ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1380 - (show annotations)
Tue Oct 15 16:49:12 2013 UTC (6 years, 1 month ago) by ph10
File size: 254980 byte(s)
Error occurred while calculating annotation data.
Explicit possessive quantifiers now use the new opcodes. Fixed an infelicity 
with EXACT in caseless mode.
1 ChangeLog for PCRE
2 ------------------
4 Version 8.34 xx-xxxx-201x
5 --------------------------
7 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8 executable memory. Patch inspired by Carsten Klein.
10 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11 this macro is never tested and has no effect, because the work to support
12 coverage involves only compiling and linking options and special targets in
13 the Makefile. The comment in config.h implied that defining the macro would
14 enable coverage support, which is totally false. There was also support for
15 setting this macro in the CMake files (my fault, I just copied it from
16 configure). SUPPORT_GCOV has now been removed.
18 3. Make a small performance improvement in strlen16() and strlen32() in
19 pcretest.
21 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23 these cases). There was also one in pcretest.c.
25 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
27 6. In UTF mode, the code for checking whether a group could match an empty
28 string (which is used for indefinitely repeated groups to allow for
29 breaking an infinite loop) was broken when the group contained a repeated
30 negated single-character class with a character that occupied more than one
31 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
32 UTF-8 mode). The effect was undefined: the group might or might not be
33 deemed as matching an empty string, or the program might have crashed.
35 7. The code for checking whether a group could match an empty string was not
36 recognizing that \h, \H, \v, \V, and \R must match a character.
38 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
39 an empty string. If it can, pcretest shows this in its information output.
41 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
42 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
43 matched by pcre_exec() without using JIT:
45 (a) If the rest of the pattern did not match after a maximal run of
46 grapheme clusters, the code for backing up to try with fewer of them
47 did not always back up over a full grapheme when characters that do not
48 have the modifier quality were involved, e.g. Hangul syllables.
50 (b) If the match point in a subject started with modifier character, and
51 there was no match, the code could incorrectly back up beyond the match
52 point, and potentially beyond the first character in the subject,
53 leading to a segfault or an incorrect match result.
55 10. A conditional group with an assertion condition could lead to PCRE
56 recording an incorrect first data item for a match if no other first data
57 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58 first data item, and therefore matched "ca" after "c" instead of at the
59 start.
61 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
62 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
64 12. The source of pcregrep now includes z/OS-specific code so that it can be
65 compiled for z/OS as part of the special z/OS distribution.
67 13. Added the -T and -TM options to pcretest.
69 14. The code in pcre_compile.c for creating the table of named capturing groups
70 has been refactored. Instead of creating the table dynamically during the
71 actual compiling pass, the information is remembered during the pre-compile
72 pass (on the stack unless there are more than 20 named groups, in which
73 case malloc() is used) and the whole table is created before the actual
74 compile happens. This has simplified the code (it is now nearly 150 lines
75 shorter) and prepared the way for better handling of references to groups
76 with duplicate names.
78 15. A back reference to a named subpattern when there is more than one of the
79 same name now checks them in the order in which they appear in the pattern.
80 The first one that is set is used for the reference. Previously only the
81 first one was inspected. This change makes PCRE more compatible with Perl.
83 16. Unicode character properties were updated from Unicode 6.3.0.
85 17. The compile-time code for auto-possessification has been refactored, based
86 on a patch by Zoltan Herczeg. It now happens after instead of during
87 compilation. The code is cleaner, and more cases are handled. The option
88 PCRE_NO_AUTO_POSSESSIFY is added for testing purposes, and the -O and /O
89 options in pcretest are provided to set it.
91 18. The character VT has been added to the set of characters that match \s and
92 are generally treated as white space, following this same change in Perl
93 5.18. There is now no difference between "Perl space" and "POSIX space".
95 19. The code for checking named groups as conditions, either for being set or
96 for being recursed, has been refactored (this is related to 14 and 15
97 above). Processing unduplicated named groups should now be as fast at
98 numerical groups, and processing duplicated groups should be faster than
99 before.
101 20. Two patches to the CMake build system, by Alexander Barkov:
103 (1) Replace the "source" command by "." in CMakeLists.txt because
104 "source" is a bash-ism.
106 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
107 without these the CMake build does not work on Solaris.
109 21. Perl has changed its handling of \8 and \9. If there is no previously
110 encountered capturing group of those numbers, they are treated as the
111 literal characters 8 and 9 instead of a binary zero followed by the
112 literals. PCRE now does the same.
114 22. Following Perl, added \o{} to specify codepoints in octal, making it
115 possible to specify values greater than 0777 and also making them
116 unambiguous.
118 23. Perl now gives an error for missing closing braces after \x{... instead of
119 treating the string as literal. PCRE now does the same.
121 24. RunTest used to grumble if an inappropriate test was selected explicitly,
122 but just skip it when running all tests. This make it awkward to run ranges
123 of tests when one of them was inappropriate. Now it just skips any
124 inappropriate tests, as it always did when running all tests.
126 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
127 character types such as \d or \w, too many callouts were inserted, and the
128 data that they returned was rubbish.
130 26. In UCP mode, \s was not matching two of the characters that Perl matches,
131 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
132 were matched by \h. The code has now been refactored so that the lists of
133 the horizontal and vertical whitespace characters used for \h and \v (which
134 are defined only in one place) are now also used for \s.
136 27. Add JIT support for the 64 bit TileGX architecture.
137 Patch by Jiong Wang (Tilera Corporation).
139 28. Possessive quantifiers for classes (both explicit and automatically
140 generated) now use special opcodes instead of wrapping in ONCE brackets.
142 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
143 (because it's meaningless), this was not happening when PCRE_CASELESS was
144 set. Not wrong, but inefficient.
147 Version 8.33 28-May-2013
148 ------------------------
150 1. Added 'U' to some constants that are compared to unsigned integers, to
151 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
152 variables that are added to signed variables, to ensure the result is
153 signed and can be negated.
155 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
156 CMake config files.
158 3. Revise the creation of config.h.generic so that all boolean macros are
159 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
160 overriding via -D on the command line possible.
162 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
163 to unsigned int is reported to make a quite noticeable speed difference in
164 a specific Windows environment. Testing on Linux did also appear to show
165 some benefit (and it is clearly not harmful). Also fixed the definition of
166 Xop which should be unsigned.
168 5. Related to (4), changing the definition of the intermediate variable cc
169 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
170 improvements.
172 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
173 unnecessary spaces.
175 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
176 and later.
178 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
180 9. Optimizing fast_forward_start_bits in JIT.
182 10. Adding support for callouts in JIT, and fixing some issues revealed
183 during this work. Namely:
185 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
187 (b) Minimum length was not checked before the matching is started.
189 11. The value of capture_last that is passed to callouts was incorrect in some
190 cases when there was a capture on one path that was subsequently abandoned
191 after a backtrack. Also, the capture_last value is now reset after a
192 recursion, since all captures are also reset in this case.
194 12. The interpreter no longer returns the "too many substrings" error in the
195 case when an overflowing capture is in a branch that is subsequently
196 abandoned after a backtrack.
198 13. In the pathological case when an offset vector of size 2 is used, pcretest
199 now prints out the matched string after a yield of 0 or 1.
201 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
202 Only supported by the JIT compiler at the moment.
204 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
206 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
207 offset of the starting point of the matching process, provided the offsets
208 vector is large enough.
210 17. The \A escape now records a lookbehind value of 1, though its execution
211 does not actually inspect the previous character. This is to ensure that,
212 in partial multi-segment matching, at least one character from the old
213 segment is retained when a new segment is processed. Otherwise, if there
214 are no lookbehinds in the pattern, \A might match incorrectly at the start
215 of a new segment.
217 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
219 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
220 gives some modest performance improvement in 8-bit mode.
222 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
223 be expressed in certain programming languages using Universal Character
224 Names.
226 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
227 which points out that "non characters" are not "characters that may not
228 appear in Unicode strings" but rather "characters that are reserved for
229 internal use and have only local meaning".
231 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
232 there was a conditional group that depended on an assertion, if the
233 assertion was false, the callout that immediately followed the alternation
234 in the condition was skipped when pcre_exec() was used for matching.
236 23. Allow an explicit callout to be inserted before an assertion that is the
237 condition for a conditional group, for compatibility with automatic
238 callouts, which always insert a callout at this point.
240 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
241 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
243 25. (*PRUNE) is now supported by the JIT compiler.
245 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
247 27. Fix the case where there are two or more SKIPs with arguments that may be
248 ignored.
250 28. (*SKIP) is now supported by the JIT compiler.
252 29. (*THEN) is now supported by the JIT compiler.
254 30. Update RunTest with additional test selector options.
256 31. The way PCRE handles backtracking verbs has been changed in two ways.
258 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
259 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
260 backtracking. In some cases this makes it more Perl-compatible, but Perl's
261 rather obscure rules do not always do the same thing.
263 (2) Previously, backtracking verbs were confined within assertions. This is
264 no longer the case for positive assertions, except for (*ACCEPT). Again,
265 this sometimes improves Perl compatibility, and sometimes does not.
267 32. A number of tests that were in test 2 because Perl did things differently
268 have been moved to test 1, because either Perl or PCRE has changed, and
269 these tests are now compatible.
271 32. Backtracking control verbs are now handled in the same way in JIT and
272 interpreter.
274 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
275 contained a forward subroutine reference caused a compile error.
277 34. Auto-detect and optimize limited repetitions in JIT.
279 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
280 blocking (*UTF) etc.
282 36. In the interpreter, maximizing pattern repetitions for characters and
283 character types now use tail recursion, which reduces stack usage.
285 37. The value of the max lookbehind was not correctly preserved if a compiled
286 and saved regex was reloaded on a host of different endianness.
288 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
289 of the compiled pattern block, expand the flags field from 16 to 32 bits
290 because it was almost full.
292 39. Try madvise first before posix_madvise.
294 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
295 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
296 original change locked out all matches of empty strings. This has been
297 changed so that one match of an empty string per line is recognized.
298 Subsequent searches on the same line (for colouring or for --only-matching,
299 for example) do not recognize empty strings.
301 41. Applied a user patch to fix a number of spelling mistakes in comments.
303 42. Data lines longer than 65536 caused pcretest to crash.
305 43. Clarified the data type for length and startoffset arguments for pcre_exec
306 and pcre_dfa_exec in the function-specific man pages, where they were
307 explicitly stated to be in bytes, never having been updated. I also added
308 some clarification to the pcreapi man page.
310 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
311 a segmentation fault.
314 Version 8.32 30-November-2012
315 -----------------------------
317 1. Improved JIT compiler optimizations for first character search and single
318 character iterators.
320 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
321 Patch by Daniel Richard G.
323 3. Single character iterator optimizations in the JIT compiler.
325 4. Improved JIT compiler optimizations for character ranges.
327 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
328 Reported by Giuseppe D'Angelo.
330 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
331 of a line, was being set incorrectly in cases where .* appeared inside
332 atomic brackets at the start of a pattern, or where there was a subsequent
333 *PRUNE or *SKIP.
335 7. Improved instruction cache flush for POWER/PowerPC.
336 Patch by Daniel Richard G.
338 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
339 grep:
341 (a) There is now no limit to the number of patterns to be matched.
343 (b) An error is given if a pattern is too long.
345 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
346 are now supported.
348 (d) --exclude-from and --include-from (multiple use) have been added.
350 (e) Exclusions and inclusions now apply to all files and directories, not
351 just to those obtained from scanning a directory recursively.
353 (f) Multiple uses of -f and --file-list are now supported.
355 (g) In a Windows environment, the default for -d has been changed from
356 "read" (the GNU grep default) to "skip", because otherwise the presence
357 of a directory in the file list provokes an error.
359 (h) The documentation has been revised and clarified in places.
361 9. Improve the matching speed of capturing brackets.
363 10. Changed the meaning of \X so that it now matches a Unicode extended
364 grapheme cluster.
366 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
367 out POSIX threads when JIT support is configured.
369 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
371 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
372 structure could go wrong in environments where size_t is not the same size
373 as int.
375 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
376 set.
378 15. The EBCDIC support had decayed; later updates to the code had included
379 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
380 general tidy up of EBCDIC-related issues, and the documentation was also
381 not quite right. There is now a test that can be run on ASCII systems to
382 check some of the EBCDIC-related things (but is it not a full test).
384 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
385 in a small tidy to the code.
387 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
389 18. If the --only-matching (-o) option in pcregrep is specified multiple
390 times, each one causes appropriate output. For example, -o1 -o2 outputs the
391 substrings matched by the 1st and 2nd capturing parentheses. A separating
392 string can be specified by --om-separator (default empty).
394 19. Improving the first n character searches.
396 20. Turn case lists for horizontal and vertical white space into macros so that
397 they are defined only once.
399 21. This set of changes together give more compatible Unicode case-folding
400 behaviour for characters that have more than one other case when UCP
401 support is available.
403 (a) The Unicode property table now has offsets into a new table of sets of
404 three or more characters that are case-equivalent. The MultiStage2.py
405 script that generates these tables (the pcre_ucd.c file) now scans
406 CaseFolding.txt instead of UnicodeData.txt for character case
407 information.
409 (b) The code for adding characters or ranges of characters to a character
410 class has been abstracted into a generalized function that also handles
411 case-independence. In UTF-mode with UCP support, this uses the new data
412 to handle characters with more than one other case.
414 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
415 whose other case is greater than 256 are now correctly matched
416 caselessly. Previously, the high codepoint matched the low one, but not
417 vice versa.
419 (d) The processing of \h, \H, \v, and \ in character classes now makes use
420 of the new class addition function, using character lists defined as
421 macros alongside the case definitions of 20 above.
423 (e) Caseless back references now work with characters that have more than
424 one other case.
426 (f) General caseless matching of characters with more than one other case
427 is supported.
429 22. Unicode character properties were updated from Unicode 6.2.0
431 23. Improved CMake support under Windows. Patch by Daniel Richard G.
433 24. Add support for 32-bit character strings, and UTF-32
435 25. Major JIT compiler update (code refactoring and bugfixing).
436 Experimental Sparc 32 support is added.
438 26. Applied a modified version of Daniel Richard G's patch to create
439 pcre.h.generic and config.h.generic by "make" instead of in the
440 PrepareRelease script.
442 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
443 pcre_compile.c when checking for a zero character.
445 28. Introducing a native interface for JIT. Through this interface, the compiled
446 machine code can be directly executed. The purpose of this interface is to
447 provide fast pattern matching, so several sanity checks are not performed.
448 However, feature tests are still performed. The new interface provides
449 1.4x speedup compared to the old one.
451 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
452 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
453 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
455 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
456 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
457 the "old" RFC 2279). Instead, it ended up passing a negative length to
458 pcre_exec().
460 31. Add support for GCC's visibility feature to hide internal functions.
462 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
463 "unknown -C option" after outputting 0 or 1.
465 33. There is now support for generating a code coverage report for the test
466 suite in environments where gcc is the compiler and lcov is installed. This
467 is mainly for the benefit of the developers.
469 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
470 unaddressable using valgrind annotations, allowing valgrind to detect
471 invalid memory accesses. This is mainly for the benefit of the developers.
473 25. (*UTF) can now be used to start a pattern in any of the three libraries.
475 26. Give configure error if --enable-cpp but no C++ compiler found.
478 Version 8.31 06-July-2012
479 -------------------------
481 1. Fixing a wrong JIT test case and some compiler warnings.
483 2. Removed a bashism from the RunTest script.
485 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
486 to unsigned type, result still unsigned" that was given by an MS compiler
487 on encountering the code "-sizeof(xxx)".
489 4. Partial matching support is added to the JIT compiler.
491 5. Fixed several bugs concerned with partial matching of items that consist
492 of more than one character:
494 (a) /^(..)\1/ did not partially match "aba" because checking references was
495 done on an "all or nothing" basis. This also applied to repeated
496 references.
498 (b) \R did not give a hard partial match if \r was found at the end of the
499 subject.
501 (c) \X did not give a hard partial match after matching one or more
502 characters at the end of the subject.
504 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
505 a partial match for the string "\r".
507 (e) When newline was set to CRLF, the metacharacter "." did not recognize
508 a partial match for a CR character at the end of the subject string.
510 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
511 running pcretest, the text "(JIT)" added to the output whenever JIT is
512 actually used to run the match.
514 7. Individual JIT compile options can be set in pcretest by following -s+[+]
515 or /S+[+] with a digit between 1 and 7.
517 8. OP_NOT now supports any UTF character not just single-byte ones.
519 9. (*MARK) control verb is now supported by the JIT compiler.
521 10. The command "./RunTest list" lists the available tests without actually
522 running any of them. (Because I keep forgetting what they all are.)
526 12. Applied a (slightly modified) user-supplied patch that improves performance
527 when the heap is used for recursion (compiled with --disable-stack-for-
528 recursion). Instead of malloc and free for each heap frame each time a
529 logical recursion happens, frames are retained on a chain and re-used where
530 possible. This sometimes gives as much as 30% improvement.
532 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
533 call.
535 14. As documented, (*COMMIT) is now confined to within a positive assertion.
537 15. It is now possible to link pcretest with libedit as an alternative to
538 libreadline.
540 16. (*COMMIT) control verb is now supported by the JIT compiler.
542 17. The Unicode data tables have been updated to Unicode 6.1.0.
544 18. Added --file-list option to pcregrep.
546 19. Added binary file support to pcregrep, including the -a, --binary-files,
547 -I, and --text options.
549 20. The madvise function is renamed for posix_madvise for QNX compatibility
550 reasons. Fixed by Giuseppe D'Angelo.
552 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
554 22. Changed the option for creating symbolic links for 16-bit man pages from
555 -s to -sf so that re-installing does not cause issues.
557 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
559 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
560 matches in certain environments (the workspace was not being correctly
561 retained). Also added to pcre_dfa_exec() a simple plausibility check on
562 some of the workspace data at the beginning of a restart.
564 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
565 was not doing so when it should - probably a typo introduced by SVN 528
566 (change 8.10/14).
568 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
569 \w+ when the character tables indicated that \x{c4} was a word character.
570 There were several related cases, all because the tests for doing a table
571 lookup were testing for characters less than 127 instead of 255.
573 27. If a pattern contains capturing parentheses that are not used in a match,
574 their slots in the ovector are set to -1. For those that are higher than
575 any matched groups, this happens at the end of processing. In the case when
576 there were back references that the ovector was too small to contain
577 (causing temporary malloc'd memory to be used during matching), and the
578 highest capturing number was not used, memory off the end of the ovector
579 was incorrectly being set to -1. (It was using the size of the temporary
580 memory instead of the true size.)
582 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
583 ovector size, it uses memory at the end of the block that it has got.
585 29. Check for an overlong MARK name and give an error at compile time. The
586 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
588 30. JIT compiler update.
590 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
591 Rill for the patch.
593 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
595 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
597 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
598 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
599 was enabled.
601 35. Improve JIT code generation for greedy plus quantifier.
603 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
604 1 to "aa" instead of to an empty string. The bug affected repeated groups
605 that could potentially match an empty string.
607 37. Optimizing single character iterators in JIT.
609 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
610 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
611 codepoints that are too big for the mode are faulted, and in a UTF mode,
612 disallowed codepoints are also faulted.
614 39. If PCRE was compiled with UTF support, in three places in the DFA
615 matcher there was code that should only have been obeyed in UTF mode, but
616 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
617 processing when bytes with values greater than 127 were present. In 16-bit
618 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
619 both cases the values are those that cannot be the first data item in a UTF
620 character. The three items that might have provoked this were recursions,
621 possessively repeated groups, and atomic groups.
623 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
624 and pcregrep, because some OS require shared objects to be explicitly
625 passed to ld, causing the link step to fail if they are not.
627 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
628 mode, patterns that started with \h* or \R* might be incorrectly matched.
631 Version 8.30 04-February-2012
632 -----------------------------
634 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
635 name is defined in ctype.h.
637 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
638 only in quite long subpatterns.
640 3. Removed the function pcre_info(), which has been obsolete and deprecated
641 since it was replaced by pcre_fullinfo() in February 2000.
643 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
644 match a (*MARK), and the match failed at the start of the subject, a
645 reference to memory before the start of the subject could occur. This bug
646 was introduced by fix 17 of release 8.21.
648 5. A reference to an unset group with zero minimum repetition was giving
649 totally wrong answers (in non-JavaScript-compatibility mode). For example,
650 /(another)?(\1?)test/ matched against "hello world test". This bug was
651 introduced in release 8.13.
653 6. Add support for 16-bit character strings (a large amount of work involving
654 many changes and refactorings).
656 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
657 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
658 from a file.
660 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
661 rounding is not applied in this particular case).
663 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
664 if they appear, or are escaped, in patterns.
666 10. Get rid of a number of -Wunused-but-set-variable warnings.
668 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
669 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
670 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
671 also returns the mark "x". This bug applied to capturing parentheses,
672 non-capturing parentheses, and atomic parentheses. It also applied to some
673 assertions.
675 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
676 information out of configure.ac instead of relying on pcre.h.generic, which
677 is not stored in the repository.
679 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
680 -lreadline.
682 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
684 15. Applied Graycode's patch to put the top-level frame on the stack rather
685 than the heap when not using the stack for recursion. This gives a
686 performance improvement in many cases when recursion is not deep.
688 16. Experimental code added to "pcretest -C" to output the stack frame size.
691 Version 8.21 12-Dec-2011
692 ------------------------
694 1. Updating the JIT compiler.
696 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
697 are added as well.
699 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
700 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
701 calling _pcre_jit_exec. Some extra comments are added.
703 4. (*MARK) settings inside atomic groups that do not contain any capturing
704 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
705 was introduced by change 18 for 8.20.
707 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
708 ECMA-262 standard.
710 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
711 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
712 This bug was probably introduced by change 9 of 8.13.
714 7. While fixing 6 above, I noticed that a number of other items were being
715 incorrectly rejected as "not fixed length". This arose partly because newer
716 opcodes had not been added to the fixed-length checking code. I have (a)
717 corrected the bug and added tests for these items, and (b) arranged for an
718 error to occur if an unknown opcode is encountered while checking for fixed
719 length instead of just assuming "not fixed length". The items that were
720 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
721 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
722 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
724 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
725 being incorrectly compiled and would have given unpredicatble results.
727 9. A possessively repeated subpattern with minimum repeat count greater than
728 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
729 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
730 the first (A) could occur when it should not.
732 10. Add a cast and remove a redundant test from the code.
734 11. JIT should use pcre_malloc/pcre_free for allocation.
736 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
737 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
738 is anything other than /usr, -L is still shown).
740 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
742 14. Perl does not support \N without a following name in a [] class; PCRE now
743 also gives an error.
745 15. If a forward reference was repeated with an upper limit of around 2000,
746 it caused the error "internal error: overran compiling workspace". The
747 maximum number of forward references (including repeats) was limited by the
748 internal workspace, and dependent on the LINK_SIZE. The code has been
749 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
750 the default depends on LINK_SIZE. There is a new upper limit (for safety)
751 of around 200,000 forward references. While doing this, I also speeded up
752 the filling in of repeated forward references.
754 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
755 incorrectly expecting the subject to contain another "a" after the start.
757 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
758 in the match, the SKIP should be ignored. This was not happening; instead
759 the SKIP was being treated as NOMATCH. For patterns such as
760 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
761 tested.
763 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
764 now much more compatible with Perl, in particular in cases where the result
765 is a non-match for a non-anchored pattern. For example, if
766 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
767 "m", where previously it did not return a name. A side effect of this
768 change is that for partial matches, the last encountered mark name is
769 returned, as for non matches. A number of tests that were previously not
770 Perl-compatible have been moved into the Perl-compatible test files. The
771 refactoring has had the pleasing side effect of removing one argument from
772 the match() function, thus reducing its stack requirements.
774 19. If the /S+ option was used in pcretest to study a pattern using JIT,
775 subsequent uses of /S (without +) incorrectly behaved like /S+.
777 21. Retrieve executable code size support for the JIT compiler and fixing
778 some warnings.
780 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
781 not work when the shorter character appeared right at the end of the
782 subject string.
784 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
785 systems.
787 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
788 output it when the /M option is used in pcretest.
790 25. The CheckMan script was not being included in the distribution. Also, added
791 an explicit "perl" to run Perl scripts from the PrepareRelease script
792 because this is reportedly needed in Windows.
794 26. If study data was being save in a file and studying had not found a set of
795 "starts with" bytes for the pattern, the data written to the file (though
796 never used) was taken from uninitialized memory and so caused valgrind to
797 complain.
799 27. Updated RunTest.bat as provided by Sheri Pierce.
801 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
803 29. Computation of memory usage for the table of capturing group names was
804 giving an unnecessarily large value.
807 Version 8.20 21-Oct-2011
808 ------------------------
810 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
811 a POSIX class. After further experiments with Perl, which convinced me that
812 Perl has bugs and confusions, a closing square bracket is no longer allowed
813 in a POSIX name. This bug also affected patterns with classes that started
814 with full stops.
816 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
817 captured substring, but while checking the failing first alternative,
818 substring 1 is temporarily captured. If the output vector supplied to
819 pcre_exec() was not big enough for this capture, the yield of the function
820 was still zero ("insufficient space for captured substrings"). This cannot
821 be totally fixed without adding another stack variable, which seems a lot
822 of expense for a edge case. However, I have improved the situation in cases
823 such as /(a)(b)x|abc/ matched against "abc", where the return code
824 indicates that fewer than the maximum number of slots in the ovector have
825 been set.
827 3. Related to (2) above: when there are more back references in a pattern than
828 slots in the output vector, pcre_exec() uses temporary memory during
829 matching, and copies in the captures as far as possible afterwards. It was
830 using the entire output vector, but this conflicts with the specification
831 that only 2/3 is used for passing back captured substrings. Now it uses
832 only the first 2/3, for compatibility. This is, of course, another edge
833 case.
835 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
836 main code base, and can be used by building with --enable-jit. When this is
837 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
838 runtime --no-jit option is given.
840 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
841 ovector, the return from the function was zero, implying that there were
842 other matches that did not fit. The correct "exactly full" value is now
843 returned.
845 6. If a subpattern that was called recursively or as a subroutine contained
846 (*PRUNE) or any other control that caused it to give a non-standard return,
847 invalid errors such as "Error -26 (nested recursion at the same subject
848 position)" or even infinite loops could occur.
850 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
851 computing the minimum length on reaching *ACCEPT, and so ended up with the
852 wrong value of 1 rather than 0. Further investigation indicates that
853 computing a minimum subject length in the presence of *ACCEPT is difficult
854 (think back references, subroutine calls), and so I have changed the code
855 so that no minimum is registered for a pattern that contains *ACCEPT.
857 8. If (*THEN) was present in the first (true) branch of a conditional group,
858 it was not handled as intended. [But see 16 below.]
860 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
861 Sheri Pierce.
863 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
864 the first byte in a match must be "a".
866 11. Change 17 for 8.13 increased the recursion depth for patterns like
867 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
868 pattern contains any instances of (*THEN). If it does not, the old
869 optimizations are restored. It would be nice to do this on a per-group
870 basis, but at the moment that is not feasible.
872 12. In some environments, the output of pcretest -C is CRLF terminated. This
873 broke RunTest's code that checks for the link size. A single white space
874 character after the value is now allowed for.
876 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
877 For "fr", it uses the Windows-specific input and output files.
879 14. If (*THEN) appeared in a group that was called recursively or as a
880 subroutine, it did not work as intended. [But see next item.]
882 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
883 pattern fragments (but not containing any | characters). If A and B are
884 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
885 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
886 D. In other words, Perl considers parentheses that do not contain any |
887 characters to be part of a surrounding alternative, whereas PCRE was
888 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
889 differently. PCRE now behaves in the same way as Perl, except in the case
890 of subroutine/recursion calls such as (?1) which have in any case always
891 been different (but PCRE had them first :-).
893 16. Related to 15 above: Perl does not treat the | in a conditional group as
894 creating alternatives. Such a group is treated in the same way as an
895 ordinary group without any | characters when processing (*THEN). PCRE has
896 been changed to match Perl's behaviour.
898 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
899 RunGrepTest script failed.
901 18. Change 22 for version 13 caused atomic groups to use more stack. This is
902 inevitable for groups that contain captures, but it can lead to a lot of
903 stack use in large patterns. The old behaviour has been restored for atomic
904 groups that do not contain any capturing parentheses.
906 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
907 suppress the check for a minimum subject length at run time. (If it was
908 given to pcre_exec() or pcre_dfa_exec() it did work.)
910 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
911 fail to work when decoding hex characters in data strings in EBCDIC
912 environments.
914 21. It appears that in at least one Mac OS environment, the isxdigit() function
915 is implemented as a macro that evaluates to its argument more than once,
916 contravening the C 90 Standard (I haven't checked a later standard). There
917 was an instance in pcretest which caused it to go wrong when processing
918 \x{...} escapes in subject strings. The has been rewritten to avoid using
919 things like p++ in the argument of isxdigit().
922 Version 8.13 16-Aug-2011
923 ------------------------
925 1. The Unicode data tables have been updated to Unicode 6.0.0.
927 2. Two minor typos in pcre_internal.h have been fixed.
929 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
930 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
931 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
933 4. There were a number of related bugs in the code for matching backrefences
934 caselessly in UTF-8 mode when codes for the characters concerned were
935 different numbers of bytes. For example, U+023A and U+2C65 are an upper
936 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
937 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
938 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
939 2-byte code at the end of the subject (it thought there wasn't enough data
940 left).
942 5. Comprehensive information about what went wrong is now returned by
943 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
944 as the output vector has at least 2 elements. The offset of the start of
945 the failing character and a reason code are placed in the vector.
947 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
948 now returned is for the first byte of the failing character, instead of the
949 last byte inspected. This is an incompatible change, but I hope it is small
950 enough not to be a problem. It makes the returned offset consistent with
951 pcre_exec() and pcre_dfa_exec().
953 7. pcretest now gives a text phrase as well as the error number when
954 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
955 failure, the offset and reason code are output.
957 8. When \R was used with a maximizing quantifier it failed to skip backwards
958 over a \r\n pair if the subsequent match failed. Instead, it just skipped
959 back over a single character (\n). This seems wrong (because it treated the
960 two characters as a single entity when going forwards), conflicts with the
961 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
962 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
963 has been changed.
965 9. Some internal refactoring has changed the processing so that the handling
966 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
967 time (the PCRE_DOTALL option was changed this way some time ago: version
968 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
969 which was always a bit of a fudge. It also means that there is one less
970 argument for the match() function, which reduces its stack requirements
971 slightly. This change also fixes an incompatibility with Perl: the pattern
972 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
974 10. More internal refactoring has drastically reduced the number of recursive
975 calls to match() for possessively repeated groups such as (abc)++ when
976 using pcre_exec().
978 11. While implementing 10, a number of bugs in the handling of groups were
979 discovered and fixed:
981 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
982 (a|)*(?1) gave a compile-time internal error.
983 ((a|)+)+ did not notice that the outer group could match an empty string.
984 (^a|^)+ was not marked as anchored.
985 (.*a|.*)+ was not marked as matching at start or after a newline.
987 12. Yet more internal refactoring has removed another argument from the match()
988 function. Special calls to this function are now indicated by setting a
989 value in a variable in the "match data" data block.
991 13. Be more explicit in pcre_study() instead of relying on "default" for
992 opcodes that mean there is no starting character; this means that when new
993 ones are added and accidentally left out of pcre_study(), testing should
994 pick them up.
996 14. The -s option of pcretest has been documented for ages as being an old
997 synonym of -m (show memory usage). I have changed it to mean "force study
998 for every regex", that is, assume /S for every regex. This is similar to -i
999 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1000 using it. It makes it easier to run collections of tests with and without
1001 study enabled, and thereby test pcre_study() more easily. All the standard
1002 tests are now run with and without -s (but some patterns can be marked as
1003 "never study" - see 20 below).
1005 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1006 restoration of the capturing data to the outer values was not happening
1007 correctly.
1009 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1010 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1011 pattern had matched an empty string, and so incorrectly returned a no
1012 match.
1014 17. There was optimizing code for the last branch of non-capturing parentheses,
1015 and also for the obeyed branch of a conditional subexpression, which used
1016 tail recursion to cut down on stack usage. Unfortunately, now that there is
1017 the possibility of (*THEN) occurring in these branches, tail recursion is
1018 no longer possible because the return has to be checked for (*THEN). These
1019 two optimizations have therefore been removed. [But see 8.20/11 above.]
1021 18. If a pattern containing \R was studied, it was assumed that \R always
1022 matched two bytes, thus causing the minimum subject length to be
1023 incorrectly computed because \R can also match just one byte.
1025 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1026 was incorrectly computed.
1028 20. If /S is present twice on a test pattern in pcretest input, it now
1029 *disables* studying, thereby overriding the use of -s on the command line
1030 (see 14 above). This is necessary for one or two tests to keep the output
1031 identical in both cases.
1033 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1034 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1036 22. When an atomic group that contained a capturing parenthesis was
1037 successfully matched, but the branch in which it appeared failed, the
1038 capturing was not being forgotten if a higher numbered group was later
1039 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1040 group 1 to "a", when in fact it should be unset. This applied to multi-
1041 branched capturing and non-capturing groups, repeated or not, and also to
1042 positive assertions (capturing in negative assertions does not happen
1043 in PCRE) and also to nested atomic groups.
1045 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1046 subject after a captured substring, to make it easier to tell which of a
1047 number of identical substrings has been captured.
1049 24. The way atomic groups are processed by pcre_exec() has been changed so that
1050 if they are repeated, backtracking one repetition now resets captured
1051 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1052 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1053 "aaa". Previously, it would have been "a". As part of this code
1054 refactoring, the way recursive calls are handled has also been changed.
1056 25. If an assertion condition captured any substrings, they were not passed
1057 back unless some other capturing happened later. For example, if
1058 (?(?=(a))a) was matched against "a", no capturing was returned.
1060 26. When studying a pattern that contained subroutine calls or assertions,
1061 the code for finding the minimum length of a possible match was handling
1062 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1063 group 1 called group 2 while simultaneously a separate group 2 called group
1064 1). A stack overflow occurred in this case. I have fixed this by limiting
1065 the recursion depth to 10.
1067 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1068 Fortmann. This supports explicit test numbers on the command line, and has
1069 argument validation and error reporting.
1071 28. An instance of \X with an unlimited repeat could fail if at any point the
1072 first character it looked at was a mark character.
1074 29. Some minor code refactoring concerning Unicode properties and scripts
1075 should reduce the stack requirement of match() slightly.
1077 30. Added the '=' option to pcretest to check the setting of unused capturing
1078 slots at the end of the pattern, which are documented as being -1, but are
1079 not included in the return count.
1081 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1082 compiled something random. Now it gives a compile-time error (as does
1083 Perl).
1085 32. A *MARK encountered during the processing of a positive assertion is now
1086 recorded and passed back (compatible with Perl).
1088 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1089 had alternative anchored branches, the search for a second match in a line
1090 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1091 matched the line "0102" twice. The same bug affected patterns that started
1092 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1093 twice.
1095 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1096 does, and because of capturing effects, quantifying parenthesized
1097 assertions may at times be useful. Quantifiers are now allowed for
1098 parenthesized assertions.
1100 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1102 36. \g was being checked for fancy things in a character class, when it should
1103 just be a literal "g".
1105 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1106 appearance of a nested POSIX class supersedes an apparent external class.
1107 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1108 unescaped square brackets may also appear as part of class names. For
1109 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1110 more like Perl. (But see 8.20/1 above.)
1112 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1113 was because it thought it was \N{name}, which is not supported).
1115 39. Add minix to OS list not supporting the -S option in pcretest.
1117 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1118 cannot analyze patterns in sufficient detail to catch mutual recursions
1119 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1120 subgroup is called recursively as a subpattern for a second time at the
1121 same position in the subject string. In previous releases this might have
1122 been caught by the recursion limit, or it might have run out of stack.
1124 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1125 happen only once. PCRE was, however incorrectly giving a compile time error
1126 "recursive call could loop indefinitely" because it cannot analyze the
1127 pattern in sufficient detail. The compile time test no longer happens when
1128 PCRE is compiling a conditional subpattern, but actual runaway loops are
1129 now caught at runtime (see 40 above).
1131 42. It seems that Perl allows any characters other than a closing parenthesis
1132 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1133 has been changed to be the same.
1135 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1136 as not to get warnings when autogen.sh is called. Also changed
1137 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1139 44. To help people who use pcregrep to scan files containing exceedingly long
1140 lines, the following changes have been made:
1142 (a) The default value of the buffer size parameter has been increased from
1143 8K to 20K. (The actual buffer used is three times this size.)
1145 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1146 PCRE is built.
1148 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1149 to be set at run time.
1151 (d) Numerical values in pcregrep options can be followed by K or M, for
1152 example --buffer-size=50K.
1154 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1155 given and the return code is set to 2.
1157 45. Add a pointer to the latest mark to the callout data block.
1159 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1160 partial match of an empty string instead of no match. This was specific to
1161 the use of ".".
1163 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1164 complete match instead of a partial match. This bug was dependent on both
1165 the PCRE_UTF8 and PCRE_DOTALL options being set.
1167 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1168 starting byte set, because \b was not being ignored.
1171 Version 8.12 15-Jan-2011
1172 ------------------------
1174 1. Fixed some typos in the markup of the man pages, and wrote a script that
1175 checks for such things as part of the documentation building process.
1177 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1178 --match-limit and --recursion-limit options (added for 8.11). In
1179 particular, this made one of the standard tests fail. (The integer value
1180 went into the wrong half of a long int.)
1182 3. If the --colour option was given to pcregrep with -v (invert match), it
1183 did strange things, either producing crazy output, or crashing. It should,
1184 of course, ignore a request for colour when reporting lines that do not
1185 match.
1187 4. Another pcregrep bug caused similar problems if --colour was specified with
1188 -M (multiline) and the pattern match finished with a line ending.
1190 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1191 matched in multiline mode, the following line was shown as part of the
1192 match. This seems wrong, so I have changed it.
1194 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1195 the check for further matches in the same line (so they could be coloured)
1196 to overrun the end of the current line. If another match was found, it was
1197 incorrectly shown (and then shown again when found in the next line).
1199 7. If pcregrep was compiled under Windows, there was a reference to the
1200 function pcregrep_exit() before it was defined. I am assuming this was
1201 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1202 reported by a user. I've moved the definition above the reference.
1205 Version 8.11 10-Dec-2010
1206 ------------------------
1208 1. (*THEN) was not working properly if there were untried alternatives prior
1209 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1210 backtracked to try for "b" instead of moving to the next alternative branch
1211 at the same level (in this case, to look for "c"). The Perl documentation
1212 is clear that when (*THEN) is backtracked onto, it goes to the "next
1213 alternative in the innermost enclosing group".
1215 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1216 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1217 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1218 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1219 (*THEN).
1221 3. If \s appeared in a character class, it removed the VT character from
1222 the class, even if it had been included by some previous item, for example
1223 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1224 of \s, but is part of the POSIX "space" class.)
1226 4. A partial match never returns an empty string (because you can always
1227 match an empty string at the end of the subject); however the checking for
1228 an empty string was starting at the "start of match" point. This has been
1229 changed to the "earliest inspected character" point, because the returned
1230 data for a partial match starts at this character. This means that, for
1231 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1232 (previously it gave "no match").
1234 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1235 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1236 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1237 has an implication that the given string is incomplete (because a partial
1238 match is preferred over a full match). For this reason, these items now
1239 give a partial match in this situation. [Aside: previously, the one case
1240 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1241 match rather than a full match, which was wrong by the old rules, but is
1242 now correct.]
1244 6. There was a bug in the handling of #-introduced comments, recognized when
1245 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1246 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1247 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1248 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1249 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1250 places in pcre_compile().
1252 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1253 comments when looking ahead for named forward references to subpatterns,
1254 the only newline sequence it recognized was NL. It now handles newlines
1255 according to the set newline convention.
1257 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1258 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1259 cater for a lack of strerror(). These oversights have been fixed.
1261 9. Added --match-limit and --recursion-limit to pcregrep.
1263 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1265 11. When the -o option was used, pcregrep was setting a return code of 1, even
1266 when matches were found, and --line-buffered was not being honoured.
1268 12. Added an optional parentheses number to the -o and --only-matching options
1269 of pcregrep.
1271 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1272 can match an empty string. The code to do it in pcretest and pcredemo
1273 needed fixing:
1275 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1276 only one byte after an empty string match just before CRLF (this case
1277 just got forgotten; "any" and "anycrlf" were OK).
1279 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1280 mode when an empty string match preceded an ASCII character followed by
1281 a non-ASCII character. (The code for advancing by one character rather
1282 than one byte was nonsense.)
1284 (c) The pcredemo.c sample program did not have any code at all to handle
1285 the cases when CRLF is a valid newline sequence.
1287 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1288 as a starting offset was within the subject string. There is now a new
1289 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1290 negative or greater than the length of the string. In order to test this,
1291 pcretest is extended to allow the setting of negative starting offsets.
1293 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1294 starting offset points to the beginning of a UTF-8 character was
1295 unnecessarily clumsy. I tidied it up.
1297 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1298 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1300 17. Nobody had reported that the --include_dir option, which was added in
1301 release 7.7 should have been called --include-dir (hyphen, not underscore)
1302 for compatibility with GNU grep. I have changed it to --include-dir, but
1303 left --include_dir as an undocumented synonym, and the same for
1304 --exclude-dir, though that is not available in GNU grep, at least as of
1305 release 2.5.4.
1307 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1308 characters from a string of bytes) have been redefined so as not to use
1309 loops, in order to improve performance in some environments. At the same
1310 time, I abstracted some of the common code into auxiliary macros to save
1311 repetition (this should not affect the compiled code).
1313 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1314 compile-time error is now given if \c is not followed by an ASCII
1315 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1316 different, and any byte value is allowed.)
1318 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1319 START_OPTIMIZE option, which is now allowed at compile time - but just
1320 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1321 to pcregrep and other applications that have no direct access to PCRE
1322 options. The new /Y option in pcretest sets this option when calling
1323 pcre_compile().
1325 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1326 back references. Groups containing recursive back references were forced to
1327 be atomic by that change, but in the case of named groups, the amount of
1328 memory required was incorrectly computed, leading to "Failed: internal
1329 error: code overflow". This has been fixed.
1331 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1332 pcretest.c, to avoid build problems in some Borland environments.
1335 Version 8.10 25-Jun-2010
1336 ------------------------
1338 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1339 THEN.
1341 2. (*ACCEPT) was not working when inside an atomic group.
1343 3. Inside a character class, \B is treated as a literal by default, but
1344 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1345 causes the error). The code is unchanged, but I tidied the documentation.
1347 4. Inside a character class, PCRE always treated \R and \X as literals,
1348 whereas Perl faults them if its -w option is set. I have changed PCRE so
1349 that it faults them when PCRE_EXTRA is set.
1351 5. Added support for \N, which always matches any character other than
1352 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1354 6. When compiling pcregrep with newer versions of gcc which may have
1355 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1356 declared with attribute warn_unused_result" were given. Just casting the
1357 result to (void) does not stop the warnings; a more elaborate fudge is
1358 needed. I've used a macro to implement this.
1360 7. Minor change to pcretest.c to avoid a compiler warning.
1362 8. Added four artifical Unicode properties to help with an option to make
1363 \s etc use properties (see next item). The new properties are: Xan
1364 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1366 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1367 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1368 this option. Modified pcretest to add /W to test this facility. Added
1369 REG_UCP to make it available via the POSIX interface.
1371 10. Added --line-buffered to pcregrep.
1373 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1374 studied, and the match started with a letter with a code point greater than
1375 127 whose first byte was different to the first byte of the other case of
1376 the letter, the other case of this starting letter was not recognized
1377 (#976).
1379 12. If a pattern that was studied started with a repeated Unicode property
1380 test, for example, \p{Nd}+, there was the theoretical possibility of
1381 setting up an incorrect bitmap of starting bytes, but fortunately it could
1382 not have actually happened in practice until change 8 above was made (it
1383 added property types that matched character-matching opcodes).
1385 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1386 possible starting bytes for non-anchored patterns.
1388 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1389 \R, and also a number of cases that involve Unicode properties, both
1390 explicit and implicit when PCRE_UCP is set.
1392 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1393 input, it could crash or give wrong results if characters with values
1394 greater than 0xc0 were present in the subject string. (Detail: it assumed
1395 UTF-8 input when processing these items.)
1397 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1398 size_t is 64-bit (#991).
1400 17. Added a check for running out of memory when PCRE is compiled with
1401 --disable-stack-for-recursion (#990).
1403 18. If the last data line in a file for pcretest does not have a newline on
1404 the end, a newline was missing in the output.
1406 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1407 less than 128) in its various bitmaps. However, there is a facility for
1408 generating tables according to the current locale when PCRE is compiled. It
1409 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1410 characters, are recognized by isspace() and therefore were getting set in
1411 these tables, and indeed these tables seem to approximate to ISO 8859. This
1412 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1413 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1414 which of course cannot start UTF-8 characters. I have changed the code so
1415 that only real ASCII characters (less than 128) and the correct starting
1416 bytes for UTF-8 encodings are set for characters greater than 127 when in
1417 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1418 altogether.)
1420 20. Added the /T option to pcretest so as to be able to run tests with non-
1421 standard character tables, thus making it possible to include the tests
1422 used for 19 above in the standard set of tests.
1424 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1425 reference to a subpattern the other side of a comment that contains an
1426 opening parenthesis caused either an internal compiling error, or a
1427 reference to the wrong subpattern.
1430 Version 8.02 19-Mar-2010
1431 ------------------------
1433 1. The Unicode data tables have been updated to Unicode 5.2.0.
1435 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1436 configured.
1438 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1439 original author of that file, following a query about its status.
1441 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1442 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1444 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1445 quantifier applied to a forward-referencing subroutine call, could compile
1446 incorrect code or give the error "internal error: previously-checked
1447 referenced subpattern not found".
1449 6. Both MS Visual Studio and Symbian OS have problems with initializing
1450 variables to point to external functions. For these systems, therefore,
1451 pcre_malloc etc. are now initialized to local functions that call the
1452 relevant global functions.
1454 7. There were two entries missing in the vectors called coptable and poptable
1455 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1456 I've fixed the data, and added a kludgy way of testing at compile time that
1457 the lengths are correct (equal to the number of opcodes).
1459 8. Following on from 7, I added a similar kludge to check the length of the
1460 eint vector in pcreposix.c.
1462 9. Error texts for pcre_compile() are held as one long string to avoid too
1463 much relocation at load time. To find a text, the string is searched,
1464 counting zeros. There was no check for running off the end of the string,
1465 which could happen if a new error number was added without updating the
1466 string.
1468 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1470 11. \K was not working if it appeared in an atomic group or in a group that
1471 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1472 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1473 the assertion is positive, but not if it is negative.
1475 12. Change 11 fortuitously reduced the size of the stack frame used in the
1476 "match()" function of pcre_exec.c by one pointer. Forthcoming
1477 implementation of support for (*MARK) will need an extra pointer on the
1478 stack; I have reserved it now, so that the stack frame size does not
1479 decrease.
1481 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1482 item in branch that calls a recursion is a subroutine call - as in the
1483 second branch in the above example - was incorrectly given the compile-
1484 time error "recursive call could loop indefinitely" because pcre_compile()
1485 was not correctly checking the subroutine for matching a non-empty string.
1487 14. The checks for overrunning compiling workspace could trigger after an
1488 overrun had occurred. This is a "should never occur" error, but it can be
1489 triggered by pathological patterns such as hundreds of nested parentheses.
1490 The checks now trigger 100 bytes before the end of the workspace.
1492 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1495 Version 8.01 19-Jan-2010
1496 ------------------------
1498 1. If a pattern contained a conditional subpattern with only one branch (in
1499 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1500 computed the wrong minimum data length (which is of course zero for such
1501 subpatterns). This could cause incorrect "no match" results.
1503 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1504 the pattern is reset in the first branch, pcre_compile() failed with
1505 "internal error: code overflow at offset...". This happened only when
1506 the reset was to the original external option setting. (An optimization
1507 abstracts leading options settings into an external setting, which was the
1508 cause of this.)
1510 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1511 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1512 assertion pattern did not match (meaning that the assertion was true), it
1513 was incorrectly treated as false if the SKIP had been reached during the
1514 matching. This also applied to assertions used as conditions.
1516 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1517 assertion subpattern, including such a pattern used as a condition,
1518 unpredictable results occurred, instead of the error return
1521 5. The C++ GlobalReplace function was not working like Perl for the special
1522 situation when an empty string is matched. It now does the fancy magic
1523 stuff that is necessary.
1525 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1526 removed. (These were left over from very, very early versions of PCRE.)
1528 7. Some cosmetic changes to the code to make life easier when compiling it
1529 as part of something else:
1531 (a) Change DEBUG to PCRE_DEBUG.
1533 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1534 called "current" as "current_branch", to prevent a collision with the
1535 Linux macro when compiled as a kernel module.
1537 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1538 prevent a collision with the Linux macro when compiled as a kernel
1539 module.
1541 8. In pcre_compile() there are some checks for integer overflows that used to
1542 cast potentially large values to (double). This has been changed to that
1543 when building, a check for int64_t is made, and if it is found, it is used
1544 instead, thus avoiding the use of floating point arithmetic. (There is no
1545 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1546 double.
1548 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1549 2005 (difference between two addresses compared to an unsigned value).
1551 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1552 custom one, because of the following reported problem in Windows:
1554 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1555 under Win32.
1556 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1557 therefore missing the function definition.
1558 - The compiler thus generates a "C" signature for the test function.
1559 - The linker fails to find the "C" function.
1560 - PCRE fails to configure if asked to do so against libbz2.
1562 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1563 messages were output:
1565 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1566 rerunning libtoolize, to keep the correct libtool macros in-tree.
1567 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1569 I have done both of these things.
1571 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1572 most of the time, it *can* run out if it is given a pattern that contains a
1573 runaway infinite recursion. I updated the discussion in the pcrestack man
1574 page.
1576 13. Now that we have gone to the x.xx style of version numbers, the minor
1577 version may start with zero. Using 08 or 09 is a bad idea because users
1578 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1579 interpreted as invalid octal numbers. I've updated the previous comment in
1580 configure.ac, and also added a check that gives an error if 08 or 09 are
1581 used.
1583 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1584 causing partial matching to fail when the end of the subject matched \W
1585 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1587 15. There were some discrepancies between the declarations in pcre_internal.h
1588 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1589 their definitions. The declarations used "const uschar *" and the
1590 definitions used USPTR. Even though USPTR is normally defined as "const
1591 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1592 reported that: "This difference in casting confuses some C++ compilers, for
1593 example, SunCC recognizes above declarations as different functions and
1594 generates broken code for hbpcre." I have changed the declarations to use
1595 USPTR.
1597 16. GNU libtool is named differently on some systems. The autogen.sh script now
1598 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1599 (FreeBSD).
1601 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1602 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1603 comment: "Figure out how to create a longlong from a string: strtoll and
1604 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1605 instance, but it only takes 2 args instead of 3!"
1607 18. A subtle bug concerned with back references has been fixed by a change of
1608 specification, with a corresponding code fix. A pattern such as
1609 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1610 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1611 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1612 same bug. Such groups have to be quantified to be useful, or contained
1613 inside another quantified group. (If there's no repetition, the reference
1614 can never match.) The problem arises because, having left the group and
1615 moved on to the rest of the pattern, a later failure that backtracks into
1616 the group uses the captured value from the final iteration of the group
1617 rather than the correct earlier one. I have fixed this in PCRE by forcing
1618 any group that contains a reference to itself to be an atomic group; that
1619 is, there cannot be any backtracking into it once it has completed. This is
1620 similar to recursive and subroutine calls.
1623 Version 8.00 19-Oct-09
1624 ----------------------
1626 1. The table for translating pcre_compile() error codes into POSIX error codes
1627 was out-of-date, and there was no check on the pcre_compile() error code
1628 being within the table. This could lead to an OK return being given in
1629 error.
1631 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1632 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1633 in a Windows environment.
1635 3. The pcregrep --count option prints the count for each file even when it is
1636 zero, as does GNU grep. However, pcregrep was also printing all files when
1637 --files-with-matches was added. Now, when both options are given, it prints
1638 counts only for those files that have at least one match. (GNU grep just
1639 prints the file name in this circumstance, but including the count seems
1640 more useful - otherwise, why use --count?) Also ensured that the
1641 combination -clh just lists non-zero counts, with no names.
1643 4. The long form of the pcregrep -F option was incorrectly implemented as
1644 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1645 but it seems right to fix it, and I didn't think it was worth preserving
1646 the old behaviour.
1648 5. The command line items --regex=pattern and --regexp=pattern were not
1649 recognized by pcregrep, which required --regex pattern or --regexp pattern
1650 (with a space rather than an '='). The man page documented the '=' forms,
1651 which are compatible with GNU grep; these now work.
1653 6. No libpcreposix.pc file was created for pkg-config; there was just
1654 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1656 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1657 when UCP support is not needed, by modifying the Python script that
1658 generates it from Unicode data files. This should not matter if the module
1659 is correctly used as a library, but I received one complaint about 50K of
1660 unwanted data. My guess is that the person linked everything into his
1661 program rather than using a library. Anyway, it does no harm.
1663 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1664 was a minimum greater than 1 for a wide character in a possessive
1665 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1666 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1667 character. Chaos in the form of incorrect output or a compiling loop could
1668 result.
1670 9. The restrictions on what a pattern can contain when partial matching is
1671 requested for pcre_exec() have been removed. All patterns can now be
1672 partially matched by this function. In addition, if there are at least two
1673 slots in the offset vector, the offset of the earliest inspected character
1674 for the match and the offset of the end of the subject are set in them when
1675 PCRE_ERROR_PARTIAL is returned.
1677 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1678 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1679 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1680 and may be more useful for multi-segment matching.
1682 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1683 used to be given if ever the end of the subject was reached; now it is
1684 given only if matching could not proceed because another character was
1685 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1686 string "Z", which now yields "no match" instead of "partial match". In the
1687 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1688 final character ended with (*FAIL).
1690 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1691 if the pattern had a "must contain" character that was already found in the
1692 earlier partial match, unless partial matching was again requested. For
1693 example, with the pattern /dog.(body)?/, the "must contain" character is
1694 "g". If the first part-match was for the string "dog", restarting with
1695 "sbody" failed. This bug has been fixed.
1697 13. The string returned by pcre_dfa_exec() after a partial match has been
1698 changed so that it starts at the first inspected character rather than the
1699 first character of the match. This makes a difference only if the pattern
1700 starts with a lookbehind assertion or \b or \B (\K is not supported by
1701 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1702 matching functions compatible, and I think it's the right thing to do.
1704 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1705 so that the demonstration program is easily available in environments where
1706 PCRE has not been installed from source.
1708 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1709 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1710 library.
1712 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1713 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1714 is not the first non-POSIX option to be added. Clearly some people find
1715 these options useful.
1717 17. If a caller to the POSIX matching function regexec() passes a non-zero
1718 value for nmatch with a NULL value for pmatch, the value of
1719 nmatch is forced to zero.
1721 18. RunGrepTest did not have a test for the availability of the -u option of
1722 the diff command, as RunTest does. It now checks in the same way as
1723 RunTest, and also checks for the -b option.
1725 19. If an odd number of negated classes containing just a single character
1726 interposed, within parentheses, between a forward reference to a named
1727 subpattern and the definition of the subpattern, compilation crashed with
1728 an internal error, complaining that it could not find the referenced
1729 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1730 [The bug was that it was starting one character too far in when skipping
1731 over the character class, thus treating the ] as data rather than
1732 terminating the class. This meant it could skip too much.]
1734 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1735 /g option in pcretest when the pattern contains \K, which makes it possible
1736 to have an empty string match not at the start, even when the pattern is
1737 anchored. Updated pcretest and pcredemo to use this option.
1739 21. If the maximum number of capturing subpatterns in a recursion was greater
1740 than the maximum at the outer level, the higher number was returned, but
1741 with unset values at the outer level. The correct (outer level) value is
1742 now given.
1744 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1745 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1746 make it do so. The string so far is captured, making this feature
1747 compatible with Perl.
1749 23. The tests have been re-organized, adding tests 11 and 12, to make it
1750 possible to check the Perl 5.10 features against Perl 5.10.
1752 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1753 pattern matches a fixed length string. PCRE did not allow this; now it
1754 does. Neither allows recursion.
1756 25. I finally figured out how to implement a request to provide the minimum
1757 length of subject string that was needed in order to match a given pattern.
1758 (It was back references and recursion that I had previously got hung up
1759 on.) This code has now been added to pcre_study(); it finds a lower bound
1760 to the length of subject needed. It is not necessarily the greatest lower
1761 bound, but using it to avoid searching strings that are too short does give
1762 some useful speed-ups. The value is available to calling programs via
1763 pcre_fullinfo().
1765 26. While implementing 25, I discovered to my embarrassment that pcretest had
1766 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1767 study optimizations had never been tested with that matching function.
1768 Oops. What is worse, even when it was passed study data, there was a bug in
1769 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1770 were also very few tests of studied patterns with pcre_dfa_exec().
1772 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1773 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1774 on the other side of the coin, they are no longer allowed to have different
1775 names, because these cannot be distinguished in PCRE, and this has caused
1776 confusion. (This is a difference from Perl.)
1778 28. When duplicate subpattern names are present (necessarily with different
1779 numbers, as required by 27 above), and a test is made by name in a
1780 conditional pattern, either for a subpattern having been matched, or for
1781 recursion in such a pattern, all the associated numbered subpatterns are
1782 tested, and the overall condition is true if the condition is true for any
1783 one of them. This is the way Perl works, and is also more like the way
1784 testing by number works.
1787 Version 7.9 11-Apr-09
1788 ---------------------
1790 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1791 (pcretest), all targets were linked against these libraries. This included
1792 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1793 libraries. This caused unwanted dependencies to be created. This problem
1794 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1795 pcretest is linked with readline.
1797 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1798 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1799 moved outside it again, because FALSE and TRUE are already defined in AIX,
1800 but BOOL is not.
1802 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1803 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1805 4. The pcregrep documentation said spaces were inserted as well as colons (or
1806 hyphens) following file names and line numbers when outputting matching
1807 lines. This is not true; no spaces are inserted. I have also clarified the
1808 wording for the --colour (or --color) option.
1810 5. In pcregrep, when --colour was used with -o, the list of matching strings
1811 was not coloured; this is different to GNU grep, so I have changed it to be
1812 the same.
1814 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1815 each matching line was coloured. Now it goes on to look for further matches
1816 of any of the test patterns, which is the same behaviour as GNU grep.
1818 7. A pattern that could match an empty string could cause pcregrep to loop; it
1819 doesn't make sense to accept an empty string match in pcregrep, so I have
1820 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1821 seems to be how GNU grep behaves. [But see later change 40 for release
1822 8.33.]
1824 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1825 start or after a newline", because the conditional assertion was not being
1826 correctly handled. The rule now is that both the assertion and what follows
1827 in the first alternative must satisfy the test.
1829 9. If auto-callout was enabled in a pattern with a conditional group whose
1830 condition was an assertion, PCRE could crash during matching, both with
1831 pcre_exec() and pcre_dfa_exec().
1833 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1834 used for matching.
1836 11. Unicode property support in character classes was not working for
1837 characters (bytes) greater than 127 when not in UTF-8 mode.
1839 12. Added the -M command line option to pcretest.
1841 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1843 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1845 16. Added comments and documentation about mis-use of no_arg in the C++
1846 wrapper.
1848 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1849 from Martin Jerabek that uses macro names for all relevant character and
1850 string constants.
1852 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1853 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1854 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1855 these, but not everybody uses configure.
1857 19. A conditional group that had only one branch was not being correctly
1858 recognized as an item that could match an empty string. This meant that an
1859 enclosing group might also not be so recognized, causing infinite looping
1860 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1861 with the subject "ab", where knowledge that the repeated group can match
1862 nothing is needed in order to break the loop.
1864 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1865 exec(), but without supplying a callout function, matching went wrong.
1867 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1868 leak if the size of the offset vector was greater than 30. When the vector
1869 is smaller, the saved offsets during recursion go onto a local stack
1870 vector, but for larger vectors malloc() is used. It was failing to free
1871 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1872 error, in fact).
1874 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1875 heapframe that is used only when UTF-8 support is enabled. This caused no
1876 problem, but was untidy.
1878 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1879 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1880 included within another project.
1882 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1883 slightly modified by me:
1885 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1886 not building pcregrep.
1888 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1889 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1891 25. Forward references, both numeric and by name, in patterns that made use of
1892 duplicate group numbers, could behave incorrectly or give incorrect errors,
1893 because when scanning forward to find the reference group, PCRE was not
1894 taking into account the duplicate group numbers. A pattern such as
1895 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1897 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1898 the feature of a custom pointer more persuasive (as requested by a user).
1900 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1901 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1902 pre-defined.
1904 28. Added support for (*UTF8) at the start of a pattern.
1906 29. Arrange for flags added by the "release type" setting in CMake to be shown
1907 in the configuration summary.
1910 Version 7.8 05-Sep-08
1911 ---------------------
1913 1. Replaced UCP searching code with optimized version as implemented for Ad
1914 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1915 stage table and inline lookup instead of a function, giving speed ups of 2
1916 to 5 times on some simple patterns that I tested. Permission was given to
1917 distribute the MultiStage2.py script that generates the tables (it's not in
1918 the tarball, but is in the Subversion repository).
1920 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1921 scripts.
1923 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1924 a group with a zero qualifier. The result of the study could be incorrect,
1925 or the function might crash, depending on the pattern.
1927 4. Caseless matching was not working for non-ASCII characters in back
1928 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1929 It now works when Unicode Property Support is available.
1931 5. In pcretest, an escape such as \x{de} in the data was always generating
1932 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1933 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1934 truncation.
1936 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1938 7. Added two (int) casts to pcregrep when printing the difference of two
1939 pointers, in case they are 64-bit values.
1941 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1942 test 2 if it fails.
1944 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1945 and a #define of that name to empty if it is not externally set. This is to
1946 allow users of MSVC to set it if necessary.
1948 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1949 the convenience functions in the pcre_get.c source file.
1951 11. An option change at the start of a pattern that had top-level alternatives
1952 could cause overwriting and/or a crash. This command provoked a crash in
1953 some environments:
1955 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1957 This potential security problem was recorded as CVE-2008-2371.
1959 12. For a pattern where the match had to start at the beginning or immediately
1960 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1961 pcre_dfa_exec() could read past the end of the passed subject if there was
1962 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1963 pcretest so that it places the subject at the end of its malloc-ed buffer.
1965 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1966 exec() might read past the end of the data buffer in UTF-8 mode.
1968 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1969 the data contained the byte 0x85 as part of a UTF-8 character within its
1970 first line. This applied both to normal and DFA matching.
1972 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1973 /^[^d]*?$/8 failed to match "abc".
1975 16. Added a missing copyright notice to pcrecpp_internal.h.
1977 17. Make it more clear in the documentation that values returned from
1978 pcre_exec() in ovector are byte offsets, not character counts.
1980 18. Tidied a few places to stop certain compilers from issuing warnings.
1982 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1983 supplied by Stefan Weber. I made a further small update for 7.8 because
1984 there is a change of source arrangements: the pcre_searchfuncs.c module is
1985 replaced by pcre_ucd.c.
1988 Version 7.7 07-May-08
1989 ---------------------
1991 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1992 a string to a long long, pretend we don't even have a long long." This is
1993 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1995 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1996 pre-7.6 versions, which defined a global no_arg variable instead of putting
1997 it in the RE class. (See also #8 below.)
1999 3. Remove a line of dead code, identified by coverity and reported by Nuno
2000 Lopes.
2002 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2004 (1) The include/exclude patterns were being applied to the whole pathnames
2005 of files, instead of just to the final components.
2007 (2) If there was more than one level of directory, the subdirectories were
2008 skipped unless they satisfied the include/exclude conditions. This is
2009 inconsistent with GNU grep (and could even be seen as contrary to the
2010 pcregrep specification - which I improved to make it absolutely clear).
2011 The action now is always to scan all levels of directory, and just
2012 apply the include/exclude patterns to regular files.
2014 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2015 --exclude_dir in the tests to avoid scanning .svn directories.
2017 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2018 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2019 doesn't support NULs in patterns.
2021 7. Added some missing "const"s to declarations of static tables in
2022 pcre_compile.c and pcre_dfa_exec.c.
2024 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2025 caused by fix #2 above. (Subsequently also a second patch to fix the
2026 first patch. And a third patch - this was a messy problem.)
2028 9. Applied Craig's patch to remove the use of push_back().
2030 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2031 matching function regexec().
2033 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2034 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2035 references. PCRE supports relative numbers with this syntax (I don't think
2036 Oniguruma does).
2038 12. Previously, a group with a zero repeat such as (...){0} was completely
2039 omitted from the compiled regex. However, this means that if the group
2040 was called as a subroutine from elsewhere in the pattern, things went wrong
2041 (an internal error was given). Such groups are now left in the compiled
2042 pattern, with a new opcode that causes them to be skipped at execution
2043 time.
2045 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2046 to the way PCRE behaves:
2048 (a) A lone ] character is dis-allowed (Perl treats it as data).
2050 (b) A back reference to an unmatched subpattern matches an empty string
2051 (Perl fails the current match path).
2053 (c) A data ] in a character class must be notated as \] because if the
2054 first data character in a class is ], it defines an empty class. (In
2055 Perl it is not possible to have an empty class.) The empty class []
2056 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2057 The negative empty class [^] matches any one character, independently
2058 of the DOTALL setting.
2060 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2061 non-existent subpattern following a character class starting with ']' and
2062 containing () gave an internal compiling error instead of "reference to
2063 non-existent subpattern". Fortunately, when the pattern did exist, the
2064 compiled code was correct. (When scanning forwards to check for the
2065 existence of the subpattern, it was treating the data ']' as terminating
2066 the class, so got the count wrong. When actually compiling, the reference
2067 was subsequently set up correctly.)
2069 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2070 it was being rejected as not supported by pcre_dfa_exec(), even though
2071 other assertions are supported. I have made pcre_dfa_exec() support
2072 (*FAIL).
2074 16. The implementation of 13c above involved the invention of a new opcode,
2075 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2076 cannot be changed at match time, I realized I could make a small
2077 improvement to matching performance by compiling OP_ALLANY instead of
2078 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2079 on the OP_ANY path.
2081 17. Compiling pcretest on Windows with readline support failed without the
2082 following two fixes: (1) Make the unistd.h include conditional on
2083 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2085 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2086 ncurses library to be included for pcretest when ReadLine support is
2087 requested, but also to allow for it to be overridden. This patch came from
2088 Daniel Bergström.
2090 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2091 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2092 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2093 spotting this.
2096 Version 7.6 28-Jan-08
2097 ---------------------
2099 1. A character class containing a very large number of characters with
2100 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2101 overflow.
2103 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2104 HAVE_LONG_LONG is not defined.
2106 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2107 bring them up to date and include new features. This patch includes:
2109 - Fixed PH's badly added libz and libbz2 support.
2110 - Fixed a problem with static linking.
2111 - Added pcredemo. [But later removed - see 7 below.]
2112 - Fixed dftables problem and added an option.
2113 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2115 - Added readline support for pcretest.
2116 - Added an listing of the option settings after cmake has run.
2118 4. A user submitted a patch to Makefile that makes it easy to create
2119 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2120 Makefile.am that cause it to include this special target, without
2121 affecting anything else. Note that the same mingw target plus all
2122 the other distribution libraries and programs are now supported
2123 when configuring with CMake (see 6 below) instead of with
2124 Configure/Make.
2126 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2127 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2128 exported in the Windows port". It has not yet been confirmed that the patch
2129 solves the problem, but it does no harm.
2131 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2132 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2133 with CMake, and also correct the comment about stack recursion.
2135 7. Remove the automatic building of pcredemo from the ./configure system and
2136 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2137 of a program that users should build themselves after PCRE is installed, so
2138 building it automatically is not really right. What is more, it gave
2139 trouble in some build environments.
2141 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2144 Version 7.5 10-Jan-08
2145 ---------------------
2147 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2148 values in parens when parsing an RE using the C++ wrapper."
2150 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2151 Characters greater than 255 were excluded from the class instead of being
2152 included.
2154 3. The same bug as (2) above applied to negated POSIX classes such as
2155 [:^space:].
2157 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2158 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2159 I have changed it.
2161 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2162 first named subpattern) and a construct such as (?&a) would reference the
2163 first named subpattern whose name started with "a" (in other words, the
2164 length check was missing). Both these problems are fixed. "Subpattern name
2165 expected" is now given for (?&) (a zero-length name), and this patch also
2166 makes it give the same error for \k'' (previously it complained that that
2167 was a reference to a non-existent subpattern).
2169 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2170 this is right because (?- can be followed by option settings as well as by
2171 digits. I have, however, made the messages clearer.
2173 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2174 than the number used in the conditional) now cause a compile-time error.
2175 This is actually not compatible with Perl, which accepts such patterns, but
2176 treats the conditional as always being FALSE (as PCRE used to), but it
2177 seems to me that giving a diagnostic is better.
2179 8. Change "alphameric" to the more common word "alphanumeric" in comments
2180 and messages.
2182 9. Fix two occurrences of "backslash" in comments that should have been
2183 "backspace".
2185 10. Remove two redundant lines of code that can never be obeyed (their function
2186 was moved elsewhere).
2188 11. The program that makes PCRE's Unicode character property table had a bug
2189 which caused it to generate incorrect table entries for sequences of
2190 characters that have the same character type, but are in different scripts.
2191 It amalgamated them into a single range, with the script of the first of
2192 them. In other words, some characters were in the wrong script. There were
2193 thirteen such cases, affecting characters in the following ranges:
2195 U+002b0 - U+002c1
2196 U+0060c - U+0060d
2197 U+0061e - U+00612
2198 U+0064b - U+0065e
2199 U+0074d - U+0076d
2200 U+01800 - U+01805
2201 U+01d00 - U+01d77
2202 U+01d9b - U+01dbf
2203 U+0200b - U+0200f
2204 U+030fc - U+030fe
2205 U+03260 - U+0327f
2206 U+0fb46 - U+0fbb1
2207 U+10450 - U+1049d
2209 12. The -o option (show only the matching part of a line) for pcregrep was not
2210 compatible with GNU grep in that, if there was more than one match in a
2211 line, it showed only the first of them. It now behaves in the same way as
2212 GNU grep.
2214 13. If the -o and -v options were combined for pcregrep, it printed a blank
2215 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2216 does the same. The return code can be used to tell if there were any
2217 non-matching lines.
2219 14. Added --file-offsets and --line-offsets to pcregrep.
2221 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2222 infinitely looping recursion. The bug was that positive lookaheads were not
2223 being skipped when checking for a possible empty match (negative lookaheads
2224 and both kinds of lookbehind were skipped).
2226 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2227 inclusion of <windows.h> to before rather than after the definition of
2228 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2230 17. Specifying a possessive quantifier with a specific limit for a Unicode
2231 character property caused pcre_compile() to compile bad code, which led at
2232 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2233 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2234 caused the error; without that there was no problem.
2236 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2238 19. Added --enable-pcretest-libreadline.
2240 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2241 RE::GlobalReplace(). As a result, the number of replacements returned was
2242 double what it should be. I removed one of the increments, but Craig sent a
2243 later patch that removed the other one (the right fix) and added unit tests
2244 that check the return values (which was not done before).
2246 21. Several CMake things:
2248 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2249 the names libpcre and libpcreposix, not just pcre and pcreposix.
2251 (2) The above change means that pcretest and pcregrep are now correctly
2252 linked with the newly-built libraries, not previously installed ones.
2256 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2257 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2258 UTF-8 newline character). The key issue is that the pattern starts .*;
2259 this means that the match must be either at the beginning, or after a
2260 newline. The bug was in the code for advancing after a failed match and
2261 checking that the new position followed a newline. It was not taking
2262 account of UTF-8 characters correctly.
2264 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2265 character classes. PCRE was not treating the sequence [:...:] as a
2266 character class unless the ... were all letters. Perl, however, seems to
2267 allow any characters between [: and :], though of course it rejects as
2268 unknown any "names" that contain non-letters, because all the known class
2269 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2270 for example, whereas PCRE did not - it did not recognize a POSIX character
2271 class. This seemed a bit dangerous, so the code has been changed to be
2272 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2273 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2274 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2275 Perl does, and where it didn't before.
2277 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2278 Windows environments %n is disabled by default.
2281 Version 7.4 21-Sep-07
2282 ---------------------
2284 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2285 means that a class such as [\s] counted as "explicit reference to CR or
2286 LF". That isn't really right - the whole point of the change was to try to
2287 help when there was an actual mention of one of the two characters. So now
2288 the change happens only if \r or \n (or a literal CR or LF) character is
2289 encountered.
2291 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2292 of both had grown to the point where there were only 3 bits left.
2293 Fortunately, there was spare space in the data structure, and so I have
2294 moved the internal flags into a new 16-bit field to free up more option
2295 bits.
2297 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2298 but did not set the internal JCHANGED flag - either of these is enough to
2299 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2300 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2301 start sets both bits.
2303 4. Added options (at build time, compile time, exec time) to change \R from
2304 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2306 5. doc/pcresyntax.html was missing from the distribution.
2308 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2309 compatibility, even though it is no longer used.
2311 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2312 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2313 windows.h file is present (where different names are used). [This was
2314 reversed later after testing - see 16 below.]
2316 8. Changed all #include <config.h> to #include "config.h". There were also
2317 some further <pcre.h> cases that I changed to "pcre.h".
2319 9. When pcregrep was used with the --colour option, it missed the line ending
2320 sequence off the lines that it output.
2322 10. It was pointed out to me that arrays of string pointers cause lots of
2323 relocations when a shared library is dynamically loaded. A technique of
2324 using a single long string with a table of offsets can drastically reduce
2325 these. I have refactored PCRE in four places to do this. The result is
2326 dramatic:
2328 Originally: 290
2329 After changing UCP table: 187
2330 After changing error message table: 43
2331 After changing table of "verbs" 36
2332 After changing table of Posix names 22
2334 Thanks to the folks working on Gregex for glib for this insight.
2336 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2337 unicode-properties was also set.
2339 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2341 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2342 checked only for CRLF.
2344 14. Added casts to pcretest.c to avoid compiler warnings.
2346 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2348 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2349 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2350 entirely. This removes changes made in 7 above.
2352 17. The CMake files have been updated, and there is now more information about
2353 building with CMake in the NON-UNIX-USE document.
2356 Version 7.3 28-Aug-07
2357 ---------------------
2359 1. In the rejigging of the build system that eventually resulted in 7.1, the
2360 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2361 brackets there is not right, since it causes compilers to look for an
2362 installed pcre.h, not the version that is in the source that is being
2363 compiled (which of course may be different). I have changed it back to:
2365 #include "pcre.h"
2367 I have a vague recollection that the change was concerned with compiling in
2368 different directories, but in the new build system, that is taken care of
2369 by the VPATH setting the Makefile.
2371 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2372 when the subject happened to end in the byte 0x85 (e.g. if the last
2373 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2374 characters but of course it shouldn't be taken as a newline when it is part
2375 of another character. The bug was that, for an unlimited repeat of . in
2376 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2377 characters when looking for a newline.
2379 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2381 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2382 in debug output.
2384 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2385 long printing in the pcrecpp unittest when running under MinGW.
2387 6. ESC_K was left out of the EBCDIC table.
2389 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2390 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2391 limit also applies to "virtual nesting" when a pattern is recursive, and in
2392 this case 1000 isn't so big. I have been able to remove this limit at the
2393 expense of backing off one optimization in certain circumstances. Normally,
2394 when pcre_exec() would call its internal match() function recursively and
2395 immediately return the result unconditionally, it uses a "tail recursion"
2396 feature to save stack. However, when a subpattern that can match an empty
2397 string has an unlimited repetition quantifier, it no longer makes this
2398 optimization. That gives it a stack frame in which to save the data for
2399 checking that an empty string has been matched. Previously this was taken
2400 from the 1000-entry workspace that had been reserved. So now there is no
2401 explicit limit, but more stack is used.
2403 8. Applied Daniel's patches to solve problems with the import/export magic
2404 syntax that is required for Windows, and which was going wrong for the
2405 pcreposix and pcrecpp parts of the library. These were overlooked when this
2406 problem was solved for the main library.
2408 9. There were some crude static tests to avoid integer overflow when computing
2409 the size of patterns that contain repeated groups with explicit upper
2410 limits. As the maximum quantifier is 65535, the maximum group length was
2411 set at 30,000 so that the product of these two numbers did not overflow a
2412 32-bit integer. However, it turns out that people want to use groups that
2413 are longer than 30,000 bytes (though not repeat them that many times).
2414 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2415 made it possible to implement the integer overflow checks in a much more
2416 dynamic way, which I have now done. The artificial limitation on group
2417 length has been removed - we now have only the limit on the total length of
2418 the compiled pattern, which depends on the LINK_SIZE setting.
2420 10. Fixed a bug in the documentation for get/copy named substring when
2421 duplicate names are permitted. If none of the named substrings are set, the
2422 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2423 empty string.
2425 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2426 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2427 because the ] is interpreted as the first data character and the
2428 terminating ] is not found. PCRE has been made compatible with Perl in this
2429 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2430 cause memory overwriting.
2432 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2433 string has been matched (to stop an infinite loop). It was not recognizing
2434 a conditional subpattern that could match an empty string if that
2435 subpattern was within another subpattern. For example, it looped when
2436 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2437 condition was not nested. This bug has been fixed.
2439 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2440 past the start of the subject in the presence of bytes with the top bit
2441 set, for example "\x8aBCD".
2443 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2444 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2446 14. Optimized (?!) to (*FAIL).
2448 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2449 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2450 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2451 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2452 does: it's just the validity check that is more restrictive.
2454 16. Inserted checks for integer overflows during escape sequence (backslash)
2455 processing, and also fixed erroneous offset values for syntax errors during
2456 backslash processing.
2458 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2459 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2461 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2462 caused an overrun.
2464 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2465 something other than just ASCII characters) inside a group that had an
2466 unlimited repeat caused a loop at compile time (while checking to see
2467 whether the group could match an empty string).
2469 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2470 [\P{Any}] did so. (Error in the code for printing property names.)
2472 21. An orphan \E inside a character class could cause a crash.
2474 22. A repeated capturing bracket such as (A)? could cause a wild memory
2475 reference during compilation.
2477 23. There are several functions in pcre_compile() that scan along a compiled
2478 expression for various reasons (e.g. to see if it's fixed length for look
2479 behind). There were bugs in these functions when a repeated \p or \P was
2480 present in the pattern. These operators have additional parameters compared
2481 with \d, etc, and these were not being taken into account when moving along
2482 the compiled data. Specifically:
2484 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2485 length.
2487 (b) An item such as \pL+ within a repeated group could cause crashes or
2488 loops.
2490 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2491 "reference to non-existent subpattern" error.
2493 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2495 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2496 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2498 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2500 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2501 character were causing crashes (broken optimization).
2503 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2504 \p or \P) caused a compile-time loop.
2506 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2507 break. For example, the unstudied pattern [\r\n]A does not match the string
2508 "\r\nA" because change 7.0/46 below moves the current point on by two
2509 characters after failing to match at the start. However, the pattern \nA
2510 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2511 the same is true. There doesn't seem any very clean way out of this, but
2512 what I have chosen to do makes the common cases work: PCRE now takes note
2513 of whether there can be an explicit match for \r or \n anywhere in the
2514 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2515 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2516 pattern has explicit CR or LF references.
2518 29. Added (*CR) etc for changing newline setting at start of pattern.
2521 Version 7.2 19-Jun-07
2522 ---------------------
2524 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2525 which is apparently normally available under Windows.
2527 2. Re-jig the pcregrep tests with different newline settings in an attempt
2528 to make them independent of the local environment's newline setting.
2530 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2532 4. Some of the "internals" tests were previously cut out when the link size
2533 was not 2, because the output contained actual offsets. The recent new
2534 "Z" feature of pcretest means that these can be cut out, making the tests
2535 usable with all link sizes.
2537 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2538 stack recursion. This gives a massive performance boost under BSD, but just
2539 a small improvement under Linux. However, it saves one field in the frame
2540 in all cases.
2542 6. Added more features from the forthcoming Perl 5.10:
2544 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2545 recursion call. It refers to the nth most recently opened parentheses.
2547 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2548 to be opened parentheses.
2550 (c) Conditions that refer to capturing parentheses can be specified
2551 relatively, for example, (?(-2)... or (?(+3)...
2553 (d) \K resets the start of the current match so that everything before
2554 is not part of it.
2556 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2558 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2559 reference syntax.
2561 (g) (?| introduces a group in which the numbering of parentheses in each
2562 alternative starts with the same number.
2564 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2566 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2569 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2570 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2571 for detecting groups that can match an empty string.
2573 9. A pattern with a very large number of alternatives (more than several
2574 hundred) was running out of internal workspace during the pre-compile
2575 phase, where pcre_compile() figures out how much memory will be needed. A
2576 bit of new cunning has reduced the workspace needed for groups with
2577 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2578 workspace instead of running out of the 4096 that are available.
2580 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2582 11. Applied patch from Google to remove an optimization that didn't quite work.
2583 The report of the bug said:
2585 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2586 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2587 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2589 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2590 it matched the wrong number of bytes.
2593 Version 7.1 24-Apr-07
2594 ---------------------
2596 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2597 that is more "standard", making use of automake and other Autotools. There
2598 is some re-arrangement of the files and adjustment of comments consequent
2599 on this.
2601 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2602 for recursive directory scanning broke on some systems because the files
2603 are not scanned in any specific order and on different systems the order
2604 was different. A call to "sort" has been inserted into RunGrepTest for the
2605 approprate test as a short-term fix. In the longer term there may be an
2606 alternative.
2608 3. I had an email from Eric Raymond about problems translating some of PCRE's
2609 man pages to HTML (despite the fact that I distribute HTML pages, some
2610 people do their own conversions for various reasons). The problems
2611 concerned the use of low-level troff macros .br and .in. I have therefore
2612 removed all such uses from the man pages (some were redundant, some could
2613 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2614 HTML has been updated to handle .nf/.fi and to complain if it encounters
2615 .br or .in.
2617 4. Updated comments in configure.ac that get placed in config.h.in and also
2618 arranged for config.h to be included in the distribution, with the name
2619 config.h.generic, for the benefit of those who have to compile without
2620 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2622 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2623 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2624 makevp.bat for the current PCRE, using the additional files
2625 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2627 6. A Windows user reported a minor discrepancy with test 2, which turned out
2628 to be caused by a trailing space on an input line that had got lost in his
2629 copy. The trailing space was an accident, so I've just removed it.
2631 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2632 that is needed.
2634 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2635 as "const" (a) because they are and (b) because it helps the PHP
2636 maintainers who have recently made a script to detect big data structures
2637 in the php code that should be moved to the .rodata section. I remembered
2638 to update Builducptable as well, so it won't revert if ucptable.h is ever
2639 re-created.
2641 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2642 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2643 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2644 support is not required. This saves 1.5-2K of code, which is important in
2645 some applications.
2647 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2648 so as not to refer to the tables, even though these functions will never be
2649 called when UTF-8 support is disabled. Otherwise there are problems with a
2650 shared library.
2652 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2654 (a) It was defining its arguments as char * instead of void *.
2656 (b) It was assuming that all moves were upwards in memory; this was true
2657 a long time ago when I wrote it, but is no longer the case.
2659 The emulated memove() is provided for those environments that have neither
2660 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2661 is clearly not the case, as these two bugs were recently reported.
2663 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2664 and Detrail to create the HTML documentation, the .txt form of the man
2665 pages, and it removes trailing spaces from listed files. It also creates
2666 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2667 case, it wraps all the #defines with #ifndefs. This script should be run
2668 before "make dist".
2670 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2671 with Unicode property support.
2673 (a) For a maximizing quantifier, if the two different cases of the
2674 character were of different lengths in their UTF-8 codings (there are
2675 some cases like this - I found 11), and the matching function had to
2676 back up over a mixture of the two cases, it incorrectly assumed they
2677 were both the same length.
2679 (b) When PCRE was configured to use the heap rather than the stack for
2680 recursion during matching, it was not correctly preserving the data for
2681 the other case of a UTF-8 character when checking ahead for a match
2682 while processing a minimizing repeat. If the check also involved
2683 matching a wide character, but failed, corruption could cause an
2684 erroneous result when trying to check for a repeat of the original
2685 character.
2687 13. Some tidying changes to the testing mechanism:
2689 (a) The RunTest script now detects the internal link size and whether there
2690 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2691 values substituted by "configure". (The RunGrepTest script already did
2692 this for UTF-8.) The configure.ac script no longer substitutes the
2693 relevant variables.
2695 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2696 with length and offset values. This means that the output is different
2697 for different internal link sizes. Test 2 is skipped for link sizes
2698 other than 2 because of this, bypassing the problem. Unfortunately,
2699 there was also a test in test 3 (the locale tests) that used /B and
2700 failed for link sizes other than 2. Rather than cut the whole test out,
2701 I have added a new /Z option to pcretest that replaces the length and
2702 offset values with spaces. This is now used to make test 3 independent
2703 of link size. (Test 2 will be tidied up later.)
2705 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2706 segmentation fault instead of returning the appropriate error message.
2708 15. In multiline mode when the newline sequence was set to "any", the pattern
2709 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2710 This doesn't seem right; it now treats the CRLF combination as the line
2711 ending, and so does not match in that case. It's only a pattern such as ^$
2712 that would hit this one: something like ^ABC$ would have failed after \r
2713 and then tried again after \r\n.
2715 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2716 in an attempt to make files that differ only in their line terminators
2717 compare equal. This works on Linux.
2719 17. Under certain error circumstances pcregrep might try to free random memory
2720 as it exited. This is now fixed, thanks to valgrind.
2722 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2723 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2724 was because its rules for how to advance for /g after matching an empty
2725 string at the end of a line did not allow for this case. They now check for
2726 it specially.
2728 20. pcretest is supposed to handle patterns and data of any length, by
2729 extending its buffers when necessary. It was getting this wrong when the
2730 buffer for a data line had to be extended.
2732 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2733 CRLF as a newline sequence.
2735 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2736 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2737 I have nevertheless tidied it up.
2739 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2741 24. Added a man page for pcre-config.
2744 Version 7.0 19-Dec-06
2745 ---------------------
2747 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2748 moving to gcc 4.1.1.
2750 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2751 sys/time.h, which is documented as needed for this function. It doesn't
2752 seem to matter on Linux, but it showed up on some releases of OS X.
2754 3. It seems that there are systems where bytes whose values are greater than
2755 127 match isprint() in the "C" locale. The "C" locale should be the
2756 default when a C program starts up. In most systems, only ASCII printing
2757 characters match isprint(). This difference caused the output from pcretest
2758 to vary, making some of the tests fail. I have changed pcretest so that:
2760 (a) When it is outputting text in the compiled version of a pattern, bytes
2761 other than 32-126 are always shown as hex escapes.
2763 (b) When it is outputting text that is a matched part of a subject string,
2764 it does the same, unless a different locale has been set for the match
2765 (using the /L modifier). In this case, it uses isprint() to decide.
2767 4. Fixed a major bug that caused incorrect computation of the amount of memory
2768 required for a compiled pattern when options that changed within the
2769 pattern affected the logic of the preliminary scan that determines the
2770 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2771 that the computed length was too small. The symptoms of this bug were
2772 either the PCRE error "internal error: code overflow" from pcre_compile(),
2773 or a glibc crash with a message such as "pcretest: free(): invalid next
2774 size (fast)". Examples of patterns that provoked this bug (shown in
2775 pcretest format) are:
2777 /(?-x: )/x
2778 /(?x)(?-x: \s*#\s*)/
2779 /((?i)[\x{c0}])/8
2780 /(?i:[\x{c0}])/8
2782 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2783 is now done differently.
2785 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2786 wrapper classes; (b) implement a new function in the C++ scanner that is
2787 more efficient than the old way of doing things because it avoids levels of
2788 recursion in the regex matching; (c) add a paragraph to the documentation
2789 for the FullMatch() function.
2791 6. The escape sequence \n was being treated as whatever was defined as
2792 "newline". Not only was this contrary to the documentation, which states
2793 that \n is character 10 (hex 0A), but it also went horribly wrong when
2794 "newline" was defined as CRLF. This has been fixed.
2796 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2797 was being set to -1 for the "end of line" case (supposedly a value that no
2798 character can have). Though this value is never used (the check for end of
2799 line is "zero bytes in current character"), it caused compiler complaints.
2800 I've changed it to 0xffffffff.
2802 8. In pcre_version.c, the version string was being built by a sequence of
2803 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2804 string (as it is for production releases) called a macro with an empty
2805 argument. The C standard says the result of this is undefined. The gcc
2806 compiler treats it as an empty string (which was what was wanted) but it is
2807 reported that Visual C gives an error. The source has been hacked around to
2808 avoid this problem.
2810 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2811 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2812 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2813 of them did).
2815 10. Originally, pcretest opened its input and output without "b"; then I was
2816 told that "b" was needed in some environments, so it was added for release
2817 5.0 to both the input and output. (It makes no difference on Unix-like
2818 systems.) Later I was told that it is wrong for the input on Windows. I've
2819 now abstracted the modes into two macros, to make it easier to fiddle with
2820 them, and removed "b" from the input mode under Windows.
2822 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2824 12. Added -help and --help to pcretest as an official way of being reminded
2825 of the options.
2827 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2828 and pcrecpp.cc because they annoy compilers at high warning levels.
2830 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2832 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2833 scripts are not C programs :-) and which was not noticed because it works
2834 on Linux.
2836 16. pcretest is supposed to handle any length of pattern and data line (as one
2837 line or as a continued sequence of lines) by extending its input buffer if
2838 necessary. This feature was broken for very long pattern lines, leading to
2839 a string of junk being passed to pcre_compile() if the pattern was longer
2840 than about 50K.
2842 17. I have done a major re-factoring of the way pcre_compile() computes the
2843 amount of memory needed for a compiled pattern. Previously, there was code
2844 that made a preliminary scan of the pattern in order to do this. That was
2845 OK when PCRE was new, but as the facilities have expanded, it has become
2846 harder and harder to keep it in step with the real compile phase, and there
2847 have been a number of bugs (see for example, 4 above). I have now found a
2848 cunning way of running the real compile function in a "fake" mode that
2849 enables it to compute how much memory it would need, while actually only
2850 ever using a few hundred bytes of working memory and without too many
2851 tests of the mode. This should make future maintenance and development
2852 easier. A side effect of this work is that the limit of 200 on the nesting
2853 depth of parentheses has been removed (though this was never a serious
2854 limitation, I suspect). However, there is a downside: pcre_compile() now
2855 runs more slowly than before (30% or more, depending on the pattern). I
2856 hope this isn't a big issue. There is no effect on runtime performance.
2858 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2859 newline (only possible for the last line of a file) and it was a
2860 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2862 19. Added additional timing features to pcretest. (1) The -tm option now times
2863 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2864 separate command line item, by a number that specifies the number of
2865 repeats to use when timing. The default is 50000; this gives better
2866 precision, but takes uncomfortably long for very large patterns.
2868 20. Extended pcre_study() to be more clever in cases where a branch of a
2869 subpattern has no definite first character. For example, (a*|b*)[cd] would
2870 previously give no result from pcre_study(). Now it recognizes that the
2871 first character must be a, b, c, or d.
2873 21. There was an incorrect error "recursive call could loop indefinitely" if
2874 a subpattern (or the entire pattern) that was being tested for matching an
2875 empty string contained only one non-empty item after a nested subpattern.
2876 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2877 incorrectly, because the \d was being skipped in the check.
2879 22. The pcretest program now has a new pattern option /B and a command line
2880 option -b, which is equivalent to adding /B to every pattern. This causes
2881 it to show the compiled bytecode, without the additional information that
2882 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2883 is the same as /B/I).
2885 23. A new optimization is now able automatically to treat some sequences such
2886 as a*b as a*+b. More specifically, if something simple (such as a character
2887 or a simple class like \d) has an unlimited quantifier, and is followed by
2888 something that cannot possibly match the quantified thing, the quantifier
2889 is automatically "possessified".
2891 24. A recursive reference to a subpattern whose number was greater than 39
2892 went wrong under certain circumstances in UTF-8 mode. This bug could also
2893 have affected the operation of pcre_study().
2895 25. Realized that a little bit of performance could be had by replacing
2896 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2898 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2900 27. Possessive quantifiers such as a++ were previously implemented by turning
2901 them into atomic groups such as ($>a+). Now they have their own opcodes,
2902 which improves performance. This includes the automatically created ones
2903 from 23 above.
2905 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2906 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2907 the first matched character to be a colon. This applied both to named and
2908 numbered groups.
2910 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2912 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2913 building PCRE on Windows easier, so I have included it in the distribution.
2915 31. There is now a check in pcretest against a ridiculously large number being
2916 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2917 loop, the loop is abandoned.
2919 32. Forward references to subpatterns in conditions such as (?(2)...) where
2920 subpattern 2 is defined later cause pcre_compile() to search forwards in
2921 the pattern for the relevant set of parentheses. This search went wrong
2922 when there were unescaped parentheses in a character class, parentheses
2923 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2925 33. "Subroutine" calls and backreferences were previously restricted to
2926 referencing subpatterns earlier in the regex. This restriction has now
2927 been removed.
2929 34. Added a number of extra features that are going to be in Perl 5.10. On the
2930 whole, these are just syntactic alternatives for features that PCRE had
2931 previously implemented using the Python syntax or my own invention. The
2932 other formats are all retained for compatibility.
2934 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2935 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2936 also .NET compatible.
2938 (b) A recursion or subroutine call to a named group can now be defined as
2939 (?&name) as well as (?P>name).
2941 (c) A backreference to a named group can now be defined as \k<name> or
2942 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2943 5.10, are also .NET compatible.
2945 (d) A conditional reference to a named group can now use the syntax
2946 (?(<name>) or (?('name') as well as (?(name).
2948 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2949 groups (named and numbered) that are never evaluated inline, but can be
2950 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2951 is always false. There may be only one alternative in such a group.
2953 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2954 as the simple (?(R). The condition is true only if the most recent
2955 recursion is that of the given number or name. It does not search out
2956 through the entire recursion stack.
2958 (g) The escape \gN or \g{N} has been added, where N is a positive or
2959 negative number, specifying an absolute or relative reference.
2961 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2962 some "unreachable code" warnings.
2964 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2965 things, this adds five new scripts.
2967 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2968 There were also incompatibilities regarding the handling of \Q..\E inside
2969 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2970 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2972 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2973 matches an empty string, and forcibly breaks the loop. There were bugs in
2974 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2975 against aaaa the result was just "a" rather than "aaaa", for example. Two
2976 separate and independent bugs (that affected different cases) have been
2977 fixed.
2979 39. Refactored the code to abolish the use of different opcodes for small
2980 capturing bracket numbers. This is a tidy that I avoided doing when I
2981 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2982 The new approach is not only tidier, it makes it possible to reduce the
2983 memory needed to fix the previous bug (38).
2985 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2986 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2987 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2988 mode.
2990 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2991 report.
2993 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2994 copy construction and assignment in the C++ wrapper.
2996 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2997 couple of bugs that could have given wrong results in the "--newline=crlf"
2998 case.
3000 44. Added a number of casts and did some reorganization of signed/unsigned int
3001 variables following suggestions from Dair Grant. Also renamed the variable
3002 "this" as "item" because it is a C++ keyword.
3004 45. Arranged for dftables to add
3006 #include "pcre_internal.h"
3008 to pcre_chartables.c because without it, gcc 4.x may remove the array
3009 definition from the final binary if PCRE is built into a static library and
3010 dead code stripping is activated.
3012 46. For an unanchored pattern, if a match attempt fails at the start of a
3013 newline sequence, and the newline setting is CRLF or ANY, and the next two
3014 characters are CRLF, advance by two characters instead of one.
3017 Version 6.7 04-Jul-06
3018 ---------------------
3020 1. In order to handle tests when input lines are enormously long, pcretest has
3021 been re-factored so that it automatically extends its buffers when
3022 necessary. The code is crude, but this _is_ just a test program. The
3023 default size has been increased from 32K to 50K.
3025 2. The code in pcre_study() was using the value of the re argument before
3026 testing it for NULL. (Of course, in any sensible call of the function, it
3027 won't be NULL.)
3029 3. The memmove() emulation function in pcre_internal.h, which is used on
3030 systems that lack both memmove() and bcopy() - that is, hardly ever -
3031 was missing a "static" storage class specifier.
3033 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3034 containing an extended class (one that cannot be represented by a bitmap
3035 because it contains high-valued characters or Unicode property items, e.g.
3036 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3037 pattern, but PCRE should not loop if you do not (it no longer does).
3038 [Detail: two cases were found: (a) a repeated subpattern containing an
3039 extended class; (b) a recursive reference to a subpattern that followed a
3040 previous extended class. It wasn't skipping over the extended class
3041 correctly when UTF-8 mode was not set.]
3043 5. A negated single-character class was not being recognized as fixed-length
3044 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3045 compile error "lookbehind assertion is not fixed length".
3047 6. The RunPerlTest auxiliary script was showing an unexpected difference
3048 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3049 write a Perl script that can interpret lines of an input file either as
3050 byte characters or as UTF-8, which is what "perltest" was being required to
3051 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3052 can't do is switch easily at run time between having the "use utf8;" pragma
3053 or not. In the end, I fudged it by using the RunPerlTest script to insert
3054 "use utf8;" explicitly for the UTF-8 tests.
3056 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3057 the end of the subject string, contrary to the documentation and to what
3058 Perl does. This was true of both matching functions. Now it matches only at
3059 the start of the subject and immediately after *internal* newlines.
3061 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3062 a pointer to an int instead of a pointer to an unsigned long int. This
3063 caused problems on 64-bit systems.
3065 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3066 instance of the 'standard' template library not being so standard".
3068 10. There was no check on the number of named subpatterns nor the maximum
3069 length of a subpattern name. The product of these values is used to compute
3070 the size of the memory block for a compiled pattern. By supplying a very
3071 long subpattern name and a large number of named subpatterns, the size
3072 computation could be caused to overflow. This is now prevented by limiting
3073 the length of names to 32 characters, and the number of named subpatterns
3074 to 10,000.
3076 11. Subpatterns that are repeated with specific counts have to be replicated in
3077 the compiled pattern. The size of memory for this was computed from the
3078 length of the subpattern and the repeat count. The latter is limited to
3079 65535, but there was no limit on the former, meaning that integer overflow
3080 could in principle occur. The compiled length of a repeated subpattern is
3081 now limited to 30,000 bytes in order to prevent this.
3083 12. Added the optional facility to have named substrings with the same name.
3085 13. Added the ability to use a named substring as a condition, using the
3086 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3087 are numbers (not recommended). Forward references are permitted.
3089 14. Added forward references in named backreferences (if you see what I mean).
3091 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3092 pattern could run off the end of the subject. For example, the pattern
3093 "(?s)(.{1,5})"8 did this with the subject "ab".
3095 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3096 PCRE_CASELESS was set when matching characters that were quantified with ?
3097 or *.
3099 17. A character class other than a single negated character that had a minimum
3100 but no maximum quantifier - for example [ab]{6,} - was not handled
3101 correctly by pce_dfa_exec(). It would match only one character.
3103 18. A valid (though odd) pattern that looked like a POSIX character
3104 class but used an invalid character after [ (for example [[,abc,]]) caused
3105 pcre_compile() to give the error "Failed: internal error: code overflow" or
3106 in some cases to crash with a glibc free() error. This could even happen if
3107 the pattern terminated after [[ but there just happened to be a sequence of
3108 letters, a binary zero, and a closing ] in the memory that followed.
3110 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3111 over the years. Originally (before any Unicode support), just the bottom 8
3112 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3113 output from "man perlunicode" includes this:
3115 The regular expression compiler produces polymorphic opcodes. That
3116 is, the pattern adapts to the data and automatically switches to
3117 the Unicode character scheme when presented with Unicode data--or
3118 instead uses a traditional byte scheme when presented with byte
3119 data.
3121 Sadly, a wide octal escape does not cause a switch, and in a string with
3122 no other multibyte characters, these octal escapes are treated as before.
3123 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3124 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3125 Unicode string.
3127 I have not perpetrated such confusion in PCRE. Up till now, it took just
3128 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3129 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3130 translate to the appropriate multibyte character.
3132 29. Applied some refactoring to reduce the number of warnings from Microsoft
3133 and Borland compilers. This has included removing the fudge introduced
3134 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3135 a warning about an unused variable.
3137 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3138 characters since release 4.0, because Perl (from release 5.004) does not.
3139 [Or at least, is documented not to: some releases seem to be in conflict
3140 with the documentation.] However, when a pattern was studied with
3141 pcre_study() and all its branches started with \s, PCRE still included VT
3142 as a possible starting character. Of course, this did no harm; it just
3143 caused an unnecessary match attempt.
3145 22. Removed a now-redundant internal flag bit that recorded the fact that case
3146 dependency changed within the pattern. This was once needed for "required
3147 byte" processing, but is no longer used. This recovers a now-scarce options
3148 bit. Also moved the least significant internal flag bit to the most-
3149 significant bit of the word, which was not previously used (hangover from
3150 the days when it was an int rather than a uint) to free up another bit for
3151 the future.
3153 23. Added support for CRLF line endings as well as CR and LF. As well as the
3154 default being selectable at build time, it can now be changed at runtime
3155 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3156 specify that it is scanning data with non-default line endings.
3158 24. Changed the definition of CXXLINK to make it agree with the definition of
3159 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3161 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3162 recursions. This makes a big different to stack usage for some patterns.
3164 26. If a subpattern containing a named recursion or subroutine reference such
3165 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3166 the space required for the compiled pattern went wrong and gave too small a
3167 value. Depending on the environment, this could lead to "Failed: internal
3168 error: code overflow at offset 49" or "glibc detected double free or
3169 corruption" errors.
3171 27. Applied patches from Google (a) to support the new newline modes and (b) to
3172 advance over multibyte UTF-8 characters in GlobalReplace.
3174 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3175 difference for some implementation of PCRE in some Windows version.
3177 29. Added some extra testing facilities to pcretest:
3179 \q<number> in a data line sets the "match limit" value
3180 \Q<number> in a data line sets the "match recursion limt" value
3181 -S <number> sets the stack size, where <number> is in megabytes
3183 The -S option isn't available for Windows.
3186 Version 6.6 06-Feb-06
3187 ---------------------
3189 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3190 in pcreposix.h. I have copied the definition from pcre.h.
3192 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3193 because pcre.h is no longer a built file.
3195 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3196 not normally included in the compiled code.
3199 Version 6.5 01-Feb-06
3200 ---------------------
3202 1. When using the partial match feature with pcre_dfa_exec(), it was not
3203 anchoring the second and subsequent partial matches at the new starting
3204 point. This could lead to incorrect results. For example, with the pattern
3205 /1234/, partially matching against "123" and then "a4" gave a match.
3207 2. Changes to pcregrep:
3209 (a) All non-match returns from pcre_exec() were being treated as failures
3210 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3211 error message is output. Some extra information is given for the
3213 probably the only errors that are likely to be caused by users (by
3214 specifying a regex that has nested indefinite repeats, for instance).
3215 If there are more than 20 of these errors, pcregrep is abandoned.
3217 (b) A binary zero was treated as data while matching, but terminated the
3218 output line if it was written out. This has been fixed: binary zeroes
3219 are now no different to any other data bytes.
3221 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3222 used to set a locale for matching. The --locale=xxxx long option has
3223 been added (no short equivalent) to specify a locale explicitly on the
3224 pcregrep command, overriding the environment variables.
3226 (d) When -B was used with -n, some line numbers in the output were one less
3227 than they should have been.
3229 (e) Added the -o (--only-matching) option.
3231 (f) If -A or -C was used with -c (count only), some lines of context were
3232 accidentally printed for the final match.
3234 (g) Added the -H (--with-filename) option.
3236 (h) The combination of options -rh failed to suppress file names for files
3237 that were found from directory arguments.
3239 (i) Added the -D (--devices) and -d (--directories) options.
3241 (j) Added the -F (--fixed-strings) option.
3243 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3245 (l) Added the --colo(u)r option.
3247 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3248 is not present by default.
3250 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3251 items such as (?R) or (?1), when the recursion could match a number of
3252 alternatives. If it matched one of the alternatives, but subsequently,
3253 outside the recursion, there was a failure, the code tried to back up into
3254 the recursion. However, because of the way PCRE is implemented, this is not
3255 possible, and the result was an incorrect result from the match.
3257 In order to prevent this happening, the specification of recursion has
3258 been changed so that all such subpatterns are automatically treated as
3259 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3261 4. I had overlooked the fact that, in some locales, there are characters for
3262 which isalpha() is true but neither isupper() nor islower() are true. In
3263 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3264 and ordfeminine) are like this. This affected the treatment of \w and \W
3265 when they appeared in character classes, but not when they appeared outside
3266 a character class. The bit map for "word" characters is now created
3267 separately from the results of isalnum() instead of just taking it from the
3268 upper, lower, and digit maps. (Plus the underscore character, of course.)
3270 5. The above bug also affected the handling of POSIX character classes such as
3271 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3272 permanent tables. Instead, the bit maps for such a class were previously
3273 created as the appropriate unions of the upper, lower, and digit bitmaps.
3274 Now they are created by subtraction from the [[:word:]] class, which has
3275 its own bitmap.
3277 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3278 It is created by subtracting the vertical space characters (\x09, \x0a,
3279 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3280 subtraction was done in the overall bitmap for a character class, meaning
3281 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3282 be recognized. This bug has been fixed.
3284 7. Patches from the folks at Google:
3286 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3287 real life, but is still worth protecting against".
3289 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3290 regular expressions".
3292 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3293 have it.
3295 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3296 "configure" and the latter not, in order to fix a problem somebody had
3297 with compiling the Arg class on HP-UX.
3299 (e) Improve the error-handling of the C++ wrapper a little bit.
3301 (f) New tests for checking recursion limiting.
3303 8. The pcre_memmove() function, which is used only if the environment does not
3304 have a standard memmove() function (and is therefore rarely compiled),
3305 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3306 returning a result (though PCRE never actually uses the result).
3308 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3309 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3310 returned instead of calling malloc() with an overflowing number that would
3311 most likely cause subsequent chaos.
3313 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3315 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3316 with this option is matched, the nmatch and pmatch options of regexec() are
3317 ignored.
3319 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3320 provided in case anyone wants to the the POSIX interface with UTF-8
3321 strings.
3323 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3324 C++ linking (needed for some HP-UX environments).
3326 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3327 (unused parameter) and in the pcre_printint() function (omitted "default"
3328 switch label when the default is to do nothing).
3330 15. Added some code to make it possible, when PCRE is compiled as a C++
3331 library, to replace subject pointers for pcre_exec() with a smart pointer
3332 class, thus making it possible to process discontinuous strings.
3334 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3335 much the same function. They were added by different people who were trying
3336 to make PCRE easy to compile on non-Unix systems. It has been suggested
3337 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3338 for compiling on Windows systems. I have therefore replaced it with
3339 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3340 defaults to "extern" for C or "extern C" for C++, which works fine on
3341 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3342 SCOPE with something explicit in config.h. In addition:
3344 (a) pcreposix.h still had just "extern" instead of either of these macros;
3345 I have replaced it with PCRE_DATA_SCOPE.
3347 (b) Functions such as _pcre_xclass(), which are internal to the library,
3348 but external in the C sense, all had PCRE_EXPORT in their definitions.
3349 This is apparently wrong for the Windows case, so I have removed it.
3350 (It makes no difference on Unix-like systems.)
3352 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3353 of recursive calls to match(). This is different to MATCH_LIMIT because
3354 that limits the total number of calls to match(), not all of which increase
3355 the depth of recursion. Limiting the recursion depth limits the amount of
3356 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3357 when PCRE is compiled, and changed at run time. A patch from Google adds
3358 this functionality to the C++ interface.
3360 18. Changes to the handling of Unicode character properties:
3362 (a) Updated the table to Unicode 4.1.0.
3364 (b) Recognize characters that are not in the table as "Cn" (undefined).
3366 (c) I revised the way the table is implemented to a much improved format
3367 which includes recognition of ranges. It now supports the ranges that
3368 are defined in UnicodeData.txt, and it also amalgamates other
3369 characters into ranges. This has reduced the number of entries in the
3370 table from around 16,000 to around 3,000, thus reducing its size
3371 considerably. I realized I did not need to use a tree structure after
3372 all - a binary chop search is just as efficient. Having reduced the
3373 number of entries, I extended their size from 6 bytes to 8 bytes to
3374 allow for more data.
3376 (d) Added support for Unicode script names via properties such as \p{Han}.
3378 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3379 matching that character.
3381 20. When matching a repeated Unicode property with a minimum greater than zero,
3382 (for example \pL{2,}), PCRE could look past the end of the subject if it
3383 reached it while seeking the minimum number of characters. This could
3384 happen only if some of the characters were more than one byte long, because
3385 there is a check for at least the minimum number of bytes.
3387 21. Refactored the implementation of \p and \P so as to be more general, to
3388 allow for more different types of property in future. This has changed the
3389 compiled form incompatibly. Anybody with saved compiled patterns that use
3390 \p or \P will have to recompile them.
3392 22. Added "Any" and "L&" to the supported property types.
3394 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3395 but give a compile time error if the value is greater than 0xff.
3397 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3398 accidentally not being installed or uninstalled.
3400 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3401 made were to insert the current release number. This seemed silly, because
3402 it made things harder for people building PCRE on systems that don't run
3403 "configure". I have turned pcre.h into a distributed file, no longer built
3404 by "configure", with the version identification directly included. There is
3405 no longer a pcre.h.in file.
3407 However, this change necessitated a change to the pcre-config script as
3408 well. It is built from pcre-config.in, and one of the substitutions was the
3409 release number. I have updated configure.ac so that ./configure now finds
3410 the release number by grepping pcre.h.
3412 26. Added the ability to run the tests under valgrind.
3415 Version 6.4 05-Sep-05
3416 ---------------------
3418 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3419 "--" to be printed when multiple files were scanned, even when none of the
3420 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3421 consider it to be a bug, and have restored the previous behaviour.
3423 2. A couple of code tidies to get rid of compiler warnings.
3425 3. The pcretest program used to cheat by referring to symbols in the library
3426 whose names begin with _pcre_. These are internal symbols that are not
3427 really supposed to be visible externally, and in some environments it is
3428 possible to suppress them. The cheating is now confined to including
3429 certain files from the library's source, which is a bit cleaner.
3431 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3432 file's purpose clearer.
3434 5. Reorganized pcre_ucp_findchar().
3437 Version 6.3 15-Aug-05
3438 ---------------------
3440 1. The file libpcre.pc.in did not have general read permission in the tarball.
3442 2. There were some problems when building without C++ support:
3444 (a) If C++ support was not built, "make install" and "make test" still
3445 tried to test it.
3447 (b) There were problems when the value of CXX was explicitly set. Some
3448 changes have been made to try to fix these, and ...
3450 (c) --disable-cpp can now be used to explicitly disable C++ support.
3452 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3453 backslash in a target when C++ was disabled. This confuses some
3454 versions of "make", apparently. Using an intermediate variable solves
3455 this. (Same for CPP_LOBJ.)
3457 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3458 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3459 necessary on certain architectures.
3461 4. Added a setting of -export-symbols-regex to the link command to remove
3462 those symbols that are exported in the C sense, but actually are local
3463 within the library, and not documented. Their names all begin with
3464 "_pcre_". This is not a perfect job, because (a) we have to except some
3465 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3466 available (and never for static libraries). I have made a note to try to
3467 find a way round (a) in the future.
3470 Version 6.2 01-Aug-05
3471 ---------------------
3473 1. There was no test for integer overflow of quantifier values. A construction
3474 such as {1111111111111111} would give undefined results. What is worse, if
3475 a minimum quantifier for a parenthesized subpattern overflowed and became
3476 negative, the calculation of the memory size went wrong. This could have
3477 led to memory overwriting.
3479 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3481 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3482 operating environments where this matters.
3484 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3485 PCRE options from within the C++ wrapper.
3487 5. Named capturing subpatterns were not being correctly counted when a pattern
3488 was compiled. This caused two problems: (a) If there were more than 100
3489 such subpatterns, the calculation of the memory needed for the whole
3490 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3491 back references of the form \12, where the number was greater than 9, were
3492 not recognized as back references, even though there were sufficient
3493 previous subpatterns.
3495 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3496 versions of gcc, e.g. 2.95.4.
3499 Version 6.1 21-Jun-05
3500 ---------------------
3502 1. There was one reference to the variable "posix" in pcretest.c that was not
3503 surrounded by "#if !defined NOPOSIX".
3505 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3506 the cross-check on the old pcre_info() function, for the benefit of the
3507 cut-down version of PCRE that is currently imported into Exim.
3509 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3510 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3511 patterns that start (?i) but I don't think that matters. The use of (?i) is
3512 just an example; this all applies to the other options as well.
3514 4. Since libtool seems to echo the compile commands it is issuing, the output
3515 from "make" can be reduced a bit by putting "@" in front of each libtool
3516 compile command.
3518 5. Patch from the folks at Google for configure.in to be a bit more thorough
3519 in checking for a suitable C++ installation before trying to compile the
3520 C++ stuff. This should fix a reported problem when a compiler was present,
3521 but no suitable headers.
3523 6. The man pages all had just "PCRE" as their title. I have changed them to
3524 be the relevant file name. I have also arranged that these names are
3525 retained in the file doc/pcre.txt, which is a concatenation in text format
3526 of all the man pages except the little individual ones for each function.
3528 7. The NON-UNIX-USE file had not been updated for the different set of source
3529 files that come with release 6. I also added a few comments about the C++
3530 wrapper.
3533 Version 6.0 07-Jun-05
3534 ---------------------
3536 1. Some minor internal re-organization to help with my DFA experiments.
3538 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3539 didn't matter for the library itself when fully configured, but did matter
3540 when compiling without UCP support, or within Exim, where the ucp files are
3541 not imported.
3543 3. Refactoring of the library code to split up the various functions into
3544 different source modules. The addition of the new DFA matching code (see
3545 below) to a single monolithic source would have made it really too
3546 unwieldy, quite apart from causing all the code to be include in a
3547 statically linked application, when only some functions are used. This is
3548 relevant even without the DFA addition now that patterns can be compiled in
3549 one application and matched in another.
3551 The downside of splitting up is that there have to be some external
3552 functions and data tables that are used internally in different modules of
3553 the library but which are not part of the API. These have all had their
3554 names changed to start with "_pcre_" so that they are unlikely to clash
3555 with other external names.
3557 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3558 a different (DFA) algorithm. Although it is slower than the original
3559 function, it does have some advantages for certain types of matching
3560 problem.
3562 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3563 including restarting after a partial match.
3565 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3566 defined when compiling for Windows was sent to me. I have put it into the
3567 code, though I have no means of testing or verifying it.
3569 7. Added the pcre_refcount() auxiliary function.
3571 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3572 match before or at the first newline in the subject string. In pcretest,
3573 the /f option on a pattern can be used to set this.
3575 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3576 would behave wrongly. This has been present in PCRE since release 4.0.
3578 10. A number of changes to the pcregrep command:
3580 (a) Refactored how -x works; insert ^(...)$ instead of setting
3581 PCRE_ANCHORED and checking the length, in preparation for adding
3582 something similar for -w.
3584 (b) Added the -w (match as a word) option.
3586 (c) Refactored the way lines are read and buffered so as to have more
3587 than one at a time available.
3589 (d) Implemented a pcregrep test script.
3591 (e) Added the -M (multiline match) option. This allows patterns to match
3592 over several lines of the subject. The buffering ensures that at least
3593 8K, or the rest of the document (whichever is the shorter) is available
3594 for matching (and similarly the previous 8K for lookbehind assertions).
3596 (f) Changed the --help output so that it now says
3598 -w, --word-regex(p)
3600 instead of two lines, one with "regex" and the other with "regexp"
3601 because that confused at least one person since the short forms are the
3602 same. (This required a bit of code, as the output is generated
3603 automatically from a table. It wasn't just a text change.)
3605 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3606 option but starts with a hyphen. Could be a pattern or a path name
3607 starting with a hyphen, for instance.
3609 (h) "-" can be given as a file name to represent stdin.
3611 (i) When file names are being printed, "(standard input)" is used for
3612 the standard input, for compatibility with GNU grep. Previously
3613 "<stdin>" was used.
3615 (j) The option --label=xxx can be used to supply a name to be used for
3616 stdin when file names are being printed. There is no short form.
3618 (k) Re-factored the options decoding logic because we are going to add
3619 two more options that take data. Such options can now be given in four
3620 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3622 (l) Added the -A, -B, and -C options for requesting that lines of context
3623 around matches be printed.
3625 (m) Added the -L option to print the names of files that do not contain
3626 any matching lines, that is, the complement of -l.
3628 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3629 continue to scan other files.
3631 (o) The -s option was incorrectly implemented. For compatibility with other
3632 greps, it now suppresses the error message for a non-existent or non-
3633 accessible file (but not the return code). There is a new option called
3634 -q that suppresses the output of matching lines, which was what -s was
3635 previously doing.
3637 (p) Added --include and --exclude options to specify files for inclusion
3638 and exclusion when recursing.
3640 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3641 Hopefully, it now does.
3643 12. Missing cast in pcre_study().
3645 13. Added an "uninstall" target to the makefile.
3647 14. Replaced "extern" in the function prototypes in Makefile.in with
3648 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3649 world, but is set differently for Windows.
3651 15. Added a second compiling function called pcre_compile2(). The only
3652 difference is that it has an extra argument, which is a pointer to an
3653 integer error code. When there is a compile-time failure, this is set
3654 non-zero, in addition to the error test pointer being set to point to an
3655 error message. The new argument may be NULL if no error number is required
3656 (but then you may as well call pcre_compile(), which is now just a
3657 wrapper). This facility is provided because some applications need a
3658 numeric error indication, but it has also enabled me to tidy up the way
3659 compile-time errors are handled in the POSIX wrapper.
3661 16. Added VPATH=.libs to the makefile; this should help when building with one
3662 prefix path and installing with another. (Or so I'm told by someone who
3663 knows more about this stuff than I do.)
3665 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3666 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3667 match everything, including newlines. This is not POSIX-compatible, but
3668 somebody wanted the feature. From pcretest it can be activated by using
3669 both the P and the s flags.
3671 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3673 19. libpcre.pc was being incorrectly installed as executable.
3675 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3676 it now also looks for '\r' so that it will work unmodified on Windows.
3678 21. Added Google's contributed C++ wrapper to the distribution.
3680 22. Added some untidy missing memory free() calls in pcretest, to keep
3681 Electric Fence happy when testing.
3685 Version 5.0 13-Sep-04
3686 ---------------------
3688 1. Internal change: literal characters are no longer packed up into items
3689 containing multiple characters in a single byte-string. Each character
3690 is now matched using a separate opcode. However, there may be more than one
3691 byte in the character in UTF-8 mode.
3693 2. The pcre_callout_block structure has two new fields: pattern_position and
3694 next_item_length. These contain the offset in the pattern to the next match
3695 item, and its length, respectively.
3697 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3698 insertion of callouts before each pattern item. Added the /C option to
3699 pcretest to make use of this.
3701 4. On the advice of a Windows user, the lines
3703 #if defined(_WIN32) || defined(WIN32)
3704 _setmode( _fileno( stdout ), 0x8000 );
3705 #endif /* defined(_WIN32) || defined(WIN32) */
3707 have been added to the source of pcretest. This apparently does useful
3708 magic in relation to line terminators.
3710 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3711 for the benefit of those environments where the "b" makes a difference.
3713 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3714 to know about it. I have put a hack into configure.in that adds in code
3715 to set GCC=yes if CC=icc. This seems to end up at a point in the
3716 generated configure script that is early enough to affect the setting of
3717 compiler options, which is what is needed, but I have no means of testing
3718 whether it really works. (The user who reported this had patched the
3719 generated configure script, which of course I cannot do.)
3721 LATER: After change 22 below (new libtool files), the configure script
3722 seems to know about icc (and also ecc). Therefore, I have commented out
3723 this hack in configure.in.
3725 7. Added support for pkg-config (2 patches were sent in).
3727 8. Negated POSIX character classes that used a combination of internal tables
3728 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3729 [[:^ascii]]. Typically, they would match almost any characters. The other
3730 POSIX classes were not broken in this way.
3732 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3733 to find the match, as PCRE was deluded into thinking that the match had to
3734 start at the start point or following a newline. The same bug applied to
3735 patterns with negative forward assertions or any backward assertions
3736 preceding ".*" at the start, unless the pattern required a fixed first
3737 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3739 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3740 starting at the last subject character, bytes beyond the end of the subject
3741 string were read.
3743 11. Renamed the variable "class" as "classbits" to make life easier for C++
3744 users. (Previously there was a macro definition, but it apparently wasn't
3745 enough.)
3747 12. Added the new field "tables" to the extra data so that tables can be passed
3748 in at exec time, or the internal tables can be re-selected. This allows
3749 a compiled regex to be saved and re-used at a later time by a different
3750 program that might have everything at different addresses.
3752 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3753 -R library as well as a -L library.
3755 14. The debugging options of pcretest (-d on the command line or D on a
3756 pattern) showed incorrect output for anything following an extended class
3757 that contained multibyte characters and which was followed by a quantifier.
3759 15. Added optional support for general category Unicode character properties
3760 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3761 support. It adds about 90K to the size of the library. The meanings of the
3762 inbuilt class escapes such as \d and \s have NOT been changed.
3764 16. Updated pcredemo.c to include calls to free() to release the memory for the
3765 compiled pattern.
3767 17. The generated file chartables.c was being created in the source directory
3768 instead of in the building directory. This caused the build to fail if the
3769 source directory was different from the building directory, and was
3770 read-only.
3772 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3773 file. No doubt somebody will tell me if they don't make sense... Also added
3774 Dan Mooney's comments about building on OpenVMS.
3776 19. Added support for partial matching via the PCRE_PARTIAL option for
3777 pcre_exec() and the \P data escape in pcretest.
3779 20. Extended pcretest with 3 new pattern features:
3781 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3782 write the compiled pattern to the file whose name is "rest-of-line".
3783 This is a straight binary dump of the data, with the saved pointer to
3784 the character tables forced to be NULL. The study data, if any, is
3785 written too. After writing, pcretest reads a new pattern.
3787 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3788 compiled pattern from the given file. There must not be any
3789 occurrences of "<" in the file name (pretty unlikely); if there are,
3790 pcretest will instead treat the initial "<" as a pattern delimiter.
3791 After reading in the pattern, pcretest goes on to read data lines as
3792 usual.
3794 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3795 and 16-bit fields in a compiled pattern, to simulate a pattern that
3796 was compiled on a host of opposite endianness.
3798 21. The pcre-exec() function can now cope with patterns that were compiled on
3799 hosts of opposite endianness, with this restriction:
3801 As for any compiled expression that is saved and used later, the tables
3802 pointer field cannot be preserved; the extra_data field in the arguments
3803 to pcre_exec() should be used to pass in a tables address if a value
3804 other than the default internal tables were used at compile time.
3806 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3807 now diagnosed as an error. Previously, most of the time, a negative number
3808 would have been treated as zero, but if in addition "ovector" was passed as
3809 NULL, a crash could occur.
3811 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3812 new versions from the libtool 1.5 distribution (the last one is a copy of
3813 a file called libtool.m4). This seems to have fixed the need to patch
3814 "configure" to support Darwin 1.3 (which I used to do). However, I still
3815 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3816 workstation).
3818 24. Changed the PCRE licence to be the more standard "BSD" licence.
3821 Version 4.5 01-Dec-03
3822 ---------------------
3824 1. There has been some re-arrangement of the code for the match() function so
3825 that it can be compiled in a version that does not call itself recursively.
3826 Instead, it keeps those local variables that need separate instances for
3827 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3828 needs to "recurse". Keeping track of where control must go is done by means
3829 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3830 hide most of the details from the main code, and operates only if
3831 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3832 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3833 operating.
3835 To make it easier for callers to provide specially tailored get/free
3836 functions for this usage, two new functions, pcre_stack_malloc, and
3837 pcre_stack_free, are used. They are always called in strict stacking order,
3838 and the size of block requested is always the same.
3840 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3841 PCRE has been compiled to use the stack or the heap for recursion. The
3842 -C option of pcretest uses this to show which version is compiled.
3844 A new data escape \S, is added to pcretest; it causes the amounts of store
3845 obtained and freed by both kinds of malloc/free at match time to be added
3846 to the output.
3848 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3849 what's available on my current Linux desktop machine.
3851 3. When matching a UTF-8 string, the test for a valid string at the start has
3852 been extended. If start_offset is not zero, PCRE now checks that it points
3853 to a byte that is the start of a UTF-8 character. If not, it returns
3854 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3855 this is necessary because there may be backward assertions in the pattern.
3856 When matching the same subject several times, it may save resources to use
3857 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3859 4. The code for checking the validity of UTF-8 strings has been tightened so
3860 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3861 containing "overlong sequences".
3863 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3864 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3865 should have been "&", but it just so happened that all the cases this let
3866 through by mistake were picked up later in the function.
3868 6. I had used a variable called "isblank" - this is a C99 function, causing
3869 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3871 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3872 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3873 so the version is shown; (c) stop "make test" echoing "./RunTest".
3875 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3877 9. The prototype for memmove() for systems that don't have it was using
3878 size_t, but the inclusion of the header that defines size_t was later. I've
3879 moved the #includes for the C headers earlier to avoid this.
3881 10. Added some adjustments to the code to make it easier to compiler on certain
3882 special systems:
3884 (a) Some "const" qualifiers were missing.
3885 (b) Added the macro EXPORT before all exported functions; by default this
3886 is defined to be empty.
3887 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3888 that it reads its output file name as an argument instead of writing
3889 to the standard output and assuming this can be redirected.
3891 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3892 class containing characters with values greater than 255, PCRE compilation
3893 went into a loop.
3895 12. A recursive reference to a subpattern that was within another subpattern
3896 that had a minimum quantifier of zero caused PCRE to crash. For example,
3897 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3898 recursion. If the recursively-called subpattern itself had a zero repeat,
3899 that was OK.
3901 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3902 buffer into which it was copied (for escape processing) was still set at
3903 1024, so long lines caused crashes.
3905 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3906 "internal error: code overflow...". This applied to any character class
3907 that was followed by a possessive quantifier.
3909 15. Modified the Makefile to add libpcre.la as a prerequisite for
3910 libpcreposix.la because I was told this is needed for a parallel build to
3911 work.
3913 16. If a pattern that contained .* following optional items at the start was
3914 studied, the wrong optimizing data was generated, leading to matching
3915 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3916 matching string must start with a or b or c. The correct conclusion for
3917 this pattern is that a match can start with any character.
3920 Version 4.4 13-Aug-03
3921 ---------------------
3923 1. In UTF-8 mode, a character class containing characters with values between
3924 127 and 255 was not handled correctly if the compiled pattern was studied.
3925 In fixing this, I have also improved the studying algorithm for such
3926 classes (slightly).
3928 2. Three internal functions had redundant arguments passed to them. Removal
3929 might give a very teeny performance improvement.
3931 3. Documentation bug: the value of the capture_top field in a callout is *one
3932 more than* the number of the hightest numbered captured substring.
3934 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3935 in incorrectly linking with a previously installed version. They now link
3936 explicitly with libpcre.la.
3938 5. configure.in no longer needs to recognize Cygwin specially.
3940 6. A problem in pcre.in for Windows platforms is fixed.
3942 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3943 pcretest, it used to include the size of the study block as part of its
3944 output. Unfortunately, the structure contains a field that has a different
3945 size on different hardware architectures. This meant that the tests that
3946 showed this size failed. As the block is currently always of a fixed size,
3947 this information isn't actually particularly useful in pcretest output, so
3948 I have just removed it.
3950 8. Three pre-processor statements accidentally did not start in column 1.
3951 Sadly, there are *still* compilers around that complain, even though
3952 standard C has not required this for well over a decade. Sigh.
3954 9. In pcretest, the code for checking callouts passed small integers in the
3955 callout_data field, which is a void * field. However, some picky compilers
3956 complained about the casts involved for this on 64-bit systems. Now
3957 pcretest passes the address of the small integer instead, which should get
3958 rid of the warnings.
3960 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3961 both compile and run time, and gives an error if an invalid UTF-8 sequence
3962 is found. There is a option for disabling this check in cases where the
3963 string is known to be correct and/or the maximum performance is wanted.
3965 11. In response to a bug report, I changed one line in Makefile.in from
3967 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3968 to
3969 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3971 to look similar to other lines, but I have no way of telling whether this
3972 is the right thing to do, as I do not use Windows. No doubt I'll get told
3973 if it's wrong...
3976 Version 4.3 21-May-03
3977 ---------------------
3979 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3980 Makefile.
3982 2. Some refactoring to improve the quality of the code:
3984 (i) The utf8_table... variables are now declared "const".
3986 (ii) The code for \cx, which used the "case flipping" table to upper case
3987 lower case letters, now just substracts 32. This is ASCII-specific,
3988 but the whole concept of \cx is ASCII-specific, so it seems
3989 reasonable.
3991 (iii) PCRE was using its character types table to recognize decimal and
3992 hexadecimal digits in the pattern. This is silly, because it handles
3993 only 0-9, a-f, and A-F, but the character types table is locale-
3994 specific, which means strange things might happen. A private
3995 table is now used for this - though it costs 256 bytes, a table is
3996 much faster than multiple explicit tests. Of course, the standard
3997 character types table is still used for matching digits in subject
3998 strings against \d.
4000 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4001 ending in _t are). So I've renamed it as ESC_tee.
4003 3. The first argument for regexec() in the POSIX wrapper should have been
4004 defined as "const".
4006 4. Changed pcretest to use malloc() for its buffers so that they can be
4007 Electric Fenced for debugging.
4009 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4010 to read one or more bytes before the start of the subject string. Often this
4011 had no effect on PCRE's behaviour, but in some circumstances it could
4012 provoke a segmentation fault.
4014 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4015 to try to read one or more bytes before the start of the subject string.
4017 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4018 UTF-8 support could misbehave in various ways if the subject string
4019 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4020 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4021 back over UTF-8 characters.)
4024 Version 4.2 14-Apr-03
4025 ---------------------
4027 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4029 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4030 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4031 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4032 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4033 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4035 Note: automatic setting of the BUILD variables is not yet working
4036 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4037 compile-time but not at link-time
4038 [LINK]: use for linking executables only
4039 make different versions for Windows and non-Windows
4040 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4041 libraries
4042 [LINK_FOR_BUILD]: new variable
4043 [OBJEXT]: use throughout
4044 [EXEEXT]: use throughout
4045 <winshared>: new target
4046 <wininstall>: new target
4047 <dftables.o>: use native compiler
4048 <dftables>: use native linker
4049 <install>: handle Windows platform correctly
4050 <clean>: ditto
4051 <check>: ditto
4052 copy DLL to top builddir before testing
4054 As part of these changes, -no-undefined was removed again. This was reported
4055 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4056 in any case.
4058 3. Some tidies to get rid of compiler warnings:
4060 . In the match_data structure, match_limit was an unsigned long int, whereas
4061 match_call_count was an int. I've made them both unsigned long ints.
4063 . In pcretest the fact that a const uschar * doesn't automatically cast to
4064 a void * provoked a warning.
4066 . Turning on some more compiler warnings threw up some "shadow" variables
4067 and a few more missing casts.
4069 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4070 option, a class that contained a single character with a value between 128
4071 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4073 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4074 option, a class that contained several characters, but with at least one
4075 whose value was between 128 and 255 caused PCRE to crash.
4078 Version 4.1 12-Mar-03
4079 ---------------------
4081 1. Compiling with gcc -pedantic found a couple of places where casts were
4082 needed, and a string in dftables.c that was longer than standard compilers are
4083 required to support.
4085 2. Compiling with Sun's compiler found a few more places where the code could
4086 be tidied up in order to avoid warnings.
4088 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4089 first of these names is deprecated in the latest Autoconf in favour of the name
4090 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4091 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4092 analogy I have changed it to CFLAGS_FOR_BUILD.
4094 4. Added -no-undefined to the linking command in the Makefile, because this is
4095 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4096 linking step for the pcreposix library.
4098 5. PCRE was failing to diagnose the case of two named groups with the same
4099 name.
4101 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4102 literal character that is needed in the subject for a match, and scans along to
4103 ensure that it is present before embarking on the full matching process. This
4104 saves time in cases of nested unlimited repeats that are never going to match.
4105 Problem: the scan can take a lot of time if the subject is very long (e.g.
4106 megabytes), thus penalizing straightforward matches. It is now done only if the
4107 amount of subject to be scanned is less than 1000 bytes.
4109 7. A lesser problem with the same optimization is that it was recording the
4110 first character of an anchored pattern as "needed", thus provoking a search
4111 right along the subject, even when the first match of the pattern was going to
4112 fail. The "needed" character is now not set for anchored patterns, unless it
4113 follows something in the pattern that is of non-fixed length. Thus, it still
4114 fulfils its original purpose of finding quick non-matches in cases of nested
4115 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4118 Version 4.0 17-Feb-03
4119 ---------------------
4121 1. If a comment in an extended regex that started immediately after a meta-item
4122 extended to the end of string, PCRE compiled incorrect data. This could lead to
4123 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4125 2. Moved to autoconf 2.53 and libtool 1.4.2.
4127 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4128 the special perltest8 script is no longer needed - all the tests can be run
4129 from a single perltest script.
4131 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4132 by \s. It has now been removed in PCRE. This means it isn't recognized as
4133 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4134 class [:space:] *does* include VT, thereby creating a mess.
4136 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4137 space and tab.
4139 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4140 its new features into the main test script, reducing the number of scripts.
4142 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4143 were backward compatible, and made the (?i) apply to the whole pattern, as if
4144 /i were given. Now it behaves more logically, and applies the option setting
4145 only to what follows. PCRE has been changed to follow suit. However, if it
4146 finds options settings right at the start of the pattern, it extracts them into
4147 the global options, as before. Thus, they show up in the info data.
4149 8. Added support for the \Q...\E escape sequence. Characters in between are
4150 treated as literals. This is slightly different from Perl in that $ and @ are
4151 also handled as literals inside the quotes. In Perl, they will cause variable
4152 interpolation. Note the following examples:
4154 Pattern PCRE matches Perl matches
4156 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4157 \Qabc\$xyz\E abc\$xyz abc\$xyz
4158 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4160 For compatibility with Perl, \Q...\E sequences are recognized inside character
4161 classes as well as outside them.
4163 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4164 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4165 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4166 signed/unsigned warnings.
4168 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4169 option for pcretest, so I've replaced it by a simple function that does just
4170 that job.
4172 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4173 "pcregrep -".
4175 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4176 Java package. This provides some syntactic sugar for simple cases of what my
4177 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4178 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4179 item, you can use this simplified notation. Note that only makes sense with
4180 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4181 greediness, whatever the setting of the PCRE_UNGREEDY option.
4183 13. A change of greediness default within a pattern was not taking effect at
4184 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4185 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4186 was abstracted outside.
4188 14. PCRE now supports the \G assertion. It is true when the current matching
4189 position is at the start point of the match. This differs from \A when the
4190 starting offset is non-zero. Used with the /g option of pcretest (or similar
4191 code), it works in the same way as it does for Perl's /g option. If all
4192 alternatives of a regex begin with \G, the expression is anchored to the start
4193 match position, and the "anchored" flag is set in the compiled expression.
4195 15. Some bugs concerning the handling of certain option changes within patterns
4196 have been fixed. These applied to options other than (?ims). For example,
4197 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4198 been the other way round. Some of this was related to change 7 above.
4200 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4201 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4202 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4203 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4205 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4206 mode. Unlike ".", it always matches newline, whatever the setting of
4207 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4208 assertions. Perl allows it, but it doesn't (in general) work because it can't
4209 calculate the length of the lookbehind. At least, that's the case for Perl
4210 5.8.0 - I've been told they are going to document that it doesn't work in
4211 future.
4213 18. Added an error diagnosis for escapes that PCRE does not support: these are
4214 \L, \l, \N, \P, \p, \U, \u, and \X.
4216 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4217 reading past the end of the pattern in cases such as /[abcd/.
4219 20. PCRE was getting more memory than necessary for patterns with classes that
4220 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4222 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4223 compiling PCRE for use with Virtual Pascal.
4225 22. Small fix to the Makefile to make it work properly if the build is done
4226 outside the source tree.
4228 23. Added a new extension: a condition to go with recursion. If a conditional
4229 subpattern starts with (?(R) the "true" branch is used if recursion has
4230 happened, whereas the "false" branch is used only at the top level.
4232 24. When there was a very long string of literal characters (over 255 bytes
4233 without UTF support, over 250 bytes with UTF support), the computation of how
4234 much memory was required could be incorrect, leading to segfaults or other
4235 strange effects.
4237 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4238 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4239 there was a subsequent back reference to those brackets. This meant that, for
4240 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4241 possible to check for precisely this case. All we can do is abandon the
4242 optimization if .* occurs inside capturing brackets when there are any back
4243 references whatsoever. (See below for a better fix that came later.)
4245 26. The handling of the optimization for finding the first character of a
4246 non-anchored pattern, and for finding a character that is required later in the
4247 match were failing in some cases. This didn't break the matching; it just
4248 failed to optimize when it could. The way this is done has been re-implemented.
4250 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4252 28. Added a new feature that provides some of the functionality that Perl
4253 provides with (?{...}). The facility is termed a "callout". The way it is done
4254 in PCRE is for the caller to provide an optional function, by setting
4255 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4256 global variable. By default it is unset, which disables all calling out. To get
4257 the function called, the regex must include (?C) at appropriate points. This
4258 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4259 This provides a means of identifying different callout points. When PCRE
4260 reaches such a point in the regex, if pcre_callout has been set, the external
4261 function is called. It is provided with data in a structure called
4262 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4263 matching continues; if it returns a non-zero value, the match at the current
4264 point fails. However, backtracking will occur if possible. [This was changed
4265 later and other features added - see item 49 below.]
4267 29. pcretest is upgraded to test the callout functionality. It provides a
4268 callout function that displays information. By default, it shows the start of
4269 the match and the current position in the text. There are some new data escapes
4270 to vary what happens:
4272 \C+ in addition, show current contents of captured substrings
4273 \C- do not supply a callout function
4274 \C!n return 1 when callout number n is reached
4275 \C!n!m return 1 when callout number n is reached for the mth time
4277 30. If pcregrep was called with the -l option and just a single file name, it
4278 output "<stdin>" if a match was found, instead of the file name.
4280 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4281 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4282 pcre_exec(). This saves a malloc/free per call. The default value of
4283 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4284 when configuring.
4286 32. The default maximum size of a compiled pattern is 64K. There have been a
4287 few cases of people hitting this limit. The code now uses macros to handle the
4288 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4289 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4290 configuring. Tests 2 and 5 work only with 2-byte links because they output
4291 debugging information about compiled patterns.
4293 33. Internal code re-arrangements:
4295 (a) Moved the debugging function for printing out a compiled regex into
4296 its own source file (printint.c) and used #include to pull it into
4297 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4298 separate copies.
4300 (b) Defined the list of op-code names for debugging as a macro in
4301 internal.h so that it is next to the definition of the opcodes.
4303 (c) Defined a table of op-code lengths for simpler skipping along compiled
4304 code. This is again a macro in internal.h so that it is next to the
4305 definition of the opcodes.
4307 34. Added support for recursive calls to individual subpatterns, along the
4308 lines of Robin Houston's patch (but implemented somewhat differently).
4310 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4311 allow it to read and process whole directories in Win32. This code was
4312 contributed by Lionel Fourquaux; it has not been tested by me.
4314 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4315 used to name a group. Names consist of alphanumerics and underscores, and must
4316 be unique. Back references use the syntax (?P=name) and recursive calls use
4317 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4318 numbers. The function pcre_fullinfo() can be used after compilation to extract
4319 a name/number map. There are three relevant calls:
4321 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4322 PCRE_INFO_NAMECOUNT yields the number of entries
4323 PCRE_INFO_NAMETABLE yields a pointer to the map.
4325 The map is a vector of fixed-size entries. The size of each entry depends on
4326 the length of the longest name used. The first two bytes of each entry are the
4327 group number, most significant byte first. There follows the corresponding
4328 name, zero terminated. The names are in alphabetical order.
4330 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4331 case instead of 255. Making it the same both with and without UTF-8 support
4332 means that the same test output works with both.
4334 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4335 calling malloc() with a zero argument.
4337 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4338 optimization. I've improved things by keeping a bitmap of backreferences with
4339 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4340 fact referenced, the optimization can be applied. It is unlikely that a
4341 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4342 the match to follow \n) will appear inside brackets with a number greater than
4343 31, but if it does, any back reference > 31 suppresses the optimization.
4345 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4346 of disabling numbered capturing parentheses. Any opening parenthesis that is
4347 not followed by ? behaves as if it were followed by ?: but named parentheses
4348 can still be used for capturing (and they will acquire numbers in the usual
4349 way).
4351 41. Redesigned the return codes from the match() function into yes/no/error so
4352 that errors can be passed back from deep inside the nested calls. A malloc
4353 failure while inside a recursive subpattern call now causes the
4354 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4356 42. It is now possible to set a limit on the number of times the match()
4357 function is called in a call to pcre_exec(). This facility makes it possible to
4358 limit the amount of recursion and backtracking, though not in a directly
4359 obvious way, because the match() function is used in a number of different
4360 circumstances. The count starts from zero for each position in the subject
4361 string (for non-anchored patterns). The default limit is, for compatibility, a
4362 large number, namely 10 000 000. You can change this in two ways:
4364 (a) When configuring PCRE before making, you can use --with-match-limit=n
4365 to set a default value for the compiled library.
4367 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4368 a different value is set. See 45 below.
4370 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4372 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4373 of things that can be changed at compile time. The first argument specifies
4374 what is wanted and the second points to where the information is to be placed.
4375 The current list of available information is:
4379 The output is an integer that is set to one if UTF-8 support is available;
4380 otherwise it is set to zero.
4384 The output is an integer that it set to the value of the code that is used for
4385 newline. It is either LF (10) or CR (13).
4389 The output is an integer that contains the number of bytes used for internal
4390 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4394 The output is an integer that contains the threshold above which the POSIX
4395 interface uses malloc() for output vectors. See item 31 above.
4399 The output is an unsigned integer that contains the default limit of the number
4400 of match() calls in a pcre_exec() execution. See 42 above.
4402 44. pcretest has been upgraded by the addition of the -C option. This causes it
4403 to extract all the available output from the new pcre_config() function, and to
4404 output it. The program then exits immediately.
4406 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4407 order to support additional features. One way would have been to define
4408 pcre_exec2() (for example) with extra arguments, but this would not have been
4409 extensible, and would also have required all calls to the original function to
4410 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4411 is used for passing in "extra" data from pcre_study().
4413 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4414 contains the following fields:
4416 flags a bitmap indicating which of the following fields are set
4417 study_data opaque data from pcre_study()
4418 match_limit a way of specifying a limit on match() calls for a specific
4419 call to pcre_exec()
4420 callout_data data for callouts (see 49 below)
4422 The flag bits are also defined in pcre.h, and are
4428 The pcre_study() function now returns one of these new pcre_extra blocks, with
4429 the actual study data pointed to by the study_data field, and the
4430 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4431 before. That is, this change is entirely upwards-compatible and requires no
4432 change to existing code.
4434 If you want to pass in additional data to pcre_exec(), you can either place it
4435 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4436 block.
4438 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4439 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4440 times with different match limits, until it finds the minimum value needed for
4441 pcre_exec() to complete. The value is then output. This can be instructive; for
4442 most simple matches the number is quite small, but for pathological cases it
4443 gets very large very quickly.
4445 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4446 returns the size of the data block pointed to by the study_data field in a
4447 pcre_extra block, that is, the value that was passed as the argument to
4448 pcre_malloc() when PCRE was getting memory in which to place the information
4449 created by pcre_study(). The fourth argument should point to a size_t variable.
4450 pcretest has been extended so that this information is shown after a successful
4451 pcre_study() call when information about the compiled regex is being displayed.
4453 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4454 because what follows is always an absolute path. (Later: it turns out that this
4455 is more than cosmetic for MinGW, because it doesn't like empty path
4456 components.)
4458 49. Some changes have been made to the callout feature (see 28 above):
4460 (i) A callout function now has three choices for what it returns:
4462 0 => success, carry on matching
4463 > 0 => failure at this point, but backtrack if possible
4464 < 0 => serious error, return this value from pcre_exec()
4466 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4467 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4468 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4469 use by callout functions. It will never be used by PCRE itself.
4471 (ii) The pcre_extra structure (see 45 above) has a void * field called
4472 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4473 pcre_callout_block structure has a field of the same name. The contents of
4474 the field passed in the pcre_extra structure are passed to the callout
4475 function in the corresponding field in the callout block. This makes it
4476 easier to use the same callout-containing regex from multiple threads. For
4477 testing, the pcretest program has a new data escape
4479 \C*n pass the number n (may be negative) as callout_data
4481 If the callout function in pcretest receives a non-zero value as
4482 callout_data, it returns that value.
4484 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4485 there were some redundant $(CFLAGS) in commands that are now specified as
4486 $(LINK), which already includes $(CFLAGS).
4488 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4489 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4490 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4491 one-byte characters throughout. Note that case-insensitive matching applies
4492 only to characters whose values are less than 256. PCRE doesn't support the
4493 notion of cases for higher-valued characters.
4495 (i) A character class whose characters are all within 0-255 is handled as
4496 a bit map, and the map is inverted for negative classes. Previously, a
4497 character > 255 always failed to match such a class; however it should
4498 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4500 (ii) A negated character class with a single character < 255 is coded as
4501 "not this character" (OP_NOT). This wasn't working properly when the test
4502 character was multibyte, either singly or repeated.
4504 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4505 mode, for example: \x{100}{2,3}.
4507 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4508 singly or repeated) now correctly test multibyte characters. However,
4509 PCRE doesn't recognize any characters with values greater than 255 as
4510 digits, spaces, or word characters. Such characters always match \D, \S,
4511 and \W, and never match \d, \s, or \w.
4513 (v) Classes may now contain characters and character ranges with values
4514 greater than 255. For example: [ab\x{100}-\x{400}].
4516 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4517 PCRE in UTF-8 mode.
4519 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4520 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4521 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4522 value.)
4524 53. The single man page has become too large. I have therefore split it up into
4525 a number of separate man pages. These also give rise to individual HTML pages;
4526 these are now put in a separate directory, and there is an index.html page that
4527 lists them all. Some hyperlinking between the pages has been installed.
4529 54. Added convenience functions for handling named capturing parentheses.
4531 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4532 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4533 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4534 are faulted.
4536 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4537 calling configure. These values are used when compiling the dftables.c program
4538 which is run to generate the source of the default character tables. They
4539 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4540 you will need to set these values.
4542 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4545 Version 3.9 02-Jan-02
4546 ---------------------
4548 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4550 2. If --disable-static was given, the building process failed when trying to
4551 build pcretest and pcregrep. (For some reason it was using libtool to compile
4552 them, which is not right, as they aren't part of the library.)
4555 Version 3.8 18-Dec-01
4556 ---------------------
4558 1. The experimental UTF-8 code was completely screwed up. It was packing the
4559 bytes in the wrong order. How dumb can you get?
4562 Version 3.7 29-Oct-01
4563 ---------------------
4565 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4566 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4567 this didn't happen under Solaris 8, where I normally test things.
4569 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4570 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4571 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4572 this fix an hour or so after the initial 3.7 release.)
4575 Version 3.6 23-Oct-01
4576 ---------------------
4578 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4579 offsets passed as NULL with zero offset count.
4581 2. The config.guess and config.sub files had not been updated when I moved to
4582 the latest autoconf.
4585 Version 3.5 15-Aug-01
4586 ---------------------
4588 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4589 had been forgotten.
4591 2. By using declared but undefined structures, we can avoid using "void"
4592 definitions in pcre.h while keeping the internal definitions of the structures
4593 private.
4595 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4596 user point of view, this means that both static and shared libraries are built
4597 by default, but this can be individually controlled. More of the work of
4598 handling this static/shared cases is now inside libtool instead of PCRE's make
4599 file.
4601 4. The pcretest utility is now installed along with pcregrep because it is
4602 useful for users (to test regexs) and by doing this, it automatically gets
4603 relinked by libtool. The documentation has been turned into a man page, so
4604 there are now .1, .txt, and .html versions in /doc.
4606 5. Upgrades to pcregrep:
4607 (i) Added long-form option names like gnu grep.
4608 (ii) Added --help to list all options with an explanatory phrase.
4609 (iii) Added -r, --recursive to recurse into sub-directories.
4610 (iv) Added -f, --file to read patterns from a file.
4612 6. pcre_exec() was referring to its "code" argument before testing that
4613 argument for NULL (and giving an error if it was NULL).
4615 7. Upgraded Makefile.in to allow for compiling in a different directory from
4616 the source directory.
4618 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4619 options bits, the pointer it was passed was to an int instead of to an unsigned
4620 long int. This mattered only on 64-bit systems.
4622 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4623 generated) instead of pcre.in, which it its source. Also made the same change
4624 in several of the .c files.
4626 10. A new release of gcc defines printf() as a macro, which broke pcretest
4627 because it had an ifdef in the middle of a string argument for printf(). Fixed
4628 by using separate calls to printf().
4630 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4631 script, to force use of CR or LF instead of \n in the source. On non-Unix
4632 systems, the value can be set in config.h.
4634 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4635 absolute limit. Changed the text of the error message to make this clear, and
4636 likewise updated the man page.
4638 13. The limit of 99 on the number of capturing subpatterns has been removed.
4639 The new limit is 65535, which I hope will not be a "real" limit.
4642 Version 3.4 22-Aug-00
4643 ---------------------
4645 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4647 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4650 Version 3.3 01-Aug-00
4651 ---------------------
4653 1. If an octal character was given, but the value was greater than \377, it
4654 was not getting masked to the least significant bits, as documented. This could
4655 lead to crashes in some systems.
4657 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4658 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4660 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4661 These just pass their arguments on to (pcre_free)(), but they are provided
4662 because some uses of PCRE bind it to non-C systems that can call its functions,
4663 but cannot call free() or pcre_free() directly.
4665 4. Add "make test" as a synonym for "make check". Corrected some comments in
4666 the Makefile.
4668 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4669 Makefile.
4671 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4672 command called pgrep for grepping around the active processes.
4674 7. Added the beginnings of support for UTF-8 character strings.
4676 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4677 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4678 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4679 out for the ar command.)
4682 Version 3.2 12-May-00
4683 ---------------------
4685 This is purely a bug fixing release.
4687 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4688 of ZA. This was just one example of several cases that could provoke this bug,
4689 which was introduced by change 9 of version 2.00. The code for breaking
4690 infinite loops after an iteration that matches an empty string was't working
4691 correctly.
4693 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4694 when matched against abbab (for example). After matching an empty string, it
4695 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4696 caused it to match further down the string than it should.
4698 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4699 was there because it doesn't seem to be needed, and it causes trouble on some
4700 systems, as it is not a Standard C header. It has been removed.
4702 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4703 were reported on the Macintosh. The changes were from
4705 while ((c = *(++ptr)) != 0 && c != '\n');
4706 to
4707 while ((c = *(++ptr)) != 0 && c != '\n') ;
4709 Totally extraordinary, but if that's what it takes...
4711 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4712 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4713 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4714 assumes the way PCRE uses memmove() (always moving upwards).
4716 6. PCRE is being used in one environment where strchr() is not available. There
4717 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4718 faster code anyway.
4721 Version 3.1 09-Feb-00
4722 ---------------------
4724 The only change in this release is the fixing of some bugs in Makefile.in for
4725 the "install" target:
4727 (1) It was failing to install pcreposix.h.
4729 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4732 Version 3.0 01-Feb-00
4733 ---------------------
4735 1. Add support for the /+ modifier to perltest (to output $` like it does in
4736 pcretest).
4738 2. Add support for the /g modifier to perltest.
4740 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4741 matches null strings.
4743 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4744 pattern. Perl treats empty patterns specially - it reuses the most recent
4745 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4746 effect.
4748 5. The POSIX interface was broken in that it was just handing over the POSIX
4749 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4750 required a bigger vector, with some working space on the end. This means that
4751 the POSIX wrapper now has to get and free some memory, and copy the results.
4753 6. Added some simple autoconf support, placing the test data and the
4754 documentation in separate directories, re-organizing some of the
4755 information files, and making it build pcre-config (a GNU standard). Also added
4756 libtool support for building PCRE as a shared library, which is now the
4757 default.
4759 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4760 09 are not valid octal constants. Single digits will be used for minor values
4761 less than 10.
4763 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4764 existing programs that set these in the POSIX interface can use PCRE without
4765 modification.
4767 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4768 return all that pcre_info() returns, plus additional data. The pcre_info()
4769 function is retained for compatibility, but is considered to be obsolete.
4771 10. Added experimental recursion feature (?R) to handle one common case that
4772 Perl 5.6 will be able to do with (?p{...}).
4774 11. Added support for POSIX character classes like [:alpha:], which Perl is
4775 adopting.
4778 Version 2.08 31-Aug-99
4779 ----------------------
4781 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4782 trying to match at the startoffset position, but instead was moving forward to
4783 the next newline as if a previous match had failed.
4785 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4786 and could get into a loop if a null string was matched other than at the start
4787 of the subject.
4789 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4790 be distinguished at compile time, and for completeness also added PCRE_DATE.
4792 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4793 in GnuWin32 environments.
4796 Version 2.07 29-Jul-99
4797 ----------------------
4799 1. The documentation is now supplied in plain text form and HTML as well as in
4800 the form of man page sources.
4802 2. C++ compilers don't like assigning (void *) values to other pointer types.
4803 In particular this affects malloc(). Although there is no problem in Standard
4804 C, I've put in casts to keep C++ compilers happy.
4806 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4807 should be (const char *).
4809 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4810 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4811 However, I haven't made this a standard facility. The documentation doesn't
4812 mention it, and the Makefile doesn't support it.
4814 5. The Makefile now contains an "install" target, with editable destinations at
4815 the top of the file. The pcretest program is not installed.
4817 6. pgrep -V now gives the PCRE version number and date.
4819 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4820 causing the entire string to be ignored, instead of just the last character.
4822 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4823 non-matching string, it can take a very, very long time, even for strings of
4824 quite modest length, because of the nested recursion. PCRE now does better in
4825 some of these cases. It does this by remembering the last required literal
4826 character in the pattern, and pre-searching the subject to ensure it is present
4827 before running the real match. In other words, it applies a heuristic to detect
4828 some types of certain failure quickly, and in the above example, if presented
4829 with a string that has no trailing " it gives "no match" very quickly.
4831 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4832 other alternatives are tried instead.
4835 Version 2.06 09-Jun-99
4836 ----------------------
4838 1. Change pcretest's output for amount of store used to show just the code
4839 space, because the remainder (the data block) varies in size between 32-bit and
4840 64-bit systems.
4842 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4843 start matching at. This allows lookbehinds to work when searching for multiple
4844 occurrences in a string.
4846 3. Added additional options to pcretest for testing multiple occurrences:
4848 /+ outputs the rest of the string that follows a match
4849 /g loops for multiple occurrences, using the new startoffset argument
4850 /G loops for multiple occurrences by passing an incremented pointer
4852 4. PCRE wasn't doing the "first character" optimization for patterns starting
4853 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4854 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4855 the letter 'x'. On long subject strings, this gives a significant speed-up.
4858 Version 2.05 21-Apr-99
4859 ----------------------
4861 1. Changed the type of magic_number from int to long int so that it works
4862 properly on 16-bit systems.
4864 2. Fixed a bug which caused patterns starting with .* not to work correctly
4865 when the subject string contained newline characters. PCRE was assuming
4866 anchoring for such patterns in all cases, which is not correct because .* will
4867 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4868 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4869 must be retried after every newline in the subject.
4872 Version 2.04 18-Feb-99
4873 ----------------------
4875 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4876 computation of the store needed to hold the pattern was incorrect (too large).
4877 If such patterns were nested a few deep, this could multiply and become a real
4878 problem.
4880 2. Added /M option to pcretest to show the memory requirement of a specific
4881 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4883 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4884 compiled in such a way that the backtracking after subsequent failure was
4885 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4886 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4889 Version 2.03 02-Feb-99
4890 ----------------------
4892 1. Fixed typo and small mistake in man page.
4894 2. Added 4th condition (GPL supersedes if conflict) and created separate
4895 LICENCE file containing the conditions.
4897 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4898 Perl, that is the internal \ allows the delimiter to be included in the
4899 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4900 the final delimiter, add \ to the end of the pattern (to test the error).
4902 4. Added the convenience functions for extracting substrings after a successful
4903 match. Updated pcretest to make it able to test these functions.
4906 Version 2.02 14-Jan-99
4907 ----------------------
4909 1. Initialized the working variables associated with each extraction so that
4910 their saving and restoring doesn't refer to uninitialized store.
4912 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4913 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4914 fix the problem.
4916 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4917 calls, and wasn't printing the correct value for compiling calls. Increased the
4918 default value of LOOPREPEAT, and the number of significant figures in the
4919 times.
4921 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4923 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4924 a building problem on Windows NT with a FAT file system.
4927 Version 2.01 21-Oct-98
4928 ----------------------
4930 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4931 to character tables built by pcre_maketables() in the current locale. If NULL
4932 is passed, the default tables are used.
4935 Version 2.00 24-Sep-98
4936 ----------------------
4938 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4939 it any more.
4941 2. Allow quantification of (?>) groups, and make it work correctly.
4943 3. The first character computation wasn't working for (?>) groups.
4945 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4946 end of the subject) and add 5.005's \z, which really does match only at the
4947 very end of the subject.
4949 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4951 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4952 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4953 localized options. All options to pcre_study() were also removed.
4955 7. Add other new features from 5.005:
4957 $(?<= positive lookbehind
4958 $(?<! negative lookbehind
4959 (?imsx-imsx) added the unsetting capability
4960 such a setting is global if at outer level; local otherwise
4961 (?imsx-imsx:) non-capturing groups with option setting
4962 (?(cond)re|re) conditional pattern matching
4964 A backreference to itself in a repeated group matches the previous
4965 captured string.
4967 8. General tidying up of studying (both automatic and via "study")
4968 consequential on the addition of new assertions.
4970 9. As in 5.005, unlimited repeated groups that could match an empty substring
4971 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4972 runtime if any iteration does actually match an empty substring.
4974 10. Include the RunTest script in the distribution.
4976 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4977 discrepancies, some of which were old and were also with respect to 5.004. They
4978 have now been fixed.
4981 Version 1.09 28-Apr-98
4982 ----------------------
4984 1. A negated single character class followed by a quantifier with a minimum
4985 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4986 program crashes, or just wrong answers. This did not apply to negated classes
4987 containing more than one character, or to minima other than one.
4990 Version 1.08 27-Mar-98
4991 ----------------------
4993 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4995 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4996 latter must appear before anything that relies on it in the pattern.
4999 Version 1.07 16-Feb-98
5000 ----------------------
5002 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5003 repeat of a potentially empty string).
5006 Version 1.06 23-Jan-98
5007 ----------------------
5009 1. Added Markus Oberhumer's little patches for C++.
5011 2. Literal strings longer than 255 characters were broken.
5014 Version 1.05 23-Dec-97
5015 ----------------------
5017 1. Negated character classes containing more than one character were failing if
5018 PCRE_CASELESS was set at run time.
5021 Version 1.04 19-Dec-97
5022 ----------------------
5024 1. Corrected the man page, where some "const" qualifiers had been omitted.
5026 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5027 input syntax.
5029 3. Fixed memory leak which occurred when a regex with back references was
5030 matched with an offsets vector that wasn't big enough. The temporary memory
5031 that is used in this case wasn't being freed if the match failed.
5033 4. Tidied pcretest to ensure it frees memory that it gets.
5035 5. Temporary memory was being obtained in the case where the passed offsets
5036 vector was exactly big enough.
5038 6. Corrected definition of offsetof() from change 5 below.
5040 7. I had screwed up change 6 below and broken the rules for the use of
5041 setjmp(). Now fixed.
5044 Version 1.03 18-Dec-97
5045 ----------------------
5047 1. A erroneous regex with a missing opening parenthesis was correctly
5048 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5049 on some systems.
5051 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5052 it was reported that one broken compiler failed on the former because "code" is
5053 also an independent variable.
5055 3. The erroneous regex a[]b caused an array overrun reference.
5057 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5058 fail on data ending with that character. (It was going on too far, and checking
5059 the next character, typically a binary zero.) This was specific to the
5060 optimized code for single-character negative classes.
5062 5. Added a contributed patch from the TIN world which does the following:
5064 + Add an undef for memmove, in case the the system defines a macro for it.
5066 + Add a definition of offsetof(), in case there isn't one. (I don't know
5067 the reason behind this - offsetof() is part of the ANSI standard - but
5068 it does no harm).
5070 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5071 most of the places where whitespace preceded '#'. I have given up and
5072 allowed the remaining 2 cases to be at the margin.
5074 + Rename some variables in pcre to eliminate shadowing. This seems very
5075 pedantic, but does no harm, of course.
5077 6. Moved the call to setjmp() into its own function, to get rid of warnings
5078 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5080 7. Constructs such as \d{8,} were compiling into the equivalent of
5081 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5082 outcome, but in this particular case used more store than had been allocated,
5083 which caused the bug to be discovered because it threw up an internal error.
5085 8. The debugging code in both pcre and pcretest for outputting the compiled
5086 form of a regex was going wrong in the case of back references followed by
5087 curly-bracketed repeats.
5090 Version 1.02 12-Dec-97
5091 ----------------------
5093 1. Typos in pcre.3 and comments in the source fixed.
5095 2. Applied a contributed patch to get rid of places where it used to remove
5096 'const' from variables, and fixed some signed/unsigned and uninitialized
5097 variable warnings.
5099 3. Added the "runtest" target to Makefile.
5101 4. Set default compiler flag to -O2 rather than just -O.
5104 Version 1.01 19-Nov-97
5105 ----------------------
5107 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5108 like /([ab]*)*/, that is, for classes with more than one character in them.
5110 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5111 as /((?>a*))*/ (a PCRE_EXTRA facility).
5114 Version 1.00 18-Nov-97
5115 ----------------------
5117 1. Added compile-time macros to support systems such as SunOS4 which don't have
5118 memmove() or strerror() but have other things that can be used instead.
5120 2. Arranged that "make clean" removes the executables.
5123 Version 0.99 27-Oct-97
5124 ----------------------
5126 1. Fixed bug in code for optimizing classes with only one character. It was
5127 initializing a 32-byte map regardless, which could cause it to run off the end
5128 of the memory it had got.
5130 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5133 Version 0.98 22-Oct-97
5134 ----------------------
5136 1. Fixed bug in code for handling temporary memory usage when there are more
5137 back references than supplied space in the ovector. This could cause segfaults.
5140 Version 0.97 21-Oct-97
5141 ----------------------
5143 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5145 2. Optimized negated single characters not to use a bit map.
5147 3. Brought error texts together as macro definitions; clarified some of them;
5148 fixed one that was wrong - it said "range out of order" when it meant "invalid
5149 escape sequence".
5151 4. Changed some char * arguments to const char *.
5153 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5155 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5156 pcretest.
5159 Version 0.96 16-Oct-97
5160 ----------------------
5162 1. Added a simple "pgrep" utility to the distribution.
5164 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5165 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5166 where "ddd" means "one or more decimal digits".
5168 3. Fixed serious bug. If a pattern had a back reference, but the call to
5169 pcre_exec() didn't supply a large enough ovector to record the related
5170 identifying subpattern, the match always failed. PCRE now remembers the number
5171 of the largest back reference, and gets some temporary memory in which to save
5172 the offsets during matching if necessary, in order to ensure that
5173 backreferences always work.
5175 4. Increased the compatibility with Perl in a number of ways:
5177 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5178 to request this handling. The option can be set at compile or exec time.
5180 (b) $ matches before a terminating newline by default; an option
5181 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5182 mode). The option can be set at compile or exec time.
5184 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5185 the same as Perl's. If the decimal number it represents is less than 10
5186 or there aren't that many previous left capturing parentheses, an octal
5187 escape is read. Inside a character class, it's always an octal escape,
5188 even if it is a single digit.
5190 (d) An escaped but undefined alphabetic character is taken as a literal,
5191 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5192 escapes.
5194 (e) {0} is now permitted. (The previous item is removed from the compiled
5195 pattern).
5197 5. Changed all the names of code files so that the basic parts are no longer
5198 than 10 characters, and abolished the teeny "globals.c" file.
5200 6. Changed the handling of character classes; they are now done with a 32-byte
5201 bit map always.
5203 7. Added the -d and /D options to pcretest to make it possible to look at the
5204 internals of compilation without having to recompile pcre.
5207 Version 0.95 23-Sep-97
5208 ----------------------
5210 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5211 \x20 at the start of a run of normal characters. These were being treated as
5212 real characters, instead of the source characters being re-checked.
5215 Version 0.94 18-Sep-97
5216 ----------------------
5218 1. The functions are now thread-safe, with the caveat that the global variables
5219 containing pointers to malloc() and free() or alternative functions are the
5220 same for all threads.
5222 2. Get pcre_study() to generate a bitmap of initial characters for non-
5223 anchored patterns when this is possible, and use it if passed to pcre_exec().
5226 Version 0.93 15-Sep-97
5227 ----------------------
5229 1. /(b)|(:+)/ was computing an incorrect first character.
5231 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5232 but not actually doing anything yet.
5234 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5235 as Perl does (e.g. [-az] or [az-]).
5237 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5238 all possible positions.
5240 5. Split up into different modules to avoid including unneeded functions in a
5241 compiled binary. However, compile and exec are still in one module. The "study"
5242 function is split off.
5244 6. The character tables are now in a separate module whose source is generated
5245 by an auxiliary program - but can then be edited by hand if required. There are
5246 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5247 toupper() in the code.
5249 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5250 make them global. Abolish the function for setting them, as the caller can now
5251 set them directly.
5254 Version 0.92 11-Sep-97
5255 ----------------------
5257 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5258 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5260 2. Caseless matching was not working in character classes if the characters in
5261 the pattern were in upper case.
5263 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5265 4. Make PCRE_ANCHORED public and accept as a compile option.
5267 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5268 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5269 pass them.
5271 6. Give an error if bad option bits passed at compile or run time.
5273 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5274 pcretest to cause it to pass that flag.
5276 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5277 options, and the first character, if set.
5279 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5282 Version 0.91 10-Sep-97
5283 ----------------------
5285 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5286 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5288 2. PCRE was looping on encountering an indefinitely repeated back reference to
5289 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5290 Perl does - treats the match as successful.
5292 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5