognized as fixed-length
3086 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3087 compile error "lookbehind assertion is not fixed length".
3089 6. The RunPerlTest auxiliary script was showing an unexpected difference
3090 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3091 write a Perl script that can interpret lines of an input file either as
3092 byte characters or as UTF-8, which is what "perltest" was being required to
3093 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3094 can't do is switch easily at run time between having the "use utf8;" pragma
3095 or not. In the end, I fudged it by using the RunPerlTest script to insert
3096 "use utf8;" explicitly for the UTF-8 tests.
3098 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3099 the end of the subject string, contrary to the documentation and to what
3100 Perl does. This was true of both matching functions. Now it matches only at
3101 the start of the subject and immediately after *internal* newlines.
3103 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3104 a pointer to an int instead of a pointer to an unsigned long int. This
3105 caused problems on 64-bit systems.
3107 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3108 instance of the 'standard' template library not being so standard".
3110 10. There was no check on the number of named subpatterns nor the maximum
3111 length of a subpattern name. The product of these values is used to compute
3112 the size of the memory block for a compiled pattern. By supplying a very
3113 long subpattern name and a large number of named subpatterns, the size
3114 computation could be caused to overflow. This is now prevented by limiting
3115 the length of names to 32 characters, and the number of named subpatterns
3116 to 10,000.
3118 11. Subpatterns that are repeated with specific counts have to be replicated in
3119 the compiled pattern. The size of memory for this was computed from the
3120 length of the subpattern and the repeat count. The latter is limited to
3121 65535, but there was no limit on the former, meaning that integer overflow
3122 could in principle occur. The compiled length of a repeated subpattern is
3123 now limited to 30,000 bytes in order to prevent this.
3125 12. Added the optional facility to have named substrings with the same name.
3127 13. Added the ability to use a named substring as a condition, using the
3128 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3129 are numbers (not recommended). Forward references are permitted.
3131 14. Added forward references in named backreferences (if you see what I mean).
3133 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3134 pattern could run off the end of the subject. For example, the pattern
3135 "(?s)(.{1,5})"8 did this with the subject "ab".
3137 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3138 PCRE_CASELESS was set when matching characters that were quantified with ?
3139 or *.
3141 17. A character class other than a single negated character that had a minimum
3142 but no maximum quantifier - for example [ab]{6,} - was not handled
3143 correctly by pce_dfa_exec(). It would match only one character.
3145 18. A valid (though odd) pattern that looked like a POSIX character
3146 class but used an invalid character after [ (for example [[,abc,]]) caused
3147 pcre_compile() to give the error "Failed: internal error: code overflow" or
3148 in some cases to crash with a glibc free() error. This could even happen if
3149 the pattern terminated after [[ but there just happened to be a sequence of
3150 letters, a binary zero, and a closing ] in the memory that followed.
3152 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3153 over the years. Originally (before any Unicode support), just the bottom 8
3154 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3155 output from "man perlunicode" includes this:
3157 The regular expression compiler produces polymorphic opcodes. That
3158 is, the pattern adapts to the data and automatically switches to
3159 the Unicode character scheme when presented with Unicode data--or
3160 instead uses a traditional byte scheme when presented with byte
3161 data.
3163 Sadly, a wide octal escape does not cause a switch, and in a string with
3164 no other multibyte characters, these octal escapes are treated as before.
3165 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3166 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3167 Unicode string.
3169 I have not perpetrated such confusion in PCRE. Up till now, it took just
3170 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3171 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3172 translate to the appropriate multibyte character.
3174 29. Applied some refactoring to reduce the number of warnings from Microsoft
3175 and Borland compilers. This has included removing the fudge introduced
3176 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3177 a warning about an unused variable.
3179 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3180 characters since release 4.0, because Perl (from release 5.004) does not.
3181 [Or at least, is documented not to: some releases seem to be in conflict
3182 with the documentation.] However, when a pattern was studied with
3183 pcre_study() and all its branches started with \s, PCRE still included VT
3184 as a possible starting character. Of course, this did no harm; it just
3185 caused an unnecessary match attempt.
3187 22. Removed a now-redundant internal flag bit that recorded the fact that case
3188 dependency changed within the pattern. This was once needed for "required
3189 byte" processing, but is no longer used. This recovers a now-scarce options
3190 bit. Also moved the least significant internal flag bit to the most-
3191 significant bit of the word, which was not previously used (hangover from
3192 the days when it was an int rather than a uint) to free up another bit for
3193 the future.
3195 23. Added support for CRLF line endings as well as CR and LF. As well as the
3196 default being selectable at build time, it can now be changed at runtime
3197 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3198 specify that it is scanning data with non-default line endings.
3200 24. Changed the definition of CXXLINK to make it agree with the definition of
3201 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3203 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3204 recursions. This makes a big different to stack usage for some patterns.
3206 26. If a subpattern containing a named recursion or subroutine reference such
3207 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3208 the space required for the compiled pattern went wrong and gave too small a
3209 value. Depending on the environment, this could lead to "Failed: internal
3210 error: code overflow at offset 49" or "glibc detected double free or
3211 corruption" errors.
3213 27. Applied patches from Google (a) to support the new newline modes and (b) to
3214 advance over multibyte UTF-8 characters in GlobalReplace.
3216 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3217 difference for some implementation of PCRE in some Windows version.
3219 29. Added some extra testing facilities to pcretest:
3221 \q<number> in a data line sets the "match limit" value
3222 \Q<number> in a data line sets the "match recursion limt" value
3223 -S <number> sets the stack size, where <number> is in megabytes
3225 The -S option isn't available for Windows.
3228 Version 6.6 06-Feb-06
3229 ---------------------
3231 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3232 in pcreposix.h. I have copied the definition from pcre.h.
3234 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3235 because pcre.h is no longer a built file.
3237 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3238 not normally included in the compiled code.
3241 Version 6.5 01-Feb-06
3242 ---------------------
3244 1. When using the partial match feature with pcre_dfa_exec(), it was not
3245 anchoring the second and subsequent partial matches at the new starting
3246 point. This could lead to incorrect results. For example, with the pattern
3247 /1234/, partially matching against "123" and then "a4" gave a match.
3249 2. Changes to pcregrep:
3251 (a) All non-match returns from pcre_exec() were being treated as failures
3252 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3253 error message is output. Some extra information is given for the
3255 probably the only errors that are likely to be caused by users (by
3256 specifying a regex that has nested indefinite repeats, for instance).
3257 If there are more than 20 of these errors, pcregrep is abandoned.
3259 (b) A binary zero was treated as data while matching, but terminated the
3260 output line if it was written out. This has been fixed: binary zeroes
3261 are now no different to any other data bytes.
3263 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3264 used to set a locale for matching. The --locale=xxxx long option has
3265 been added (no short equivalent) to specify a locale explicitly on the
3266 pcregrep command, overriding the environment variables.
3268 (d) When -B was used with -n, some line numbers in the output were one less
3269 than they should have been.
3271 (e) Added the -o (--only-matching) option.
3273 (f) If -A or -C was used with -c (count only), some lines of context were
3274 accidentally printed for the final match.
3276 (g) Added the -H (--with-filename) option.
3278 (h) The combination of options -rh failed to suppress file names for files
3279 that were found from directory arguments.
3281 (i) Added the -D (--devices) and -d (--directories) options.
3283 (j) Added the -F (--fixed-strings) option.
3285 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3287 (l) Added the --colo(u)r option.
3289 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3290 is not present by default.
3292 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3293 items such as (?R) or (?1), when the recursion could match a number of
3294 alternatives. If it matched one of the alternatives, but subsequently,
3295 outside the recursion, there was a failure, the code tried to back up into
3296 the recursion. However, because of the way PCRE is implemented, this is not
3297 possible, and the result was an incorrect result from the match.
3299 In order to prevent this happening, the specification of recursion has
3300 been changed so that all such subpatterns are automatically treated as
3301 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3303 4. I had overlooked the fact that, in some locales, there are characters for
3304 which isalpha() is true but neither isupper() nor islower() are true. In
3305 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3306 and ordfeminine) are like this. This affected the treatment of \w and \W
3307 when they appeared in character classes, but not when they appeared outside
3308 a character class. The bit map for "word" characters is now created
3309 separately from the results of isalnum() instead of just taking it from the
3310 upper, lower, and digit maps. (Plus the underscore character, of course.)
3312 5. The above bug also affected the handling of POSIX character classes such as
3313 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3314 permanent tables. Instead, the bit maps for such a class were previously
3315 created as the appropriate unions of the upper, lower, and digit bitmaps.
3316 Now they are created by subtraction from the [[:word:]] class, which has
3317 its own bitmap.
3319 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3320 It is created by subtracting the vertical space characters (\x09, \x0a,
3321 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3322 subtraction was done in the overall bitmap for a character class, meaning
3323 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3324 be recognized. This bug has been fixed.
3326 7. Patches from the folks at Google:
3328 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3329 real life, but is still worth protecting against".
3331 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3332 regular expressions".
3334 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3335 have it.
3337 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3338 "configure" and the latter not, in order to fix a problem somebody had
3339 with compiling the Arg class on HP-UX.
3341 (e) Improve the error-handling of the C++ wrapper a little bit.
3343 (f) New tests for checking recursion limiting.
3345 8. The pcre_memmove() function, which is used only if the environment does not
3346 have a standard memmove() function (and is therefore rarely compiled),
3347 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3348 returning a result (though PCRE never actually uses the result).
3350 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3351 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3352 returned instead of calling malloc() with an overflowing number that would
3353 most likely cause subsequent chaos.
3355 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3357 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3358 with this option is matched, the nmatch and pmatch options of regexec() are
3359 ignored.
3361 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3362 provided in case anyone wants to the the POSIX interface with UTF-8
3363 strings.
3365 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3366 C++ linking (needed for some HP-UX environments).
3368 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3369 (unused parameter) and in the pcre_printint() function (omitted "default"
3370 switch label when the default is to do nothing).
3372 15. Added some code to make it possible, when PCRE is compiled as a C++
3373 library, to replace subject pointers for pcre_exec() with a smart pointer
3374 class, thus making it possible to process discontinuous strings.
3376 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3377 much the same function. They were added by different people who were trying
3378 to make PCRE easy to compile on non-Unix systems. It has been suggested
3379 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3380 for compiling on Windows systems. I have therefore replaced it with
3381 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3382 defaults to "extern" for C or "extern C" for C++, which works fine on
3383 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3384 SCOPE with something explicit in config.h. In addition:
3386 (a) pcreposix.h still had just "extern" instead of either of these macros;
3387 I have replaced it with PCRE_DATA_SCOPE.
3389 (b) Functions such as _pcre_xclass(), which are internal to the library,
3390 but external in the C sense, all had PCRE_EXPORT in their definitions.
3391 This is apparently wrong for the Windows case, so I have removed it.
3392 (It makes no difference on Unix-like systems.)
3394 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3395 of recursive calls to match(). This is different to MATCH_LIMIT because
3396 that limits the total number of calls to match(), not all of which increase
3397 the depth of recursion. Limiting the recursion depth limits the amount of
3398 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3399 when PCRE is compiled, and changed at run time. A patch from Google adds
3400 this functionality to the C++ interface.
3402 18. Changes to the handling of Unicode character properties:
3404 (a) Updated the table to Unicode 4.1.0.
3406 (b) Recognize characters that are not in the table as "Cn" (undefined).
3408 (c) I revised the way the table is implemented to a much improved format
3409 which includes recognition of ranges. It now supports the ranges that
3410 are defined in UnicodeData.txt, and it also amalgamates other
3411 characters into ranges. This has reduced the number of entries in the
3412 table from around 16,000 to around 3,000, thus reducing its size
3413 considerably. I realized I did not need to use a tree structure after
3414 all - a binary chop search is just as efficient. Having reduced the
3415 number of entries, I extended their size from 6 bytes to 8 bytes to
3416 allow for more data.
3418 (d) Added support for Unicode script names via properties such as \p{Han}.
3420 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3421 matching that character.
3423 20. When matching a repeated Unicode property with a minimum greater than zero,
3424 (for example \pL{2,}), PCRE could look past the end of the subject if it
3425 reached it while seeking the minimum number of characters. This could
3426 happen only if some of the characters were more than one byte long, because
3427 there is a check for at least the minimum number of bytes.
3429 21. Refactored the implementation of \p and \P so as to be more general, to
3430 allow for more different types of property in future. This has changed the
3431 compiled form incompatibly. Anybody with saved compiled patterns that use
3432 \p or \P will have to recompile them.
3434 22. Added "Any" and "L&" to the supported property types.
3436 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3437 but give a compile time error if the value is greater than 0xff.
3439 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3440 accidentally not being installed or uninstalled.
3442 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3443 made were to insert the current release number. This seemed silly, because
3444 it made things harder for people building PCRE on systems that don't run
3445 "configure". I have turned pcre.h into a distributed file, no longer built
3446 by "configure", with the version identification directly included. There is
3447 no longer a pcre.h.in file.
3449 However, this change necessitated a change to the pcre-config script as
3450 well. It is built from pcre-config.in, and one of the substitutions was the
3451 release number. I have updated configure.ac so that ./configure now finds
3452 the release number by grepping pcre.h.
3454 26. Added the ability to run the tests under valgrind.
3457 Version 6.4 05-Sep-05
3458 ---------------------
3460 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3461 "--" to be printed when multiple files were scanned, even when none of the
3462 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3463 consider it to be a bug, and have restored the previous behaviour.
3465 2. A couple of code tidies to get rid of compiler warnings.
3467 3. The pcretest program used to cheat by referring to symbols in the library
3468 whose names begin with _pcre_. These are internal symbols that are not
3469 really supposed to be visible externally, and in some environments it is
3470 possible to suppress them. The cheating is now confined to including
3471 certain files from the library's source, which is a bit cleaner.
3473 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3474 file's purpose clearer.
3476 5. Reorganized pcre_ucp_findchar().
3479 Version 6.3 15-Aug-05
3480 ---------------------
3482 1. The file libpcre.pc.in did not have general read permission in the tarball.
3484 2. There were some problems when building without C++ support:
3486 (a) If C++ support was not built, "make install" and "make test" still
3487 tried to test it.
3489 (b) There were problems when the value of CXX was explicitly set. Some
3490 changes have been made to try to fix these, and ...
3492 (c) --disable-cpp can now be used to explicitly disable C++ support.
3494 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3495 backslash in a target when C++ was disabled. This confuses some
3496 versions of "make", apparently. Using an intermediate variable solves
3497 this. (Same for CPP_LOBJ.)
3499 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3500 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3501 necessary on certain architectures.
3503 4. Added a setting of -export-symbols-regex to the link command to remove
3504 those symbols that are exported in the C sense, but actually are local
3505 within the library, and not documented. Their names all begin with
3506 "_pcre_". This is not a perfect job, because (a) we have to except some
3507 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3508 available (and never for static libraries). I have made a note to try to
3509 find a way round (a) in the future.
3512 Version 6.2 01-Aug-05
3513 ---------------------
3515 1. There was no test for integer overflow of quantifier values. A construction
3516 such as {1111111111111111} would give undefined results. What is worse, if
3517 a minimum quantifier for a parenthesized subpattern overflowed and became
3518 negative, the calculation of the memory size went wrong. This could have
3519 led to memory overwriting.
3521 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3523 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3524 operating environments where this matters.
3526 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3527 PCRE options from within the C++ wrapper.
3529 5. Named capturing subpatterns were not being correctly counted when a pattern
3530 was compiled. This caused two problems: (a) If there were more than 100
3531 such subpatterns, the calculation of the memory needed for the whole
3532 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3533 back references of the form \12, where the number was greater than 9, were
3534 not recognized as back references, even though there were sufficient
3535 previous subpatterns.
3537 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3538 versions of gcc, e.g. 2.95.4.
3541 Version 6.1 21-Jun-05
3542 ---------------------
3544 1. There was one reference to the variable "posix" in pcretest.c that was not
3545 surrounded by "#if !defined NOPOSIX".
3547 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3548 the cross-check on the old pcre_info() function, for the benefit of the
3549 cut-down version of PCRE that is currently imported into Exim.
3551 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3552 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3553 patterns that start (?i) but I don't think that matters. The use of (?i) is
3554 just an example; this all applies to the other options as well.
3556 4. Since libtool seems to echo the compile commands it is issuing, the output
3557 from "make" can be reduced a bit by putting "@" in front of each libtool
3558 compile command.
3560 5. Patch from the folks at Google for configure.in to be a bit more thorough
3561 in checking for a suitable C++ installation before trying to compile the
3562 C++ stuff. This should fix a reported problem when a compiler was present,
3563 but no suitable headers.
3565 6. The man pages all had just "PCRE" as their title. I have changed them to
3566 be the relevant file name. I have also arranged that these names are
3567 retained in the file doc/pcre.txt, which is a concatenation in text format
3568 of all the man pages except the little individual ones for each function.
3570 7. The NON-UNIX-USE file had not been updated for the different set of source
3571 files that come with release 6. I also added a few comments about the C++
3572 wrapper.
3575 Version 6.0 07-Jun-05
3576 ---------------------
3578 1. Some minor internal re-organization to help with my DFA experiments.
3580 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3581 didn't matter for the library itself when fully configured, but did matter
3582 when compiling without UCP support, or within Exim, where the ucp files are
3583 not imported.
3585 3. Refactoring of the library code to split up the various functions into
3586 different source modules. The addition of the new DFA matching code (see
3587 below) to a single monolithic source would have made it really too
3588 unwieldy, quite apart from causing all the code to be include in a
3589 statically linked application, when only some functions are used. This is
3590 relevant even without the DFA addition now that patterns can be compiled in
3591 one application and matched in another.
3593 The downside of splitting up is that there have to be some external
3594 functions and data tables that are used internally in different modules of
3595 the library but which are not part of the API. These have all had their
3596 names changed to start with "_pcre_" so that they are unlikely to clash
3597 with other external names.
3599 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3600 a different (DFA) algorithm. Although it is slower than the original
3601 function, it does have some advantages for certain types of matching
3602 problem.
3604 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3605 including restarting after a partial match.
3607 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3608 defined when compiling for Windows was sent to me. I have put it into the
3609 code, though I have no means of testing or verifying it.
3611 7. Added the pcre_refcount() auxiliary function.
3613 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3614 match before or at the first newline in the subject string. In pcretest,
3615 the /f option on a pattern can be used to set this.
3617 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3618 would behave wrongly. This has been present in PCRE since release 4.0.
3620 10. A number of changes to the pcregrep command:
3622 (a) Refactored how -x works; insert ^(...)$ instead of setting
3623 PCRE_ANCHORED and checking the length, in preparation for adding
3624 something similar for -w.
3626 (b) Added the -w (match as a word) option.
3628 (c) Refactored the way lines are read and buffered so as to have more
3629 than one at a time available.
3631 (d) Implemented a pcregrep test script.
3633 (e) Added the -M (multiline match) option. This allows patterns to match
3634 over several lines of the subject. The buffering ensures that at least
3635 8K, or the rest of the document (whichever is the shorter) is available
3636 for matching (and similarly the previous 8K for lookbehind assertions).
3638 (f) Changed the --help output so that it now says
3640 -w, --word-regex(p)
3642 instead of two lines, one with "regex" and the other with "regexp"
3643 because that confused at least one person since the short forms are the
3644 same. (This required a bit of code, as the output is generated
3645 automatically from a table. It wasn't just a text change.)
3647 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3648 option but starts with a hyphen. Could be a pattern or a path name
3649 starting with a hyphen, for instance.
3651 (h) "-" can be given as a file name to represent stdin.
3653 (i) When file names are being printed, "(standard input)" is used for
3654 the standard input, for compatibility with GNU grep. Previously
3655 "<stdin>" was used.
3657 (j) The option --label=xxx can be used to supply a name to be used for
3658 stdin when file names are being printed. There is no short form.
3660 (k) Re-factored the options decoding logic because we are going to add
3661 two more options that take data. Such options can now be given in four
3662 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3664 (l) Added the -A, -B, and -C options for requesting that lines of context
3665 around matches be printed.
3667 (m) Added the -L option to print the names of files that do not contain
3668 any matching lines, that is, the complement of -l.
3670 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3671 continue to scan other files.
3673 (o) The -s option was incorrectly implemented. For compatibility with other
3674 greps, it now suppresses the error message for a non-existent or non-
3675 accessible file (but not the return code). There is a new option called
3676 -q that suppresses the output of matching lines, which was what -s was
3677 previously doing.
3679 (p) Added --include and --exclude options to specify files for inclusion
3680 and exclusion when recursing.
3682 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3683 Hopefully, it now does.
3685 12. Missing cast in pcre_study().
3687 13. Added an "uninstall" target to the makefile.
3689 14. Replaced "extern" in the function prototypes in Makefile.in with
3690 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3691 world, but is set differently for Windows.
3693 15. Added a second compiling function called pcre_compile2(). The only
3694 difference is that it has an extra argument, which is a pointer to an
3695 integer error code. When there is a compile-time failure, this is set
3696 non-zero, in addition to the error test pointer being set to point to an
3697 error message. The new argument may be NULL if no error number is required
3698 (but then you may as well call pcre_compile(), which is now just a
3699 wrapper). This facility is provided because some applications need a
3700 numeric error indication, but it has also enabled me to tidy up the way
3701 compile-time errors are handled in the POSIX wrapper.
3703 16. Added VPATH=.libs to the makefile; this should help when building with one
3704 prefix path and installing with another. (Or so I'm told by someone who
3705 knows more about this stuff than I do.)
3707 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3708 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3709 match everything, including newlines. This is not POSIX-compatible, but
3710 somebody wanted the feature. From pcretest it can be activated by using
3711 both the P and the s flags.
3713 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3715 19. libpcre.pc was being incorrectly installed as executable.
3717 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3718 it now also looks for '\r' so that it will work unmodified on Windows.
3720 21. Added Google's contributed C++ wrapper to the distribution.
3722 22. Added some untidy missing memory free() calls in pcretest, to keep
3723 Electric Fence happy when testing.
3727 Version 5.0 13-Sep-04
3728 ---------------------
3730 1. Internal change: literal characters are no longer packed up into items
3731 containing multiple characters in a single byte-string. Each character
3732 is now matched using a separate opcode. However, there may be more than one
3733 byte in the character in UTF-8 mode.
3735 2. The pcre_callout_block structure has two new fields: pattern_position and
3736 next_item_length. These contain the offset in the pattern to the next match
3737 item, and its length, respectively.
3739 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3740 insertion of callouts before each pattern item. Added the /C option to
3741 pcretest to make use of this.
3743 4. On the advice of a Windows user, the lines
3745 #if defined(_WIN32) || defined(WIN32)
3746 _setmode( _fileno( stdout ), 0x8000 );
3747 #endif /* defined(_WIN32) || defined(WIN32) */
3749 have been added to the source of pcretest. This apparently does useful
3750 magic in relation to line terminators.
3752 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3753 for the benefit of those environments where the "b" makes a difference.
3755 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3756 to know about it. I have put a hack into configure.in that adds in code
3757 to set GCC=yes if CC=icc. This seems to end up at a point in the
3758 generated configure script that is early enough to affect the setting of
3759 compiler options, which is what is needed, but I have no means of testing
3760 whether it really works. (The user who reported this had patched the
3761 generated configure script, which of course I cannot do.)
3763 LATER: After change 22 below (new libtool files), the configure script
3764 seems to know about icc (and also ecc). Therefore, I have commented out
3765 this hack in configure.in.
3767 7. Added support for pkg-config (2 patches were sent in).
3769 8. Negated POSIX character classes that used a combination of internal tables
3770 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3771 [[:^ascii]]. Typically, they would match almost any characters. The other
3772 POSIX classes were not broken in this way.
3774 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3775 to find the match, as PCRE was deluded into thinking that the match had to
3776 start at the start point or following a newline. The same bug applied to
3777 patterns with negative forward assertions or any backward assertions
3778 preceding ".*" at the start, unless the pattern required a fixed first
3779 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3781 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3782 starting at the last subject character, bytes beyond the end of the subject
3783 string were read.
3785 11. Renamed the variable "class" as "classbits" to make life easier for C++
3786 users. (Previously there was a macro definition, but it apparently wasn't
3787 enough.)
3789 12. Added the new field "tables" to the extra data so that tables can be passed
3790 in at exec time, or the internal tables can be re-selected. This allows
3791 a compiled regex to be saved and re-used at a later time by a different
3792 program that might have everything at different addresses.
3794 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3795 -R library as well as a -L library.
3797 14. The debugging options of pcretest (-d on the command line or D on a
3798 pattern) showed incorrect output for anything following an extended class
3799 that contained multibyte characters and which was followed by a quantifier.
3801 15. Added optional support for general category Unicode character properties
3802 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3803 support. It adds about 90K to the size of the library. The meanings of the
3804 inbuilt class escapes such as \d and \s have NOT been changed.
3806 16. Updated pcredemo.c to include calls to free() to release the memory for the
3807 compiled pattern.
3809 17. The generated file chartables.c was being created in the source directory
3810 instead of in the building directory. This caused the build to fail if the
3811 source directory was different from the building directory, and was
3812 read-only.
3814 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3815 file. No doubt somebody will tell me if they don't make sense... Also added
3816 Dan Mooney's comments about building on OpenVMS.
3818 19. Added support for partial matching via the PCRE_PARTIAL option for
3819 pcre_exec() and the \P data escape in pcretest.
3821 20. Extended pcretest with 3 new pattern features:
3823 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3824 write the compiled pattern to the file whose name is "rest-of-line".
3825 This is a straight binary dump of the data, with the saved pointer to
3826 the character tables forced to be NULL. The study data, if any, is
3827 written too. After writing, pcretest reads a new pattern.
3829 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3830 compiled pattern from the given file. There must not be any
3831 occurrences of "<" in the file name (pretty unlikely); if there are,
3832 pcretest will instead treat the initial "<" as a pattern delimiter.
3833 After reading in the pattern, pcretest goes on to read data lines as
3834 usual.
3836 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3837 and 16-bit fields in a compiled pattern, to simulate a pattern that
3838 was compiled on a host of opposite endianness.
3840 21. The pcre-exec() function can now cope with patterns that were compiled on
3841 hosts of opposite endianness, with this restriction:
3843 As for any compiled expression that is saved and used later, the tables
3844 pointer field cannot be preserved; the extra_data field in the arguments
3845 to pcre_exec() should be used to pass in a tables address if a value
3846 other than the default internal tables were used at compile time.
3848 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3849 now diagnosed as an error. Previously, most of the time, a negative number
3850 would have been treated as zero, but if in addition "ovector" was passed as
3851 NULL, a crash could occur.
3853 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3854 new versions from the libtool 1.5 distribution (the last one is a copy of
3855 a file called libtool.m4). This seems to have fixed the need to patch
3856 "configure" to support Darwin 1.3 (which I used to do). However, I still
3857 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3858 workstation).
3860 24. Changed the PCRE licence to be the more standard "BSD" licence.
3863 Version 4.5 01-Dec-03
3864 ---------------------
3866 1. There has been some re-arrangement of the code for the match() function so
3867 that it can be compiled in a version that does not call itself recursively.
3868 Instead, it keeps those local variables that need separate instances for
3869 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3870 needs to "recurse". Keeping track of where control must go is done by means
3871 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3872 hide most of the details from the main code, and operates only if
3873 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3874 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3875 operating.
3877 To make it easier for callers to provide specially tailored get/free
3878 functions for this usage, two new functions, pcre_stack_malloc, and
3879 pcre_stack_free, are used. They are always called in strict stacking order,
3880 and the size of block requested is always the same.
3882 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3883 PCRE has been compiled to use the stack or the heap for recursion. The
3884 -C option of pcretest uses this to show which version is compiled.
3886 A new data escape \S, is added to pcretest; it causes the amounts of store
3887 obtained and freed by both kinds of malloc/free at match time to be added
3888 to the output.
3890 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3891 what's available on my current Linux desktop machine.
3893 3. When matching a UTF-8 string, the test for a valid string at the start has
3894 been extended. If start_offset is not zero, PCRE now checks that it points
3895 to a byte that is the start of a UTF-8 character. If not, it returns
3896 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3897 this is necessary because there may be backward assertions in the pattern.
3898 When matching the same subject several times, it may save resources to use
3899 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3901 4. The code for checking the validity of UTF-8 strings has been tightened so
3902 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3903 containing "overlong sequences".
3905 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3906 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3907 should have been "&", but it just so happened that all the cases this let
3908 through by mistake were picked up later in the function.
3910 6. I had used a variable called "isblank" - this is a C99 function, causing
3911 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3913 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3914 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3915 so the version is shown; (c) stop "make test" echoing "./RunTest".
3917 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3919 9. The prototype for memmove() for systems that don't have it was using
3920 size_t, but the inclusion of the header that defines size_t was later. I've
3921 moved the #includes for the C headers earlier to avoid this.
3923 10. Added some adjustments to the code to make it easier to compiler on certain
3924 special systems:
3926 (a) Some "const" qualifiers were missing.
3927 (b) Added the macro EXPORT before all exported functions; by default this
3928 is defined to be empty.
3929 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3930 that it reads its output file name as an argument instead of writing
3931 to the standard output and assuming this can be redirected.
3933 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3934 class containing characters with values greater than 255, PCRE compilation
3935 went into a loop.
3937 12. A recursive reference to a subpattern that was within another subpattern
3938 that had a minimum quantifier of zero caused PCRE to crash. For example,
3939 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3940 recursion. If the recursively-called subpattern itself had a zero repeat,
3941 that was OK.
3943 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3944 buffer into which it was copied (for escape processing) was still set at
3945 1024, so long lines caused crashes.
3947 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3948 "internal error: code overflow...". This applied to any character class
3949 that was followed by a possessive quantifier.
3951 15. Modified the Makefile to add libpcre.la as a prerequisite for
3952 libpcreposix.la because I was told this is needed for a parallel build to
3953 work.
3955 16. If a pattern that contained .* following optional items at the start was
3956 studied, the wrong optimizing data was generated, leading to matching
3957 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3958 matching string must start with a or b or c. The correct conclusion for
3959 this pattern is that a match can start with any character.
3962 Version 4.4 13-Aug-03
3963 ---------------------
3965 1. In UTF-8 mode, a character class containing characters with values between
3966 127 and 255 was not handled correctly if the compiled pattern was studied.
3967 In fixing this, I have also improved the studying algorithm for such
3968 classes (slightly).
3970 2. Three internal functions had redundant arguments passed to them. Removal
3971 might give a very teeny performance improvement.
3973 3. Documentation bug: the value of the capture_top field in a callout is *one
3974 more than* the number of the hightest numbered captured substring.
3976 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3977 in incorrectly linking with a previously installed version. They now link
3978 explicitly with libpcre.la.
3980 5. configure.in no longer needs to recognize Cygwin specially.
3982 6. A problem in pcre.in for Windows platforms is fixed.
3984 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3985 pcretest, it used to include the size of the study block as part of its
3986 output. Unfortunately, the structure contains a field that has a different
3987 size on different hardware architectures. This meant that the tests that
3988 showed this size failed. As the block is currently always of a fixed size,
3989 this information isn't actually particularly useful in pcretest output, so
3990 I have just removed it.
3992 8. Three pre-processor statements accidentally did not start in column 1.
3993 Sadly, there are *still* compilers around that complain, even though
3994 standard C has not required this for well over a decade. Sigh.
3996 9. In pcretest, the code for checking callouts passed small integers in the
3997 callout_data field, which is a void * field. However, some picky compilers
3998 complained about the casts involved for this on 64-bit systems. Now
3999 pcretest passes the address of the small integer instead, which should get
4000 rid of the warnings.
4002 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4003 both compile and run time, and gives an error if an invalid UTF-8 sequence
4004 is found. There is a option for disabling this check in cases where the
4005 string is known to be correct and/or the maximum performance is wanted.
4007 11. In response to a bug report, I changed one line in Makefile.in from
4009 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4010 to
4011 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4013 to look similar to other lines, but I have no way of telling whether this
4014 is the right thing to do, as I do not use Windows. No doubt I'll get told
4015 if it's wrong...
4018 Version 4.3 21-May-03
4019 ---------------------
4021 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4022 Makefile.
4024 2. Some refactoring to improve the quality of the code:
4026 (i) The utf8_table... variables are now declared "const".
4028 (ii) The code for \cx, which used the "case flipping" table to upper case
4029 lower case letters, now just substracts 32. This is ASCII-specific,
4030 but the whole concept of \cx is ASCII-specific, so it seems
4031 reasonable.
4033 (iii) PCRE was using its character types table to recognize decimal and
4034 hexadecimal digits in the pattern. This is silly, because it handles
4035 only 0-9, a-f, and A-F, but the character types table is locale-
4036 specific, which means strange things might happen. A private
4037 table is now used for this - though it costs 256 bytes, a table is
4038 much faster than multiple explicit tests. Of course, the standard
4039 character types table is still used for matching digits in subject
4040 strings against \d.
4042 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4043 ending in _t are). So I've renamed it as ESC_tee.
4045 3. The first argument for regexec() in the POSIX wrapper should have been
4046 defined as "const".
4048 4. Changed pcretest to use malloc() for its buffers so that they can be
4049 Electric Fenced for debugging.
4051 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4052 to read one or more bytes before the start of the subject string. Often this
4053 had no effect on PCRE's behaviour, but in some circumstances it could
4054 provoke a segmentation fault.
4056 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4057 to try to read one or more bytes before the start of the subject string.
4059 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4060 UTF-8 support could misbehave in various ways if the subject string
4061 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4062 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4063 back over UTF-8 characters.)
4066 Version 4.2 14-Apr-03
4067 ---------------------
4069 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4071 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4072 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4073 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4074 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4075 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4077 Note: automatic setting of the BUILD variables is not yet working
4078 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4079 compile-time but not at link-time
4080 [LINK]: use for linking executables only
4081 make different versions for Windows and non-Windows
4082 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4083 libraries
4084 [LINK_FOR_BUILD]: new variable
4085 [OBJEXT]: use throughout
4086 [EXEEXT]: use throughout
4087 <winshared>: new target
4088 <wininstall>: new target
4089 <dftables.o>: use native compiler
4090 <dftables>: use native linker
4091 <install>: handle Windows platform correctly
4092 <clean>: ditto
4093 <check>: ditto
4094 copy DLL to top builddir before testing
4096 As part of these changes, -no-undefined was removed again. This was reported
4097 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4098 in any case.
4100 3. Some tidies to get rid of compiler warnings:
4102 . In the match_data structure, match_limit was an unsigned long int, whereas
4103 match_call_count was an int. I've made them both unsigned long ints.
4105 . In pcretest the fact that a const uschar * doesn't automatically cast to
4106 a void * provoked a warning.
4108 . Turning on some more compiler warnings threw up some "shadow" variables
4109 and a few more missing casts.
4111 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4112 option, a class that contained a single character with a value between 128
4113 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4115 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4116 option, a class that contained several characters, but with at least one
4117 whose value was between 128 and 255 caused PCRE to crash.
4120 Version 4.1 12-Mar-03
4121 ---------------------
4123 1. Compiling with gcc -pedantic found a couple of places where casts were
4124 needed, and a string in dftables.c that was longer than standard compilers are
4125 required to support.
4127 2. Compiling with Sun's compiler found a few more places where the code could
4128 be tidied up in order to avoid warnings.
4130 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4131 first of these names is deprecated in the latest Autoconf in favour of the name
4132 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4133 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4134 analogy I have changed it to CFLAGS_FOR_BUILD.
4136 4. Added -no-undefined to the linking command in the Makefile, because this is
4137 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4138 linking step for the pcreposix library.
4140 5. PCRE was failing to diagnose the case of two named groups with the same
4141 name.
4143 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4144 literal character that is needed in the subject for a match, and scans along to
4145 ensure that it is present before embarking on the full matching process. This
4146 saves time in cases of nested unlimited repeats that are never going to match.
4147 Problem: the scan can take a lot of time if the subject is very long (e.g.
4148 megabytes), thus penalizing straightforward matches. It is now done only if the
4149 amount of subject to be scanned is less than 1000 bytes.
4151 7. A lesser problem with the same optimization is that it was recording the
4152 first character of an anchored pattern as "needed", thus provoking a search
4153 right along the subject, even when the first match of the pattern was going to
4154 fail. The "needed" character is now not set for anchored patterns, unless it
4155 follows something in the pattern that is of non-fixed length. Thus, it still
4156 fulfils its original purpose of finding quick non-matches in cases of nested
4157 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4160 Version 4.0 17-Feb-03
4161 ---------------------
4163 1. If a comment in an extended regex that started immediately after a meta-item
4164 extended to the end of string, PCRE compiled incorrect data. This could lead to
4165 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4167 2. Moved to autoconf 2.53 and libtool 1.4.2.
4169 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4170 the special perltest8 script is no longer needed - all the tests can be run
4171 from a single perltest script.
4173 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4174 by \s. It has now been removed in PCRE. This means it isn't recognized as
4175 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4176 class [:space:] *does* include VT, thereby creating a mess.
4178 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4179 space and tab.
4181 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4182 its new features into the main test script, reducing the number of scripts.
4184 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4185 were backward compatible, and made the (?i) apply to the whole pattern, as if
4186 /i were given. Now it behaves more logically, and applies the option setting
4187 only to what follows. PCRE has been changed to follow suit. However, if it
4188 finds options settings right at the start of the pattern, it extracts them into
4189 the global options, as before. Thus, they show up in the info data.
4191 8. Added support for the \Q...\E escape sequence. Characters in between are
4192 treated as literals. This is slightly different from Perl in that $ and @ are
4193 also handled as literals inside the quotes. In Perl, they will cause variable
4194 interpolation. Note the following examples:
4196 Pattern PCRE matches Perl matches
4198 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4199 \Qabc\$xyz\E abc\$xyz abc\$xyz
4200 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4202 For compatibility with Perl, \Q...\E sequences are recognized inside character
4203 classes as well as outside them.
4205 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4206 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4207 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4208 signed/unsigned warnings.
4210 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4211 option for pcretest, so I've replaced it by a simple function that does just
4212 that job.
4214 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4215 "pcregrep -".
4217 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4218 Java package. This provides some syntactic sugar for simple cases of what my
4219 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4220 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4221 item, you can use this simplified notation. Note that only makes sense with
4222 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4223 greediness, whatever the setting of the PCRE_UNGREEDY option.
4225 13. A change of greediness default within a pattern was not taking effect at
4226 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4227 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4228 was abstracted outside.
4230 14. PCRE now supports the \G assertion. It is true when the current matching
4231 position is at the start point of the match. This differs from \A when the
4232 starting offset is non-zero. Used with the /g option of pcretest (or similar
4233 code), it works in the same way as it does for Perl's /g option. If all
4234 alternatives of a regex begin with \G, the expression is anchored to the start
4235 match position, and the "anchored" flag is set in the compiled expression.
4237 15. Some bugs concerning the handling of certain option changes within patterns
4238 have been fixed. These applied to options other than (?ims). For example,
4239 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4240 been the other way round. Some of this was related to change 7 above.
4242 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4243 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4244 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4245 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4247 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4248 mode. Unlike ".", it always matches newline, whatever the setting of
4249 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4250 assertions. Perl allows it, but it doesn't (in general) work because it can't
4251 calculate the length of the lookbehind. At least, that's the case for Perl
4252 5.8.0 - I've been told they are going to document that it doesn't work in
4253 future.
4255 18. Added an error diagnosis for escapes that PCRE does not support: these are
4256 \L, \l, \N, \P, \p, \U, \u, and \X.
4258 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4259 reading past the end of the pattern in cases such as /[abcd/.
4261 20. PCRE was getting more memory than necessary for patterns with classes that
4262 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4264 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4265 compiling PCRE for use with Virtual Pascal.
4267 22. Small fix to the Makefile to make it work properly if the build is done
4268 outside the source tree.
4270 23. Added a new extension: a condition to go with recursion. If a conditional
4271 subpattern starts with (?(R) the "true" branch is used if recursion has
4272 happened, whereas the "false" branch is used only at the top level.
4274 24. When there was a very long string of literal characters (over 255 bytes
4275 without UTF support, over 250 bytes with UTF support), the computation of how
4276 much memory was required could be incorrect, leading to segfaults or other
4277 strange effects.
4279 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4280 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4281 there was a subsequent back reference to those brackets. This meant that, for
4282 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4283 possible to check for precisely this case. All we can do is abandon the
4284 optimization if .* occurs inside capturing brackets when there are any back
4285 references whatsoever. (See below for a better fix that came later.)
4287 26. The handling of the optimization for finding the first character of a
4288 non-anchored pattern, and for finding a character that is required later in the
4289 match were failing in some cases. This didn't break the matching; it just
4290 failed to optimize when it could. The way this is done has been re-implemented.
4292 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4294 28. Added a new feature that provides some of the functionality that Perl
4295 provides with (?{...}). The facility is termed a "callout". The way it is done
4296 in PCRE is for the caller to provide an optional function, by setting
4297 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4298 global variable. By default it is unset, which disables all calling out. To get
4299 the function called, the regex must include (?C) at appropriate points. This
4300 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4301 This provides a means of identifying different callout points. When PCRE
4302 reaches such a point in the regex, if pcre_callout has been set, the external
4303 function is called. It is provided with data in a structure called
4304 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4305 matching continues; if it returns a non-zero value, the match at the current
4306 point fails. However, backtracking will occur if possible. [This was changed
4307 later and other features added - see item 49 below.]
4309 29. pcretest is upgraded to test the callout functionality. It provides a
4310 callout function that displays information. By default, it shows the start of
4311 the match and the current position in the text. There are some new data escapes
4312 to vary what happens:
4314 \C+ in addition, show current contents of captured substrings
4315 \C- do not supply a callout function
4316 \C!n return 1 when callout number n is reached
4317 \C!n!m return 1 when callout number n is reached for the mth time
4319 30. If pcregrep was called with the -l option and just a single file name, it
4320 output "<stdin>" if a match was found, instead of the file name.
4322 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4323 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4324 pcre_exec(). This saves a malloc/free per call. The default value of
4325 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4326 when configuring.
4328 32. The default maximum size of a compiled pattern is 64K. There have been a
4329 few cases of people hitting this limit. The code now uses macros to handle the
4330 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4331 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4332 configuring. Tests 2 and 5 work only with 2-byte links because they output
4333 debugging information about compiled patterns.
4335 33. Internal code re-arrangements:
4337 (a) Moved the debugging function for printing out a compiled regex into
4338 its own source file (printint.c) and used #include to pull it into
4339 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4340 separate copies.
4342 (b) Defined the list of op-code names for debugging as a macro in
4343 internal.h so that it is next to the definition of the opcodes.
4345 (c) Defined a table of op-code lengths for simpler skipping along compiled
4346 code. This is again a macro in internal.h so that it is next to the
4347 definition of the opcodes.
4349 34. Added support for recursive calls to individual subpatterns, along the
4350 lines of Robin Houston's patch (but implemented somewhat differently).
4352 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4353 allow it to read and process whole directories in Win32. This code was
4354 contributed by Lionel Fourquaux; it has not been tested by me.
4356 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4357 used to name a group. Names consist of alphanumerics and underscores, and must
4358 be unique. Back references use the syntax (?P=name) and recursive calls use
4359 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4360 numbers. The function pcre_fullinfo() can be used after compilation to extract
4361 a name/number map. There are three relevant calls:
4363 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4364 PCRE_INFO_NAMECOUNT yields the number of entries
4365 PCRE_INFO_NAMETABLE yields a pointer to the map.
4367 The map is a vector of fixed-size entries. The size of each entry depends on
4368 the length of the longest name used. The first two bytes of each entry are the
4369 group number, most significant byte first. There follows the corresponding
4370 name, zero terminated. The names are in alphabetical order.
4372 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4373 case instead of 255. Making it the same both with and without UTF-8 support
4374 means that the same test output works with both.
4376 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4377 calling malloc() with a zero argument.
4379 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4380 optimization. I've improved things by keeping a bitmap of backreferences with
4381 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4382 fact referenced, the optimization can be applied. It is unlikely that a
4383 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4384 the match to follow \n) will appear inside brackets with a number greater than
4385 31, but if it does, any back reference > 31 suppresses the optimization.
4387 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4388 of disabling numbered capturing parentheses. Any opening parenthesis that is
4389 not followed by ? behaves as if it were followed by ?: but named parentheses
4390 can still be used for capturing (and they will acquire numbers in the usual
4391 way).
4393 41. Redesigned the return codes from the match() function into yes/no/error so
4394 that errors can be passed back from deep inside the nested calls. A malloc
4395 failure while inside a recursive subpattern call now causes the
4396 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4398 42. It is now possible to set a limit on the number of times the match()
4399 function is called in a call to pcre_exec(). This facility makes it possible to
4400 limit the amount of recursion and backtracking, though not in a directly
4401 obvious way, because the match() function is used in a number of different
4402 circumstances. The count starts from zero for each position in the subject
4403 string (for non-anchored patterns). The default limit is, for compatibility, a
4404 large number, namely 10 000 000. You can change this in two ways:
4406 (a) When configuring PCRE before making, you can use --with-match-limit=n
4407 to set a default value for the compiled library.
4409 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4410 a different value is set. See 45 below.
4412 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4414 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4415 of things that can be changed at compile time. The first argument specifies
4416 what is wanted and the second points to where the information is to be placed.
4417 The current list of available information is:
4421 The output is an integer that is set to one if UTF-8 support is available;
4422 otherwise it is set to zero.
4426 The output is an integer that it set to the value of the code that is used for
4427 newline. It is either LF (10) or CR (13).
4431 The output is an integer that contains the number of bytes used for internal
4432 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4436 The output is an integer that contains the threshold above which the POSIX
4437 interface uses malloc() for output vectors. See item 31 above.
4441 The output is an unsigned integer that contains the default limit of the number
4442 of match() calls in a pcre_exec() execution. See 42 above.
4444 44. pcretest has been upgraded by the addition of the -C option. This causes it
4445 to extract all the available output from the new pcre_config() function, and to
4446 output it. The program then exits immediately.
4448 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4449 order to support additional features. One way would have been to define
4450 pcre_exec2() (for example) with extra arguments, but this would not have been
4451 extensible, and would also have required all calls to the original function to
4452 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4453 is used for passing in "extra" data from pcre_study().
4455 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4456 contains the following fields:
4458 flags a bitmap indicating which of the following fields are set
4459 study_data opaque data from pcre_study()
4460 match_limit a way of specifying a limit on match() calls for a specific
4461 call to pcre_exec()
4462 callout_data data for callouts (see 49 below)
4464 The flag bits are also defined in pcre.h, and are
4470 The pcre_study() function now returns one of these new pcre_extra blocks, with
4471 the actual study data pointed to by the study_data field, and the
4472 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4473 before. That is, this change is entirely upwards-compatible and requires no
4474 change to existing code.
4476 If you want to pass in additional data to pcre_exec(), you can either place it
4477 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4478 block.
4480 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4481 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4482 times with different match limits, until it finds the minimum value needed for
4483 pcre_exec() to complete. The value is then output. This can be instructive; for
4484 most simple matches the number is quite small, but for pathological cases it
4485 gets very large very quickly.
4487 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4488 returns the size of the data block pointed to by the study_data field in a
4489 pcre_extra block, that is, the value that was passed as the argument to
4490 pcre_malloc() when PCRE was getting memory in which to place the information
4491 created by pcre_study(). The fourth argument should point to a size_t variable.
4492 pcretest has been extended so that this information is shown after a successful
4493 pcre_study() call when information about the compiled regex is being displayed.
4495 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4496 because what follows is always an absolute path. (Later: it turns out that this
4497 is more than cosmetic for MinGW, because it doesn't like empty path
4498 components.)
4500 49. Some changes have been made to the callout feature (see 28 above):
4502 (i) A callout function now has three choices for what it returns:
4504 0 => success, carry on matching
4505 > 0 => failure at this point, but backtrack if possible
4506 < 0 => serious error, return this value from pcre_exec()
4508 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4509 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4510 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4511 use by callout functions. It will never be used by PCRE itself.
4513 (ii) The pcre_extra structure (see 45 above) has a void * field called
4514 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4515 pcre_callout_block structure has a field of the same name. The contents of
4516 the field passed in the pcre_extra structure are passed to the callout
4517 function in the corresponding field in the callout block. This makes it
4518 easier to use the same callout-containing regex from multiple threads. For
4519 testing, the pcretest program has a new data escape
4521 \C*n pass the number n (may be negative) as callout_data
4523 If the callout function in pcretest receives a non-zero value as
4524 callout_data, it returns that value.
4526 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4527 there were some redundant $(CFLAGS) in commands that are now specified as
4528 $(LINK), which already includes $(CFLAGS).
4530 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4531 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4532 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4533 one-byte characters throughout. Note that case-insensitive matching applies
4534 only to characters whose values are less than 256. PCRE doesn't support the
4535 notion of cases for higher-valued characters.
4537 (i) A character class whose characters are all within 0-255 is handled as
4538 a bit map, and the map is inverted for negative classes. Previously, a
4539 character > 255 always failed to match such a class; however it should
4540 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4542 (ii) A negated character class with a single character < 255 is coded as
4543 "not this character" (OP_NOT). This wasn't working properly when the test
4544 character was multibyte, either singly or repeated.
4546 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4547 mode, for example: \x{100}{2,3}.
4549 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4550 singly or repeated) now correctly test multibyte characters. However,
4551 PCRE doesn't recognize any characters with values greater than 255 as
4552 digits, spaces, or word characters. Such characters always match \D, \S,
4553 and \W, and never match \d, \s, or \w.
4555 (v) Classes may now contain characters and character ranges with values
4556 greater than 255. For example: [ab\x{100}-\x{400}].
4558 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4559 PCRE in UTF-8 mode.
4561 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4562 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4563 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4564 value.)
4566 53. The single man page has become too large. I have therefore split it up into
4567 a number of separate man pages. These also give rise to individual HTML pages;
4568 these are now put in a separate directory, and there is an index.html page that
4569 lists them all. Some hyperlinking between the pages has been installed.
4571 54. Added convenience functions for handling named capturing parentheses.
4573 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4574 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4575 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4576 are faulted.
4578 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4579 calling configure. These values are used when compiling the dftables.c program
4580 which is run to generate the source of the default character tables. They
4581 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4582 you will need to set these values.
4584 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4587 Version 3.9 02-Jan-02
4588 ---------------------
4590 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4592 2. If --disable-static was given, the building process failed when trying to
4593 build pcretest and pcregrep. (For some reason it was using libtool to compile
4594 them, which is not right, as they aren't part of the library.)
4597 Version 3.8 18-Dec-01
4598 ---------------------
4600 1. The experimental UTF-8 code was completely screwed up. It was packing the
4601 bytes in the wrong order. How dumb can you get?
4604 Version 3.7 29-Oct-01
4605 ---------------------
4607 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4608 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4609 this didn't happen under Solaris 8, where I normally test things.
4611 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4612 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4613 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4614 this fix an hour or so after the initial 3.7 release.)
4617 Version 3.6 23-Oct-01
4618 ---------------------
4620 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4621 offsets passed as NULL with zero offset count.
4623 2. The config.guess and config.sub files had not been updated when I moved to
4624 the latest autoconf.
4627 Version 3.5 15-Aug-01
4628 ---------------------
4630 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4631 had been forgotten.
4633 2. By using declared but undefined structures, we can avoid using "void"
4634 definitions in pcre.h while keeping the internal definitions of the structures
4635 private.
4637 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4638 user point of view, this means that both static and shared libraries are built
4639 by default, but this can be individually controlled. More of the work of
4640 handling this static/shared cases is now inside libtool instead of PCRE's make
4641 file.
4643 4. The pcretest utility is now installed along with pcregrep because it is
4644 useful for users (to test regexs) and by doing this, it automatically gets
4645 relinked by libtool. The documentation has been turned into a man page, so
4646 there are now .1, .txt, and .html versions in /doc.
4648 5. Upgrades to pcregrep:
4649 (i) Added long-form option names like gnu grep.
4650 (ii) Added --help to list all options with an explanatory phrase.
4651 (iii) Added -r, --recursive to recurse into sub-directories.
4652 (iv) Added -f, --file to read patterns from a file.
4654 6. pcre_exec() was referring to its "code" argument before testing that
4655 argument for NULL (and giving an error if it was NULL).
4657 7. Upgraded Makefile.in to allow for compiling in a different directory from
4658 the source directory.
4660 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4661 options bits, the pointer it was passed was to an int instead of to an unsigned
4662 long int. This mattered only on 64-bit systems.
4664 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4665 generated) instead of pcre.in, which it its source. Also made the same change
4666 in several of the .c files.
4668 10. A new release of gcc defines printf() as a macro, which broke pcretest
4669 because it had an ifdef in the middle of a string argument for printf(). Fixed
4670 by using separate calls to printf().
4672 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4673 script, to force use of CR or LF instead of \n in the source. On non-Unix
4674 systems, the value can be set in config.h.
4676 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4677 absolute limit. Changed the text of the error message to make this clear, and
4678 likewise updated the man page.
4680 13. The limit of 99 on the number of capturing subpatterns has been removed.
4681 The new limit is 65535, which I hope will not be a "real" limit.
4684 Version 3.4 22-Aug-00
4685 ---------------------
4687 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4689 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4692 Version 3.3 01-Aug-00
4693 ---------------------
4695 1. If an octal character was given, but the value was greater than \377, it
4696 was not getting masked to the least significant bits, as documented. This could
4697 lead to crashes in some systems.
4699 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4700 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4702 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4703 These just pass their arguments on to (pcre_free)(), but they are provided
4704 because some uses of PCRE bind it to non-C systems that can call its functions,
4705 but cannot call free() or pcre_free() directly.
4707 4. Add "make test" as a synonym for "make check". Corrected some comments in
4708 the Makefile.
4710 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4711 Makefile.
4713 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4714 command called pgrep for grepping around the active processes.
4716 7. Added the beginnings of support for UTF-8 character strings.
4718 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4719 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4720 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4721 out for the ar command.)
4724 Version 3.2 12-May-00
4725 ---------------------
4727 This is purely a bug fixing release.
4729 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4730 of ZA. This was just one example of several cases that could provoke this bug,
4731 which was introduced by change 9 of version 2.00. The code for breaking
4732 infinite loops after an iteration that matches an empty string was't working
4733 correctly.
4735 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4736 when matched against abbab (for example). After matching an empty string, it
4737 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4738 caused it to match further down the string than it should.
4740 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4741 was there because it doesn't seem to be needed, and it causes trouble on some
4742 systems, as it is not a Standard C header. It has been removed.
4744 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4745 were reported on the Macintosh. The changes were from
4747 while ((c = *(++ptr)) != 0 && c != '\n');
4748 to
4749 while ((c = *(++ptr)) != 0 && c != '\n') ;
4751 Totally extraordinary, but if that's what it takes...
4753 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4754 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4755 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4756 assumes the way PCRE uses memmove() (always moving upwards).
4758 6. PCRE is being used in one environment where strchr() is not available. There
4759 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4760 faster code anyway.
4763 Version 3.1 09-Feb-00
4764 ---------------------
4766 The only change in this release is the fixing of some bugs in Makefile.in for
4767 the "install" target:
4769 (1) It was failing to install pcreposix.h.
4771 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4774 Version 3.0 01-Feb-00
4775 ---------------------
4777 1. Add support for the /+ modifier to perltest (to output $` like it does in
4778 pcretest).
4780 2. Add support for the /g modifier to perltest.
4782 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4783 matches null strings.
4785 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4786 pattern. Perl treats empty patterns specially - it reuses the most recent
4787 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4788 effect.
4790 5. The POSIX interface was broken in that it was just handing over the POSIX
4791 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4792 required a bigger vector, with some working space on the end. This means that
4793 the POSIX wrapper now has to get and free some memory, and copy the results.
4795 6. Added some simple autoconf support, placing the test data and the
4796 documentation in separate directories, re-organizing some of the
4797 information files, and making it build pcre-config (a GNU standard). Also added
4798 libtool support for building PCRE as a shared library, which is now the
4799 default.
4801 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4802 09 are not valid octal constants. Single digits will be used for minor values
4803 less than 10.
4805 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4806 existing programs that set these in the POSIX interface can use PCRE without
4807 modification.
4809 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4810 return all that pcre_info() returns, plus additional data. The pcre_info()
4811 function is retained for compatibility, but is considered to be obsolete.
4813 10. Added experimental recursion feature (?R) to handle one common case that
4814 Perl 5.6 will be able to do with (?p{...}).
4816 11. Added support for POSIX character classes like [:alpha:], which Perl is
4817 adopting.
4820 Version 2.08 31-Aug-99
4821 ----------------------
4823 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4824 trying to match at the startoffset position, but instead was moving forward to
4825 the next newline as if a previous match had failed.
4827 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4828 and could get into a loop if a null string was matched other than at the start
4829 of the subject.
4831 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4832 be distinguished at compile time, and for completeness also added PCRE_DATE.
4834 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4835 in GnuWin32 environments.
4838 Version 2.07 29-Jul-99
4839 ----------------------
4841 1. The documentation is now supplied in plain text form and HTML as well as in
4842 the form of man page sources.
4844 2. C++ compilers don't like assigning (void *) values to other pointer types.
4845 In particular this affects malloc(). Although there is no problem in Standard
4846 C, I've put in casts to keep C++ compilers happy.
4848 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4849 should be (const char *).
4851 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4852 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4853 However, I haven't made this a standard facility. The documentation doesn't
4854 mention it, and the Makefile doesn't support it.
4856 5. The Makefile now contains an "install" target, with editable destinations at
4857 the top of the file. The pcretest program is not installed.
4859 6. pgrep -V now gives the PCRE version number and date.
4861 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4862 causing the entire string to be ignored, instead of just the last character.
4864 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4865 non-matching string, it can take a very, very long time, even for strings of
4866 quite modest length, because of the nested recursion. PCRE now does better in
4867 some of these cases. It does this by remembering the last required literal
4868 character in the pattern, and pre-searching the subject to ensure it is present
4869 before running the real match. In other words, it applies a heuristic to detect
4870 some types of certain failure quickly, and in the above example, if presented
4871 with a string that has no trailing " it gives "no match" very quickly.
4873 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4874 other alternatives are tried instead.
4877 Version 2.06 09-Jun-99
4878 ----------------------
4880 1. Change pcretest's output for amount of store used to show just the code
4881 space, because the remainder (the data block) varies in size between 32-bit and
4882 64-bit systems.
4884 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4885 start matching at. This allows lookbehinds to work when searching for multiple
4886 occurrences in a string.
4888 3. Added additional options to pcretest for testing multiple occurrences:
4890 /+ outputs the rest of the string that follows a match
4891 /g loops for multiple occurrences, using the new startoffset argument
4892 /G loops for multiple occurrences by passing an incremented pointer
4894 4. PCRE wasn't doing the "first character" optimization for patterns starting
4895 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4896 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4897 the letter 'x'. On long subject strings, this gives a significant speed-up.
4900 Version 2.05 21-Apr-99
4901 ----------------------
4903 1. Changed the type of magic_number from int to long int so that it works
4904 properly on 16-bit systems.
4906 2. Fixed a bug which caused patterns starting with .* not to work correctly
4907 when the subject string contained newline characters. PCRE was assuming
4908 anchoring for such patterns in all cases, which is not correct because .* will
4909 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4910 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4911 must be retried after every newline in the subject.
4914 Version 2.04 18-Feb-99
4915 ----------------------
4917 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4918 computation of the store needed to hold the pattern was incorrect (too large).
4919 If such patterns were nested a few deep, this could multiply and become a real
4920 problem.
4922 2. Added /M option to pcretest to show the memory requirement of a specific
4923 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4925 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4926 compiled in such a way that the backtracking after subsequent failure was
4927 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4928 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4931 Version 2.03 02-Feb-99
4932 ----------------------
4934 1. Fixed typo and small mistake in man page.
4936 2. Added 4th condition (GPL supersedes if conflict) and created separate
4937 LICENCE file containing the conditions.
4939 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4940 Perl, that is the internal \ allows the delimiter to be included in the
4941 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4942 the final delimiter, add \ to the end of the pattern (to test the error).
4944 4. Added the convenience functions for extracting substrings after a successful
4945 match. Updated pcretest to make it able to test these functions.
4948 Version 2.02 14-Jan-99
4949 ----------------------
4951 1. Initialized the working variables associated with each extraction so that
4952 their saving and restoring doesn't refer to uninitialized store.
4954 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4955 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4956 fix the problem.
4958 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4959 calls, and wasn't printing the correct value for compiling calls. Increased the
4960 default value of LOOPREPEAT, and the number of significant figures in the
4961 times.
4963 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4965 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4966 a building problem on Windows NT with a FAT file system.
4969 Version 2.01 21-Oct-98
4970 ----------------------
4972 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4973 to character tables built by pcre_maketables() in the current locale. If NULL
4974 is passed, the default tables are used.
4977 Version 2.00 24-Sep-98
4978 ----------------------
4980 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4981 it any more.
4983 2. Allow quantification of (?>) groups, and make it work correctly.
4985 3. The first character computation wasn't working for (?>) groups.
4987 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4988 end of the subject) and add 5.005's \z, which really does match only at the
4989 very end of the subject.
4991 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4993 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4994 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4995 localized options. All options to pcre_study() were also removed.
4997 7. Add other new features from 5.005:
4999 $(?<= positive lookbehind
5000 $(?<! negative lookbehind
5001 (?imsx-imsx) added the unsetting capability
5002 such a setting is global if at outer level; local otherwise
5003 (?imsx-imsx:) non-capturing groups with option setting
5004 (?(cond)re|re) conditional pattern matching
5006 A backreference to itself in a repeated group matches the previous
5007 captured string.
5009 8. General tidying up of studying (both automatic and via "study")
5010 consequential on the addition of new assertions.
5012 9. As in 5.005, unlimited repeated groups that could match an empty substring
5013 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5014 runtime if any iteration does actually match an empty substring.
5016 10. Include the RunTest script in the distribution.
5018 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5019 discrepancies, some of which were old and were also with respect to 5.004. They
5020 have now been fixed.
5023 Version 1.09 28-Apr-98
5024 ----------------------
5026 1. A negated single character class followed by a quantifier with a minimum
5027 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5028 program crashes, or just wrong answers. This did not apply to negated classes
5029 containing more than one character, or to minima other than one.
5032 Version 1.08 27-Mar-98
5033 ----------------------
5035 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5037 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5038 latter must appear before anything that relies on it in the pattern.
5041 Version 1.07 16-Feb-98
5042 ----------------------
5044 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5045 repeat of a potentially empty string).
5048 Version 1.06 23-Jan-98
5049 ----------------------
5051 1. Added Markus Oberhumer's little patches for C++.
5053 2. Literal strings longer than 255 characters were broken.
5056 Version 1.05 23-Dec-97
5057 ----------------------
5059 1. Negated character classes containing more than one character were failing if
5060 PCRE_CASELESS was set at run time.
5063 Version 1.04 19-Dec-97
5064 ----------------------
5066 1. Corrected the man page, where some "const" qualifiers had been omitted.
5068 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5069 input syntax.
5071 3. Fixed memory leak which occurred when a regex with back references was
5072 matched with an offsets vector that wasn't big enough. The temporary memory
5073 that is used in this case wasn't being freed if the match failed.
5075 4. Tidied pcretest to ensure it frees memory that it gets.
5077 5. Temporary memory was being obtained in the case where the passed offsets
5078 vector was exactly big enough.
5080 6. Corrected definition of offsetof() from change 5 below.
5082 7. I had screwed up change 6 below and broken the rules for the use of
5083 setjmp(). Now fixed.
5086 Version 1.03 18-Dec-97
5087 ----------------------
5089 1. A erroneous regex with a missing opening parenthesis was correctly
5090 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5091 on some systems.
5093 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5094 it was reported that one broken compiler failed on the former because "code" is
5095 also an independent variable.
5097 3. The erroneous regex a[]b caused an array overrun reference.
5099 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5100 fail on data ending with that character. (It was going on too far, and checking
5101 the next character, typically a binary zero.) This was specific to the
5102 optimized code for single-character negative classes.
5104 5. Added a contributed patch from the TIN world which does the following:
5106 + Add an undef for memmove, in case the the system defines a macro for it.
5108 + Add a definition of offsetof(), in case there isn't one. (I don't know
5109 the reason behind this - offsetof() is part of the ANSI standard - but
5110 it does no harm).
5112 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5113 most of the places where whitespace preceded '#'. I have given up and
5114 allowed the remaining 2 cases to be at the margin.
5116 + Rename some variables in pcre to eliminate shadowing. This seems very
5117 pedantic, but does no harm, of course.
5119 6. Moved the call to setjmp() into its own function, to get rid of warnings
5120 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5122 7. Constructs such as \d{8,} were compiling into the equivalent of
5123 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5124 outcome, but in this particular case used more store than had been allocated,
5125 which caused the bug to be discovered because it threw up an internal error.
5127 8. The debugging code in both pcre and pcretest for outputting the compiled
5128 form of a regex was going wrong in the case of back references followed by
5129 curly-bracketed repeats.
5132 Version 1.02 12-Dec-97
5133 ----------------------
5135 1. Typos in pcre.3 and comments in the source fixed.
5137 2. Applied a contributed patch to get rid of places where it used to remove
5138 'const' from variables, and fixed some signed/unsigned and uninitialized
5139 variable warnings.
5141 3. Added the "runtest" target to Makefile.
5143 4. Set default compiler flag to -O2 rather than just -O.
5146 Version 1.01 19-Nov-97
5147 ----------------------
5149 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5150 like /([ab]*)*/, that is, for classes with more than one character in them.
5152 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5153 as /((?>a*))*/ (a PCRE_EXTRA facility).
5156 Version 1.00 18-Nov-97
5157 ----------------------
5159 1. Added compile-time macros to support systems such as SunOS4 which don't have
5160 memmove() or strerror() but have other things that can be used instead.
5162 2. Arranged that "make clean" removes the executables.
5165 Version 0.99 27-Oct-97
5166 ----------------------
5168 1. Fixed bug in code for optimizing classes with only one character. It was
5169 initializing a 32-byte map regardless, which could cause it to run off the end
5170 of the memory it had got.
5172 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5175 Version 0.98 22-Oct-97
5176 ----------------------
5178 1. Fixed bug in code for handling temporary memory usage when there are more
5179 back references than supplied space in the ovector. This could cause segfaults.
5182 Version 0.97 21-Oct-97
5183 ----------------------
5185 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5187 2. Optimized negated single characters not to use a bit map.
5189 3. Brought error texts together as macro definitions; clarified some of them;
5190 fixed one that was wrong - it said "range out of order" when it meant "invalid
5191 escape sequence".
5193 4. Changed some char * arguments to const char *.
5195 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5197 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5198 pcretest.
5201 Version 0.96 16-Oct-97
5202 ----------------------
5204 1. Added a simple "pgrep" utility to the distribution.
5206 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5207 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5208 where "ddd" means "one or more decimal digits".
5210 3. Fixed serious bug. If a pattern had a back reference, but the call to
5211 pcre_exec() didn't supply a large enough ovector to record the related
5212 identifying subpattern, the match always failed. PCRE now remembers the number
5213 of the largest back reference, and gets some temporary memory in which to save
5214 the offsets during matching if necessary, in order to ensure that
5215 backreferences always work.
5217 4. Increased the compatibility with Perl in a number of ways:
5219 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5220 to request this handling. The option can be set at compile or exec time.
5222 (b) $ matches before a terminating newline by default; an option
5223 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5224 mode). The option can be set at compile or exec time.
5226 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5227 the same as Perl's. If the decimal number it represents is less than 10
5228 or there aren't that many previous left capturing parentheses, an octal
5229 escape is read. Inside a character class, it's always an octal escape,
5230 even if it is a single digit.
5232 (d) An escaped but undefined alphabetic character is taken as a literal,
5233 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5234 escapes.
5236 (e) {0} is now permitted. (The previous item is removed from the compiled
5237 pattern).
5239 5. Changed all the names of code files so that the basic parts are no longer
5240 than 10 characters, and abolished the teeny "globals.c" file.
5242 6. Changed the handling of character classes; they are now done with a 32-byte
5243 bit map always.
5245 7. Added the -d and /D options to pcretest to make it possible to look at the
5246 internals of compilation without having to recompile pcre.
5249 Version 0.95 23-Sep-97
5250 ----------------------
5252 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5253 \x20 at the start of a run of normal characters. These were being treated as
5254 real characters, instead of the source characters being re-checked.
5257 Version 0.94 18-Sep-97
5258 ----------------------
5260 1. The functions are now thread-safe, with the caveat that the global variables
5261 containing pointers to malloc() and free() or alternative functions are the
5262 same for all threads.
5264 2. Get pcre_study() to generate a bitmap of initial characters for non-
5265 anchored patterns when this is possible, and use it if passed to pcre_exec().
5268 Version 0.93 15-Sep-97
5269 ----------------------
5271 1. /(b)|(:+)/ was computing an incorrect first character.
5273 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5274 but not actually doing anything yet.
5276 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5277 as Perl does (e.g. [-az] or [az-]).
5279 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5280 all possible positions.
5282 5. Split up into different modules to avoid including unneeded functions in a
5283 compiled binary. However, compile and exec are still in one module. The "study"
5284 function is split off.
5286 6. The character tables are now in a separate module whose source is generated
5287 by an auxiliary program - but can then be edited by hand if required. There are
5288 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5289 toupper() in the code.
5291 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5292 make them global. Abolish the function for setting them, as the caller can now
5293 set them directly.
5296 Version 0.92 11-Sep-97
5297 ----------------------
5299 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5300 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5302 2. Caseless matching was not working in character classes if the characters in
5303 the pattern were in upper case.
5305 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5307 4. Make PCRE_ANCHORED public and accept as a compile option.
5309 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5310 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5311 pass them.
5313 6. Give an error if bad option bits passed at compile or run time.
5315 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5316 pcretest to cause it to pass that flag.
5318 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5319 options, and the first character, if set.
5321 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5324 Version 0.91 10-Sep-97
5325 ----------------------
5327 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5328 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5330 2. PCRE was looping on encountering an indefinitely repeated back reference to
5331 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5332 Perl does - treats the match as successful.
5334 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5