ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1434 - (show annotations)
Mon Jan 6 20:04:50 2014 UTC (6 years ago) by zherczeg
File size: 260500 byte(s)
JIT: Optimize brackets with more than four alternatives.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.35-RC1 xx-xxxx-201x
5 -----------------------------
7 1. A new flag is set, when property checks are present in an XCLASS.
8 When this flag is not set, PCRE can perform certain optimizations
9 such as studying these XCLASS-es.
11 2. The auto-possessification of character sets were improved: a normal
12 and an extended character set can be compared now. Furthermore
13 the JIT compiler optimizes more character set checks.
15 3. Got rid of some compiler warnings for potentially uninitialized variables
16 that show up only when compiled with -O2.
18 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
19 of a match later then the end of the match. The pcretest program was not
20 handling the case sensibly - it was outputting from the start to the next
21 binary zero. It now reports this situation in a message, and outputs the
22 text from the end to the start.
24 5. Fast forward search is improved in JIT. Instead of the first three
25 characters, any three characters with fixed position can be searched.
26 Search order: first, last, middle.
28 6. Improve character range checks in JIT. Characters are read by an inprecise
29 function now, which returns with an unknown value if the character code is
30 above a certain treshold (e.g: 256). The only limitation is that the value
31 must be bigger than the treshold as well. This function is useful, when
32 the characters above the treshold are handled in the same way.
34 7. The macros whose names start with RAWUCHAR are placeholders for a future
35 mode in which only the bottom 21 bits of 32-bit data items are used. To
36 make this more memorable for those maintaining the code, the names have
37 been changed to start with UCHAR21, and an extensive comment has been added
38 to their definition.
40 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
41 to the export list in Makefile.am (they were accidentally omitted from the
42 8.34 tarball).
44 9. The informational output from pcretest used the phrase "starting byte set"
45 which is inappropriate for the 16-bit and 32-bit libraries. As the output
46 for "first char" and "need char" really means "non-UTF-char", I've changed
47 "byte" to "char", and slightly reworded the output. The documentation about
48 these values has also been (I hope) clarified.
50 10. Another JIT related optimization: use table jumps for selecting the correct
51 backtracking path, when more than four alternatives are present inside a
52 bracket.
55 Version 8.34 15-December-2013
56 -----------------------------
58 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
59 executable memory. Patch inspired by Carsten Klein.
61 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
62 this macro is never tested and has no effect, because the work to support
63 coverage involves only compiling and linking options and special targets in
64 the Makefile. The comment in config.h implied that defining the macro would
65 enable coverage support, which is totally false. There was also support for
66 setting this macro in the CMake files (my fault, I just copied it from
67 configure). SUPPORT_GCOV has now been removed.
69 3. Make a small performance improvement in strlen16() and strlen32() in
70 pcretest.
72 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
73 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
74 these cases). There was also one in pcretest.c.
76 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
78 6. In UTF mode, the code for checking whether a group could match an empty
79 string (which is used for indefinitely repeated groups to allow for
80 breaking an infinite loop) was broken when the group contained a repeated
81 negated single-character class with a character that occupied more than one
82 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
83 UTF-8 mode). The effect was undefined: the group might or might not be
84 deemed as matching an empty string, or the program might have crashed.
86 7. The code for checking whether a group could match an empty string was not
87 recognizing that \h, \H, \v, \V, and \R must match a character.
89 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
90 an empty string. If it can, pcretest shows this in its information output.
92 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
93 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
94 matched by pcre_exec() without using JIT:
96 (a) If the rest of the pattern did not match after a maximal run of
97 grapheme clusters, the code for backing up to try with fewer of them
98 did not always back up over a full grapheme when characters that do not
99 have the modifier quality were involved, e.g. Hangul syllables.
101 (b) If the match point in a subject started with modifier character, and
102 there was no match, the code could incorrectly back up beyond the match
103 point, and potentially beyond the first character in the subject,
104 leading to a segfault or an incorrect match result.
106 10. A conditional group with an assertion condition could lead to PCRE
107 recording an incorrect first data item for a match if no other first data
108 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
109 first data item, and therefore matched "ca" after "c" instead of at the
110 start.
112 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
113 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
115 12. The source of pcregrep now includes z/OS-specific code so that it can be
116 compiled for z/OS as part of the special z/OS distribution.
118 13. Added the -T and -TM options to pcretest.
120 14. The code in pcre_compile.c for creating the table of named capturing groups
121 has been refactored. Instead of creating the table dynamically during the
122 actual compiling pass, the information is remembered during the pre-compile
123 pass (on the stack unless there are more than 20 named groups, in which
124 case malloc() is used) and the whole table is created before the actual
125 compile happens. This has simplified the code (it is now nearly 150 lines
126 shorter) and prepared the way for better handling of references to groups
127 with duplicate names.
129 15. A back reference to a named subpattern when there is more than one of the
130 same name now checks them in the order in which they appear in the pattern.
131 The first one that is set is used for the reference. Previously only the
132 first one was inspected. This change makes PCRE more compatible with Perl.
134 16. Unicode character properties were updated from Unicode 6.3.0.
136 17. The compile-time code for auto-possessification has been refactored, based
137 on a patch by Zoltan Herczeg. It now happens after instead of during
138 compilation. The code is cleaner, and more cases are handled. The option
139 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
140 options in pcretest are provided to set it. It can also be set by
141 (*NO_AUTO_POSSESS) at the start of a pattern.
143 18. The character VT has been added to the default ("C" locale) set of
144 characters that match \s and are generally treated as white space,
145 following this same change in Perl 5.18. There is now no difference between
146 "Perl space" and "POSIX space". Whether VT is treated as white space in
147 other locales depends on the locale.
149 19. The code for checking named groups as conditions, either for being set or
150 for being recursed, has been refactored (this is related to 14 and 15
151 above). Processing unduplicated named groups should now be as fast at
152 numerical groups, and processing duplicated groups should be faster than
153 before.
155 20. Two patches to the CMake build system, by Alexander Barkov:
157 (1) Replace the "source" command by "." in CMakeLists.txt because
158 "source" is a bash-ism.
160 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
161 without these the CMake build does not work on Solaris.
163 21. Perl has changed its handling of \8 and \9. If there is no previously
164 encountered capturing group of those numbers, they are treated as the
165 literal characters 8 and 9 instead of a binary zero followed by the
166 literals. PCRE now does the same.
168 22. Following Perl, added \o{} to specify codepoints in octal, making it
169 possible to specify values greater than 0777 and also making them
170 unambiguous.
172 23. Perl now gives an error for missing closing braces after \x{... instead of
173 treating the string as literal. PCRE now does the same.
175 24. RunTest used to grumble if an inappropriate test was selected explicitly,
176 but just skip it when running all tests. This make it awkward to run ranges
177 of tests when one of them was inappropriate. Now it just skips any
178 inappropriate tests, as it always did when running all tests.
180 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
181 character types such as \d or \w, too many callouts were inserted, and the
182 data that they returned was rubbish.
184 26. In UCP mode, \s was not matching two of the characters that Perl matches,
185 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
186 were matched by \h. The code has now been refactored so that the lists of
187 the horizontal and vertical whitespace characters used for \h and \v (which
188 are defined only in one place) are now also used for \s.
190 27. Add JIT support for the 64 bit TileGX architecture.
191 Patch by Jiong Wang (Tilera Corporation).
193 28. Possessive quantifiers for classes (both explicit and automatically
194 generated) now use special opcodes instead of wrapping in ONCE brackets.
196 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
197 (because it's meaningless), this was not happening when PCRE_CASELESS was
198 set. Not wrong, but inefficient.
200 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
201 properties for \w, \d, etc) is present in a test regex. Otherwise if the
202 test contains no characters greater than 255, Perl doesn't realise it
203 should be using Unicode semantics.
205 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
206 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
207 does in Unicode mode.
209 32. Added the "forbid" facility to pcretest so that putting tests into the
210 wrong test files can sometimes be quickly detected.
212 33. There is now a limit (default 250) on the depth of nesting of parentheses.
213 This limit is imposed to control the amount of system stack used at compile
214 time. It can be changed at build time by --with-parens-nest-limit=xxx or
215 the equivalent in CMake.
217 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
218 errors. Perl warns for these when in warning mode, but PCRE has no facility
219 for giving warnings.
221 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
222 However, this was not working for (?!) because it is optimized to (*FAIL),
223 for which PCRE does not allow quantifiers. The optimization is now disabled
224 when a quantifier follows (?!). I can't see any use for this, but it makes
225 things uniform.
227 36. Perl no longer allows group names to start with digits, so I have made this
228 change also in PCRE. It simplifies the code a bit.
230 37. In extended mode, Perl ignores spaces before a + that indicates a
231 possessive quantifier. PCRE allowed a space before the quantifier, but not
232 before the possessive +. It now does.
234 38. The use of \K (reset reported match start) within a repeated possessive
235 group such as (a\Kb)*+ was not working.
237 40. Document that the same character tables must be used at compile time and
238 run time, and that the facility to pass tables to pcre_exec() and
239 pcre_dfa_exec() is for use only with saved/restored patterns.
241 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
242 features for Builds with MSVC:
244 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
245 builds). The .rc files can be used to set FileDescription and many other
246 attributes.
248 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
249 This allows higher-level build scripts which want .pdb files to avoid
250 hard-coding the exact files needed."
252 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
253 mean "start of word" and "end of word", respectively, as a transition aid.
255 43. A minimizing repeat of a class containing codepoints greater than 255 in
256 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
257 compiled to use the heap for recursion.
259 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
260 is configured.
263 Version 8.33 28-May-2013
264 ------------------------
266 1. Added 'U' to some constants that are compared to unsigned integers, to
267 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
268 variables that are added to signed variables, to ensure the result is
269 signed and can be negated.
271 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
272 CMake config files.
274 3. Revise the creation of config.h.generic so that all boolean macros are
275 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
276 overriding via -D on the command line possible.
278 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
279 to unsigned int is reported to make a quite noticeable speed difference in
280 a specific Windows environment. Testing on Linux did also appear to show
281 some benefit (and it is clearly not harmful). Also fixed the definition of
282 Xop which should be unsigned.
284 5. Related to (4), changing the definition of the intermediate variable cc
285 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
286 improvements.
288 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
289 unnecessary spaces.
291 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
292 and later.
294 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
296 9. Optimizing fast_forward_start_bits in JIT.
298 10. Adding support for callouts in JIT, and fixing some issues revealed
299 during this work. Namely:
301 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
303 (b) Minimum length was not checked before the matching is started.
305 11. The value of capture_last that is passed to callouts was incorrect in some
306 cases when there was a capture on one path that was subsequently abandoned
307 after a backtrack. Also, the capture_last value is now reset after a
308 recursion, since all captures are also reset in this case.
310 12. The interpreter no longer returns the "too many substrings" error in the
311 case when an overflowing capture is in a branch that is subsequently
312 abandoned after a backtrack.
314 13. In the pathological case when an offset vector of size 2 is used, pcretest
315 now prints out the matched string after a yield of 0 or 1.
317 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
318 Only supported by the JIT compiler at the moment.
320 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
322 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
323 offset of the starting point of the matching process, provided the offsets
324 vector is large enough.
326 17. The \A escape now records a lookbehind value of 1, though its execution
327 does not actually inspect the previous character. This is to ensure that,
328 in partial multi-segment matching, at least one character from the old
329 segment is retained when a new segment is processed. Otherwise, if there
330 are no lookbehinds in the pattern, \A might match incorrectly at the start
331 of a new segment.
333 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
335 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
336 gives some modest performance improvement in 8-bit mode.
338 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
339 be expressed in certain programming languages using Universal Character
340 Names.
342 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
343 which points out that "non characters" are not "characters that may not
344 appear in Unicode strings" but rather "characters that are reserved for
345 internal use and have only local meaning".
347 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
348 there was a conditional group that depended on an assertion, if the
349 assertion was false, the callout that immediately followed the alternation
350 in the condition was skipped when pcre_exec() was used for matching.
352 23. Allow an explicit callout to be inserted before an assertion that is the
353 condition for a conditional group, for compatibility with automatic
354 callouts, which always insert a callout at this point.
356 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
357 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
359 25. (*PRUNE) is now supported by the JIT compiler.
361 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
363 27. Fix the case where there are two or more SKIPs with arguments that may be
364 ignored.
366 28. (*SKIP) is now supported by the JIT compiler.
368 29. (*THEN) is now supported by the JIT compiler.
370 30. Update RunTest with additional test selector options.
372 31. The way PCRE handles backtracking verbs has been changed in two ways.
374 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
375 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
376 backtracking. In some cases this makes it more Perl-compatible, but Perl's
377 rather obscure rules do not always do the same thing.
379 (2) Previously, backtracking verbs were confined within assertions. This is
380 no longer the case for positive assertions, except for (*ACCEPT). Again,
381 this sometimes improves Perl compatibility, and sometimes does not.
383 32. A number of tests that were in test 2 because Perl did things differently
384 have been moved to test 1, because either Perl or PCRE has changed, and
385 these tests are now compatible.
387 32. Backtracking control verbs are now handled in the same way in JIT and
388 interpreter.
390 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
391 contained a forward subroutine reference caused a compile error.
393 34. Auto-detect and optimize limited repetitions in JIT.
395 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
396 blocking (*UTF) etc.
398 36. In the interpreter, maximizing pattern repetitions for characters and
399 character types now use tail recursion, which reduces stack usage.
401 37. The value of the max lookbehind was not correctly preserved if a compiled
402 and saved regex was reloaded on a host of different endianness.
404 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
405 of the compiled pattern block, expand the flags field from 16 to 32 bits
406 because it was almost full.
408 39. Try madvise first before posix_madvise.
410 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
411 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
412 original change locked out all matches of empty strings. This has been
413 changed so that one match of an empty string per line is recognized.
414 Subsequent searches on the same line (for colouring or for --only-matching,
415 for example) do not recognize empty strings.
417 41. Applied a user patch to fix a number of spelling mistakes in comments.
419 42. Data lines longer than 65536 caused pcretest to crash.
421 43. Clarified the data type for length and startoffset arguments for pcre_exec
422 and pcre_dfa_exec in the function-specific man pages, where they were
423 explicitly stated to be in bytes, never having been updated. I also added
424 some clarification to the pcreapi man page.
426 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
427 a segmentation fault.
430 Version 8.32 30-November-2012
431 -----------------------------
433 1. Improved JIT compiler optimizations for first character search and single
434 character iterators.
436 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
437 Patch by Daniel Richard G.
439 3. Single character iterator optimizations in the JIT compiler.
441 4. Improved JIT compiler optimizations for character ranges.
443 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
444 Reported by Giuseppe D'Angelo.
446 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
447 of a line, was being set incorrectly in cases where .* appeared inside
448 atomic brackets at the start of a pattern, or where there was a subsequent
449 *PRUNE or *SKIP.
451 7. Improved instruction cache flush for POWER/PowerPC.
452 Patch by Daniel Richard G.
454 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
455 grep:
457 (a) There is now no limit to the number of patterns to be matched.
459 (b) An error is given if a pattern is too long.
461 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
462 are now supported.
464 (d) --exclude-from and --include-from (multiple use) have been added.
466 (e) Exclusions and inclusions now apply to all files and directories, not
467 just to those obtained from scanning a directory recursively.
469 (f) Multiple uses of -f and --file-list are now supported.
471 (g) In a Windows environment, the default for -d has been changed from
472 "read" (the GNU grep default) to "skip", because otherwise the presence
473 of a directory in the file list provokes an error.
475 (h) The documentation has been revised and clarified in places.
477 9. Improve the matching speed of capturing brackets.
479 10. Changed the meaning of \X so that it now matches a Unicode extended
480 grapheme cluster.
482 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
483 out POSIX threads when JIT support is configured.
485 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
487 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
488 structure could go wrong in environments where size_t is not the same size
489 as int.
491 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
492 set.
494 15. The EBCDIC support had decayed; later updates to the code had included
495 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
496 general tidy up of EBCDIC-related issues, and the documentation was also
497 not quite right. There is now a test that can be run on ASCII systems to
498 check some of the EBCDIC-related things (but is it not a full test).
500 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
501 in a small tidy to the code.
503 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
505 18. If the --only-matching (-o) option in pcregrep is specified multiple
506 times, each one causes appropriate output. For example, -o1 -o2 outputs the
507 substrings matched by the 1st and 2nd capturing parentheses. A separating
508 string can be specified by --om-separator (default empty).
510 19. Improving the first n character searches.
512 20. Turn case lists for horizontal and vertical white space into macros so that
513 they are defined only once.
515 21. This set of changes together give more compatible Unicode case-folding
516 behaviour for characters that have more than one other case when UCP
517 support is available.
519 (a) The Unicode property table now has offsets into a new table of sets of
520 three or more characters that are case-equivalent. The MultiStage2.py
521 script that generates these tables (the pcre_ucd.c file) now scans
522 CaseFolding.txt instead of UnicodeData.txt for character case
523 information.
525 (b) The code for adding characters or ranges of characters to a character
526 class has been abstracted into a generalized function that also handles
527 case-independence. In UTF-mode with UCP support, this uses the new data
528 to handle characters with more than one other case.
530 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
531 whose other case is greater than 256 are now correctly matched
532 caselessly. Previously, the high codepoint matched the low one, but not
533 vice versa.
535 (d) The processing of \h, \H, \v, and \ in character classes now makes use
536 of the new class addition function, using character lists defined as
537 macros alongside the case definitions of 20 above.
539 (e) Caseless back references now work with characters that have more than
540 one other case.
542 (f) General caseless matching of characters with more than one other case
543 is supported.
545 22. Unicode character properties were updated from Unicode 6.2.0
547 23. Improved CMake support under Windows. Patch by Daniel Richard G.
549 24. Add support for 32-bit character strings, and UTF-32
551 25. Major JIT compiler update (code refactoring and bugfixing).
552 Experimental Sparc 32 support is added.
554 26. Applied a modified version of Daniel Richard G's patch to create
555 pcre.h.generic and config.h.generic by "make" instead of in the
556 PrepareRelease script.
558 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
559 pcre_compile.c when checking for a zero character.
561 28. Introducing a native interface for JIT. Through this interface, the compiled
562 machine code can be directly executed. The purpose of this interface is to
563 provide fast pattern matching, so several sanity checks are not performed.
564 However, feature tests are still performed. The new interface provides
565 1.4x speedup compared to the old one.
567 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
568 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
569 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
571 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
572 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
573 the "old" RFC 2279). Instead, it ended up passing a negative length to
574 pcre_exec().
576 31. Add support for GCC's visibility feature to hide internal functions.
578 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
579 "unknown -C option" after outputting 0 or 1.
581 33. There is now support for generating a code coverage report for the test
582 suite in environments where gcc is the compiler and lcov is installed. This
583 is mainly for the benefit of the developers.
585 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
586 unaddressable using valgrind annotations, allowing valgrind to detect
587 invalid memory accesses. This is mainly for the benefit of the developers.
589 25. (*UTF) can now be used to start a pattern in any of the three libraries.
591 26. Give configure error if --enable-cpp but no C++ compiler found.
594 Version 8.31 06-July-2012
595 -------------------------
597 1. Fixing a wrong JIT test case and some compiler warnings.
599 2. Removed a bashism from the RunTest script.
601 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
602 to unsigned type, result still unsigned" that was given by an MS compiler
603 on encountering the code "-sizeof(xxx)".
605 4. Partial matching support is added to the JIT compiler.
607 5. Fixed several bugs concerned with partial matching of items that consist
608 of more than one character:
610 (a) /^(..)\1/ did not partially match "aba" because checking references was
611 done on an "all or nothing" basis. This also applied to repeated
612 references.
614 (b) \R did not give a hard partial match if \r was found at the end of the
615 subject.
617 (c) \X did not give a hard partial match after matching one or more
618 characters at the end of the subject.
620 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
621 a partial match for the string "\r".
623 (e) When newline was set to CRLF, the metacharacter "." did not recognize
624 a partial match for a CR character at the end of the subject string.
626 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
627 running pcretest, the text "(JIT)" added to the output whenever JIT is
628 actually used to run the match.
630 7. Individual JIT compile options can be set in pcretest by following -s+[+]
631 or /S+[+] with a digit between 1 and 7.
633 8. OP_NOT now supports any UTF character not just single-byte ones.
635 9. (*MARK) control verb is now supported by the JIT compiler.
637 10. The command "./RunTest list" lists the available tests without actually
638 running any of them. (Because I keep forgetting what they all are.)
642 12. Applied a (slightly modified) user-supplied patch that improves performance
643 when the heap is used for recursion (compiled with --disable-stack-for-
644 recursion). Instead of malloc and free for each heap frame each time a
645 logical recursion happens, frames are retained on a chain and re-used where
646 possible. This sometimes gives as much as 30% improvement.
648 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
649 call.
651 14. As documented, (*COMMIT) is now confined to within a positive assertion.
653 15. It is now possible to link pcretest with libedit as an alternative to
654 libreadline.
656 16. (*COMMIT) control verb is now supported by the JIT compiler.
658 17. The Unicode data tables have been updated to Unicode 6.1.0.
660 18. Added --file-list option to pcregrep.
662 19. Added binary file support to pcregrep, including the -a, --binary-files,
663 -I, and --text options.
665 20. The madvise function is renamed for posix_madvise for QNX compatibility
666 reasons. Fixed by Giuseppe D'Angelo.
668 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
670 22. Changed the option for creating symbolic links for 16-bit man pages from
671 -s to -sf so that re-installing does not cause issues.
673 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
675 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
676 matches in certain environments (the workspace was not being correctly
677 retained). Also added to pcre_dfa_exec() a simple plausibility check on
678 some of the workspace data at the beginning of a restart.
680 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
681 was not doing so when it should - probably a typo introduced by SVN 528
682 (change 8.10/14).
684 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
685 \w+ when the character tables indicated that \x{c4} was a word character.
686 There were several related cases, all because the tests for doing a table
687 lookup were testing for characters less than 127 instead of 255.
689 27. If a pattern contains capturing parentheses that are not used in a match,
690 their slots in the ovector are set to -1. For those that are higher than
691 any matched groups, this happens at the end of processing. In the case when
692 there were back references that the ovector was too small to contain
693 (causing temporary malloc'd memory to be used during matching), and the
694 highest capturing number was not used, memory off the end of the ovector
695 was incorrectly being set to -1. (It was using the size of the temporary
696 memory instead of the true size.)
698 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
699 ovector size, it uses memory at the end of the block that it has got.
701 29. Check for an overlong MARK name and give an error at compile time. The
702 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
704 30. JIT compiler update.
706 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
707 Rill for the patch.
709 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
711 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
713 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
714 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
715 was enabled.
717 35. Improve JIT code generation for greedy plus quantifier.
719 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
720 1 to "aa" instead of to an empty string. The bug affected repeated groups
721 that could potentially match an empty string.
723 37. Optimizing single character iterators in JIT.
725 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
726 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
727 codepoints that are too big for the mode are faulted, and in a UTF mode,
728 disallowed codepoints are also faulted.
730 39. If PCRE was compiled with UTF support, in three places in the DFA
731 matcher there was code that should only have been obeyed in UTF mode, but
732 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
733 processing when bytes with values greater than 127 were present. In 16-bit
734 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
735 both cases the values are those that cannot be the first data item in a UTF
736 character. The three items that might have provoked this were recursions,
737 possessively repeated groups, and atomic groups.
739 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
740 and pcregrep, because some OS require shared objects to be explicitly
741 passed to ld, causing the link step to fail if they are not.
743 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
744 mode, patterns that started with \h* or \R* might be incorrectly matched.
747 Version 8.30 04-February-2012
748 -----------------------------
750 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
751 name is defined in ctype.h.
753 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
754 only in quite long subpatterns.
756 3. Removed the function pcre_info(), which has been obsolete and deprecated
757 since it was replaced by pcre_fullinfo() in February 2000.
759 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
760 match a (*MARK), and the match failed at the start of the subject, a
761 reference to memory before the start of the subject could occur. This bug
762 was introduced by fix 17 of release 8.21.
764 5. A reference to an unset group with zero minimum repetition was giving
765 totally wrong answers (in non-JavaScript-compatibility mode). For example,
766 /(another)?(\1?)test/ matched against "hello world test". This bug was
767 introduced in release 8.13.
769 6. Add support for 16-bit character strings (a large amount of work involving
770 many changes and refactorings).
772 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
773 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
774 from a file.
776 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
777 rounding is not applied in this particular case).
779 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
780 if they appear, or are escaped, in patterns.
782 10. Get rid of a number of -Wunused-but-set-variable warnings.
784 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
785 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
786 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
787 also returns the mark "x". This bug applied to capturing parentheses,
788 non-capturing parentheses, and atomic parentheses. It also applied to some
789 assertions.
791 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
792 information out of configure.ac instead of relying on pcre.h.generic, which
793 is not stored in the repository.
795 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
796 -lreadline.
798 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
800 15. Applied Graycode's patch to put the top-level frame on the stack rather
801 than the heap when not using the stack for recursion. This gives a
802 performance improvement in many cases when recursion is not deep.
804 16. Experimental code added to "pcretest -C" to output the stack frame size.
807 Version 8.21 12-Dec-2011
808 ------------------------
810 1. Updating the JIT compiler.
812 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
813 are added as well.
815 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
816 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
817 calling _pcre_jit_exec. Some extra comments are added.
819 4. (*MARK) settings inside atomic groups that do not contain any capturing
820 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
821 was introduced by change 18 for 8.20.
823 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
824 ECMA-262 standard.
826 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
827 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
828 This bug was probably introduced by change 9 of 8.13.
830 7. While fixing 6 above, I noticed that a number of other items were being
831 incorrectly rejected as "not fixed length". This arose partly because newer
832 opcodes had not been added to the fixed-length checking code. I have (a)
833 corrected the bug and added tests for these items, and (b) arranged for an
834 error to occur if an unknown opcode is encountered while checking for fixed
835 length instead of just assuming "not fixed length". The items that were
836 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
837 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
838 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
840 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
841 being incorrectly compiled and would have given unpredicatble results.
843 9. A possessively repeated subpattern with minimum repeat count greater than
844 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
845 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
846 the first (A) could occur when it should not.
848 10. Add a cast and remove a redundant test from the code.
850 11. JIT should use pcre_malloc/pcre_free for allocation.
852 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
853 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
854 is anything other than /usr, -L is still shown).
856 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
858 14. Perl does not support \N without a following name in a [] class; PCRE now
859 also gives an error.
861 15. If a forward reference was repeated with an upper limit of around 2000,
862 it caused the error "internal error: overran compiling workspace". The
863 maximum number of forward references (including repeats) was limited by the
864 internal workspace, and dependent on the LINK_SIZE. The code has been
865 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
866 the default depends on LINK_SIZE. There is a new upper limit (for safety)
867 of around 200,000 forward references. While doing this, I also speeded up
868 the filling in of repeated forward references.
870 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
871 incorrectly expecting the subject to contain another "a" after the start.
873 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
874 in the match, the SKIP should be ignored. This was not happening; instead
875 the SKIP was being treated as NOMATCH. For patterns such as
876 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
877 tested.
879 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
880 now much more compatible with Perl, in particular in cases where the result
881 is a non-match for a non-anchored pattern. For example, if
882 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
883 "m", where previously it did not return a name. A side effect of this
884 change is that for partial matches, the last encountered mark name is
885 returned, as for non matches. A number of tests that were previously not
886 Perl-compatible have been moved into the Perl-compatible test files. The
887 refactoring has had the pleasing side effect of removing one argument from
888 the match() function, thus reducing its stack requirements.
890 19. If the /S+ option was used in pcretest to study a pattern using JIT,
891 subsequent uses of /S (without +) incorrectly behaved like /S+.
893 21. Retrieve executable code size support for the JIT compiler and fixing
894 some warnings.
896 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
897 not work when the shorter character appeared right at the end of the
898 subject string.
900 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
901 systems.
903 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
904 output it when the /M option is used in pcretest.
906 25. The CheckMan script was not being included in the distribution. Also, added
907 an explicit "perl" to run Perl scripts from the PrepareRelease script
908 because this is reportedly needed in Windows.
910 26. If study data was being save in a file and studying had not found a set of
911 "starts with" bytes for the pattern, the data written to the file (though
912 never used) was taken from uninitialized memory and so caused valgrind to
913 complain.
915 27. Updated RunTest.bat as provided by Sheri Pierce.
917 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
919 29. Computation of memory usage for the table of capturing group names was
920 giving an unnecessarily large value.
923 Version 8.20 21-Oct-2011
924 ------------------------
926 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
927 a POSIX class. After further experiments with Perl, which convinced me that
928 Perl has bugs and confusions, a closing square bracket is no longer allowed
929 in a POSIX name. This bug also affected patterns with classes that started
930 with full stops.
932 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
933 captured substring, but while checking the failing first alternative,
934 substring 1 is temporarily captured. If the output vector supplied to
935 pcre_exec() was not big enough for this capture, the yield of the function
936 was still zero ("insufficient space for captured substrings"). This cannot
937 be totally fixed without adding another stack variable, which seems a lot
938 of expense for a edge case. However, I have improved the situation in cases
939 such as /(a)(b)x|abc/ matched against "abc", where the return code
940 indicates that fewer than the maximum number of slots in the ovector have
941 been set.
943 3. Related to (2) above: when there are more back references in a pattern than
944 slots in the output vector, pcre_exec() uses temporary memory during
945 matching, and copies in the captures as far as possible afterwards. It was
946 using the entire output vector, but this conflicts with the specification
947 that only 2/3 is used for passing back captured substrings. Now it uses
948 only the first 2/3, for compatibility. This is, of course, another edge
949 case.
951 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
952 main code base, and can be used by building with --enable-jit. When this is
953 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
954 runtime --no-jit option is given.
956 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
957 ovector, the return from the function was zero, implying that there were
958 other matches that did not fit. The correct "exactly full" value is now
959 returned.
961 6. If a subpattern that was called recursively or as a subroutine contained
962 (*PRUNE) or any other control that caused it to give a non-standard return,
963 invalid errors such as "Error -26 (nested recursion at the same subject
964 position)" or even infinite loops could occur.
966 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
967 computing the minimum length on reaching *ACCEPT, and so ended up with the
968 wrong value of 1 rather than 0. Further investigation indicates that
969 computing a minimum subject length in the presence of *ACCEPT is difficult
970 (think back references, subroutine calls), and so I have changed the code
971 so that no minimum is registered for a pattern that contains *ACCEPT.
973 8. If (*THEN) was present in the first (true) branch of a conditional group,
974 it was not handled as intended. [But see 16 below.]
976 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
977 Sheri Pierce.
979 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
980 the first byte in a match must be "a".
982 11. Change 17 for 8.13 increased the recursion depth for patterns like
983 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
984 pattern contains any instances of (*THEN). If it does not, the old
985 optimizations are restored. It would be nice to do this on a per-group
986 basis, but at the moment that is not feasible.
988 12. In some environments, the output of pcretest -C is CRLF terminated. This
989 broke RunTest's code that checks for the link size. A single white space
990 character after the value is now allowed for.
992 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
993 For "fr", it uses the Windows-specific input and output files.
995 14. If (*THEN) appeared in a group that was called recursively or as a
996 subroutine, it did not work as intended. [But see next item.]
998 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
999 pattern fragments (but not containing any | characters). If A and B are
1000 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1001 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1002 D. In other words, Perl considers parentheses that do not contain any |
1003 characters to be part of a surrounding alternative, whereas PCRE was
1004 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1005 differently. PCRE now behaves in the same way as Perl, except in the case
1006 of subroutine/recursion calls such as (?1) which have in any case always
1007 been different (but PCRE had them first :-).
1009 16. Related to 15 above: Perl does not treat the | in a conditional group as
1010 creating alternatives. Such a group is treated in the same way as an
1011 ordinary group without any | characters when processing (*THEN). PCRE has
1012 been changed to match Perl's behaviour.
1014 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1015 RunGrepTest script failed.
1017 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1018 inevitable for groups that contain captures, but it can lead to a lot of
1019 stack use in large patterns. The old behaviour has been restored for atomic
1020 groups that do not contain any capturing parentheses.
1022 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1023 suppress the check for a minimum subject length at run time. (If it was
1024 given to pcre_exec() or pcre_dfa_exec() it did work.)
1026 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1027 fail to work when decoding hex characters in data strings in EBCDIC
1028 environments.
1030 21. It appears that in at least one Mac OS environment, the isxdigit() function
1031 is implemented as a macro that evaluates to its argument more than once,
1032 contravening the C 90 Standard (I haven't checked a later standard). There
1033 was an instance in pcretest which caused it to go wrong when processing
1034 \x{...} escapes in subject strings. The has been rewritten to avoid using
1035 things like p++ in the argument of isxdigit().
1038 Version 8.13 16-Aug-2011
1039 ------------------------
1041 1. The Unicode data tables have been updated to Unicode 6.0.0.
1043 2. Two minor typos in pcre_internal.h have been fixed.
1045 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1046 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1047 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1049 4. There were a number of related bugs in the code for matching backrefences
1050 caselessly in UTF-8 mode when codes for the characters concerned were
1051 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1052 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1053 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1054 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1055 2-byte code at the end of the subject (it thought there wasn't enough data
1056 left).
1058 5. Comprehensive information about what went wrong is now returned by
1059 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1060 as the output vector has at least 2 elements. The offset of the start of
1061 the failing character and a reason code are placed in the vector.
1063 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1064 now returned is for the first byte of the failing character, instead of the
1065 last byte inspected. This is an incompatible change, but I hope it is small
1066 enough not to be a problem. It makes the returned offset consistent with
1067 pcre_exec() and pcre_dfa_exec().
1069 7. pcretest now gives a text phrase as well as the error number when
1070 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1071 failure, the offset and reason code are output.
1073 8. When \R was used with a maximizing quantifier it failed to skip backwards
1074 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1075 back over a single character (\n). This seems wrong (because it treated the
1076 two characters as a single entity when going forwards), conflicts with the
1077 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1078 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1079 has been changed.
1081 9. Some internal refactoring has changed the processing so that the handling
1082 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1083 time (the PCRE_DOTALL option was changed this way some time ago: version
1084 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1085 which was always a bit of a fudge. It also means that there is one less
1086 argument for the match() function, which reduces its stack requirements
1087 slightly. This change also fixes an incompatibility with Perl: the pattern
1088 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1090 10. More internal refactoring has drastically reduced the number of recursive
1091 calls to match() for possessively repeated groups such as (abc)++ when
1092 using pcre_exec().
1094 11. While implementing 10, a number of bugs in the handling of groups were
1095 discovered and fixed:
1097 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1098 (a|)*(?1) gave a compile-time internal error.
1099 ((a|)+)+ did not notice that the outer group could match an empty string.
1100 (^a|^)+ was not marked as anchored.
1101 (.*a|.*)+ was not marked as matching at start or after a newline.
1103 12. Yet more internal refactoring has removed another argument from the match()
1104 function. Special calls to this function are now indicated by setting a
1105 value in a variable in the "match data" data block.
1107 13. Be more explicit in pcre_study() instead of relying on "default" for
1108 opcodes that mean there is no starting character; this means that when new
1109 ones are added and accidentally left out of pcre_study(), testing should
1110 pick them up.
1112 14. The -s option of pcretest has been documented for ages as being an old
1113 synonym of -m (show memory usage). I have changed it to mean "force study
1114 for every regex", that is, assume /S for every regex. This is similar to -i
1115 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1116 using it. It makes it easier to run collections of tests with and without
1117 study enabled, and thereby test pcre_study() more easily. All the standard
1118 tests are now run with and without -s (but some patterns can be marked as
1119 "never study" - see 20 below).
1121 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1122 restoration of the capturing data to the outer values was not happening
1123 correctly.
1125 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1126 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1127 pattern had matched an empty string, and so incorrectly returned a no
1128 match.
1130 17. There was optimizing code for the last branch of non-capturing parentheses,
1131 and also for the obeyed branch of a conditional subexpression, which used
1132 tail recursion to cut down on stack usage. Unfortunately, now that there is
1133 the possibility of (*THEN) occurring in these branches, tail recursion is
1134 no longer possible because the return has to be checked for (*THEN). These
1135 two optimizations have therefore been removed. [But see 8.20/11 above.]
1137 18. If a pattern containing \R was studied, it was assumed that \R always
1138 matched two bytes, thus causing the minimum subject length to be
1139 incorrectly computed because \R can also match just one byte.
1141 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1142 was incorrectly computed.
1144 20. If /S is present twice on a test pattern in pcretest input, it now
1145 *disables* studying, thereby overriding the use of -s on the command line
1146 (see 14 above). This is necessary for one or two tests to keep the output
1147 identical in both cases.
1149 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1150 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1152 22. When an atomic group that contained a capturing parenthesis was
1153 successfully matched, but the branch in which it appeared failed, the
1154 capturing was not being forgotten if a higher numbered group was later
1155 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1156 group 1 to "a", when in fact it should be unset. This applied to multi-
1157 branched capturing and non-capturing groups, repeated or not, and also to
1158 positive assertions (capturing in negative assertions does not happen
1159 in PCRE) and also to nested atomic groups.
1161 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1162 subject after a captured substring, to make it easier to tell which of a
1163 number of identical substrings has been captured.
1165 24. The way atomic groups are processed by pcre_exec() has been changed so that
1166 if they are repeated, backtracking one repetition now resets captured
1167 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1168 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1169 "aaa". Previously, it would have been "a". As part of this code
1170 refactoring, the way recursive calls are handled has also been changed.
1172 25. If an assertion condition captured any substrings, they were not passed
1173 back unless some other capturing happened later. For example, if
1174 (?(?=(a))a) was matched against "a", no capturing was returned.
1176 26. When studying a pattern that contained subroutine calls or assertions,
1177 the code for finding the minimum length of a possible match was handling
1178 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1179 group 1 called group 2 while simultaneously a separate group 2 called group
1180 1). A stack overflow occurred in this case. I have fixed this by limiting
1181 the recursion depth to 10.
1183 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1184 Fortmann. This supports explicit test numbers on the command line, and has
1185 argument validation and error reporting.
1187 28. An instance of \X with an unlimited repeat could fail if at any point the
1188 first character it looked at was a mark character.
1190 29. Some minor code refactoring concerning Unicode properties and scripts
1191 should reduce the stack requirement of match() slightly.
1193 30. Added the '=' option to pcretest to check the setting of unused capturing
1194 slots at the end of the pattern, which are documented as being -1, but are
1195 not included in the return count.
1197 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1198 compiled something random. Now it gives a compile-time error (as does
1199 Perl).
1201 32. A *MARK encountered during the processing of a positive assertion is now
1202 recorded and passed back (compatible with Perl).
1204 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1205 had alternative anchored branches, the search for a second match in a line
1206 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1207 matched the line "0102" twice. The same bug affected patterns that started
1208 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1209 twice.
1211 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1212 does, and because of capturing effects, quantifying parenthesized
1213 assertions may at times be useful. Quantifiers are now allowed for
1214 parenthesized assertions.
1216 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1218 36. \g was being checked for fancy things in a character class, when it should
1219 just be a literal "g".
1221 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1222 appearance of a nested POSIX class supersedes an apparent external class.
1223 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1224 unescaped square brackets may also appear as part of class names. For
1225 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1226 more like Perl. (But see 8.20/1 above.)
1228 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1229 was because it thought it was \N{name}, which is not supported).
1231 39. Add minix to OS list not supporting the -S option in pcretest.
1233 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1234 cannot analyze patterns in sufficient detail to catch mutual recursions
1235 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1236 subgroup is called recursively as a subpattern for a second time at the
1237 same position in the subject string. In previous releases this might have
1238 been caught by the recursion limit, or it might have run out of stack.
1240 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1241 happen only once. PCRE was, however incorrectly giving a compile time error
1242 "recursive call could loop indefinitely" because it cannot analyze the
1243 pattern in sufficient detail. The compile time test no longer happens when
1244 PCRE is compiling a conditional subpattern, but actual runaway loops are
1245 now caught at runtime (see 40 above).
1247 42. It seems that Perl allows any characters other than a closing parenthesis
1248 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1249 has been changed to be the same.
1251 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1252 as not to get warnings when autogen.sh is called. Also changed
1253 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1255 44. To help people who use pcregrep to scan files containing exceedingly long
1256 lines, the following changes have been made:
1258 (a) The default value of the buffer size parameter has been increased from
1259 8K to 20K. (The actual buffer used is three times this size.)
1261 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1262 PCRE is built.
1264 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1265 to be set at run time.
1267 (d) Numerical values in pcregrep options can be followed by K or M, for
1268 example --buffer-size=50K.
1270 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1271 given and the return code is set to 2.
1273 45. Add a pointer to the latest mark to the callout data block.
1275 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1276 partial match of an empty string instead of no match. This was specific to
1277 the use of ".".
1279 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1280 complete match instead of a partial match. This bug was dependent on both
1281 the PCRE_UTF8 and PCRE_DOTALL options being set.
1283 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1284 starting byte set, because \b was not being ignored.
1287 Version 8.12 15-Jan-2011
1288 ------------------------
1290 1. Fixed some typos in the markup of the man pages, and wrote a script that
1291 checks for such things as part of the documentation building process.
1293 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1294 --match-limit and --recursion-limit options (added for 8.11). In
1295 particular, this made one of the standard tests fail. (The integer value
1296 went into the wrong half of a long int.)
1298 3. If the --colour option was given to pcregrep with -v (invert match), it
1299 did strange things, either producing crazy output, or crashing. It should,
1300 of course, ignore a request for colour when reporting lines that do not
1301 match.
1303 4. Another pcregrep bug caused similar problems if --colour was specified with
1304 -M (multiline) and the pattern match finished with a line ending.
1306 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1307 matched in multiline mode, the following line was shown as part of the
1308 match. This seems wrong, so I have changed it.
1310 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1311 the check for further matches in the same line (so they could be coloured)
1312 to overrun the end of the current line. If another match was found, it was
1313 incorrectly shown (and then shown again when found in the next line).
1315 7. If pcregrep was compiled under Windows, there was a reference to the
1316 function pcregrep_exit() before it was defined. I am assuming this was
1317 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1318 reported by a user. I've moved the definition above the reference.
1321 Version 8.11 10-Dec-2010
1322 ------------------------
1324 1. (*THEN) was not working properly if there were untried alternatives prior
1325 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1326 backtracked to try for "b" instead of moving to the next alternative branch
1327 at the same level (in this case, to look for "c"). The Perl documentation
1328 is clear that when (*THEN) is backtracked onto, it goes to the "next
1329 alternative in the innermost enclosing group".
1331 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1332 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1333 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1334 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1335 (*THEN).
1337 3. If \s appeared in a character class, it removed the VT character from
1338 the class, even if it had been included by some previous item, for example
1339 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1340 of \s, but is part of the POSIX "space" class.)
1342 4. A partial match never returns an empty string (because you can always
1343 match an empty string at the end of the subject); however the checking for
1344 an empty string was starting at the "start of match" point. This has been
1345 changed to the "earliest inspected character" point, because the returned
1346 data for a partial match starts at this character. This means that, for
1347 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1348 (previously it gave "no match").
1350 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1351 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1352 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1353 has an implication that the given string is incomplete (because a partial
1354 match is preferred over a full match). For this reason, these items now
1355 give a partial match in this situation. [Aside: previously, the one case
1356 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1357 match rather than a full match, which was wrong by the old rules, but is
1358 now correct.]
1360 6. There was a bug in the handling of #-introduced comments, recognized when
1361 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1362 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1363 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1364 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1365 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1366 places in pcre_compile().
1368 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1369 comments when looking ahead for named forward references to subpatterns,
1370 the only newline sequence it recognized was NL. It now handles newlines
1371 according to the set newline convention.
1373 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1374 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1375 cater for a lack of strerror(). These oversights have been fixed.
1377 9. Added --match-limit and --recursion-limit to pcregrep.
1379 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1381 11. When the -o option was used, pcregrep was setting a return code of 1, even
1382 when matches were found, and --line-buffered was not being honoured.
1384 12. Added an optional parentheses number to the -o and --only-matching options
1385 of pcregrep.
1387 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1388 can match an empty string. The code to do it in pcretest and pcredemo
1389 needed fixing:
1391 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1392 only one byte after an empty string match just before CRLF (this case
1393 just got forgotten; "any" and "anycrlf" were OK).
1395 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1396 mode when an empty string match preceded an ASCII character followed by
1397 a non-ASCII character. (The code for advancing by one character rather
1398 than one byte was nonsense.)
1400 (c) The pcredemo.c sample program did not have any code at all to handle
1401 the cases when CRLF is a valid newline sequence.
1403 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1404 as a starting offset was within the subject string. There is now a new
1405 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1406 negative or greater than the length of the string. In order to test this,
1407 pcretest is extended to allow the setting of negative starting offsets.
1409 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1410 starting offset points to the beginning of a UTF-8 character was
1411 unnecessarily clumsy. I tidied it up.
1413 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1414 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1416 17. Nobody had reported that the --include_dir option, which was added in
1417 release 7.7 should have been called --include-dir (hyphen, not underscore)
1418 for compatibility with GNU grep. I have changed it to --include-dir, but
1419 left --include_dir as an undocumented synonym, and the same for
1420 --exclude-dir, though that is not available in GNU grep, at least as of
1421 release 2.5.4.
1423 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1424 characters from a string of bytes) have been redefined so as not to use
1425 loops, in order to improve performance in some environments. At the same
1426 time, I abstracted some of the common code into auxiliary macros to save
1427 repetition (this should not affect the compiled code).
1429 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1430 compile-time error is now given if \c is not followed by an ASCII
1431 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1432 different, and any byte value is allowed.)
1434 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1435 START_OPTIMIZE option, which is now allowed at compile time - but just
1436 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1437 to pcregrep and other applications that have no direct access to PCRE
1438 options. The new /Y option in pcretest sets this option when calling
1439 pcre_compile().
1441 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1442 back references. Groups containing recursive back references were forced to
1443 be atomic by that change, but in the case of named groups, the amount of
1444 memory required was incorrectly computed, leading to "Failed: internal
1445 error: code overflow". This has been fixed.
1447 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1448 pcretest.c, to avoid build problems in some Borland environments.
1451 Version 8.10 25-Jun-2010
1452 ------------------------
1454 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1455 THEN.
1457 2. (*ACCEPT) was not working when inside an atomic group.
1459 3. Inside a character class, \B is treated as a literal by default, but
1460 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1461 causes the error). The code is unchanged, but I tidied the documentation.
1463 4. Inside a character class, PCRE always treated \R and \X as literals,
1464 whereas Perl faults them if its -w option is set. I have changed PCRE so
1465 that it faults them when PCRE_EXTRA is set.
1467 5. Added support for \N, which always matches any character other than
1468 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1470 6. When compiling pcregrep with newer versions of gcc which may have
1471 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1472 declared with attribute warn_unused_result" were given. Just casting the
1473 result to (void) does not stop the warnings; a more elaborate fudge is
1474 needed. I've used a macro to implement this.
1476 7. Minor change to pcretest.c to avoid a compiler warning.
1478 8. Added four artifical Unicode properties to help with an option to make
1479 \s etc use properties (see next item). The new properties are: Xan
1480 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1482 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1483 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1484 this option. Modified pcretest to add /W to test this facility. Added
1485 REG_UCP to make it available via the POSIX interface.
1487 10. Added --line-buffered to pcregrep.
1489 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1490 studied, and the match started with a letter with a code point greater than
1491 127 whose first byte was different to the first byte of the other case of
1492 the letter, the other case of this starting letter was not recognized
1493 (#976).
1495 12. If a pattern that was studied started with a repeated Unicode property
1496 test, for example, \p{Nd}+, there was the theoretical possibility of
1497 setting up an incorrect bitmap of starting bytes, but fortunately it could
1498 not have actually happened in practice until change 8 above was made (it
1499 added property types that matched character-matching opcodes).
1501 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1502 possible starting bytes for non-anchored patterns.
1504 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1505 \R, and also a number of cases that involve Unicode properties, both
1506 explicit and implicit when PCRE_UCP is set.
1508 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1509 input, it could crash or give wrong results if characters with values
1510 greater than 0xc0 were present in the subject string. (Detail: it assumed
1511 UTF-8 input when processing these items.)
1513 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1514 size_t is 64-bit (#991).
1516 17. Added a check for running out of memory when PCRE is compiled with
1517 --disable-stack-for-recursion (#990).
1519 18. If the last data line in a file for pcretest does not have a newline on
1520 the end, a newline was missing in the output.
1522 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1523 less than 128) in its various bitmaps. However, there is a facility for
1524 generating tables according to the current locale when PCRE is compiled. It
1525 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1526 characters, are recognized by isspace() and therefore were getting set in
1527 these tables, and indeed these tables seem to approximate to ISO 8859. This
1528 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1529 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1530 which of course cannot start UTF-8 characters. I have changed the code so
1531 that only real ASCII characters (less than 128) and the correct starting
1532 bytes for UTF-8 encodings are set for characters greater than 127 when in
1533 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1534 altogether.)
1536 20. Added the /T option to pcretest so as to be able to run tests with non-
1537 standard character tables, thus making it possible to include the tests
1538 used for 19 above in the standard set of tests.
1540 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1541 reference to a subpattern the other side of a comment that contains an
1542 opening parenthesis caused either an internal compiling error, or a
1543 reference to the wrong subpattern.
1546 Version 8.02 19-Mar-2010
1547 ------------------------
1549 1. The Unicode data tables have been updated to Unicode 5.2.0.
1551 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1552 configured.
1554 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1555 original author of that file, following a query about its status.
1557 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1558 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1560 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1561 quantifier applied to a forward-referencing subroutine call, could compile
1562 incorrect code or give the error "internal error: previously-checked
1563 referenced subpattern not found".
1565 6. Both MS Visual Studio and Symbian OS have problems with initializing
1566 variables to point to external functions. For these systems, therefore,
1567 pcre_malloc etc. are now initialized to local functions that call the
1568 relevant global functions.
1570 7. There were two entries missing in the vectors called coptable and poptable
1571 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1572 I've fixed the data, and added a kludgy way of testing at compile time that
1573 the lengths are correct (equal to the number of opcodes).
1575 8. Following on from 7, I added a similar kludge to check the length of the
1576 eint vector in pcreposix.c.
1578 9. Error texts for pcre_compile() are held as one long string to avoid too
1579 much relocation at load time. To find a text, the string is searched,
1580 counting zeros. There was no check for running off the end of the string,
1581 which could happen if a new error number was added without updating the
1582 string.
1584 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1586 11. \K was not working if it appeared in an atomic group or in a group that
1587 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1588 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1589 the assertion is positive, but not if it is negative.
1591 12. Change 11 fortuitously reduced the size of the stack frame used in the
1592 "match()" function of pcre_exec.c by one pointer. Forthcoming
1593 implementation of support for (*MARK) will need an extra pointer on the
1594 stack; I have reserved it now, so that the stack frame size does not
1595 decrease.
1597 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1598 item in branch that calls a recursion is a subroutine call - as in the
1599 second branch in the above example - was incorrectly given the compile-
1600 time error "recursive call could loop indefinitely" because pcre_compile()
1601 was not correctly checking the subroutine for matching a non-empty string.
1603 14. The checks for overrunning compiling workspace could trigger after an
1604 overrun had occurred. This is a "should never occur" error, but it can be
1605 triggered by pathological patterns such as hundreds of nested parentheses.
1606 The checks now trigger 100 bytes before the end of the workspace.
1608 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1611 Version 8.01 19-Jan-2010
1612 ------------------------
1614 1. If a pattern contained a conditional subpattern with only one branch (in
1615 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1616 computed the wrong minimum data length (which is of course zero for such
1617 subpatterns). This could cause incorrect "no match" results.
1619 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1620 the pattern is reset in the first branch, pcre_compile() failed with
1621 "internal error: code overflow at offset...". This happened only when
1622 the reset was to the original external option setting. (An optimization
1623 abstracts leading options settings into an external setting, which was the
1624 cause of this.)
1626 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1627 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1628 assertion pattern did not match (meaning that the assertion was true), it
1629 was incorrectly treated as false if the SKIP had been reached during the
1630 matching. This also applied to assertions used as conditions.
1632 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1633 assertion subpattern, including such a pattern used as a condition,
1634 unpredictable results occurred, instead of the error return
1637 5. The C++ GlobalReplace function was not working like Perl for the special
1638 situation when an empty string is matched. It now does the fancy magic
1639 stuff that is necessary.
1641 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1642 removed. (These were left over from very, very early versions of PCRE.)
1644 7. Some cosmetic changes to the code to make life easier when compiling it
1645 as part of something else:
1647 (a) Change DEBUG to PCRE_DEBUG.
1649 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1650 called "current" as "current_branch", to prevent a collision with the
1651 Linux macro when compiled as a kernel module.
1653 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1654 prevent a collision with the Linux macro when compiled as a kernel
1655 module.
1657 8. In pcre_compile() there are some checks for integer overflows that used to
1658 cast potentially large values to (double). This has been changed to that
1659 when building, a check for int64_t is made, and if it is found, it is used
1660 instead, thus avoiding the use of floating point arithmetic. (There is no
1661 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1662 double.
1664 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1665 2005 (difference between two addresses compared to an unsigned value).
1667 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1668 custom one, because of the following reported problem in Windows:
1670 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1671 under Win32.
1672 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1673 therefore missing the function definition.
1674 - The compiler thus generates a "C" signature for the test function.
1675 - The linker fails to find the "C" function.
1676 - PCRE fails to configure if asked to do so against libbz2.
1678 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1679 messages were output:
1681 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1682 rerunning libtoolize, to keep the correct libtool macros in-tree.
1683 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1685 I have done both of these things.
1687 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1688 most of the time, it *can* run out if it is given a pattern that contains a
1689 runaway infinite recursion. I updated the discussion in the pcrestack man
1690 page.
1692 13. Now that we have gone to the x.xx style of version numbers, the minor
1693 version may start with zero. Using 08 or 09 is a bad idea because users
1694 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1695 interpreted as invalid octal numbers. I've updated the previous comment in
1696 configure.ac, and also added a check that gives an error if 08 or 09 are
1697 used.
1699 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1700 causing partial matching to fail when the end of the subject matched \W
1701 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1703 15. There were some discrepancies between the declarations in pcre_internal.h
1704 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1705 their definitions. The declarations used "const uschar *" and the
1706 definitions used USPTR. Even though USPTR is normally defined as "const
1707 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1708 reported that: "This difference in casting confuses some C++ compilers, for
1709 example, SunCC recognizes above declarations as different functions and
1710 generates broken code for hbpcre." I have changed the declarations to use
1711 USPTR.
1713 16. GNU libtool is named differently on some systems. The autogen.sh script now
1714 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1715 (FreeBSD).
1717 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1718 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1719 comment: "Figure out how to create a longlong from a string: strtoll and
1720 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1721 instance, but it only takes 2 args instead of 3!"
1723 18. A subtle bug concerned with back references has been fixed by a change of
1724 specification, with a corresponding code fix. A pattern such as
1725 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1726 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1727 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1728 same bug. Such groups have to be quantified to be useful, or contained
1729 inside another quantified group. (If there's no repetition, the reference
1730 can never match.) The problem arises because, having left the group and
1731 moved on to the rest of the pattern, a later failure that backtracks into
1732 the group uses the captured value from the final iteration of the group
1733 rather than the correct earlier one. I have fixed this in PCRE by forcing
1734 any group that contains a reference to itself to be an atomic group; that
1735 is, there cannot be any backtracking into it once it has completed. This is
1736 similar to recursive and subroutine calls.
1739 Version 8.00 19-Oct-09
1740 ----------------------
1742 1. The table for translating pcre_compile() error codes into POSIX error codes
1743 was out-of-date, and there was no check on the pcre_compile() error code
1744 being within the table. This could lead to an OK return being given in
1745 error.
1747 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1748 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1749 in a Windows environment.
1751 3. The pcregrep --count option prints the count for each file even when it is
1752 zero, as does GNU grep. However, pcregrep was also printing all files when
1753 --files-with-matches was added. Now, when both options are given, it prints
1754 counts only for those files that have at least one match. (GNU grep just
1755 prints the file name in this circumstance, but including the count seems
1756 more useful - otherwise, why use --count?) Also ensured that the
1757 combination -clh just lists non-zero counts, with no names.
1759 4. The long form of the pcregrep -F option was incorrectly implemented as
1760 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1761 but it seems right to fix it, and I didn't think it was worth preserving
1762 the old behaviour.
1764 5. The command line items --regex=pattern and --regexp=pattern were not
1765 recognized by pcregrep, which required --regex pattern or --regexp pattern
1766 (with a space rather than an '='). The man page documented the '=' forms,
1767 which are compatible with GNU grep; these now work.
1769 6. No libpcreposix.pc file was created for pkg-config; there was just
1770 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1772 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1773 when UCP support is not needed, by modifying the Python script that
1774 generates it from Unicode data files. This should not matter if the module
1775 is correctly used as a library, but I received one complaint about 50K of
1776 unwanted data. My guess is that the person linked everything into his
1777 program rather than using a library. Anyway, it does no harm.
1779 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1780 was a minimum greater than 1 for a wide character in a possessive
1781 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1782 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1783 character. Chaos in the form of incorrect output or a compiling loop could
1784 result.
1786 9. The restrictions on what a pattern can contain when partial matching is
1787 requested for pcre_exec() have been removed. All patterns can now be
1788 partially matched by this function. In addition, if there are at least two
1789 slots in the offset vector, the offset of the earliest inspected character
1790 for the match and the offset of the end of the subject are set in them when
1791 PCRE_ERROR_PARTIAL is returned.
1793 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1794 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1795 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1796 and may be more useful for multi-segment matching.
1798 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1799 used to be given if ever the end of the subject was reached; now it is
1800 given only if matching could not proceed because another character was
1801 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1802 string "Z", which now yields "no match" instead of "partial match". In the
1803 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1804 final character ended with (*FAIL).
1806 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1807 if the pattern had a "must contain" character that was already found in the
1808 earlier partial match, unless partial matching was again requested. For
1809 example, with the pattern /dog.(body)?/, the "must contain" character is
1810 "g". If the first part-match was for the string "dog", restarting with
1811 "sbody" failed. This bug has been fixed.
1813 13. The string returned by pcre_dfa_exec() after a partial match has been
1814 changed so that it starts at the first inspected character rather than the
1815 first character of the match. This makes a difference only if the pattern
1816 starts with a lookbehind assertion or \b or \B (\K is not supported by
1817 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1818 matching functions compatible, and I think it's the right thing to do.
1820 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1821 so that the demonstration program is easily available in environments where
1822 PCRE has not been installed from source.
1824 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1825 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1826 library.
1828 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1829 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1830 is not the first non-POSIX option to be added. Clearly some people find
1831 these options useful.
1833 17. If a caller to the POSIX matching function regexec() passes a non-zero
1834 value for nmatch with a NULL value for pmatch, the value of
1835 nmatch is forced to zero.
1837 18. RunGrepTest did not have a test for the availability of the -u option of
1838 the diff command, as RunTest does. It now checks in the same way as
1839 RunTest, and also checks for the -b option.
1841 19. If an odd number of negated classes containing just a single character
1842 interposed, within parentheses, between a forward reference to a named
1843 subpattern and the definition of the subpattern, compilation crashed with
1844 an internal error, complaining that it could not find the referenced
1845 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1846 [The bug was that it was starting one character too far in when skipping
1847 over the character class, thus treating the ] as data rather than
1848 terminating the class. This meant it could skip too much.]
1850 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1851 /g option in pcretest when the pattern contains \K, which makes it possible
1852 to have an empty string match not at the start, even when the pattern is
1853 anchored. Updated pcretest and pcredemo to use this option.
1855 21. If the maximum number of capturing subpatterns in a recursion was greater
1856 than the maximum at the outer level, the higher number was returned, but
1857 with unset values at the outer level. The correct (outer level) value is
1858 now given.
1860 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1861 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1862 make it do so. The string so far is captured, making this feature
1863 compatible with Perl.
1865 23. The tests have been re-organized, adding tests 11 and 12, to make it
1866 possible to check the Perl 5.10 features against Perl 5.10.
1868 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1869 pattern matches a fixed length string. PCRE did not allow this; now it
1870 does. Neither allows recursion.
1872 25. I finally figured out how to implement a request to provide the minimum
1873 length of subject string that was needed in order to match a given pattern.
1874 (It was back references and recursion that I had previously got hung up
1875 on.) This code has now been added to pcre_study(); it finds a lower bound
1876 to the length of subject needed. It is not necessarily the greatest lower
1877 bound, but using it to avoid searching strings that are too short does give
1878 some useful speed-ups. The value is available to calling programs via
1879 pcre_fullinfo().
1881 26. While implementing 25, I discovered to my embarrassment that pcretest had
1882 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1883 study optimizations had never been tested with that matching function.
1884 Oops. What is worse, even when it was passed study data, there was a bug in
1885 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1886 were also very few tests of studied patterns with pcre_dfa_exec().
1888 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1889 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1890 on the other side of the coin, they are no longer allowed to have different
1891 names, because these cannot be distinguished in PCRE, and this has caused
1892 confusion. (This is a difference from Perl.)
1894 28. When duplicate subpattern names are present (necessarily with different
1895 numbers, as required by 27 above), and a test is made by name in a
1896 conditional pattern, either for a subpattern having been matched, or for
1897 recursion in such a pattern, all the associated numbered subpatterns are
1898 tested, and the overall condition is true if the condition is true for any
1899 one of them. This is the way Perl works, and is also more like the way
1900 testing by number works.
1903 Version 7.9 11-Apr-09
1904 ---------------------
1906 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1907 (pcretest), all targets were linked against these libraries. This included
1908 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1909 libraries. This caused unwanted dependencies to be created. This problem
1910 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1911 pcretest is linked with readline.
1913 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1914 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1915 moved outside it again, because FALSE and TRUE are already defined in AIX,
1916 but BOOL is not.
1918 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1919 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1921 4. The pcregrep documentation said spaces were inserted as well as colons (or
1922 hyphens) following file names and line numbers when outputting matching
1923 lines. This is not true; no spaces are inserted. I have also clarified the
1924 wording for the --colour (or --color) option.
1926 5. In pcregrep, when --colour was used with -o, the list of matching strings
1927 was not coloured; this is different to GNU grep, so I have changed it to be
1928 the same.
1930 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1931 each matching line was coloured. Now it goes on to look for further matches
1932 of any of the test patterns, which is the same behaviour as GNU grep.
1934 7. A pattern that could match an empty string could cause pcregrep to loop; it
1935 doesn't make sense to accept an empty string match in pcregrep, so I have
1936 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1937 seems to be how GNU grep behaves. [But see later change 40 for release
1938 8.33.]
1940 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1941 start or after a newline", because the conditional assertion was not being
1942 correctly handled. The rule now is that both the assertion and what follows
1943 in the first alternative must satisfy the test.
1945 9. If auto-callout was enabled in a pattern with a conditional group whose
1946 condition was an assertion, PCRE could crash during matching, both with
1947 pcre_exec() and pcre_dfa_exec().
1949 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1950 used for matching.
1952 11. Unicode property support in character classes was not working for
1953 characters (bytes) greater than 127 when not in UTF-8 mode.
1955 12. Added the -M command line option to pcretest.
1957 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1959 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1961 16. Added comments and documentation about mis-use of no_arg in the C++
1962 wrapper.
1964 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1965 from Martin Jerabek that uses macro names for all relevant character and
1966 string constants.
1968 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1969 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1970 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1971 these, but not everybody uses configure.
1973 19. A conditional group that had only one branch was not being correctly
1974 recognized as an item that could match an empty string. This meant that an
1975 enclosing group might also not be so recognized, causing infinite looping
1976 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1977 with the subject "ab", where knowledge that the repeated group can match
1978 nothing is needed in order to break the loop.
1980 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1981 exec(), but without supplying a callout function, matching went wrong.
1983 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1984 leak if the size of the offset vector was greater than 30. When the vector
1985 is smaller, the saved offsets during recursion go onto a local stack
1986 vector, but for larger vectors malloc() is used. It was failing to free
1987 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1988 error, in fact).
1990 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1991 heapframe that is used only when UTF-8 support is enabled. This caused no
1992 problem, but was untidy.
1994 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1995 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1996 included within another project.
1998 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1999 slightly modified by me:
2001 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2002 not building pcregrep.
2004 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2005 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2007 25. Forward references, both numeric and by name, in patterns that made use of
2008 duplicate group numbers, could behave incorrectly or give incorrect errors,
2009 because when scanning forward to find the reference group, PCRE was not
2010 taking into account the duplicate group numbers. A pattern such as
2011 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2013 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2014 the feature of a custom pointer more persuasive (as requested by a user).
2016 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2017 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2018 pre-defined.
2020 28. Added support for (*UTF8) at the start of a pattern.
2022 29. Arrange for flags added by the "release type" setting in CMake to be shown
2023 in the configuration summary.
2026 Version 7.8 05-Sep-08
2027 ---------------------
2029 1. Replaced UCP searching code with optimized version as implemented for Ad
2030 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2031 stage table and inline lookup instead of a function, giving speed ups of 2
2032 to 5 times on some simple patterns that I tested. Permission was given to
2033 distribute the MultiStage2.py script that generates the tables (it's not in
2034 the tarball, but is in the Subversion repository).
2036 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2037 scripts.
2039 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2040 a group with a zero qualifier. The result of the study could be incorrect,
2041 or the function might crash, depending on the pattern.
2043 4. Caseless matching was not working for non-ASCII characters in back
2044 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2045 It now works when Unicode Property Support is available.
2047 5. In pcretest, an escape such as \x{de} in the data was always generating
2048 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2049 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2050 truncation.
2052 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2054 7. Added two (int) casts to pcregrep when printing the difference of two
2055 pointers, in case they are 64-bit values.
2057 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2058 test 2 if it fails.
2060 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2061 and a #define of that name to empty if it is not externally set. This is to
2062 allow users of MSVC to set it if necessary.
2064 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2065 the convenience functions in the pcre_get.c source file.
2067 11. An option change at the start of a pattern that had top-level alternatives
2068 could cause overwriting and/or a crash. This command provoked a crash in
2069 some environments:
2071 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2073 This potential security problem was recorded as CVE-2008-2371.
2075 12. For a pattern where the match had to start at the beginning or immediately
2076 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2077 pcre_dfa_exec() could read past the end of the passed subject if there was
2078 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2079 pcretest so that it places the subject at the end of its malloc-ed buffer.
2081 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2082 exec() might read past the end of the data buffer in UTF-8 mode.
2084 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2085 the data contained the byte 0x85 as part of a UTF-8 character within its
2086 first line. This applied both to normal and DFA matching.
2088 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2089 /^[^d]*?$/8 failed to match "abc".
2091 16. Added a missing copyright notice to pcrecpp_internal.h.
2093 17. Make it more clear in the documentation that values returned from
2094 pcre_exec() in ovector are byte offsets, not character counts.
2096 18. Tidied a few places to stop certain compilers from issuing warnings.
2098 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2099 supplied by Stefan Weber. I made a further small update for 7.8 because
2100 there is a change of source arrangements: the pcre_searchfuncs.c module is
2101 replaced by pcre_ucd.c.
2104 Version 7.7 07-May-08
2105 ---------------------
2107 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2108 a string to a long long, pretend we don't even have a long long." This is
2109 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2111 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2112 pre-7.6 versions, which defined a global no_arg variable instead of putting
2113 it in the RE class. (See also #8 below.)
2115 3. Remove a line of dead code, identified by coverity and reported by Nuno
2116 Lopes.
2118 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2120 (1) The include/exclude patterns were being applied to the whole pathnames
2121 of files, instead of just to the final components.
2123 (2) If there was more than one level of directory, the subdirectories were
2124 skipped unless they satisfied the include/exclude conditions. This is
2125 inconsistent with GNU grep (and could even be seen as contrary to the
2126 pcregrep specification - which I improved to make it absolutely clear).
2127 The action now is always to scan all levels of directory, and just
2128 apply the include/exclude patterns to regular files.
2130 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2131 --exclude_dir in the tests to avoid scanning .svn directories.
2133 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2134 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2135 doesn't support NULs in patterns.
2137 7. Added some missing "const"s to declarations of static tables in
2138 pcre_compile.c and pcre_dfa_exec.c.
2140 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2141 caused by fix #2 above. (Subsequently also a second patch to fix the
2142 first patch. And a third patch - this was a messy problem.)
2144 9. Applied Craig's patch to remove the use of push_back().
2146 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2147 matching function regexec().
2149 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2150 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2151 references. PCRE supports relative numbers with this syntax (I don't think
2152 Oniguruma does).
2154 12. Previously, a group with a zero repeat such as (...){0} was completely
2155 omitted from the compiled regex. However, this means that if the group
2156 was called as a subroutine from elsewhere in the pattern, things went wrong
2157 (an internal error was given). Such groups are now left in the compiled
2158 pattern, with a new opcode that causes them to be skipped at execution
2159 time.
2161 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2162 to the way PCRE behaves:
2164 (a) A lone ] character is dis-allowed (Perl treats it as data).
2166 (b) A back reference to an unmatched subpattern matches an empty string
2167 (Perl fails the current match path).
2169 (c) A data ] in a character class must be notated as \] because if the
2170 first data character in a class is ], it defines an empty class. (In
2171 Perl it is not possible to have an empty class.) The empty class []
2172 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2173 The negative empty class [^] matches any one character, independently
2174 of the DOTALL setting.
2176 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2177 non-existent subpattern following a character class starting with ']' and
2178 containing () gave an internal compiling error instead of "reference to
2179 non-existent subpattern". Fortunately, when the pattern did exist, the
2180 compiled code was correct. (When scanning forwards to check for the
2181 existence of the subpattern, it was treating the data ']' as terminating
2182 the class, so got the count wrong. When actually compiling, the reference
2183 was subsequently set up correctly.)
2185 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2186 it was being rejected as not supported by pcre_dfa_exec(), even though
2187 other assertions are supported. I have made pcre_dfa_exec() support
2188 (*FAIL).
2190 16. The implementation of 13c above involved the invention of a new opcode,
2191 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2192 cannot be changed at match time, I realized I could make a small
2193 improvement to matching performance by compiling OP_ALLANY instead of
2194 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2195 on the OP_ANY path.
2197 17. Compiling pcretest on Windows with readline support failed without the
2198 following two fixes: (1) Make the unistd.h include conditional on
2199 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2201 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2202 ncurses library to be included for pcretest when ReadLine support is
2203 requested, but also to allow for it to be overridden. This patch came from
2204 Daniel Bergström.
2206 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2207 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2208 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2209 spotting this.
2212 Version 7.6 28-Jan-08
2213 ---------------------
2215 1. A character class containing a very large number of characters with
2216 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2217 overflow.
2219 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2220 HAVE_LONG_LONG is not defined.
2222 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2223 bring them up to date and include new features. This patch includes:
2225 - Fixed PH's badly added libz and libbz2 support.
2226 - Fixed a problem with static linking.
2227 - Added pcredemo. [But later removed - see 7 below.]
2228 - Fixed dftables problem and added an option.
2229 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2231 - Added readline support for pcretest.
2232 - Added an listing of the option settings after cmake has run.
2234 4. A user submitted a patch to Makefile that makes it easy to create
2235 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2236 Makefile.am that cause it to include this special target, without
2237 affecting anything else. Note that the same mingw target plus all
2238 the other distribution libraries and programs are now supported
2239 when configuring with CMake (see 6 below) instead of with
2240 Configure/Make.
2242 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2243 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2244 exported in the Windows port". It has not yet been confirmed that the patch
2245 solves the problem, but it does no harm.
2247 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2248 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2249 with CMake, and also correct the comment about stack recursion.
2251 7. Remove the automatic building of pcredemo from the ./configure system and
2252 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2253 of a program that users should build themselves after PCRE is installed, so
2254 building it automatically is not really right. What is more, it gave
2255 trouble in some build environments.
2257 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2260 Version 7.5 10-Jan-08
2261 ---------------------
2263 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2264 values in parens when parsing an RE using the C++ wrapper."
2266 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2267 Characters greater than 255 were excluded from the class instead of being
2268 included.
2270 3. The same bug as (2) above applied to negated POSIX classes such as
2271 [:^space:].
2273 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2274 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2275 I have changed it.
2277 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2278 first named subpattern) and a construct such as (?&a) would reference the
2279 first named subpattern whose name started with "a" (in other words, the
2280 length check was missing). Both these problems are fixed. "Subpattern name
2281 expected" is now given for (?&) (a zero-length name), and this patch also
2282 makes it give the same error for \k'' (previously it complained that that
2283 was a reference to a non-existent subpattern).
2285 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2286 this is right because (?- can be followed by option settings as well as by
2287 digits. I have, however, made the messages clearer.
2289 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2290 than the number used in the conditional) now cause a compile-time error.
2291 This is actually not compatible with Perl, which accepts such patterns, but
2292 treats the conditional as always being FALSE (as PCRE used to), but it
2293 seems to me that giving a diagnostic is better.
2295 8. Change "alphameric" to the more common word "alphanumeric" in comments
2296 and messages.
2298 9. Fix two occurrences of "backslash" in comments that should have been
2299 "backspace".
2301 10. Remove two redundant lines of code that can never be obeyed (their function
2302 was moved elsewhere).
2304 11. The program that makes PCRE's Unicode character property table had a bug
2305 which caused it to generate incorrect table entries for sequences of
2306 characters that have the same character type, but are in different scripts.
2307 It amalgamated them into a single range, with the script of the first of
2308 them. In other words, some characters were in the wrong script. There were
2309 thirteen such cases, affecting characters in the following ranges:
2311 U+002b0 - U+002c1
2312 U+0060c - U+0060d
2313 U+0061e - U+00612
2314 U+0064b - U+0065e
2315 U+0074d - U+0076d
2316 U+01800 - U+01805
2317 U+01d00 - U+01d77
2318 U+01d9b - U+01dbf
2319 U+0200b - U+0200f
2320 U+030fc - U+030fe
2321 U+03260 - U+0327f
2322 U+0fb46 - U+0fbb1
2323 U+10450 - U+1049d
2325 12. The -o option (show only the matching part of a line) for pcregrep was not
2326 compatible with GNU grep in that, if there was more than one match in a
2327 line, it showed only the first of them. It now behaves in the same way as
2328 GNU grep.
2330 13. If the -o and -v options were combined for pcregrep, it printed a blank
2331 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2332 does the same. The return code can be used to tell if there were any
2333 non-matching lines.
2335 14. Added --file-offsets and --line-offsets to pcregrep.
2337 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2338 infinitely looping recursion. The bug was that positive lookaheads were not
2339 being skipped when checking for a possible empty match (negative lookaheads
2340 and both kinds of lookbehind were skipped).
2342 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2343 inclusion of <windows.h> to before rather than after the definition of
2344 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2346 17. Specifying a possessive quantifier with a specific limit for a Unicode
2347 character property caused pcre_compile() to compile bad code, which led at
2348 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2349 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2350 caused the error; without that there was no problem.
2352 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2354 19. Added --enable-pcretest-libreadline.
2356 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2357 RE::GlobalReplace(). As a result, the number of replacements returned was
2358 double what it should be. I removed one of the increments, but Craig sent a
2359 later patch that removed the other one (the right fix) and added unit tests
2360 that check the return values (which was not done before).
2362 21. Several CMake things:
2364 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2365 the names libpcre and libpcreposix, not just pcre and pcreposix.
2367 (2) The above change means that pcretest and pcregrep are now correctly
2368 linked with the newly-built libraries, not previously installed ones.
2372 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2373 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2374 UTF-8 newline character). The key issue is that the pattern starts .*;
2375 this means that the match must be either at the beginning, or after a
2376 newline. The bug was in the code for advancing after a failed match and
2377 checking that the new position followed a newline. It was not taking
2378 account of UTF-8 characters correctly.
2380 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2381 character classes. PCRE was not treating the sequence [:...:] as a
2382 character class unless the ... were all letters. Perl, however, seems to
2383 allow any characters between [: and :], though of course it rejects as
2384 unknown any "names" that contain non-letters, because all the known class
2385 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2386 for example, whereas PCRE did not - it did not recognize a POSIX character
2387 class. This seemed a bit dangerous, so the code has been changed to be
2388 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2389 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2390 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2391 Perl does, and where it didn't before.
2393 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2394 Windows environments %n is disabled by default.
2397 Version 7.4 21-Sep-07
2398 ---------------------
2400 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2401 means that a class such as [\s] counted as "explicit reference to CR or
2402 LF". That isn't really right - the whole point of the change was to try to
2403 help when there was an actual mention of one of the two characters. So now
2404 the change happens only if \r or \n (or a literal CR or LF) character is
2405 encountered.
2407 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2408 of both had grown to the point where there were only 3 bits left.
2409 Fortunately, there was spare space in the data structure, and so I have
2410 moved the internal flags into a new 16-bit field to free up more option
2411 bits.
2413 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2414 but did not set the internal JCHANGED flag - either of these is enough to
2415 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2416 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2417 start sets both bits.
2419 4. Added options (at build time, compile time, exec time) to change \R from
2420 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2422 5. doc/pcresyntax.html was missing from the distribution.
2424 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2425 compatibility, even though it is no longer used.
2427 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2428 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2429 windows.h file is present (where different names are used). [This was
2430 reversed later after testing - see 16 below.]
2432 8. Changed all #include <config.h> to #include "config.h". There were also
2433 some further <pcre.h> cases that I changed to "pcre.h".
2435 9. When pcregrep was used with the --colour option, it missed the line ending
2436 sequence off the lines that it output.
2438 10. It was pointed out to me that arrays of string pointers cause lots of
2439 relocations when a shared library is dynamically loaded. A technique of
2440 using a single long string with a table of offsets can drastically reduce
2441 these. I have refactored PCRE in four places to do this. The result is
2442 dramatic:
2444 Originally: 290
2445 After changing UCP table: 187
2446 After changing error message table: 43
2447 After changing table of "verbs" 36
2448 After changing table of Posix names 22
2450 Thanks to the folks working on Gregex for glib for this insight.
2452 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2453 unicode-properties was also set.
2455 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2457 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2458 checked only for CRLF.
2460 14. Added casts to pcretest.c to avoid compiler warnings.
2462 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2464 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2465 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2466 entirely. This removes changes made in 7 above.
2468 17. The CMake files have been updated, and there is now more information about
2469 building with CMake in the NON-UNIX-USE document.
2472 Version 7.3 28-Aug-07
2473 ---------------------
2475 1. In the rejigging of the build system that eventually resulted in 7.1, the
2476 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2477 brackets there is not right, since it causes compilers to look for an
2478 installed pcre.h, not the version that is in the source that is being
2479 compiled (which of course may be different). I have changed it back to:
2481 #include "pcre.h"
2483 I have a vague recollection that the change was concerned with compiling in
2484 different directories, but in the new build system, that is taken care of
2485 by the VPATH setting the Makefile.
2487 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2488 when the subject happened to end in the byte 0x85 (e.g. if the last
2489 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2490 characters but of course it shouldn't be taken as a newline when it is part
2491 of another character. The bug was that, for an unlimited repeat of . in
2492 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2493 characters when looking for a newline.
2495 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2497 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2498 in debug output.
2500 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2501 long printing in the pcrecpp unittest when running under MinGW.
2503 6. ESC_K was left out of the EBCDIC table.
2505 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2506 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2507 limit also applies to "virtual nesting" when a pattern is recursive, and in
2508 this case 1000 isn't so big. I have been able to remove this limit at the
2509 expense of backing off one optimization in certain circumstances. Normally,
2510 when pcre_exec() would call its internal match() function recursively and
2511 immediately return the result unconditionally, it uses a "tail recursion"
2512 feature to save stack. However, when a subpattern that can match an empty
2513 string has an unlimited repetition quantifier, it no longer makes this
2514 optimization. That gives it a stack frame in which to save the data for
2515 checking that an empty string has been matched. Previously this was taken
2516 from the 1000-entry workspace that had been reserved. So now there is no
2517 explicit limit, but more stack is used.
2519 8. Applied Daniel's patches to solve problems with the import/export magic
2520 syntax that is required for Windows, and which was going wrong for the
2521 pcreposix and pcrecpp parts of the library. These were overlooked when this
2522 problem was solved for the main library.
2524 9. There were some crude static tests to avoid integer overflow when computing
2525 the size of patterns that contain repeated groups with explicit upper
2526 limits. As the maximum quantifier is 65535, the maximum group length was
2527 set at 30,000 so that the product of these two numbers did not overflow a
2528 32-bit integer. However, it turns out that people want to use groups that
2529 are longer than 30,000 bytes (though not repeat them that many times).
2530 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2531 made it possible to implement the integer overflow checks in a much more
2532 dynamic way, which I have now done. The artificial limitation on group
2533 length has been removed - we now have only the limit on the total length of
2534 the compiled pattern, which depends on the LINK_SIZE setting.
2536 10. Fixed a bug in the documentation for get/copy named substring when
2537 duplicate names are permitted. If none of the named substrings are set, the
2538 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2539 empty string.
2541 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2542 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2543 because the ] is interpreted as the first data character and the
2544 terminating ] is not found. PCRE has been made compatible with Perl in this
2545 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2546 cause memory overwriting.
2548 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2549 string has been matched (to stop an infinite loop). It was not recognizing
2550 a conditional subpattern that could match an empty string if that
2551 subpattern was within another subpattern. For example, it looped when
2552 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2553 condition was not nested. This bug has been fixed.
2555 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2556 past the start of the subject in the presence of bytes with the top bit
2557 set, for example "\x8aBCD".
2559 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2560 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2562 14. Optimized (?!) to (*FAIL).
2564 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2565 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2566 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2567 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2568 does: it's just the validity check that is more restrictive.
2570 16. Inserted checks for integer overflows during escape sequence (backslash)
2571 processing, and also fixed erroneous offset values for syntax errors during
2572 backslash processing.
2574 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2575 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2577 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2578 caused an overrun.
2580 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2581 something other than just ASCII characters) inside a group that had an
2582 unlimited repeat caused a loop at compile time (while checking to see
2583 whether the group could match an empty string).
2585 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2586 [\P{Any}] did so. (Error in the code for printing property names.)
2588 21. An orphan \E inside a character class could cause a crash.
2590 22. A repeated capturing bracket such as (A)? could cause a wild memory
2591 reference during compilation.
2593 23. There are several functions in pcre_compile() that scan along a compiled
2594 expression for various reasons (e.g. to see if it's fixed length for look
2595 behind). There were bugs in these functions when a repeated \p or \P was
2596 present in the pattern. These operators have additional parameters compared
2597 with \d, etc, and these were not being taken into account when moving along
2598 the compiled data. Specifically:
2600 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2601 length.
2603 (b) An item such as \pL+ within a repeated group could cause crashes or
2604 loops.
2606 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2607 "reference to non-existent subpattern" error.
2609 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2611 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2612 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2614 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2616 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2617 character were causing crashes (broken optimization).
2619 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2620 \p or \P) caused a compile-time loop.
2622 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2623 break. For example, the unstudied pattern [\r\n]A does not match the string
2624 "\r\nA" because change 7.0/46 below moves the current point on by two
2625 characters after failing to match at the start. However, the pattern \nA
2626 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2627 the same is true. There doesn't seem any very clean way out of this, but
2628 what I have chosen to do makes the common cases work: PCRE now takes note
2629 of whether there can be an explicit match for \r or \n anywhere in the
2630 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2631 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2632 pattern has explicit CR or LF references.
2634 29. Added (*CR) etc for changing newline setting at start of pattern.
2637 Version 7.2 19-Jun-07
2638 ---------------------
2640 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2641 which is apparently normally available under Windows.
2643 2. Re-jig the pcregrep tests with different newline settings in an attempt
2644 to make them independent of the local environment's newline setting.
2646 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2648 4. Some of the "internals" tests were previously cut out when the link size
2649 was not 2, because the output contained actual offsets. The recent new
2650 "Z" feature of pcretest means that these can be cut out, making the tests
2651 usable with all link sizes.
2653 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2654 stack recursion. This gives a massive performance boost under BSD, but just
2655 a small improvement under Linux. However, it saves one field in the frame
2656 in all cases.
2658 6. Added more features from the forthcoming Perl 5.10:
2660 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2661 recursion call. It refers to the nth most recently opened parentheses.
2663 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2664 to be opened parentheses.
2666 (c) Conditions that refer to capturing parentheses can be specified
2667 relatively, for example, (?(-2)... or (?(+3)...
2669 (d) \K resets the start of the current match so that everything before
2670 is not part of it.
2672 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2674 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2675 reference syntax.
2677 (g) (?| introduces a group in which the numbering of parentheses in each
2678 alternative starts with the same number.
2680 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2682 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2685 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2686 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2687 for detecting groups that can match an empty string.
2689 9. A pattern with a very large number of alternatives (more than several
2690 hundred) was running out of internal workspace during the pre-compile
2691 phase, where pcre_compile() figures out how much memory will be needed. A
2692 bit of new cunning has reduced the workspace needed for groups with
2693 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2694 workspace instead of running out of the 4096 that are available.
2696 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2698 11. Applied patch from Google to remove an optimization that didn't quite work.
2699 The report of the bug said:
2701 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2702 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2703 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2705 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2706 it matched the wrong number of bytes.
2709 Version 7.1 24-Apr-07
2710 ---------------------
2712 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2713 that is more "standard", making use of automake and other Autotools. There
2714 is some re-arrangement of the files and adjustment of comments consequent
2715 on this.
2717 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2718 for recursive directory scanning broke on some systems because the files
2719 are not scanned in any specific order and on different systems the order
2720 was different. A call to "sort" has been inserted into RunGrepTest for the
2721 approprate test as a short-term fix. In the longer term there may be an
2722 alternative.
2724 3. I had an email from Eric Raymond about problems translating some of PCRE's
2725 man pages to HTML (despite the fact that I distribute HTML pages, some
2726 people do their own conversions for various reasons). The problems
2727 concerned the use of low-level troff macros .br and .in. I have therefore
2728 removed all such uses from the man pages (some were redundant, some could
2729 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2730 HTML has been updated to handle .nf/.fi and to complain if it encounters
2731 .br or .in.
2733 4. Updated comments in configure.ac that get placed in config.h.in and also
2734 arranged for config.h to be included in the distribution, with the name
2735 config.h.generic, for the benefit of those who have to compile without
2736 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2738 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2739 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2740 makevp.bat for the current PCRE, using the additional files
2741 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2743 6. A Windows user reported a minor discrepancy with test 2, which turned out
2744 to be caused by a trailing space on an input line that had got lost in his
2745 copy. The trailing space was an accident, so I've just removed it.
2747 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2748 that is needed.
2750 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2751 as "const" (a) because they are and (b) because it helps the PHP
2752 maintainers who have recently made a script to detect big data structures
2753 in the php code that should be moved to the .rodata section. I remembered
2754 to update Builducptable as well, so it won't revert if ucptable.h is ever
2755 re-created.
2757 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2758 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2759 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2760 support is not required. This saves 1.5-2K of code, which is important in
2761 some applications.
2763 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2764 so as not to refer to the tables, even though these functions will never be
2765 called when UTF-8 support is disabled. Otherwise there are problems with a
2766 shared library.
2768 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2770 (a) It was defining its arguments as char * instead of void *.
2772 (b) It was assuming that all moves were upwards in memory; this was true
2773 a long time ago when I wrote it, but is no longer the case.
2775 The emulated memove() is provided for those environments that have neither
2776 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2777 is clearly not the case, as these two bugs were recently reported.
2779 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2780 and Detrail to create the HTML documentation, the .txt form of the man
2781 pages, and it removes trailing spaces from listed files. It also creates
2782 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2783 case, it wraps all the #defines with #ifndefs. This script should be run
2784 before "make dist".
2786 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2787 with Unicode property support.
2789 (a) For a maximizing quantifier, if the two different cases of the
2790 character were of different lengths in their UTF-8 codings (there are
2791 some cases like this - I found 11), and the matching function had to
2792 back up over a mixture of the two cases, it incorrectly assumed they
2793 were both the same length.
2795 (b) When PCRE was configured to use the heap rather than the stack for
2796 recursion during matching, it was not correctly preserving the data for
2797 the other case of a UTF-8 character when checking ahead for a match
2798 while processing a minimizing repeat. If the check also involved
2799 matching a wide character, but failed, corruption could cause an
2800 erroneous result when trying to check for a repeat of the original
2801 character.
2803 13. Some tidying changes to the testing mechanism:
2805 (a) The RunTest script now detects the internal link size and whether there
2806 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2807 values substituted by "configure". (The RunGrepTest script already did
2808 this for UTF-8.) The configure.ac script no longer substitutes the
2809 relevant variables.
2811 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2812 with length and offset values. This means that the output is different
2813 for different internal link sizes. Test 2 is skipped for link sizes
2814 other than 2 because of this, bypassing the problem. Unfortunately,
2815 there was also a test in test 3 (the locale tests) that used /B and
2816 failed for link sizes other than 2. Rather than cut the whole test out,
2817 I have added a new /Z option to pcretest that replaces the length and
2818 offset values with spaces. This is now used to make test 3 independent
2819 of link size. (Test 2 will be tidied up later.)
2821 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2822 segmentation fault instead of returning the appropriate error message.
2824 15. In multiline mode when the newline sequence was set to "any", the pattern
2825 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2826 This doesn't seem right; it now treats the CRLF combination as the line
2827 ending, and so does not match in that case. It's only a pattern such as ^$
2828 that would hit this one: something like ^ABC$ would have failed after \r
2829 and then tried again after \r\n.
2831 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2832 in an attempt to make files that differ only in their line terminators
2833 compare equal. This works on Linux.
2835 17. Under certain error circumstances pcregrep might try to free random memory
2836 as it exited. This is now fixed, thanks to valgrind.
2838 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2839 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2840 was because its rules for how to advance for /g after matching an empty
2841 string at the end of a line did not allow for this case. They now check for
2842 it specially.
2844 20. pcretest is supposed to handle patterns and data of any length, by
2845 extending its buffers when necessary. It was getting this wrong when the
2846 buffer for a data line had to be extended.
2848 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2849 CRLF as a newline sequence.
2851 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2852 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2853 I have nevertheless tidied it up.
2855 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2857 24. Added a man page for pcre-config.
2860 Version 7.0 19-Dec-06
2861 ---------------------
2863 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2864 moving to gcc 4.1.1.
2866 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2867 sys/time.h, which is documented as needed for this function. It doesn't
2868 seem to matter on Linux, but it showed up on some releases of OS X.
2870 3. It seems that there are systems where bytes whose values are greater than
2871 127 match isprint() in the "C" locale. The "C" locale should be the
2872 default when a C program starts up. In most systems, only ASCII printing
2873 characters match isprint(). This difference caused the output from pcretest
2874 to vary, making some of the tests fail. I have changed pcretest so that:
2876 (a) When it is outputting text in the compiled version of a pattern, bytes
2877 other than 32-126 are always shown as hex escapes.
2879 (b) When it is outputting text that is a matched part of a subject string,
2880 it does the same, unless a different locale has been set for the match
2881 (using the /L modifier). In this case, it uses isprint() to decide.
2883 4. Fixed a major bug that caused incorrect computation of the amount of memory
2884 required for a compiled pattern when options that changed within the
2885 pattern affected the logic of the preliminary scan that determines the
2886 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2887 that the computed length was too small. The symptoms of this bug were
2888 either the PCRE error "internal error: code overflow" from pcre_compile(),
2889 or a glibc crash with a message such as "pcretest: free(): invalid next
2890 size (fast)". Examples of patterns that provoked this bug (shown in
2891 pcretest format) are:
2893 /(?-x: )/x
2894 /(?x)(?-x: \s*#\s*)/
2895 /((?i)[\x{c0}])/8
2896 /(?i:[\x{c0}])/8
2898 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2899 is now done differently.
2901 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2902 wrapper classes; (b) implement a new function in the C++ scanner that is
2903 more efficient than the old way of doing things because it avoids levels of
2904 recursion in the regex matching; (c) add a paragraph to the documentation
2905 for the FullMatch() function.
2907 6. The escape sequence \n was being treated as whatever was defined as
2908 "newline". Not only was this contrary to the documentation, which states
2909 that \n is character 10 (hex 0A), but it also went horribly wrong when
2910 "newline" was defined as CRLF. This has been fixed.
2912 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2913 was being set to -1 for the "end of line" case (supposedly a value that no
2914 character can have). Though this value is never used (the check for end of
2915 line is "zero bytes in current character"), it caused compiler complaints.
2916 I've changed it to 0xffffffff.
2918 8. In pcre_version.c, the version string was being built by a sequence of
2919 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2920 string (as it is for production releases) called a macro with an empty
2921 argument. The C standard says the result of this is undefined. The gcc
2922 compiler treats it as an empty string (which was what was wanted) but it is
2923 reported that Visual C gives an error. The source has been hacked around to
2924 avoid this problem.
2926 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2927 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2928 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2929 of them did).
2931 10. Originally, pcretest opened its input and output without "b"; then I was
2932 told that "b" was needed in some environments, so it was added for release
2933 5.0 to both the input and output. (It makes no difference on Unix-like
2934 systems.) Later I was told that it is wrong for the input on Windows. I've
2935 now abstracted the modes into two macros, to make it easier to fiddle with
2936 them, and removed "b" from the input mode under Windows.
2938 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2940 12. Added -help and --help to pcretest as an official way of being reminded
2941 of the options.
2943 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2944 and pcrecpp.cc because they annoy compilers at high warning levels.
2946 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2948 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2949 scripts are not C programs :-) and which was not noticed because it works
2950 on Linux.
2952 16. pcretest is supposed to handle any length of pattern and data line (as one
2953 line or as a continued sequence of lines) by extending its input buffer if
2954 necessary. This feature was broken for very long pattern lines, leading to
2955 a string of junk being passed to pcre_compile() if the pattern was longer
2956 than about 50K.
2958 17. I have done a major re-factoring of the way pcre_compile() computes the
2959 amount of memory needed for a compiled pattern. Previously, there was code
2960 that made a preliminary scan of the pattern in order to do this. That was
2961 OK when PCRE was new, but as the facilities have expanded, it has become
2962 harder and harder to keep it in step with the real compile phase, and there
2963 have been a number of bugs (see for example, 4 above). I have now found a
2964 cunning way of running the real compile function in a "fake" mode that
2965 enables it to compute how much memory it would need, while actually only
2966 ever using a few hundred bytes of working memory and without too many
2967 tests of the mode. This should make future maintenance and development
2968 easier. A side effect of this work is that the limit of 200 on the nesting
2969 depth of parentheses has been removed (though this was never a serious
2970 limitation, I suspect). However, there is a downside: pcre_compile() now
2971 runs more slowly than before (30% or more, depending on the pattern). I
2972 hope this isn't a big issue. There is no effect on runtime performance.
2974 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2975 newline (only possible for the last line of a file) and it was a
2976 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2978 19. Added additional timing features to pcretest. (1) The -tm option now times
2979 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2980 separate command line item, by a number that specifies the number of
2981 repeats to use when timing. The default is 50000; this gives better
2982 precision, but takes uncomfortably long for very large patterns.
2984 20. Extended pcre_study() to be more clever in cases where a branch of a
2985 subpattern has no definite first character. For example, (a*|b*)[cd] would
2986 previously give no result from pcre_study(). Now it recognizes that the
2987 first character must be a, b, c, or d.
2989 21. There was an incorrect error "recursive call could loop indefinitely" if
2990 a subpattern (or the entire pattern) that was being tested for matching an
2991 empty string contained only one non-empty item after a nested subpattern.
2992 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2993 incorrectly, because the \d was being skipped in the check.
2995 22. The pcretest program now has a new pattern option /B and a command line
2996 option -b, which is equivalent to adding /B to every pattern. This causes
2997 it to show the compiled bytecode, without the additional information that
2998 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2999 is the same as /B/I).
3001 23. A new optimization is now able automatically to treat some sequences such
3002 as a*b as a*+b. More specifically, if something simple (such as a character
3003 or a simple class like \d) has an unlimited quantifier, and is followed by
3004 something that cannot possibly match the quantified thing, the quantifier
3005 is automatically "possessified".
3007 24. A recursive reference to a subpattern whose number was greater than 39
3008 went wrong under certain circumstances in UTF-8 mode. This bug could also
3009 have affected the operation of pcre_study().
3011 25. Realized that a little bit of performance could be had by replacing
3012 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3014 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3016 27. Possessive quantifiers such as a++ were previously implemented by turning
3017 them into atomic groups such as ($>a+). Now they have their own opcodes,
3018 which improves performance. This includes the automatically created ones
3019 from 23 above.
3021 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3022 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3023 the first matched character to be a colon. This applied both to named and
3024 numbered groups.
3026 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3028 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3029 building PCRE on Windows easier, so I have included it in the distribution.
3031 31. There is now a check in pcretest against a ridiculously large number being
3032 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3033 loop, the loop is abandoned.
3035 32. Forward references to subpatterns in conditions such as (?(2)...) where
3036 subpattern 2 is defined later cause pcre_compile() to search forwards in
3037 the pattern for the relevant set of parentheses. This search went wrong
3038 when there were unescaped parentheses in a character class, parentheses
3039 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3041 33. "Subroutine" calls and backreferences were previously restricted to
3042 referencing subpatterns earlier in the regex. This restriction has now
3043 been removed.
3045 34. Added a number of extra features that are going to be in Perl 5.10. On the
3046 whole, these are just syntactic alternatives for features that PCRE had
3047 previously implemented using the Python syntax or my own invention. The
3048 other formats are all retained for compatibility.
3050 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3051 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3052 also .NET compatible.
3054 (b) A recursion or subroutine call to a named group can now be defined as
3055 (?&name) as well as (?P>name).
3057 (c) A backreference to a named group can now be defined as \k<name> or
3058 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3059 5.10, are also .NET compatible.
3061 (d) A conditional reference to a named group can now use the syntax
3062 (?(<name>) or (?('name') as well as (?(name).
3064 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3065 groups (named and numbered) that are never evaluated inline, but can be
3066 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3067 is always false. There may be only one alternative in such a group.
3069 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3070 as the simple (?(R). The condition is true only if the most recent
3071 recursion is that of the given number or name. It does not search out
3072 through the entire recursion stack.
3074 (g) The escape \gN or \g{N} has been added, where N is a positive or
3075 negative number, specifying an absolute or relative reference.
3077 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3078 some "unreachable code" warnings.
3080 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3081 things, this adds five new scripts.
3083 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3084 There were also incompatibilities regarding the handling of \Q..\E inside
3085 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3086 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3088 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3089 matches an empty string, and forcibly breaks the loop. There were bugs in
3090 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3091 against aaaa the result was just "a" rather than "aaaa", for example. Two
3092 separate and independent bugs (that affected different cases) have been
3093 fixed.
3095 39. Refactored the code to abolish the use of different opcodes for small
3096 capturing bracket numbers. This is a tidy that I avoided doing when I
3097 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3098 The new approach is not only tidier, it makes it possible to reduce the
3099 memory needed to fix the previous bug (38).
3101 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3102 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3103 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3104 mode.
3106 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3107 report.
3109 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3110 copy construction and assignment in the C++ wrapper.
3112 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3113 couple of bugs that could have given wrong results in the "--newline=crlf"
3114 case.
3116 44. Added a number of casts and did some reorganization of signed/unsigned int
3117 variables following suggestions from Dair Grant. Also renamed the variable
3118 "this" as "item" because it is a C++ keyword.
3120 45. Arranged for dftables to add
3122 #include "pcre_internal.h"
3124 to pcre_chartables.c because without it, gcc 4.x may remove the array
3125 definition from the final binary if PCRE is built into a static library and
3126 dead code stripping is activated.
3128 46. For an unanchored pattern, if a match attempt fails at the start of a
3129 newline sequence, and the newline setting is CRLF or ANY, and the next two
3130 characters are CRLF, advance by two characters instead of one.
3133 Version 6.7 04-Jul-06
3134 ---------------------
3136 1. In order to handle tests when input lines are enormously long, pcretest has
3137 been re-factored so that it automatically extends its buffers when
3138 necessary. The code is crude, but this _is_ just a test program. The
3139 default size has been increased from 32K to 50K.
3141 2. The code in pcre_study() was using the value of the re argument before
3142 testing it for NULL. (Of course, in any sensible call of the function, it
3143 won't be NULL.)
3145 3. The memmove() emulation function in pcre_internal.h, which is used on
3146 systems that lack both memmove() and bcopy() - that is, hardly ever -
3147 was missing a "static" storage class specifier.
3149 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3150 containing an extended class (one that cannot be represented by a bitmap
3151 because it contains high-valued characters or Unicode property items, e.g.
3152 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3153 pattern, but PCRE should not loop if you do not (it no longer does).
3154 [Detail: two cases were found: (a) a repeated subpattern containing an
3155 extended class; (b) a recursive reference to a subpattern that followed a
3156 previous extended class. It wasn't skipping over the extended class
3157 correctly when UTF-8 mode was not set.]
3159 5. A negated single-character class was not being recognized as fixed-length
3160 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3161 compile error "lookbehind assertion is not fixed length".
3163 6. The RunPerlTest auxiliary script was showing an unexpected difference
3164 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3165 write a Perl script that can interpret lines of an input file either as
3166 byte characters or as UTF-8, which is what "perltest" was being required to
3167 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3168 can't do is switch easily at run time between having the "use utf8;" pragma
3169 or not. In the end, I fudged it by using the RunPerlTest script to insert
3170 "use utf8;" explicitly for the UTF-8 tests.
3172 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3173 the end of the subject string, contrary to the documentation and to what
3174 Perl does. This was true of both matching functions. Now it matches only at
3175 the start of the subject and immediately after *internal* newlines.
3177 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3178 a pointer to an int instead of a pointer to an unsigned long int. This
3179 caused problems on 64-bit systems.
3181 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3182 instance of the 'standard' template library not being so standard".
3184 10. There was no check on the number of named subpatterns nor the maximum
3185 length of a subpattern name. The product of these values is used to compute
3186 the size of the memory block for a compiled pattern. By supplying a very
3187 long subpattern name and a large number of named subpatterns, the size
3188 computation could be caused to overflow. This is now prevented by limiting
3189 the length of names to 32 characters, and the number of named subpatterns
3190 to 10,000.
3192 11. Subpatterns that are repeated with specific counts have to be replicated in
3193 the compiled pattern. The size of memory for this was computed from the
3194 length of the subpattern and the repeat count. The latter is limited to
3195 65535, but there was no limit on the former, meaning that integer overflow
3196 could in principle occur. The compiled length of a repeated subpattern is
3197 now limited to 30,000 bytes in order to prevent this.
3199 12. Added the optional facility to have named substrings with the same name.
3201 13. Added the ability to use a named substring as a condition, using the
3202 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3203 are numbers (not recommended). Forward references are permitted.
3205 14. Added forward references in named backreferences (if you see what I mean).
3207 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3208 pattern could run off the end of the subject. For example, the pattern
3209 "(?s)(.{1,5})"8 did this with the subject "ab".
3211 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3212 PCRE_CASELESS was set when matching characters that were quantified with ?
3213 or *.
3215 17. A character class other than a single negated character that had a minimum
3216 but no maximum quantifier - for example [ab]{6,} - was not handled
3217 correctly by pce_dfa_exec(). It would match only one character.
3219 18. A valid (though odd) pattern that looked like a POSIX character
3220 class but used an invalid character after [ (for example [[,abc,]]) caused
3221 pcre_compile() to give the error "Failed: internal error: code overflow" or
3222 in some cases to crash with a glibc free() error. This could even happen if
3223 the pattern terminated after [[ but there just happened to be a sequence of
3224 letters, a binary zero, and a closing ] in the memory that followed.
3226 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3227 over the years. Originally (before any Unicode support), just the bottom 8
3228 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3229 output from "man perlunicode" includes this:
3231 The regular expression compiler produces polymorphic opcodes. That
3232 is, the pattern adapts to the data and automatically switches to
3233 the Unicode character scheme when presented with Unicode data--or
3234 instead uses a traditional byte scheme when presented with byte
3235 data.
3237 Sadly, a wide octal escape does not cause a switch, and in a string with
3238 no other multibyte characters, these octal escapes are treated as before.
3239 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3240 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3241 Unicode string.
3243 I have not perpetrated such confusion in PCRE. Up till now, it took just
3244 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3245 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3246 translate to the appropriate multibyte character.
3248 29. Applied some refactoring to reduce the number of warnings from Microsoft
3249 and Borland compilers. This has included removing the fudge introduced
3250 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3251 a warning about an unused variable.
3253 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3254 characters since release 4.0, because Perl (from release 5.004) does not.
3255 [Or at least, is documented not to: some releases seem to be in conflict
3256 with the documentation.] However, when a pattern was studied with
3257 pcre_study() and all its branches started with \s, PCRE still included VT
3258 as a possible starting character. Of course, this did no harm; it just
3259 caused an unnecessary match attempt.
3261 22. Removed a now-redundant internal flag bit that recorded the fact that case
3262 dependency changed within the pattern. This was once needed for "required
3263 byte" processing, but is no longer used. This recovers a now-scarce options
3264 bit. Also moved the least significant internal flag bit to the most-
3265 significant bit of the word, which was not previously used (hangover from
3266 the days when it was an int rather than a uint) to free up another bit for
3267 the future.
3269 23. Added support for CRLF line endings as well as CR and LF. As well as the
3270 default being selectable at build time, it can now be changed at runtime
3271 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3272 specify that it is scanning data with non-default line endings.
3274 24. Changed the definition of CXXLINK to make it agree with the definition of
3275 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3277 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3278 recursions. This makes a big different to stack usage for some patterns.
3280 26. If a subpattern containing a named recursion or subroutine reference such
3281 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3282 the space required for the compiled pattern went wrong and gave too small a
3283 value. Depending on the environment, this could lead to "Failed: internal
3284 error: code overflow at offset 49" or "glibc detected double free or
3285 corruption" errors.
3287 27. Applied patches from Google (a) to support the new newline modes and (b) to
3288 advance over multibyte UTF-8 characters in GlobalReplace.
3290 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3291 difference for some implementation of PCRE in some Windows version.
3293 29. Added some extra testing facilities to pcretest:
3295 \q<number> in a data line sets the "match limit" value
3296 \Q<number> in a data line sets the "match recursion limt" value
3297 -S <number> sets the stack size, where <number> is in megabytes
3299 The -S option isn't available for Windows.
3302 Version 6.6 06-Feb-06
3303 ---------------------
3305 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3306 in pcreposix.h. I have copied the definition from pcre.h.
3308 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3309 because pcre.h is no longer a built file.
3311 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3312 not normally included in the compiled code.
3315 Version 6.5 01-Feb-06
3316 ---------------------
3318 1. When using the partial match feature with pcre_dfa_exec(), it was not
3319 anchoring the second and subsequent partial matches at the new starting
3320 point. This could lead to incorrect results. For example, with the pattern
3321 /1234/, partially matching against "123" and then "a4" gave a match.
3323 2. Changes to pcregrep:
3325 (a) All non-match returns from pcre_exec() were being treated as failures
3326 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3327 error message is output. Some extra information is given for the
3329 probably the only errors that are likely to be caused by users (by
3330 specifying a regex that has nested indefinite repeats, for instance).
3331 If there are more than 20 of these errors, pcregrep is abandoned.
3333 (b) A binary zero was treated as data while matching, but terminated the
3334 output line if it was written out. This has been fixed: binary zeroes
3335 are now no different to any other data bytes.
3337 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3338 used to set a locale for matching. The --locale=xxxx long option has
3339 been added (no short equivalent) to specify a locale explicitly on the
3340 pcregrep command, overriding the environment variables.
3342 (d) When -B was used with -n, some line numbers in the output were one less
3343 than they should have been.
3345 (e) Added the -o (--only-matching) option.
3347 (f) If -A or -C was used with -c (count only), some lines of context were
3348 accidentally printed for the final match.
3350 (g) Added the -H (--with-filename) option.
3352 (h) The combination of options -rh failed to suppress file names for files
3353 that were found from directory arguments.
3355 (i) Added the -D (--devices) and -d (--directories) options.
3357 (j) Added the -F (--fixed-strings) option.
3359 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3361 (l) Added the --colo(u)r option.
3363 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3364 is not present by default.
3366 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3367 items such as (?R) or (?1), when the recursion could match a number of
3368 alternatives. If it matched one of the alternatives, but subsequently,
3369 outside the recursion, there was a failure, the code tried to back up into
3370 the recursion. However, because of the way PCRE is implemented, this is not
3371 possible, and the result was an incorrect result from the match.
3373 In order to prevent this happening, the specification of recursion has
3374 been changed so that all such subpatterns are automatically treated as
3375 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3377 4. I had overlooked the fact that, in some locales, there are characters for
3378 which isalpha() is true but neither isupper() nor islower() are true. In
3379 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3380 and ordfeminine) are like this. This affected the treatment of \w and \W
3381 when they appeared in character classes, but not when they appeared outside
3382 a character class. The bit map for "word" characters is now created
3383 separately from the results of isalnum() instead of just taking it from the
3384 upper, lower, and digit maps. (Plus the underscore character, of course.)
3386 5. The above bug also affected the handling of POSIX character classes such as
3387 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3388 permanent tables. Instead, the bit maps for such a class were previously
3389 created as the appropriate unions of the upper, lower, and digit bitmaps.
3390 Now they are created by subtraction from the [[:word:]] class, which has
3391 its own bitmap.
3393 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3394 It is created by subtracting the vertical space characters (\x09, \x0a,
3395 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3396 subtraction was done in the overall bitmap for a character class, meaning
3397 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3398 be recognized. This bug has been fixed.
3400 7. Patches from the folks at Google:
3402 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3403 real life, but is still worth protecting against".
3405 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3406 regular expressions".
3408 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3409 have it.
3411 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3412 "configure" and the latter not, in order to fix a problem somebody had
3413 with compiling the Arg class on HP-UX.
3415 (e) Improve the error-handling of the C++ wrapper a little bit.
3417 (f) New tests for checking recursion limiting.
3419 8. The pcre_memmove() function, which is used only if the environment does not
3420 have a standard memmove() function (and is therefore rarely compiled),
3421 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3422 returning a result (though PCRE never actually uses the result).
3424 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3425 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3426 returned instead of calling malloc() with an overflowing number that would
3427 most likely cause subsequent chaos.
3429 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3431 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3432 with this option is matched, the nmatch and pmatch options of regexec() are
3433 ignored.
3435 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3436 provided in case anyone wants to the the POSIX interface with UTF-8
3437 strings.
3439 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3440 C++ linking (needed for some HP-UX environments).
3442 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3443 (unused parameter) and in the pcre_printint() function (omitted "default"
3444 switch label when the default is to do nothing).
3446 15. Added some code to make it possible, when PCRE is compiled as a C++
3447 library, to replace subject pointers for pcre_exec() with a smart pointer
3448 class, thus making it possible to process discontinuous strings.
3450 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3451 much the same function. They were added by different people who were trying
3452 to make PCRE easy to compile on non-Unix systems. It has been suggested
3453 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3454 for compiling on Windows systems. I have therefore replaced it with
3455 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3456 defaults to "extern" for C or "extern C" for C++, which works fine on
3457 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3458 SCOPE with something explicit in config.h. In addition:
3460 (a) pcreposix.h still had just "extern" instead of either of these macros;
3461 I have replaced it with PCRE_DATA_SCOPE.
3463 (b) Functions such as _pcre_xclass(), which are internal to the library,
3464 but external in the C sense, all had PCRE_EXPORT in their definitions.
3465 This is apparently wrong for the Windows case, so I have removed it.
3466 (It makes no difference on Unix-like systems.)
3468 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3469 of recursive calls to match(). This is different to MATCH_LIMIT because
3470 that limits the total number of calls to match(), not all of which increase
3471 the depth of recursion. Limiting the recursion depth limits the amount of
3472 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3473 when PCRE is compiled, and changed at run time. A patch from Google adds
3474 this functionality to the C++ interface.
3476 18. Changes to the handling of Unicode character properties:
3478 (a) Updated the table to Unicode 4.1.0.
3480 (b) Recognize characters that are not in the table as "Cn" (undefined).
3482 (c) I revised the way the table is implemented to a much improved format
3483 which includes recognition of ranges. It now supports the ranges that
3484 are defined in UnicodeData.txt, and it also amalgamates other
3485 characters into ranges. This has reduced the number of entries in the
3486 table from around 16,000 to around 3,000, thus reducing its size
3487 considerably. I realized I did not need to use a tree structure after
3488 all - a binary chop search is just as efficient. Having reduced the
3489 number of entries, I extended their size from 6 bytes to 8 bytes to
3490 allow for more data.
3492 (d) Added support for Unicode script names via properties such as \p{Han}.
3494 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3495 matching that character.
3497 20. When matching a repeated Unicode property with a minimum greater than zero,
3498 (for example \pL{2,}), PCRE could look past the end of the subject if it
3499 reached it while seeking the minimum number of characters. This could
3500 happen only if some of the characters were more than one byte long, because
3501 there is a check for at least the minimum number of bytes.
3503 21. Refactored the implementation of \p and \P so as to be more general, to
3504 allow for more different types of property in future. This has changed the
3505 compiled form incompatibly. Anybody with saved compiled patterns that use
3506 \p or \P will have to recompile them.
3508 22. Added "Any" and "L&" to the supported property types.
3510 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3511 but give a compile time error if the value is greater than 0xff.
3513 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3514 accidentally not being installed or uninstalled.
3516 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3517 made were to insert the current release number. This seemed silly, because
3518 it made things harder for people building PCRE on systems that don't run
3519 "configure". I have turned pcre.h into a distributed file, no longer built
3520 by "configure", with the version identification directly included. There is
3521 no longer a pcre.h.in file.
3523 However, this change necessitated a change to the pcre-config script as
3524 well. It is built from pcre-config.in, and one of the substitutions was the
3525 release number. I have updated configure.ac so that ./configure now finds
3526 the release number by grepping pcre.h.
3528 26. Added the ability to run the tests under valgrind.
3531 Version 6.4 05-Sep-05
3532 ---------------------
3534 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3535 "--" to be printed when multiple files were scanned, even when none of the
3536 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3537 consider it to be a bug, and have restored the previous behaviour.
3539 2. A couple of code tidies to get rid of compiler warnings.
3541 3. The pcretest program used to cheat by referring to symbols in the library
3542 whose names begin with _pcre_. These are internal symbols that are not
3543 really supposed to be visible externally, and in some environments it is
3544 possible to suppress them. The cheating is now confined to including
3545 certain files from the library's source, which is a bit cleaner.
3547 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3548 file's purpose clearer.
3550 5. Reorganized pcre_ucp_findchar().
3553 Version 6.3 15-Aug-05
3554 ---------------------
3556 1. The file libpcre.pc.in did not have general read permission in the tarball.
3558 2. There were some problems when building without C++ support:
3560 (a) If C++ support was not built, "make install" and "make test" still
3561 tried to test it.
3563 (b) There were problems when the value of CXX was explicitly set. Some
3564 changes have been made to try to fix these, and ...
3566 (c) --disable-cpp can now be used to explicitly disable C++ support.
3568 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3569 backslash in a target when C++ was disabled. This confuses some
3570 versions of "make", apparently. Using an intermediate variable solves
3571 this. (Same for CPP_LOBJ.)
3573 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3574 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3575 necessary on certain architectures.
3577 4. Added a setting of -export-symbols-regex to the link command to remove
3578 those symbols that are exported in the C sense, but actually are local
3579 within the library, and not documented. Their names all begin with
3580 "_pcre_". This is not a perfect job, because (a) we have to except some
3581 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3582 available (and never for static libraries). I have made a note to try to
3583 find a way round (a) in the future.
3586 Version 6.2 01-Aug-05
3587 ---------------------
3589 1. There was no test for integer overflow of quantifier values. A construction
3590 such as {1111111111111111} would give undefined results. What is worse, if
3591 a minimum quantifier for a parenthesized subpattern overflowed and became
3592 negative, the calculation of the memory size went wrong. This could have
3593 led to memory overwriting.
3595 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3597 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3598 operating environments where this matters.
3600 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3601 PCRE options from within the C++ wrapper.
3603 5. Named capturing subpatterns were not being correctly counted when a pattern
3604 was compiled. This caused two problems: (a) If there were more than 100
3605 such subpatterns, the calculation of the memory needed for the whole
3606 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3607 back references of the form \12, where the number was greater than 9, were
3608 not recognized as back references, even though there were sufficient
3609 previous subpatterns.
3611 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3612 versions of gcc, e.g. 2.95.4.
3615 Version 6.1 21-Jun-05
3616 ---------------------
3618 1. There was one reference to the variable "posix" in pcretest.c that was not
3619 surrounded by "#if !defined NOPOSIX".
3621 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3622 the cross-check on the old pcre_info() function, for the benefit of the
3623 cut-down version of PCRE that is currently imported into Exim.
3625 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3626 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3627 patterns that start (?i) but I don't think that matters. The use of (?i) is
3628 just an example; this all applies to the other options as well.
3630 4. Since libtool seems to echo the compile commands it is issuing, the output
3631 from "make" can be reduced a bit by putting "@" in front of each libtool
3632 compile command.
3634 5. Patch from the folks at Google for configure.in to be a bit more thorough
3635 in checking for a suitable C++ installation before trying to compile the
3636 C++ stuff. This should fix a reported problem when a compiler was present,
3637 but no suitable headers.
3639 6. The man pages all had just "PCRE" as their title. I have changed them to
3640 be the relevant file name. I have also arranged that these names are
3641 retained in the file doc/pcre.txt, which is a concatenation in text format
3642 of all the man pages except the little individual ones for each function.
3644 7. The NON-UNIX-USE file had not been updated for the different set of source
3645 files that come with release 6. I also added a few comments about the C++
3646 wrapper.
3649 Version 6.0 07-Jun-05
3650 ---------------------
3652 1. Some minor internal re-organization to help with my DFA experiments.
3654 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3655 didn't matter for the library itself when fully configured, but did matter
3656 when compiling without UCP support, or within Exim, where the ucp files are
3657 not imported.
3659 3. Refactoring of the library code to split up the various functions into
3660 different source modules. The addition of the new DFA matching code (see
3661 below) to a single monolithic source would have made it really too
3662 unwieldy, quite apart from causing all the code to be include in a
3663 statically linked application, when only some functions are used. This is
3664 relevant even without the DFA addition now that patterns can be compiled in
3665 one application and matched in another.
3667 The downside of splitting up is that there have to be some external
3668 functions and data tables that are used internally in different modules of
3669 the library but which are not part of the API. These have all had their
3670 names changed to start with "_pcre_" so that they are unlikely to clash
3671 with other external names.
3673 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3674 a different (DFA) algorithm. Although it is slower than the original
3675 function, it does have some advantages for certain types of matching
3676 problem.
3678 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3679 including restarting after a partial match.
3681 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3682 defined when compiling for Windows was sent to me. I have put it into the
3683 code, though I have no means of testing or verifying it.
3685 7. Added the pcre_refcount() auxiliary function.
3687 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3688 match before or at the first newline in the subject string. In pcretest,
3689 the /f option on a pattern can be used to set this.
3691 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3692 would behave wrongly. This has been present in PCRE since release 4.0.
3694 10. A number of changes to the pcregrep command:
3696 (a) Refactored how -x works; insert ^(...)$ instead of setting
3697 PCRE_ANCHORED and checking the length, in preparation for adding
3698 something similar for -w.
3700 (b) Added the -w (match as a word) option.
3702 (c) Refactored the way lines are read and buffered so as to have more
3703 than one at a time available.
3705 (d) Implemented a pcregrep test script.
3707 (e) Added the -M (multiline match) option. This allows patterns to match
3708 over several lines of the subject. The buffering ensures that at least
3709 8K, or the rest of the document (whichever is the shorter) is available
3710 for matching (and similarly the previous 8K for lookbehind assertions).
3712 (f) Changed the --help output so that it now says
3714 -w, --word-regex(p)
3716 instead of two lines, one with "regex" and the other with "regexp"
3717 because that confused at least one person since the short forms are the
3718 same. (This required a bit of code, as the output is generated
3719 automatically from a table. It wasn't just a text change.)
3721 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3722 option but starts with a hyphen. Could be a pattern or a path name
3723 starting with a hyphen, for instance.
3725 (h) "-" can be given as a file name to represent stdin.
3727 (i) When file names are being printed, "(standard input)" is used for
3728 the standard input, for compatibility with GNU grep. Previously
3729 "<stdin>" was used.
3731 (j) The option --label=xxx can be used to supply a name to be used for
3732 stdin when file names are being printed. There is no short form.
3734 (k) Re-factored the options decoding logic because we are going to add
3735 two more options that take data. Such options can now be given in four
3736 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3738 (l) Added the -A, -B, and -C options for requesting that lines of context
3739 around matches be printed.
3741 (m) Added the -L option to print the names of files that do not contain
3742 any matching lines, that is, the complement of -l.
3744 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3745 continue to scan other files.
3747 (o) The -s option was incorrectly implemented. For compatibility with other
3748 greps, it now suppresses the error message for a non-existent or non-
3749 accessible file (but not the return code). There is a new option called
3750 -q that suppresses the output of matching lines, which was what -s was
3751 previously doing.
3753 (p) Added --include and --exclude options to specify files for inclusion
3754 and exclusion when recursing.
3756 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3757 Hopefully, it now does.
3759 12. Missing cast in pcre_study().
3761 13. Added an "uninstall" target to the makefile.
3763 14. Replaced "extern" in the function prototypes in Makefile.in with
3764 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3765 world, but is set differently for Windows.
3767 15. Added a second compiling function called pcre_compile2(). The only
3768 difference is that it has an extra argument, which is a pointer to an
3769 integer error code. When there is a compile-time failure, this is set
3770 non-zero, in addition to the error test pointer being set to point to an
3771 error message. The new argument may be NULL if no error number is required
3772 (but then you may as well call pcre_compile(), which is now just a
3773 wrapper). This facility is provided because some applications need a
3774 numeric error indication, but it has also enabled me to tidy up the way
3775 compile-time errors are handled in the POSIX wrapper.
3777 16. Added VPATH=.libs to the makefile; this should help when building with one
3778 prefix path and installing with another. (Or so I'm told by someone who
3779 knows more about this stuff than I do.)
3781 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3782 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3783 match everything, including newlines. This is not POSIX-compatible, but
3784 somebody wanted the feature. From pcretest it can be activated by using
3785 both the P and the s flags.
3787 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3789 19. libpcre.pc was being incorrectly installed as executable.
3791 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3792 it now also looks for '\r' so that it will work unmodified on Windows.
3794 21. Added Google's contributed C++ wrapper to the distribution.
3796 22. Added some untidy missing memory free() calls in pcretest, to keep
3797 Electric Fence happy when testing.
3801 Version 5.0 13-Sep-04
3802 ---------------------
3804 1. Internal change: literal characters are no longer packed up into items
3805 containing multiple characters in a single byte-string. Each character
3806 is now matched using a separate opcode. However, there may be more than one
3807 byte in the character in UTF-8 mode.
3809 2. The pcre_callout_block structure has two new fields: pattern_position and
3810 next_item_length. These contain the offset in the pattern to the next match
3811 item, and its length, respectively.
3813 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3814 insertion of callouts before each pattern item. Added the /C option to
3815 pcretest to make use of this.
3817 4. On the advice of a Windows user, the lines
3819 #if defined(_WIN32) || defined(WIN32)
3820 _setmode( _fileno( stdout ), 0x8000 );
3821 #endif /* defined(_WIN32) || defined(WIN32) */
3823 have been added to the source of pcretest. This apparently does useful
3824 magic in relation to line terminators.
3826 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3827 for the benefit of those environments where the "b" makes a difference.
3829 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3830 to know about it. I have put a hack into configure.in that adds in code
3831 to set GCC=yes if CC=icc. This seems to end up at a point in the
3832 generated configure script that is early enough to affect the setting of
3833 compiler options, which is what is needed, but I have no means of testing
3834 whether it really works. (The user who reported this had patched the
3835 generated configure script, which of course I cannot do.)
3837 LATER: After change 22 below (new libtool files), the configure script
3838 seems to know about icc (and also ecc). Therefore, I have commented out
3839 this hack in configure.in.
3841 7. Added support for pkg-config (2 patches were sent in).
3843 8. Negated POSIX character classes that used a combination of internal tables
3844 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3845 [[:^ascii]]. Typically, they would match almost any characters. The other
3846 POSIX classes were not broken in this way.
3848 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3849 to find the match, as PCRE was deluded into thinking that the match had to
3850 start at the start point or following a newline. The same bug applied to
3851 patterns with negative forward assertions or any backward assertions
3852 preceding ".*" at the start, unless the pattern required a fixed first
3853 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3855 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3856 starting at the last subject character, bytes beyond the end of the subject
3857 string were read.
3859 11. Renamed the variable "class" as "classbits" to make life easier for C++
3860 users. (Previously there was a macro definition, but it apparently wasn't
3861 enough.)
3863 12. Added the new field "tables" to the extra data so that tables can be passed
3864 in at exec time, or the internal tables can be re-selected. This allows
3865 a compiled regex to be saved and re-used at a later time by a different
3866 program that might have everything at different addresses.
3868 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3869 -R library as well as a -L library.
3871 14. The debugging options of pcretest (-d on the command line or D on a
3872 pattern) showed incorrect output for anything following an extended class
3873 that contained multibyte characters and which was followed by a quantifier.
3875 15. Added optional support for general category Unicode character properties
3876 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3877 support. It adds about 90K to the size of the library. The meanings of the
3878 inbuilt class escapes such as \d and \s have NOT been changed.
3880 16. Updated pcredemo.c to include calls to free() to release the memory for the
3881 compiled pattern.
3883 17. The generated file chartables.c was being created in the source directory
3884 instead of in the building directory. This caused the build to fail if the
3885 source directory was different from the building directory, and was
3886 read-only.
3888 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3889 file. No doubt somebody will tell me if they don't make sense... Also added
3890 Dan Mooney's comments about building on OpenVMS.
3892 19. Added support for partial matching via the PCRE_PARTIAL option for
3893 pcre_exec() and the \P data escape in pcretest.
3895 20. Extended pcretest with 3 new pattern features:
3897 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3898 write the compiled pattern to the file whose name is "rest-of-line".
3899 This is a straight binary dump of the data, with the saved pointer to
3900 the character tables forced to be NULL. The study data, if any, is
3901 written too. After writing, pcretest reads a new pattern.
3903 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3904 compiled pattern from the given file. There must not be any
3905 occurrences of "<" in the file name (pretty unlikely); if there are,
3906 pcretest will instead treat the initial "<" as a pattern delimiter.
3907 After reading in the pattern, pcretest goes on to read data lines as
3908 usual.
3910 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3911 and 16-bit fields in a compiled pattern, to simulate a pattern that
3912 was compiled on a host of opposite endianness.
3914 21. The pcre-exec() function can now cope with patterns that were compiled on
3915 hosts of opposite endianness, with this restriction:
3917 As for any compiled expression that is saved and used later, the tables
3918 pointer field cannot be preserved; the extra_data field in the arguments
3919 to pcre_exec() should be used to pass in a tables address if a value
3920 other than the default internal tables were used at compile time.
3922 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3923 now diagnosed as an error. Previously, most of the time, a negative number
3924 would have been treated as zero, but if in addition "ovector" was passed as
3925 NULL, a crash could occur.
3927 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3928 new versions from the libtool 1.5 distribution (the last one is a copy of
3929 a file called libtool.m4). This seems to have fixed the need to patch
3930 "configure" to support Darwin 1.3 (which I used to do). However, I still
3931 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3932 workstation).
3934 24. Changed the PCRE licence to be the more standard "BSD" licence.
3937 Version 4.5 01-Dec-03
3938 ---------------------
3940 1. There has been some re-arrangement of the code for the match() function so
3941 that it can be compiled in a version that does not call itself recursively.
3942 Instead, it keeps those local variables that need separate instances for
3943 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3944 needs to "recurse". Keeping track of where control must go is done by means
3945 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3946 hide most of the details from the main code, and operates only if
3947 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3948 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3949 operating.
3951 To make it easier for callers to provide specially tailored get/free
3952 functions for this usage, two new functions, pcre_stack_malloc, and
3953 pcre_stack_free, are used. They are always called in strict stacking order,
3954 and the size of block requested is always the same.
3956 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3957 PCRE has been compiled to use the stack or the heap for recursion. The
3958 -C option of pcretest uses this to show which version is compiled.
3960 A new data escape \S, is added to pcretest; it causes the amounts of store
3961 obtained and freed by both kinds of malloc/free at match time to be added
3962 to the output.
3964 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3965 what's available on my current Linux desktop machine.
3967 3. When matching a UTF-8 string, the test for a valid string at the start has
3968 been extended. If start_offset is not zero, PCRE now checks that it points
3969 to a byte that is the start of a UTF-8 character. If not, it returns
3970 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3971 this is necessary because there may be backward assertions in the pattern.
3972 When matching the same subject several times, it may save resources to use
3973 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3975 4. The code for checking the validity of UTF-8 strings has been tightened so
3976 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3977 containing "overlong sequences".
3979 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3980 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3981 should have been "&", but it just so happened that all the cases this let
3982 through by mistake were picked up later in the function.
3984 6. I had used a variable called "isblank" - this is a C99 function, causing
3985 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3987 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3988 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3989 so the version is shown; (c) stop "make test" echoing "./RunTest".
3991 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3993 9. The prototype for memmove() for systems that don't have it was using
3994 size_t, but the inclusion of the header that defines size_t was later. I've
3995 moved the #includes for the C headers earlier to avoid this.
3997 10. Added some adjustments to the code to make it easier to compiler on certain
3998 special systems:
4000 (a) Some "const" qualifiers were missing.
4001 (b) Added the macro EXPORT before all exported functions; by default this
4002 is defined to be empty.
4003 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4004 that it reads its output file name as an argument instead of writing
4005 to the standard output and assuming this can be redirected.
4007 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4008 class containing characters with values greater than 255, PCRE compilation
4009 went into a loop.
4011 12. A recursive reference to a subpattern that was within another subpattern
4012 that had a minimum quantifier of zero caused PCRE to crash. For example,
4013 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4014 recursion. If the recursively-called subpattern itself had a zero repeat,
4015 that was OK.
4017 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4018 buffer into which it was copied (for escape processing) was still set at
4019 1024, so long lines caused crashes.
4021 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4022 "internal error: code overflow...". This applied to any character class
4023 that was followed by a possessive quantifier.
4025 15. Modified the Makefile to add libpcre.la as a prerequisite for
4026 libpcreposix.la because I was told this is needed for a parallel build to
4027 work.
4029 16. If a pattern that contained .* following optional items at the start was
4030 studied, the wrong optimizing data was generated, leading to matching
4031 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4032 matching string must start with a or b or c. The correct conclusion for
4033 this pattern is that a match can start with any character.
4036 Version 4.4 13-Aug-03
4037 ---------------------
4039 1. In UTF-8 mode, a character class containing characters with values between
4040 127 and 255 was not handled correctly if the compiled pattern was studied.
4041 In fixing this, I have also improved the studying algorithm for such
4042 classes (slightly).
4044 2. Three internal functions had redundant arguments passed to them. Removal
4045 might give a very teeny performance improvement.
4047 3. Documentation bug: the value of the capture_top field in a callout is *one
4048 more than* the number of the hightest numbered captured substring.
4050 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4051 in incorrectly linking with a previously installed version. They now link
4052 explicitly with libpcre.la.
4054 5. configure.in no longer needs to recognize Cygwin specially.
4056 6. A problem in pcre.in for Windows platforms is fixed.
4058 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4059 pcretest, it used to include the size of the study block as part of its
4060 output. Unfortunately, the structure contains a field that has a different
4061 size on different hardware architectures. This meant that the tests that
4062 showed this size failed. As the block is currently always of a fixed size,
4063 this information isn't actually particularly useful in pcretest output, so
4064 I have just removed it.
4066 8. Three pre-processor statements accidentally did not start in column 1.
4067 Sadly, there are *still* compilers around that complain, even though
4068 standard C has not required this for well over a decade. Sigh.
4070 9. In pcretest, the code for checking callouts passed small integers in the
4071 callout_data field, which is a void * field. However, some picky compilers
4072 complained about the casts involved for this on 64-bit systems. Now
4073 pcretest passes the address of the small integer instead, which should get
4074 rid of the warnings.
4076 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4077 both compile and run time, and gives an error if an invalid UTF-8 sequence
4078 is found. There is a option for disabling this check in cases where the
4079 string is known to be correct and/or the maximum performance is wanted.
4081 11. In response to a bug report, I changed one line in Makefile.in from
4083 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4084 to
4085 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4087 to look similar to other lines, but I have no way of telling whether this
4088 is the right thing to do, as I do not use Windows. No doubt I'll get told
4089 if it's wrong...
4092 Version 4.3 21-May-03
4093 ---------------------
4095 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4096 Makefile.
4098 2. Some refactoring to improve the quality of the code:
4100 (i) The utf8_table... variables are now declared "const".
4102 (ii) The code for \cx, which used the "case flipping" table to upper case
4103 lower case letters, now just substracts 32. This is ASCII-specific,
4104 but the whole concept of \cx is ASCII-specific, so it seems
4105 reasonable.
4107 (iii) PCRE was using its character types table to recognize decimal and
4108 hexadecimal digits in the pattern. This is silly, because it handles
4109 only 0-9, a-f, and A-F, but the character types table is locale-
4110 specific, which means strange things might happen. A private
4111 table is now used for this - though it costs 256 bytes, a table is
4112 much faster than multiple explicit tests. Of course, the standard
4113 character types table is still used for matching digits in subject
4114 strings against \d.
4116 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4117 ending in _t are). So I've renamed it as ESC_tee.
4119 3. The first argument for regexec() in the POSIX wrapper should have been
4120 defined as "const".
4122 4. Changed pcretest to use malloc() for its buffers so that they can be
4123 Electric Fenced for debugging.
4125 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4126 to read one or more bytes before the start of the subject string. Often this
4127 had no effect on PCRE's behaviour, but in some circumstances it could
4128 provoke a segmentation fault.
4130 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4131 to try to read one or more bytes before the start of the subject string.
4133 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4134 UTF-8 support could misbehave in various ways if the subject string
4135 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4136 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4137 back over UTF-8 characters.)
4140 Version 4.2 14-Apr-03
4141 ---------------------
4143 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4145 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4146 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4147 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4148 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4149 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4151 Note: automatic setting of the BUILD variables is not yet working
4152 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4153 compile-time but not at link-time
4154 [LINK]: use for linking executables only
4155 make different versions for Windows and non-Windows
4156 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4157 libraries
4158 [LINK_FOR_BUILD]: new variable
4159 [OBJEXT]: use throughout
4160 [EXEEXT]: use throughout
4161 <winshared>: new target
4162 <wininstall>: new target
4163 <dftables.o>: use native compiler
4164 <dftables>: use native linker
4165 <install>: handle Windows platform correctly
4166 <clean>: ditto
4167 <check>: ditto
4168 copy DLL to top builddir before testing
4170 As part of these changes, -no-undefined was removed again. This was reported
4171 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4172 in any case.
4174 3. Some tidies to get rid of compiler warnings:
4176 . In the match_data structure, match_limit was an unsigned long int, whereas
4177 match_call_count was an int. I've made them both unsigned long ints.
4179 . In pcretest the fact that a const uschar * doesn't automatically cast to
4180 a void * provoked a warning.
4182 . Turning on some more compiler warnings threw up some "shadow" variables
4183 and a few more missing casts.
4185 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4186 option, a class that contained a single character with a value between 128
4187 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4189 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4190 option, a class that contained several characters, but with at least one
4191 whose value was between 128 and 255 caused PCRE to crash.
4194 Version 4.1 12-Mar-03
4195 ---------------------
4197 1. Compiling with gcc -pedantic found a couple of places where casts were
4198 needed, and a string in dftables.c that was longer than standard compilers are
4199 required to support.
4201 2. Compiling with Sun's compiler found a few more places where the code could
4202 be tidied up in order to avoid warnings.
4204 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4205 first of these names is deprecated in the latest Autoconf in favour of the name
4206 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4207 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4208 analogy I have changed it to CFLAGS_FOR_BUILD.
4210 4. Added -no-undefined to the linking command in the Makefile, because this is
4211 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4212 linking step for the pcreposix library.
4214 5. PCRE was failing to diagnose the case of two named groups with the same
4215 name.
4217 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4218 literal character that is needed in the subject for a match, and scans along to
4219 ensure that it is present before embarking on the full matching process. This
4220 saves time in cases of nested unlimited repeats that are never going to match.
4221 Problem: the scan can take a lot of time if the subject is very long (e.g.
4222 megabytes), thus penalizing straightforward matches. It is now done only if the
4223 amount of subject to be scanned is less than 1000 bytes.
4225 7. A lesser problem with the same optimization is that it was recording the
4226 first character of an anchored pattern as "needed", thus provoking a search
4227 right along the subject, even when the first match of the pattern was going to
4228 fail. The "needed" character is now not set for anchored patterns, unless it
4229 follows something in the pattern that is of non-fixed length. Thus, it still
4230 fulfils its original purpose of finding quick non-matches in cases of nested
4231 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4234 Version 4.0 17-Feb-03
4235 ---------------------
4237 1. If a comment in an extended regex that started immediately after a meta-item
4238 extended to the end of string, PCRE compiled incorrect data. This could lead to
4239 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4241 2. Moved to autoconf 2.53 and libtool 1.4.2.
4243 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4244 the special perltest8 script is no longer needed - all the tests can be run
4245 from a single perltest script.
4247 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4248 by \s. It has now been removed in PCRE. This means it isn't recognized as
4249 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4250 class [:space:] *does* include VT, thereby creating a mess.
4252 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4253 space and tab.
4255 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4256 its new features into the main test script, reducing the number of scripts.
4258 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4259 were backward compatible, and made the (?i) apply to the whole pattern, as if
4260 /i were given. Now it behaves more logically, and applies the option setting
4261 only to what follows. PCRE has been changed to follow suit. However, if it
4262 finds options settings right at the start of the pattern, it extracts them into
4263 the global options, as before. Thus, they show up in the info data.
4265 8. Added support for the \Q...\E escape sequence. Characters in between are
4266 treated as literals. This is slightly different from Perl in that $ and @ are
4267 also handled as literals inside the quotes. In Perl, they will cause variable
4268 interpolation. Note the following examples:
4270 Pattern PCRE matches Perl matches
4272 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4273 \Qabc\$xyz\E abc\$xyz abc\$xyz
4274 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4276 For compatibility with Perl, \Q...\E sequences are recognized inside character
4277 classes as well as outside them.
4279 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4280 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4281 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4282 signed/unsigned warnings.
4284 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4285 option for pcretest, so I've replaced it by a simple function that does just
4286 that job.
4288 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4289 "pcregrep -".
4291 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4292 Java package. This provides some syntactic sugar for simple cases of what my
4293 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4294 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4295 item, you can use this simplified notation. Note that only makes sense with
4296 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4297 greediness, whatever the setting of the PCRE_UNGREEDY option.
4299 13. A change of greediness default within a pattern was not taking effect at
4300 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4301 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4302 was abstracted outside.
4304 14. PCRE now supports the \G assertion. It is true when the current matching
4305 position is at the start point of the match. This differs from \A when the
4306 starting offset is non-zero. Used with the /g option of pcretest (or similar
4307 code), it works in the same way as it does for Perl's /g option. If all
4308 alternatives of a regex begin with \G, the expression is anchored to the start
4309 match position, and the "anchored" flag is set in the compiled expression.
4311 15. Some bugs concerning the handling of certain option changes within patterns
4312 have been fixed. These applied to options other than (?ims). For example,
4313 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4314 been the other way round. Some of this was related to change 7 above.
4316 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4317 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4318 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4319 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4321 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4322 mode. Unlike ".", it always matches newline, whatever the setting of
4323 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4324 assertions. Perl allows it, but it doesn't (in general) work because it can't
4325 calculate the length of the lookbehind. At least, that's the case for Perl
4326 5.8.0 - I've been told they are going to document that it doesn't work in
4327 future.
4329 18. Added an error diagnosis for escapes that PCRE does not support: these are
4330 \L, \l, \N, \P, \p, \U, \u, and \X.
4332 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4333 reading past the end of the pattern in cases such as /[abcd/.
4335 20. PCRE was getting more memory than necessary for patterns with classes that
4336 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4338 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4339 compiling PCRE for use with Virtual Pascal.
4341 22. Small fix to the Makefile to make it work properly if the build is done
4342 outside the source tree.
4344 23. Added a new extension: a condition to go with recursion. If a conditional
4345 subpattern starts with (?(R) the "true" branch is used if recursion has
4346 happened, whereas the "false" branch is used only at the top level.
4348 24. When there was a very long string of literal characters (over 255 bytes
4349 without UTF support, over 250 bytes with UTF support), the computation of how
4350 much memory was required could be incorrect, leading to segfaults or other
4351 strange effects.
4353 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4354 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4355 there was a subsequent back reference to those brackets. This meant that, for
4356 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4357 possible to check for precisely this case. All we can do is abandon the
4358 optimization if .* occurs inside capturing brackets when there are any back
4359 references whatsoever. (See below for a better fix that came later.)
4361 26. The handling of the optimization for finding the first character of a
4362 non-anchored pattern, and for finding a character that is required later in the
4363 match were failing in some cases. This didn't break the matching; it just
4364 failed to optimize when it could. The way this is done has been re-implemented.
4366 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4368 28. Added a new feature that provides some of the functionality that Perl
4369 provides with (?{...}). The facility is termed a "callout". The way it is done
4370 in PCRE is for the caller to provide an optional function, by setting
4371 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4372 global variable. By default it is unset, which disables all calling out. To get
4373 the function called, the regex must include (?C) at appropriate points. This
4374 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4375 This provides a means of identifying different callout points. When PCRE
4376 reaches such a point in the regex, if pcre_callout has been set, the external
4377 function is called. It is provided with data in a structure called
4378 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4379 matching continues; if it returns a non-zero value, the match at the current
4380 point fails. However, backtracking will occur if possible. [This was changed
4381 later and other features added - see item 49 below.]
4383 29. pcretest is upgraded to test the callout functionality. It provides a
4384 callout function that displays information. By default, it shows the start of
4385 the match and the current position in the text. There are some new data escapes
4386 to vary what happens:
4388 \C+ in addition, show current contents of captured substrings
4389 \C- do not supply a callout function
4390 \C!n return 1 when callout number n is reached
4391 \C!n!m return 1 when callout number n is reached for the mth time
4393 30. If pcregrep was called with the -l option and just a single file name, it
4394 output "<stdin>" if a match was found, instead of the file name.
4396 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4397 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4398 pcre_exec(). This saves a malloc/free per call. The default value of
4399 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4400 when configuring.
4402 32. The default maximum size of a compiled pattern is 64K. There have been a
4403 few cases of people hitting this limit. The code now uses macros to handle the
4404 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4405 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4406 configuring. Tests 2 and 5 work only with 2-byte links because they output
4407 debugging information about compiled patterns.
4409 33. Internal code re-arrangements:
4411 (a) Moved the debugging function for printing out a compiled regex into
4412 its own source file (printint.c) and used #include to pull it into
4413 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4414 separate copies.
4416 (b) Defined the list of op-code names for debugging as a macro in
4417 internal.h so that it is next to the definition of the opcodes.
4419 (c) Defined a table of op-code lengths for simpler skipping along compiled
4420 code. This is again a macro in internal.h so that it is next to the
4421 definition of the opcodes.
4423 34. Added support for recursive calls to individual subpatterns, along the
4424 lines of Robin Houston's patch (but implemented somewhat differently).
4426 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4427 allow it to read and process whole directories in Win32. This code was
4428 contributed by Lionel Fourquaux; it has not been tested by me.
4430 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4431 used to name a group. Names consist of alphanumerics and underscores, and must
4432 be unique. Back references use the syntax (?P=name) and recursive calls use
4433 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4434 numbers. The function pcre_fullinfo() can be used after compilation to extract
4435 a name/number map. There are three relevant calls:
4437 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4438 PCRE_INFO_NAMECOUNT yields the number of entries
4439 PCRE_INFO_NAMETABLE yields a pointer to the map.
4441 The map is a vector of fixed-size entries. The size of each entry depends on
4442 the length of the longest name used. The first two bytes of each entry are the
4443 group number, most significant byte first. There follows the corresponding
4444 name, zero terminated. The names are in alphabetical order.
4446 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4447 case instead of 255. Making it the same both with and without UTF-8 support
4448 means that the same test output works with both.
4450 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4451 calling malloc() with a zero argument.
4453 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4454 optimization. I've improved things by keeping a bitmap of backreferences with
4455 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4456 fact referenced, the optimization can be applied. It is unlikely that a
4457 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4458 the match to follow \n) will appear inside brackets with a number greater than
4459 31, but if it does, any back reference > 31 suppresses the optimization.
4461 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4462 of disabling numbered capturing parentheses. Any opening parenthesis that is
4463 not followed by ? behaves as if it were followed by ?: but named parentheses
4464 can still be used for capturing (and they will acquire numbers in the usual
4465 way).
4467 41. Redesigned the return codes from the match() function into yes/no/error so
4468 that errors can be passed back from deep inside the nested calls. A malloc
4469 failure while inside a recursive subpattern call now causes the
4470 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4472 42. It is now possible to set a limit on the number of times the match()
4473 function is called in a call to pcre_exec(). This facility makes it possible to
4474 limit the amount of recursion and backtracking, though not in a directly
4475 obvious way, because the match() function is used in a number of different
4476 circumstances. The count starts from zero for each position in the subject
4477 string (for non-anchored patterns). The default limit is, for compatibility, a
4478 large number, namely 10 000 000. You can change this in two ways:
4480 (a) When configuring PCRE before making, you can use --with-match-limit=n
4481 to set a default value for the compiled library.
4483 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4484 a different value is set. See 45 below.
4486 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4488 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4489 of things that can be changed at compile time. The first argument specifies
4490 what is wanted and the second points to where the information is to be placed.
4491 The current list of available information is:
4495 The output is an integer that is set to one if UTF-8 support is available;
4496 otherwise it is set to zero.
4500 The output is an integer that it set to the value of the code that is used for
4501 newline. It is either LF (10) or CR (13).
4505 The output is an integer that contains the number of bytes used for internal
4506 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4510 The output is an integer that contains the threshold above which the POSIX
4511 interface uses malloc() for output vectors. See item 31 above.
4515 The output is an unsigned integer that contains the default limit of the number
4516 of match() calls in a pcre_exec() execution. See 42 above.
4518 44. pcretest has been upgraded by the addition of the -C option. This causes it
4519 to extract all the available output from the new pcre_config() function, and to
4520 output it. The program then exits immediately.
4522 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4523 order to support additional features. One way would have been to define
4524 pcre_exec2() (for example) with extra arguments, but this would not have been
4525 extensible, and would also have required all calls to the original function to
4526 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4527 is used for passing in "extra" data from pcre_study().
4529 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4530 contains the following fields:
4532 flags a bitmap indicating which of the following fields are set
4533 study_data opaque data from pcre_study()
4534 match_limit a way of specifying a limit on match() calls for a specific
4535 call to pcre_exec()
4536 callout_data data for callouts (see 49 below)
4538 The flag bits are also defined in pcre.h, and are
4544 The pcre_study() function now returns one of these new pcre_extra blocks, with
4545 the actual study data pointed to by the study_data field, and the
4546 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4547 before. That is, this change is entirely upwards-compatible and requires no
4548 change to existing code.
4550 If you want to pass in additional data to pcre_exec(), you can either place it
4551 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4552 block.
4554 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4555 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4556 times with different match limits, until it finds the minimum value needed for
4557 pcre_exec() to complete. The value is then output. This can be instructive; for
4558 most simple matches the number is quite small, but for pathological cases it
4559 gets very large very quickly.
4561 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4562 returns the size of the data block pointed to by the study_data field in a
4563 pcre_extra block, that is, the value that was passed as the argument to
4564 pcre_malloc() when PCRE was getting memory in which to place the information
4565 created by pcre_study(). The fourth argument should point to a size_t variable.
4566 pcretest has been extended so that this information is shown after a successful
4567 pcre_study() call when information about the compiled regex is being displayed.
4569 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4570 because what follows is always an absolute path. (Later: it turns out that this
4571 is more than cosmetic for MinGW, because it doesn't like empty path
4572 components.)
4574 49. Some changes have been made to the callout feature (see 28 above):
4576 (i) A callout function now has three choices for what it returns:
4578 0 => success, carry on matching
4579 > 0 => failure at this point, but backtrack if possible
4580 < 0 => serious error, return this value from pcre_exec()
4582 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4583 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4584 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4585 use by callout functions. It will never be used by PCRE itself.
4587 (ii) The pcre_extra structure (see 45 above) has a void * field called
4588 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4589 pcre_callout_block structure has a field of the same name. The contents of
4590 the field passed in the pcre_extra structure are passed to the callout
4591 function in the corresponding field in the callout block. This makes it
4592 easier to use the same callout-containing regex from multiple threads. For
4593 testing, the pcretest program has a new data escape
4595 \C*n pass the number n (may be negative) as callout_data
4597 If the callout function in pcretest receives a non-zero value as
4598 callout_data, it returns that value.
4600 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4601 there were some redundant $(CFLAGS) in commands that are now specified as
4602 $(LINK), which already includes $(CFLAGS).
4604 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4605 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4606 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4607 one-byte characters throughout. Note that case-insensitive matching applies
4608 only to characters whose values are less than 256. PCRE doesn't support the
4609 notion of cases for higher-valued characters.
4611 (i) A character class whose characters are all within 0-255 is handled as
4612 a bit map, and the map is inverted for negative classes. Previously, a
4613 character > 255 always failed to match such a class; however it should
4614 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4616 (ii) A negated character class with a single character < 255 is coded as
4617 "not this character" (OP_NOT). This wasn't working properly when the test
4618 character was multibyte, either singly or repeated.
4620 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4621 mode, for example: \x{100}{2,3}.
4623 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4624 singly or repeated) now correctly test multibyte characters. However,
4625 PCRE doesn't recognize any characters with values greater than 255 as
4626 digits, spaces, or word characters. Such characters always match \D, \S,
4627 and \W, and never match \d, \s, or \w.
4629 (v) Classes may now contain characters and character ranges with values
4630 greater than 255. For example: [ab\x{100}-\x{400}].
4632 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4633 PCRE in UTF-8 mode.
4635 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4636 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4637 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4638 value.)
4640 53. The single man page has become too large. I have therefore split it up into
4641 a number of separate man pages. These also give rise to individual HTML pages;
4642 these are now put in a separate directory, and there is an index.html page that
4643 lists them all. Some hyperlinking between the pages has been installed.
4645 54. Added convenience functions for handling named capturing parentheses.
4647 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4648 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4649 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4650 are faulted.
4652 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4653 calling configure. These values are used when compiling the dftables.c program
4654 which is run to generate the source of the default character tables. They
4655 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4656 you will need to set these values.
4658 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4661 Version 3.9 02-Jan-02
4662 ---------------------
4664 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4666 2. If --disable-static was given, the building process failed when trying to
4667 build pcretest and pcregrep. (For some reason it was using libtool to compile
4668 them, which is not right, as they aren't part of the library.)
4671 Version 3.8 18-Dec-01
4672 ---------------------
4674 1. The experimental UTF-8 code was completely screwed up. It was packing the
4675 bytes in the wrong order. How dumb can you get?
4678 Version 3.7 29-Oct-01
4679 ---------------------
4681 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4682 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4683 this didn't happen under Solaris 8, where I normally test things.
4685 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4686 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4687 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4688 this fix an hour or so after the initial 3.7 release.)
4691 Version 3.6 23-Oct-01
4692 ---------------------
4694 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4695 offsets passed as NULL with zero offset count.
4697 2. The config.guess and config.sub files had not been updated when I moved to
4698 the latest autoconf.
4701 Version 3.5 15-Aug-01
4702 ---------------------
4704 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4705 had been forgotten.
4707 2. By using declared but undefined structures, we can avoid using "void"
4708 definitions in pcre.h while keeping the internal definitions of the structures
4709 private.
4711 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4712 user point of view, this means that both static and shared libraries are built
4713 by default, but this can be individually controlled. More of the work of
4714 handling this static/shared cases is now inside libtool instead of PCRE's make
4715 file.
4717 4. The pcretest utility is now installed along with pcregrep because it is
4718 useful for users (to test regexs) and by doing this, it automatically gets
4719 relinked by libtool. The documentation has been turned into a man page, so
4720 there are now .1, .txt, and .html versions in /doc.
4722 5. Upgrades to pcregrep:
4723 (i) Added long-form option names like gnu grep.
4724 (ii) Added --help to list all options with an explanatory phrase.
4725 (iii) Added -r, --recursive to recurse into sub-directories.
4726 (iv) Added -f, --file to read patterns from a file.
4728 6. pcre_exec() was referring to its "code" argument before testing that
4729 argument for NULL (and giving an error if it was NULL).
4731 7. Upgraded Makefile.in to allow for compiling in a different directory from
4732 the source directory.
4734 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4735 options bits, the pointer it was passed was to an int instead of to an unsigned
4736 long int. This mattered only on 64-bit systems.
4738 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4739 generated) instead of pcre.in, which it its source. Also made the same change
4740 in several of the .c files.
4742 10. A new release of gcc defines printf() as a macro, which broke pcretest
4743 because it had an ifdef in the middle of a string argument for printf(). Fixed
4744 by using separate calls to printf().
4746 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4747 script, to force use of CR or LF instead of \n in the source. On non-Unix
4748 systems, the value can be set in config.h.
4750 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4751 absolute limit. Changed the text of the error message to make this clear, and
4752 likewise updated the man page.
4754 13. The limit of 99 on the number of capturing subpatterns has been removed.
4755 The new limit is 65535, which I hope will not be a "real" limit.
4758 Version 3.4 22-Aug-00
4759 ---------------------
4761 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4763 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4766 Version 3.3 01-Aug-00
4767 ---------------------
4769 1. If an octal character was given, but the value was greater than \377, it
4770 was not getting masked to the least significant bits, as documented. This could
4771 lead to crashes in some systems.
4773 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4774 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4776 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4777 These just pass their arguments on to (pcre_free)(), but they are provided
4778 because some uses of PCRE bind it to non-C systems that can call its functions,
4779 but cannot call free() or pcre_free() directly.
4781 4. Add "make test" as a synonym for "make check". Corrected some comments in
4782 the Makefile.
4784 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4785 Makefile.
4787 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4788 command called pgrep for grepping around the active processes.
4790 7. Added the beginnings of support for UTF-8 character strings.
4792 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4793 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4794 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4795 out for the ar command.)
4798 Version 3.2 12-May-00
4799 ---------------------
4801 This is purely a bug fixing release.
4803 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4804 of ZA. This was just one example of several cases that could provoke this bug,
4805 which was introduced by change 9 of version 2.00. The code for breaking
4806 infinite loops after an iteration that matches an empty string was't working
4807 correctly.
4809 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4810 when matched against abbab (for example). After matching an empty string, it
4811 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4812 caused it to match further down the string than it should.
4814 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4815 was there because it doesn't seem to be needed, and it causes trouble on some
4816 systems, as it is not a Standard C header. It has been removed.
4818 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4819 were reported on the Macintosh. The changes were from
4821 while ((c = *(++ptr)) != 0 && c != '\n');
4822 to
4823 while ((c = *(++ptr)) != 0 && c != '\n') ;
4825 Totally extraordinary, but if that's what it takes...
4827 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4828 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4829 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4830 assumes the way PCRE uses memmove() (always moving upwards).
4832 6. PCRE is being used in one environment where strchr() is not available. There
4833 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4834 faster code anyway.
4837 Version 3.1 09-Feb-00
4838 ---------------------
4840 The only change in this release is the fixing of some bugs in Makefile.in for
4841 the "install" target:
4843 (1) It was failing to install pcreposix.h.
4845 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4848 Version 3.0 01-Feb-00
4849 ---------------------
4851 1. Add support for the /+ modifier to perltest (to output $` like it does in
4852 pcretest).
4854 2. Add support for the /g modifier to perltest.
4856 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4857 matches null strings.
4859 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4860 pattern. Perl treats empty patterns specially - it reuses the most recent
4861 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4862 effect.
4864 5. The POSIX interface was broken in that it was just handing over the POSIX
4865 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4866 required a bigger vector, with some working space on the end. This means that
4867 the POSIX wrapper now has to get and free some memory, and copy the results.
4869 6. Added some simple autoconf support, placing the test data and the
4870 documentation in separate directories, re-organizing some of the
4871 information files, and making it build pcre-config (a GNU standard). Also added
4872 libtool support for building PCRE as a shared library, which is now the
4873 default.
4875 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4876 09 are not valid octal constants. Single digits will be used for minor values
4877 less than 10.
4879 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4880 existing programs that set these in the POSIX interface can use PCRE without
4881 modification.
4883 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4884 return all that pcre_info() returns, plus additional data. The pcre_info()
4885 function is retained for compatibility, but is considered to be obsolete.
4887 10. Added experimental recursion feature (?R) to handle one common case that
4888 Perl 5.6 will be able to do with (?p{...}).
4890 11. Added support for POSIX character classes like [:alpha:], which Perl is
4891 adopting.
4894 Version 2.08 31-Aug-99
4895 ----------------------
4897 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4898 trying to match at the startoffset position, but instead was moving forward to
4899 the next newline as if a previous match had failed.
4901 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4902 and could get into a loop if a null string was matched other than at the start
4903 of the subject.
4905 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4906 be distinguished at compile time, and for completeness also added PCRE_DATE.
4908 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4909 in GnuWin32 environments.
4912 Version 2.07 29-Jul-99
4913 ----------------------
4915 1. The documentation is now supplied in plain text form and HTML as well as in
4916 the form of man page sources.
4918 2. C++ compilers don't like assigning (void *) values to other pointer types.
4919 In particular this affects malloc(). Although there is no problem in Standard
4920 C, I've put in casts to keep C++ compilers happy.
4922 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4923 should be (const char *).
4925 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4926 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4927 However, I haven't made this a standard facility. The documentation doesn't
4928 mention it, and the Makefile doesn't support it.
4930 5. The Makefile now contains an "install" target, with editable destinations at
4931 the top of the file. The pcretest program is not installed.
4933 6. pgrep -V now gives the PCRE version number and date.
4935 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4936 causing the entire string to be ignored, instead of just the last character.
4938 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4939 non-matching string, it can take a very, very long time, even for strings of
4940 quite modest length, because of the nested recursion. PCRE now does better in
4941 some of these cases. It does this by remembering the last required literal
4942 character in the pattern, and pre-searching the subject to ensure it is present
4943 before running the real match. In other words, it applies a heuristic to detect
4944 some types of certain failure quickly, and in the above example, if presented
4945 with a string that has no trailing " it gives "no match" very quickly.
4947 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4948 other alternatives are tried instead.
4951 Version 2.06 09-Jun-99
4952 ----------------------
4954 1. Change pcretest's output for amount of store used to show just the code
4955 space, because the remainder (the data block) varies in size between 32-bit and
4956 64-bit systems.
4958 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4959 start matching at. This allows lookbehinds to work when searching for multiple
4960 occurrences in a string.
4962 3. Added additional options to pcretest for testing multiple occurrences:
4964 /+ outputs the rest of the string that follows a match
4965 /g loops for multiple occurrences, using the new startoffset argument
4966 /G loops for multiple occurrences by passing an incremented pointer
4968 4. PCRE wasn't doing the "first character" optimization for patterns starting
4969 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4970 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4971 the letter 'x'. On long subject strings, this gives a significant speed-up.
4974 Version 2.05 21-Apr-99
4975 ----------------------
4977 1. Changed the type of magic_number from int to long int so that it works
4978 properly on 16-bit systems.
4980 2. Fixed a bug which caused patterns starting with .* not to work correctly
4981 when the subject string contained newline characters. PCRE was assuming
4982 anchoring for such patterns in all cases, which is not correct because .* will
4983 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4984 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4985 must be retried after every newline in the subject.
4988 Version 2.04 18-Feb-99
4989 ----------------------
4991 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4992 computation of the store needed to hold the pattern was incorrect (too large).
4993 If such patterns were nested a few deep, this could multiply and become a real
4994 problem.
4996 2. Added /M option to pcretest to show the memory requirement of a specific
4997 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4999 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5000 compiled in such a way that the backtracking after subsequent failure was
5001 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5002 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5005 Version 2.03 02-Feb-99
5006 ----------------------
5008 1. Fixed typo and small mistake in man page.
5010 2. Added 4th condition (GPL supersedes if conflict) and created separate
5011 LICENCE file containing the conditions.
5013 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5014 Perl, that is the internal \ allows the delimiter to be included in the
5015 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5016 the final delimiter, add \ to the end of the pattern (to test the error).
5018 4. Added the convenience functions for extracting substrings after a successful
5019 match. Updated pcretest to make it able to test these functions.
5022 Version 2.02 14-Jan-99
5023 ----------------------
5025 1. Initialized the working variables associated with each extraction so that
5026 their saving and restoring doesn't refer to uninitialized store.
5028 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5029 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5030 fix the problem.
5032 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5033 calls, and wasn't printing the correct value for compiling calls. Increased the
5034 default value of LOOPREPEAT, and the number of significant figures in the
5035 times.
5037 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5039 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5040 a building problem on Windows NT with a FAT file system.
5043 Version 2.01 21-Oct-98
5044 ----------------------
5046 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5047 to character tables built by pcre_maketables() in the current locale. If NULL
5048 is passed, the default tables are used.
5051 Version 2.00 24-Sep-98
5052 ----------------------
5054 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5055 it any more.
5057 2. Allow quantification of (?>) groups, and make it work correctly.
5059 3. The first character computation wasn't working for (?>) groups.
5061 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5062 end of the subject) and add 5.005's \z, which really does match only at the
5063 very end of the subject.
5065 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5067 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5068 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5069 localized options. All options to pcre_study() were also removed.
5071 7. Add other new features from 5.005:
5073 $(?<= positive lookbehind
5074 $(?<! negative lookbehind
5075 (?imsx-imsx) added the unsetting capability
5076 such a setting is global if at outer level; local otherwise
5077 (?imsx-imsx:) non-capturing groups with option setting
5078 (?(cond)re|re) conditional pattern matching
5080 A backreference to itself in a repeated group matches the previous
5081 captured string.
5083 8. General tidying up of studying (both automatic and via "study")
5084 consequential on the addition of new assertions.
5086 9. As in 5.005, unlimited repeated groups that could match an empty substring
5087 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5088 runtime if any iteration does actually match an empty substring.
5090 10. Include the RunTest script in the distribution.
5092 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5093 discrepancies, some of which were old and were also with respect to 5.004. They
5094 have now been fixed.
5097 Version 1.09 28-Apr-98
5098 ----------------------
5100 1. A negated single character class followed by a quantifier with a minimum
5101 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5102 program crashes, or just wrong answers. This did not apply to negated classes
5103 containing more than one character, or to minima other than one.
5106 Version 1.08 27-Mar-98
5107 ----------------------
5109 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5111 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5112 latter must appear before anything that relies on it in the pattern.
5115 Version 1.07 16-Feb-98
5116 ----------------------
5118 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5119 repeat of a potentially empty string).
5122 Version 1.06 23-Jan-98
5123 ----------------------
5125 1. Added Markus Oberhumer's little patches for C++.
5127 2. Literal strings longer than 255 characters were broken.
5130 Version 1.05 23-Dec-97
5131 ----------------------
5133 1. Negated character classes containing more than one character were failing if
5134 PCRE_CASELESS was set at run time.
5137 Version 1.04 19-Dec-97
5138 ----------------------
5140 1. Corrected the man page, where some "const" qualifiers had been omitted.
5142 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5143 input syntax.
5145 3. Fixed memory leak which occurred when a regex with back references was
5146 matched with an offsets vector that wasn't big enough. The temporary memory
5147 that is used in this case wasn't being freed if the match failed.
5149 4. Tidied pcretest to ensure it frees memory that it gets.
5151 5. Temporary memory was being obtained in the case where the passed offsets
5152 vector was exactly big enough.
5154 6. Corrected definition of offsetof() from change 5 below.
5156 7. I had screwed up change 6 below and broken the rules for the use of
5157 setjmp(). Now fixed.
5160 Version 1.03 18-Dec-97
5161 ----------------------
5163 1. A erroneous regex with a missing opening parenthesis was correctly
5164 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5165 on some systems.
5167 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5168 it was reported that one broken compiler failed on the former because "code" is
5169 also an independent variable.
5171 3. The erroneous regex a[]b caused an array overrun reference.
5173 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5174 fail on data ending with that character. (It was going on too far, and checking
5175 the next character, typically a binary zero.) This was specific to the
5176 optimized code for single-character negative classes.
5178 5. Added a contributed patch from the TIN world which does the following:
5180 + Add an undef for memmove, in case the the system defines a macro for it.
5182 + Add a definition of offsetof(), in case there isn't one. (I don't know
5183 the reason behind this - offsetof() is part of the ANSI standard - but
5184 it does no harm).
5186 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5187 most of the places where whitespace preceded '#'. I have given up and
5188 allowed the remaining 2 cases to be at the margin.
5190 + Rename some variables in pcre to eliminate shadowing. This seems very
5191 pedantic, but does no harm, of course.
5193 6. Moved the call to setjmp() into its own function, to get rid of warnings
5194 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5196 7. Constructs such as \d{8,} were compiling into the equivalent of
5197 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5198 outcome, but in this particular case used more store than had been allocated,
5199 which caused the bug to be discovered because it threw up an internal error.
5201 8. The debugging code in both pcre and pcretest for outputting the compiled
5202 form of a regex was going wrong in the case of back references followed by
5203 curly-bracketed repeats.
5206 Version 1.02 12-Dec-97
5207 ----------------------
5209 1. Typos in pcre.3 and comments in the source fixed.
5211 2. Applied a contributed patch to get rid of places where it used to remove
5212 'const' from variables, and fixed some signed/unsigned and uninitialized
5213 variable warnings.
5215 3. Added the "runtest" target to Makefile.
5217 4. Set default compiler flag to -O2 rather than just -O.
5220 Version 1.01 19-Nov-97
5221 ----------------------
5223 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5224 like /([ab]*)*/, that is, for classes with more than one character in them.
5226 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5227 as /((?>a*))*/ (a PCRE_EXTRA facility).
5230 Version 1.00 18-Nov-97
5231 ----------------------
5233 1. Added compile-time macros to support systems such as SunOS4 which don't have
5234 memmove() or strerror() but have other things that can be used instead.
5236 2. Arranged that "make clean" removes the executables.
5239 Version 0.99 27-Oct-97
5240 ----------------------
5242 1. Fixed bug in code for optimizing classes with only one character. It was
5243 initializing a 32-byte map regardless, which could cause it to run off the end
5244 of the memory it had got.
5246 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5249 Version 0.98 22-Oct-97
5250 ----------------------
5252 1. Fixed bug in code for handling temporary memory usage when there are more
5253 back references than supplied space in the ovector. This could cause segfaults.
5256 Version 0.97 21-Oct-97
5257 ----------------------
5259 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5261 2. Optimized negated single characters not to use a bit map.
5263 3. Brought error texts together as macro definitions; clarified some of them;
5264 fixed one that was wrong - it said "range out of order" when it meant "invalid
5265 escape sequence".
5267 4. Changed some char * arguments to const char *.
5269 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5271 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5272 pcretest.
5275 Version 0.96 16-Oct-97
5276 ----------------------
5278 1. Added a simple "pgrep" utility to the distribution.
5280 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5281 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5282 where "ddd" means "one or more decimal digits".
5284 3. Fixed serious bug. If a pattern had a back reference, but the call to
5285 pcre_exec() didn't supply a large enough ovector to record the related
5286 identifying subpattern, the match always failed. PCRE now remembers the number
5287 of the largest back reference, and gets some temporary memory in which to save
5288 the offsets during matching if necessary, in order to ensure that
5289 backreferences always work.
5291 4. Increased the compatibility with Perl in a number of ways:
5293 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5294 to request this handling. The option can be set at compile or exec time.
5296 (b) $ matches before a terminating newline by default; an option
5297 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5298 mode). The option can be set at compile or exec time.
5300 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5301 the same as Perl's. If the decimal number it represents is less than 10
5302 or there aren't that many previous left capturing parentheses, an octal
5303 escape is read. Inside a character class, it's always an octal escape,
5304 even if it is a single digit.
5306 (d) An escaped but undefined alphabetic character is taken as a literal,
5307 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5308 escapes.
5310 (e) {0} is now permitted. (The previous item is removed from the compiled
5311 pattern).
5313 5. Changed all the names of code files so that the basic parts are no longer
5314 than 10 characters, and abolished the teeny "globals.c" file.
5316 6. Changed the handling of character classes; they are now done with a 32-byte
5317 bit map always.
5319 7. Added the -d and /D options to pcretest to make it possible to look at the
5320 internals of compilation without having to recompile pcre.
5323 Version 0.95 23-Sep-97
5324 ----------------------
5326 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5327 \x20 at the start of a run of normal characters. These were being treated as
5328 real characters, instead of the source characters being re-checked.
5331 Version 0.94 18-Sep-97
5332 ----------------------
5334 1. The functions are now thread-safe, with the caveat that the global variables
5335 containing pointers to malloc() and free() or alternative functions are the
5336 same for all threads.
5338 2. Get pcre_study() to generate a bitmap of initial characters for non-
5339 anchored patterns when this is possible, and use it if passed to pcre_exec().
5342 Version 0.93 15-Sep-97
5343 ----------------------
5345 1. /(b)|(:+)/ was computing an incorrect first character.
5347 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5348 but not actually doing anything yet.
5350 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5351 as Perl does (e.g. [-az] or [az-]).
5353 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5354 all possible positions.
5356 5. Split up into different modules to avoid including unneeded functions in a
5357 compiled binary. However, compile and exec are still in one module. The "study"
5358 function is split off.
5360 6. The character tables are now in a separate module whose source is generated
5361 by an auxiliary program - but can then be edited by hand if required. There are
5362 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5363 toupper() in the code.
5365 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5366 make them global. Abolish the function for setting them, as the caller can now
5367 set them directly.
5370 Version 0.92 11-Sep-97
5371 ----------------------
5373 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5374 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5376 2. Caseless matching was not working in character classes if the characters in
5377 the pattern were in upper case.
5379 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5381 4. Make PCRE_ANCHORED public and accept as a compile option.
5383 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5384 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5385 pass them.
5387 6. Give an error if bad option bits passed at compile or run time.
5389 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5390 pcretest to cause it to pass that flag.
5392 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5393 options, and the first character, if set.
5395 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5398 Version 0.91 10-Sep-97
5399 ----------------------
5401 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5402 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5404 2. PCRE was looping on encountering an indefinitely repeated back reference to
5405 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5406 Perl does - treats the match as successful.
5408 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5