ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1420 - (show annotations)
Sun Dec 29 11:43:10 2013 UTC (6 years ago) by zherczeg
File size: 259425 byte(s)
Improve character range checks in JIT.
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.
35 Version 8.34 15-December-2013
36 -----------------------------
38 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
39 executable memory. Patch inspired by Carsten Klein.
41 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
42 this macro is never tested and has no effect, because the work to support
43 coverage involves only compiling and linking options and special targets in
44 the Makefile. The comment in config.h implied that defining the macro would
45 enable coverage support, which is totally false. There was also support for
46 setting this macro in the CMake files (my fault, I just copied it from
47 configure). SUPPORT_GCOV has now been removed.
49 3. Make a small performance improvement in strlen16() and strlen32() in
50 pcretest.
52 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
53 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
54 these cases). There was also one in pcretest.c.
56 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
58 6. In UTF mode, the code for checking whether a group could match an empty
59 string (which is used for indefinitely repeated groups to allow for
60 breaking an infinite loop) was broken when the group contained a repeated
61 negated single-character class with a character that occupied more than one
62 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
63 UTF-8 mode). The effect was undefined: the group might or might not be
64 deemed as matching an empty string, or the program might have crashed.
66 7. The code for checking whether a group could match an empty string was not
67 recognizing that \h, \H, \v, \V, and \R must match a character.
69 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
70 an empty string. If it can, pcretest shows this in its information output.
72 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
73 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
74 matched by pcre_exec() without using JIT:
76 (a) If the rest of the pattern did not match after a maximal run of
77 grapheme clusters, the code for backing up to try with fewer of them
78 did not always back up over a full grapheme when characters that do not
79 have the modifier quality were involved, e.g. Hangul syllables.
81 (b) If the match point in a subject started with modifier character, and
82 there was no match, the code could incorrectly back up beyond the match
83 point, and potentially beyond the first character in the subject,
84 leading to a segfault or an incorrect match result.
86 10. A conditional group with an assertion condition could lead to PCRE
87 recording an incorrect first data item for a match if no other first data
88 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
89 first data item, and therefore matched "ca" after "c" instead of at the
90 start.
92 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
93 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
95 12. The source of pcregrep now includes z/OS-specific code so that it can be
96 compiled for z/OS as part of the special z/OS distribution.
98 13. Added the -T and -TM options to pcretest.
100 14. The code in pcre_compile.c for creating the table of named capturing groups
101 has been refactored. Instead of creating the table dynamically during the
102 actual compiling pass, the information is remembered during the pre-compile
103 pass (on the stack unless there are more than 20 named groups, in which
104 case malloc() is used) and the whole table is created before the actual
105 compile happens. This has simplified the code (it is now nearly 150 lines
106 shorter) and prepared the way for better handling of references to groups
107 with duplicate names.
109 15. A back reference to a named subpattern when there is more than one of the
110 same name now checks them in the order in which they appear in the pattern.
111 The first one that is set is used for the reference. Previously only the
112 first one was inspected. This change makes PCRE more compatible with Perl.
114 16. Unicode character properties were updated from Unicode 6.3.0.
116 17. The compile-time code for auto-possessification has been refactored, based
117 on a patch by Zoltan Herczeg. It now happens after instead of during
118 compilation. The code is cleaner, and more cases are handled. The option
119 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
120 options in pcretest are provided to set it. It can also be set by
121 (*NO_AUTO_POSSESS) at the start of a pattern.
123 18. The character VT has been added to the default ("C" locale) set of
124 characters that match \s and are generally treated as white space,
125 following this same change in Perl 5.18. There is now no difference between
126 "Perl space" and "POSIX space". Whether VT is treated as white space in
127 other locales depends on the locale.
129 19. The code for checking named groups as conditions, either for being set or
130 for being recursed, has been refactored (this is related to 14 and 15
131 above). Processing unduplicated named groups should now be as fast at
132 numerical groups, and processing duplicated groups should be faster than
133 before.
135 20. Two patches to the CMake build system, by Alexander Barkov:
137 (1) Replace the "source" command by "." in CMakeLists.txt because
138 "source" is a bash-ism.
140 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
141 without these the CMake build does not work on Solaris.
143 21. Perl has changed its handling of \8 and \9. If there is no previously
144 encountered capturing group of those numbers, they are treated as the
145 literal characters 8 and 9 instead of a binary zero followed by the
146 literals. PCRE now does the same.
148 22. Following Perl, added \o{} to specify codepoints in octal, making it
149 possible to specify values greater than 0777 and also making them
150 unambiguous.
152 23. Perl now gives an error for missing closing braces after \x{... instead of
153 treating the string as literal. PCRE now does the same.
155 24. RunTest used to grumble if an inappropriate test was selected explicitly,
156 but just skip it when running all tests. This make it awkward to run ranges
157 of tests when one of them was inappropriate. Now it just skips any
158 inappropriate tests, as it always did when running all tests.
160 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
161 character types such as \d or \w, too many callouts were inserted, and the
162 data that they returned was rubbish.
164 26. In UCP mode, \s was not matching two of the characters that Perl matches,
165 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
166 were matched by \h. The code has now been refactored so that the lists of
167 the horizontal and vertical whitespace characters used for \h and \v (which
168 are defined only in one place) are now also used for \s.
170 27. Add JIT support for the 64 bit TileGX architecture.
171 Patch by Jiong Wang (Tilera Corporation).
173 28. Possessive quantifiers for classes (both explicit and automatically
174 generated) now use special opcodes instead of wrapping in ONCE brackets.
176 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
177 (because it's meaningless), this was not happening when PCRE_CASELESS was
178 set. Not wrong, but inefficient.
180 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
181 properties for \w, \d, etc) is present in a test regex. Otherwise if the
182 test contains no characters greater than 255, Perl doesn't realise it
183 should be using Unicode semantics.
185 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
186 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
187 does in Unicode mode.
189 32. Added the "forbid" facility to pcretest so that putting tests into the
190 wrong test files can sometimes be quickly detected.
192 33. There is now a limit (default 250) on the depth of nesting of parentheses.
193 This limit is imposed to control the amount of system stack used at compile
194 time. It can be changed at build time by --with-parens-nest-limit=xxx or
195 the equivalent in CMake.
197 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
198 errors. Perl warns for these when in warning mode, but PCRE has no facility
199 for giving warnings.
201 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
202 However, this was not working for (?!) because it is optimized to (*FAIL),
203 for which PCRE does not allow quantifiers. The optimization is now disabled
204 when a quantifier follows (?!). I can't see any use for this, but it makes
205 things uniform.
207 36. Perl no longer allows group names to start with digits, so I have made this
208 change also in PCRE. It simplifies the code a bit.
210 37. In extended mode, Perl ignores spaces before a + that indicates a
211 possessive quantifier. PCRE allowed a space before the quantifier, but not
212 before the possessive +. It now does.
214 38. The use of \K (reset reported match start) within a repeated possessive
215 group such as (a\Kb)*+ was not working.
217 40. Document that the same character tables must be used at compile time and
218 run time, and that the facility to pass tables to pcre_exec() and
219 pcre_dfa_exec() is for use only with saved/restored patterns.
221 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
222 features for Builds with MSVC:
224 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
225 builds). The .rc files can be used to set FileDescription and many other
226 attributes.
228 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
229 This allows higher-level build scripts which want .pdb files to avoid
230 hard-coding the exact files needed."
232 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
233 mean "start of word" and "end of word", respectively, as a transition aid.
235 43. A minimizing repeat of a class containing codepoints greater than 255 in
236 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
237 compiled to use the heap for recursion.
239 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
240 is configured.
243 Version 8.33 28-May-2013
244 ------------------------
246 1. Added 'U' to some constants that are compared to unsigned integers, to
247 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
248 variables that are added to signed variables, to ensure the result is
249 signed and can be negated.
251 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
252 CMake config files.
254 3. Revise the creation of config.h.generic so that all boolean macros are
255 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
256 overriding via -D on the command line possible.
258 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
259 to unsigned int is reported to make a quite noticeable speed difference in
260 a specific Windows environment. Testing on Linux did also appear to show
261 some benefit (and it is clearly not harmful). Also fixed the definition of
262 Xop which should be unsigned.
264 5. Related to (4), changing the definition of the intermediate variable cc
265 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
266 improvements.
268 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
269 unnecessary spaces.
271 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
272 and later.
274 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
276 9. Optimizing fast_forward_start_bits in JIT.
278 10. Adding support for callouts in JIT, and fixing some issues revealed
279 during this work. Namely:
281 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
283 (b) Minimum length was not checked before the matching is started.
285 11. The value of capture_last that is passed to callouts was incorrect in some
286 cases when there was a capture on one path that was subsequently abandoned
287 after a backtrack. Also, the capture_last value is now reset after a
288 recursion, since all captures are also reset in this case.
290 12. The interpreter no longer returns the "too many substrings" error in the
291 case when an overflowing capture is in a branch that is subsequently
292 abandoned after a backtrack.
294 13. In the pathological case when an offset vector of size 2 is used, pcretest
295 now prints out the matched string after a yield of 0 or 1.
297 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
298 Only supported by the JIT compiler at the moment.
300 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
302 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
303 offset of the starting point of the matching process, provided the offsets
304 vector is large enough.
306 17. The \A escape now records a lookbehind value of 1, though its execution
307 does not actually inspect the previous character. This is to ensure that,
308 in partial multi-segment matching, at least one character from the old
309 segment is retained when a new segment is processed. Otherwise, if there
310 are no lookbehinds in the pattern, \A might match incorrectly at the start
311 of a new segment.
313 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
315 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
316 gives some modest performance improvement in 8-bit mode.
318 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
319 be expressed in certain programming languages using Universal Character
320 Names.
322 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
323 which points out that "non characters" are not "characters that may not
324 appear in Unicode strings" but rather "characters that are reserved for
325 internal use and have only local meaning".
327 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
328 there was a conditional group that depended on an assertion, if the
329 assertion was false, the callout that immediately followed the alternation
330 in the condition was skipped when pcre_exec() was used for matching.
332 23. Allow an explicit callout to be inserted before an assertion that is the
333 condition for a conditional group, for compatibility with automatic
334 callouts, which always insert a callout at this point.
336 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
337 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
339 25. (*PRUNE) is now supported by the JIT compiler.
341 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
343 27. Fix the case where there are two or more SKIPs with arguments that may be
344 ignored.
346 28. (*SKIP) is now supported by the JIT compiler.
348 29. (*THEN) is now supported by the JIT compiler.
350 30. Update RunTest with additional test selector options.
352 31. The way PCRE handles backtracking verbs has been changed in two ways.
354 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
355 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
356 backtracking. In some cases this makes it more Perl-compatible, but Perl's
357 rather obscure rules do not always do the same thing.
359 (2) Previously, backtracking verbs were confined within assertions. This is
360 no longer the case for positive assertions, except for (*ACCEPT). Again,
361 this sometimes improves Perl compatibility, and sometimes does not.
363 32. A number of tests that were in test 2 because Perl did things differently
364 have been moved to test 1, because either Perl or PCRE has changed, and
365 these tests are now compatible.
367 32. Backtracking control verbs are now handled in the same way in JIT and
368 interpreter.
370 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
371 contained a forward subroutine reference caused a compile error.
373 34. Auto-detect and optimize limited repetitions in JIT.
375 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
376 blocking (*UTF) etc.
378 36. In the interpreter, maximizing pattern repetitions for characters and
379 character types now use tail recursion, which reduces stack usage.
381 37. The value of the max lookbehind was not correctly preserved if a compiled
382 and saved regex was reloaded on a host of different endianness.
384 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
385 of the compiled pattern block, expand the flags field from 16 to 32 bits
386 because it was almost full.
388 39. Try madvise first before posix_madvise.
390 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
391 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
392 original change locked out all matches of empty strings. This has been
393 changed so that one match of an empty string per line is recognized.
394 Subsequent searches on the same line (for colouring or for --only-matching,
395 for example) do not recognize empty strings.
397 41. Applied a user patch to fix a number of spelling mistakes in comments.
399 42. Data lines longer than 65536 caused pcretest to crash.
401 43. Clarified the data type for length and startoffset arguments for pcre_exec
402 and pcre_dfa_exec in the function-specific man pages, where they were
403 explicitly stated to be in bytes, never having been updated. I also added
404 some clarification to the pcreapi man page.
406 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
407 a segmentation fault.
410 Version 8.32 30-November-2012
411 -----------------------------
413 1. Improved JIT compiler optimizations for first character search and single
414 character iterators.
416 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
417 Patch by Daniel Richard G.
419 3. Single character iterator optimizations in the JIT compiler.
421 4. Improved JIT compiler optimizations for character ranges.
423 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
424 Reported by Giuseppe D'Angelo.
426 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
427 of a line, was being set incorrectly in cases where .* appeared inside
428 atomic brackets at the start of a pattern, or where there was a subsequent
429 *PRUNE or *SKIP.
431 7. Improved instruction cache flush for POWER/PowerPC.
432 Patch by Daniel Richard G.
434 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
435 grep:
437 (a) There is now no limit to the number of patterns to be matched.
439 (b) An error is given if a pattern is too long.
441 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
442 are now supported.
444 (d) --exclude-from and --include-from (multiple use) have been added.
446 (e) Exclusions and inclusions now apply to all files and directories, not
447 just to those obtained from scanning a directory recursively.
449 (f) Multiple uses of -f and --file-list are now supported.
451 (g) In a Windows environment, the default for -d has been changed from
452 "read" (the GNU grep default) to "skip", because otherwise the presence
453 of a directory in the file list provokes an error.
455 (h) The documentation has been revised and clarified in places.
457 9. Improve the matching speed of capturing brackets.
459 10. Changed the meaning of \X so that it now matches a Unicode extended
460 grapheme cluster.
462 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
463 out POSIX threads when JIT support is configured.
465 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
467 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
468 structure could go wrong in environments where size_t is not the same size
469 as int.
471 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
472 set.
474 15. The EBCDIC support had decayed; later updates to the code had included
475 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
476 general tidy up of EBCDIC-related issues, and the documentation was also
477 not quite right. There is now a test that can be run on ASCII systems to
478 check some of the EBCDIC-related things (but is it not a full test).
480 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
481 in a small tidy to the code.
483 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
485 18. If the --only-matching (-o) option in pcregrep is specified multiple
486 times, each one causes appropriate output. For example, -o1 -o2 outputs the
487 substrings matched by the 1st and 2nd capturing parentheses. A separating
488 string can be specified by --om-separator (default empty).
490 19. Improving the first n character searches.
492 20. Turn case lists for horizontal and vertical white space into macros so that
493 they are defined only once.
495 21. This set of changes together give more compatible Unicode case-folding
496 behaviour for characters that have more than one other case when UCP
497 support is available.
499 (a) The Unicode property table now has offsets into a new table of sets of
500 three or more characters that are case-equivalent. The MultiStage2.py
501 script that generates these tables (the pcre_ucd.c file) now scans
502 CaseFolding.txt instead of UnicodeData.txt for character case
503 information.
505 (b) The code for adding characters or ranges of characters to a character
506 class has been abstracted into a generalized function that also handles
507 case-independence. In UTF-mode with UCP support, this uses the new data
508 to handle characters with more than one other case.
510 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
511 whose other case is greater than 256 are now correctly matched
512 caselessly. Previously, the high codepoint matched the low one, but not
513 vice versa.
515 (d) The processing of \h, \H, \v, and \ in character classes now makes use
516 of the new class addition function, using character lists defined as
517 macros alongside the case definitions of 20 above.
519 (e) Caseless back references now work with characters that have more than
520 one other case.
522 (f) General caseless matching of characters with more than one other case
523 is supported.
525 22. Unicode character properties were updated from Unicode 6.2.0
527 23. Improved CMake support under Windows. Patch by Daniel Richard G.
529 24. Add support for 32-bit character strings, and UTF-32
531 25. Major JIT compiler update (code refactoring and bugfixing).
532 Experimental Sparc 32 support is added.
534 26. Applied a modified version of Daniel Richard G's patch to create
535 pcre.h.generic and config.h.generic by "make" instead of in the
536 PrepareRelease script.
538 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
539 pcre_compile.c when checking for a zero character.
541 28. Introducing a native interface for JIT. Through this interface, the compiled
542 machine code can be directly executed. The purpose of this interface is to
543 provide fast pattern matching, so several sanity checks are not performed.
544 However, feature tests are still performed. The new interface provides
545 1.4x speedup compared to the old one.
547 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
548 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
549 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
551 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
552 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
553 the "old" RFC 2279). Instead, it ended up passing a negative length to
554 pcre_exec().
556 31. Add support for GCC's visibility feature to hide internal functions.
558 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
559 "unknown -C option" after outputting 0 or 1.
561 33. There is now support for generating a code coverage report for the test
562 suite in environments where gcc is the compiler and lcov is installed. This
563 is mainly for the benefit of the developers.
565 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
566 unaddressable using valgrind annotations, allowing valgrind to detect
567 invalid memory accesses. This is mainly for the benefit of the developers.
569 25. (*UTF) can now be used to start a pattern in any of the three libraries.
571 26. Give configure error if --enable-cpp but no C++ compiler found.
574 Version 8.31 06-July-2012
575 -------------------------
577 1. Fixing a wrong JIT test case and some compiler warnings.
579 2. Removed a bashism from the RunTest script.
581 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
582 to unsigned type, result still unsigned" that was given by an MS compiler
583 on encountering the code "-sizeof(xxx)".
585 4. Partial matching support is added to the JIT compiler.
587 5. Fixed several bugs concerned with partial matching of items that consist
588 of more than one character:
590 (a) /^(..)\1/ did not partially match "aba" because checking references was
591 done on an "all or nothing" basis. This also applied to repeated
592 references.
594 (b) \R did not give a hard partial match if \r was found at the end of the
595 subject.
597 (c) \X did not give a hard partial match after matching one or more
598 characters at the end of the subject.
600 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
601 a partial match for the string "\r".
603 (e) When newline was set to CRLF, the metacharacter "." did not recognize
604 a partial match for a CR character at the end of the subject string.
606 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
607 running pcretest, the text "(JIT)" added to the output whenever JIT is
608 actually used to run the match.
610 7. Individual JIT compile options can be set in pcretest by following -s+[+]
611 or /S+[+] with a digit between 1 and 7.
613 8. OP_NOT now supports any UTF character not just single-byte ones.
615 9. (*MARK) control verb is now supported by the JIT compiler.
617 10. The command "./RunTest list" lists the available tests without actually
618 running any of them. (Because I keep forgetting what they all are.)
622 12. Applied a (slightly modified) user-supplied patch that improves performance
623 when the heap is used for recursion (compiled with --disable-stack-for-
624 recursion). Instead of malloc and free for each heap frame each time a
625 logical recursion happens, frames are retained on a chain and re-used where
626 possible. This sometimes gives as much as 30% improvement.
628 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
629 call.
631 14. As documented, (*COMMIT) is now confined to within a positive assertion.
633 15. It is now possible to link pcretest with libedit as an alternative to
634 libreadline.
636 16. (*COMMIT) control verb is now supported by the JIT compiler.
638 17. The Unicode data tables have been updated to Unicode 6.1.0.
640 18. Added --file-list option to pcregrep.
642 19. Added binary file support to pcregrep, including the -a, --binary-files,
643 -I, and --text options.
645 20. The madvise function is renamed for posix_madvise for QNX compatibility
646 reasons. Fixed by Giuseppe D'Angelo.
648 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
650 22. Changed the option for creating symbolic links for 16-bit man pages from
651 -s to -sf so that re-installing does not cause issues.
653 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
655 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
656 matches in certain environments (the workspace was not being correctly
657 retained). Also added to pcre_dfa_exec() a simple plausibility check on
658 some of the workspace data at the beginning of a restart.
660 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
661 was not doing so when it should - probably a typo introduced by SVN 528
662 (change 8.10/14).
664 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
665 \w+ when the character tables indicated that \x{c4} was a word character.
666 There were several related cases, all because the tests for doing a table
667 lookup were testing for characters less than 127 instead of 255.
669 27. If a pattern contains capturing parentheses that are not used in a match,
670 their slots in the ovector are set to -1. For those that are higher than
671 any matched groups, this happens at the end of processing. In the case when
672 there were back references that the ovector was too small to contain
673 (causing temporary malloc'd memory to be used during matching), and the
674 highest capturing number was not used, memory off the end of the ovector
675 was incorrectly being set to -1. (It was using the size of the temporary
676 memory instead of the true size.)
678 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
679 ovector size, it uses memory at the end of the block that it has got.
681 29. Check for an overlong MARK name and give an error at compile time. The
682 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
684 30. JIT compiler update.
686 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
687 Rill for the patch.
689 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
691 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
693 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
694 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
695 was enabled.
697 35. Improve JIT code generation for greedy plus quantifier.
699 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
700 1 to "aa" instead of to an empty string. The bug affected repeated groups
701 that could potentially match an empty string.
703 37. Optimizing single character iterators in JIT.
705 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
706 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
707 codepoints that are too big for the mode are faulted, and in a UTF mode,
708 disallowed codepoints are also faulted.
710 39. If PCRE was compiled with UTF support, in three places in the DFA
711 matcher there was code that should only have been obeyed in UTF mode, but
712 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
713 processing when bytes with values greater than 127 were present. In 16-bit
714 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
715 both cases the values are those that cannot be the first data item in a UTF
716 character. The three items that might have provoked this were recursions,
717 possessively repeated groups, and atomic groups.
719 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
720 and pcregrep, because some OS require shared objects to be explicitly
721 passed to ld, causing the link step to fail if they are not.
723 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
724 mode, patterns that started with \h* or \R* might be incorrectly matched.
727 Version 8.30 04-February-2012
728 -----------------------------
730 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
731 name is defined in ctype.h.
733 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
734 only in quite long subpatterns.
736 3. Removed the function pcre_info(), which has been obsolete and deprecated
737 since it was replaced by pcre_fullinfo() in February 2000.
739 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
740 match a (*MARK), and the match failed at the start of the subject, a
741 reference to memory before the start of the subject could occur. This bug
742 was introduced by fix 17 of release 8.21.
744 5. A reference to an unset group with zero minimum repetition was giving
745 totally wrong answers (in non-JavaScript-compatibility mode). For example,
746 /(another)?(\1?)test/ matched against "hello world test". This bug was
747 introduced in release 8.13.
749 6. Add support for 16-bit character strings (a large amount of work involving
750 many changes and refactorings).
752 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
753 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
754 from a file.
756 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
757 rounding is not applied in this particular case).
759 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
760 if they appear, or are escaped, in patterns.
762 10. Get rid of a number of -Wunused-but-set-variable warnings.
764 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
765 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
766 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
767 also returns the mark "x". This bug applied to capturing parentheses,
768 non-capturing parentheses, and atomic parentheses. It also applied to some
769 assertions.
771 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
772 information out of configure.ac instead of relying on pcre.h.generic, which
773 is not stored in the repository.
775 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
776 -lreadline.
778 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
780 15. Applied Graycode's patch to put the top-level frame on the stack rather
781 than the heap when not using the stack for recursion. This gives a
782 performance improvement in many cases when recursion is not deep.
784 16. Experimental code added to "pcretest -C" to output the stack frame size.
787 Version 8.21 12-Dec-2011
788 ------------------------
790 1. Updating the JIT compiler.
792 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
793 are added as well.
795 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
796 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
797 calling _pcre_jit_exec. Some extra comments are added.
799 4. (*MARK) settings inside atomic groups that do not contain any capturing
800 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
801 was introduced by change 18 for 8.20.
803 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
804 ECMA-262 standard.
806 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
807 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
808 This bug was probably introduced by change 9 of 8.13.
810 7. While fixing 6 above, I noticed that a number of other items were being
811 incorrectly rejected as "not fixed length". This arose partly because newer
812 opcodes had not been added to the fixed-length checking code. I have (a)
813 corrected the bug and added tests for these items, and (b) arranged for an
814 error to occur if an unknown opcode is encountered while checking for fixed
815 length instead of just assuming "not fixed length". The items that were
816 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
817 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
818 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
820 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
821 being incorrectly compiled and would have given unpredicatble results.
823 9. A possessively repeated subpattern with minimum repeat count greater than
824 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
825 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
826 the first (A) could occur when it should not.
828 10. Add a cast and remove a redundant test from the code.
830 11. JIT should use pcre_malloc/pcre_free for allocation.
832 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
833 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
834 is anything other than /usr, -L is still shown).
836 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
838 14. Perl does not support \N without a following name in a [] class; PCRE now
839 also gives an error.
841 15. If a forward reference was repeated with an upper limit of around 2000,
842 it caused the error "internal error: overran compiling workspace". The
843 maximum number of forward references (including repeats) was limited by the
844 internal workspace, and dependent on the LINK_SIZE. The code has been
845 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
846 the default depends on LINK_SIZE. There is a new upper limit (for safety)
847 of around 200,000 forward references. While doing this, I also speeded up
848 the filling in of repeated forward references.
850 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
851 incorrectly expecting the subject to contain another "a" after the start.
853 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
854 in the match, the SKIP should be ignored. This was not happening; instead
855 the SKIP was being treated as NOMATCH. For patterns such as
856 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
857 tested.
859 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
860 now much more compatible with Perl, in particular in cases where the result
861 is a non-match for a non-anchored pattern. For example, if
862 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
863 "m", where previously it did not return a name. A side effect of this
864 change is that for partial matches, the last encountered mark name is
865 returned, as for non matches. A number of tests that were previously not
866 Perl-compatible have been moved into the Perl-compatible test files. The
867 refactoring has had the pleasing side effect of removing one argument from
868 the match() function, thus reducing its stack requirements.
870 19. If the /S+ option was used in pcretest to study a pattern using JIT,
871 subsequent uses of /S (without +) incorrectly behaved like /S+.
873 21. Retrieve executable code size support for the JIT compiler and fixing
874 some warnings.
876 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
877 not work when the shorter character appeared right at the end of the
878 subject string.
880 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
881 systems.
883 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
884 output it when the /M option is used in pcretest.
886 25. The CheckMan script was not being included in the distribution. Also, added
887 an explicit "perl" to run Perl scripts from the PrepareRelease script
888 because this is reportedly needed in Windows.
890 26. If study data was being save in a file and studying had not found a set of
891 "starts with" bytes for the pattern, the data written to the file (though
892 never used) was taken from uninitialized memory and so caused valgrind to
893 complain.
895 27. Updated RunTest.bat as provided by Sheri Pierce.
897 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
899 29. Computation of memory usage for the table of capturing group names was
900 giving an unnecessarily large value.
903 Version 8.20 21-Oct-2011
904 ------------------------
906 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
907 a POSIX class. After further experiments with Perl, which convinced me that
908 Perl has bugs and confusions, a closing square bracket is no longer allowed
909 in a POSIX name. This bug also affected patterns with classes that started
910 with full stops.
912 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
913 captured substring, but while checking the failing first alternative,
914 substring 1 is temporarily captured. If the output vector supplied to
915 pcre_exec() was not big enough for this capture, the yield of the function
916 was still zero ("insufficient space for captured substrings"). This cannot
917 be totally fixed without adding another stack variable, which seems a lot
918 of expense for a edge case. However, I have improved the situation in cases
919 such as /(a)(b)x|abc/ matched against "abc", where the return code
920 indicates that fewer than the maximum number of slots in the ovector have
921 been set.
923 3. Related to (2) above: when there are more back references in a pattern than
924 slots in the output vector, pcre_exec() uses temporary memory during
925 matching, and copies in the captures as far as possible afterwards. It was
926 using the entire output vector, but this conflicts with the specification
927 that only 2/3 is used for passing back captured substrings. Now it uses
928 only the first 2/3, for compatibility. This is, of course, another edge
929 case.
931 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
932 main code base, and can be used by building with --enable-jit. When this is
933 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
934 runtime --no-jit option is given.
936 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
937 ovector, the return from the function was zero, implying that there were
938 other matches that did not fit. The correct "exactly full" value is now
939 returned.
941 6. If a subpattern that was called recursively or as a subroutine contained
942 (*PRUNE) or any other control that caused it to give a non-standard return,
943 invalid errors such as "Error -26 (nested recursion at the same subject
944 position)" or even infinite loops could occur.
946 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
947 computing the minimum length on reaching *ACCEPT, and so ended up with the
948 wrong value of 1 rather than 0. Further investigation indicates that
949 computing a minimum subject length in the presence of *ACCEPT is difficult
950 (think back references, subroutine calls), and so I have changed the code
951 so that no minimum is registered for a pattern that contains *ACCEPT.
953 8. If (*THEN) was present in the first (true) branch of a conditional group,
954 it was not handled as intended. [But see 16 below.]
956 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
957 Sheri Pierce.
959 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
960 the first byte in a match must be "a".
962 11. Change 17 for 8.13 increased the recursion depth for patterns like
963 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
964 pattern contains any instances of (*THEN). If it does not, the old
965 optimizations are restored. It would be nice to do this on a per-group
966 basis, but at the moment that is not feasible.
968 12. In some environments, the output of pcretest -C is CRLF terminated. This
969 broke RunTest's code that checks for the link size. A single white space
970 character after the value is now allowed for.
972 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
973 For "fr", it uses the Windows-specific input and output files.
975 14. If (*THEN) appeared in a group that was called recursively or as a
976 subroutine, it did not work as intended. [But see next item.]
978 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
979 pattern fragments (but not containing any | characters). If A and B are
980 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
981 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
982 D. In other words, Perl considers parentheses that do not contain any |
983 characters to be part of a surrounding alternative, whereas PCRE was
984 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
985 differently. PCRE now behaves in the same way as Perl, except in the case
986 of subroutine/recursion calls such as (?1) which have in any case always
987 been different (but PCRE had them first :-).
989 16. Related to 15 above: Perl does not treat the | in a conditional group as
990 creating alternatives. Such a group is treated in the same way as an
991 ordinary group without any | characters when processing (*THEN). PCRE has
992 been changed to match Perl's behaviour.
994 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
995 RunGrepTest script failed.
997 18. Change 22 for version 13 caused atomic groups to use more stack. This is
998 inevitable for groups that contain captures, but it can lead to a lot of
999 stack use in large patterns. The old behaviour has been restored for atomic
1000 groups that do not contain any capturing parentheses.
1002 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1003 suppress the check for a minimum subject length at run time. (If it was
1004 given to pcre_exec() or pcre_dfa_exec() it did work.)
1006 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1007 fail to work when decoding hex characters in data strings in EBCDIC
1008 environments.
1010 21. It appears that in at least one Mac OS environment, the isxdigit() function
1011 is implemented as a macro that evaluates to its argument more than once,
1012 contravening the C 90 Standard (I haven't checked a later standard). There
1013 was an instance in pcretest which caused it to go wrong when processing
1014 \x{...} escapes in subject strings. The has been rewritten to avoid using
1015 things like p++ in the argument of isxdigit().
1018 Version 8.13 16-Aug-2011
1019 ------------------------
1021 1. The Unicode data tables have been updated to Unicode 6.0.0.
1023 2. Two minor typos in pcre_internal.h have been fixed.
1025 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1026 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1027 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1029 4. There were a number of related bugs in the code for matching backrefences
1030 caselessly in UTF-8 mode when codes for the characters concerned were
1031 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1032 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1033 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1034 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1035 2-byte code at the end of the subject (it thought there wasn't enough data
1036 left).
1038 5. Comprehensive information about what went wrong is now returned by
1039 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1040 as the output vector has at least 2 elements. The offset of the start of
1041 the failing character and a reason code are placed in the vector.
1043 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1044 now returned is for the first byte of the failing character, instead of the
1045 last byte inspected. This is an incompatible change, but I hope it is small
1046 enough not to be a problem. It makes the returned offset consistent with
1047 pcre_exec() and pcre_dfa_exec().
1049 7. pcretest now gives a text phrase as well as the error number when
1050 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1051 failure, the offset and reason code are output.
1053 8. When \R was used with a maximizing quantifier it failed to skip backwards
1054 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1055 back over a single character (\n). This seems wrong (because it treated the
1056 two characters as a single entity when going forwards), conflicts with the
1057 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1058 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1059 has been changed.
1061 9. Some internal refactoring has changed the processing so that the handling
1062 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1063 time (the PCRE_DOTALL option was changed this way some time ago: version
1064 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1065 which was always a bit of a fudge. It also means that there is one less
1066 argument for the match() function, which reduces its stack requirements
1067 slightly. This change also fixes an incompatibility with Perl: the pattern
1068 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1070 10. More internal refactoring has drastically reduced the number of recursive
1071 calls to match() for possessively repeated groups such as (abc)++ when
1072 using pcre_exec().
1074 11. While implementing 10, a number of bugs in the handling of groups were
1075 discovered and fixed:
1077 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1078 (a|)*(?1) gave a compile-time internal error.
1079 ((a|)+)+ did not notice that the outer group could match an empty string.
1080 (^a|^)+ was not marked as anchored.
1081 (.*a|.*)+ was not marked as matching at start or after a newline.
1083 12. Yet more internal refactoring has removed another argument from the match()
1084 function. Special calls to this function are now indicated by setting a
1085 value in a variable in the "match data" data block.
1087 13. Be more explicit in pcre_study() instead of relying on "default" for
1088 opcodes that mean there is no starting character; this means that when new
1089 ones are added and accidentally left out of pcre_study(), testing should
1090 pick them up.
1092 14. The -s option of pcretest has been documented for ages as being an old
1093 synonym of -m (show memory usage). I have changed it to mean "force study
1094 for every regex", that is, assume /S for every regex. This is similar to -i
1095 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1096 using it. It makes it easier to run collections of tests with and without
1097 study enabled, and thereby test pcre_study() more easily. All the standard
1098 tests are now run with and without -s (but some patterns can be marked as
1099 "never study" - see 20 below).
1101 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1102 restoration of the capturing data to the outer values was not happening
1103 correctly.
1105 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1106 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1107 pattern had matched an empty string, and so incorrectly returned a no
1108 match.
1110 17. There was optimizing code for the last branch of non-capturing parentheses,
1111 and also for the obeyed branch of a conditional subexpression, which used
1112 tail recursion to cut down on stack usage. Unfortunately, now that there is
1113 the possibility of (*THEN) occurring in these branches, tail recursion is
1114 no longer possible because the return has to be checked for (*THEN). These
1115 two optimizations have therefore been removed. [But see 8.20/11 above.]
1117 18. If a pattern containing \R was studied, it was assumed that \R always
1118 matched two bytes, thus causing the minimum subject length to be
1119 incorrectly computed because \R can also match just one byte.
1121 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1122 was incorrectly computed.
1124 20. If /S is present twice on a test pattern in pcretest input, it now
1125 *disables* studying, thereby overriding the use of -s on the command line
1126 (see 14 above). This is necessary for one or two tests to keep the output
1127 identical in both cases.
1129 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1130 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1132 22. When an atomic group that contained a capturing parenthesis was
1133 successfully matched, but the branch in which it appeared failed, the
1134 capturing was not being forgotten if a higher numbered group was later
1135 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1136 group 1 to "a", when in fact it should be unset. This applied to multi-
1137 branched capturing and non-capturing groups, repeated or not, and also to
1138 positive assertions (capturing in negative assertions does not happen
1139 in PCRE) and also to nested atomic groups.
1141 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1142 subject after a captured substring, to make it easier to tell which of a
1143 number of identical substrings has been captured.
1145 24. The way atomic groups are processed by pcre_exec() has been changed so that
1146 if they are repeated, backtracking one repetition now resets captured
1147 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1148 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1149 "aaa". Previously, it would have been "a". As part of this code
1150 refactoring, the way recursive calls are handled has also been changed.
1152 25. If an assertion condition captured any substrings, they were not passed
1153 back unless some other capturing happened later. For example, if
1154 (?(?=(a))a) was matched against "a", no capturing was returned.
1156 26. When studying a pattern that contained subroutine calls or assertions,
1157 the code for finding the minimum length of a possible match was handling
1158 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1159 group 1 called group 2 while simultaneously a separate group 2 called group
1160 1). A stack overflow occurred in this case. I have fixed this by limiting
1161 the recursion depth to 10.
1163 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1164 Fortmann. This supports explicit test numbers on the command line, and has
1165 argument validation and error reporting.
1167 28. An instance of \X with an unlimited repeat could fail if at any point the
1168 first character it looked at was a mark character.
1170 29. Some minor code refactoring concerning Unicode properties and scripts
1171 should reduce the stack requirement of match() slightly.
1173 30. Added the '=' option to pcretest to check the setting of unused capturing
1174 slots at the end of the pattern, which are documented as being -1, but are
1175 not included in the return count.
1177 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1178 compiled something random. Now it gives a compile-time error (as does
1179 Perl).
1181 32. A *MARK encountered during the processing of a positive assertion is now
1182 recorded and passed back (compatible with Perl).
1184 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1185 had alternative anchored branches, the search for a second match in a line
1186 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1187 matched the line "0102" twice. The same bug affected patterns that started
1188 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1189 twice.
1191 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1192 does, and because of capturing effects, quantifying parenthesized
1193 assertions may at times be useful. Quantifiers are now allowed for
1194 parenthesized assertions.
1196 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1198 36. \g was being checked for fancy things in a character class, when it should
1199 just be a literal "g".
1201 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1202 appearance of a nested POSIX class supersedes an apparent external class.
1203 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1204 unescaped square brackets may also appear as part of class names. For
1205 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1206 more like Perl. (But see 8.20/1 above.)
1208 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1209 was because it thought it was \N{name}, which is not supported).
1211 39. Add minix to OS list not supporting the -S option in pcretest.
1213 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1214 cannot analyze patterns in sufficient detail to catch mutual recursions
1215 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1216 subgroup is called recursively as a subpattern for a second time at the
1217 same position in the subject string. In previous releases this might have
1218 been caught by the recursion limit, or it might have run out of stack.
1220 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1221 happen only once. PCRE was, however incorrectly giving a compile time error
1222 "recursive call could loop indefinitely" because it cannot analyze the
1223 pattern in sufficient detail. The compile time test no longer happens when
1224 PCRE is compiling a conditional subpattern, but actual runaway loops are
1225 now caught at runtime (see 40 above).
1227 42. It seems that Perl allows any characters other than a closing parenthesis
1228 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1229 has been changed to be the same.
1231 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1232 as not to get warnings when autogen.sh is called. Also changed
1233 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1235 44. To help people who use pcregrep to scan files containing exceedingly long
1236 lines, the following changes have been made:
1238 (a) The default value of the buffer size parameter has been increased from
1239 8K to 20K. (The actual buffer used is three times this size.)
1241 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1242 PCRE is built.
1244 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1245 to be set at run time.
1247 (d) Numerical values in pcregrep options can be followed by K or M, for
1248 example --buffer-size=50K.
1250 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1251 given and the return code is set to 2.
1253 45. Add a pointer to the latest mark to the callout data block.
1255 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1256 partial match of an empty string instead of no match. This was specific to
1257 the use of ".".
1259 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1260 complete match instead of a partial match. This bug was dependent on both
1261 the PCRE_UTF8 and PCRE_DOTALL options being set.
1263 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1264 starting byte set, because \b was not being ignored.
1267 Version 8.12 15-Jan-2011
1268 ------------------------
1270 1. Fixed some typos in the markup of the man pages, and wrote a script that
1271 checks for such things as part of the documentation building process.
1273 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1274 --match-limit and --recursion-limit options (added for 8.11). In
1275 particular, this made one of the standard tests fail. (The integer value
1276 went into the wrong half of a long int.)
1278 3. If the --colour option was given to pcregrep with -v (invert match), it
1279 did strange things, either producing crazy output, or crashing. It should,
1280 of course, ignore a request for colour when reporting lines that do not
1281 match.
1283 4. Another pcregrep bug caused similar problems if --colour was specified with
1284 -M (multiline) and the pattern match finished with a line ending.
1286 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1287 matched in multiline mode, the following line was shown as part of the
1288 match. This seems wrong, so I have changed it.
1290 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1291 the check for further matches in the same line (so they could be coloured)
1292 to overrun the end of the current line. If another match was found, it was
1293 incorrectly shown (and then shown again when found in the next line).
1295 7. If pcregrep was compiled under Windows, there was a reference to the
1296 function pcregrep_exit() before it was defined. I am assuming this was
1297 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1298 reported by a user. I've moved the definition above the reference.
1301 Version 8.11 10-Dec-2010
1302 ------------------------
1304 1. (*THEN) was not working properly if there were untried alternatives prior
1305 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1306 backtracked to try for "b" instead of moving to the next alternative branch
1307 at the same level (in this case, to look for "c"). The Perl documentation
1308 is clear that when (*THEN) is backtracked onto, it goes to the "next
1309 alternative in the innermost enclosing group".
1311 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1312 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1313 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1314 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1315 (*THEN).
1317 3. If \s appeared in a character class, it removed the VT character from
1318 the class, even if it had been included by some previous item, for example
1319 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1320 of \s, but is part of the POSIX "space" class.)
1322 4. A partial match never returns an empty string (because you can always
1323 match an empty string at the end of the subject); however the checking for
1324 an empty string was starting at the "start of match" point. This has been
1325 changed to the "earliest inspected character" point, because the returned
1326 data for a partial match starts at this character. This means that, for
1327 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1328 (previously it gave "no match").
1330 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1331 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1332 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1333 has an implication that the given string is incomplete (because a partial
1334 match is preferred over a full match). For this reason, these items now
1335 give a partial match in this situation. [Aside: previously, the one case
1336 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1337 match rather than a full match, which was wrong by the old rules, but is
1338 now correct.]
1340 6. There was a bug in the handling of #-introduced comments, recognized when
1341 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1342 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1343 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1344 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1345 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1346 places in pcre_compile().
1348 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1349 comments when looking ahead for named forward references to subpatterns,
1350 the only newline sequence it recognized was NL. It now handles newlines
1351 according to the set newline convention.
1353 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1354 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1355 cater for a lack of strerror(). These oversights have been fixed.
1357 9. Added --match-limit and --recursion-limit to pcregrep.
1359 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1361 11. When the -o option was used, pcregrep was setting a return code of 1, even
1362 when matches were found, and --line-buffered was not being honoured.
1364 12. Added an optional parentheses number to the -o and --only-matching options
1365 of pcregrep.
1367 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1368 can match an empty string. The code to do it in pcretest and pcredemo
1369 needed fixing:
1371 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1372 only one byte after an empty string match just before CRLF (this case
1373 just got forgotten; "any" and "anycrlf" were OK).
1375 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1376 mode when an empty string match preceded an ASCII character followed by
1377 a non-ASCII character. (The code for advancing by one character rather
1378 than one byte was nonsense.)
1380 (c) The pcredemo.c sample program did not have any code at all to handle
1381 the cases when CRLF is a valid newline sequence.
1383 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1384 as a starting offset was within the subject string. There is now a new
1385 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1386 negative or greater than the length of the string. In order to test this,
1387 pcretest is extended to allow the setting of negative starting offsets.
1389 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1390 starting offset points to the beginning of a UTF-8 character was
1391 unnecessarily clumsy. I tidied it up.
1393 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1394 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1396 17. Nobody had reported that the --include_dir option, which was added in
1397 release 7.7 should have been called --include-dir (hyphen, not underscore)
1398 for compatibility with GNU grep. I have changed it to --include-dir, but
1399 left --include_dir as an undocumented synonym, and the same for
1400 --exclude-dir, though that is not available in GNU grep, at least as of
1401 release 2.5.4.
1403 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1404 characters from a string of bytes) have been redefined so as not to use
1405 loops, in order to improve performance in some environments. At the same
1406 time, I abstracted some of the common code into auxiliary macros to save
1407 repetition (this should not affect the compiled code).
1409 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1410 compile-time error is now given if \c is not followed by an ASCII
1411 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1412 different, and any byte value is allowed.)
1414 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1415 START_OPTIMIZE option, which is now allowed at compile time - but just
1416 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1417 to pcregrep and other applications that have no direct access to PCRE
1418 options. The new /Y option in pcretest sets this option when calling
1419 pcre_compile().
1421 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1422 back references. Groups containing recursive back references were forced to
1423 be atomic by that change, but in the case of named groups, the amount of
1424 memory required was incorrectly computed, leading to "Failed: internal
1425 error: code overflow". This has been fixed.
1427 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1428 pcretest.c, to avoid build problems in some Borland environments.
1431 Version 8.10 25-Jun-2010
1432 ------------------------
1434 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1435 THEN.
1437 2. (*ACCEPT) was not working when inside an atomic group.
1439 3. Inside a character class, \B is treated as a literal by default, but
1440 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1441 causes the error). The code is unchanged, but I tidied the documentation.
1443 4. Inside a character class, PCRE always treated \R and \X as literals,
1444 whereas Perl faults them if its -w option is set. I have changed PCRE so
1445 that it faults them when PCRE_EXTRA is set.
1447 5. Added support for \N, which always matches any character other than
1448 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1450 6. When compiling pcregrep with newer versions of gcc which may have
1451 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1452 declared with attribute warn_unused_result" were given. Just casting the
1453 result to (void) does not stop the warnings; a more elaborate fudge is
1454 needed. I've used a macro to implement this.
1456 7. Minor change to pcretest.c to avoid a compiler warning.
1458 8. Added four artifical Unicode properties to help with an option to make
1459 \s etc use properties (see next item). The new properties are: Xan
1460 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1462 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1463 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1464 this option. Modified pcretest to add /W to test this facility. Added
1465 REG_UCP to make it available via the POSIX interface.
1467 10. Added --line-buffered to pcregrep.
1469 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1470 studied, and the match started with a letter with a code point greater than
1471 127 whose first byte was different to the first byte of the other case of
1472 the letter, the other case of this starting letter was not recognized
1473 (#976).
1475 12. If a pattern that was studied started with a repeated Unicode property
1476 test, for example, \p{Nd}+, there was the theoretical possibility of
1477 setting up an incorrect bitmap of starting bytes, but fortunately it could
1478 not have actually happened in practice until change 8 above was made (it
1479 added property types that matched character-matching opcodes).
1481 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1482 possible starting bytes for non-anchored patterns.
1484 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1485 \R, and also a number of cases that involve Unicode properties, both
1486 explicit and implicit when PCRE_UCP is set.
1488 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1489 input, it could crash or give wrong results if characters with values
1490 greater than 0xc0 were present in the subject string. (Detail: it assumed
1491 UTF-8 input when processing these items.)
1493 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1494 size_t is 64-bit (#991).
1496 17. Added a check for running out of memory when PCRE is compiled with
1497 --disable-stack-for-recursion (#990).
1499 18. If the last data line in a file for pcretest does not have a newline on
1500 the end, a newline was missing in the output.
1502 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1503 less than 128) in its various bitmaps. However, there is a facility for
1504 generating tables according to the current locale when PCRE is compiled. It
1505 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1506 characters, are recognized by isspace() and therefore were getting set in
1507 these tables, and indeed these tables seem to approximate to ISO 8859. This
1508 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1509 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1510 which of course cannot start UTF-8 characters. I have changed the code so
1511 that only real ASCII characters (less than 128) and the correct starting
1512 bytes for UTF-8 encodings are set for characters greater than 127 when in
1513 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1514 altogether.)
1516 20. Added the /T option to pcretest so as to be able to run tests with non-
1517 standard character tables, thus making it possible to include the tests
1518 used for 19 above in the standard set of tests.
1520 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1521 reference to a subpattern the other side of a comment that contains an
1522 opening parenthesis caused either an internal compiling error, or a
1523 reference to the wrong subpattern.
1526 Version 8.02 19-Mar-2010
1527 ------------------------
1529 1. The Unicode data tables have been updated to Unicode 5.2.0.
1531 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1532 configured.
1534 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1535 original author of that file, following a query about its status.
1537 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1538 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1540 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1541 quantifier applied to a forward-referencing subroutine call, could compile
1542 incorrect code or give the error "internal error: previously-checked
1543 referenced subpattern not found".
1545 6. Both MS Visual Studio and Symbian OS have problems with initializing
1546 variables to point to external functions. For these systems, therefore,
1547 pcre_malloc etc. are now initialized to local functions that call the
1548 relevant global functions.
1550 7. There were two entries missing in the vectors called coptable and poptable
1551 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1552 I've fixed the data, and added a kludgy way of testing at compile time that
1553 the lengths are correct (equal to the number of opcodes).
1555 8. Following on from 7, I added a similar kludge to check the length of the
1556 eint vector in pcreposix.c.
1558 9. Error texts for pcre_compile() are held as one long string to avoid too
1559 much relocation at load time. To find a text, the string is searched,
1560 counting zeros. There was no check for running off the end of the string,
1561 which could happen if a new error number was added without updating the
1562 string.
1564 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1566 11. \K was not working if it appeared in an atomic group or in a group that
1567 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1568 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1569 the assertion is positive, but not if it is negative.
1571 12. Change 11 fortuitously reduced the size of the stack frame used in the
1572 "match()" function of pcre_exec.c by one pointer. Forthcoming
1573 implementation of support for (*MARK) will need an extra pointer on the
1574 stack; I have reserved it now, so that the stack frame size does not
1575 decrease.
1577 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1578 item in branch that calls a recursion is a subroutine call - as in the
1579 second branch in the above example - was incorrectly given the compile-
1580 time error "recursive call could loop indefinitely" because pcre_compile()
1581 was not correctly checking the subroutine for matching a non-empty string.
1583 14. The checks for overrunning compiling workspace could trigger after an
1584 overrun had occurred. This is a "should never occur" error, but it can be
1585 triggered by pathological patterns such as hundreds of nested parentheses.
1586 The checks now trigger 100 bytes before the end of the workspace.
1588 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1591 Version 8.01 19-Jan-2010
1592 ------------------------
1594 1. If a pattern contained a conditional subpattern with only one branch (in
1595 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1596 computed the wrong minimum data length (which is of course zero for such
1597 subpatterns). This could cause incorrect "no match" results.
1599 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1600 the pattern is reset in the first branch, pcre_compile() failed with
1601 "internal error: code overflow at offset...". This happened only when
1602 the reset was to the original external option setting. (An optimization
1603 abstracts leading options settings into an external setting, which was the
1604 cause of this.)
1606 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1607 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1608 assertion pattern did not match (meaning that the assertion was true), it
1609 was incorrectly treated as false if the SKIP had been reached during the
1610 matching. This also applied to assertions used as conditions.
1612 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1613 assertion subpattern, including such a pattern used as a condition,
1614 unpredictable results occurred, instead of the error return
1617 5. The C++ GlobalReplace function was not working like Perl for the special
1618 situation when an empty string is matched. It now does the fancy magic
1619 stuff that is necessary.
1621 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1622 removed. (These were left over from very, very early versions of PCRE.)
1624 7. Some cosmetic changes to the code to make life easier when compiling it
1625 as part of something else:
1627 (a) Change DEBUG to PCRE_DEBUG.
1629 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1630 called "current" as "current_branch", to prevent a collision with the
1631 Linux macro when compiled as a kernel module.
1633 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1634 prevent a collision with the Linux macro when compiled as a kernel
1635 module.
1637 8. In pcre_compile() there are some checks for integer overflows that used to
1638 cast potentially large values to (double). This has been changed to that
1639 when building, a check for int64_t is made, and if it is found, it is used
1640 instead, thus avoiding the use of floating point arithmetic. (There is no
1641 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1642 double.
1644 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1645 2005 (difference between two addresses compared to an unsigned value).
1647 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1648 custom one, because of the following reported problem in Windows:
1650 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1651 under Win32.
1652 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1653 therefore missing the function definition.
1654 - The compiler thus generates a "C" signature for the test function.
1655 - The linker fails to find the "C" function.
1656 - PCRE fails to configure if asked to do so against libbz2.
1658 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1659 messages were output:
1661 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1662 rerunning libtoolize, to keep the correct libtool macros in-tree.
1663 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1665 I have done both of these things.
1667 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1668 most of the time, it *can* run out if it is given a pattern that contains a
1669 runaway infinite recursion. I updated the discussion in the pcrestack man
1670 page.
1672 13. Now that we have gone to the x.xx style of version numbers, the minor
1673 version may start with zero. Using 08 or 09 is a bad idea because users
1674 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1675 interpreted as invalid octal numbers. I've updated the previous comment in
1676 configure.ac, and also added a check that gives an error if 08 or 09 are
1677 used.
1679 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1680 causing partial matching to fail when the end of the subject matched \W
1681 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1683 15. There were some discrepancies between the declarations in pcre_internal.h
1684 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1685 their definitions. The declarations used "const uschar *" and the
1686 definitions used USPTR. Even though USPTR is normally defined as "const
1687 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1688 reported that: "This difference in casting confuses some C++ compilers, for
1689 example, SunCC recognizes above declarations as different functions and
1690 generates broken code for hbpcre." I have changed the declarations to use
1691 USPTR.
1693 16. GNU libtool is named differently on some systems. The autogen.sh script now
1694 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1695 (FreeBSD).
1697 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1698 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1699 comment: "Figure out how to create a longlong from a string: strtoll and
1700 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1701 instance, but it only takes 2 args instead of 3!"
1703 18. A subtle bug concerned with back references has been fixed by a change of
1704 specification, with a corresponding code fix. A pattern such as
1705 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1706 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1707 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1708 same bug. Such groups have to be quantified to be useful, or contained
1709 inside another quantified group. (If there's no repetition, the reference
1710 can never match.) The problem arises because, having left the group and
1711 moved on to the rest of the pattern, a later failure that backtracks into
1712 the group uses the captured value from the final iteration of the group
1713 rather than the correct earlier one. I have fixed this in PCRE by forcing
1714 any group that contains a reference to itself to be an atomic group; that
1715 is, there cannot be any backtracking into it once it has completed. This is
1716 similar to recursive and subroutine calls.
1719 Version 8.00 19-Oct-09
1720 ----------------------
1722 1. The table for translating pcre_compile() error codes into POSIX error codes
1723 was out-of-date, and there was no check on the pcre_compile() error code
1724 being within the table. This could lead to an OK return being given in
1725 error.
1727 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1728 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1729 in a Windows environment.
1731 3. The pcregrep --count option prints the count for each file even when it is
1732 zero, as does GNU grep. However, pcregrep was also printing all files when
1733 --files-with-matches was added. Now, when both options are given, it prints
1734 counts only for those files that have at least one match. (GNU grep just
1735 prints the file name in this circumstance, but including the count seems
1736 more useful - otherwise, why use --count?) Also ensured that the
1737 combination -clh just lists non-zero counts, with no names.
1739 4. The long form of the pcregrep -F option was incorrectly implemented as
1740 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1741 but it seems right to fix it, and I didn't think it was worth preserving
1742 the old behaviour.
1744 5. The command line items --regex=pattern and --regexp=pattern were not
1745 recognized by pcregrep, which required --regex pattern or --regexp pattern
1746 (with a space rather than an '='). The man page documented the '=' forms,
1747 which are compatible with GNU grep; these now work.
1749 6. No libpcreposix.pc file was created for pkg-config; there was just
1750 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1752 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1753 when UCP support is not needed, by modifying the Python script that
1754 generates it from Unicode data files. This should not matter if the module
1755 is correctly used as a library, but I received one complaint about 50K of
1756 unwanted data. My guess is that the person linked everything into his
1757 program rather than using a library. Anyway, it does no harm.
1759 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1760 was a minimum greater than 1 for a wide character in a possessive
1761 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1762 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1763 character. Chaos in the form of incorrect output or a compiling loop could
1764 result.
1766 9. The restrictions on what a pattern can contain when partial matching is
1767 requested for pcre_exec() have been removed. All patterns can now be
1768 partially matched by this function. In addition, if there are at least two
1769 slots in the offset vector, the offset of the earliest inspected character
1770 for the match and the offset of the end of the subject are set in them when
1771 PCRE_ERROR_PARTIAL is returned.
1773 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1774 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1775 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1776 and may be more useful for multi-segment matching.
1778 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1779 used to be given if ever the end of the subject was reached; now it is
1780 given only if matching could not proceed because another character was
1781 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1782 string "Z", which now yields "no match" instead of "partial match". In the
1783 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1784 final character ended with (*FAIL).
1786 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1787 if the pattern had a "must contain" character that was already found in the
1788 earlier partial match, unless partial matching was again requested. For
1789 example, with the pattern /dog.(body)?/, the "must contain" character is
1790 "g". If the first part-match was for the string "dog", restarting with
1791 "sbody" failed. This bug has been fixed.
1793 13. The string returned by pcre_dfa_exec() after a partial match has been
1794 changed so that it starts at the first inspected character rather than the
1795 first character of the match. This makes a difference only if the pattern
1796 starts with a lookbehind assertion or \b or \B (\K is not supported by
1797 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1798 matching functions compatible, and I think it's the right thing to do.
1800 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1801 so that the demonstration program is easily available in environments where
1802 PCRE has not been installed from source.
1804 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1805 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1806 library.
1808 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1809 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1810 is not the first non-POSIX option to be added. Clearly some people find
1811 these options useful.
1813 17. If a caller to the POSIX matching function regexec() passes a non-zero
1814 value for nmatch with a NULL value for pmatch, the value of
1815 nmatch is forced to zero.
1817 18. RunGrepTest did not have a test for the availability of the -u option of
1818 the diff command, as RunTest does. It now checks in the same way as
1819 RunTest, and also checks for the -b option.
1821 19. If an odd number of negated classes containing just a single character
1822 interposed, within parentheses, between a forward reference to a named
1823 subpattern and the definition of the subpattern, compilation crashed with
1824 an internal error, complaining that it could not find the referenced
1825 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1826 [The bug was that it was starting one character too far in when skipping
1827 over the character class, thus treating the ] as data rather than
1828 terminating the class. This meant it could skip too much.]
1830 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1831 /g option in pcretest when the pattern contains \K, which makes it possible
1832 to have an empty string match not at the start, even when the pattern is
1833 anchored. Updated pcretest and pcredemo to use this option.
1835 21. If the maximum number of capturing subpatterns in a recursion was greater
1836 than the maximum at the outer level, the higher number was returned, but
1837 with unset values at the outer level. The correct (outer level) value is
1838 now given.
1840 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1841 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1842 make it do so. The string so far is captured, making this feature
1843 compatible with Perl.
1845 23. The tests have been re-organized, adding tests 11 and 12, to make it
1846 possible to check the Perl 5.10 features against Perl 5.10.
1848 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1849 pattern matches a fixed length string. PCRE did not allow this; now it
1850 does. Neither allows recursion.
1852 25. I finally figured out how to implement a request to provide the minimum
1853 length of subject string that was needed in order to match a given pattern.
1854 (It was back references and recursion that I had previously got hung up
1855 on.) This code has now been added to pcre_study(); it finds a lower bound
1856 to the length of subject needed. It is not necessarily the greatest lower
1857 bound, but using it to avoid searching strings that are too short does give
1858 some useful speed-ups. The value is available to calling programs via
1859 pcre_fullinfo().
1861 26. While implementing 25, I discovered to my embarrassment that pcretest had
1862 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1863 study optimizations had never been tested with that matching function.
1864 Oops. What is worse, even when it was passed study data, there was a bug in
1865 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1866 were also very few tests of studied patterns with pcre_dfa_exec().
1868 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1869 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1870 on the other side of the coin, they are no longer allowed to have different
1871 names, because these cannot be distinguished in PCRE, and this has caused
1872 confusion. (This is a difference from Perl.)
1874 28. When duplicate subpattern names are present (necessarily with different
1875 numbers, as required by 27 above), and a test is made by name in a
1876 conditional pattern, either for a subpattern having been matched, or for
1877 recursion in such a pattern, all the associated numbered subpatterns are
1878 tested, and the overall condition is true if the condition is true for any
1879 one of them. This is the way Perl works, and is also more like the way
1880 testing by number works.
1883 Version 7.9 11-Apr-09
1884 ---------------------
1886 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1887 (pcretest), all targets were linked against these libraries. This included
1888 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1889 libraries. This caused unwanted dependencies to be created. This problem
1890 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1891 pcretest is linked with readline.
1893 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1894 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1895 moved outside it again, because FALSE and TRUE are already defined in AIX,
1896 but BOOL is not.
1898 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1899 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1901 4. The pcregrep documentation said spaces were inserted as well as colons (or
1902 hyphens) following file names and line numbers when outputting matching
1903 lines. This is not true; no spaces are inserted. I have also clarified the
1904 wording for the --colour (or --color) option.
1906 5. In pcregrep, when --colour was used with -o, the list of matching strings
1907 was not coloured; this is different to GNU grep, so I have changed it to be
1908 the same.
1910 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1911 each matching line was coloured. Now it goes on to look for further matches
1912 of any of the test patterns, which is the same behaviour as GNU grep.
1914 7. A pattern that could match an empty string could cause pcregrep to loop; it
1915 doesn't make sense to accept an empty string match in pcregrep, so I have
1916 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1917 seems to be how GNU grep behaves. [But see later change 40 for release
1918 8.33.]
1920 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1921 start or after a newline", because the conditional assertion was not being
1922 correctly handled. The rule now is that both the assertion and what follows
1923 in the first alternative must satisfy the test.
1925 9. If auto-callout was enabled in a pattern with a conditional group whose
1926 condition was an assertion, PCRE could crash during matching, both with
1927 pcre_exec() and pcre_dfa_exec().
1929 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1930 used for matching.
1932 11. Unicode property support in character classes was not working for
1933 characters (bytes) greater than 127 when not in UTF-8 mode.
1935 12. Added the -M command line option to pcretest.
1937 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1939 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1941 16. Added comments and documentation about mis-use of no_arg in the C++
1942 wrapper.
1944 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1945 from Martin Jerabek that uses macro names for all relevant character and
1946 string constants.
1948 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1949 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1950 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1951 these, but not everybody uses configure.
1953 19. A conditional group that had only one branch was not being correctly
1954 recognized as an item that could match an empty string. This meant that an
1955 enclosing group might also not be so recognized, causing infinite looping
1956 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1957 with the subject "ab", where knowledge that the repeated group can match
1958 nothing is needed in order to break the loop.
1960 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1961 exec(), but without supplying a callout function, matching went wrong.
1963 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1964 leak if the size of the offset vector was greater than 30. When the vector
1965 is smaller, the saved offsets during recursion go onto a local stack
1966 vector, but for larger vectors malloc() is used. It was failing to free
1967 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1968 error, in fact).
1970 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1971 heapframe that is used only when UTF-8 support is enabled. This caused no
1972 problem, but was untidy.
1974 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1975 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1976 included within another project.
1978 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1979 slightly modified by me:
1981 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1982 not building pcregrep.
1984 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1985 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1987 25. Forward references, both numeric and by name, in patterns that made use of
1988 duplicate group numbers, could behave incorrectly or give incorrect errors,
1989 because when scanning forward to find the reference group, PCRE was not
1990 taking into account the duplicate group numbers. A pattern such as
1991 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1993 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1994 the feature of a custom pointer more persuasive (as requested by a user).
1996 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1997 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1998 pre-defined.
2000 28. Added support for (*UTF8) at the start of a pattern.
2002 29. Arrange for flags added by the "release type" setting in CMake to be shown
2003 in the configuration summary.
2006 Version 7.8 05-Sep-08
2007 ---------------------
2009 1. Replaced UCP searching code with optimized version as implemented for Ad
2010 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2011 stage table and inline lookup instead of a function, giving speed ups of 2
2012 to 5 times on some simple patterns that I tested. Permission was given to
2013 distribute the MultiStage2.py script that generates the tables (it's not in
2014 the tarball, but is in the Subversion repository).
2016 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2017 scripts.
2019 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2020 a group with a zero qualifier. The result of the study could be incorrect,
2021 or the function might crash, depending on the pattern.
2023 4. Caseless matching was not working for non-ASCII characters in back
2024 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2025 It now works when Unicode Property Support is available.
2027 5. In pcretest, an escape such as \x{de} in the data was always generating
2028 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2029 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2030 truncation.
2032 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2034 7. Added two (int) casts to pcregrep when printing the difference of two
2035 pointers, in case they are 64-bit values.
2037 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2038 test 2 if it fails.
2040 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2041 and a #define of that name to empty if it is not externally set. This is to
2042 allow users of MSVC to set it if necessary.
2044 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2045 the convenience functions in the pcre_get.c source file.
2047 11. An option change at the start of a pattern that had top-level alternatives
2048 could cause overwriting and/or a crash. This command provoked a crash in
2049 some environments:
2051 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2053 This potential security problem was recorded as CVE-2008-2371.
2055 12. For a pattern where the match had to start at the beginning or immediately
2056 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2057 pcre_dfa_exec() could read past the end of the passed subject if there was
2058 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2059 pcretest so that it places the subject at the end of its malloc-ed buffer.
2061 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2062 exec() might read past the end of the data buffer in UTF-8 mode.
2064 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2065 the data contained the byte 0x85 as part of a UTF-8 character within its
2066 first line. This applied both to normal and DFA matching.
2068 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2069 /^[^d]*?$/8 failed to match "abc".
2071 16. Added a missing copyright notice to pcrecpp_internal.h.
2073 17. Make it more clear in the documentation that values returned from
2074 pcre_exec() in ovector are byte offsets, not character counts.
2076 18. Tidied a few places to stop certain compilers from issuing warnings.
2078 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2079 supplied by Stefan Weber. I made a further small update for 7.8 because
2080 there is a change of source arrangements: the pcre_searchfuncs.c module is
2081 replaced by pcre_ucd.c.
2084 Version 7.7 07-May-08
2085 ---------------------
2087 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2088 a string to a long long, pretend we don't even have a long long." This is
2089 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2091 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2092 pre-7.6 versions, which defined a global no_arg variable instead of putting
2093 it in the RE class. (See also #8 below.)
2095 3. Remove a line of dead code, identified by coverity and reported by Nuno
2096 Lopes.
2098 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2100 (1) The include/exclude patterns were being applied to the whole pathnames
2101 of files, instead of just to the final components.
2103 (2) If there was more than one level of directory, the subdirectories were
2104 skipped unless they satisfied the include/exclude conditions. This is
2105 inconsistent with GNU grep (and could even be seen as contrary to the
2106 pcregrep specification - which I improved to make it absolutely clear).
2107 The action now is always to scan all levels of directory, and just
2108 apply the include/exclude patterns to regular files.
2110 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2111 --exclude_dir in the tests to avoid scanning .svn directories.
2113 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2114 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2115 doesn't support NULs in patterns.
2117 7. Added some missing "const"s to declarations of static tables in
2118 pcre_compile.c and pcre_dfa_exec.c.
2120 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2121 caused by fix #2 above. (Subsequently also a second patch to fix the
2122 first patch. And a third patch - this was a messy problem.)
2124 9. Applied Craig's patch to remove the use of push_back().
2126 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2127 matching function regexec().
2129 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2130 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2131 references. PCRE supports relative numbers with this syntax (I don't think
2132 Oniguruma does).
2134 12. Previously, a group with a zero repeat such as (...){0} was completely
2135 omitted from the compiled regex. However, this means that if the group
2136 was called as a subroutine from elsewhere in the pattern, things went wrong
2137 (an internal error was given). Such groups are now left in the compiled
2138 pattern, with a new opcode that causes them to be skipped at execution
2139 time.
2141 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2142 to the way PCRE behaves:
2144 (a) A lone ] character is dis-allowed (Perl treats it as data).
2146 (b) A back reference to an unmatched subpattern matches an empty string
2147 (Perl fails the current match path).
2149 (c) A data ] in a character class must be notated as \] because if the
2150 first data character in a class is ], it defines an empty class. (In
2151 Perl it is not possible to have an empty class.) The empty class []
2152 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2153 The negative empty class [^] matches any one character, independently
2154 of the DOTALL setting.
2156 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2157 non-existent subpattern following a character class starting with ']' and
2158 containing () gave an internal compiling error instead of "reference to
2159 non-existent subpattern". Fortunately, when the pattern did exist, the
2160 compiled code was correct. (When scanning forwards to check for the
2161 existence of the subpattern, it was treating the data ']' as terminating
2162 the class, so got the count wrong. When actually compiling, the reference
2163 was subsequently set up correctly.)
2165 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2166 it was being rejected as not supported by pcre_dfa_exec(), even though
2167 other assertions are supported. I have made pcre_dfa_exec() support
2168 (*FAIL).
2170 16. The implementation of 13c above involved the invention of a new opcode,
2171 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2172 cannot be changed at match time, I realized I could make a small
2173 improvement to matching performance by compiling OP_ALLANY instead of
2174 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2175 on the OP_ANY path.
2177 17. Compiling pcretest on Windows with readline support failed without the
2178 following two fixes: (1) Make the unistd.h include conditional on
2179 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2181 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2182 ncurses library to be included for pcretest when ReadLine support is
2183 requested, but also to allow for it to be overridden. This patch came from
2184 Daniel Bergström.
2186 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2187 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2188 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2189 spotting this.
2192 Version 7.6 28-Jan-08
2193 ---------------------
2195 1. A character class containing a very large number of characters with
2196 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2197 overflow.
2199 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2200 HAVE_LONG_LONG is not defined.
2202 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2203 bring them up to date and include new features. This patch includes:
2205 - Fixed PH's badly added libz and libbz2 support.
2206 - Fixed a problem with static linking.
2207 - Added pcredemo. [But later removed - see 7 below.]
2208 - Fixed dftables problem and added an option.
2209 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2211 - Added readline support for pcretest.
2212 - Added an listing of the option settings after cmake has run.
2214 4. A user submitted a patch to Makefile that makes it easy to create
2215 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2216 Makefile.am that cause it to include this special target, without
2217 affecting anything else. Note that the same mingw target plus all
2218 the other distribution libraries and programs are now supported
2219 when configuring with CMake (see 6 below) instead of with
2220 Configure/Make.
2222 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2223 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2224 exported in the Windows port". It has not yet been confirmed that the patch
2225 solves the problem, but it does no harm.
2227 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2228 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2229 with CMake, and also correct the comment about stack recursion.
2231 7. Remove the automatic building of pcredemo from the ./configure system and
2232 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2233 of a program that users should build themselves after PCRE is installed, so
2234 building it automatically is not really right. What is more, it gave
2235 trouble in some build environments.
2237 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2240 Version 7.5 10-Jan-08
2241 ---------------------
2243 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2244 values in parens when parsing an RE using the C++ wrapper."
2246 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2247 Characters greater than 255 were excluded from the class instead of being
2248 included.
2250 3. The same bug as (2) above applied to negated POSIX classes such as
2251 [:^space:].
2253 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2254 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2255 I have changed it.
2257 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2258 first named subpattern) and a construct such as (?&a) would reference the
2259 first named subpattern whose name started with "a" (in other words, the
2260 length check was missing). Both these problems are fixed. "Subpattern name
2261 expected" is now given for (?&) (a zero-length name), and this patch also
2262 makes it give the same error for \k'' (previously it complained that that
2263 was a reference to a non-existent subpattern).
2265 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2266 this is right because (?- can be followed by option settings as well as by
2267 digits. I have, however, made the messages clearer.
2269 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2270 than the number used in the conditional) now cause a compile-time error.
2271 This is actually not compatible with Perl, which accepts such patterns, but
2272 treats the conditional as always being FALSE (as PCRE used to), but it
2273 seems to me that giving a diagnostic is better.
2275 8. Change "alphameric" to the more common word "alphanumeric" in comments
2276 and messages.
2278 9. Fix two occurrences of "backslash" in comments that should have been
2279 "backspace".
2281 10. Remove two redundant lines of code that can never be obeyed (their function
2282 was moved elsewhere).
2284 11. The program that makes PCRE's Unicode character property table had a bug
2285 which caused it to generate incorrect table entries for sequences of
2286 characters that have the same character type, but are in different scripts.
2287 It amalgamated them into a single range, with the script of the first of
2288 them. In other words, some characters were in the wrong script. There were
2289 thirteen such cases, affecting characters in the following ranges:
2291 U+002b0 - U+002c1
2292 U+0060c - U+0060d
2293 U+0061e - U+00612
2294 U+0064b - U+0065e
2295 U+0074d - U+0076d
2296 U+01800 - U+01805
2297 U+01d00 - U+01d77
2298 U+01d9b - U+01dbf
2299 U+0200b - U+0200f
2300 U+030fc - U+030fe
2301 U+03260 - U+0327f
2302 U+0fb46 - U+0fbb1
2303 U+10450 - U+1049d
2305 12. The -o option (show only the matching part of a line) for pcregrep was not
2306 compatible with GNU grep in that, if there was more than one match in a
2307 line, it showed only the first of them. It now behaves in the same way as
2308 GNU grep.
2310 13. If the -o and -v options were combined for pcregrep, it printed a blank
2311 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2312 does the same. The return code can be used to tell if there were any
2313 non-matching lines.
2315 14. Added --file-offsets and --line-offsets to pcregrep.
2317 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2318 infinitely looping recursion. The bug was that positive lookaheads were not
2319 being skipped when checking for a possible empty match (negative lookaheads
2320 and both kinds of lookbehind were skipped).
2322 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2323 inclusion of <windows.h> to before rather than after the definition of
2324 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2326 17. Specifying a possessive quantifier with a specific limit for a Unicode
2327 character property caused pcre_compile() to compile bad code, which led at
2328 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2329 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2330 caused the error; without that there was no problem.
2332 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2334 19. Added --enable-pcretest-libreadline.
2336 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2337 RE::GlobalReplace(). As a result, the number of replacements returned was
2338 double what it should be. I removed one of the increments, but Craig sent a
2339 later patch that removed the other one (the right fix) and added unit tests
2340 that check the return values (which was not done before).
2342 21. Several CMake things:
2344 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2345 the names libpcre and libpcreposix, not just pcre and pcreposix.
2347 (2) The above change means that pcretest and pcregrep are now correctly
2348 linked with the newly-built libraries, not previously installed ones.
2352 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2353 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2354 UTF-8 newline character). The key issue is that the pattern starts .*;
2355 this means that the match must be either at the beginning, or after a
2356 newline. The bug was in the code for advancing after a failed match and
2357 checking that the new position followed a newline. It was not taking
2358 account of UTF-8 characters correctly.
2360 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2361 character classes. PCRE was not treating the sequence [:...:] as a
2362 character class unless the ... were all letters. Perl, however, seems to
2363 allow any characters between [: and :], though of course it rejects as
2364 unknown any "names" that contain non-letters, because all the known class
2365 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2366 for example, whereas PCRE did not - it did not recognize a POSIX character
2367 class. This seemed a bit dangerous, so the code has been changed to be
2368 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2369 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2370 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2371 Perl does, and where it didn't before.
2373 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2374 Windows environments %n is disabled by default.
2377 Version 7.4 21-Sep-07
2378 ---------------------
2380 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2381 means that a class such as [\s] counted as "explicit reference to CR or
2382 LF". That isn't really right - the whole point of the change was to try to
2383 help when there was an actual mention of one of the two characters. So now
2384 the change happens only if \r or \n (or a literal CR or LF) character is
2385 encountered.
2387 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2388 of both had grown to the point where there were only 3 bits left.
2389 Fortunately, there was spare space in the data structure, and so I have
2390 moved the internal flags into a new 16-bit field to free up more option
2391 bits.
2393 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2394 but did not set the internal JCHANGED flag - either of these is enough to
2395 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2396 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2397 start sets both bits.
2399 4. Added options (at build time, compile time, exec time) to change \R from
2400 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2402 5. doc/pcresyntax.html was missing from the distribution.
2404 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2405 compatibility, even though it is no longer used.
2407 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2408 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2409 windows.h file is present (where different names are used). [This was
2410 reversed later after testing - see 16 below.]
2412 8. Changed all #include <config.h> to #include "config.h". There were also
2413 some further <pcre.h> cases that I changed to "pcre.h".
2415 9. When pcregrep was used with the --colour option, it missed the line ending
2416 sequence off the lines that it output.
2418 10. It was pointed out to me that arrays of string pointers cause lots of
2419 relocations when a shared library is dynamically loaded. A technique of
2420 using a single long string with a table of offsets can drastically reduce
2421 these. I have refactored PCRE in four places to do this. The result is
2422 dramatic:
2424 Originally: 290
2425 After changing UCP table: 187
2426 After changing error message table: 43
2427 After changing table of "verbs" 36
2428 After changing table of Posix names 22
2430 Thanks to the folks working on Gregex for glib for this insight.
2432 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2433 unicode-properties was also set.
2435 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2437 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2438 checked only for CRLF.
2440 14. Added casts to pcretest.c to avoid compiler warnings.
2442 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2444 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2445 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2446 entirely. This removes changes made in 7 above.
2448 17. The CMake files have been updated, and there is now more information about
2449 building with CMake in the NON-UNIX-USE document.
2452 Version 7.3 28-Aug-07
2453 ---------------------
2455 1. In the rejigging of the build system that eventually resulted in 7.1, the
2456 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2457 brackets there is not right, since it causes compilers to look for an
2458 installed pcre.h, not the version that is in the source that is being
2459 compiled (which of course may be different). I have changed it back to:
2461 #include "pcre.h"
2463 I have a vague recollection that the change was concerned with compiling in
2464 different directories, but in the new build system, that is taken care of
2465 by the VPATH setting the Makefile.
2467 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2468 when the subject happened to end in the byte 0x85 (e.g. if the last
2469 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2470 characters but of course it shouldn't be taken as a newline when it is part
2471 of another character. The bug was that, for an unlimited repeat of . in
2472 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2473 characters when looking for a newline.
2475 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2477 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2478 in debug output.
2480 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2481 long printing in the pcrecpp unittest when running under MinGW.
2483 6. ESC_K was left out of the EBCDIC table.
2485 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2486 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2487 limit also applies to "virtual nesting" when a pattern is recursive, and in
2488 this case 1000 isn't so big. I have been able to remove this limit at the
2489 expense of backing off one optimization in certain circumstances. Normally,
2490 when pcre_exec() would call its internal match() function recursively and
2491 immediately return the result unconditionally, it uses a "tail recursion"
2492 feature to save stack. However, when a subpattern that can match an empty
2493 string has an unlimited repetition quantifier, it no longer makes this
2494 optimization. That gives it a stack frame in which to save the data for
2495 checking that an empty string has been matched. Previously this was taken
2496 from the 1000-entry workspace that had been reserved. So now there is no
2497 explicit limit, but more stack is used.
2499 8. Applied Daniel's patches to solve problems with the import/export magic
2500 syntax that is required for Windows, and which was going wrong for the
2501 pcreposix and pcrecpp parts of the library. These were overlooked when this
2502 problem was solved for the main library.
2504 9. There were some crude static tests to avoid integer overflow when computing
2505 the size of patterns that contain repeated groups with explicit upper
2506 limits. As the maximum quantifier is 65535, the maximum group length was
2507 set at 30,000 so that the product of these two numbers did not overflow a
2508 32-bit integer. However, it turns out that people want to use groups that
2509 are longer than 30,000 bytes (though not repeat them that many times).
2510 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2511 made it possible to implement the integer overflow checks in a much more
2512 dynamic way, which I have now done. The artificial limitation on group
2513 length has been removed - we now have only the limit on the total length of
2514 the compiled pattern, which depends on the LINK_SIZE setting.
2516 10. Fixed a bug in the documentation for get/copy named substring when
2517 duplicate names are permitted. If none of the named substrings are set, the
2518 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2519 empty string.
2521 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2522 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2523 because the ] is interpreted as the first data character and the
2524 terminating ] is not found. PCRE has been made compatible with Perl in this
2525 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2526 cause memory overwriting.
2528 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2529 string has been matched (to stop an infinite loop). It was not recognizing
2530 a conditional subpattern that could match an empty string if that
2531 subpattern was within another subpattern. For example, it looped when
2532 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2533 condition was not nested. This bug has been fixed.
2535 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2536 past the start of the subject in the presence of bytes with the top bit
2537 set, for example "\x8aBCD".
2539 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2540 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2542 14. Optimized (?!) to (*FAIL).
2544 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2545 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2546 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2547 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2548 does: it's just the validity check that is more restrictive.
2550 16. Inserted checks for integer overflows during escape sequence (backslash)
2551 processing, and also fixed erroneous offset values for syntax errors during
2552 backslash processing.
2554 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2555 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2557 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2558 caused an overrun.
2560 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2561 something other than just ASCII characters) inside a group that had an
2562 unlimited repeat caused a loop at compile time (while checking to see
2563 whether the group could match an empty string).
2565 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2566 [\P{Any}] did so. (Error in the code for printing property names.)
2568 21. An orphan \E inside a character class could cause a crash.
2570 22. A repeated capturing bracket such as (A)? could cause a wild memory
2571 reference during compilation.
2573 23. There are several functions in pcre_compile() that scan along a compiled
2574 expression for various reasons (e.g. to see if it's fixed length for look
2575 behind). There were bugs in these functions when a repeated \p or \P was
2576 present in the pattern. These operators have additional parameters compared
2577 with \d, etc, and these were not being taken into account when moving along
2578 the compiled data. Specifically:
2580 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2581 length.
2583 (b) An item such as \pL+ within a repeated group could cause crashes or
2584 loops.
2586 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2587 "reference to non-existent subpattern" error.
2589 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2591 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2592 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2594 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2596 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2597 character were causing crashes (broken optimization).
2599 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2600 \p or \P) caused a compile-time loop.
2602 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2603 break. For example, the unstudied pattern [\r\n]A does not match the string
2604 "\r\nA" because change 7.0/46 below moves the current point on by two
2605 characters after failing to match at the start. However, the pattern \nA
2606 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2607 the same is true. There doesn't seem any very clean way out of this, but
2608 what I have chosen to do makes the common cases work: PCRE now takes note
2609 of whether there can be an explicit match for \r or \n anywhere in the
2610 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2611 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2612 pattern has explicit CR or LF references.
2614 29. Added (*CR) etc for changing newline setting at start of pattern.
2617 Version 7.2 19-Jun-07
2618 ---------------------
2620 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2621 which is apparently normally available under Windows.
2623 2. Re-jig the pcregrep tests with different newline settings in an attempt
2624 to make them independent of the local environment's newline setting.
2626 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2628 4. Some of the "internals" tests were previously cut out when the link size
2629 was not 2, because the output contained actual offsets. The recent new
2630 "Z" feature of pcretest means that these can be cut out, making the tests
2631 usable with all link sizes.
2633 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2634 stack recursion. This gives a massive performance boost under BSD, but just
2635 a small improvement under Linux. However, it saves one field in the frame
2636 in all cases.
2638 6. Added more features from the forthcoming Perl 5.10:
2640 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2641 recursion call. It refers to the nth most recently opened parentheses.
2643 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2644 to be opened parentheses.
2646 (c) Conditions that refer to capturing parentheses can be specified
2647 relatively, for example, (?(-2)... or (?(+3)...
2649 (d) \K resets the start of the current match so that everything before
2650 is not part of it.
2652 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2654 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2655 reference syntax.
2657 (g) (?| introduces a group in which the numbering of parentheses in each
2658 alternative starts with the same number.
2660 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2662 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2665 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2666 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2667 for detecting groups that can match an empty string.
2669 9. A pattern with a very large number of alternatives (more than several
2670 hundred) was running out of internal workspace during the pre-compile
2671 phase, where pcre_compile() figures out how much memory will be needed. A
2672 bit of new cunning has reduced the workspace needed for groups with
2673 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2674 workspace instead of running out of the 4096 that are available.
2676 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2678 11. Applied patch from Google to remove an optimization that didn't quite work.
2679 The report of the bug said:
2681 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2682 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2683 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2685 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2686 it matched the wrong number of bytes.
2689 Version 7.1 24-Apr-07
2690 ---------------------
2692 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2693 that is more "standard", making use of automake and other Autotools. There
2694 is some re-arrangement of the files and adjustment of comments consequent
2695 on this.
2697 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2698 for recursive directory scanning broke on some systems because the files
2699 are not scanned in any specific order and on different systems the order
2700 was different. A call to "sort" has been inserted into RunGrepTest for the
2701 approprate test as a short-term fix. In the longer term there may be an
2702 alternative.
2704 3. I had an email from Eric Raymond about problems translating some of PCRE's
2705 man pages to HTML (despite the fact that I distribute HTML pages, some
2706 people do their own conversions for various reasons). The problems
2707 concerned the use of low-level troff macros .br and .in. I have therefore
2708 removed all such uses from the man pages (some were redundant, some could
2709 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2710 HTML has been updated to handle .nf/.fi and to complain if it encounters
2711 .br or .in.
2713 4. Updated comments in configure.ac that get placed in config.h.in and also
2714 arranged for config.h to be included in the distribution, with the name
2715 config.h.generic, for the benefit of those who have to compile without
2716 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2718 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2719 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2720 makevp.bat for the current PCRE, using the additional files
2721 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2723 6. A Windows user reported a minor discrepancy with test 2, which turned out
2724 to be caused by a trailing space on an input line that had got lost in his
2725 copy. The trailing space was an accident, so I've just removed it.
2727 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2728 that is needed.
2730 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2731 as "const" (a) because they are and (b) because it helps the PHP
2732 maintainers who have recently made a script to detect big data structures
2733 in the php code that should be moved to the .rodata section. I remembered
2734 to update Builducptable as well, so it won't revert if ucptable.h is ever
2735 re-created.
2737 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2738 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2739 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2740 support is not required. This saves 1.5-2K of code, which is important in
2741 some applications.
2743 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2744 so as not to refer to the tables, even though these functions will never be
2745 called when UTF-8 support is disabled. Otherwise there are problems with a
2746 shared library.
2748 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2750 (a) It was defining its arguments as char * instead of void *.
2752 (b) It was assuming that all moves were upwards in memory; this was true
2753 a long time ago when I wrote it, but is no longer the case.
2755 The emulated memove() is provided for those environments that have neither
2756 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2757 is clearly not the case, as these two bugs were recently reported.
2759 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2760 and Detrail to create the HTML documentation, the .txt form of the man
2761 pages, and it removes trailing spaces from listed files. It also creates
2762 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2763 case, it wraps all the #defines with #ifndefs. This script should be run
2764 before "make dist".
2766 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2767 with Unicode property support.
2769 (a) For a maximizing quantifier, if the two different cases of the
2770 character were of different lengths in their UTF-8 codings (there are
2771 some cases like this - I found 11), and the matching function had to
2772 back up over a mixture of the two cases, it incorrectly assumed they
2773 were both the same length.
2775 (b) When PCRE was configured to use the heap rather than the stack for
2776 recursion during matching, it was not correctly preserving the data for
2777 the other case of a UTF-8 character when checking ahead for a match
2778 while processing a minimizing repeat. If the check also involved
2779 matching a wide character, but failed, corruption could cause an
2780 erroneous result when trying to check for a repeat of the original
2781 character.
2783 13. Some tidying changes to the testing mechanism:
2785 (a) The RunTest script now detects the internal link size and whether there
2786 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2787 values substituted by "configure". (The RunGrepTest script already did
2788 this for UTF-8.) The configure.ac script no longer substitutes the
2789 relevant variables.
2791 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2792 with length and offset values. This means that the output is different
2793 for different internal link sizes. Test 2 is skipped for link sizes
2794 other than 2 because of this, bypassing the problem. Unfortunately,
2795 there was also a test in test 3 (the locale tests) that used /B and
2796 failed for link sizes other than 2. Rather than cut the whole test out,
2797 I have added a new /Z option to pcretest that replaces the length and
2798 offset values with spaces. This is now used to make test 3 independent
2799 of link size. (Test 2 will be tidied up later.)
2801 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2802 segmentation fault instead of returning the appropriate error message.
2804 15. In multiline mode when the newline sequence was set to "any", the pattern
2805 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2806 This doesn't seem right; it now treats the CRLF combination as the line
2807 ending, and so does not match in that case. It's only a pattern such as ^$
2808 that would hit this one: something like ^ABC$ would have failed after \r
2809 and then tried again after \r\n.
2811 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2812 in an attempt to make files that differ only in their line terminators
2813 compare equal. This works on Linux.
2815 17. Under certain error circumstances pcregrep might try to free random memory
2816 as it exited. This is now fixed, thanks to valgrind.
2818 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2819 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2820 was because its rules for how to advance for /g after matching an empty
2821 string at the end of a line did not allow for this case. They now check for
2822 it specially.
2824 20. pcretest is supposed to handle patterns and data of any length, by
2825 extending its buffers when necessary. It was getting this wrong when the
2826 buffer for a data line had to be extended.
2828 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2829 CRLF as a newline sequence.
2831 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2832 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2833 I have nevertheless tidied it up.
2835 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2837 24. Added a man page for pcre-config.
2840 Version 7.0 19-Dec-06
2841 ---------------------
2843 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2844 moving to gcc 4.1.1.
2846 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2847 sys/time.h, which is documented as needed for this function. It doesn't
2848 seem to matter on Linux, but it showed up on some releases of OS X.
2850 3. It seems that there are systems where bytes whose values are greater than
2851 127 match isprint() in the "C" locale. The "C" locale should be the
2852 default when a C program starts up. In most systems, only ASCII printing
2853 characters match isprint(). This difference caused the output from pcretest
2854 to vary, making some of the tests fail. I have changed pcretest so that:
2856 (a) When it is outputting text in the compiled version of a pattern, bytes
2857 other than 32-126 are always shown as hex escapes.
2859 (b) When it is outputting text that is a matched part of a subject string,
2860 it does the same, unless a different locale has been set for the match
2861 (using the /L modifier). In this case, it uses isprint() to decide.
2863 4. Fixed a major bug that caused incorrect computation of the amount of memory
2864 required for a compiled pattern when options that changed within the
2865 pattern affected the logic of the preliminary scan that determines the
2866 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2867 that the computed length was too small. The symptoms of this bug were
2868 either the PCRE error "internal error: code overflow" from pcre_compile(),
2869 or a glibc crash with a message such as "pcretest: free(): invalid next
2870 size (fast)". Examples of patterns that provoked this bug (shown in
2871 pcretest format) are:
2873 /(?-x: )/x
2874 /(?x)(?-x: \s*#\s*)/
2875 /((?i)[\x{c0}])/8
2876 /(?i:[\x{c0}])/8
2878 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2879 is now done differently.
2881 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2882 wrapper classes; (b) implement a new function in the C++ scanner that is
2883 more efficient than the old way of doing things because it avoids levels of
2884 recursion in the regex matching; (c) add a paragraph to the documentation
2885 for the FullMatch() function.
2887 6. The escape sequence \n was being treated as whatever was defined as
2888 "newline". Not only was this contrary to the documentation, which states
2889 that \n is character 10 (hex 0A), but it also went horribly wrong when
2890 "newline" was defined as CRLF. This has been fixed.
2892 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2893 was being set to -1 for the "end of line" case (supposedly a value that no
2894 character can have). Though this value is never used (the check for end of
2895 line is "zero bytes in current character"), it caused compiler complaints.
2896 I've changed it to 0xffffffff.
2898 8. In pcre_version.c, the version string was being built by a sequence of
2899 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2900 string (as it is for production releases) called a macro with an empty
2901 argument. The C standard says the result of this is undefined. The gcc
2902 compiler treats it as an empty string (which was what was wanted) but it is
2903 reported that Visual C gives an error. The source has been hacked around to
2904 avoid this problem.
2906 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2907 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2908 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2909 of them did).
2911 10. Originally, pcretest opened its input and output without "b"; then I was
2912 told that "b" was needed in some environments, so it was added for release
2913 5.0 to both the input and output. (It makes no difference on Unix-like
2914 systems.) Later I was told that it is wrong for the input on Windows. I've
2915 now abstracted the modes into two macros, to make it easier to fiddle with
2916 them, and removed "b" from the input mode under Windows.
2918 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2920 12. Added -help and --help to pcretest as an official way of being reminded
2921 of the options.
2923 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2924 and pcrecpp.cc because they annoy compilers at high warning levels.
2926 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2928 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2929 scripts are not C programs :-) and which was not noticed because it works
2930 on Linux.
2932 16. pcretest is supposed to handle any length of pattern and data line (as one
2933 line or as a continued sequence of lines) by extending its input buffer if
2934 necessary. This feature was broken for very long pattern lines, leading to
2935 a string of junk being passed to pcre_compile() if the pattern was longer
2936 than about 50K.
2938 17. I have done a major re-factoring of the way pcre_compile() computes the
2939 amount of memory needed for a compiled pattern. Previously, there was code
2940 that made a preliminary scan of the pattern in order to do this. That was
2941 OK when PCRE was new, but as the facilities have expanded, it has become
2942 harder and harder to keep it in step with the real compile phase, and there
2943 have been a number of bugs (see for example, 4 above). I have now found a
2944 cunning way of running the real compile function in a "fake" mode that
2945 enables it to compute how much memory it would need, while actually only
2946 ever using a few hundred bytes of working memory and without too many
2947 tests of the mode. This should make future maintenance and development
2948 easier. A side effect of this work is that the limit of 200 on the nesting
2949 depth of parentheses has been removed (though this was never a serious
2950 limitation, I suspect). However, there is a downside: pcre_compile() now
2951 runs more slowly than before (30% or more, depending on the pattern). I
2952 hope this isn't a big issue. There is no effect on runtime performance.
2954 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2955 newline (only possible for the last line of a file) and it was a
2956 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2958 19. Added additional timing features to pcretest. (1) The -tm option now times
2959 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2960 separate command line item, by a number that specifies the number of
2961 repeats to use when timing. The default is 50000; this gives better
2962 precision, but takes uncomfortably long for very large patterns.
2964 20. Extended pcre_study() to be more clever in cases where a branch of a
2965 subpattern has no definite first character. For example, (a*|b*)[cd] would
2966 previously give no result from pcre_study(). Now it recognizes that the
2967 first character must be a, b, c, or d.
2969 21. There was an incorrect error "recursive call could loop indefinitely" if
2970 a subpattern (or the entire pattern) that was being tested for matching an
2971 empty string contained only one non-empty item after a nested subpattern.
2972 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2973 incorrectly, because the \d was being skipped in the check.
2975 22. The pcretest program now has a new pattern option /B and a command line
2976 option -b, which is equivalent to adding /B to every pattern. This causes
2977 it to show the compiled bytecode, without the additional information that
2978 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2979 is the same as /B/I).
2981 23. A new optimization is now able automatically to treat some sequences such
2982 as a*b as a*+b. More specifically, if something simple (such as a character
2983 or a simple class like \d) has an unlimited quantifier, and is followed by
2984 something that cannot possibly match the quantified thing, the quantifier
2985 is automatically "possessified".
2987 24. A recursive reference to a subpattern whose number was greater than 39
2988 went wrong under certain circumstances in UTF-8 mode. This bug could also
2989 have affected the operation of pcre_study().
2991 25. Realized that a little bit of performance could be had by replacing
2992 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2994 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2996 27. Possessive quantifiers such as a++ were previously implemented by turning
2997 them into atomic groups such as ($>a+). Now they have their own opcodes,
2998 which improves performance. This includes the automatically created ones
2999 from 23 above.
3001 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3002 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3003 the first matched character to be a colon. This applied both to named and
3004 numbered groups.
3006 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3008 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3009 building PCRE on Windows easier, so I have included it in the distribution.
3011 31. There is now a check in pcretest against a ridiculously large number being
3012 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3013 loop, the loop is abandoned.
3015 32. Forward references to subpatterns in conditions such as (?(2)...) where
3016 subpattern 2 is defined later cause pcre_compile() to search forwards in
3017 the pattern for the relevant set of parentheses. This search went wrong
3018 when there were unescaped parentheses in a character class, parentheses
3019 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3021 33. "Subroutine" calls and backreferences were previously restricted to
3022 referencing subpatterns earlier in the regex. This restriction has now
3023 been removed.
3025 34. Added a number of extra features that are going to be in Perl 5.10. On the
3026 whole, these are just syntactic alternatives for features that PCRE had
3027 previously implemented using the Python syntax or my own invention. The
3028 other formats are all retained for compatibility.
3030 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3031 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3032 also .NET compatible.
3034 (b) A recursion or subroutine call to a named group can now be defined as
3035 (?&name) as well as (?P>name).
3037 (c) A backreference to a named group can now be defined as \k<name> or
3038 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3039 5.10, are also .NET compatible.
3041 (d) A conditional reference to a named group can now use the syntax
3042 (?(<name>) or (?('name') as well as (?(name).
3044 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3045 groups (named and numbered) that are never evaluated inline, but can be
3046 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3047 is always false. There may be only one alternative in such a group.
3049 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3050 as the simple (?(R). The condition is true only if the most recent
3051 recursion is that of the given number or name. It does not search out
3052 through the entire recursion stack.
3054 (g) The escape \gN or \g{N} has been added, where N is a positive or
3055 negative number, specifying an absolute or relative reference.
3057 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3058 some "unreachable code" warnings.
3060 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3061 things, this adds five new scripts.
3063 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3064 There were also incompatibilities regarding the handling of \Q..\E inside
3065 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3066 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3068 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3069 matches an empty string, and forcibly breaks the loop. There were bugs in
3070 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3071 against aaaa the result was just "a" rather than "aaaa", for example. Two
3072 separate and independent bugs (that affected different cases) have been
3073 fixed.
3075 39. Refactored the code to abolish the use of different opcodes for small
3076 capturing bracket numbers. This is a tidy that I avoided doing when I
3077 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3078 The new approach is not only tidier, it makes it possible to reduce the
3079 memory needed to fix the previous bug (38).
3081 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3082 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3083 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3084 mode.
3086 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3087 report.
3089 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3090 copy construction and assignment in the C++ wrapper.
3092 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3093 couple of bugs that could have given wrong results in the "--newline=crlf"
3094 case.
3096 44. Added a number of casts and did some reorganization of signed/unsigned int
3097 variables following suggestions from Dair Grant. Also renamed the variable
3098 "this" as "item" because it is a C++ keyword.
3100 45. Arranged for dftables to add
3102 #include "pcre_internal.h"
3104 to pcre_chartables.c because without it, gcc 4.x may remove the array
3105 definition from the final binary if PCRE is built into a static library and
3106 dead code stripping is activated.
3108 46. For an unanchored pattern, if a match attempt fails at the start of a
3109 newline sequence, and the newline setting is CRLF or ANY, and the next two
3110 characters are CRLF, advance by two characters instead of one.
3113 Version 6.7 04-Jul-06
3114 ---------------------
3116 1. In order to handle tests when input lines are enormously long, pcretest has
3117 been re-factored so that it automatically extends its buffers when
3118 necessary. The code is crude, but this _is_ just a test program. The
3119 default size has been increased from 32K to 50K.
3121 2. The code in pcre_study() was using the value of the re argument before
3122 testing it for NULL. (Of course, in any sensible call of the function, it
3123 won't be NULL.)
3125 3. The memmove() emulation function in pcre_internal.h, which is used on
3126 systems that lack both memmove() and bcopy() - that is, hardly ever -
3127 was missing a "static" storage class specifier.
3129 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3130 containing an extended class (one that cannot be represented by a bitmap
3131 because it contains high-valued characters or Unicode property items, e.g.
3132 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3133 pattern, but PCRE should not loop if you do not (it no longer does).
3134 [Detail: two cases were found: (a) a repeated subpattern containing an
3135 extended class; (b) a recursive reference to a subpattern that followed a
3136 previous extended class. It wasn't skipping over the extended class
3137 correctly when UTF-8 mode was not set.]
3139 5. A negated single-character class was not being recognized as fixed-length
3140 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3141 compile error "lookbehind assertion is not fixed length".
3143 6. The RunPerlTest auxiliary script was showing an unexpected difference
3144 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3145 write a Perl script that can interpret lines of an input file either as
3146 byte characters or as UTF-8, which is what "perltest" was being required to
3147 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3148 can't do is switch easily at run time between having the "use utf8;" pragma
3149 or not. In the end, I fudged it by using the RunPerlTest script to insert
3150 "use utf8;" explicitly for the UTF-8 tests.
3152 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3153 the end of the subject string, contrary to the documentation and to what
3154 Perl does. This was true of both matching functions. Now it matches only at
3155 the start of the subject and immediately after *internal* newlines.
3157 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3158 a pointer to an int instead of a pointer to an unsigned long int. This
3159 caused problems on 64-bit systems.
3161 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3162 instance of the 'standard' template library not being so standard".
3164 10. There was no check on the number of named subpatterns nor the maximum
3165 length of a subpattern name. The product of these values is used to compute
3166 the size of the memory block for a compiled pattern. By supplying a very
3167 long subpattern name and a large number of named subpatterns, the size
3168 computation could be caused to overflow. This is now prevented by limiting
3169 the length of names to 32 characters, and the number of named subpatterns
3170 to 10,000.
3172 11. Subpatterns that are repeated with specific counts have to be replicated in
3173 the compiled pattern. The size of memory for this was computed from the
3174 length of the subpattern and the repeat count. The latter is limited to
3175 65535, but there was no limit on the former, meaning that integer overflow
3176 could in principle occur. The compiled length of a repeated subpattern is
3177 now limited to 30,000 bytes in order to prevent this.
3179 12. Added the optional facility to have named substrings with the same name.
3181 13. Added the ability to use a named substring as a condition, using the
3182 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3183 are numbers (not recommended). Forward references are permitted.
3185 14. Added forward references in named backreferences (if you see what I mean).
3187 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3188 pattern could run off the end of the subject. For example, the pattern
3189 "(?s)(.{1,5})"8 did this with the subject "ab".
3191 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3192 PCRE_CASELESS was set when matching characters that were quantified with ?
3193 or *.
3195 17. A character class other than a single negated character that had a minimum
3196 but no maximum quantifier - for example [ab]{6,} - was not handled
3197 correctly by pce_dfa_exec(). It would match only one character.
3199 18. A valid (though odd) pattern that looked like a POSIX character
3200 class but used an invalid character after [ (for example [[,abc,]]) caused
3201 pcre_compile() to give the error "Failed: internal error: code overflow" or
3202 in some cases to crash with a glibc free() error. This could even happen if
3203 the pattern terminated after [[ but there just happened to be a sequence of
3204 letters, a binary zero, and a closing ] in the memory that followed.
3206 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3207 over the years. Originally (before any Unicode support), just the bottom 8
3208 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3209 output from "man perlunicode" includes this:
3211 The regular expression compiler produces polymorphic opcodes. That
3212 is, the pattern adapts to the data and automatically switches to
3213 the Unicode character scheme when presented with Unicode data--or
3214 instead uses a traditional byte scheme when presented with byte
3215 data.
3217 Sadly, a wide octal escape does not cause a switch, and in a string with
3218 no other multibyte characters, these octal escapes are treated as before.
3219 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3220 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3221 Unicode string.
3223 I have not perpetrated such confusion in PCRE. Up till now, it took just
3224 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3225 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3226 translate to the appropriate multibyte character.
3228 29. Applied some refactoring to reduce the number of warnings from Microsoft
3229 and Borland compilers. This has included removing the fudge introduced
3230 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3231 a warning about an unused variable.
3233 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3234 characters since release 4.0, because Perl (from release 5.004) does not.
3235 [Or at least, is documented not to: some releases seem to be in conflict
3236 with the documentation.] However, when a pattern was studied with
3237 pcre_study() and all its branches started with \s, PCRE still included VT
3238 as a possible starting character. Of course, this did no harm; it just
3239 caused an unnecessary match attempt.
3241 22. Removed a now-redundant internal flag bit that recorded the fact that case
3242 dependency changed within the pattern. This was once needed for "required
3243 byte" processing, but is no longer used. This recovers a now-scarce options
3244 bit. Also moved the least significant internal flag bit to the most-
3245 significant bit of the word, which was not previously used (hangover from
3246 the days when it was an int rather than a uint) to free up another bit for
3247 the future.
3249 23. Added support for CRLF line endings as well as CR and LF. As well as the
3250 default being selectable at build time, it can now be changed at runtime
3251 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3252 specify that it is scanning data with non-default line endings.
3254 24. Changed the definition of CXXLINK to make it agree with the definition of
3255 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3257 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3258 recursions. This makes a big different to stack usage for some patterns.
3260 26. If a subpattern containing a named recursion or subroutine reference such
3261 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3262 the space required for the compiled pattern went wrong and gave too small a
3263 value. Depending on the environment, this could lead to "Failed: internal
3264 error: code overflow at offset 49" or "glibc detected double free or
3265 corruption" errors.
3267 27. Applied patches from Google (a) to support the new newline modes and (b) to
3268 advance over multibyte UTF-8 characters in GlobalReplace.
3270 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3271 difference for some implementation of PCRE in some Windows version.
3273 29. Added some extra testing facilities to pcretest:
3275 \q<number> in a data line sets the "match limit" value
3276 \Q<number> in a data line sets the "match recursion limt" value
3277 -S <number> sets the stack size, where <number> is in megabytes
3279 The -S option isn't available for Windows.
3282 Version 6.6 06-Feb-06
3283 ---------------------
3285 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3286 in pcreposix.h. I have copied the definition from pcre.h.
3288 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3289 because pcre.h is no longer a built file.
3291 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3292 not normally included in the compiled code.
3295 Version 6.5 01-Feb-06
3296 ---------------------
3298 1. When using the partial match feature with pcre_dfa_exec(), it was not
3299 anchoring the second and subsequent partial matches at the new starting
3300 point. This could lead to incorrect results. For example, with the pattern
3301 /1234/, partially matching against "123" and then "a4" gave a match.
3303 2. Changes to pcregrep:
3305 (a) All non-match returns from pcre_exec() were being treated as failures
3306 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3307 error message is output. Some extra information is given for the
3309 probably the only errors that are likely to be caused by users (by
3310 specifying a regex that has nested indefinite repeats, for instance).
3311 If there are more than 20 of these errors, pcregrep is abandoned.
3313 (b) A binary zero was treated as data while matching, but terminated the
3314 output line if it was written out. This has been fixed: binary zeroes
3315 are now no different to any other data bytes.
3317 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3318 used to set a locale for matching. The --locale=xxxx long option has
3319 been added (no short equivalent) to specify a locale explicitly on the
3320 pcregrep command, overriding the environment variables.
3322 (d) When -B was used with -n, some line numbers in the output were one less
3323 than they should have been.
3325 (e) Added the -o (--only-matching) option.
3327 (f) If -A or -C was used with -c (count only), some lines of context were
3328 accidentally printed for the final match.
3330 (g) Added the -H (--with-filename) option.
3332 (h) The combination of options -rh failed to suppress file names for files
3333 that were found from directory arguments.
3335 (i) Added the -D (--devices) and -d (--directories) options.
3337 (j) Added the -F (--fixed-strings) option.
3339 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3341 (l) Added the --colo(u)r option.
3343 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3344 is not present by default.
3346 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3347 items such as (?R) or (?1), when the recursion could match a number of
3348 alternatives. If it matched one of the alternatives, but subsequently,
3349 outside the recursion, there was a failure, the code tried to back up into
3350 the recursion. However, because of the way PCRE is implemented, this is not
3351 possible, and the result was an incorrect result from the match.
3353 In order to prevent this happening, the specification of recursion has
3354 been changed so that all such subpatterns are automatically treated as
3355 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3357 4. I had overlooked the fact that, in some locales, there are characters for
3358 which isalpha() is true but neither isupper() nor islower() are true. In
3359 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3360 and ordfeminine) are like this. This affected the treatment of \w and \W
3361 when they appeared in character classes, but not when they appeared outside
3362 a character class. The bit map for "word" characters is now created
3363 separately from the results of isalnum() instead of just taking it from the
3364 upper, lower, and digit maps. (Plus the underscore character, of course.)
3366 5. The above bug also affected the handling of POSIX character classes such as
3367 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3368 permanent tables. Instead, the bit maps for such a class were previously
3369 created as the appropriate unions of the upper, lower, and digit bitmaps.
3370 Now they are created by subtraction from the [[:word:]] class, which has
3371 its own bitmap.
3373 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3374 It is created by subtracting the vertical space characters (\x09, \x0a,
3375 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3376 subtraction was done in the overall bitmap for a character class, meaning
3377 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3378 be recognized. This bug has been fixed.
3380 7. Patches from the folks at Google:
3382 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3383 real life, but is still worth protecting against".
3385 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3386 regular expressions".
3388 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3389 have it.
3391 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3392 "configure" and the latter not, in order to fix a problem somebody had
3393 with compiling the Arg class on HP-UX.
3395 (e) Improve the error-handling of the C++ wrapper a little bit.
3397 (f) New tests for checking recursion limiting.
3399 8. The pcre_memmove() function, which is used only if the environment does not
3400 have a standard memmove() function (and is therefore rarely compiled),
3401 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3402 returning a result (though PCRE never actually uses the result).
3404 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3405 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3406 returned instead of calling malloc() with an overflowing number that would
3407 most likely cause subsequent chaos.
3409 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3411 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3412 with this option is matched, the nmatch and pmatch options of regexec() are
3413 ignored.
3415 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3416 provided in case anyone wants to the the POSIX interface with UTF-8
3417 strings.
3419 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3420 C++ linking (needed for some HP-UX environments).
3422 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3423 (unused parameter) and in the pcre_printint() function (omitted "default"
3424 switch label when the default is to do nothing).
3426 15. Added some code to make it possible, when PCRE is compiled as a C++
3427 library, to replace subject pointers for pcre_exec() with a smart pointer
3428 class, thus making it possible to process discontinuous strings.
3430 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3431 much the same function. They were added by different people who were trying
3432 to make PCRE easy to compile on non-Unix systems. It has been suggested
3433 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3434 for compiling on Windows systems. I have therefore replaced it with
3435 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3436 defaults to "extern" for C or "extern C" for C++, which works fine on
3437 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3438 SCOPE with something explicit in config.h. In addition:
3440 (a) pcreposix.h still had just "extern" instead of either of these macros;
3441 I have replaced it with PCRE_DATA_SCOPE.
3443 (b) Functions such as _pcre_xclass(), which are internal to the library,
3444 but external in the C sense, all had PCRE_EXPORT in their definitions.
3445 This is apparently wrong for the Windows case, so I have removed it.
3446 (It makes no difference on Unix-like systems.)
3448 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3449 of recursive calls to match(). This is different to MATCH_LIMIT because
3450 that limits the total number of calls to match(), not all of which increase
3451 the depth of recursion. Limiting the recursion depth limits the amount of
3452 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3453 when PCRE is compiled, and changed at run time. A patch from Google adds
3454 this functionality to the C++ interface.
3456 18. Changes to the handling of Unicode character properties:
3458 (a) Updated the table to Unicode 4.1.0.
3460 (b) Recognize characters that are not in the table as "Cn" (undefined).
3462 (c) I revised the way the table is implemented to a much improved format
3463 which includes recognition of ranges. It now supports the ranges that
3464 are defined in UnicodeData.txt, and it also amalgamates other
3465 characters into ranges. This has reduced the number of entries in the
3466 table from around 16,000 to around 3,000, thus reducing its size
3467 considerably. I realized I did not need to use a tree structure after
3468 all - a binary chop search is just as efficient. Having reduced the
3469 number of entries, I extended their size from 6 bytes to 8 bytes to
3470 allow for more data.
3472 (d) Added support for Unicode script names via properties such as \p{Han}.
3474 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3475 matching that character.
3477 20. When matching a repeated Unicode property with a minimum greater than zero,
3478 (for example \pL{2,}), PCRE could look past the end of the subject if it
3479 reached it while seeking the minimum number of characters. This could
3480 happen only if some of the characters were more than one byte long, because
3481 there is a check for at least the minimum number of bytes.
3483 21. Refactored the implementation of \p and \P so as to be more general, to
3484 allow for more different types of property in future. This has changed the
3485 compiled form incompatibly. Anybody with saved compiled patterns that use
3486 \p or \P will have to recompile them.
3488 22. Added "Any" and "L&" to the supported property types.
3490 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3491 but give a compile time error if the value is greater than 0xff.
3493 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3494 accidentally not being installed or uninstalled.
3496 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3497 made were to insert the current release number. This seemed silly, because
3498 it made things harder for people building PCRE on systems that don't run
3499 "configure". I have turned pcre.h into a distributed file, no longer built
3500 by "configure", with the version identification directly included. There is
3501 no longer a pcre.h.in file.
3503 However, this change necessitated a change to the pcre-config script as
3504 well. It is built from pcre-config.in, and one of the substitutions was the
3505 release number. I have updated configure.ac so that ./configure now finds
3506 the release number by grepping pcre.h.
3508 26. Added the ability to run the tests under valgrind.
3511 Version 6.4 05-Sep-05
3512 ---------------------
3514 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3515 "--" to be printed when multiple files were scanned, even when none of the
3516 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3517 consider it to be a bug, and have restored the previous behaviour.
3519 2. A couple of code tidies to get rid of compiler warnings.
3521 3. The pcretest program used to cheat by referring to symbols in the library
3522 whose names begin with _pcre_. These are internal symbols that are not
3523 really supposed to be visible externally, and in some environments it is
3524 possible to suppress them. The cheating is now confined to including
3525 certain files from the library's source, which is a bit cleaner.
3527 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3528 file's purpose clearer.
3530 5. Reorganized pcre_ucp_findchar().
3533 Version 6.3 15-Aug-05
3534 ---------------------
3536 1. The file libpcre.pc.in did not have general read permission in the tarball.
3538 2. There were some problems when building without C++ support:
3540 (a) If C++ support was not built, "make install" and "make test" still
3541 tried to test it.
3543 (b) There were problems when the value of CXX was explicitly set. Some
3544 changes have been made to try to fix these, and ...
3546 (c) --disable-cpp can now be used to explicitly disable C++ support.
3548 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3549 backslash in a target when C++ was disabled. This confuses some
3550 versions of "make", apparently. Using an intermediate variable solves
3551 this. (Same for CPP_LOBJ.)
3553 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3554 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3555 necessary on certain architectures.
3557 4. Added a setting of -export-symbols-regex to the link command to remove
3558 those symbols that are exported in the C sense, but actually are local
3559 within the library, and not documented. Their names all begin with
3560 "_pcre_". This is not a perfect job, because (a) we have to except some
3561 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3562 available (and never for static libraries). I have made a note to try to
3563 find a way round (a) in the future.
3566 Version 6.2 01-Aug-05
3567 ---------------------
3569 1. There was no test for integer overflow of quantifier values. A construction
3570 such as {1111111111111111} would give undefined results. What is worse, if
3571 a minimum quantifier for a parenthesized subpattern overflowed and became
3572 negative, the calculation of the memory size went wrong. This could have
3573 led to memory overwriting.
3575 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3577 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3578 operating environments where this matters.
3580 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3581 PCRE options from within the C++ wrapper.
3583 5. Named capturing subpatterns were not being correctly counted when a pattern
3584 was compiled. This caused two problems: (a) If there were more than 100
3585 such subpatterns, the calculation of the memory needed for the whole
3586 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3587 back references of the form \12, where the number was greater than 9, were
3588 not recognized as back references, even though there were sufficient
3589 previous subpatterns.
3591 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3592 versions of gcc, e.g. 2.95.4.
3595 Version 6.1 21-Jun-05
3596 ---------------------
3598 1. There was one reference to the variable "posix" in pcretest.c that was not
3599 surrounded by "#if !defined NOPOSIX".
3601 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3602 the cross-check on the old pcre_info() function, for the benefit of the
3603 cut-down version of PCRE that is currently imported into Exim.
3605 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3606 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3607 patterns that start (?i) but I don't think that matters. The use of (?i) is
3608 just an example; this all applies to the other options as well.
3610 4. Since libtool seems to echo the compile commands it is issuing, the output
3611 from "make" can be reduced a bit by putting "@" in front of each libtool
3612 compile command.
3614 5. Patch from the folks at Google for configure.in to be a bit more thorough
3615 in checking for a suitable C++ installation before trying to compile the
3616 C++ stuff. This should fix a reported problem when a compiler was present,
3617 but no suitable headers.
3619 6. The man pages all had just "PCRE" as their title. I have changed them to
3620 be the relevant file name. I have also arranged that these names are
3621 retained in the file doc/pcre.txt, which is a concatenation in text format
3622 of all the man pages except the little individual ones for each function.
3624 7. The NON-UNIX-USE file had not been updated for the different set of source
3625 files that come with release 6. I also added a few comments about the C++
3626 wrapper.
3629 Version 6.0 07-Jun-05
3630 ---------------------
3632 1. Some minor internal re-organization to help with my DFA experiments.
3634 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3635 didn't matter for the library itself when fully configured, but did matter
3636 when compiling without UCP support, or within Exim, where the ucp files are
3637 not imported.
3639 3. Refactoring of the library code to split up the various functions into
3640 different source modules. The addition of the new DFA matching code (see
3641 below) to a single monolithic source would have made it really too
3642 unwieldy, quite apart from causing all the code to be include in a
3643 statically linked application, when only some functions are used. This is
3644 relevant even without the DFA addition now that patterns can be compiled in
3645 one application and matched in another.
3647 The downside of splitting up is that there have to be some external
3648 functions and data tables that are used internally in different modules of
3649 the library but which are not part of the API. These have all had their
3650 names changed to start with "_pcre_" so that they are unlikely to clash
3651 with other external names.
3653 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3654 a different (DFA) algorithm. Although it is slower than the original
3655 function, it does have some advantages for certain types of matching
3656 problem.
3658 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3659 including restarting after a partial match.
3661 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3662 defined when compiling for Windows was sent to me. I have put it into the
3663 code, though I have no means of testing or verifying it.
3665 7. Added the pcre_refcount() auxiliary function.
3667 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3668 match before or at the first newline in the subject string. In pcretest,
3669 the /f option on a pattern can be used to set this.
3671 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3672 would behave wrongly. This has been present in PCRE since release 4.0.
3674 10. A number of changes to the pcregrep command:
3676 (a) Refactored how -x works; insert ^(...)$ instead of setting
3677 PCRE_ANCHORED and checking the length, in preparation for adding
3678 something similar for -w.
3680 (b) Added the -w (match as a word) option.
3682 (c) Refactored the way lines are read and buffered so as to have more
3683 than one at a time available.
3685 (d) Implemented a pcregrep test script.
3687 (e) Added the -M (multiline match) option. This allows patterns to match
3688 over several lines of the subject. The buffering ensures that at least
3689 8K, or the rest of the document (whichever is the shorter) is available
3690 for matching (and similarly the previous 8K for lookbehind assertions).
3692 (f) Changed the --help output so that it now says
3694 -w, --word-regex(p)
3696 instead of two lines, one with "regex" and the other with "regexp"
3697 because that confused at least one person since the short forms are the
3698 same. (This required a bit of code, as the output is generated
3699 automatically from a table. It wasn't just a text change.)
3701 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3702 option but starts with a hyphen. Could be a pattern or a path name
3703 starting with a hyphen, for instance.
3705 (h) "-" can be given as a file name to represent stdin.
3707 (i) When file names are being printed, "(standard input)" is used for
3708 the standard input, for compatibility with GNU grep. Previously
3709 "<stdin>" was used.
3711 (j) The option --label=xxx can be used to supply a name to be used for
3712 stdin when file names are being printed. There is no short form.
3714 (k) Re-factored the options decoding logic because we are going to add
3715 two more options that take data. Such options can now be given in four
3716 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3718 (l) Added the -A, -B, and -C options for requesting that lines of context
3719 around matches be printed.
3721 (m) Added the -L option to print the names of files that do not contain
3722 any matching lines, that is, the complement of -l.
3724 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3725 continue to scan other files.
3727 (o) The -s option was incorrectly implemented. For compatibility with other
3728 greps, it now suppresses the error message for a non-existent or non-
3729 accessible file (but not the return code). There is a new option called
3730 -q that suppresses the output of matching lines, which was what -s was
3731 previously doing.
3733 (p) Added --include and --exclude options to specify files for inclusion
3734 and exclusion when recursing.
3736 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3737 Hopefully, it now does.
3739 12. Missing cast in pcre_study().
3741 13. Added an "uninstall" target to the makefile.
3743 14. Replaced "extern" in the function prototypes in Makefile.in with
3744 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3745 world, but is set differently for Windows.
3747 15. Added a second compiling function called pcre_compile2(). The only
3748 difference is that it has an extra argument, which is a pointer to an
3749 integer error code. When there is a compile-time failure, this is set
3750 non-zero, in addition to the error test pointer being set to point to an
3751 error message. The new argument may be NULL if no error number is required
3752 (but then you may as well call pcre_compile(), which is now just a
3753 wrapper). This facility is provided because some applications need a
3754 numeric error indication, but it has also enabled me to tidy up the way
3755 compile-time errors are handled in the POSIX wrapper.
3757 16. Added VPATH=.libs to the makefile; this should help when building with one
3758 prefix path and installing with another. (Or so I'm told by someone who
3759 knows more about this stuff than I do.)
3761 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3762 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3763 match everything, including newlines. This is not POSIX-compatible, but
3764 somebody wanted the feature. From pcretest it can be activated by using
3765 both the P and the s flags.
3767 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3769 19. libpcre.pc was being incorrectly installed as executable.
3771 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3772 it now also looks for '\r' so that it will work unmodified on Windows.
3774 21. Added Google's contributed C++ wrapper to the distribution.
3776 22. Added some untidy missing memory free() calls in pcretest, to keep
3777 Electric Fence happy when testing.
3781 Version 5.0 13-Sep-04
3782 ---------------------
3784 1. Internal change: literal characters are no longer packed up into items
3785 containing multiple characters in a single byte-string. Each character
3786 is now matched using a separate opcode. However, there may be more than one
3787 byte in the character in UTF-8 mode.
3789 2. The pcre_callout_block structure has two new fields: pattern_position and
3790 next_item_length. These contain the offset in the pattern to the next match
3791 item, and its length, respectively.
3793 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3794 insertion of callouts before each pattern item. Added the /C option to
3795 pcretest to make use of this.
3797 4. On the advice of a Windows user, the lines
3799 #if defined(_WIN32) || defined(WIN32)
3800 _setmode( _fileno( stdout ), 0x8000 );
3801 #endif /* defined(_WIN32) || defined(WIN32) */
3803 have been added to the source of pcretest. This apparently does useful
3804 magic in relation to line terminators.
3806 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3807 for the benefit of those environments where the "b" makes a difference.
3809 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3810 to know about it. I have put a hack into configure.in that adds in code
3811 to set GCC=yes if CC=icc. This seems to end up at a point in the
3812 generated configure script that is early enough to affect the setting of
3813 compiler options, which is what is needed, but I have no means of testing
3814 whether it really works. (The user who reported this had patched the
3815 generated configure script, which of course I cannot do.)
3817 LATER: After change 22 below (new libtool files), the configure script
3818 seems to know about icc (and also ecc). Therefore, I have commented out
3819 this hack in configure.in.
3821 7. Added support for pkg-config (2 patches were sent in).
3823 8. Negated POSIX character classes that used a combination of internal tables
3824 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3825 [[:^ascii]]. Typically, they would match almost any characters. The other
3826 POSIX classes were not broken in this way.
3828 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3829 to find the match, as PCRE was deluded into thinking that the match had to
3830 start at the start point or following a newline. The same bug applied to
3831 patterns with negative forward assertions or any backward assertions
3832 preceding ".*" at the start, unless the pattern required a fixed first
3833 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3835 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3836 starting at the last subject character, bytes beyond the end of the subject
3837 string were read.
3839 11. Renamed the variable "class" as "classbits" to make life easier for C++
3840 users. (Previously there was a macro definition, but it apparently wasn't
3841 enough.)
3843 12. Added the new field "tables" to the extra data so that tables can be passed
3844 in at exec time, or the internal tables can be re-selected. This allows
3845 a compiled regex to be saved and re-used at a later time by a different
3846 program that might have everything at different addresses.
3848 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3849 -R library as well as a -L library.
3851 14. The debugging options of pcretest (-d on the command line or D on a
3852 pattern) showed incorrect output for anything following an extended class
3853 that contained multibyte characters and which was followed by a quantifier.
3855 15. Added optional support for general category Unicode character properties
3856 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3857 support. It adds about 90K to the size of the library. The meanings of the
3858 inbuilt class escapes such as \d and \s have NOT been changed.
3860 16. Updated pcredemo.c to include calls to free() to release the memory for the
3861 compiled pattern.
3863 17. The generated file chartables.c was being created in the source directory
3864 instead of in the building directory. This caused the build to fail if the
3865 source directory was different from the building directory, and was
3866 read-only.
3868 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3869 file. No doubt somebody will tell me if they don't make sense... Also added
3870 Dan Mooney's comments about building on OpenVMS.
3872 19. Added support for partial matching via the PCRE_PARTIAL option for
3873 pcre_exec() and the \P data escape in pcretest.
3875 20. Extended pcretest with 3 new pattern features:
3877 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3878 write the compiled pattern to the file whose name is "rest-of-line".
3879 This is a straight binary dump of the data, with the saved pointer to
3880 the character tables forced to be NULL. The study data, if any, is
3881 written too. After writing, pcretest reads a new pattern.
3883 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3884 compiled pattern from the given file. There must not be any
3885 occurrences of "<" in the file name (pretty unlikely); if there are,
3886 pcretest will instead treat the initial "<" as a pattern delimiter.
3887 After reading in the pattern, pcretest goes on to read data lines as
3888 usual.
3890 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3891 and 16-bit fields in a compiled pattern, to simulate a pattern that
3892 was compiled on a host of opposite endianness.
3894 21. The pcre-exec() function can now cope with patterns that were compiled on
3895 hosts of opposite endianness, with this restriction:
3897 As for any compiled expression that is saved and used later, the tables
3898 pointer field cannot be preserved; the extra_data field in the arguments
3899 to pcre_exec() should be used to pass in a tables address if a value
3900 other than the default internal tables were used at compile time.
3902 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3903 now diagnosed as an error. Previously, most of the time, a negative number
3904 would have been treated as zero, but if in addition "ovector" was passed as
3905 NULL, a crash could occur.
3907 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3908 new versions from the libtool 1.5 distribution (the last one is a copy of
3909 a file called libtool.m4). This seems to have fixed the need to patch
3910 "configure" to support Darwin 1.3 (which I used to do). However, I still
3911 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3912 workstation).
3914 24. Changed the PCRE licence to be the more standard "BSD" licence.
3917 Version 4.5 01-Dec-03
3918 ---------------------
3920 1. There has been some re-arrangement of the code for the match() function so
3921 that it can be compiled in a version that does not call itself recursively.
3922 Instead, it keeps those local variables that need separate instances for
3923 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3924 needs to "recurse". Keeping track of where control must go is done by means
3925 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3926 hide most of the details from the main code, and operates only if
3927 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3928 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3929 operating.
3931 To make it easier for callers to provide specially tailored get/free
3932 functions for this usage, two new functions, pcre_stack_malloc, and
3933 pcre_stack_free, are used. They are always called in strict stacking order,
3934 and the size of block requested is always the same.
3936 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3937 PCRE has been compiled to use the stack or the heap for recursion. The
3938 -C option of pcretest uses this to show which version is compiled.
3940 A new data escape \S, is added to pcretest; it causes the amounts of store
3941 obtained and freed by both kinds of malloc/free at match time to be added
3942 to the output.
3944 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3945 what's available on my current Linux desktop machine.
3947 3. When matching a UTF-8 string, the test for a valid string at the start has
3948 been extended. If start_offset is not zero, PCRE now checks that it points
3949 to a byte that is the start of a UTF-8 character. If not, it returns
3950 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3951 this is necessary because there may be backward assertions in the pattern.
3952 When matching the same subject several times, it may save resources to use
3953 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3955 4. The code for checking the validity of UTF-8 strings has been tightened so
3956 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3957 containing "overlong sequences".
3959 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3960 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3961 should have been "&", but it just so happened that all the cases this let
3962 through by mistake were picked up later in the function.
3964 6. I had used a variable called "isblank" - this is a C99 function, causing
3965 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3967 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3968 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3969 so the version is shown; (c) stop "make test" echoing "./RunTest".
3971 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3973 9. The prototype for memmove() for systems that don't have it was using
3974 size_t, but the inclusion of the header that defines size_t was later. I've
3975 moved the #includes for the C headers earlier to avoid this.
3977 10. Added some adjustments to the code to make it easier to compiler on certain
3978 special systems:
3980 (a) Some "const" qualifiers were missing.
3981 (b) Added the macro EXPORT before all exported functions; by default this
3982 is defined to be empty.
3983 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3984 that it reads its output file name as an argument instead of writing
3985 to the standard output and assuming this can be redirected.
3987 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3988 class containing characters with values greater than 255, PCRE compilation
3989 went into a loop.
3991 12. A recursive reference to a subpattern that was within another subpattern
3992 that had a minimum quantifier of zero caused PCRE to crash. For example,
3993 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3994 recursion. If the recursively-called subpattern itself had a zero repeat,
3995 that was OK.
3997 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3998 buffer into which it was copied (for escape processing) was still set at
3999 1024, so long lines caused crashes.
4001 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4002 "internal error: code overflow...". This applied to any character class
4003 that was followed by a possessive quantifier.
4005 15. Modified the Makefile to add libpcre.la as a prerequisite for
4006 libpcreposix.la because I was told this is needed for a parallel build to
4007 work.
4009 16. If a pattern that contained .* following optional items at the start was
4010 studied, the wrong optimizing data was generated, leading to matching
4011 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4012 matching string must start with a or b or c. The correct conclusion for
4013 this pattern is that a match can start with any character.
4016 Version 4.4 13-Aug-03
4017 ---------------------
4019 1. In UTF-8 mode, a character class containing characters with values between
4020 127 and 255 was not handled correctly if the compiled pattern was studied.
4021 In fixing this, I have also improved the studying algorithm for such
4022 classes (slightly).
4024 2. Three internal functions had redundant arguments passed to them. Removal
4025 might give a very teeny performance improvement.
4027 3. Documentation bug: the value of the capture_top field in a callout is *one
4028 more than* the number of the hightest numbered captured substring.
4030 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4031 in incorrectly linking with a previously installed version. They now link
4032 explicitly with libpcre.la.
4034 5. configure.in no longer needs to recognize Cygwin specially.
4036 6. A problem in pcre.in for Windows platforms is fixed.
4038 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4039 pcretest, it used to include the size of the study block as part of its
4040 output. Unfortunately, the structure contains a field that has a different
4041 size on different hardware architectures. This meant that the tests that
4042 showed this size failed. As the block is currently always of a fixed size,
4043 this information isn't actually particularly useful in pcretest output, so
4044 I have just removed it.
4046 8. Three pre-processor statements accidentally did not start in column 1.
4047 Sadly, there are *still* compilers around that complain, even though
4048 standard C has not required this for well over a decade. Sigh.
4050 9. In pcretest, the code for checking callouts passed small integers in the
4051 callout_data field, which is a void * field. However, some picky compilers
4052 complained about the casts involved for this on 64-bit systems. Now
4053 pcretest passes the address of the small integer instead, which should get
4054 rid of the warnings.
4056 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4057 both compile and run time, and gives an error if an invalid UTF-8 sequence
4058 is found. There is a option for disabling this check in cases where the
4059 string is known to be correct and/or the maximum performance is wanted.
4061 11. In response to a bug report, I changed one line in Makefile.in from
4063 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4064 to
4065 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4067 to look similar to other lines, but I have no way of telling whether this
4068 is the right thing to do, as I do not use Windows. No doubt I'll get told
4069 if it's wrong...
4072 Version 4.3 21-May-03
4073 ---------------------
4075 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4076 Makefile.
4078 2. Some refactoring to improve the quality of the code:
4080 (i) The utf8_table... variables are now declared "const".
4082 (ii) The code for \cx, which used the "case flipping" table to upper case
4083 lower case letters, now just substracts 32. This is ASCII-specific,
4084 but the whole concept of \cx is ASCII-specific, so it seems
4085 reasonable.
4087 (iii) PCRE was using its character types table to recognize decimal and
4088 hexadecimal digits in the pattern. This is silly, because it handles
4089 only 0-9, a-f, and A-F, but the character types table is locale-
4090 specific, which means strange things might happen. A private
4091 table is now used for this - though it costs 256 bytes, a table is
4092 much faster than multiple explicit tests. Of course, the standard
4093 character types table is still used for matching digits in subject
4094 strings against \d.
4096 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4097 ending in _t are). So I've renamed it as ESC_tee.
4099 3. The first argument for regexec() in the POSIX wrapper should have been
4100 defined as "const".
4102 4. Changed pcretest to use malloc() for its buffers so that they can be
4103 Electric Fenced for debugging.
4105 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4106 to read one or more bytes before the start of the subject string. Often this
4107 had no effect on PCRE's behaviour, but in some circumstances it could
4108 provoke a segmentation fault.
4110 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4111 to try to read one or more bytes before the start of the subject string.
4113 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4114 UTF-8 support could misbehave in various ways if the subject string
4115 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4116 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4117 back over UTF-8 characters.)
4120 Version 4.2 14-Apr-03
4121 ---------------------
4123 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4125 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4126 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4127 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4128 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4129 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4131 Note: automatic setting of the BUILD variables is not yet working
4132 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4133 compile-time but not at link-time
4134 [LINK]: use for linking executables only
4135 make different versions for Windows and non-Windows
4136 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4137 libraries
4138 [LINK_FOR_BUILD]: new variable
4139 [OBJEXT]: use throughout
4140 [EXEEXT]: use throughout
4141 <winshared>: new target
4142 <wininstall>: new target
4143 <dftables.o>: use native compiler
4144 <dftables>: use native linker
4145 <install>: handle Windows platform correctly
4146 <clean>: ditto
4147 <check>: ditto
4148 copy DLL to top builddir before testing
4150 As part of these changes, -no-undefined was removed again. This was reported
4151 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4152 in any case.
4154 3. Some tidies to get rid of compiler warnings:
4156 . In the match_data structure, match_limit was an unsigned long int, whereas
4157 match_call_count was an int. I've made them both unsigned long ints.
4159 . In pcretest the fact that a const uschar * doesn't automatically cast to
4160 a void * provoked a warning.
4162 . Turning on some more compiler warnings threw up some "shadow" variables
4163 and a few more missing casts.
4165 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4166 option, a class that contained a single character with a value between 128
4167 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4169 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4170 option, a class that contained several characters, but with at least one
4171 whose value was between 128 and 255 caused PCRE to crash.
4174 Version 4.1 12-Mar-03
4175 ---------------------
4177 1. Compiling with gcc -pedantic found a couple of places where casts were
4178 needed, and a string in dftables.c that was longer than standard compilers are
4179 required to support.
4181 2. Compiling with Sun's compiler found a few more places where the code could
4182 be tidied up in order to avoid warnings.
4184 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4185 first of these names is deprecated in the latest Autoconf in favour of the name
4186 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4187 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4188 analogy I have changed it to CFLAGS_FOR_BUILD.
4190 4. Added -no-undefined to the linking command in the Makefile, because this is
4191 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4192 linking step for the pcreposix library.
4194 5. PCRE was failing to diagnose the case of two named groups with the same
4195 name.
4197 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4198 literal character that is needed in the subject for a match, and scans along to
4199 ensure that it is present before embarking on the full matching process. This
4200 saves time in cases of nested unlimited repeats that are never going to match.
4201 Problem: the scan can take a lot of time if the subject is very long (e.g.
4202 megabytes), thus penalizing straightforward matches. It is now done only if the
4203 amount of subject to be scanned is less than 1000 bytes.
4205 7. A lesser problem with the same optimization is that it was recording the
4206 first character of an anchored pattern as "needed", thus provoking a search
4207 right along the subject, even when the first match of the pattern was going to
4208 fail. The "needed" character is now not set for anchored patterns, unless it
4209 follows something in the pattern that is of non-fixed length. Thus, it still
4210 fulfils its original purpose of finding quick non-matches in cases of nested
4211 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4214 Version 4.0 17-Feb-03
4215 ---------------------
4217 1. If a comment in an extended regex that started immediately after a meta-item
4218 extended to the end of string, PCRE compiled incorrect data. This could lead to
4219 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4221 2. Moved to autoconf 2.53 and libtool 1.4.2.
4223 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4224 the special perltest8 script is no longer needed - all the tests can be run
4225 from a single perltest script.
4227 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4228 by \s. It has now been removed in PCRE. This means it isn't recognized as
4229 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4230 class [:space:] *does* include VT, thereby creating a mess.
4232 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4233 space and tab.
4235 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4236 its new features into the main test script, reducing the number of scripts.
4238 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4239 were backward compatible, and made the (?i) apply to the whole pattern, as if
4240 /i were given. Now it behaves more logically, and applies the option setting
4241 only to what follows. PCRE has been changed to follow suit. However, if it
4242 finds options settings right at the start of the pattern, it extracts them into
4243 the global options, as before. Thus, they show up in the info data.
4245 8. Added support for the \Q...\E escape sequence. Characters in between are
4246 treated as literals. This is slightly different from Perl in that $ and @ are
4247 also handled as literals inside the quotes. In Perl, they will cause variable
4248 interpolation. Note the following examples:
4250 Pattern PCRE matches Perl matches
4252 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4253 \Qabc\$xyz\E abc\$xyz abc\$xyz
4254 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4256 For compatibility with Perl, \Q...\E sequences are recognized inside character
4257 classes as well as outside them.
4259 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4260 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4261 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4262 signed/unsigned warnings.
4264 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4265 option for pcretest, so I've replaced it by a simple function that does just
4266 that job.
4268 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4269 "pcregrep -".
4271 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4272 Java package. This provides some syntactic sugar for simple cases of what my
4273 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4274 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4275 item, you can use this simplified notation. Note that only makes sense with
4276 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4277 greediness, whatever the setting of the PCRE_UNGREEDY option.
4279 13. A change of greediness default within a pattern was not taking effect at
4280 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4281 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4282 was abstracted outside.
4284 14. PCRE now supports the \G assertion. It is true when the current matching
4285 position is at the start point of the match. This differs from \A when the
4286 starting offset is non-zero. Used with the /g option of pcretest (or similar
4287 code), it works in the same way as it does for Perl's /g option. If all
4288 alternatives of a regex begin with \G, the expression is anchored to the start
4289 match position, and the "anchored" flag is set in the compiled expression.
4291 15. Some bugs concerning the handling of certain option changes within patterns
4292 have been fixed. These applied to options other than (?ims). For example,
4293 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4294 been the other way round. Some of this was related to change 7 above.
4296 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4297 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4298 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4299 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4301 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4302 mode. Unlike ".", it always matches newline, whatever the setting of
4303 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4304 assertions. Perl allows it, but it doesn't (in general) work because it can't
4305 calculate the length of the lookbehind. At least, that's the case for Perl
4306 5.8.0 - I've been told they are going to document that it doesn't work in
4307 future.
4309 18. Added an error diagnosis for escapes that PCRE does not support: these are
4310 \L, \l, \N, \P, \p, \U, \u, and \X.
4312 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4313 reading past the end of the pattern in cases such as /[abcd/.
4315 20. PCRE was getting more memory than necessary for patterns with classes that
4316 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4318 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4319 compiling PCRE for use with Virtual Pascal.
4321 22. Small fix to the Makefile to make it work properly if the build is done
4322 outside the source tree.
4324 23. Added a new extension: a condition to go with recursion. If a conditional
4325 subpattern starts with (?(R) the "true" branch is used if recursion has
4326 happened, whereas the "false" branch is used only at the top level.
4328 24. When there was a very long string of literal characters (over 255 bytes
4329 without UTF support, over 250 bytes with UTF support), the computation of how
4330 much memory was required could be incorrect, leading to segfaults or other
4331 strange effects.
4333 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4334 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4335 there was a subsequent back reference to those brackets. This meant that, for
4336 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4337 possible to check for precisely this case. All we can do is abandon the
4338 optimization if .* occurs inside capturing brackets when there are any back
4339 references whatsoever. (See below for a better fix that came later.)
4341 26. The handling of the optimization for finding the first character of a
4342 non-anchored pattern, and for finding a character that is required later in the
4343 match were failing in some cases. This didn't break the matching; it just
4344 failed to optimize when it could. The way this is done has been re-implemented.
4346 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4348 28. Added a new feature that provides some of the functionality that Perl
4349 provides with (?{...}). The facility is termed a "callout". The way it is done
4350 in PCRE is for the caller to provide an optional function, by setting
4351 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4352 global variable. By default it is unset, which disables all calling out. To get
4353 the function called, the regex must include (?C) at appropriate points. This
4354 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4355 This provides a means of identifying different callout points. When PCRE
4356 reaches such a point in the regex, if pcre_callout has been set, the external
4357 function is called. It is provided with data in a structure called
4358 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4359 matching continues; if it returns a non-zero value, the match at the current
4360 point fails. However, backtracking will occur if possible. [This was changed
4361 later and other features added - see item 49 below.]
4363 29. pcretest is upgraded to test the callout functionality. It provides a
4364 callout function that displays information. By default, it shows the start of
4365 the match and the current position in the text. There are some new data escapes
4366 to vary what happens:
4368 \C+ in addition, show current contents of captured substrings
4369 \C- do not supply a callout function
4370 \C!n return 1 when callout number n is reached
4371 \C!n!m return 1 when callout number n is reached for the mth time
4373 30. If pcregrep was called with the -l option and just a single file name, it
4374 output "<stdin>" if a match was found, instead of the file name.
4376 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4377 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4378 pcre_exec(). This saves a malloc/free per call. The default value of
4379 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4380 when configuring.
4382 32. The default maximum size of a compiled pattern is 64K. There have been a
4383 few cases of people hitting this limit. The code now uses macros to handle the
4384 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4385 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4386 configuring. Tests 2 and 5 work only with 2-byte links because they output
4387 debugging information about compiled patterns.
4389 33. Internal code re-arrangements:
4391 (a) Moved the debugging function for printing out a compiled regex into
4392 its own source file (printint.c) and used #include to pull it into
4393 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4394 separate copies.
4396 (b) Defined the list of op-code names for debugging as a macro in
4397 internal.h so that it is next to the definition of the opcodes.
4399 (c) Defined a table of op-code lengths for simpler skipping along compiled
4400 code. This is again a macro in internal.h so that it is next to the
4401 definition of the opcodes.
4403 34. Added support for recursive calls to individual subpatterns, along the
4404 lines of Robin Houston's patch (but implemented somewhat differently).
4406 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4407 allow it to read and process whole directories in Win32. This code was
4408 contributed by Lionel Fourquaux; it has not been tested by me.
4410 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4411 used to name a group. Names consist of alphanumerics and underscores, and must
4412 be unique. Back references use the syntax (?P=name) and recursive calls use
4413 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4414 numbers. The function pcre_fullinfo() can be used after compilation to extract
4415 a name/number map. There are three relevant calls:
4417 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4418 PCRE_INFO_NAMECOUNT yields the number of entries
4419 PCRE_INFO_NAMETABLE yields a pointer to the map.
4421 The map is a vector of fixed-size entries. The size of each entry depends on
4422 the length of the longest name used. The first two bytes of each entry are the
4423 group number, most significant byte first. There follows the corresponding
4424 name, zero terminated. The names are in alphabetical order.
4426 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4427 case instead of 255. Making it the same both with and without UTF-8 support
4428 means that the same test output works with both.
4430 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4431 calling malloc() with a zero argument.
4433 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4434 optimization. I've improved things by keeping a bitmap of backreferences with
4435 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4436 fact referenced, the optimization can be applied. It is unlikely that a
4437 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4438 the match to follow \n) will appear inside brackets with a number greater than
4439 31, but if it does, any back reference > 31 suppresses the optimization.
4441 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4442 of disabling numbered capturing parentheses. Any opening parenthesis that is
4443 not followed by ? behaves as if it were followed by ?: but named parentheses
4444 can still be used for capturing (and they will acquire numbers in the usual
4445 way).
4447 41. Redesigned the return codes from the match() function into yes/no/error so
4448 that errors can be passed back from deep inside the nested calls. A malloc
4449 failure while inside a recursive subpattern call now causes the
4450 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4452 42. It is now possible to set a limit on the number of times the match()
4453 function is called in a call to pcre_exec(). This facility makes it possible to
4454 limit the amount of recursion and backtracking, though not in a directly
4455 obvious way, because the match() function is used in a number of different
4456 circumstances. The count starts from zero for each position in the subject
4457 string (for non-anchored patterns). The default limit is, for compatibility, a
4458 large number, namely 10 000 000. You can change this in two ways:
4460 (a) When configuring PCRE before making, you can use --with-match-limit=n
4461 to set a default value for the compiled library.
4463 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4464 a different value is set. See 45 below.
4466 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4468 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4469 of things that can be changed at compile time. The first argument specifies
4470 what is wanted and the second points to where the information is to be placed.
4471 The current list of available information is:
4475 The output is an integer that is set to one if UTF-8 support is available;
4476 otherwise it is set to zero.
4480 The output is an integer that it set to the value of the code that is used for
4481 newline. It is either LF (10) or CR (13).
4485 The output is an integer that contains the number of bytes used for internal
4486 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4490 The output is an integer that contains the threshold above which the POSIX
4491 interface uses malloc() for output vectors. See item 31 above.
4495 The output is an unsigned integer that contains the default limit of the number
4496 of match() calls in a pcre_exec() execution. See 42 above.
4498 44. pcretest has been upgraded by the addition of the -C option. This causes it
4499 to extract all the available output from the new pcre_config() function, and to
4500 output it. The program then exits immediately.
4502 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4503 order to support additional features. One way would have been to define
4504 pcre_exec2() (for example) with extra arguments, but this would not have been
4505 extensible, and would also have required all calls to the original function to
4506 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4507 is used for passing in "extra" data from pcre_study().
4509 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4510 contains the following fields:
4512 flags a bitmap indicating which of the following fields are set
4513 study_data opaque data from pcre_study()
4514 match_limit a way of specifying a limit on match() calls for a specific
4515 call to pcre_exec()
4516 callout_data data for callouts (see 49 below)
4518 The flag bits are also defined in pcre.h, and are
4524 The pcre_study() function now returns one of these new pcre_extra blocks, with
4525 the actual study data pointed to by the study_data field, and the
4526 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4527 before. That is, this change is entirely upwards-compatible and requires no
4528 change to existing code.
4530 If you want to pass in additional data to pcre_exec(), you can either place it
4531 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4532 block.
4534 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4535 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4536 times with different match limits, until it finds the minimum value needed for
4537 pcre_exec() to complete. The value is then output. This can be instructive; for
4538 most simple matches the number is quite small, but for pathological cases it
4539 gets very large very quickly.
4541 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4542 returns the size of the data block pointed to by the study_data field in a
4543 pcre_extra block, that is, the value that was passed as the argument to
4544 pcre_malloc() when PCRE was getting memory in which to place the information
4545 created by pcre_study(). The fourth argument should point to a size_t variable.
4546 pcretest has been extended so that this information is shown after a successful
4547 pcre_study() call when information about the compiled regex is being displayed.
4549 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4550 because what follows is always an absolute path. (Later: it turns out that this
4551 is more than cosmetic for MinGW, because it doesn't like empty path
4552 components.)
4554 49. Some changes have been made to the callout feature (see 28 above):
4556 (i) A callout function now has three choices for what it returns:
4558 0 => success, carry on matching
4559 > 0 => failure at this point, but backtrack if possible
4560 < 0 => serious error, return this value from pcre_exec()
4562 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4563 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4564 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4565 use by callout functions. It will never be used by PCRE itself.
4567 (ii) The pcre_extra structure (see 45 above) has a void * field called
4568 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4569 pcre_callout_block structure has a field of the same name. The contents of
4570 the field passed in the pcre_extra structure are passed to the callout
4571 function in the corresponding field in the callout block. This makes it
4572 easier to use the same callout-containing regex from multiple threads. For
4573 testing, the pcretest program has a new data escape
4575 \C*n pass the number n (may be negative) as callout_data
4577 If the callout function in pcretest receives a non-zero value as
4578 callout_data, it returns that value.
4580 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4581 there were some redundant $(CFLAGS) in commands that are now specified as
4582 $(LINK), which already includes $(CFLAGS).
4584 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4585 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4586 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4587 one-byte characters throughout. Note that case-insensitive matching applies
4588 only to characters whose values are less than 256. PCRE doesn't support the
4589 notion of cases for higher-valued characters.
4591 (i) A character class whose characters are all within 0-255 is handled as
4592 a bit map, and the map is inverted for negative classes. Previously, a
4593 character > 255 always failed to match such a class; however it should
4594 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4596 (ii) A negated character class with a single character < 255 is coded as
4597 "not this character" (OP_NOT). This wasn't working properly when the test
4598 character was multibyte, either singly or repeated.
4600 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4601 mode, for example: \x{100}{2,3}.
4603 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4604 singly or repeated) now correctly test multibyte characters. However,
4605 PCRE doesn't recognize any characters with values greater than 255 as
4606 digits, spaces, or word characters. Such characters always match \D, \S,
4607 and \W, and never match \d, \s, or \w.
4609 (v) Classes may now contain characters and character ranges with values
4610 greater than 255. For example: [ab\x{100}-\x{400}].
4612 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4613 PCRE in UTF-8 mode.
4615 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4616 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4617 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4618 value.)
4620 53. The single man page has become too large. I have therefore split it up into
4621 a number of separate man pages. These also give rise to individual HTML pages;
4622 these are now put in a separate directory, and there is an index.html page that
4623 lists them all. Some hyperlinking between the pages has been installed.
4625 54. Added convenience functions for handling named capturing parentheses.
4627 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4628 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4629 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4630 are faulted.
4632 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4633 calling configure. These values are used when compiling the dftables.c program
4634 which is run to generate the source of the default character tables. They
4635 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4636 you will need to set these values.
4638 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4641 Version 3.9 02-Jan-02
4642 ---------------------
4644 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4646 2. If --disable-static was given, the building process failed when trying to
4647 build pcretest and pcregrep. (For some reason it was using libtool to compile
4648 them, which is not right, as they aren't part of the library.)
4651 Version 3.8 18-Dec-01
4652 ---------------------
4654 1. The experimental UTF-8 code was completely screwed up. It was packing the
4655 bytes in the wrong order. How dumb can you get?
4658 Version 3.7 29-Oct-01
4659 ---------------------
4661 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4662 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4663 this didn't happen under Solaris 8, where I normally test things.
4665 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4666 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4667 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4668 this fix an hour or so after the initial 3.7 release.)
4671 Version 3.6 23-Oct-01
4672 ---------------------
4674 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4675 offsets passed as NULL with zero offset count.
4677 2. The config.guess and config.sub files had not been updated when I moved to
4678 the latest autoconf.
4681 Version 3.5 15-Aug-01
4682 ---------------------
4684 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4685 had been forgotten.
4687 2. By using declared but undefined structures, we can avoid using "void"
4688 definitions in pcre.h while keeping the internal definitions of the structures
4689 private.
4691 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4692 user point of view, this means that both static and shared libraries are built
4693 by default, but this can be individually controlled. More of the work of
4694 handling this static/shared cases is now inside libtool instead of PCRE's make
4695 file.
4697 4. The pcretest utility is now installed along with pcregrep because it is
4698 useful for users (to test regexs) and by doing this, it automatically gets
4699 relinked by libtool. The documentation has been turned into a man page, so
4700 there are now .1, .txt, and .html versions in /doc.
4702 5. Upgrades to pcregrep:
4703 (i) Added long-form option names like gnu grep.
4704 (ii) Added --help to list all options with an explanatory phrase.
4705 (iii) Added -r, --recursive to recurse into sub-directories.
4706 (iv) Added -f, --file to read patterns from a file.
4708 6. pcre_exec() was referring to its "code" argument before testing that
4709 argument for NULL (and giving an error if it was NULL).
4711 7. Upgraded Makefile.in to allow for compiling in a different directory from
4712 the source directory.
4714 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4715 options bits, the pointer it was passed was to an int instead of to an unsigned
4716 long int. This mattered only on 64-bit systems.
4718 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4719 generated) instead of pcre.in, which it its source. Also made the same change
4720 in several of the .c files.
4722 10. A new release of gcc defines printf() as a macro, which broke pcretest
4723 because it had an ifdef in the middle of a string argument for printf(). Fixed
4724 by using separate calls to printf().
4726 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4727 script, to force use of CR or LF instead of \n in the source. On non-Unix
4728 systems, the value can be set in config.h.
4730 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4731 absolute limit. Changed the text of the error message to make this clear, and
4732 likewise updated the man page.
4734 13. The limit of 99 on the number of capturing subpatterns has been removed.
4735 The new limit is 65535, which I hope will not be a "real" limit.
4738 Version 3.4 22-Aug-00
4739 ---------------------
4741 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4743 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4746 Version 3.3 01-Aug-00
4747 ---------------------
4749 1. If an octal character was given, but the value was greater than \377, it
4750 was not getting masked to the least significant bits, as documented. This could
4751 lead to crashes in some systems.
4753 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4754 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4756 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4757 These just pass their arguments on to (pcre_free)(), but they are provided
4758 because some uses of PCRE bind it to non-C systems that can call its functions,
4759 but cannot call free() or pcre_free() directly.
4761 4. Add "make test" as a synonym for "make check". Corrected some comments in
4762 the Makefile.
4764 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4765 Makefile.
4767 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4768 command called pgrep for grepping around the active processes.
4770 7. Added the beginnings of support for UTF-8 character strings.
4772 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4773 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4774 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4775 out for the ar command.)
4778 Version 3.2 12-May-00
4779 ---------------------
4781 This is purely a bug fixing release.
4783 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4784 of ZA. This was just one example of several cases that could provoke this bug,
4785 which was introduced by change 9 of version 2.00. The code for breaking
4786 infinite loops after an iteration that matches an empty string was't working
4787 correctly.
4789 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4790 when matched against abbab (for example). After matching an empty string, it
4791 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4792 caused it to match further down the string than it should.
4794 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4795 was there because it doesn't seem to be needed, and it causes trouble on some
4796 systems, as it is not a Standard C header. It has been removed.
4798 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4799 were reported on the Macintosh. The changes were from
4801 while ((c = *(++ptr)) != 0 && c != '\n');
4802 to
4803 while ((c = *(++ptr)) != 0 && c != '\n') ;
4805 Totally extraordinary, but if that's what it takes...
4807 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4808 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4809 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4810 assumes the way PCRE uses memmove() (always moving upwards).
4812 6. PCRE is being used in one environment where strchr() is not available. There
4813 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4814 faster code anyway.
4817 Version 3.1 09-Feb-00
4818 ---------------------
4820 The only change in this release is the fixing of some bugs in Makefile.in for
4821 the "install" target:
4823 (1) It was failing to install pcreposix.h.
4825 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4828 Version 3.0 01-Feb-00
4829 ---------------------
4831 1. Add support for the /+ modifier to perltest (to output $` like it does in
4832 pcretest).
4834 2. Add support for the /g modifier to perltest.
4836 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4837 matches null strings.
4839 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4840 pattern. Perl treats empty patterns specially - it reuses the most recent
4841 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4842 effect.
4844 5. The POSIX interface was broken in that it was just handing over the POSIX
4845 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4846 required a bigger vector, with some working space on the end. This means that
4847 the POSIX wrapper now has to get and free some memory, and copy the results.
4849 6. Added some simple autoconf support, placing the test data and the
4850 documentation in separate directories, re-organizing some of the
4851 information files, and making it build pcre-config (a GNU standard). Also added
4852 libtool support for building PCRE as a shared library, which is now the
4853 default.
4855 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4856 09 are not valid octal constants. Single digits will be used for minor values
4857 less than 10.
4859 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4860 existing programs that set these in the POSIX interface can use PCRE without
4861 modification.
4863 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4864 return all that pcre_info() returns, plus additional data. The pcre_info()
4865 function is retained for compatibility, but is considered to be obsolete.
4867 10. Added experimental recursion feature (?R) to handle one common case that
4868 Perl 5.6 will be able to do with (?p{...}).
4870 11. Added support for POSIX character classes like [:alpha:], which Perl is
4871 adopting.
4874 Version 2.08 31-Aug-99
4875 ----------------------
4877 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4878 trying to match at the startoffset position, but instead was moving forward to
4879 the next newline as if a previous match had failed.
4881 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4882 and could get into a loop if a null string was matched other than at the start
4883 of the subject.
4885 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4886 be distinguished at compile time, and for completeness also added PCRE_DATE.
4888 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4889 in GnuWin32 environments.
4892 Version 2.07 29-Jul-99
4893 ----------------------
4895 1. The documentation is now supplied in plain text form and HTML as well as in
4896 the form of man page sources.
4898 2. C++ compilers don't like assigning (void *) values to other pointer types.
4899 In particular this affects malloc(). Although there is no problem in Standard
4900 C, I've put in casts to keep C++ compilers happy.
4902 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4903 should be (const char *).
4905 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4906 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4907 However, I haven't made this a standard facility. The documentation doesn't
4908 mention it, and the Makefile doesn't support it.
4910 5. The Makefile now contains an "install" target, with editable destinations at
4911 the top of the file. The pcretest program is not installed.
4913 6. pgrep -V now gives the PCRE version number and date.
4915 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4916 causing the entire string to be ignored, instead of just the last character.
4918 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4919 non-matching string, it can take a very, very long time, even for strings of
4920 quite modest length, because of the nested recursion. PCRE now does better in
4921 some of these cases. It does this by remembering the last required literal
4922 character in the pattern, and pre-searching the subject to ensure it is present
4923 before running the real match. In other words, it applies a heuristic to detect
4924 some types of certain failure quickly, and in the above example, if presented
4925 with a string that has no trailing " it gives "no match" very quickly.
4927 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4928 other alternatives are tried instead.
4931 Version 2.06 09-Jun-99
4932 ----------------------
4934 1. Change pcretest's output for amount of store used to show just the code
4935 space, because the remainder (the data block) varies in size between 32-bit and
4936 64-bit systems.
4938 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4939 start matching at. This allows lookbehinds to work when searching for multiple
4940 occurrences in a string.
4942 3. Added additional options to pcretest for testing multiple occurrences:
4944 /+ outputs the rest of the string that follows a match
4945 /g loops for multiple occurrences, using the new startoffset argument
4946 /G loops for multiple occurrences by passing an incremented pointer
4948 4. PCRE wasn't doing the "first character" optimization for patterns starting
4949 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4950 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4951 the letter 'x'. On long subject strings, this gives a significant speed-up.
4954 Version 2.05 21-Apr-99
4955 ----------------------
4957 1. Changed the type of magic_number from int to long int so that it works
4958 properly on 16-bit systems.
4960 2. Fixed a bug which caused patterns starting with .* not to work correctly
4961 when the subject string contained newline characters. PCRE was assuming
4962 anchoring for such patterns in all cases, which is not correct because .* will
4963 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4964 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4965 must be retried after every newline in the subject.
4968 Version 2.04 18-Feb-99
4969 ----------------------
4971 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4972 computation of the store needed to hold the pattern was incorrect (too large).
4973 If such patterns were nested a few deep, this could multiply and become a real
4974 problem.
4976 2. Added /M option to pcretest to show the memory requirement of a specific
4977 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4979 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4980 compiled in such a way that the backtracking after subsequent failure was
4981 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4982 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4985 Version 2.03 02-Feb-99
4986 ----------------------
4988 1. Fixed typo and small mistake in man page.
4990 2. Added 4th condition (GPL supersedes if conflict) and created separate
4991 LICENCE file containing the conditions.
4993 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4994 Perl, that is the internal \ allows the delimiter to be included in the
4995 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4996 the final delimiter, add \ to the end of the pattern (to test the error).
4998 4. Added the convenience functions for extracting substrings after a successful
4999 match. Updated pcretest to make it able to test these functions.
5002 Version 2.02 14-Jan-99
5003 ----------------------
5005 1. Initialized the working variables associated with each extraction so that
5006 their saving and restoring doesn't refer to uninitialized store.
5008 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5009 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5010 fix the problem.
5012 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5013 calls, and wasn't printing the correct value for compiling calls. Increased the
5014 default value of LOOPREPEAT, and the number of significant figures in the
5015 times.
5017 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5019 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5020 a building problem on Windows NT with a FAT file system.
5023 Version 2.01 21-Oct-98
5024 ----------------------
5026 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5027 to character tables built by pcre_maketables() in the current locale. If NULL
5028 is passed, the default tables are used.
5031 Version 2.00 24-Sep-98
5032 ----------------------
5034 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5035 it any more.
5037 2. Allow quantification of (?>) groups, and make it work correctly.
5039 3. The first character computation wasn't working for (?>) groups.
5041 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5042 end of the subject) and add 5.005's \z, which really does match only at the
5043 very end of the subject.
5045 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5047 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5048 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5049 localized options. All options to pcre_study() were also removed.
5051 7. Add other new features from 5.005:
5053 $(?<= positive lookbehind
5054 $(?<! negative lookbehind
5055 (?imsx-imsx) added the unsetting capability
5056 such a setting is global if at outer level; local otherwise
5057 (?imsx-imsx:) non-capturing groups with option setting
5058 (?(cond)re|re) conditional pattern matching
5060 A backreference to itself in a repeated group matches the previous
5061 captured string.
5063 8. General tidying up of studying (both automatic and via "study")
5064 consequential on the addition of new assertions.
5066 9. As in 5.005, unlimited repeated groups that could match an empty substring
5067 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5068 runtime if any iteration does actually match an empty substring.
5070 10. Include the RunTest script in the distribution.
5072 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5073 discrepancies, some of which were old and were also with respect to 5.004. They
5074 have now been fixed.
5077 Version 1.09 28-Apr-98
5078 ----------------------
5080 1. A negated single character class followed by a quantifier with a minimum
5081 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5082 program crashes, or just wrong answers. This did not apply to negated classes
5083 containing more than one character, or to minima other than one.
5086 Version 1.08 27-Mar-98
5087 ----------------------
5089 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5091 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5092 latter must appear before anything that relies on it in the pattern.
5095 Version 1.07 16-Feb-98
5096 ----------------------
5098 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5099 repeat of a potentially empty string).
5102 Version 1.06 23-Jan-98
5103 ----------------------
5105 1. Added Markus Oberhumer's little patches for C++.
5107 2. Literal strings longer than 255 characters were broken.
5110 Version 1.05 23-Dec-97
5111 ----------------------
5113 1. Negated character classes containing more than one character were failing if
5114 PCRE_CASELESS was set at run time.
5117 Version 1.04 19-Dec-97
5118 ----------------------
5120 1. Corrected the man page, where some "const" qualifiers had been omitted.
5122 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5123 input syntax.
5125 3. Fixed memory leak which occurred when a regex with back references was
5126 matched with an offsets vector that wasn't big enough. The temporary memory
5127 that is used in this case wasn't being freed if the match failed.
5129 4. Tidied pcretest to ensure it frees memory that it gets.
5131 5. Temporary memory was being obtained in the case where the passed offsets
5132 vector was exactly big enough.
5134 6. Corrected definition of offsetof() from change 5 below.
5136 7. I had screwed up change 6 below and broken the rules for the use of
5137 setjmp(). Now fixed.
5140 Version 1.03 18-Dec-97
5141 ----------------------
5143 1. A erroneous regex with a missing opening parenthesis was correctly
5144 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5145 on some systems.
5147 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5148 it was reported that one broken compiler failed on the former because "code" is
5149 also an independent variable.
5151 3. The erroneous regex a[]b caused an array overrun reference.
5153 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5154 fail on data ending with that character. (It was going on too far, and checking
5155 the next character, typically a binary zero.) This was specific to the
5156 optimized code for single-character negative classes.
5158 5. Added a contributed patch from the TIN world which does the following:
5160 + Add an undef for memmove, in case the the system defines a macro for it.
5162 + Add a definition of offsetof(), in case there isn't one. (I don't know
5163 the reason behind this - offsetof() is part of the ANSI standard - but
5164 it does no harm).
5166 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5167 most of the places where whitespace preceded '#'. I have given up and
5168 allowed the remaining 2 cases to be at the margin.
5170 + Rename some variables in pcre to eliminate shadowing. This seems very
5171 pedantic, but does no harm, of course.
5173 6. Moved the call to setjmp() into its own function, to get rid of warnings
5174 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5176 7. Constructs such as \d{8,} were compiling into the equivalent of
5177 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5178 outcome, but in this particular case used more store than had been allocated,
5179 which caused the bug to be discovered because it threw up an internal error.
5181 8. The debugging code in both pcre and pcretest for outputting the compiled
5182 form of a regex was going wrong in the case of back references followed by
5183 curly-bracketed repeats.
5186 Version 1.02 12-Dec-97
5187 ----------------------
5189 1. Typos in pcre.3 and comments in the source fixed.
5191 2. Applied a contributed patch to get rid of places where it used to remove
5192 'const' from variables, and fixed some signed/unsigned and uninitialized
5193 variable warnings.
5195 3. Added the "runtest" target to Makefile.
5197 4. Set default compiler flag to -O2 rather than just -O.
5200 Version 1.01 19-Nov-97
5201 ----------------------
5203 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5204 like /([ab]*)*/, that is, for classes with more than one character in them.
5206 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5207 as /((?>a*))*/ (a PCRE_EXTRA facility).
5210 Version 1.00 18-Nov-97
5211 ----------------------
5213 1. Added compile-time macros to support systems such as SunOS4 which don't have
5214 memmove() or strerror() but have other things that can be used instead.
5216 2. Arranged that "make clean" removes the executables.
5219 Version 0.99 27-Oct-97
5220 ----------------------
5222 1. Fixed bug in code for optimizing classes with only one character. It was
5223 initializing a 32-byte map regardless, which could cause it to run off the end
5224 of the memory it had got.
5226 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5229 Version 0.98 22-Oct-97
5230 ----------------------
5232 1. Fixed bug in code for handling temporary memory usage when there are more
5233 back references than supplied space in the ovector. This could cause segfaults.
5236 Version 0.97 21-Oct-97
5237 ----------------------
5239 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5241 2. Optimized negated single characters not to use a bit map.
5243 3. Brought error texts together as macro definitions; clarified some of them;
5244 fixed one that was wrong - it said "range out of order" when it meant "invalid
5245 escape sequence".
5247 4. Changed some char * arguments to const char *.
5249 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5251 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5252 pcretest.
5255 Version 0.96 16-Oct-97
5256 ----------------------
5258 1. Added a simple "pgrep" utility to the distribution.
5260 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5261 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5262 where "ddd" means "one or more decimal digits".
5264 3. Fixed serious bug. If a pattern had a back reference, but the call to
5265 pcre_exec() didn't supply a large enough ovector to record the related
5266 identifying subpattern, the match always failed. PCRE now remembers the number
5267 of the largest back reference, and gets some temporary memory in which to save
5268 the offsets during matching if necessary, in order to ensure that
5269 backreferences always work.
5271 4. Increased the compatibility with Perl in a number of ways:
5273 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5274 to request this handling. The option can be set at compile or exec time.
5276 (b) $ matches before a terminating newline by default; an option
5277 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5278 mode). The option can be set at compile or exec time.
5280 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5281 the same as Perl's. If the decimal number it represents is less than 10
5282 or there aren't that many previous left capturing parentheses, an octal
5283 escape is read. Inside a character class, it's always an octal escape,
5284 even if it is a single digit.
5286 (d) An escaped but undefined alphabetic character is taken as a literal,
5287 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5288 escapes.
5290 (e) {0} is now permitted. (The previous item is removed from the compiled
5291 pattern).
5293 5. Changed all the names of code files so that the basic parts are no longer
5294 than 10 characters, and abolished the teeny "globals.c" file.
5296 6. Changed the handling of character classes; they are now done with a 32-byte
5297 bit map always.
5299 7. Added the -d and /D options to pcretest to make it possible to look at the
5300 internals of compilation without having to recompile pcre.
5303 Version 0.95 23-Sep-97
5304 ----------------------
5306 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5307 \x20 at the start of a run of normal characters. These were being treated as
5308 real characters, instead of the source characters being re-checked.
5311 Version 0.94 18-Sep-97
5312 ----------------------
5314 1. The functions are now thread-safe, with the caveat that the global variables
5315 containing pointers to malloc() and free() or alternative functions are the
5316 same for all threads.
5318 2. Get pcre_study() to generate a bitmap of initial characters for non-
5319 anchored patterns when this is possible, and use it if passed to pcre_exec().
5322 Version 0.93 15-Sep-97
5323 ----------------------
5325 1. /(b)|(:+)/ was computing an incorrect first character.
5327 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5328 but not actually doing anything yet.
5330 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5331 as Perl does (e.g. [-az] or [az-]).
5333 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5334 all possible positions.
5336 5. Split up into different modules to avoid including unneeded functions in a
5337 compiled binary. However, compile and exec are still in one module. The "study"
5338 function is split off.
5340 6. The character tables are now in a separate module whose source is generated
5341 by an auxiliary program - but can then be edited by hand if required. There are
5342 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5343 toupper() in the code.
5345 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5346 make them global. Abolish the function for setting them, as the caller can now
5347 set them directly.
5350 Version 0.92 11-Sep-97
5351 ----------------------
5353 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5354 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5356 2. Caseless matching was not working in character classes if the characters in
5357 the pattern were in upper case.
5359 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5361 4. Make PCRE_ANCHORED public and accept as a compile option.
5363 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5364 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5365 pass them.
5367 6. Give an error if bad option bits passed at compile or run time.
5369 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5370 pcretest to cause it to pass that flag.
5372 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5373 options, and the first character, if set.
5375 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5378 Version 0.91 10-Sep-97
5379 ----------------------
5381 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5382 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5384 2. PCRE was looping on encountering an indefinitely repeated back reference to
5385 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5386 Perl does - treats the match as successful.
5388 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5