ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1354 - (show annotations)
Mon Aug 12 14:31:07 2013 UTC (8 years, 1 month ago) by ph10
File size: 250772 byte(s)
Add z/OS-specific code to pcregrep.
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 firse 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.
68 Version 8.33 28-May-2013
69 ------------------------
71 1. Added 'U' to some constants that are compared to unsigned integers, to
72 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
73 variables that are added to signed variables, to ensure the result is
74 signed and can be negated.
76 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
77 CMake config files.
79 3. Revise the creation of config.h.generic so that all boolean macros are
80 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
81 overriding via -D on the command line possible.
83 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
84 to unsigned int is reported to make a quite noticeable speed difference in
85 a specific Windows environment. Testing on Linux did also appear to show
86 some benefit (and it is clearly not harmful). Also fixed the definition of
87 Xop which should be unsigned.
89 5. Related to (4), changing the definition of the intermediate variable cc
90 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
91 improvements.
93 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
94 unnecessary spaces.
96 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
97 and later.
99 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
101 9. Optimizing fast_forward_start_bits in JIT.
103 10. Adding support for callouts in JIT, and fixing some issues revealed
104 during this work. Namely:
106 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
108 (b) Minimum length was not checked before the matching is started.
110 11. The value of capture_last that is passed to callouts was incorrect in some
111 cases when there was a capture on one path that was subsequently abandoned
112 after a backtrack. Also, the capture_last value is now reset after a
113 recursion, since all captures are also reset in this case.
115 12. The interpreter no longer returns the "too many substrings" error in the
116 case when an overflowing capture is in a branch that is subsequently
117 abandoned after a backtrack.
119 13. In the pathological case when an offset vector of size 2 is used, pcretest
120 now prints out the matched string after a yield of 0 or 1.
122 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
123 Only supported by the JIT compiler at the moment.
125 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
127 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
128 offset of the starting point of the matching process, provided the offsets
129 vector is large enough.
131 17. The \A escape now records a lookbehind value of 1, though its execution
132 does not actually inspect the previous character. This is to ensure that,
133 in partial multi-segment matching, at least one character from the old
134 segment is retained when a new segment is processed. Otherwise, if there
135 are no lookbehinds in the pattern, \A might match incorrectly at the start
136 of a new segment.
138 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
140 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
141 gives some modest performance improvement in 8-bit mode.
143 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
144 be expressed in certain programming languages using Universal Character
145 Names.
147 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
148 which points out that "non characters" are not "characters that may not
149 appear in Unicode strings" but rather "characters that are reserved for
150 internal use and have only local meaning".
152 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
153 there was a conditional group that depended on an assertion, if the
154 assertion was false, the callout that immediately followed the alternation
155 in the condition was skipped when pcre_exec() was used for matching.
157 23. Allow an explicit callout to be inserted before an assertion that is the
158 condition for a conditional group, for compatibility with automatic
159 callouts, which always insert a callout at this point.
161 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
162 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
164 25. (*PRUNE) is now supported by the JIT compiler.
166 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
168 27. Fix the case where there are two or more SKIPs with arguments that may be
169 ignored.
171 28. (*SKIP) is now supported by the JIT compiler.
173 29. (*THEN) is now supported by the JIT compiler.
175 30. Update RunTest with additional test selector options.
177 31. The way PCRE handles backtracking verbs has been changed in two ways.
179 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
180 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
181 backtracking. In some cases this makes it more Perl-compatible, but Perl's
182 rather obscure rules do not always do the same thing.
184 (2) Previously, backtracking verbs were confined within assertions. This is
185 no longer the case for positive assertions, except for (*ACCEPT). Again,
186 this sometimes improves Perl compatibility, and sometimes does not.
188 32. A number of tests that were in test 2 because Perl did things differently
189 have been moved to test 1, because either Perl or PCRE has changed, and
190 these tests are now compatible.
192 32. Backtracking control verbs are now handled in the same way in JIT and
193 interpreter.
195 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
196 contained a forward subroutine reference caused a compile error.
198 34. Auto-detect and optimize limited repetitions in JIT.
200 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
201 blocking (*UTF) etc.
203 36. In the interpreter, maximizing pattern repetitions for characters and
204 character types now use tail recursion, which reduces stack usage.
206 37. The value of the max lookbehind was not correctly preserved if a compiled
207 and saved regex was reloaded on a host of different endianness.
209 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
210 of the compiled pattern block, expand the flags field from 16 to 32 bits
211 because it was almost full.
213 39. Try madvise first before posix_madvise.
215 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
216 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
217 original change locked out all matches of empty strings. This has been
218 changed so that one match of an empty string per line is recognized.
219 Subsequent searches on the same line (for colouring or for --only-matching,
220 for example) do not recognize empty strings.
222 41. Applied a user patch to fix a number of spelling mistakes in comments.
224 42. Data lines longer than 65536 caused pcretest to crash.
226 43. Clarified the data type for length and startoffset arguments for pcre_exec
227 and pcre_dfa_exec in the function-specific man pages, where they were
228 explicitly stated to be in bytes, never having been updated. I also added
229 some clarification to the pcreapi man page.
231 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
232 a segmentation fault.
235 Version 8.32 30-November-2012
236 -----------------------------
238 1. Improved JIT compiler optimizations for first character search and single
239 character iterators.
241 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
242 Patch by Daniel Richard G.
244 3. Single character iterator optimizations in the JIT compiler.
246 4. Improved JIT compiler optimizations for character ranges.
248 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
249 Reported by Giuseppe D'Angelo.
251 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
252 of a line, was being set incorrectly in cases where .* appeared inside
253 atomic brackets at the start of a pattern, or where there was a subsequent
254 *PRUNE or *SKIP.
256 7. Improved instruction cache flush for POWER/PowerPC.
257 Patch by Daniel Richard G.
259 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
260 grep:
262 (a) There is now no limit to the number of patterns to be matched.
264 (b) An error is given if a pattern is too long.
266 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
267 are now supported.
269 (d) --exclude-from and --include-from (multiple use) have been added.
271 (e) Exclusions and inclusions now apply to all files and directories, not
272 just to those obtained from scanning a directory recursively.
274 (f) Multiple uses of -f and --file-list are now supported.
276 (g) In a Windows environment, the default for -d has been changed from
277 "read" (the GNU grep default) to "skip", because otherwise the presence
278 of a directory in the file list provokes an error.
280 (h) The documentation has been revised and clarified in places.
282 9. Improve the matching speed of capturing brackets.
284 10. Changed the meaning of \X so that it now matches a Unicode extended
285 grapheme cluster.
287 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
288 out POSIX threads when JIT support is configured.
290 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
292 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
293 structure could go wrong in environments where size_t is not the same size
294 as int.
296 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
297 set.
299 15. The EBCDIC support had decayed; later updates to the code had included
300 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
301 general tidy up of EBCDIC-related issues, and the documentation was also
302 not quite right. There is now a test that can be run on ASCII systems to
303 check some of the EBCDIC-related things (but is it not a full test).
305 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
306 in a small tidy to the code.
308 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
310 18. If the --only-matching (-o) option in pcregrep is specified multiple
311 times, each one causes appropriate output. For example, -o1 -o2 outputs the
312 substrings matched by the 1st and 2nd capturing parentheses. A separating
313 string can be specified by --om-separator (default empty).
315 19. Improving the first n character searches.
317 20. Turn case lists for horizontal and vertical white space into macros so that
318 they are defined only once.
320 21. This set of changes together give more compatible Unicode case-folding
321 behaviour for characters that have more than one other case when UCP
322 support is available.
324 (a) The Unicode property table now has offsets into a new table of sets of
325 three or more characters that are case-equivalent. The MultiStage2.py
326 script that generates these tables (the pcre_ucd.c file) now scans
327 CaseFolding.txt instead of UnicodeData.txt for character case
328 information.
330 (b) The code for adding characters or ranges of characters to a character
331 class has been abstracted into a generalized function that also handles
332 case-independence. In UTF-mode with UCP support, this uses the new data
333 to handle characters with more than one other case.
335 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
336 whose other case is greater than 256 are now correctly matched
337 caselessly. Previously, the high codepoint matched the low one, but not
338 vice versa.
340 (d) The processing of \h, \H, \v, and \ in character classes now makes use
341 of the new class addition function, using character lists defined as
342 macros alongside the case definitions of 20 above.
344 (e) Caseless back references now work with characters that have more than
345 one other case.
347 (f) General caseless matching of characters with more than one other case
348 is supported.
350 22. Unicode character properties were updated from Unicode 6.2.0
352 23. Improved CMake support under Windows. Patch by Daniel Richard G.
354 24. Add support for 32-bit character strings, and UTF-32
356 25. Major JIT compiler update (code refactoring and bugfixing).
357 Experimental Sparc 32 support is added.
359 26. Applied a modified version of Daniel Richard G's patch to create
360 pcre.h.generic and config.h.generic by "make" instead of in the
361 PrepareRelease script.
363 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
364 pcre_compile.c when checking for a zero character.
366 28. Introducing a native interface for JIT. Through this interface, the compiled
367 machine code can be directly executed. The purpose of this interface is to
368 provide fast pattern matching, so several sanity checks are not performed.
369 However, feature tests are still performed. The new interface provides
370 1.4x speedup compared to the old one.
372 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
373 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
374 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
376 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
377 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
378 the "old" RFC 2279). Instead, it ended up passing a negative length to
379 pcre_exec().
381 31. Add support for GCC's visibility feature to hide internal functions.
383 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
384 "unknown -C option" after outputting 0 or 1.
386 33. There is now support for generating a code coverage report for the test
387 suite in environments where gcc is the compiler and lcov is installed. This
388 is mainly for the benefit of the developers.
390 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
391 unaddressable using valgrind annotations, allowing valgrind to detect
392 invalid memory accesses. This is mainly for the benefit of the developers.
394 25. (*UTF) can now be used to start a pattern in any of the three libraries.
396 26. Give configure error if --enable-cpp but no C++ compiler found.
399 Version 8.31 06-July-2012
400 -------------------------
402 1. Fixing a wrong JIT test case and some compiler warnings.
404 2. Removed a bashism from the RunTest script.
406 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
407 to unsigned type, result still unsigned" that was given by an MS compiler
408 on encountering the code "-sizeof(xxx)".
410 4. Partial matching support is added to the JIT compiler.
412 5. Fixed several bugs concerned with partial matching of items that consist
413 of more than one character:
415 (a) /^(..)\1/ did not partially match "aba" because checking references was
416 done on an "all or nothing" basis. This also applied to repeated
417 references.
419 (b) \R did not give a hard partial match if \r was found at the end of the
420 subject.
422 (c) \X did not give a hard partial match after matching one or more
423 characters at the end of the subject.
425 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
426 a partial match for the string "\r".
428 (e) When newline was set to CRLF, the metacharacter "." did not recognize
429 a partial match for a CR character at the end of the subject string.
431 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
432 running pcretest, the text "(JIT)" added to the output whenever JIT is
433 actually used to run the match.
435 7. Individual JIT compile options can be set in pcretest by following -s+[+]
436 or /S+[+] with a digit between 1 and 7.
438 8. OP_NOT now supports any UTF character not just single-byte ones.
440 9. (*MARK) control verb is now supported by the JIT compiler.
442 10. The command "./RunTest list" lists the available tests without actually
443 running any of them. (Because I keep forgetting what they all are.)
447 12. Applied a (slightly modified) user-supplied patch that improves performance
448 when the heap is used for recursion (compiled with --disable-stack-for-
449 recursion). Instead of malloc and free for each heap frame each time a
450 logical recursion happens, frames are retained on a chain and re-used where
451 possible. This sometimes gives as much as 30% improvement.
453 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
454 call.
456 14. As documented, (*COMMIT) is now confined to within a positive assertion.
458 15. It is now possible to link pcretest with libedit as an alternative to
459 libreadline.
461 16. (*COMMIT) control verb is now supported by the JIT compiler.
463 17. The Unicode data tables have been updated to Unicode 6.1.0.
465 18. Added --file-list option to pcregrep.
467 19. Added binary file support to pcregrep, including the -a, --binary-files,
468 -I, and --text options.
470 20. The madvise function is renamed for posix_madvise for QNX compatibility
471 reasons. Fixed by Giuseppe D'Angelo.
473 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
475 22. Changed the option for creating symbolic links for 16-bit man pages from
476 -s to -sf so that re-installing does not cause issues.
478 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
480 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
481 matches in certain environments (the workspace was not being correctly
482 retained). Also added to pcre_dfa_exec() a simple plausibility check on
483 some of the workspace data at the beginning of a restart.
485 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
486 was not doing so when it should - probably a typo introduced by SVN 528
487 (change 8.10/14).
489 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
490 \w+ when the character tables indicated that \x{c4} was a word character.
491 There were several related cases, all because the tests for doing a table
492 lookup were testing for characters less than 127 instead of 255.
494 27. If a pattern contains capturing parentheses that are not used in a match,
495 their slots in the ovector are set to -1. For those that are higher than
496 any matched groups, this happens at the end of processing. In the case when
497 there were back references that the ovector was too small to contain
498 (causing temporary malloc'd memory to be used during matching), and the
499 highest capturing number was not used, memory off the end of the ovector
500 was incorrectly being set to -1. (It was using the size of the temporary
501 memory instead of the true size.)
503 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
504 ovector size, it uses memory at the end of the block that it has got.
506 29. Check for an overlong MARK name and give an error at compile time. The
507 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
509 30. JIT compiler update.
511 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
512 Rill for the patch.
514 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
516 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
518 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
519 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
520 was enabled.
522 35. Improve JIT code generation for greedy plus quantifier.
524 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
525 1 to "aa" instead of to an empty string. The bug affected repeated groups
526 that could potentially match an empty string.
528 37. Optimizing single character iterators in JIT.
530 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
531 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
532 codepoints that are too big for the mode are faulted, and in a UTF mode,
533 disallowed codepoints are also faulted.
535 39. If PCRE was compiled with UTF support, in three places in the DFA
536 matcher there was code that should only have been obeyed in UTF mode, but
537 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
538 processing when bytes with values greater than 127 were present. In 16-bit
539 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
540 both cases the values are those that cannot be the first data item in a UTF
541 character. The three items that might have provoked this were recursions,
542 possessively repeated groups, and atomic groups.
544 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
545 and pcregrep, because some OS require shared objects to be explicitly
546 passed to ld, causing the link step to fail if they are not.
548 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
549 mode, patterns that started with \h* or \R* might be incorrectly matched.
552 Version 8.30 04-February-2012
553 -----------------------------
555 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
556 name is defined in ctype.h.
558 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
559 only in quite long subpatterns.
561 3. Removed the function pcre_info(), which has been obsolete and deprecated
562 since it was replaced by pcre_fullinfo() in February 2000.
564 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
565 match a (*MARK), and the match failed at the start of the subject, a
566 reference to memory before the start of the subject could occur. This bug
567 was introduced by fix 17 of release 8.21.
569 5. A reference to an unset group with zero minimum repetition was giving
570 totally wrong answers (in non-JavaScript-compatibility mode). For example,
571 /(another)?(\1?)test/ matched against "hello world test". This bug was
572 introduced in release 8.13.
574 6. Add support for 16-bit character strings (a large amount of work involving
575 many changes and refactorings).
577 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
578 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
579 from a file.
581 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
582 rounding is not applied in this particular case).
584 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
585 if they appear, or are escaped, in patterns.
587 10. Get rid of a number of -Wunused-but-set-variable warnings.
589 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
590 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
591 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
592 also returns the mark "x". This bug applied to capturing parentheses,
593 non-capturing parentheses, and atomic parentheses. It also applied to some
594 assertions.
596 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
597 information out of configure.ac instead of relying on pcre.h.generic, which
598 is not stored in the repository.
600 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
601 -lreadline.
603 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
605 15. Applied Graycode's patch to put the top-level frame on the stack rather
606 than the heap when not using the stack for recursion. This gives a
607 performance improvement in many cases when recursion is not deep.
609 16. Experimental code added to "pcretest -C" to output the stack frame size.
612 Version 8.21 12-Dec-2011
613 ------------------------
615 1. Updating the JIT compiler.
617 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
618 are added as well.
620 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
621 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
622 calling _pcre_jit_exec. Some extra comments are added.
624 4. (*MARK) settings inside atomic groups that do not contain any capturing
625 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
626 was introduced by change 18 for 8.20.
628 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
629 ECMA-262 standard.
631 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
632 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
633 This bug was probably introduced by change 9 of 8.13.
635 7. While fixing 6 above, I noticed that a number of other items were being
636 incorrectly rejected as "not fixed length". This arose partly because newer
637 opcodes had not been added to the fixed-length checking code. I have (a)
638 corrected the bug and added tests for these items, and (b) arranged for an
639 error to occur if an unknown opcode is encountered while checking for fixed
640 length instead of just assuming "not fixed length". The items that were
641 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
642 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
643 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
645 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
646 being incorrectly compiled and would have given unpredicatble results.
648 9. A possessively repeated subpattern with minimum repeat count greater than
649 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
650 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
651 the first (A) could occur when it should not.
653 10. Add a cast and remove a redundant test from the code.
655 11. JIT should use pcre_malloc/pcre_free for allocation.
657 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
658 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
659 is anything other than /usr, -L is still shown).
661 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
663 14. Perl does not support \N without a following name in a [] class; PCRE now
664 also gives an error.
666 15. If a forward reference was repeated with an upper limit of around 2000,
667 it caused the error "internal error: overran compiling workspace". The
668 maximum number of forward references (including repeats) was limited by the
669 internal workspace, and dependent on the LINK_SIZE. The code has been
670 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
671 the default depends on LINK_SIZE. There is a new upper limit (for safety)
672 of around 200,000 forward references. While doing this, I also speeded up
673 the filling in of repeated forward references.
675 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
676 incorrectly expecting the subject to contain another "a" after the start.
678 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
679 in the match, the SKIP should be ignored. This was not happening; instead
680 the SKIP was being treated as NOMATCH. For patterns such as
681 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
682 tested.
684 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
685 now much more compatible with Perl, in particular in cases where the result
686 is a non-match for a non-anchored pattern. For example, if
687 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
688 "m", where previously it did not return a name. A side effect of this
689 change is that for partial matches, the last encountered mark name is
690 returned, as for non matches. A number of tests that were previously not
691 Perl-compatible have been moved into the Perl-compatible test files. The
692 refactoring has had the pleasing side effect of removing one argument from
693 the match() function, thus reducing its stack requirements.
695 19. If the /S+ option was used in pcretest to study a pattern using JIT,
696 subsequent uses of /S (without +) incorrectly behaved like /S+.
698 21. Retrieve executable code size support for the JIT compiler and fixing
699 some warnings.
701 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
702 not work when the shorter character appeared right at the end of the
703 subject string.
705 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
706 systems.
708 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
709 output it when the /M option is used in pcretest.
711 25. The CheckMan script was not being included in the distribution. Also, added
712 an explicit "perl" to run Perl scripts from the PrepareRelease script
713 because this is reportedly needed in Windows.
715 26. If study data was being save in a file and studying had not found a set of
716 "starts with" bytes for the pattern, the data written to the file (though
717 never used) was taken from uninitialized memory and so caused valgrind to
718 complain.
720 27. Updated RunTest.bat as provided by Sheri Pierce.
722 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
724 29. Computation of memory usage for the table of capturing group names was
725 giving an unnecessarily large value.
728 Version 8.20 21-Oct-2011
729 ------------------------
731 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
732 a POSIX class. After further experiments with Perl, which convinced me that
733 Perl has bugs and confusions, a closing square bracket is no longer allowed
734 in a POSIX name. This bug also affected patterns with classes that started
735 with full stops.
737 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
738 captured substring, but while checking the failing first alternative,
739 substring 1 is temporarily captured. If the output vector supplied to
740 pcre_exec() was not big enough for this capture, the yield of the function
741 was still zero ("insufficient space for captured substrings"). This cannot
742 be totally fixed without adding another stack variable, which seems a lot
743 of expense for a edge case. However, I have improved the situation in cases
744 such as /(a)(b)x|abc/ matched against "abc", where the return code
745 indicates that fewer than the maximum number of slots in the ovector have
746 been set.
748 3. Related to (2) above: when there are more back references in a pattern than
749 slots in the output vector, pcre_exec() uses temporary memory during
750 matching, and copies in the captures as far as possible afterwards. It was
751 using the entire output vector, but this conflicts with the specification
752 that only 2/3 is used for passing back captured substrings. Now it uses
753 only the first 2/3, for compatibility. This is, of course, another edge
754 case.
756 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
757 main code base, and can be used by building with --enable-jit. When this is
758 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
759 runtime --no-jit option is given.
761 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
762 ovector, the return from the function was zero, implying that there were
763 other matches that did not fit. The correct "exactly full" value is now
764 returned.
766 6. If a subpattern that was called recursively or as a subroutine contained
767 (*PRUNE) or any other control that caused it to give a non-standard return,
768 invalid errors such as "Error -26 (nested recursion at the same subject
769 position)" or even infinite loops could occur.
771 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
772 computing the minimum length on reaching *ACCEPT, and so ended up with the
773 wrong value of 1 rather than 0. Further investigation indicates that
774 computing a minimum subject length in the presence of *ACCEPT is difficult
775 (think back references, subroutine calls), and so I have changed the code
776 so that no minimum is registered for a pattern that contains *ACCEPT.
778 8. If (*THEN) was present in the first (true) branch of a conditional group,
779 it was not handled as intended. [But see 16 below.]
781 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
782 Sheri Pierce.
784 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
785 the first byte in a match must be "a".
787 11. Change 17 for 8.13 increased the recursion depth for patterns like
788 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
789 pattern contains any instances of (*THEN). If it does not, the old
790 optimizations are restored. It would be nice to do this on a per-group
791 basis, but at the moment that is not feasible.
793 12. In some environments, the output of pcretest -C is CRLF terminated. This
794 broke RunTest's code that checks for the link size. A single white space
795 character after the value is now allowed for.
797 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
798 For "fr", it uses the Windows-specific input and output files.
800 14. If (*THEN) appeared in a group that was called recursively or as a
801 subroutine, it did not work as intended. [But see next item.]
803 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
804 pattern fragments (but not containing any | characters). If A and B are
805 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
806 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
807 D. In other words, Perl considers parentheses that do not contain any |
808 characters to be part of a surrounding alternative, whereas PCRE was
809 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
810 differently. PCRE now behaves in the same way as Perl, except in the case
811 of subroutine/recursion calls such as (?1) which have in any case always
812 been different (but PCRE had them first :-).
814 16. Related to 15 above: Perl does not treat the | in a conditional group as
815 creating alternatives. Such a group is treated in the same way as an
816 ordinary group without any | characters when processing (*THEN). PCRE has
817 been changed to match Perl's behaviour.
819 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
820 RunGrepTest script failed.
822 18. Change 22 for version 13 caused atomic groups to use more stack. This is
823 inevitable for groups that contain captures, but it can lead to a lot of
824 stack use in large patterns. The old behaviour has been restored for atomic
825 groups that do not contain any capturing parentheses.
827 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
828 suppress the check for a minimum subject length at run time. (If it was
829 given to pcre_exec() or pcre_dfa_exec() it did work.)
831 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
832 fail to work when decoding hex characters in data strings in EBCDIC
833 environments.
835 21. It appears that in at least one Mac OS environment, the isxdigit() function
836 is implemented as a macro that evaluates to its argument more than once,
837 contravening the C 90 Standard (I haven't checked a later standard). There
838 was an instance in pcretest which caused it to go wrong when processing
839 \x{...} escapes in subject strings. The has been rewritten to avoid using
840 things like p++ in the argument of isxdigit().
843 Version 8.13 16-Aug-2011
844 ------------------------
846 1. The Unicode data tables have been updated to Unicode 6.0.0.
848 2. Two minor typos in pcre_internal.h have been fixed.
850 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
851 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
852 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
854 4. There were a number of related bugs in the code for matching backrefences
855 caselessly in UTF-8 mode when codes for the characters concerned were
856 different numbers of bytes. For example, U+023A and U+2C65 are an upper
857 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
858 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
859 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
860 2-byte code at the end of the subject (it thought there wasn't enough data
861 left).
863 5. Comprehensive information about what went wrong is now returned by
864 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
865 as the output vector has at least 2 elements. The offset of the start of
866 the failing character and a reason code are placed in the vector.
868 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
869 now returned is for the first byte of the failing character, instead of the
870 last byte inspected. This is an incompatible change, but I hope it is small
871 enough not to be a problem. It makes the returned offset consistent with
872 pcre_exec() and pcre_dfa_exec().
874 7. pcretest now gives a text phrase as well as the error number when
875 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
876 failure, the offset and reason code are output.
878 8. When \R was used with a maximizing quantifier it failed to skip backwards
879 over a \r\n pair if the subsequent match failed. Instead, it just skipped
880 back over a single character (\n). This seems wrong (because it treated the
881 two characters as a single entity when going forwards), conflicts with the
882 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
883 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
884 has been changed.
886 9. Some internal refactoring has changed the processing so that the handling
887 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
888 time (the PCRE_DOTALL option was changed this way some time ago: version
889 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
890 which was always a bit of a fudge. It also means that there is one less
891 argument for the match() function, which reduces its stack requirements
892 slightly. This change also fixes an incompatibility with Perl: the pattern
893 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
895 10. More internal refactoring has drastically reduced the number of recursive
896 calls to match() for possessively repeated groups such as (abc)++ when
897 using pcre_exec().
899 11. While implementing 10, a number of bugs in the handling of groups were
900 discovered and fixed:
902 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
903 (a|)*(?1) gave a compile-time internal error.
904 ((a|)+)+ did not notice that the outer group could match an empty string.
905 (^a|^)+ was not marked as anchored.
906 (.*a|.*)+ was not marked as matching at start or after a newline.
908 12. Yet more internal refactoring has removed another argument from the match()
909 function. Special calls to this function are now indicated by setting a
910 value in a variable in the "match data" data block.
912 13. Be more explicit in pcre_study() instead of relying on "default" for
913 opcodes that mean there is no starting character; this means that when new
914 ones are added and accidentally left out of pcre_study(), testing should
915 pick them up.
917 14. The -s option of pcretest has been documented for ages as being an old
918 synonym of -m (show memory usage). I have changed it to mean "force study
919 for every regex", that is, assume /S for every regex. This is similar to -i
920 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
921 using it. It makes it easier to run collections of tests with and without
922 study enabled, and thereby test pcre_study() more easily. All the standard
923 tests are now run with and without -s (but some patterns can be marked as
924 "never study" - see 20 below).
926 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
927 restoration of the capturing data to the outer values was not happening
928 correctly.
930 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
931 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
932 pattern had matched an empty string, and so incorrectly returned a no
933 match.
935 17. There was optimizing code for the last branch of non-capturing parentheses,
936 and also for the obeyed branch of a conditional subexpression, which used
937 tail recursion to cut down on stack usage. Unfortunately, now that there is
938 the possibility of (*THEN) occurring in these branches, tail recursion is
939 no longer possible because the return has to be checked for (*THEN). These
940 two optimizations have therefore been removed. [But see 8.20/11 above.]
942 18. If a pattern containing \R was studied, it was assumed that \R always
943 matched two bytes, thus causing the minimum subject length to be
944 incorrectly computed because \R can also match just one byte.
946 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
947 was incorrectly computed.
949 20. If /S is present twice on a test pattern in pcretest input, it now
950 *disables* studying, thereby overriding the use of -s on the command line
951 (see 14 above). This is necessary for one or two tests to keep the output
952 identical in both cases.
954 21. When (*ACCEPT) was used in an assertion that matched an empty string and
955 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
957 22. When an atomic group that contained a capturing parenthesis was
958 successfully matched, but the branch in which it appeared failed, the
959 capturing was not being forgotten if a higher numbered group was later
960 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
961 group 1 to "a", when in fact it should be unset. This applied to multi-
962 branched capturing and non-capturing groups, repeated or not, and also to
963 positive assertions (capturing in negative assertions does not happen
964 in PCRE) and also to nested atomic groups.
966 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
967 subject after a captured substring, to make it easier to tell which of a
968 number of identical substrings has been captured.
970 24. The way atomic groups are processed by pcre_exec() has been changed so that
971 if they are repeated, backtracking one repetition now resets captured
972 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
973 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
974 "aaa". Previously, it would have been "a". As part of this code
975 refactoring, the way recursive calls are handled has also been changed.
977 25. If an assertion condition captured any substrings, they were not passed
978 back unless some other capturing happened later. For example, if
979 (?(?=(a))a) was matched against "a", no capturing was returned.
981 26. When studying a pattern that contained subroutine calls or assertions,
982 the code for finding the minimum length of a possible match was handling
983 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
984 group 1 called group 2 while simultaneously a separate group 2 called group
985 1). A stack overflow occurred in this case. I have fixed this by limiting
986 the recursion depth to 10.
988 27. Updated RunTest.bat in the distribution to the version supplied by Tom
989 Fortmann. This supports explicit test numbers on the command line, and has
990 argument validation and error reporting.
992 28. An instance of \X with an unlimited repeat could fail if at any point the
993 first character it looked at was a mark character.
995 29. Some minor code refactoring concerning Unicode properties and scripts
996 should reduce the stack requirement of match() slightly.
998 30. Added the '=' option to pcretest to check the setting of unused capturing
999 slots at the end of the pattern, which are documented as being -1, but are
1000 not included in the return count.
1002 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1003 compiled something random. Now it gives a compile-time error (as does
1004 Perl).
1006 32. A *MARK encountered during the processing of a positive assertion is now
1007 recorded and passed back (compatible with Perl).
1009 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1010 had alternative anchored branches, the search for a second match in a line
1011 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1012 matched the line "0102" twice. The same bug affected patterns that started
1013 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1014 twice.
1016 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1017 does, and because of capturing effects, quantifying parenthesized
1018 assertions may at times be useful. Quantifiers are now allowed for
1019 parenthesized assertions.
1021 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1023 36. \g was being checked for fancy things in a character class, when it should
1024 just be a literal "g".
1026 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1027 appearance of a nested POSIX class supersedes an apparent external class.
1028 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1029 unescaped square brackets may also appear as part of class names. For
1030 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1031 more like Perl. (But see 8.20/1 above.)
1033 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1034 was because it thought it was \N{name}, which is not supported).
1036 39. Add minix to OS list not supporting the -S option in pcretest.
1038 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1039 cannot analyze patterns in sufficient detail to catch mutual recursions
1040 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1041 subgroup is called recursively as a subpattern for a second time at the
1042 same position in the subject string. In previous releases this might have
1043 been caught by the recursion limit, or it might have run out of stack.
1045 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1046 happen only once. PCRE was, however incorrectly giving a compile time error
1047 "recursive call could loop indefinitely" because it cannot analyze the
1048 pattern in sufficient detail. The compile time test no longer happens when
1049 PCRE is compiling a conditional subpattern, but actual runaway loops are
1050 now caught at runtime (see 40 above).
1052 42. It seems that Perl allows any characters other than a closing parenthesis
1053 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1054 has been changed to be the same.
1056 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1057 as not to get warnings when autogen.sh is called. Also changed
1058 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1060 44. To help people who use pcregrep to scan files containing exceedingly long
1061 lines, the following changes have been made:
1063 (a) The default value of the buffer size parameter has been increased from
1064 8K to 20K. (The actual buffer used is three times this size.)
1066 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1067 PCRE is built.
1069 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1070 to be set at run time.
1072 (d) Numerical values in pcregrep options can be followed by K or M, for
1073 example --buffer-size=50K.
1075 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1076 given and the return code is set to 2.
1078 45. Add a pointer to the latest mark to the callout data block.
1080 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1081 partial match of an empty string instead of no match. This was specific to
1082 the use of ".".
1084 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1085 complete match instead of a partial match. This bug was dependent on both
1086 the PCRE_UTF8 and PCRE_DOTALL options being set.
1088 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1089 starting byte set, because \b was not being ignored.
1092 Version 8.12 15-Jan-2011
1093 ------------------------
1095 1. Fixed some typos in the markup of the man pages, and wrote a script that
1096 checks for such things as part of the documentation building process.
1098 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1099 --match-limit and --recursion-limit options (added for 8.11). In
1100 particular, this made one of the standard tests fail. (The integer value
1101 went into the wrong half of a long int.)
1103 3. If the --colour option was given to pcregrep with -v (invert match), it
1104 did strange things, either producing crazy output, or crashing. It should,
1105 of course, ignore a request for colour when reporting lines that do not
1106 match.
1108 4. Another pcregrep bug caused similar problems if --colour was specified with
1109 -M (multiline) and the pattern match finished with a line ending.
1111 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1112 matched in multiline mode, the following line was shown as part of the
1113 match. This seems wrong, so I have changed it.
1115 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1116 the check for further matches in the same line (so they could be coloured)
1117 to overrun the end of the current line. If another match was found, it was
1118 incorrectly shown (and then shown again when found in the next line).
1120 7. If pcregrep was compiled under Windows, there was a reference to the
1121 function pcregrep_exit() before it was defined. I am assuming this was
1122 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1123 reported by a user. I've moved the definition above the reference.
1126 Version 8.11 10-Dec-2010
1127 ------------------------
1129 1. (*THEN) was not working properly if there were untried alternatives prior
1130 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1131 backtracked to try for "b" instead of moving to the next alternative branch
1132 at the same level (in this case, to look for "c"). The Perl documentation
1133 is clear that when (*THEN) is backtracked onto, it goes to the "next
1134 alternative in the innermost enclosing group".
1136 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1137 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1138 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1139 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1140 (*THEN).
1142 3. If \s appeared in a character class, it removed the VT character from
1143 the class, even if it had been included by some previous item, for example
1144 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1145 of \s, but is part of the POSIX "space" class.)
1147 4. A partial match never returns an empty string (because you can always
1148 match an empty string at the end of the subject); however the checking for
1149 an empty string was starting at the "start of match" point. This has been
1150 changed to the "earliest inspected character" point, because the returned
1151 data for a partial match starts at this character. This means that, for
1152 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1153 (previously it gave "no match").
1155 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1156 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1157 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1158 has an implication that the given string is incomplete (because a partial
1159 match is preferred over a full match). For this reason, these items now
1160 give a partial match in this situation. [Aside: previously, the one case
1161 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1162 match rather than a full match, which was wrong by the old rules, but is
1163 now correct.]
1165 6. There was a bug in the handling of #-introduced comments, recognized when
1166 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1167 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1168 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1169 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1170 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1171 places in pcre_compile().
1173 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1174 comments when looking ahead for named forward references to subpatterns,
1175 the only newline sequence it recognized was NL. It now handles newlines
1176 according to the set newline convention.
1178 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1179 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1180 cater for a lack of strerror(). These oversights have been fixed.
1182 9. Added --match-limit and --recursion-limit to pcregrep.
1184 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1186 11. When the -o option was used, pcregrep was setting a return code of 1, even
1187 when matches were found, and --line-buffered was not being honoured.
1189 12. Added an optional parentheses number to the -o and --only-matching options
1190 of pcregrep.
1192 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1193 can match an empty string. The code to do it in pcretest and pcredemo
1194 needed fixing:
1196 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1197 only one byte after an empty string match just before CRLF (this case
1198 just got forgotten; "any" and "anycrlf" were OK).
1200 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1201 mode when an empty string match preceded an ASCII character followed by
1202 a non-ASCII character. (The code for advancing by one character rather
1203 than one byte was nonsense.)
1205 (c) The pcredemo.c sample program did not have any code at all to handle
1206 the cases when CRLF is a valid newline sequence.
1208 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1209 as a starting offset was within the subject string. There is now a new
1210 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1211 negative or greater than the length of the string. In order to test this,
1212 pcretest is extended to allow the setting of negative starting offsets.
1214 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1215 starting offset points to the beginning of a UTF-8 character was
1216 unnecessarily clumsy. I tidied it up.
1218 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1219 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1221 17. Nobody had reported that the --include_dir option, which was added in
1222 release 7.7 should have been called --include-dir (hyphen, not underscore)
1223 for compatibility with GNU grep. I have changed it to --include-dir, but
1224 left --include_dir as an undocumented synonym, and the same for
1225 --exclude-dir, though that is not available in GNU grep, at least as of
1226 release 2.5.4.
1228 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1229 characters from a string of bytes) have been redefined so as not to use
1230 loops, in order to improve performance in some environments. At the same
1231 time, I abstracted some of the common code into auxiliary macros to save
1232 repetition (this should not affect the compiled code).
1234 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1235 compile-time error is now given if \c is not followed by an ASCII
1236 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1237 different, and any byte value is allowed.)
1239 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1240 START_OPTIMIZE option, which is now allowed at compile time - but just
1241 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1242 to pcregrep and other applications that have no direct access to PCRE
1243 options. The new /Y option in pcretest sets this option when calling
1244 pcre_compile().
1246 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1247 back references. Groups containing recursive back references were forced to
1248 be atomic by that change, but in the case of named groups, the amount of
1249 memory required was incorrectly computed, leading to "Failed: internal
1250 error: code overflow". This has been fixed.
1252 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1253 pcretest.c, to avoid build problems in some Borland environments.
1256 Version 8.10 25-Jun-2010
1257 ------------------------
1259 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1260 THEN.
1262 2. (*ACCEPT) was not working when inside an atomic group.
1264 3. Inside a character class, \B is treated as a literal by default, but
1265 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1266 causes the error). The code is unchanged, but I tidied the documentation.
1268 4. Inside a character class, PCRE always treated \R and \X as literals,
1269 whereas Perl faults them if its -w option is set. I have changed PCRE so
1270 that it faults them when PCRE_EXTRA is set.
1272 5. Added support for \N, which always matches any character other than
1273 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1275 6. When compiling pcregrep with newer versions of gcc which may have
1276 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1277 declared with attribute warn_unused_result" were given. Just casting the
1278 result to (void) does not stop the warnings; a more elaborate fudge is
1279 needed. I've used a macro to implement this.
1281 7. Minor change to pcretest.c to avoid a compiler warning.
1283 8. Added four artifical Unicode properties to help with an option to make
1284 \s etc use properties (see next item). The new properties are: Xan
1285 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1287 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1288 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1289 this option. Modified pcretest to add /W to test this facility. Added
1290 REG_UCP to make it available via the POSIX interface.
1292 10. Added --line-buffered to pcregrep.
1294 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1295 studied, and the match started with a letter with a code point greater than
1296 127 whose first byte was different to the first byte of the other case of
1297 the letter, the other case of this starting letter was not recognized
1298 (#976).
1300 12. If a pattern that was studied started with a repeated Unicode property
1301 test, for example, \p{Nd}+, there was the theoretical possibility of
1302 setting up an incorrect bitmap of starting bytes, but fortunately it could
1303 not have actually happened in practice until change 8 above was made (it
1304 added property types that matched character-matching opcodes).
1306 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1307 possible starting bytes for non-anchored patterns.
1309 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1310 \R, and also a number of cases that involve Unicode properties, both
1311 explicit and implicit when PCRE_UCP is set.
1313 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1314 input, it could crash or give wrong results if characters with values
1315 greater than 0xc0 were present in the subject string. (Detail: it assumed
1316 UTF-8 input when processing these items.)
1318 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1319 size_t is 64-bit (#991).
1321 17. Added a check for running out of memory when PCRE is compiled with
1322 --disable-stack-for-recursion (#990).
1324 18. If the last data line in a file for pcretest does not have a newline on
1325 the end, a newline was missing in the output.
1327 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1328 less than 128) in its various bitmaps. However, there is a facility for
1329 generating tables according to the current locale when PCRE is compiled. It
1330 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1331 characters, are recognized by isspace() and therefore were getting set in
1332 these tables, and indeed these tables seem to approximate to ISO 8859. This
1333 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1334 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1335 which of course cannot start UTF-8 characters. I have changed the code so
1336 that only real ASCII characters (less than 128) and the correct starting
1337 bytes for UTF-8 encodings are set for characters greater than 127 when in
1338 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1339 altogether.)
1341 20. Added the /T option to pcretest so as to be able to run tests with non-
1342 standard character tables, thus making it possible to include the tests
1343 used for 19 above in the standard set of tests.
1345 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1346 reference to a subpattern the other side of a comment that contains an
1347 opening parenthesis caused either an internal compiling error, or a
1348 reference to the wrong subpattern.
1351 Version 8.02 19-Mar-2010
1352 ------------------------
1354 1. The Unicode data tables have been updated to Unicode 5.2.0.
1356 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1357 configured.
1359 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1360 original author of that file, following a query about its status.
1362 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1363 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1365 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1366 quantifier applied to a forward-referencing subroutine call, could compile
1367 incorrect code or give the error "internal error: previously-checked
1368 referenced subpattern not found".
1370 6. Both MS Visual Studio and Symbian OS have problems with initializing
1371 variables to point to external functions. For these systems, therefore,
1372 pcre_malloc etc. are now initialized to local functions that call the
1373 relevant global functions.
1375 7. There were two entries missing in the vectors called coptable and poptable
1376 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1377 I've fixed the data, and added a kludgy way of testing at compile time that
1378 the lengths are correct (equal to the number of opcodes).
1380 8. Following on from 7, I added a similar kludge to check the length of the
1381 eint vector in pcreposix.c.
1383 9. Error texts for pcre_compile() are held as one long string to avoid too
1384 much relocation at load time. To find a text, the string is searched,
1385 counting zeros. There was no check for running off the end of the string,
1386 which could happen if a new error number was added without updating the
1387 string.
1389 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1391 11. \K was not working if it appeared in an atomic group or in a group that
1392 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1393 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1394 the assertion is positive, but not if it is negative.
1396 12. Change 11 fortuitously reduced the size of the stack frame used in the
1397 "match()" function of pcre_exec.c by one pointer. Forthcoming
1398 implementation of support for (*MARK) will need an extra pointer on the
1399 stack; I have reserved it now, so that the stack frame size does not
1400 decrease.
1402 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1403 item in branch that calls a recursion is a subroutine call - as in the
1404 second branch in the above example - was incorrectly given the compile-
1405 time error "recursive call could loop indefinitely" because pcre_compile()
1406 was not correctly checking the subroutine for matching a non-empty string.
1408 14. The checks for overrunning compiling workspace could trigger after an
1409 overrun had occurred. This is a "should never occur" error, but it can be
1410 triggered by pathological patterns such as hundreds of nested parentheses.
1411 The checks now trigger 100 bytes before the end of the workspace.
1413 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1416 Version 8.01 19-Jan-2010
1417 ------------------------
1419 1. If a pattern contained a conditional subpattern with only one branch (in
1420 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1421 computed the wrong minimum data length (which is of course zero for such
1422 subpatterns). This could cause incorrect "no match" results.
1424 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1425 the pattern is reset in the first branch, pcre_compile() failed with
1426 "internal error: code overflow at offset...". This happened only when
1427 the reset was to the original external option setting. (An optimization
1428 abstracts leading options settings into an external setting, which was the
1429 cause of this.)
1431 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1432 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1433 assertion pattern did not match (meaning that the assertion was true), it
1434 was incorrectly treated as false if the SKIP had been reached during the
1435 matching. This also applied to assertions used as conditions.
1437 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1438 assertion subpattern, including such a pattern used as a condition,
1439 unpredictable results occurred, instead of the error return
1442 5. The C++ GlobalReplace function was not working like Perl for the special
1443 situation when an empty string is matched. It now does the fancy magic
1444 stuff that is necessary.
1446 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1447 removed. (These were left over from very, very early versions of PCRE.)
1449 7. Some cosmetic changes to the code to make life easier when compiling it
1450 as part of something else:
1452 (a) Change DEBUG to PCRE_DEBUG.
1454 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1455 called "current" as "current_branch", to prevent a collision with the
1456 Linux macro when compiled as a kernel module.
1458 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1459 prevent a collision with the Linux macro when compiled as a kernel
1460 module.
1462 8. In pcre_compile() there are some checks for integer overflows that used to
1463 cast potentially large values to (double). This has been changed to that
1464 when building, a check for int64_t is made, and if it is found, it is used
1465 instead, thus avoiding the use of floating point arithmetic. (There is no
1466 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1467 double.
1469 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1470 2005 (difference between two addresses compared to an unsigned value).
1472 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1473 custom one, because of the following reported problem in Windows:
1475 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1476 under Win32.
1477 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1478 therefore missing the function definition.
1479 - The compiler thus generates a "C" signature for the test function.
1480 - The linker fails to find the "C" function.
1481 - PCRE fails to configure if asked to do so against libbz2.
1483 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1484 messages were output:
1486 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1487 rerunning libtoolize, to keep the correct libtool macros in-tree.
1488 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1490 I have done both of these things.
1492 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1493 most of the time, it *can* run out if it is given a pattern that contains a
1494 runaway infinite recursion. I updated the discussion in the pcrestack man
1495 page.
1497 13. Now that we have gone to the x.xx style of version numbers, the minor
1498 version may start with zero. Using 08 or 09 is a bad idea because users
1499 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1500 interpreted as invalid octal numbers. I've updated the previous comment in
1501 configure.ac, and also added a check that gives an error if 08 or 09 are
1502 used.
1504 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1505 causing partial matching to fail when the end of the subject matched \W
1506 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1508 15. There were some discrepancies between the declarations in pcre_internal.h
1509 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1510 their definitions. The declarations used "const uschar *" and the
1511 definitions used USPTR. Even though USPTR is normally defined as "const
1512 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1513 reported that: "This difference in casting confuses some C++ compilers, for
1514 example, SunCC recognizes above declarations as different functions and
1515 generates broken code for hbpcre." I have changed the declarations to use
1516 USPTR.
1518 16. GNU libtool is named differently on some systems. The autogen.sh script now
1519 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1520 (FreeBSD).
1522 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1523 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1524 comment: "Figure out how to create a longlong from a string: strtoll and
1525 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1526 instance, but it only takes 2 args instead of 3!"
1528 18. A subtle bug concerned with back references has been fixed by a change of
1529 specification, with a corresponding code fix. A pattern such as
1530 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1531 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1532 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1533 same bug. Such groups have to be quantified to be useful, or contained
1534 inside another quantified group. (If there's no repetition, the reference
1535 can never match.) The problem arises because, having left the group and
1536 moved on to the rest of the pattern, a later failure that backtracks into
1537 the group uses the captured value from the final iteration of the group
1538 rather than the correct earlier one. I have fixed this in PCRE by forcing
1539 any group that contains a reference to itself to be an atomic group; that
1540 is, there cannot be any backtracking into it once it has completed. This is
1541 similar to recursive and subroutine calls.
1544 Version 8.00 19-Oct-09
1545 ----------------------
1547 1. The table for translating pcre_compile() error codes into POSIX error codes
1548 was out-of-date, and there was no check on the pcre_compile() error code
1549 being within the table. This could lead to an OK return being given in
1550 error.
1552 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1553 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1554 in a Windows environment.
1556 3. The pcregrep --count option prints the count for each file even when it is
1557 zero, as does GNU grep. However, pcregrep was also printing all files when
1558 --files-with-matches was added. Now, when both options are given, it prints
1559 counts only for those files that have at least one match. (GNU grep just
1560 prints the file name in this circumstance, but including the count seems
1561 more useful - otherwise, why use --count?) Also ensured that the
1562 combination -clh just lists non-zero counts, with no names.
1564 4. The long form of the pcregrep -F option was incorrectly implemented as
1565 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1566 but it seems right to fix it, and I didn't think it was worth preserving
1567 the old behaviour.
1569 5. The command line items --regex=pattern and --regexp=pattern were not
1570 recognized by pcregrep, which required --regex pattern or --regexp pattern
1571 (with a space rather than an '='). The man page documented the '=' forms,
1572 which are compatible with GNU grep; these now work.
1574 6. No libpcreposix.pc file was created for pkg-config; there was just
1575 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1577 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1578 when UCP support is not needed, by modifying the Python script that
1579 generates it from Unicode data files. This should not matter if the module
1580 is correctly used as a library, but I received one complaint about 50K of
1581 unwanted data. My guess is that the person linked everything into his
1582 program rather than using a library. Anyway, it does no harm.
1584 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1585 was a minimum greater than 1 for a wide character in a possessive
1586 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1587 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1588 character. Chaos in the form of incorrect output or a compiling loop could
1589 result.
1591 9. The restrictions on what a pattern can contain when partial matching is
1592 requested for pcre_exec() have been removed. All patterns can now be
1593 partially matched by this function. In addition, if there are at least two
1594 slots in the offset vector, the offset of the earliest inspected character
1595 for the match and the offset of the end of the subject are set in them when
1596 PCRE_ERROR_PARTIAL is returned.
1598 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1599 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1600 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1601 and may be more useful for multi-segment matching.
1603 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1604 used to be given if ever the end of the subject was reached; now it is
1605 given only if matching could not proceed because another character was
1606 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1607 string "Z", which now yields "no match" instead of "partial match". In the
1608 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1609 final character ended with (*FAIL).
1611 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1612 if the pattern had a "must contain" character that was already found in the
1613 earlier partial match, unless partial matching was again requested. For
1614 example, with the pattern /dog.(body)?/, the "must contain" character is
1615 "g". If the first part-match was for the string "dog", restarting with
1616 "sbody" failed. This bug has been fixed.
1618 13. The string returned by pcre_dfa_exec() after a partial match has been
1619 changed so that it starts at the first inspected character rather than the
1620 first character of the match. This makes a difference only if the pattern
1621 starts with a lookbehind assertion or \b or \B (\K is not supported by
1622 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1623 matching functions compatible, and I think it's the right thing to do.
1625 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1626 so that the demonstration program is easily available in environments where
1627 PCRE has not been installed from source.
1629 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1630 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1631 library.
1633 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1634 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1635 is not the first non-POSIX option to be added. Clearly some people find
1636 these options useful.
1638 17. If a caller to the POSIX matching function regexec() passes a non-zero
1639 value for nmatch with a NULL value for pmatch, the value of
1640 nmatch is forced to zero.
1642 18. RunGrepTest did not have a test for the availability of the -u option of
1643 the diff command, as RunTest does. It now checks in the same way as
1644 RunTest, and also checks for the -b option.
1646 19. If an odd number of negated classes containing just a single character
1647 interposed, within parentheses, between a forward reference to a named
1648 subpattern and the definition of the subpattern, compilation crashed with
1649 an internal error, complaining that it could not find the referenced
1650 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1651 [The bug was that it was starting one character too far in when skipping
1652 over the character class, thus treating the ] as data rather than
1653 terminating the class. This meant it could skip too much.]
1655 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1656 /g option in pcretest when the pattern contains \K, which makes it possible
1657 to have an empty string match not at the start, even when the pattern is
1658 anchored. Updated pcretest and pcredemo to use this option.
1660 21. If the maximum number of capturing subpatterns in a recursion was greater
1661 than the maximum at the outer level, the higher number was returned, but
1662 with unset values at the outer level. The correct (outer level) value is
1663 now given.
1665 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1666 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1667 make it do so. The string so far is captured, making this feature
1668 compatible with Perl.
1670 23. The tests have been re-organized, adding tests 11 and 12, to make it
1671 possible to check the Perl 5.10 features against Perl 5.10.
1673 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1674 pattern matches a fixed length string. PCRE did not allow this; now it
1675 does. Neither allows recursion.
1677 25. I finally figured out how to implement a request to provide the minimum
1678 length of subject string that was needed in order to match a given pattern.
1679 (It was back references and recursion that I had previously got hung up
1680 on.) This code has now been added to pcre_study(); it finds a lower bound
1681 to the length of subject needed. It is not necessarily the greatest lower
1682 bound, but using it to avoid searching strings that are too short does give
1683 some useful speed-ups. The value is available to calling programs via
1684 pcre_fullinfo().
1686 26. While implementing 25, I discovered to my embarrassment that pcretest had
1687 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1688 study optimizations had never been tested with that matching function.
1689 Oops. What is worse, even when it was passed study data, there was a bug in
1690 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1691 were also very few tests of studied patterns with pcre_dfa_exec().
1693 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1694 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1695 on the other side of the coin, they are no longer allowed to have different
1696 names, because these cannot be distinguished in PCRE, and this has caused
1697 confusion. (This is a difference from Perl.)
1699 28. When duplicate subpattern names are present (necessarily with different
1700 numbers, as required by 27 above), and a test is made by name in a
1701 conditional pattern, either for a subpattern having been matched, or for
1702 recursion in such a pattern, all the associated numbered subpatterns are
1703 tested, and the overall condition is true if the condition is true for any
1704 one of them. This is the way Perl works, and is also more like the way
1705 testing by number works.
1708 Version 7.9 11-Apr-09
1709 ---------------------
1711 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1712 (pcretest), all targets were linked against these libraries. This included
1713 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1714 libraries. This caused unwanted dependencies to be created. This problem
1715 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1716 pcretest is linked with readline.
1718 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1719 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1720 moved outside it again, because FALSE and TRUE are already defined in AIX,
1721 but BOOL is not.
1723 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1724 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1726 4. The pcregrep documentation said spaces were inserted as well as colons (or
1727 hyphens) following file names and line numbers when outputting matching
1728 lines. This is not true; no spaces are inserted. I have also clarified the
1729 wording for the --colour (or --color) option.
1731 5. In pcregrep, when --colour was used with -o, the list of matching strings
1732 was not coloured; this is different to GNU grep, so I have changed it to be
1733 the same.
1735 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1736 each matching line was coloured. Now it goes on to look for further matches
1737 of any of the test patterns, which is the same behaviour as GNU grep.
1739 7. A pattern that could match an empty string could cause pcregrep to loop; it
1740 doesn't make sense to accept an empty string match in pcregrep, so I have
1741 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1742 seems to be how GNU grep behaves. [But see later change 40 for release
1743 8.33.]
1745 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1746 start or after a newline", because the conditional assertion was not being
1747 correctly handled. The rule now is that both the assertion and what follows
1748 in the first alternative must satisfy the test.
1750 9. If auto-callout was enabled in a pattern with a conditional group whose
1751 condition was an assertion, PCRE could crash during matching, both with
1752 pcre_exec() and pcre_dfa_exec().
1754 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1755 used for matching.
1757 11. Unicode property support in character classes was not working for
1758 characters (bytes) greater than 127 when not in UTF-8 mode.
1760 12. Added the -M command line option to pcretest.
1762 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1764 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1766 16. Added comments and documentation about mis-use of no_arg in the C++
1767 wrapper.
1769 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1770 from Martin Jerabek that uses macro names for all relevant character and
1771 string constants.
1773 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1774 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1775 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1776 these, but not everybody uses configure.
1778 19. A conditional group that had only one branch was not being correctly
1779 recognized as an item that could match an empty string. This meant that an
1780 enclosing group might also not be so recognized, causing infinite looping
1781 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1782 with the subject "ab", where knowledge that the repeated group can match
1783 nothing is needed in order to break the loop.
1785 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1786 exec(), but without supplying a callout function, matching went wrong.
1788 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1789 leak if the size of the offset vector was greater than 30. When the vector
1790 is smaller, the saved offsets during recursion go onto a local stack
1791 vector, but for larger vectors malloc() is used. It was failing to free
1792 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1793 error, in fact).
1795 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1796 heapframe that is used only when UTF-8 support is enabled. This caused no
1797 problem, but was untidy.
1799 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1800 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1801 included within another project.
1803 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1804 slightly modified by me:
1806 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1807 not building pcregrep.
1809 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1810 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1812 25. Forward references, both numeric and by name, in patterns that made use of
1813 duplicate group numbers, could behave incorrectly or give incorrect errors,
1814 because when scanning forward to find the reference group, PCRE was not
1815 taking into account the duplicate group numbers. A pattern such as
1816 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1818 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1819 the feature of a custom pointer more persuasive (as requested by a user).
1821 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1822 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1823 pre-defined.
1825 28. Added support for (*UTF8) at the start of a pattern.
1827 29. Arrange for flags added by the "release type" setting in CMake to be shown
1828 in the configuration summary.
1831 Version 7.8 05-Sep-08
1832 ---------------------
1834 1. Replaced UCP searching code with optimized version as implemented for Ad
1835 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1836 stage table and inline lookup instead of a function, giving speed ups of 2
1837 to 5 times on some simple patterns that I tested. Permission was given to
1838 distribute the MultiStage2.py script that generates the tables (it's not in
1839 the tarball, but is in the Subversion repository).
1841 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1842 scripts.
1844 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1845 a group with a zero qualifier. The result of the study could be incorrect,
1846 or the function might crash, depending on the pattern.
1848 4. Caseless matching was not working for non-ASCII characters in back
1849 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1850 It now works when Unicode Property Support is available.
1852 5. In pcretest, an escape such as \x{de} in the data was always generating
1853 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1854 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1855 truncation.
1857 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1859 7. Added two (int) casts to pcregrep when printing the difference of two
1860 pointers, in case they are 64-bit values.
1862 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1863 test 2 if it fails.
1865 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1866 and a #define of that name to empty if it is not externally set. This is to
1867 allow users of MSVC to set it if necessary.
1869 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1870 the convenience functions in the pcre_get.c source file.
1872 11. An option change at the start of a pattern that had top-level alternatives
1873 could cause overwriting and/or a crash. This command provoked a crash in
1874 some environments:
1876 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1878 This potential security problem was recorded as CVE-2008-2371.
1880 12. For a pattern where the match had to start at the beginning or immediately
1881 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1882 pcre_dfa_exec() could read past the end of the passed subject if there was
1883 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1884 pcretest so that it places the subject at the end of its malloc-ed buffer.
1886 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1887 exec() might read past the end of the data buffer in UTF-8 mode.
1889 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1890 the data contained the byte 0x85 as part of a UTF-8 character within its
1891 first line. This applied both to normal and DFA matching.
1893 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1894 /^[^d]*?$/8 failed to match "abc".
1896 16. Added a missing copyright notice to pcrecpp_internal.h.
1898 17. Make it more clear in the documentation that values returned from
1899 pcre_exec() in ovector are byte offsets, not character counts.
1901 18. Tidied a few places to stop certain compilers from issuing warnings.
1903 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1904 supplied by Stefan Weber. I made a further small update for 7.8 because
1905 there is a change of source arrangements: the pcre_searchfuncs.c module is
1906 replaced by pcre_ucd.c.
1909 Version 7.7 07-May-08
1910 ---------------------
1912 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1913 a string to a long long, pretend we don't even have a long long." This is
1914 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1916 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1917 pre-7.6 versions, which defined a global no_arg variable instead of putting
1918 it in the RE class. (See also #8 below.)
1920 3. Remove a line of dead code, identified by coverity and reported by Nuno
1921 Lopes.
1923 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1925 (1) The include/exclude patterns were being applied to the whole pathnames
1926 of files, instead of just to the final components.
1928 (2) If there was more than one level of directory, the subdirectories were
1929 skipped unless they satisfied the include/exclude conditions. This is
1930 inconsistent with GNU grep (and could even be seen as contrary to the
1931 pcregrep specification - which I improved to make it absolutely clear).
1932 The action now is always to scan all levels of directory, and just
1933 apply the include/exclude patterns to regular files.
1935 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1936 --exclude_dir in the tests to avoid scanning .svn directories.
1938 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1939 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1940 doesn't support NULs in patterns.
1942 7. Added some missing "const"s to declarations of static tables in
1943 pcre_compile.c and pcre_dfa_exec.c.
1945 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1946 caused by fix #2 above. (Subsequently also a second patch to fix the
1947 first patch. And a third patch - this was a messy problem.)
1949 9. Applied Craig's patch to remove the use of push_back().
1951 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1952 matching function regexec().
1954 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1955 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1956 references. PCRE supports relative numbers with this syntax (I don't think
1957 Oniguruma does).
1959 12. Previously, a group with a zero repeat such as (...){0} was completely
1960 omitted from the compiled regex. However, this means that if the group
1961 was called as a subroutine from elsewhere in the pattern, things went wrong
1962 (an internal error was given). Such groups are now left in the compiled
1963 pattern, with a new opcode that causes them to be skipped at execution
1964 time.
1966 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1967 to the way PCRE behaves:
1969 (a) A lone ] character is dis-allowed (Perl treats it as data).
1971 (b) A back reference to an unmatched subpattern matches an empty string
1972 (Perl fails the current match path).
1974 (c) A data ] in a character class must be notated as \] because if the
1975 first data character in a class is ], it defines an empty class. (In
1976 Perl it is not possible to have an empty class.) The empty class []
1977 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1978 The negative empty class [^] matches any one character, independently
1979 of the DOTALL setting.
1981 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1982 non-existent subpattern following a character class starting with ']' and
1983 containing () gave an internal compiling error instead of "reference to
1984 non-existent subpattern". Fortunately, when the pattern did exist, the
1985 compiled code was correct. (When scanning forwards to check for the
1986 existence of the subpattern, it was treating the data ']' as terminating
1987 the class, so got the count wrong. When actually compiling, the reference
1988 was subsequently set up correctly.)
1990 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1991 it was being rejected as not supported by pcre_dfa_exec(), even though
1992 other assertions are supported. I have made pcre_dfa_exec() support
1993 (*FAIL).
1995 16. The implementation of 13c above involved the invention of a new opcode,
1996 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1997 cannot be changed at match time, I realized I could make a small
1998 improvement to matching performance by compiling OP_ALLANY instead of
1999 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2000 on the OP_ANY path.
2002 17. Compiling pcretest on Windows with readline support failed without the
2003 following two fixes: (1) Make the unistd.h include conditional on
2004 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2006 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2007 ncurses library to be included for pcretest when ReadLine support is
2008 requested, but also to allow for it to be overridden. This patch came from
2009 Daniel Bergström.
2011 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2012 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2013 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2014 spotting this.
2017 Version 7.6 28-Jan-08
2018 ---------------------
2020 1. A character class containing a very large number of characters with
2021 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2022 overflow.
2024 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2025 HAVE_LONG_LONG is not defined.
2027 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2028 bring them up to date and include new features. This patch includes:
2030 - Fixed PH's badly added libz and libbz2 support.
2031 - Fixed a problem with static linking.
2032 - Added pcredemo. [But later removed - see 7 below.]
2033 - Fixed dftables problem and added an option.
2034 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2036 - Added readline support for pcretest.
2037 - Added an listing of the option settings after cmake has run.
2039 4. A user submitted a patch to Makefile that makes it easy to create
2040 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2041 Makefile.am that cause it to include this special target, without
2042 affecting anything else. Note that the same mingw target plus all
2043 the other distribution libraries and programs are now supported
2044 when configuring with CMake (see 6 below) instead of with
2045 Configure/Make.
2047 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2048 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2049 exported in the Windows port". It has not yet been confirmed that the patch
2050 solves the problem, but it does no harm.
2052 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2053 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2054 with CMake, and also correct the comment about stack recursion.
2056 7. Remove the automatic building of pcredemo from the ./configure system and
2057 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2058 of a program that users should build themselves after PCRE is installed, so
2059 building it automatically is not really right. What is more, it gave
2060 trouble in some build environments.
2062 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2065 Version 7.5 10-Jan-08
2066 ---------------------
2068 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2069 values in parens when parsing an RE using the C++ wrapper."
2071 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2072 Characters greater than 255 were excluded from the class instead of being
2073 included.
2075 3. The same bug as (2) above applied to negated POSIX classes such as
2076 [:^space:].
2078 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2079 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2080 I have changed it.
2082 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2083 first named subpattern) and a construct such as (?&a) would reference the
2084 first named subpattern whose name started with "a" (in other words, the
2085 length check was missing). Both these problems are fixed. "Subpattern name
2086 expected" is now given for (?&) (a zero-length name), and this patch also
2087 makes it give the same error for \k'' (previously it complained that that
2088 was a reference to a non-existent subpattern).
2090 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2091 this is right because (?- can be followed by option settings as well as by
2092 digits. I have, however, made the messages clearer.
2094 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2095 than the number used in the conditional) now cause a compile-time error.
2096 This is actually not compatible with Perl, which accepts such patterns, but
2097 treats the conditional as always being FALSE (as PCRE used to), but it
2098 seems to me that giving a diagnostic is better.
2100 8. Change "alphameric" to the more common word "alphanumeric" in comments
2101 and messages.
2103 9. Fix two occurrences of "backslash" in comments that should have been
2104 "backspace".
2106 10. Remove two redundant lines of code that can never be obeyed (their function
2107 was moved elsewhere).
2109 11. The program that makes PCRE's Unicode character property table had a bug
2110 which caused it to generate incorrect table entries for sequences of
2111 characters that have the same character type, but are in different scripts.
2112 It amalgamated them into a single range, with the script of the first of
2113 them. In other words, some characters were in the wrong script. There were
2114 thirteen such cases, affecting characters in the following ranges:
2116 U+002b0 - U+002c1
2117 U+0060c - U+0060d
2118 U+0061e - U+00612
2119 U+0064b - U+0065e
2120 U+0074d - U+0076d
2121 U+01800 - U+01805
2122 U+01d00 - U+01d77
2123 U+01d9b - U+01dbf
2124 U+0200b - U+0200f
2125 U+030fc - U+030fe
2126 U+03260 - U+0327f
2127 U+0fb46 - U+0fbb1
2128 U+10450 - U+1049d
2130 12. The -o option (show only the matching part of a line) for pcregrep was not
2131 compatible with GNU grep in that, if there was more than one match in a
2132 line, it showed only the first of them. It now behaves in the same way as
2133 GNU grep.
2135 13. If the -o and -v options were combined for pcregrep, it printed a blank
2136 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2137 does the same. The return code can be used to tell if there were any
2138 non-matching lines.
2140 14. Added --file-offsets and --line-offsets to pcregrep.
2142 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2143 infinitely looping recursion. The bug was that positive lookaheads were not
2144 being skipped when checking for a possible empty match (negative lookaheads
2145 and both kinds of lookbehind were skipped).
2147 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2148 inclusion of <windows.h> to before rather than after the definition of
2149 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2151 17. Specifying a possessive quantifier with a specific limit for a Unicode
2152 character property caused pcre_compile() to compile bad code, which led at
2153 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2154 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2155 caused the error; without that there was no problem.
2157 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2159 19. Added --enable-pcretest-libreadline.
2161 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2162 RE::GlobalReplace(). As a result, the number of replacements returned was
2163 double what it should be. I removed one of the increments, but Craig sent a
2164 later patch that removed the other one (the right fix) and added unit tests
2165 that check the return values (which was not done before).
2167 21. Several CMake things:
2169 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2170 the names libpcre and libpcreposix, not just pcre and pcreposix.
2172 (2) The above change means that pcretest and pcregrep are now correctly
2173 linked with the newly-built libraries, not previously installed ones.
2177 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2178 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2179 UTF-8 newline character). The key issue is that the pattern starts .*;
2180 this means that the match must be either at the beginning, or after a
2181 newline. The bug was in the code for advancing after a failed match and
2182 checking that the new position followed a newline. It was not taking
2183 account of UTF-8 characters correctly.
2185 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2186 character classes. PCRE was not treating the sequence [:...:] as a
2187 character class unless the ... were all letters. Perl, however, seems to
2188 allow any characters between [: and :], though of course it rejects as
2189 unknown any "names" that contain non-letters, because all the known class
2190 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2191 for example, whereas PCRE did not - it did not recognize a POSIX character
2192 class. This seemed a bit dangerous, so the code has been changed to be
2193 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2194 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2195 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2196 Perl does, and where it didn't before.
2198 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2199 Windows environments %n is disabled by default.
2202 Version 7.4 21-Sep-07
2203 ---------------------
2205 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2206 means that a class such as [\s] counted as "explicit reference to CR or
2207 LF". That isn't really right - the whole point of the change was to try to
2208 help when there was an actual mention of one of the two characters. So now
2209 the change happens only if \r or \n (or a literal CR or LF) character is
2210 encountered.
2212 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2213 of both had grown to the point where there were only 3 bits left.
2214 Fortunately, there was spare space in the data structure, and so I have
2215 moved the internal flags into a new 16-bit field to free up more option
2216 bits.
2218 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2219 but did not set the internal JCHANGED flag - either of these is enough to
2220 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2221 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2222 start sets both bits.
2224 4. Added options (at build time, compile time, exec time) to change \R from
2225 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2227 5. doc/pcresyntax.html was missing from the distribution.
2229 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2230 compatibility, even though it is no longer used.
2232 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2233 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2234 windows.h file is present (where different names are used). [This was
2235 reversed later after testing - see 16 below.]
2237 8. Changed all #include <config.h> to #include "config.h". There were also
2238 some further <pcre.h> cases that I changed to "pcre.h".
2240 9. When pcregrep was used with the --colour option, it missed the line ending
2241 sequence off the lines that it output.
2243 10. It was pointed out to me that arrays of string pointers cause lots of
2244 relocations when a shared library is dynamically loaded. A technique of
2245 using a single long string with a table of offsets can drastically reduce
2246 these. I have refactored PCRE in four places to do this. The result is
2247 dramatic:
2249 Originally: 290
2250 After changing UCP table: 187
2251 After changing error message table: 43
2252 After changing table of "verbs" 36
2253 After changing table of Posix names 22
2255 Thanks to the folks working on Gregex for glib for this insight.
2257 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2258 unicode-properties was also set.
2260 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2262 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2263 checked only for CRLF.
2265 14. Added casts to pcretest.c to avoid compiler warnings.
2267 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2269 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2270 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2271 entirely. This removes changes made in 7 above.
2273 17. The CMake files have been updated, and there is now more information about
2274 building with CMake in the NON-UNIX-USE document.
2277 Version 7.3 28-Aug-07
2278 ---------------------
2280 1. In the rejigging of the build system that eventually resulted in 7.1, the
2281 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2282 brackets there is not right, since it causes compilers to look for an
2283 installed pcre.h, not the version that is in the source that is being
2284 compiled (which of course may be different). I have changed it back to:
2286 #include "pcre.h"
2288 I have a vague recollection that the change was concerned with compiling in
2289 different directories, but in the new build system, that is taken care of
2290 by the VPATH setting the Makefile.
2292 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2293 when the subject happened to end in the byte 0x85 (e.g. if the last
2294 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2295 characters but of course it shouldn't be taken as a newline when it is part
2296 of another character. The bug was that, for an unlimited repeat of . in
2297 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2298 characters when looking for a newline.
2300 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2302 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2303 in debug output.
2305 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2306 long printing in the pcrecpp unittest when running under MinGW.
2308 6. ESC_K was left out of the EBCDIC table.
2310 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2311 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2312 limit also applies to "virtual nesting" when a pattern is recursive, and in
2313 this case 1000 isn't so big. I have been able to remove this limit at the
2314 expense of backing off one optimization in certain circumstances. Normally,
2315 when pcre_exec() would call its internal match() function recursively and
2316 immediately return the result unconditionally, it uses a "tail recursion"
2317 feature to save stack. However, when a subpattern that can match an empty
2318 string has an unlimited repetition quantifier, it no longer makes this
2319 optimization. That gives it a stack frame in which to save the data for
2320 checking that an empty string has been matched. Previously this was taken
2321 from the 1000-entry workspace that had been reserved. So now there is no
2322 explicit limit, but more stack is used.
2324 8. Applied Daniel's patches to solve problems with the import/export magic
2325 syntax that is required for Windows, and which was going wrong for the
2326 pcreposix and pcrecpp parts of the library. These were overlooked when this
2327 problem was solved for the main library.
2329 9. There were some crude static tests to avoid integer overflow when computing
2330 the size of patterns that contain repeated groups with explicit upper
2331 limits. As the maximum quantifier is 65535, the maximum group length was
2332 set at 30,000 so that the product of these two numbers did not overflow a
2333 32-bit integer. However, it turns out that people want to use groups that
2334 are longer than 30,000 bytes (though not repeat them that many times).
2335 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2336 made it possible to implement the integer overflow checks in a much more
2337 dynamic way, which I have now done. The artificial limitation on group
2338 length has been removed - we now have only the limit on the total length of
2339 the compiled pattern, which depends on the LINK_SIZE setting.
2341 10. Fixed a bug in the documentation for get/copy named substring when
2342 duplicate names are permitted. If none of the named substrings are set, the
2343 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2344 empty string.
2346 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2347 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2348 because the ] is interpreted as the first data character and the
2349 terminating ] is not found. PCRE has been made compatible with Perl in this
2350 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2351 cause memory overwriting.
2353 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2354 string has been matched (to stop an infinite loop). It was not recognizing
2355 a conditional subpattern that could match an empty string if that
2356 subpattern was within another subpattern. For example, it looped when
2357 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2358 condition was not nested. This bug has been fixed.
2360 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2361 past the start of the subject in the presence of bytes with the top bit
2362 set, for example "\x8aBCD".
2364 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2365 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2367 14. Optimized (?!) to (*FAIL).
2369 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2370 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2371 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2372 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2373 does: it's just the validity check that is more restrictive.
2375 16. Inserted checks for integer overflows during escape sequence (backslash)
2376 processing, and also fixed erroneous offset values for syntax errors during
2377 backslash processing.
2379 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2380 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2382 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2383 caused an overrun.
2385 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2386 something other than just ASCII characters) inside a group that had an
2387 unlimited repeat caused a loop at compile time (while checking to see
2388 whether the group could match an empty string).
2390 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2391 [\P{Any}] did so. (Error in the code for printing property names.)
2393 21. An orphan \E inside a character class could cause a crash.
2395 22. A repeated capturing bracket such as (A)? could cause a wild memory
2396 reference during compilation.
2398 23. There are several functions in pcre_compile() that scan along a compiled
2399 expression for various reasons (e.g. to see if it's fixed length for look
2400 behind). There were bugs in these functions when a repeated \p or \P was
2401 present in the pattern. These operators have additional parameters compared
2402 with \d, etc, and these were not being taken into account when moving along
2403 the compiled data. Specifically:
2405 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2406 length.
2408 (b) An item such as \pL+ within a repeated group could cause crashes or
2409 loops.
2411 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2412 "reference to non-existent subpattern" error.
2414 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2416 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2417 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2419 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2421 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2422 character were causing crashes (broken optimization).
2424 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2425 \p or \P) caused a compile-time loop.
2427 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2428 break. For example, the unstudied pattern [\r\n]A does not match the string
2429 "\r\nA" because change 7.0/46 below moves the current point on by two
2430 characters after failing to match at the start. However, the pattern \nA
2431 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2432 the same is true. There doesn't seem any very clean way out of this, but
2433 what I have chosen to do makes the common cases work: PCRE now takes note
2434 of whether there can be an explicit match for \r or \n anywhere in the
2435 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2436 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2437 pattern has explicit CR or LF references.
2439 29. Added (*CR) etc for changing newline setting at start of pattern.
2442 Version 7.2 19-Jun-07
2443 ---------------------
2445 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2446 which is apparently normally available under Windows.
2448 2. Re-jig the pcregrep tests with different newline settings in an attempt
2449 to make them independent of the local environment's newline setting.
2451 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2453 4. Some of the "internals" tests were previously cut out when the link size
2454 was not 2, because the output contained actual offsets. The recent new
2455 "Z" feature of pcretest means that these can be cut out, making the tests
2456 usable with all link sizes.
2458 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2459 stack recursion. This gives a massive performance boost under BSD, but just
2460 a small improvement under Linux. However, it saves one field in the frame
2461 in all cases.
2463 6. Added more features from the forthcoming Perl 5.10:
2465 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2466 recursion call. It refers to the nth most recently opened parentheses.
2468 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2469 to be opened parentheses.
2471 (c) Conditions that refer to capturing parentheses can be specified
2472 relatively, for example, (?(-2)... or (?(+3)...
2474 (d) \K resets the start of the current match so that everything before
2475 is not part of it.
2477 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2479 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2480 reference syntax.
2482 (g) (?| introduces a group in which the numbering of parentheses in each
2483 alternative starts with the same number.
2485 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2487 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2490 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2491 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2492 for detecting groups that can match an empty string.
2494 9. A pattern with a very large number of alternatives (more than several
2495 hundred) was running out of internal workspace during the pre-compile
2496 phase, where pcre_compile() figures out how much memory will be needed. A
2497 bit of new cunning has reduced the workspace needed for groups with
2498 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2499 workspace instead of running out of the 4096 that are available.
2501 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2503 11. Applied patch from Google to remove an optimization that didn't quite work.
2504 The report of the bug said:
2506 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2507 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2508 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2510 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2511 it matched the wrong number of bytes.
2514 Version 7.1 24-Apr-07
2515 ---------------------
2517 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2518 that is more "standard", making use of automake and other Autotools. There
2519 is some re-arrangement of the files and adjustment of comments consequent
2520 on this.
2522 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2523 for recursive directory scanning broke on some systems because the files
2524 are not scanned in any specific order and on different systems the order
2525 was different. A call to "sort" has been inserted into RunGrepTest for the
2526 approprate test as a short-term fix. In the longer term there may be an
2527 alternative.
2529 3. I had an email from Eric Raymond about problems translating some of PCRE's
2530 man pages to HTML (despite the fact that I distribute HTML pages, some
2531 people do their own conversions for various reasons). The problems
2532 concerned the use of low-level troff macros .br and .in. I have therefore
2533 removed all such uses from the man pages (some were redundant, some could
2534 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2535 HTML has been updated to handle .nf/.fi and to complain if it encounters
2536 .br or .in.
2538 4. Updated comments in configure.ac that get placed in config.h.in and also
2539 arranged for config.h to be included in the distribution, with the name
2540 config.h.generic, for the benefit of those who have to compile without
2541 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2543 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2544 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2545 makevp.bat for the current PCRE, using the additional files
2546 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2548 6. A Windows user reported a minor discrepancy with test 2, which turned out
2549 to be caused by a trailing space on an input line that had got lost in his
2550 copy. The trailing space was an accident, so I've just removed it.
2552 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2553 that is needed.
2555 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2556 as "const" (a) because they are and (b) because it helps the PHP
2557 maintainers who have recently made a script to detect big data structures
2558 in the php code that should be moved to the .rodata section. I remembered
2559 to update Builducptable as well, so it won't revert if ucptable.h is ever
2560 re-created.
2562 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2563 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2564 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2565 support is not required. This saves 1.5-2K of code, which is important in
2566 some applications.
2568 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2569 so as not to refer to the tables, even though these functions will never be
2570 called when UTF-8 support is disabled. Otherwise there are problems with a
2571 shared library.
2573 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2575 (a) It was defining its arguments as char * instead of void *.
2577 (b) It was assuming that all moves were upwards in memory; this was true
2578 a long time ago when I wrote it, but is no longer the case.
2580 The emulated memove() is provided for those environments that have neither
2581 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2582 is clearly not the case, as these two bugs were recently reported.
2584 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2585 and Detrail to create the HTML documentation, the .txt form of the man
2586 pages, and it removes trailing spaces from listed files. It also creates
2587 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2588 case, it wraps all the #defines with #ifndefs. This script should be run
2589 before "make dist".
2591 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2592 with Unicode property support.
2594 (a) For a maximizing quantifier, if the two different cases of the
2595 character were of different lengths in their UTF-8 codings (there are
2596 some cases like this - I found 11), and the matching function had to
2597 back up over a mixture of the two cases, it incorrectly assumed they
2598 were both the same length.
2600 (b) When PCRE was configured to use the heap rather than the stack for
2601 recursion during matching, it was not correctly preserving the data for
2602 the other case of a UTF-8 character when checking ahead for a match
2603 while processing a minimizing repeat. If the check also involved
2604 matching a wide character, but failed, corruption could cause an
2605 erroneous result when trying to check for a repeat of the original
2606 character.
2608 13. Some tidying changes to the testing mechanism:
2610 (a) The RunTest script now detects the internal link size and whether there
2611 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2612 values substituted by "configure". (The RunGrepTest script already did
2613 this for UTF-8.) The configure.ac script no longer substitutes the
2614 relevant variables.
2616 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2617 with length and offset values. This means that the output is different
2618 for different internal link sizes. Test 2 is skipped for link sizes
2619 other than 2 because of this, bypassing the problem. Unfortunately,
2620 there was also a test in test 3 (the locale tests) that used /B and
2621 failed for link sizes other than 2. Rather than cut the whole test out,
2622 I have added a new /Z option to pcretest that replaces the length and
2623 offset values with spaces. This is now used to make test 3 independent
2624 of link size. (Test 2 will be tidied up later.)
2626 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2627 segmentation fault instead of returning the appropriate error message.
2629 15. In multiline mode when the newline sequence was set to "any", the pattern
2630 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2631 This doesn't seem right; it now treats the CRLF combination as the line
2632 ending, and so does not match in that case. It's only a pattern such as ^$
2633 that would hit this one: something like ^ABC$ would have failed after \r
2634 and then tried again after \r\n.
2636 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2637 in an attempt to make files that differ only in their line terminators
2638 compare equal. This works on Linux.
2640 17. Under certain error circumstances pcregrep might try to free random memory
2641 as it exited. This is now fixed, thanks to valgrind.
2643 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2644 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2645 was because its rules for how to advance for /g after matching an empty
2646 string at the end of a line did not allow for this case. They now check for
2647 it specially.
2649 20. pcretest is supposed to handle patterns and data of any length, by
2650 extending its buffers when necessary. It was getting this wrong when the
2651 buffer for a data line had to be extended.
2653 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2654 CRLF as a newline sequence.
2656 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2657 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2658 I have nevertheless tidied it up.
2660 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2662 24. Added a man page for pcre-config.
2665 Version 7.0 19-Dec-06
2666 ---------------------
2668 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2669 moving to gcc 4.1.1.
2671 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2672 sys/time.h, which is documented as needed for this function. It doesn't
2673 seem to matter on Linux, but it showed up on some releases of OS X.
2675 3. It seems that there are systems where bytes whose values are greater than
2676 127 match isprint() in the "C" locale. The "C" locale should be the
2677 default when a C program starts up. In most systems, only ASCII printing
2678 characters match isprint(). This difference caused the output from pcretest
2679 to vary, making some of the tests fail. I have changed pcretest so that:
2681 (a) When it is outputting text in the compiled version of a pattern, bytes
2682 other than 32-126 are always shown as hex escapes.
2684 (b) When it is outputting text that is a matched part of a subject string,
2685 it does the same, unless a different locale has been set for the match
2686 (using the /L modifier). In this case, it uses isprint() to decide.
2688 4. Fixed a major bug that caused incorrect computation of the amount of memory
2689 required for a compiled pattern when options that changed within the
2690 pattern affected the logic of the preliminary scan that determines the
2691 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2692 that the computed length was too small. The symptoms of this bug were
2693 either the PCRE error "internal error: code overflow" from pcre_compile(),
2694 or a glibc crash with a message such as "pcretest: free(): invalid next
2695 size (fast)". Examples of patterns that provoked this bug (shown in
2696 pcretest format) are:
2698 /(?-x: )/x
2699 /(?x)(?-x: \s*#\s*)/
2700 /((?i)[\x{c0}])/8
2701 /(?i:[\x{c0}])/8
2703 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2704 is now done differently.
2706 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2707 wrapper classes; (b) implement a new function in the C++ scanner that is
2708 more efficient than the old way of doing things because it avoids levels of
2709 recursion in the regex matching; (c) add a paragraph to the documentation
2710 for the FullMatch() function.
2712 6. The escape sequence \n was being treated as whatever was defined as
2713 "newline". Not only was this contrary to the documentation, which states
2714 that \n is character 10 (hex 0A), but it also went horribly wrong when
2715 "newline" was defined as CRLF. This has been fixed.
2717 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2718 was being set to -1 for the "end of line" case (supposedly a value that no
2719 character can have). Though this value is never used (the check for end of
2720 line is "zero bytes in current character"), it caused compiler complaints.
2721 I've changed it to 0xffffffff.
2723 8. In pcre_version.c, the version string was being built by a sequence of
2724 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2725 string (as it is for production releases) called a macro with an empty
2726 argument. The C standard says the result of this is undefined. The gcc
2727 compiler treats it as an empty string (which was what was wanted) but it is
2728 reported that Visual C gives an error. The source has been hacked around to
2729 avoid this problem.
2731 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2732 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2733 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2734 of them did).
2736 10. Originally, pcretest opened its input and output without "b"; then I was
2737 told that "b" was needed in some environments, so it was added for release
2738 5.0 to both the input and output. (It makes no difference on Unix-like
2739 systems.) Later I was told that it is wrong for the input on Windows. I've
2740 now abstracted the modes into two macros, to make it easier to fiddle with
2741 them, and removed "b" from the input mode under Windows.
2743 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2745 12. Added -help and --help to pcretest as an official way of being reminded
2746 of the options.
2748 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2749 and pcrecpp.cc because they annoy compilers at high warning levels.
2751 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2753 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2754 scripts are not C programs :-) and which was not noticed because it works
2755 on Linux.
2757 16. pcretest is supposed to handle any length of pattern and data line (as one
2758 line or as a continued sequence of lines) by extending its input buffer if
2759 necessary. This feature was broken for very long pattern lines, leading to
2760 a string of junk being passed to pcre_compile() if the pattern was longer
2761 than about 50K.
2763 17. I have done a major re-factoring of the way pcre_compile() computes the
2764 amount of memory needed for a compiled pattern. Previously, there was code
2765 that made a preliminary scan of the pattern in order to do this. That was
2766 OK when PCRE was new, but as the facilities have expanded, it has become
2767 harder and harder to keep it in step with the real compile phase, and there
2768 have been a number of bugs (see for example, 4 above). I have now found a
2769 cunning way of running the real compile function in a "fake" mode that
2770 enables it to compute how much memory it would need, while actually only
2771 ever using a few hundred bytes of working memory and without too many
2772 tests of the mode. This should make future maintenance and development
2773 easier. A side effect of this work is that the limit of 200 on the nesting
2774 depth of parentheses has been removed (though this was never a serious
2775 limitation, I suspect). However, there is a downside: pcre_compile() now
2776 runs more slowly than before (30% or more, depending on the pattern). I
2777 hope this isn't a big issue. There is no effect on runtime performance.
2779 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2780 newline (only possible for the last line of a file) and it was a
2781 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2783 19. Added additional timing features to pcretest. (1) The -tm option now times
2784 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2785 separate command line item, by a number that specifies the number of
2786 repeats to use when timing. The default is 50000; this gives better
2787 precision, but takes uncomfortably long for very large patterns.
2789 20. Extended pcre_study() to be more clever in cases where a branch of a
2790 subpattern has no definite first character. For example, (a*|b*)[cd] would
2791 previously give no result from pcre_study(). Now it recognizes that the
2792 first character must be a, b, c, or d.
2794 21. There was an incorrect error "recursive call could loop indefinitely" if
2795 a subpattern (or the entire pattern) that was being tested for matching an
2796 empty string contained only one non-empty item after a nested subpattern.
2797 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2798 incorrectly, because the \d was being skipped in the check.
2800 22. The pcretest program now has a new pattern option /B and a command line
2801 option -b, which is equivalent to adding /B to every pattern. This causes
2802 it to show the compiled bytecode, without the additional information that
2803 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2804 is the same as /B/I).
2806 23. A new optimization is now able automatically to treat some sequences such
2807 as a*b as a*+b. More specifically, if something simple (such as a character
2808 or a simple class like \d) has an unlimited quantifier, and is followed by
2809 something that cannot possibly match the quantified thing, the quantifier
2810 is automatically "possessified".
2812 24. A recursive reference to a subpattern whose number was greater than 39
2813 went wrong under certain circumstances in UTF-8 mode. This bug could also
2814 have affected the operation of pcre_study().
2816 25. Realized that a little bit of performance could be had by replacing
2817 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2819 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2821 27. Possessive quantifiers such as a++ were previously implemented by turning
2822 them into atomic groups such as ($>a+). Now they have their own opcodes,
2823 which improves performance. This includes the automatically created ones
2824 from 23 above.
2826 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2827 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2828 the first matched character to be a colon. This applied both to named and
2829 numbered groups.
2831 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2833 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2834 building PCRE on Windows easier, so I have included it in the distribution.
2836 31. There is now a check in pcretest against a ridiculously large number being
2837 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2838 loop, the loop is abandoned.
2840 32. Forward references to subpatterns in conditions such as (?(2)...) where
2841 subpattern 2 is defined later cause pcre_compile() to search forwards in
2842 the pattern for the relevant set of parentheses. This search went wrong
2843 when there were unescaped parentheses in a character class, parentheses
2844 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2846 33. "Subroutine" calls and backreferences were previously restricted to
2847 referencing subpatterns earlier in the regex. This restriction has now
2848 been removed.
2850 34. Added a number of extra features that are going to be in Perl 5.10. On the
2851 whole, these are just syntactic alternatives for features that PCRE had
2852 previously implemented using the Python syntax or my own invention. The
2853 other formats are all retained for compatibility.
2855 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2856 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2857 also .NET compatible.
2859 (b) A recursion or subroutine call to a named group can now be defined as
2860 (?&name) as well as (?P>name).
2862 (c) A backreference to a named group can now be defined as \k<name> or
2863 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2864 5.10, are also .NET compatible.
2866 (d) A conditional reference to a named group can now use the syntax
2867 (?(<name>) or (?('name') as well as (?(name).
2869 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2870 groups (named and numbered) that are never evaluated inline, but can be
2871 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2872 is always false. There may be only one alternative in such a group.
2874 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2875 as the simple (?(R). The condition is true only if the most recent
2876 recursion is that of the given number or name. It does not search out
2877 through the entire recursion stack.
2879 (g) The escape \gN or \g{N} has been added, where N is a positive or
2880 negative number, specifying an absolute or relative reference.
2882 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2883 some "unreachable code" warnings.
2885 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2886 things, this adds five new scripts.
2888 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2889 There were also incompatibilities regarding the handling of \Q..\E inside
2890 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2891 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2893 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2894 matches an empty string, and forcibly breaks the loop. There were bugs in
2895 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2896 against aaaa the result was just "a" rather than "aaaa", for example. Two
2897 separate and independent bugs (that affected different cases) have been
2898 fixed.
2900 39. Refactored the code to abolish the use of different opcodes for small
2901 capturing bracket numbers. This is a tidy that I avoided doing when I
2902 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2903 The new approach is not only tidier, it makes it possible to reduce the
2904 memory needed to fix the previous bug (38).
2906 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2907 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2908 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2909 mode.
2911 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2912 report.
2914 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2915 copy construction and assignment in the C++ wrapper.
2917 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2918 couple of bugs that could have given wrong results in the "--newline=crlf"
2919 case.
2921 44. Added a number of casts and did some reorganization of signed/unsigned int
2922 variables following suggestions from Dair Grant. Also renamed the variable
2923 "this" as "item" because it is a C++ keyword.
2925 45. Arranged for dftables to add
2927 #include "pcre_internal.h"
2929 to pcre_chartables.c because without it, gcc 4.x may remove the array
2930 definition from the final binary if PCRE is built into a static library and
2931 dead code stripping is activated.
2933 46. For an unanchored pattern, if a match attempt fails at the start of a
2934 newline sequence, and the newline setting is CRLF or ANY, and the next two
2935 characters are CRLF, advance by two characters instead of one.
2938 Version 6.7 04-Jul-06
2939 ---------------------
2941 1. In order to handle tests when input lines are enormously long, pcretest has
2942 been re-factored so that it automatically extends its buffers when
2943 necessary. The code is crude, but this _is_ just a test program. The
2944 default size has been increased from 32K to 50K.
2946 2. The code in pcre_study() was using the value of the re argument before
2947 testing it for NULL. (Of course, in any sensible call of the function, it
2948 won't be NULL.)
2950 3. The memmove() emulation function in pcre_internal.h, which is used on
2951 systems that lack both memmove() and bcopy() - that is, hardly ever -
2952 was missing a "static" storage class specifier.
2954 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2955 containing an extended class (one that cannot be represented by a bitmap
2956 because it contains high-valued characters or Unicode property items, e.g.
2957 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2958 pattern, but PCRE should not loop if you do not (it no longer does).
2959 [Detail: two cases were found: (a) a repeated subpattern containing an
2960 extended class; (b) a recursive reference to a subpattern that followed a
2961 previous extended class. It wasn't skipping over the extended class
2962 correctly when UTF-8 mode was not set.]
2964 5. A negated single-character class was not being recognized as fixed-length
2965 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2966 compile error "lookbehind assertion is not fixed length".
2968 6. The RunPerlTest auxiliary script was showing an unexpected difference
2969 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2970 write a Perl script that can interpret lines of an input file either as
2971 byte characters or as UTF-8, which is what "perltest" was being required to
2972 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2973 can't do is switch easily at run time between having the "use utf8;" pragma
2974 or not. In the end, I fudged it by using the RunPerlTest script to insert
2975 "use utf8;" explicitly for the UTF-8 tests.
2977 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2978 the end of the subject string, contrary to the documentation and to what
2979 Perl does. This was true of both matching functions. Now it matches only at
2980 the start of the subject and immediately after *internal* newlines.
2982 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2983 a pointer to an int instead of a pointer to an unsigned long int. This
2984 caused problems on 64-bit systems.
2986 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2987 instance of the 'standard' template library not being so standard".
2989 10. There was no check on the number of named subpatterns nor the maximum
2990 length of a subpattern name. The product of these values is used to compute
2991 the size of the memory block for a compiled pattern. By supplying a very
2992 long subpattern name and a large number of named subpatterns, the size
2993 computation could be caused to overflow. This is now prevented by limiting
2994 the length of names to 32 characters, and the number of named subpatterns
2995 to 10,000.
2997 11. Subpatterns that are repeated with specific counts have to be replicated in
2998 the compiled pattern. The size of memory for this was computed from the
2999 length of the subpattern and the repeat count. The latter is limited to
3000 65535, but there was no limit on the former, meaning that integer overflow
3001 could in principle occur. The compiled length of a repeated subpattern is
3002 now limited to 30,000 bytes in order to prevent this.
3004 12. Added the optional facility to have named substrings with the same name.
3006 13. Added the ability to use a named substring as a condition, using the
3007 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3008 are numbers (not recommended). Forward references are permitted.
3010 14. Added forward references in named backreferences (if you see what I mean).
3012 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3013 pattern could run off the end of the subject. For example, the pattern
3014 "(?s)(.{1,5})"8 did this with the subject "ab".
3016 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3017 PCRE_CASELESS was set when matching characters that were quantified with ?
3018 or *.
3020 17. A character class other than a single negated character that had a minimum
3021 but no maximum quantifier - for example [ab]{6,} - was not handled
3022 correctly by pce_dfa_exec(). It would match only one character.
3024 18. A valid (though odd) pattern that looked like a POSIX character
3025 class but used an invalid character after [ (for example [[,abc,]]) caused
3026 pcre_compile() to give the error "Failed: internal error: code overflow" or
3027 in some cases to crash with a glibc free() error. This could even happen if
3028 the pattern terminated after [[ but there just happened to be a sequence of
3029 letters, a binary zero, and a closing ] in the memory that followed.
3031 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3032 over the years. Originally (before any Unicode support), just the bottom 8
3033 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3034 output from "man perlunicode" includes this:
3036 The regular expression compiler produces polymorphic opcodes. That
3037 is, the pattern adapts to the data and automatically switches to
3038 the Unicode character scheme when presented with Unicode data--or
3039 instead uses a traditional byte scheme when presented with byte
3040 data.
3042 Sadly, a wide octal escape does not cause a switch, and in a string with
3043 no other multibyte characters, these octal escapes are treated as before.
3044 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3045 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3046 Unicode string.
3048 I have not perpetrated such confusion in PCRE. Up till now, it took just
3049 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3050 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3051 translate to the appropriate multibyte character.
3053 29. Applied some refactoring to reduce the number of warnings from Microsoft
3054 and Borland compilers. This has included removing the fudge introduced
3055 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3056 a warning about an unused variable.
3058 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3059 characters since release 4.0, because Perl (from release 5.004) does not.
3060 [Or at least, is documented not to: some releases seem to be in conflict
3061 with the documentation.] However, when a pattern was studied with
3062 pcre_study() and all its branches started with \s, PCRE still included VT
3063 as a possible starting character. Of course, this did no harm; it just
3064 caused an unnecessary match attempt.
3066 22. Removed a now-redundant internal flag bit that recorded the fact that case
3067 dependency changed within the pattern. This was once needed for "required
3068 byte" processing, but is no longer used. This recovers a now-scarce options
3069 bit. Also moved the least significant internal flag bit to the most-
3070 significant bit of the word, which was not previously used (hangover from
3071 the days when it was an int rather than a uint) to free up another bit for
3072 the future.
3074 23. Added support for CRLF line endings as well as CR and LF. As well as the
3075 default being selectable at build time, it can now be changed at runtime
3076 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3077 specify that it is scanning data with non-default line endings.
3079 24. Changed the definition of CXXLINK to make it agree with the definition of
3080 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3082 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3083 recursions. This makes a big different to stack usage for some patterns.
3085 26. If a subpattern containing a named recursion or subroutine reference such
3086 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3087 the space required for the compiled pattern went wrong and gave too small a
3088 value. Depending on the environment, this could lead to "Failed: internal
3089 error: code overflow at offset 49" or "glibc detected double free or
3090 corruption" errors.
3092 27. Applied patches from Google (a) to support the new newline modes and (b) to
3093 advance over multibyte UTF-8 characters in GlobalReplace.
3095 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3096 difference for some implementation of PCRE in some Windows version.
3098 29. Added some extra testing facilities to pcretest:
3100 \q<number> in a data line sets the "match limit" value
3101 \Q<number> in a data line sets the "match recursion limt" value
3102 -S <number> sets the stack size, where <number> is in megabytes
3104 The -S option isn't available for Windows.
3107 Version 6.6 06-Feb-06
3108 ---------------------
3110 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3111 in pcreposix.h. I have copied the definition from pcre.h.
3113 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3114 because pcre.h is no longer a built file.
3116 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3117 not normally included in the compiled code.
3120 Version 6.5 01-Feb-06
3121 ---------------------
3123 1. When using the partial match feature with pcre_dfa_exec(), it was not
3124 anchoring the second and subsequent partial matches at the new starting
3125 point. This could lead to incorrect results. For example, with the pattern
3126 /1234/, partially matching against "123" and then "a4" gave a match.
3128 2. Changes to pcregrep:
3130 (a) All non-match returns from pcre_exec() were being treated as failures
3131 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3132 error message is output. Some extra information is given for the
3134 probably the only errors that are likely to be caused by users (by
3135 specifying a regex that has nested indefinite repeats, for instance).
3136 If there are more than 20 of these errors, pcregrep is abandoned.
3138 (b) A binary zero was treated as data while matching, but terminated the
3139 output line if it was written out. This has been fixed: binary zeroes
3140 are now no different to any other data bytes.
3142 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3143 used to set a locale for matching. The --locale=xxxx long option has
3144 been added (no short equivalent) to specify a locale explicitly on the
3145 pcregrep command, overriding the environment variables.
3147 (d) When -B was used with -n, some line numbers in the output were one less
3148 than they should have been.
3150 (e) Added the -o (--only-matching) option.
3152 (f) If -A or -C was used with -c (count only), some lines of context were
3153 accidentally printed for the final match.
3155 (g) Added the -H (--with-filename) option.
3157 (h) The combination of options -rh failed to suppress file names for files
3158 that were found from directory arguments.
3160 (i) Added the -D (--devices) and -d (--directories) options.
3162 (j) Added the -F (--fixed-strings) option.
3164 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3166 (l) Added the --colo(u)r option.
3168 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3169 is not present by default.
3171 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3172 items such as (?R) or (?1), when the recursion could match a number of
3173 alternatives. If it matched one of the alternatives, but subsequently,
3174 outside the recursion, there was a failure, the code tried to back up into
3175 the recursion. However, because of the way PCRE is implemented, this is not
3176 possible, and the result was an incorrect result from the match.
3178 In order to prevent this happening, the specification of recursion has
3179 been changed so that all such subpatterns are automatically treated as
3180 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3182 4. I had overlooked the fact that, in some locales, there are characters for
3183 which isalpha() is true but neither isupper() nor islower() are true. In
3184 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3185 and ordfeminine) are like this. This affected the treatment of \w and \W
3186 when they appeared in character classes, but not when they appeared outside
3187 a character class. The bit map for "word" characters is now created
3188 separately from the results of isalnum() instead of just taking it from the
3189 upper, lower, and digit maps. (Plus the underscore character, of course.)
3191 5. The above bug also affected the handling of POSIX character classes such as
3192 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3193 permanent tables. Instead, the bit maps for such a class were previously
3194 created as the appropriate unions of the upper, lower, and digit bitmaps.
3195 Now they are created by subtraction from the [[:word:]] class, which has
3196 its own bitmap.
3198 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3199 It is created by subtracting the vertical space characters (\x09, \x0a,
3200 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3201 subtraction was done in the overall bitmap for a character class, meaning
3202 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3203 be recognized. This bug has been fixed.
3205 7. Patches from the folks at Google:
3207 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3208 real life, but is still worth protecting against".
3210 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3211 regular expressions".
3213 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3214 have it.
3216 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3217 "configure" and the latter not, in order to fix a problem somebody had
3218 with compiling the Arg class on HP-UX.
3220 (e) Improve the error-handling of the C++ wrapper a little bit.
3222 (f) New tests for checking recursion limiting.
3224 8. The pcre_memmove() function, which is used only if the environment does not
3225 have a standard memmove() function (and is therefore rarely compiled),
3226 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3227 returning a result (though PCRE never actually uses the result).
3229 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3230 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3231 returned instead of calling malloc() with an overflowing number that would
3232 most likely cause subsequent chaos.
3234 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3236 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3237 with this option is matched, the nmatch and pmatch options of regexec() are
3238 ignored.
3240 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3241 provided in case anyone wants to the the POSIX interface with UTF-8
3242 strings.
3244 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3245 C++ linking (needed for some HP-UX environments).
3247 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3248 (unused parameter) and in the pcre_printint() function (omitted "default"
3249 switch label when the default is to do nothing).
3251 15. Added some code to make it possible, when PCRE is compiled as a C++
3252 library, to replace subject pointers for pcre_exec() with a smart pointer
3253 class, thus making it possible to process discontinuous strings.
3255 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3256 much the same function. They were added by different people who were trying
3257 to make PCRE easy to compile on non-Unix systems. It has been suggested
3258 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3259 for compiling on Windows systems. I have therefore replaced it with
3260 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3261 defaults to "extern" for C or "extern C" for C++, which works fine on
3262 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3263 SCOPE with something explicit in config.h. In addition:
3265 (a) pcreposix.h still had just "extern" instead of either of these macros;
3266 I have replaced it with PCRE_DATA_SCOPE.
3268 (b) Functions such as _pcre_xclass(), which are internal to the library,
3269 but external in the C sense, all had PCRE_EXPORT in their definitions.
3270 This is apparently wrong for the Windows case, so I have removed it.
3271 (It makes no difference on Unix-like systems.)
3273 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3274 of recursive calls to match(). This is different to MATCH_LIMIT because
3275 that limits the total number of calls to match(), not all of which increase
3276 the depth of recursion. Limiting the recursion depth limits the amount of
3277 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3278 when PCRE is compiled, and changed at run time. A patch from Google adds
3279 this functionality to the C++ interface.
3281 18. Changes to the handling of Unicode character properties:
3283 (a) Updated the table to Unicode 4.1.0.
3285 (b) Recognize characters that are not in the table as "Cn" (undefined).
3287 (c) I revised the way the table is implemented to a much improved format
3288 which includes recognition of ranges. It now supports the ranges that
3289 are defined in UnicodeData.txt, and it also amalgamates other
3290 characters into ranges. This has reduced the number of entries in the
3291 table from around 16,000 to around 3,000, thus reducing its size
3292 considerably. I realized I did not need to use a tree structure after
3293 all - a binary chop search is just as efficient. Having reduced the
3294 number of entries, I extended their size from 6 bytes to 8 bytes to
3295 allow for more data.
3297 (d) Added support for Unicode script names via properties such as \p{Han}.
3299 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3300 matching that character.
3302 20. When matching a repeated Unicode property with a minimum greater than zero,
3303 (for example \pL{2,}), PCRE could look past the end of the subject if it
3304 reached it while seeking the minimum number of characters. This could
3305 happen only if some of the characters were more than one byte long, because
3306 there is a check for at least the minimum number of bytes.
3308 21. Refactored the implementation of \p and \P so as to be more general, to
3309 allow for more different types of property in future. This has changed the
3310 compiled form incompatibly. Anybody with saved compiled patterns that use
3311 \p or \P will have to recompile them.
3313 22. Added "Any" and "L&" to the supported property types.
3315 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3316 but give a compile time error if the value is greater than 0xff.
3318 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3319 accidentally not being installed or uninstalled.
3321 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3322 made were to insert the current release number. This seemed silly, because
3323 it made things harder for people building PCRE on systems that don't run
3324 "configure". I have turned pcre.h into a distributed file, no longer built
3325 by "configure", with the version identification directly included. There is
3326 no longer a pcre.h.in file.
3328 However, this change necessitated a change to the pcre-config script as
3329 well. It is built from pcre-config.in, and one of the substitutions was the
3330 release number. I have updated configure.ac so that ./configure now finds
3331 the release number by grepping pcre.h.
3333 26. Added the ability to run the tests under valgrind.
3336 Version 6.4 05-Sep-05
3337 ---------------------
3339 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3340 "--" to be printed when multiple files were scanned, even when none of the
3341 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3342 consider it to be a bug, and have restored the previous behaviour.
3344 2. A couple of code tidies to get rid of compiler warnings.
3346 3. The pcretest program used to cheat by referring to symbols in the library
3347 whose names begin with _pcre_. These are internal symbols that are not
3348 really supposed to be visible externally, and in some environments it is
3349 possible to suppress them. The cheating is now confined to including
3350 certain files from the library's source, which is a bit cleaner.
3352 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3353 file's purpose clearer.
3355 5. Reorganized pcre_ucp_findchar().
3358 Version 6.3 15-Aug-05
3359 ---------------------
3361 1. The file libpcre.pc.in did not have general read permission in the tarball.
3363 2. There were some problems when building without C++ support:
3365 (a) If C++ support was not built, "make install" and "make test" still
3366 tried to test it.
3368 (b) There were problems when the value of CXX was explicitly set. Some
3369 changes have been made to try to fix these, and ...
3371 (c) --disable-cpp can now be used to explicitly disable C++ support.
3373 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3374 backslash in a target when C++ was disabled. This confuses some
3375 versions of "make", apparently. Using an intermediate variable solves
3376 this. (Same for CPP_LOBJ.)
3378 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3379 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3380 necessary on certain architectures.
3382 4. Added a setting of -export-symbols-regex to the link command to remove
3383 those symbols that are exported in the C sense, but actually are local
3384 within the library, and not documented. Their names all begin with
3385 "_pcre_". This is not a perfect job, because (a) we have to except some
3386 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3387 available (and never for static libraries). I have made a note to try to
3388 find a way round (a) in the future.
3391 Version 6.2 01-Aug-05
3392 ---------------------
3394 1. There was no test for integer overflow of quantifier values. A construction
3395 such as {1111111111111111} would give undefined results. What is worse, if
3396 a minimum quantifier for a parenthesized subpattern overflowed and became
3397 negative, the calculation of the memory size went wrong. This could have
3398 led to memory overwriting.
3400 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3402 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3403 operating environments where this matters.
3405 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3406 PCRE options from within the C++ wrapper.
3408 5. Named capturing subpatterns were not being correctly counted when a pattern
3409 was compiled. This caused two problems: (a) If there were more than 100
3410 such subpatterns, the calculation of the memory needed for the whole
3411 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3412 back references of the form \12, where the number was greater than 9, were
3413 not recognized as back references, even though there were sufficient
3414 previous subpatterns.
3416 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3417 versions of gcc, e.g. 2.95.4.
3420 Version 6.1 21-Jun-05
3421 ---------------------
3423 1. There was one reference to the variable "posix" in pcretest.c that was not
3424 surrounded by "#if !defined NOPOSIX".
3426 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3427 the cross-check on the old pcre_info() function, for the benefit of the
3428 cut-down version of PCRE that is currently imported into Exim.
3430 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3431 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3432 patterns that start (?i) but I don't think that matters. The use of (?i) is
3433 just an example; this all applies to the other options as well.
3435 4. Since libtool seems to echo the compile commands it is issuing, the output
3436 from "make" can be reduced a bit by putting "@" in front of each libtool
3437 compile command.
3439 5. Patch from the folks at Google for configure.in to be a bit more thorough
3440 in checking for a suitable C++ installation before trying to compile the
3441 C++ stuff. This should fix a reported problem when a compiler was present,
3442 but no suitable headers.
3444 6. The man pages all had just "PCRE" as their title. I have changed them to
3445 be the relevant file name. I have also arranged that these names are
3446 retained in the file doc/pcre.txt, which is a concatenation in text format
3447 of all the man pages except the little individual ones for each function.
3449 7. The NON-UNIX-USE file had not been updated for the different set of source
3450 files that come with release 6. I also added a few comments about the C++
3451 wrapper.
3454 Version 6.0 07-Jun-05
3455 ---------------------
3457 1. Some minor internal re-organization to help with my DFA experiments.
3459 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3460 didn't matter for the library itself when fully configured, but did matter
3461 when compiling without UCP support, or within Exim, where the ucp files are
3462 not imported.
3464 3. Refactoring of the library code to split up the various functions into
3465 different source modules. The addition of the new DFA matching code (see
3466 below) to a single monolithic source would have made it really too
3467 unwieldy, quite apart from causing all the code to be include in a
3468 statically linked application, when only some functions are used. This is
3469 relevant even without the DFA addition now that patterns can be compiled in
3470 one application and matched in another.
3472 The downside of splitting up is that there have to be some external
3473 functions and data tables that are used internally in different modules of
3474 the library but which are not part of the API. These have all had their
3475 names changed to start with "_pcre_" so that they are unlikely to clash
3476 with other external names.
3478 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3479 a different (DFA) algorithm. Although it is slower than the original
3480 function, it does have some advantages for certain types of matching
3481 problem.
3483 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3484 including restarting after a partial match.
3486 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3487 defined when compiling for Windows was sent to me. I have put it into the
3488 code, though I have no means of testing or verifying it.
3490 7. Added the pcre_refcount() auxiliary function.
3492 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3493 match before or at the first newline in the subject string. In pcretest,
3494 the /f option on a pattern can be used to set this.
3496 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3497 would behave wrongly. This has been present in PCRE since release 4.0.
3499 10. A number of changes to the pcregrep command:
3501 (a) Refactored how -x works; insert ^(...)$ instead of setting
3502 PCRE_ANCHORED and checking the length, in preparation for adding
3503 something similar for -w.
3505 (b) Added the -w (match as a word) option.
3507 (c) Refactored the way lines are read and buffered so as to have more
3508 than one at a time available.
3510 (d) Implemented a pcregrep test script.
3512 (e) Added the -M (multiline match) option. This allows patterns to match
3513 over several lines of the subject. The buffering ensures that at least
3514 8K, or the rest of the document (whichever is the shorter) is available
3515 for matching (and similarly the previous 8K for lookbehind assertions).
3517 (f) Changed the --help output so that it now says
3519 -w, --word-regex(p)
3521 instead of two lines, one with "regex" and the other with "regexp"
3522 because that confused at least one person since the short forms are the
3523 same. (This required a bit of code, as the output is generated
3524 automatically from a table. It wasn't just a text change.)
3526 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3527 option but starts with a hyphen. Could be a pattern or a path name
3528 starting with a hyphen, for instance.
3530 (h) "-" can be given as a file name to represent stdin.
3532 (i) When file names are being printed, "(standard input)" is used for
3533 the standard input, for compatibility with GNU grep. Previously
3534 "<stdin>" was used.
3536 (j) The option --label=xxx can be used to supply a name to be used for
3537 stdin when file names are being printed. There is no short form.
3539 (k) Re-factored the options decoding logic because we are going to add
3540 two more options that take data. Such options can now be given in four
3541 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3543 (l) Added the -A, -B, and -C options for requesting that lines of context
3544 around matches be printed.
3546 (m) Added the -L option to print the names of files that do not contain
3547 any matching lines, that is, the complement of -l.
3549 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3550 continue to scan other files.
3552 (o) The -s option was incorrectly implemented. For compatibility with other
3553 greps, it now suppresses the error message for a non-existent or non-
3554 accessible file (but not the return code). There is a new option called
3555 -q that suppresses the output of matching lines, which was what -s was
3556 previously doing.
3558 (p) Added --include and --exclude options to specify files for inclusion
3559 and exclusion when recursing.
3561 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3562 Hopefully, it now does.
3564 12. Missing cast in pcre_study().
3566 13. Added an "uninstall" target to the makefile.
3568 14. Replaced "extern" in the function prototypes in Makefile.in with
3569 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3570 world, but is set differently for Windows.
3572 15. Added a second compiling function called pcre_compile2(). The only
3573 difference is that it has an extra argument, which is a pointer to an
3574 integer error code. When there is a compile-time failure, this is set
3575 non-zero, in addition to the error test pointer being set to point to an
3576 error message. The new argument may be NULL if no error number is required
3577 (but then you may as well call pcre_compile(), which is now just a
3578 wrapper). This facility is provided because some applications need a
3579 numeric error indication, but it has also enabled me to tidy up the way
3580 compile-time errors are handled in the POSIX wrapper.
3582 16. Added VPATH=.libs to the makefile; this should help when building with one
3583 prefix path and installing with another. (Or so I'm told by someone who
3584 knows more about this stuff than I do.)
3586 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3587 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3588 match everything, including newlines. This is not POSIX-compatible, but
3589 somebody wanted the feature. From pcretest it can be activated by using
3590 both the P and the s flags.
3592 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3594 19. libpcre.pc was being incorrectly installed as executable.
3596 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3597 it now also looks for '\r' so that it will work unmodified on Windows.
3599 21. Added Google's contributed C++ wrapper to the distribution.
3601 22. Added some untidy missing memory free() calls in pcretest, to keep
3602 Electric Fence happy when testing.
3606 Version 5.0 13-Sep-04
3607 ---------------------
3609 1. Internal change: literal characters are no longer packed up into items
3610 containing multiple characters in a single byte-string. Each character
3611 is now matched using a separate opcode. However, there may be more than one
3612 byte in the character in UTF-8 mode.
3614 2. The pcre_callout_block structure has two new fields: pattern_position and
3615 next_item_length. These contain the offset in the pattern to the next match
3616 item, and its length, respectively.
3618 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3619 insertion of callouts before each pattern item. Added the /C option to
3620 pcretest to make use of this.
3622 4. On the advice of a Windows user, the lines
3624 #if defined(_WIN32) || defined(WIN32)
3625 _setmode( _fileno( stdout ), 0x8000 );
3626 #endif /* defined(_WIN32) || defined(WIN32) */
3628 have been added to the source of pcretest. This apparently does useful
3629 magic in relation to line terminators.
3631 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3632 for the benefit of those environments where the "b" makes a difference.
3634 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3635 to know about it. I have put a hack into configure.in that adds in code
3636 to set GCC=yes if CC=icc. This seems to end up at a point in the
3637 generated configure script that is early enough to affect the setting of
3638 compiler options, which is what is needed, but I have no means of testing
3639 whether it really works. (The user who reported this had patched the
3640 generated configure script, which of course I cannot do.)
3642 LATER: After change 22 below (new libtool files), the configure script
3643 seems to know about icc (and also ecc). Therefore, I have commented out
3644 this hack in configure.in.
3646 7. Added support for pkg-config (2 patches were sent in).
3648 8. Negated POSIX character classes that used a combination of internal tables
3649 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3650 [[:^ascii]]. Typically, they would match almost any characters. The other
3651 POSIX classes were not broken in this way.
3653 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3654 to find the match, as PCRE was deluded into thinking that the match had to
3655 start at the start point or following a newline. The same bug applied to
3656 patterns with negative forward assertions or any backward assertions
3657 preceding ".*" at the start, unless the pattern required a fixed first
3658 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3660 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3661 starting at the last subject character, bytes beyond the end of the subject
3662 string were read.
3664 11. Renamed the variable "class" as "classbits" to make life easier for C++
3665 users. (Previously there was a macro definition, but it apparently wasn't
3666 enough.)
3668 12. Added the new field "tables" to the extra data so that tables can be passed
3669 in at exec time, or the internal tables can be re-selected. This allows
3670 a compiled regex to be saved and re-used at a later time by a different
3671 program that might have everything at different addresses.
3673 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3674 -R library as well as a -L library.
3676 14. The debugging options of pcretest (-d on the command line or D on a
3677 pattern) showed incorrect output for anything following an extended class
3678 that contained multibyte characters and which was followed by a quantifier.
3680 15. Added optional support for general category Unicode character properties
3681 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3682 support. It adds about 90K to the size of the library. The meanings of the
3683 inbuilt class escapes such as \d and \s have NOT been changed.
3685 16. Updated pcredemo.c to include calls to free() to release the memory for the
3686 compiled pattern.
3688 17. The generated file chartables.c was being created in the source directory
3689 instead of in the building directory. This caused the build to fail if the
3690 source directory was different from the building directory, and was
3691 read-only.
3693 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3694 file. No doubt somebody will tell me if they don't make sense... Also added
3695 Dan Mooney's comments about building on OpenVMS.
3697 19. Added support for partial matching via the PCRE_PARTIAL option for
3698 pcre_exec() and the \P data escape in pcretest.
3700 20. Extended pcretest with 3 new pattern features:
3702 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3703 write the compiled pattern to the file whose name is "rest-of-line".
3704 This is a straight binary dump of the data, with the saved pointer to
3705 the character tables forced to be NULL. The study data, if any, is
3706 written too. After writing, pcretest reads a new pattern.
3708 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3709 compiled pattern from the given file. There must not be any
3710 occurrences of "<" in the file name (pretty unlikely); if there are,
3711 pcretest will instead treat the initial "<" as a pattern delimiter.
3712 After reading in the pattern, pcretest goes on to read data lines as
3713 usual.
3715 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3716 and 16-bit fields in a compiled pattern, to simulate a pattern that
3717 was compiled on a host of opposite endianness.
3719 21. The pcre-exec() function can now cope with patterns that were compiled on
3720 hosts of opposite endianness, with this restriction:
3722 As for any compiled expression that is saved and used later, the tables
3723 pointer field cannot be preserved; the extra_data field in the arguments
3724 to pcre_exec() should be used to pass in a tables address if a value
3725 other than the default internal tables were used at compile time.
3727 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3728 now diagnosed as an error. Previously, most of the time, a negative number
3729 would have been treated as zero, but if in addition "ovector" was passed as
3730 NULL, a crash could occur.
3732 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3733 new versions from the libtool 1.5 distribution (the last one is a copy of
3734 a file called libtool.m4). This seems to have fixed the need to patch
3735 "configure" to support Darwin 1.3 (which I used to do). However, I still
3736 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3737 workstation).
3739 24. Changed the PCRE licence to be the more standard "BSD" licence.
3742 Version 4.5 01-Dec-03
3743 ---------------------
3745 1. There has been some re-arrangement of the code for the match() function so
3746 that it can be compiled in a version that does not call itself recursively.
3747 Instead, it keeps those local variables that need separate instances for
3748 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3749 needs to "recurse". Keeping track of where control must go is done by means
3750 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3751 hide most of the details from the main code, and operates only if
3752 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3753 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3754 operating.
3756 To make it easier for callers to provide specially tailored get/free
3757 functions for this usage, two new functions, pcre_stack_malloc, and
3758 pcre_stack_free, are used. They are always called in strict stacking order,
3759 and the size of block requested is always the same.
3761 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3762 PCRE has been compiled to use the stack or the heap for recursion. The
3763 -C option of pcretest uses this to show which version is compiled.
3765 A new data escape \S, is added to pcretest; it causes the amounts of store
3766 obtained and freed by both kinds of malloc/free at match time to be added
3767 to the output.
3769 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3770 what's available on my current Linux desktop machine.
3772 3. When matching a UTF-8 string, the test for a valid string at the start has
3773 been extended. If start_offset is not zero, PCRE now checks that it points
3774 to a byte that is the start of a UTF-8 character. If not, it returns
3775 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3776 this is necessary because there may be backward assertions in the pattern.
3777 When matching the same subject several times, it may save resources to use
3778 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3780 4. The code for checking the validity of UTF-8 strings has been tightened so
3781 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3782 containing "overlong sequences".
3784 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3785 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3786 should have been "&", but it just so happened that all the cases this let
3787 through by mistake were picked up later in the function.
3789 6. I had used a variable called "isblank" - this is a C99 function, causing
3790 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3792 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3793 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3794 so the version is shown; (c) stop "make test" echoing "./RunTest".
3796 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3798 9. The prototype for memmove() for systems that don't have it was using
3799 size_t, but the inclusion of the header that defines size_t was later. I've
3800 moved the #includes for the C headers earlier to avoid this.
3802 10. Added some adjustments to the code to make it easier to compiler on certain
3803 special systems:
3805 (a) Some "const" qualifiers were missing.
3806 (b) Added the macro EXPORT before all exported functions; by default this
3807 is defined to be empty.
3808 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3809 that it reads its output file name as an argument instead of writing
3810 to the standard output and assuming this can be redirected.
3812 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3813 class containing characters with values greater than 255, PCRE compilation
3814 went into a loop.
3816 12. A recursive reference to a subpattern that was within another subpattern
3817 that had a minimum quantifier of zero caused PCRE to crash. For example,
3818 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3819 recursion. If the recursively-called subpattern itself had a zero repeat,
3820 that was OK.
3822 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3823 buffer into which it was copied (for escape processing) was still set at
3824 1024, so long lines caused crashes.
3826 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3827 "internal error: code overflow...". This applied to any character class
3828 that was followed by a possessive quantifier.
3830 15. Modified the Makefile to add libpcre.la as a prerequisite for
3831 libpcreposix.la because I was told this is needed for a parallel build to
3832 work.
3834 16. If a pattern that contained .* following optional items at the start was
3835 studied, the wrong optimizing data was generated, leading to matching
3836 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3837 matching string must start with a or b or c. The correct conclusion for
3838 this pattern is that a match can start with any character.
3841 Version 4.4 13-Aug-03
3842 ---------------------
3844 1. In UTF-8 mode, a character class containing characters with values between
3845 127 and 255 was not handled correctly if the compiled pattern was studied.
3846 In fixing this, I have also improved the studying algorithm for such
3847 classes (slightly).
3849 2. Three internal functions had redundant arguments passed to them. Removal
3850 might give a very teeny performance improvement.
3852 3. Documentation bug: the value of the capture_top field in a callout is *one
3853 more than* the number of the hightest numbered captured substring.
3855 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3856 in incorrectly linking with a previously installed version. They now link
3857 explicitly with libpcre.la.
3859 5. configure.in no longer needs to recognize Cygwin specially.
3861 6. A problem in pcre.in for Windows platforms is fixed.
3863 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3864 pcretest, it used to include the size of the study block as part of its
3865 output. Unfortunately, the structure contains a field that has a different
3866 size on different hardware architectures. This meant that the tests that
3867 showed this size failed. As the block is currently always of a fixed size,
3868 this information isn't actually particularly useful in pcretest output, so
3869 I have just removed it.
3871 8. Three pre-processor statements accidentally did not start in column 1.
3872 Sadly, there are *still* compilers around that complain, even though
3873 standard C has not required this for well over a decade. Sigh.
3875 9. In pcretest, the code for checking callouts passed small integers in the
3876 callout_data field, which is a void * field. However, some picky compilers
3877 complained about the casts involved for this on 64-bit systems. Now
3878 pcretest passes the address of the small integer instead, which should get
3879 rid of the warnings.
3881 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3882 both compile and run time, and gives an error if an invalid UTF-8 sequence
3883 is found. There is a option for disabling this check in cases where the
3884 string is known to be correct and/or the maximum performance is wanted.
3886 11. In response to a bug report, I changed one line in Makefile.in from
3888 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3889 to
3890 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3892 to look similar to other lines, but I have no way of telling whether this
3893 is the right thing to do, as I do not use Windows. No doubt I'll get told
3894 if it's wrong...
3897 Version 4.3 21-May-03
3898 ---------------------
3900 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3901 Makefile.
3903 2. Some refactoring to improve the quality of the code:
3905 (i) The utf8_table... variables are now declared "const".
3907 (ii) The code for \cx, which used the "case flipping" table to upper case
3908 lower case letters, now just substracts 32. This is ASCII-specific,
3909 but the whole concept of \cx is ASCII-specific, so it seems
3910 reasonable.
3912 (iii) PCRE was using its character types table to recognize decimal and
3913 hexadecimal digits in the pattern. This is silly, because it handles
3914 only 0-9, a-f, and A-F, but the character types table is locale-
3915 specific, which means strange things might happen. A private
3916 table is now used for this - though it costs 256 bytes, a table is
3917 much faster than multiple explicit tests. Of course, the standard
3918 character types table is still used for matching digits in subject
3919 strings against \d.
3921 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3922 ending in _t are). So I've renamed it as ESC_tee.
3924 3. The first argument for regexec() in the POSIX wrapper should have been
3925 defined as "const".
3927 4. Changed pcretest to use malloc() for its buffers so that they can be
3928 Electric Fenced for debugging.
3930 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3931 to read one or more bytes before the start of the subject string. Often this
3932 had no effect on PCRE's behaviour, but in some circumstances it could
3933 provoke a segmentation fault.
3935 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3936 to try to read one or more bytes before the start of the subject string.
3938 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3939 UTF-8 support could misbehave in various ways if the subject string
3940 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3941 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3942 back over UTF-8 characters.)
3945 Version 4.2 14-Apr-03
3946 ---------------------
3948 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3950 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3951 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3952 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3953 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3954 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3956 Note: automatic setting of the BUILD variables is not yet working
3957 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3958 compile-time but not at link-time
3959 [LINK]: use for linking executables only
3960 make different versions for Windows and non-Windows
3961 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3962 libraries
3963 [LINK_FOR_BUILD]: new variable
3964 [OBJEXT]: use throughout
3965 [EXEEXT]: use throughout
3966 <winshared>: new target
3967 <wininstall>: new target
3968 <dftables.o>: use native compiler
3969 <dftables>: use native linker
3970 <install>: handle Windows platform correctly
3971 <clean>: ditto
3972 <check>: ditto
3973 copy DLL to top builddir before testing
3975 As part of these changes, -no-undefined was removed again. This was reported
3976 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3977 in any case.
3979 3. Some tidies to get rid of compiler warnings:
3981 . In the match_data structure, match_limit was an unsigned long int, whereas
3982 match_call_count was an int. I've made them both unsigned long ints.
3984 . In pcretest the fact that a const uschar * doesn't automatically cast to
3985 a void * provoked a warning.
3987 . Turning on some more compiler warnings threw up some "shadow" variables
3988 and a few more missing casts.
3990 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3991 option, a class that contained a single character with a value between 128
3992 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3994 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3995 option, a class that contained several characters, but with at least one
3996 whose value was between 128 and 255 caused PCRE to crash.
3999 Version 4.1 12-Mar-03
4000 ---------------------
4002 1. Compiling with gcc -pedantic found a couple of places where casts were
4003 needed, and a string in dftables.c that was longer than standard compilers are
4004 required to support.
4006 2. Compiling with Sun's compiler found a few more places where the code could
4007 be tidied up in order to avoid warnings.
4009 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4010 first of these names is deprecated in the latest Autoconf in favour of the name
4011 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4012 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4013 analogy I have changed it to CFLAGS_FOR_BUILD.
4015 4. Added -no-undefined to the linking command in the Makefile, because this is
4016 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4017 linking step for the pcreposix library.
4019 5. PCRE was failing to diagnose the case of two named groups with the same
4020 name.
4022 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4023 literal character that is needed in the subject for a match, and scans along to
4024 ensure that it is present before embarking on the full matching process. This
4025 saves time in cases of nested unlimited repeats that are never going to match.
4026 Problem: the scan can take a lot of time if the subject is very long (e.g.
4027 megabytes), thus penalizing straightforward matches. It is now done only if the
4028 amount of subject to be scanned is less than 1000 bytes.
4030 7. A lesser problem with the same optimization is that it was recording the
4031 first character of an anchored pattern as "needed", thus provoking a search
4032 right along the subject, even when the first match of the pattern was going to
4033 fail. The "needed" character is now not set for anchored patterns, unless it
4034 follows something in the pattern that is of non-fixed length. Thus, it still
4035 fulfils its original purpose of finding quick non-matches in cases of nested
4036 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4039 Version 4.0 17-Feb-03
4040 ---------------------
4042 1. If a comment in an extended regex that started immediately after a meta-item
4043 extended to the end of string, PCRE compiled incorrect data. This could lead to
4044 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4046 2. Moved to autoconf 2.53 and libtool 1.4.2.
4048 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4049 the special perltest8 script is no longer needed - all the tests can be run
4050 from a single perltest script.
4052 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4053 by \s. It has now been removed in PCRE. This means it isn't recognized as
4054 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4055 class [:space:] *does* include VT, thereby creating a mess.
4057 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4058 space and tab.
4060 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4061 its new features into the main test script, reducing the number of scripts.
4063 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4064 were backward compatible, and made the (?i) apply to the whole pattern, as if
4065 /i were given. Now it behaves more logically, and applies the option setting
4066 only to what follows. PCRE has been changed to follow suit. However, if it
4067 finds options settings right at the start of the pattern, it extracts them into
4068 the global options, as before. Thus, they show up in the info data.
4070 8. Added support for the \Q...\E escape sequence. Characters in between are
4071 treated as literals. This is slightly different from Perl in that $ and @ are
4072 also handled as literals inside the quotes. In Perl, they will cause variable
4073 interpolation. Note the following examples:
4075 Pattern PCRE matches Perl matches
4077 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4078 \Qabc\$xyz\E abc\$xyz abc\$xyz
4079 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4081 For compatibility with Perl, \Q...\E sequences are recognized inside character
4082 classes as well as outside them.
4084 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4085 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4086 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4087 signed/unsigned warnings.
4089 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4090 option for pcretest, so I've replaced it by a simple function that does just
4091 that job.
4093 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4094 "pcregrep -".
4096 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4097 Java package. This provides some syntactic sugar for simple cases of what my
4098 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4099 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4100 item, you can use this simplified notation. Note that only makes sense with
4101 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4102 greediness, whatever the setting of the PCRE_UNGREEDY option.
4104 13. A change of greediness default within a pattern was not taking effect at
4105 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4106 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4107 was abstracted outside.
4109 14. PCRE now supports the \G assertion. It is true when the current matching
4110 position is at the start point of the match. This differs from \A when the
4111 starting offset is non-zero. Used with the /g option of pcretest (or similar
4112 code), it works in the same way as it does for Perl's /g option. If all
4113 alternatives of a regex begin with \G, the expression is anchored to the start
4114 match position, and the "anchored" flag is set in the compiled expression.
4116 15. Some bugs concerning the handling of certain option changes within patterns
4117 have been fixed. These applied to options other than (?ims). For example,
4118 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4119 been the other way round. Some of this was related to change 7 above.
4121 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4122 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4123 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4124 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4126 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4127 mode. Unlike ".", it always matches newline, whatever the setting of
4128 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4129 assertions. Perl allows it, but it doesn't (in general) work because it can't
4130 calculate the length of the lookbehind. At least, that's the case for Perl
4131 5.8.0 - I've been told they are going to document that it doesn't work in
4132 future.
4134 18. Added an error diagnosis for escapes that PCRE does not support: these are
4135 \L, \l, \N, \P, \p, \U, \u, and \X.
4137 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4138 reading past the end of the pattern in cases such as /[abcd/.
4140 20. PCRE was getting more memory than necessary for patterns with classes that
4141 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4143 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4144 compiling PCRE for use with Virtual Pascal.
4146 22. Small fix to the Makefile to make it work properly if the build is done
4147 outside the source tree.
4149 23. Added a new extension: a condition to go with recursion. If a conditional
4150 subpattern starts with (?(R) the "true" branch is used if recursion has
4151 happened, whereas the "false" branch is used only at the top level.
4153 24. When there was a very long string of literal characters (over 255 bytes
4154 without UTF support, over 250 bytes with UTF support), the computation of how
4155 much memory was required could be incorrect, leading to segfaults or other
4156 strange effects.
4158 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4159 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4160 there was a subsequent back reference to those brackets. This meant that, for
4161 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4162 possible to check for precisely this case. All we can do is abandon the
4163 optimization if .* occurs inside capturing brackets when there are any back
4164 references whatsoever. (See below for a better fix that came later.)
4166 26. The handling of the optimization for finding the first character of a
4167 non-anchored pattern, and for finding a character that is required later in the
4168 match were failing in some cases. This didn't break the matching; it just
4169 failed to optimize when it could. The way this is done has been re-implemented.
4171 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4173 28. Added a new feature that provides some of the functionality that Perl
4174 provides with (?{...}). The facility is termed a "callout". The way it is done
4175 in PCRE is for the caller to provide an optional function, by setting
4176 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4177 global variable. By default it is unset, which disables all calling out. To get
4178 the function called, the regex must include (?C) at appropriate points. This
4179 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4180 This provides a means of identifying different callout points. When PCRE
4181 reaches such a point in the regex, if pcre_callout has been set, the external
4182 function is called. It is provided with data in a structure called
4183 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4184 matching continues; if it returns a non-zero value, the match at the current
4185 point fails. However, backtracking will occur if possible. [This was changed
4186 later and other features added - see item 49 below.]
4188 29. pcretest is upgraded to test the callout functionality. It provides a
4189 callout function that displays information. By default, it shows the start of
4190 the match and the current position in the text. There are some new data escapes
4191 to vary what happens:
4193 \C+ in addition, show current contents of captured substrings
4194 \C- do not supply a callout function
4195 \C!n return 1 when callout number n is reached
4196 \C!n!m return 1 when callout number n is reached for the mth time
4198 30. If pcregrep was called with the -l option and just a single file name, it
4199 output "<stdin>" if a match was found, instead of the file name.
4201 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4202 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4203 pcre_exec(). This saves a malloc/free per call. The default value of
4204 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4205 when configuring.
4207 32. The default maximum size of a compiled pattern is 64K. There have been a
4208 few cases of people hitting this limit. The code now uses macros to handle the
4209 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4210 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4211 configuring. Tests 2 and 5 work only with 2-byte links because they output
4212 debugging information about compiled patterns.
4214 33. Internal code re-arrangements:
4216 (a) Moved the debugging function for printing out a compiled regex into
4217 its own source file (printint.c) and used #include to pull it into
4218 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4219 separate copies.
4221 (b) Defined the list of op-code names for debugging as a macro in
4222 internal.h so that it is next to the definition of the opcodes.
4224 (c) Defined a table of op-code lengths for simpler skipping along compiled
4225 code. This is again a macro in internal.h so that it is next to the
4226 definition of the opcodes.
4228 34. Added support for recursive calls to individual subpatterns, along the
4229 lines of Robin Houston's patch (but implemented somewhat differently).
4231 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4232 allow it to read and process whole directories in Win32. This code was
4233 contributed by Lionel Fourquaux; it has not been tested by me.
4235 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4236 used to name a group. Names consist of alphanumerics and underscores, and must
4237 be unique. Back references use the syntax (?P=name) and recursive calls use
4238 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4239 numbers. The function pcre_fullinfo() can be used after compilation to extract
4240 a name/number map. There are three relevant calls:
4242 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4243 PCRE_INFO_NAMECOUNT yields the number of entries
4244 PCRE_INFO_NAMETABLE yields a pointer to the map.
4246 The map is a vector of fixed-size entries. The size of each entry depends on
4247 the length of the longest name used. The first two bytes of each entry are the
4248 group number, most significant byte first. There follows the corresponding
4249 name, zero terminated. The names are in alphabetical order.
4251 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4252 case instead of 255. Making it the same both with and without UTF-8 support
4253 means that the same test output works with both.
4255 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4256 calling malloc() with a zero argument.
4258 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4259 optimization. I've improved things by keeping a bitmap of backreferences with
4260 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4261 fact referenced, the optimization can be applied. It is unlikely that a
4262 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4263 the match to follow \n) will appear inside brackets with a number greater than
4264 31, but if it does, any back reference > 31 suppresses the optimization.
4266 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4267 of disabling numbered capturing parentheses. Any opening parenthesis that is
4268 not followed by ? behaves as if it were followed by ?: but named parentheses
4269 can still be used for capturing (and they will acquire numbers in the usual
4270 way).
4272 41. Redesigned the return codes from the match() function into yes/no/error so
4273 that errors can be passed back from deep inside the nested calls. A malloc
4274 failure while inside a recursive subpattern call now causes the
4275 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4277 42. It is now possible to set a limit on the number of times the match()
4278 function is called in a call to pcre_exec(). This facility makes it possible to
4279 limit the amount of recursion and backtracking, though not in a directly
4280 obvious way, because the match() function is used in a number of different
4281 circumstances. The count starts from zero for each position in the subject
4282 string (for non-anchored patterns). The default limit is, for compatibility, a
4283 large number, namely 10 000 000. You can change this in two ways:
4285 (a) When configuring PCRE before making, you can use --with-match-limit=n
4286 to set a default value for the compiled library.
4288 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4289 a different value is set. See 45 below.
4291 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4293 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4294 of things that can be changed at compile time. The first argument specifies
4295 what is wanted and the second points to where the information is to be placed.
4296 The current list of available information is:
4300 The output is an integer that is set to one if UTF-8 support is available;
4301 otherwise it is set to zero.
4305 The output is an integer that it set to the value of the code that is used for
4306 newline. It is either LF (10) or CR (13).
4310 The output is an integer that contains the number of bytes used for internal
4311 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4315 The output is an integer that contains the threshold above which the POSIX
4316 interface uses malloc() for output vectors. See item 31 above.
4320 The output is an unsigned integer that contains the default limit of the number
4321 of match() calls in a pcre_exec() execution. See 42 above.
4323 44. pcretest has been upgraded by the addition of the -C option. This causes it
4324 to extract all the available output from the new pcre_config() function, and to
4325 output it. The program then exits immediately.
4327 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4328 order to support additional features. One way would have been to define
4329 pcre_exec2() (for example) with extra arguments, but this would not have been
4330 extensible, and would also have required all calls to the original function to
4331 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4332 is used for passing in "extra" data from pcre_study().
4334 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4335 contains the following fields:
4337 flags a bitmap indicating which of the following fields are set
4338 study_data opaque data from pcre_study()
4339 match_limit a way of specifying a limit on match() calls for a specific
4340 call to pcre_exec()
4341 callout_data data for callouts (see 49 below)
4343 The flag bits are also defined in pcre.h, and are
4349 The pcre_study() function now returns one of these new pcre_extra blocks, with
4350 the actual study data pointed to by the study_data field, and the
4351 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4352 before. That is, this change is entirely upwards-compatible and requires no
4353 change to existing code.
4355 If you want to pass in additional data to pcre_exec(), you can either place it
4356 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4357 block.
4359 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4360 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4361 times with different match limits, until it finds the minimum value needed for
4362 pcre_exec() to complete. The value is then output. This can be instructive; for
4363 most simple matches the number is quite small, but for pathological cases it
4364 gets very large very quickly.
4366 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4367 returns the size of the data block pointed to by the study_data field in a
4368 pcre_extra block, that is, the value that was passed as the argument to
4369 pcre_malloc() when PCRE was getting memory in which to place the information
4370 created by pcre_study(). The fourth argument should point to a size_t variable.
4371 pcretest has been extended so that this information is shown after a successful
4372 pcre_study() call when information about the compiled regex is being displayed.
4374 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4375 because what follows is always an absolute path. (Later: it turns out that this
4376 is more than cosmetic for MinGW, because it doesn't like empty path
4377 components.)
4379 49. Some changes have been made to the callout feature (see 28 above):
4381 (i) A callout function now has three choices for what it returns:
4383 0 => success, carry on matching
4384 > 0 => failure at this point, but backtrack if possible
4385 < 0 => serious error, return this value from pcre_exec()
4387 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4388 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4389 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4390 use by callout functions. It will never be used by PCRE itself.
4392 (ii) The pcre_extra structure (see 45 above) has a void * field called
4393 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4394 pcre_callout_block structure has a field of the same name. The contents of
4395 the field passed in the pcre_extra structure are passed to the callout
4396 function in the corresponding field in the callout block. This makes it
4397 easier to use the same callout-containing regex from multiple threads. For
4398 testing, the pcretest program has a new data escape
4400 \C*n pass the number n (may be negative) as callout_data
4402 If the callout function in pcretest receives a non-zero value as
4403 callout_data, it returns that value.
4405 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4406 there were some redundant $(CFLAGS) in commands that are now specified as
4407 $(LINK), which already includes $(CFLAGS).
4409 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4410 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4411 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4412 one-byte characters throughout. Note that case-insensitive matching applies
4413 only to characters whose values are less than 256. PCRE doesn't support the
4414 notion of cases for higher-valued characters.
4416 (i) A character class whose characters are all within 0-255 is handled as
4417 a bit map, and the map is inverted for negative classes. Previously, a
4418 character > 255 always failed to match such a class; however it should
4419 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4421 (ii) A negated character class with a single character < 255 is coded as
4422 "not this character" (OP_NOT). This wasn't working properly when the test
4423 character was multibyte, either singly or repeated.
4425 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4426 mode, for example: \x{100}{2,3}.
4428 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4429 singly or repeated) now correctly test multibyte characters. However,
4430 PCRE doesn't recognize any characters with values greater than 255 as
4431 digits, spaces, or word characters. Such characters always match \D, \S,
4432 and \W, and never match \d, \s, or \w.
4434 (v) Classes may now contain characters and character ranges with values
4435 greater than 255. For example: [ab\x{100}-\x{400}].
4437 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4438 PCRE in UTF-8 mode.
4440 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4441 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4442 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4443 value.)
4445 53. The single man page has become too large. I have therefore split it up into
4446 a number of separate man pages. These also give rise to individual HTML pages;
4447 these are now put in a separate directory, and there is an index.html page that
4448 lists them all. Some hyperlinking between the pages has been installed.
4450 54. Added convenience functions for handling named capturing parentheses.
4452 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4453 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4454 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4455 are faulted.
4457 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4458 calling configure. These values are used when compiling the dftables.c program
4459 which is run to generate the source of the default character tables. They
4460 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4461 you will need to set these values.
4463 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4466 Version 3.9 02-Jan-02
4467 ---------------------
4469 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4471 2. If --disable-static was given, the building process failed when trying to
4472 build pcretest and pcregrep. (For some reason it was using libtool to compile
4473 them, which is not right, as they aren't part of the library.)
4476 Version 3.8 18-Dec-01
4477 ---------------------
4479 1. The experimental UTF-8 code was completely screwed up. It was packing the
4480 bytes in the wrong order. How dumb can you get?
4483 Version 3.7 29-Oct-01
4484 ---------------------
4486 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4487 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4488 this didn't happen under Solaris 8, where I normally test things.
4490 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4491 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4492 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4493 this fix an hour or so after the initial 3.7 release.)
4496 Version 3.6 23-Oct-01
4497 ---------------------
4499 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4500 offsets passed as NULL with zero offset count.
4502 2. The config.guess and config.sub files had not been updated when I moved to
4503 the latest autoconf.
4506 Version 3.5 15-Aug-01
4507 ---------------------
4509 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4510 had been forgotten.
4512 2. By using declared but undefined structures, we can avoid using "void"
4513 definitions in pcre.h while keeping the internal definitions of the structures
4514 private.
4516 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4517 user point of view, this means that both static and shared libraries are built
4518 by default, but this can be individually controlled. More of the work of
4519 handling this static/shared cases is now inside libtool instead of PCRE's make
4520 file.
4522 4. The pcretest utility is now installed along with pcregrep because it is
4523 useful for users (to test regexs) and by doing this, it automatically gets
4524 relinked by libtool. The documentation has been turned into a man page, so
4525 there are now .1, .txt, and .html versions in /doc.
4527 5. Upgrades to pcregrep:
4528 (i) Added long-form option names like gnu grep.
4529 (ii) Added --help to list all options with an explanatory phrase.
4530 (iii) Added -r, --recursive to recurse into sub-directories.
4531 (iv) Added -f, --file to read patterns from a file.
4533 6. pcre_exec() was referring to its "code" argument before testing that
4534 argument for NULL (and giving an error if it was NULL).
4536 7. Upgraded Makefile.in to allow for compiling in a different directory from
4537 the source directory.
4539 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4540 options bits, the pointer it was passed was to an int instead of to an unsigned
4541 long int. This mattered only on 64-bit systems.
4543 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4544 generated) instead of pcre.in, which it its source. Also made the same change
4545 in several of the .c files.
4547 10. A new release of gcc defines printf() as a macro, which broke pcretest
4548 because it had an ifdef in the middle of a string argument for printf(). Fixed
4549 by using separate calls to printf().
4551 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4552 script, to force use of CR or LF instead of \n in the source. On non-Unix
4553 systems, the value can be set in config.h.
4555 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4556 absolute limit. Changed the text of the error message to make this clear, and
4557 likewise updated the man page.
4559 13. The limit of 99 on the number of capturing subpatterns has been removed.
4560 The new limit is 65535, which I hope will not be a "real" limit.
4563 Version 3.4 22-Aug-00
4564 ---------------------
4566 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4568 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4571 Version 3.3 01-Aug-00
4572 ---------------------
4574 1. If an octal character was given, but the value was greater than \377, it
4575 was not getting masked to the least significant bits, as documented. This could
4576 lead to crashes in some systems.
4578 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4579 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4581 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4582 These just pass their arguments on to (pcre_free)(), but they are provided
4583 because some uses of PCRE bind it to non-C systems that can call its functions,
4584 but cannot call free() or pcre_free() directly.
4586 4. Add "make test" as a synonym for "make check". Corrected some comments in
4587 the Makefile.
4589 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4590 Makefile.
4592 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4593 command called pgrep for grepping around the active processes.
4595 7. Added the beginnings of support for UTF-8 character strings.
4597 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4598 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4599 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4600 out for the ar command.)
4603 Version 3.2 12-May-00
4604 ---------------------
4606 This is purely a bug fixing release.
4608 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4609 of ZA. This was just one example of several cases that could provoke this bug,
4610 which was introduced by change 9 of version 2.00. The code for breaking
4611 infinite loops after an iteration that matches an empty string was't working
4612 correctly.
4614 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4615 when matched against abbab (for example). After matching an empty string, it
4616 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4617 caused it to match further down the string than it should.
4619 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4620 was there because it doesn't seem to be needed, and it causes trouble on some
4621 systems, as it is not a Standard C header. It has been removed.
4623 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4624 were reported on the Macintosh. The changes were from
4626 while ((c = *(++ptr)) != 0 && c != '\n');
4627 to
4628 while ((c = *(++ptr)) != 0 && c != '\n') ;
4630 Totally extraordinary, but if that's what it takes...
4632 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4633 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4634 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4635 assumes the way PCRE uses memmove() (always moving upwards).
4637 6. PCRE is being used in one environment where strchr() is not available. There
4638 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4639 faster code anyway.
4642 Version 3.1 09-Feb-00
4643 ---------------------
4645 The only change in this release is the fixing of some bugs in Makefile.in for
4646 the "install" target:
4648 (1) It was failing to install pcreposix.h.
4650 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4653 Version 3.0 01-Feb-00
4654 ---------------------
4656 1. Add support for the /+ modifier to perltest (to output $` like it does in
4657 pcretest).
4659 2. Add support for the /g modifier to perltest.
4661 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4662 matches null strings.
4664 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4665 pattern. Perl treats empty patterns specially - it reuses the most recent
4666 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4667 effect.
4669 5. The POSIX interface was broken in that it was just handing over the POSIX
4670 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4671 required a bigger vector, with some working space on the end. This means that
4672 the POSIX wrapper now has to get and free some memory, and copy the results.
4674 6. Added some simple autoconf support, placing the test data and the
4675 documentation in separate directories, re-organizing some of the
4676 information files, and making it build pcre-config (a GNU standard). Also added
4677 libtool support for building PCRE as a shared library, which is now the
4678 default.
4680 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4681 09 are not valid octal constants. Single digits will be used for minor values
4682 less than 10.
4684 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4685 existing programs that set these in the POSIX interface can use PCRE without
4686 modification.
4688 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4689 return all that pcre_info() returns, plus additional data. The pcre_info()
4690 function is retained for compatibility, but is considered to be obsolete.
4692 10. Added experimental recursion feature (?R) to handle one common case that
4693 Perl 5.6 will be able to do with (?p{...}).
4695 11. Added support for POSIX character classes like [:alpha:], which Perl is
4696 adopting.
4699 Version 2.08 31-Aug-99
4700 ----------------------
4702 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4703 trying to match at the startoffset position, but instead was moving forward to
4704 the next newline as if a previous match had failed.
4706 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4707 and could get into a loop if a null string was matched other than at the start
4708 of the subject.
4710 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4711 be distinguished at compile time, and for completeness also added PCRE_DATE.
4713 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4714 in GnuWin32 environments.
4717 Version 2.07 29-Jul-99
4718 ----------------------
4720 1. The documentation is now supplied in plain text form and HTML as well as in
4721 the form of man page sources.
4723 2. C++ compilers don't like assigning (void *) values to other pointer types.
4724 In particular this affects malloc(). Although there is no problem in Standard
4725 C, I've put in casts to keep C++ compilers happy.
4727 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4728 should be (const char *).
4730 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4731 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4732 However, I haven't made this a standard facility. The documentation doesn't
4733 mention it, and the Makefile doesn't support it.
4735 5. The Makefile now contains an "install" target, with editable destinations at
4736 the top of the file. The pcretest program is not installed.
4738 6. pgrep -V now gives the PCRE version number and date.
4740 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4741 causing the entire string to be ignored, instead of just the last character.
4743 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4744 non-matching string, it can take a very, very long time, even for strings of
4745 quite modest length, because of the nested recursion. PCRE now does better in
4746 some of these cases. It does this by remembering the last required literal
4747 character in the pattern, and pre-searching the subject to ensure it is present
4748 before running the real match. In other words, it applies a heuristic to detect
4749 some types of certain failure quickly, and in the above example, if presented
4750 with a string that has no trailing " it gives "no match" very quickly.
4752 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4753 other alternatives are tried instead.
4756 Version 2.06 09-Jun-99
4757 ----------------------
4759 1. Change pcretest's output for amount of store used to show just the code
4760 space, because the remainder (the data block) varies in size between 32-bit and
4761 64-bit systems.
4763 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4764 start matching at. This allows lookbehinds to work when searching for multiple
4765 occurrences in a string.
4767 3. Added additional options to pcretest for testing multiple occurrences:
4769 /+ outputs the rest of the string that follows a match
4770 /g loops for multiple occurrences, using the new startoffset argument
4771 /G loops for multiple occurrences by passing an incremented pointer
4773 4. PCRE wasn't doing the "first character" optimization for patterns starting
4774 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4775 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4776 the letter 'x'. On long subject strings, this gives a significant speed-up.
4779 Version 2.05 21-Apr-99
4780 ----------------------
4782 1. Changed the type of magic_number from int to long int so that it works
4783 properly on 16-bit systems.
4785 2. Fixed a bug which caused patterns starting with .* not to work correctly
4786 when the subject string contained newline characters. PCRE was assuming
4787 anchoring for such patterns in all cases, which is not correct because .* will
4788 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4789 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4790 must be retried after every newline in the subject.
4793 Version 2.04 18-Feb-99
4794 ----------------------
4796 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4797 computation of the store needed to hold the pattern was incorrect (too large).
4798 If such patterns were nested a few deep, this could multiply and become a real
4799 problem.
4801 2. Added /M option to pcretest to show the memory requirement of a specific
4802 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4804 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4805 compiled in such a way that the backtracking after subsequent failure was
4806 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4807 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4810 Version 2.03 02-Feb-99
4811 ----------------------
4813 1. Fixed typo and small mistake in man page.
4815 2. Added 4th condition (GPL supersedes if conflict) and created separate
4816 LICENCE file containing the conditions.
4818 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4819 Perl, that is the internal \ allows the delimiter to be included in the
4820 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4821 the final delimiter, add \ to the end of the pattern (to test the error).
4823 4. Added the convenience functions for extracting substrings after a successful
4824 match. Updated pcretest to make it able to test these functions.
4827 Version 2.02 14-Jan-99
4828 ----------------------
4830 1. Initialized the working variables associated with each extraction so that
4831 their saving and restoring doesn't refer to uninitialized store.
4833 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4834 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4835 fix the problem.
4837 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4838 calls, and wasn't printing the correct value for compiling calls. Increased the
4839 default value of LOOPREPEAT, and the number of significant figures in the
4840 times.
4842 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4844 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4845 a building problem on Windows NT with a FAT file system.
4848 Version 2.01 21-Oct-98
4849 ----------------------
4851 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4852 to character tables built by pcre_maketables() in the current locale. If NULL
4853 is passed, the default tables are used.
4856 Version 2.00 24-Sep-98
4857 ----------------------
4859 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4860 it any more.
4862 2. Allow quantification of (?>) groups, and make it work correctly.
4864 3. The first character computation wasn't working for (?>) groups.
4866 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4867 end of the subject) and add 5.005's \z, which really does match only at the
4868 very end of the subject.
4870 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4872 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4873 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4874 localized options. All options to pcre_study() were also removed.
4876 7. Add other new features from 5.005:
4878 $(?<= positive lookbehind
4879 $(?<! negative lookbehind
4880 (?imsx-imsx) added the unsetting capability
4881 such a setting is global if at outer level; local otherwise
4882 (?imsx-imsx:) non-capturing groups with option setting
4883 (?(cond)re|re) conditional pattern matching
4885 A backreference to itself in a repeated group matches the previous
4886 captured string.
4888 8. General tidying up of studying (both automatic and via "study")
4889 consequential on the addition of new assertions.
4891 9. As in 5.005, unlimited repeated groups that could match an empty substring
4892 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4893 runtime if any iteration does actually match an empty substring.
4895 10. Include the RunTest script in the distribution.
4897 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4898 discrepancies, some of which were old and were also with respect to 5.004. They
4899 have now been fixed.
4902 Version 1.09 28-Apr-98
4903 ----------------------
4905 1. A negated single character class followed by a quantifier with a minimum
4906 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4907 program crashes, or just wrong answers. This did not apply to negated classes
4908 containing more than one character, or to minima other than one.
4911 Version 1.08 27-Mar-98
4912 ----------------------
4914 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4916 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4917 latter must appear before anything that relies on it in the pattern.
4920 Version 1.07 16-Feb-98
4921 ----------------------
4923 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4924 repeat of a potentially empty string).
4927 Version 1.06 23-Jan-98
4928 ----------------------
4930 1. Added Markus Oberhumer's little patches for C++.
4932 2. Literal strings longer than 255 characters were broken.
4935 Version 1.05 23-Dec-97
4936 ----------------------
4938 1. Negated character classes containing more than one character were failing if
4939 PCRE_CASELESS was set at run time.
4942 Version 1.04 19-Dec-97
4943 ----------------------
4945 1. Corrected the man page, where some "const" qualifiers had been omitted.
4947 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4948 input syntax.
4950 3. Fixed memory leak which occurred when a regex with back references was
4951 matched with an offsets vector that wasn't big enough. The temporary memory
4952 that is used in this case wasn't being freed if the match failed.
4954 4. Tidied pcretest to ensure it frees memory that it gets.
4956 5. Temporary memory was being obtained in the case where the passed offsets
4957 vector was exactly big enough.
4959 6. Corrected definition of offsetof() from change 5 below.
4961 7. I had screwed up change 6 below and broken the rules for the use of
4962 setjmp(). Now fixed.
4965 Version 1.03 18-Dec-97
4966 ----------------------
4968 1. A erroneous regex with a missing opening parenthesis was correctly
4969 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4970 on some systems.
4972 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4973 it was reported that one broken compiler failed on the former because "code" is
4974 also an independent variable.
4976 3. The erroneous regex a[]b caused an array overrun reference.
4978 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4979 fail on data ending with that character. (It was going on too far, and checking
4980 the next character, typically a binary zero.) This was specific to the
4981 optimized code for single-character negative classes.
4983 5. Added a contributed patch from the TIN world which does the following:
4985 + Add an undef for memmove, in case the the system defines a macro for it.
4987 + Add a definition of offsetof(), in case there isn't one. (I don't know
4988 the reason behind this - offsetof() is part of the ANSI standard - but
4989 it does no harm).
4991 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4992 most of the places where whitespace preceded '#'. I have given up and
4993 allowed the remaining 2 cases to be at the margin.
4995 + Rename some variables in pcre to eliminate shadowing. This seems very
4996 pedantic, but does no harm, of course.
4998 6. Moved the call to setjmp() into its own function, to get rid of warnings
4999 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5001 7. Constructs such as \d{8,} were compiling into the equivalent of
5002 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5003 outcome, but in this particular case used more store than had been allocated,
5004 which caused the bug to be discovered because it threw up an internal error.
5006 8. The debugging code in both pcre and pcretest for outputting the compiled
5007 form of a regex was going wrong in the case of back references followed by
5008 curly-bracketed repeats.
5011 Version 1.02 12-Dec-97
5012 ----------------------
5014 1. Typos in pcre.3 and comments in the source fixed.
5016 2. Applied a contributed patch to get rid of places where it used to remove
5017 'const' from variables, and fixed some signed/unsigned and uninitialized
5018 variable warnings.
5020 3. Added the "runtest" target to Makefile.
5022 4. Set default compiler flag to -O2 rather than just -O.
5025 Version 1.01 19-Nov-97
5026 ----------------------
5028 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5029 like /([ab]*)*/, that is, for classes with more than one character in them.
5031 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5032 as /((?>a*))*/ (a PCRE_EXTRA facility).
5035 Version 1.00 18-Nov-97
5036 ----------------------
5038 1. Added compile-time macros to support systems such as SunOS4 which don't have
5039 memmove() or strerror() but have other things that can be used instead.
5041 2. Arranged that "make clean" removes the executables.
5044 Version 0.99 27-Oct-97
5045 ----------------------
5047 1. Fixed bug in code for optimizing classes with only one character. It was
5048 initializing a 32-byte map regardless, which could cause it to run off the end
5049 of the memory it had got.
5051 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5054 Version 0.98 22-Oct-97
5055 ----------------------
5057 1. Fixed bug in code for handling temporary memory usage when there are more
5058 back references than supplied space in the ovector. This could cause segfaults.
5061 Version 0.97 21-Oct-97
5062 ----------------------
5064 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5066 2. Optimized negated single characters not to use a bit map.
5068 3. Brought error texts together as macro definitions; clarified some of them;
5069 fixed one that was wrong - it said "range out of order" when it meant "invalid
5070 escape sequence".
5072 4. Changed some char * arguments to const char *.
5074 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5076 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5077 pcretest.
5080 Version 0.96 16-Oct-97
5081 ----------------------
5083 1. Added a simple "pgrep" utility to the distribution.
5085 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5086 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5087 where "ddd" means "one or more decimal digits".
5089 3. Fixed serious bug. If a pattern had a back reference, but the call to
5090 pcre_exec() didn't supply a large enough ovector to record the related
5091 identifying subpattern, the match always failed. PCRE now remembers the number
5092 of the largest back reference, and gets some temporary memory in which to save
5093 the offsets during matching if necessary, in order to ensure that
5094 backreferences always work.
5096 4. Increased the compatibility with Perl in a number of ways:
5098 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5099 to request this handling. The option can be set at compile or exec time.
5101 (b) $ matches before a terminating newline by default; an option
5102 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5103 mode). The option can be set at compile or exec time.
5105 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5106 the same as Perl's. If the decimal number it represents is less than 10
5107 or there aren't that many previous left capturing parentheses, an octal
5108 escape is read. Inside a character class, it's always an octal escape,
5109 even if it is a single digit.
5111 (d) An escaped but undefined alphabetic character is taken as a literal,
5112 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5113 escapes.
5115 (e) {0} is now permitted. (The previous item is removed from the compiled
5116 pattern).
5118 5. Changed all the names of code files so that the basic parts are no longer
5119 than 10 characters, and abolished the teeny "globals.c" file.
5121 6. Changed the handling of character classes; they are now done with a 32-byte
5122 bit map always.
5124 7. Added the -d and /D options to pcretest to make it possible to look at the
5125 internals of compilation without having to recompile pcre.
5128 Version 0.95 23-Sep-97
5129 ----------------------
5131 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5132 \x20 at the start of a run of normal characters. These were being treated as
5133 real characters, instead of the source characters being re-checked.
5136 Version 0.94 18-Sep-97
5137 ----------------------
5139 1. The functions are now thread-safe, with the caveat that the global variables
5140 containing pointers to malloc() and free() or alternative functions are the
5141 same for all threads.
5143 2. Get pcre_study() to generate a bitmap of initial characters for non-
5144 anchored patterns when this is possible, and use it if passed to pcre_exec().
5147 Version 0.93 15-Sep-97
5148 ----------------------
5150 1. /(b)|(:+)/ was computing an incorrect first character.
5152 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5153 but not actually doing anything yet.
5155 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5156 as Perl does (e.g. [-az] or [az-]).
5158 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5159 all possible positions.
5161 5. Split up into different modules to avoid including unneeded functions in a
5162 compiled binary. However, compile and exec are still in one module. The "study"
5163 function is split off.
5165 6. The character tables are now in a separate module whose source is generated
5166 by an auxiliary program - but can then be edited by hand if required. There are
5167 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5168 toupper() in the code.
5170 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5171 make them global. Abolish the function for setting them, as the caller can now
5172 set them directly.
5175 Version 0.92 11-Sep-97
5176 ----------------------
5178 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5179 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5181 2. Caseless matching was not working in character classes if the characters in
5182 the pattern were in upper case.
5184 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5186 4. Make PCRE_ANCHORED public and accept as a compile option.
5188 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5189 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5190 pass them.
5192 6. Give an error if bad option bits passed at compile or run time.
5194 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5195 pcretest to cause it to pass that flag.
5197 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5198 options, and the first character, if set.
5200 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5203 Version 0.91 10-Sep-97
5204 ----------------------
5206 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5207 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5209 2. PCRE was looping on encountering an indefinitely repeated back reference to
5210 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5211 Perl does - treats the match as successful.
5213 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5