/[pcre]/code/trunk/ChangeLog
ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1361 - (show annotations)
Fri Sep 6 17:47:32 2013 UTC (6 years, 3 months ago) by ph10
File size: 251725 byte(s)
Error occurred while calculating annotation data.
Make back references to duplicated named subpatterns more like Perl.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.34 xx-xxxx-201x
5 --------------------------
6
7 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8 executable memory. Patch inspired by Carsten Klein.
9
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.
17
18 3. Make a small performance improvement in strlen16() and strlen32() in
19 pcretest.
20
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.
24
25 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26
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.
34
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.
37
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.
40
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:
44
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.
49
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.
54
55 10. A conditional group with an assertion condition could lead to PCRE
56 recording an incorrect first data item for a match if no other first data
57 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
58 first data item, and therefore matched "ca" after "c" instead of at the
59 start.
60
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.
63
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.
66
67 13. Added the -T and -TM options to pcretest.
68
69 14. The code in pcre_compile.c for creating the table of named capturing groups
70 has been refactored. Instead of creating the table dynamically during the
71 actual compiling pass, the information is remembered during the pre-compile
72 pass (on the stack unless there are more than 20 named groups, in which
73 case malloc() is used) and the whole table is created before the actual
74 compile happens. This has simplified the code (it is now nearly 150 lines
75 shorter) and prepared the way for better handling of references to groups
76 with duplicate names.
77
78 15. A back reference to a named subpattern when there is more than one of the
79 same name now checks them in the order in which they appear in the pattern.
80 The first one that is set is used for the reference. Previously only the
81 first one was inspected. This change makes PCRE more compatible with Perl.
82
83
84 Version 8.33 28-May-2013
85 ------------------------
86
87 1. Added 'U' to some constants that are compared to unsigned integers, to
88 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
89 variables that are added to signed variables, to ensure the result is
90 signed and can be negated.
91
92 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
93 CMake config files.
94
95 3. Revise the creation of config.h.generic so that all boolean macros are
96 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
97 overriding via -D on the command line possible.
98
99 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
100 to unsigned int is reported to make a quite noticeable speed difference in
101 a specific Windows environment. Testing on Linux did also appear to show
102 some benefit (and it is clearly not harmful). Also fixed the definition of
103 Xop which should be unsigned.
104
105 5. Related to (4), changing the definition of the intermediate variable cc
106 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
107 improvements.
108
109 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
110 unnecessary spaces.
111
112 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
113 and later.
114
115 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
116
117 9. Optimizing fast_forward_start_bits in JIT.
118
119 10. Adding support for callouts in JIT, and fixing some issues revealed
120 during this work. Namely:
121
122 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
123
124 (b) Minimum length was not checked before the matching is started.
125
126 11. The value of capture_last that is passed to callouts was incorrect in some
127 cases when there was a capture on one path that was subsequently abandoned
128 after a backtrack. Also, the capture_last value is now reset after a
129 recursion, since all captures are also reset in this case.
130
131 12. The interpreter no longer returns the "too many substrings" error in the
132 case when an overflowing capture is in a branch that is subsequently
133 abandoned after a backtrack.
134
135 13. In the pathological case when an offset vector of size 2 is used, pcretest
136 now prints out the matched string after a yield of 0 or 1.
137
138 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
139 Only supported by the JIT compiler at the moment.
140
141 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
142
143 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
144 offset of the starting point of the matching process, provided the offsets
145 vector is large enough.
146
147 17. The \A escape now records a lookbehind value of 1, though its execution
148 does not actually inspect the previous character. This is to ensure that,
149 in partial multi-segment matching, at least one character from the old
150 segment is retained when a new segment is processed. Otherwise, if there
151 are no lookbehinds in the pattern, \A might match incorrectly at the start
152 of a new segment.
153
154 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
155
156 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
157 gives some modest performance improvement in 8-bit mode.
158
159 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
160 be expressed in certain programming languages using Universal Character
161 Names.
162
163 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
164 which points out that "non characters" are not "characters that may not
165 appear in Unicode strings" but rather "characters that are reserved for
166 internal use and have only local meaning".
167
168 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
169 there was a conditional group that depended on an assertion, if the
170 assertion was false, the callout that immediately followed the alternation
171 in the condition was skipped when pcre_exec() was used for matching.
172
173 23. Allow an explicit callout to be inserted before an assertion that is the
174 condition for a conditional group, for compatibility with automatic
175 callouts, which always insert a callout at this point.
176
177 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
178 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
179
180 25. (*PRUNE) is now supported by the JIT compiler.
181
182 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
183
184 27. Fix the case where there are two or more SKIPs with arguments that may be
185 ignored.
186
187 28. (*SKIP) is now supported by the JIT compiler.
188
189 29. (*THEN) is now supported by the JIT compiler.
190
191 30. Update RunTest with additional test selector options.
192
193 31. The way PCRE handles backtracking verbs has been changed in two ways.
194
195 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
196 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
197 backtracking. In some cases this makes it more Perl-compatible, but Perl's
198 rather obscure rules do not always do the same thing.
199
200 (2) Previously, backtracking verbs were confined within assertions. This is
201 no longer the case for positive assertions, except for (*ACCEPT). Again,
202 this sometimes improves Perl compatibility, and sometimes does not.
203
204 32. A number of tests that were in test 2 because Perl did things differently
205 have been moved to test 1, because either Perl or PCRE has changed, and
206 these tests are now compatible.
207
208 32. Backtracking control verbs are now handled in the same way in JIT and
209 interpreter.
210
211 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
212 contained a forward subroutine reference caused a compile error.
213
214 34. Auto-detect and optimize limited repetitions in JIT.
215
216 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
217 blocking (*UTF) etc.
218
219 36. In the interpreter, maximizing pattern repetitions for characters and
220 character types now use tail recursion, which reduces stack usage.
221
222 37. The value of the max lookbehind was not correctly preserved if a compiled
223 and saved regex was reloaded on a host of different endianness.
224
225 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
226 of the compiled pattern block, expand the flags field from 16 to 32 bits
227 because it was almost full.
228
229 39. Try madvise first before posix_madvise.
230
231 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
232 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
233 original change locked out all matches of empty strings. This has been
234 changed so that one match of an empty string per line is recognized.
235 Subsequent searches on the same line (for colouring or for --only-matching,
236 for example) do not recognize empty strings.
237
238 41. Applied a user patch to fix a number of spelling mistakes in comments.
239
240 42. Data lines longer than 65536 caused pcretest to crash.
241
242 43. Clarified the data type for length and startoffset arguments for pcre_exec
243 and pcre_dfa_exec in the function-specific man pages, where they were
244 explicitly stated to be in bytes, never having been updated. I also added
245 some clarification to the pcreapi man page.
246
247 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
248 a segmentation fault.
249
250
251 Version 8.32 30-November-2012
252 -----------------------------
253
254 1. Improved JIT compiler optimizations for first character search and single
255 character iterators.
256
257 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
258 Patch by Daniel Richard G.
259
260 3. Single character iterator optimizations in the JIT compiler.
261
262 4. Improved JIT compiler optimizations for character ranges.
263
264 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
265 Reported by Giuseppe D'Angelo.
266
267 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
268 of a line, was being set incorrectly in cases where .* appeared inside
269 atomic brackets at the start of a pattern, or where there was a subsequent
270 *PRUNE or *SKIP.
271
272 7. Improved instruction cache flush for POWER/PowerPC.
273 Patch by Daniel Richard G.
274
275 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
276 grep:
277
278 (a) There is now no limit to the number of patterns to be matched.
279
280 (b) An error is given if a pattern is too long.
281
282 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
283 are now supported.
284
285 (d) --exclude-from and --include-from (multiple use) have been added.
286
287 (e) Exclusions and inclusions now apply to all files and directories, not
288 just to those obtained from scanning a directory recursively.
289
290 (f) Multiple uses of -f and --file-list are now supported.
291
292 (g) In a Windows environment, the default for -d has been changed from
293 "read" (the GNU grep default) to "skip", because otherwise the presence
294 of a directory in the file list provokes an error.
295
296 (h) The documentation has been revised and clarified in places.
297
298 9. Improve the matching speed of capturing brackets.
299
300 10. Changed the meaning of \X so that it now matches a Unicode extended
301 grapheme cluster.
302
303 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
304 out POSIX threads when JIT support is configured.
305
306 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
307
308 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
309 structure could go wrong in environments where size_t is not the same size
310 as int.
311
312 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
313 set.
314
315 15. The EBCDIC support had decayed; later updates to the code had included
316 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
317 general tidy up of EBCDIC-related issues, and the documentation was also
318 not quite right. There is now a test that can be run on ASCII systems to
319 check some of the EBCDIC-related things (but is it not a full test).
320
321 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
322 in a small tidy to the code.
323
324 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
325
326 18. If the --only-matching (-o) option in pcregrep is specified multiple
327 times, each one causes appropriate output. For example, -o1 -o2 outputs the
328 substrings matched by the 1st and 2nd capturing parentheses. A separating
329 string can be specified by --om-separator (default empty).
330
331 19. Improving the first n character searches.
332
333 20. Turn case lists for horizontal and vertical white space into macros so that
334 they are defined only once.
335
336 21. This set of changes together give more compatible Unicode case-folding
337 behaviour for characters that have more than one other case when UCP
338 support is available.
339
340 (a) The Unicode property table now has offsets into a new table of sets of
341 three or more characters that are case-equivalent. The MultiStage2.py
342 script that generates these tables (the pcre_ucd.c file) now scans
343 CaseFolding.txt instead of UnicodeData.txt for character case
344 information.
345
346 (b) The code for adding characters or ranges of characters to a character
347 class has been abstracted into a generalized function that also handles
348 case-independence. In UTF-mode with UCP support, this uses the new data
349 to handle characters with more than one other case.
350
351 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
352 whose other case is greater than 256 are now correctly matched
353 caselessly. Previously, the high codepoint matched the low one, but not
354 vice versa.
355
356 (d) The processing of \h, \H, \v, and \ in character classes now makes use
357 of the new class addition function, using character lists defined as
358 macros alongside the case definitions of 20 above.
359
360 (e) Caseless back references now work with characters that have more than
361 one other case.
362
363 (f) General caseless matching of characters with more than one other case
364 is supported.
365
366 22. Unicode character properties were updated from Unicode 6.2.0
367
368 23. Improved CMake support under Windows. Patch by Daniel Richard G.
369
370 24. Add support for 32-bit character strings, and UTF-32
371
372 25. Major JIT compiler update (code refactoring and bugfixing).
373 Experimental Sparc 32 support is added.
374
375 26. Applied a modified version of Daniel Richard G's patch to create
376 pcre.h.generic and config.h.generic by "make" instead of in the
377 PrepareRelease script.
378
379 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
380 pcre_compile.c when checking for a zero character.
381
382 28. Introducing a native interface for JIT. Through this interface, the compiled
383 machine code can be directly executed. The purpose of this interface is to
384 provide fast pattern matching, so several sanity checks are not performed.
385 However, feature tests are still performed. The new interface provides
386 1.4x speedup compared to the old one.
387
388 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
389 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
390 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
391
392 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
393 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
394 the "old" RFC 2279). Instead, it ended up passing a negative length to
395 pcre_exec().
396
397 31. Add support for GCC's visibility feature to hide internal functions.
398
399 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
400 "unknown -C option" after outputting 0 or 1.
401
402 33. There is now support for generating a code coverage report for the test
403 suite in environments where gcc is the compiler and lcov is installed. This
404 is mainly for the benefit of the developers.
405
406 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
407 unaddressable using valgrind annotations, allowing valgrind to detect
408 invalid memory accesses. This is mainly for the benefit of the developers.
409
410 25. (*UTF) can now be used to start a pattern in any of the three libraries.
411
412 26. Give configure error if --enable-cpp but no C++ compiler found.
413
414
415 Version 8.31 06-July-2012
416 -------------------------
417
418 1. Fixing a wrong JIT test case and some compiler warnings.
419
420 2. Removed a bashism from the RunTest script.
421
422 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
423 to unsigned type, result still unsigned" that was given by an MS compiler
424 on encountering the code "-sizeof(xxx)".
425
426 4. Partial matching support is added to the JIT compiler.
427
428 5. Fixed several bugs concerned with partial matching of items that consist
429 of more than one character:
430
431 (a) /^(..)\1/ did not partially match "aba" because checking references was
432 done on an "all or nothing" basis. This also applied to repeated
433 references.
434
435 (b) \R did not give a hard partial match if \r was found at the end of the
436 subject.
437
438 (c) \X did not give a hard partial match after matching one or more
439 characters at the end of the subject.
440
441 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
442 a partial match for the string "\r".
443
444 (e) When newline was set to CRLF, the metacharacter "." did not recognize
445 a partial match for a CR character at the end of the subject string.
446
447 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
448 running pcretest, the text "(JIT)" added to the output whenever JIT is
449 actually used to run the match.
450
451 7. Individual JIT compile options can be set in pcretest by following -s+[+]
452 or /S+[+] with a digit between 1 and 7.
453
454 8. OP_NOT now supports any UTF character not just single-byte ones.
455
456 9. (*MARK) control verb is now supported by the JIT compiler.
457
458 10. The command "./RunTest list" lists the available tests without actually
459 running any of them. (Because I keep forgetting what they all are.)
460
461 11. Add PCRE_INFO_MAXLOOKBEHIND.
462
463 12. Applied a (slightly modified) user-supplied patch that improves performance
464 when the heap is used for recursion (compiled with --disable-stack-for-
465 recursion). Instead of malloc and free for each heap frame each time a
466 logical recursion happens, frames are retained on a chain and re-used where
467 possible. This sometimes gives as much as 30% improvement.
468
469 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
470 call.
471
472 14. As documented, (*COMMIT) is now confined to within a positive assertion.
473
474 15. It is now possible to link pcretest with libedit as an alternative to
475 libreadline.
476
477 16. (*COMMIT) control verb is now supported by the JIT compiler.
478
479 17. The Unicode data tables have been updated to Unicode 6.1.0.
480
481 18. Added --file-list option to pcregrep.
482
483 19. Added binary file support to pcregrep, including the -a, --binary-files,
484 -I, and --text options.
485
486 20. The madvise function is renamed for posix_madvise for QNX compatibility
487 reasons. Fixed by Giuseppe D'Angelo.
488
489 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
490
491 22. Changed the option for creating symbolic links for 16-bit man pages from
492 -s to -sf so that re-installing does not cause issues.
493
494 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
495
496 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
497 matches in certain environments (the workspace was not being correctly
498 retained). Also added to pcre_dfa_exec() a simple plausibility check on
499 some of the workspace data at the beginning of a restart.
500
501 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
502 was not doing so when it should - probably a typo introduced by SVN 528
503 (change 8.10/14).
504
505 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
506 \w+ when the character tables indicated that \x{c4} was a word character.
507 There were several related cases, all because the tests for doing a table
508 lookup were testing for characters less than 127 instead of 255.
509
510 27. If a pattern contains capturing parentheses that are not used in a match,
511 their slots in the ovector are set to -1. For those that are higher than
512 any matched groups, this happens at the end of processing. In the case when
513 there were back references that the ovector was too small to contain
514 (causing temporary malloc'd memory to be used during matching), and the
515 highest capturing number was not used, memory off the end of the ovector
516 was incorrectly being set to -1. (It was using the size of the temporary
517 memory instead of the true size.)
518
519 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
520 ovector size, it uses memory at the end of the block that it has got.
521
522 29. Check for an overlong MARK name and give an error at compile time. The
523 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
524
525 30. JIT compiler update.
526
527 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
528 Rill for the patch.
529
530 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
531
532 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
533
534 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
535 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
536 was enabled.
537
538 35. Improve JIT code generation for greedy plus quantifier.
539
540 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
541 1 to "aa" instead of to an empty string. The bug affected repeated groups
542 that could potentially match an empty string.
543
544 37. Optimizing single character iterators in JIT.
545
546 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
547 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
548 codepoints that are too big for the mode are faulted, and in a UTF mode,
549 disallowed codepoints are also faulted.
550
551 39. If PCRE was compiled with UTF support, in three places in the DFA
552 matcher there was code that should only have been obeyed in UTF mode, but
553 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
554 processing when bytes with values greater than 127 were present. In 16-bit
555 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
556 both cases the values are those that cannot be the first data item in a UTF
557 character. The three items that might have provoked this were recursions,
558 possessively repeated groups, and atomic groups.
559
560 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
561 and pcregrep, because some OS require shared objects to be explicitly
562 passed to ld, causing the link step to fail if they are not.
563
564 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
565 mode, patterns that started with \h* or \R* might be incorrectly matched.
566
567
568 Version 8.30 04-February-2012
569 -----------------------------
570
571 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
572 name is defined in ctype.h.
573
574 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
575 only in quite long subpatterns.
576
577 3. Removed the function pcre_info(), which has been obsolete and deprecated
578 since it was replaced by pcre_fullinfo() in February 2000.
579
580 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
581 match a (*MARK), and the match failed at the start of the subject, a
582 reference to memory before the start of the subject could occur. This bug
583 was introduced by fix 17 of release 8.21.
584
585 5. A reference to an unset group with zero minimum repetition was giving
586 totally wrong answers (in non-JavaScript-compatibility mode). For example,
587 /(another)?(\1?)test/ matched against "hello world test". This bug was
588 introduced in release 8.13.
589
590 6. Add support for 16-bit character strings (a large amount of work involving
591 many changes and refactorings).
592
593 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
594 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
595 from a file.
596
597 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
598 rounding is not applied in this particular case).
599
600 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
601 if they appear, or are escaped, in patterns.
602
603 10. Get rid of a number of -Wunused-but-set-variable warnings.
604
605 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
606 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
607 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
608 also returns the mark "x". This bug applied to capturing parentheses,
609 non-capturing parentheses, and atomic parentheses. It also applied to some
610 assertions.
611
612 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
613 information out of configure.ac instead of relying on pcre.h.generic, which
614 is not stored in the repository.
615
616 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
617 -lreadline.
618
619 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
620
621 15. Applied Graycode's patch to put the top-level frame on the stack rather
622 than the heap when not using the stack for recursion. This gives a
623 performance improvement in many cases when recursion is not deep.
624
625 16. Experimental code added to "pcretest -C" to output the stack frame size.
626
627
628 Version 8.21 12-Dec-2011
629 ------------------------
630
631 1. Updating the JIT compiler.
632
633 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
634 are added as well.
635
636 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
637 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
638 calling _pcre_jit_exec. Some extra comments are added.
639
640 4. (*MARK) settings inside atomic groups that do not contain any capturing
641 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
642 was introduced by change 18 for 8.20.
643
644 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
645 ECMA-262 standard.
646
647 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
648 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
649 This bug was probably introduced by change 9 of 8.13.
650
651 7. While fixing 6 above, I noticed that a number of other items were being
652 incorrectly rejected as "not fixed length". This arose partly because newer
653 opcodes had not been added to the fixed-length checking code. I have (a)
654 corrected the bug and added tests for these items, and (b) arranged for an
655 error to occur if an unknown opcode is encountered while checking for fixed
656 length instead of just assuming "not fixed length". The items that were
657 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
658 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
659 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
660
661 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
662 being incorrectly compiled and would have given unpredicatble results.
663
664 9. A possessively repeated subpattern with minimum repeat count greater than
665 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
666 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
667 the first (A) could occur when it should not.
668
669 10. Add a cast and remove a redundant test from the code.
670
671 11. JIT should use pcre_malloc/pcre_free for allocation.
672
673 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
674 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
675 is anything other than /usr, -L is still shown).
676
677 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
678
679 14. Perl does not support \N without a following name in a [] class; PCRE now
680 also gives an error.
681
682 15. If a forward reference was repeated with an upper limit of around 2000,
683 it caused the error "internal error: overran compiling workspace". The
684 maximum number of forward references (including repeats) was limited by the
685 internal workspace, and dependent on the LINK_SIZE. The code has been
686 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
687 the default depends on LINK_SIZE. There is a new upper limit (for safety)
688 of around 200,000 forward references. While doing this, I also speeded up
689 the filling in of repeated forward references.
690
691 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
692 incorrectly expecting the subject to contain another "a" after the start.
693
694 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
695 in the match, the SKIP should be ignored. This was not happening; instead
696 the SKIP was being treated as NOMATCH. For patterns such as
697 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
698 tested.
699
700 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
701 now much more compatible with Perl, in particular in cases where the result
702 is a non-match for a non-anchored pattern. For example, if
703 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
704 "m", where previously it did not return a name. A side effect of this
705 change is that for partial matches, the last encountered mark name is
706 returned, as for non matches. A number of tests that were previously not
707 Perl-compatible have been moved into the Perl-compatible test files. The
708 refactoring has had the pleasing side effect of removing one argument from
709 the match() function, thus reducing its stack requirements.
710
711 19. If the /S+ option was used in pcretest to study a pattern using JIT,
712 subsequent uses of /S (without +) incorrectly behaved like /S+.
713
714 21. Retrieve executable code size support for the JIT compiler and fixing
715 some warnings.
716
717 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
718 not work when the shorter character appeared right at the end of the
719 subject string.
720
721 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
722 systems.
723
724 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
725 output it when the /M option is used in pcretest.
726
727 25. The CheckMan script was not being included in the distribution. Also, added
728 an explicit "perl" to run Perl scripts from the PrepareRelease script
729 because this is reportedly needed in Windows.
730
731 26. If study data was being save in a file and studying had not found a set of
732 "starts with" bytes for the pattern, the data written to the file (though
733 never used) was taken from uninitialized memory and so caused valgrind to
734 complain.
735
736 27. Updated RunTest.bat as provided by Sheri Pierce.
737
738 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
739
740 29. Computation of memory usage for the table of capturing group names was
741 giving an unnecessarily large value.
742
743
744 Version 8.20 21-Oct-2011
745 ------------------------
746
747 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
748 a POSIX class. After further experiments with Perl, which convinced me that
749 Perl has bugs and confusions, a closing square bracket is no longer allowed
750 in a POSIX name. This bug also affected patterns with classes that started
751 with full stops.
752
753 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
754 captured substring, but while checking the failing first alternative,
755 substring 1 is temporarily captured. If the output vector supplied to
756 pcre_exec() was not big enough for this capture, the yield of the function
757 was still zero ("insufficient space for captured substrings"). This cannot
758 be totally fixed without adding another stack variable, which seems a lot
759 of expense for a edge case. However, I have improved the situation in cases
760 such as /(a)(b)x|abc/ matched against "abc", where the return code
761 indicates that fewer than the maximum number of slots in the ovector have
762 been set.
763
764 3. Related to (2) above: when there are more back references in a pattern than
765 slots in the output vector, pcre_exec() uses temporary memory during
766 matching, and copies in the captures as far as possible afterwards. It was
767 using the entire output vector, but this conflicts with the specification
768 that only 2/3 is used for passing back captured substrings. Now it uses
769 only the first 2/3, for compatibility. This is, of course, another edge
770 case.
771
772 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
773 main code base, and can be used by building with --enable-jit. When this is
774 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
775 runtime --no-jit option is given.
776
777 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
778 ovector, the return from the function was zero, implying that there were
779 other matches that did not fit. The correct "exactly full" value is now
780 returned.
781
782 6. If a subpattern that was called recursively or as a subroutine contained
783 (*PRUNE) or any other control that caused it to give a non-standard return,
784 invalid errors such as "Error -26 (nested recursion at the same subject
785 position)" or even infinite loops could occur.
786
787 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
788 computing the minimum length on reaching *ACCEPT, and so ended up with the
789 wrong value of 1 rather than 0. Further investigation indicates that
790 computing a minimum subject length in the presence of *ACCEPT is difficult
791 (think back references, subroutine calls), and so I have changed the code
792 so that no minimum is registered for a pattern that contains *ACCEPT.
793
794 8. If (*THEN) was present in the first (true) branch of a conditional group,
795 it was not handled as intended. [But see 16 below.]
796
797 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
798 Sheri Pierce.
799
800 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
801 the first byte in a match must be "a".
802
803 11. Change 17 for 8.13 increased the recursion depth for patterns like
804 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
805 pattern contains any instances of (*THEN). If it does not, the old
806 optimizations are restored. It would be nice to do this on a per-group
807 basis, but at the moment that is not feasible.
808
809 12. In some environments, the output of pcretest -C is CRLF terminated. This
810 broke RunTest's code that checks for the link size. A single white space
811 character after the value is now allowed for.
812
813 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
814 For "fr", it uses the Windows-specific input and output files.
815
816 14. If (*THEN) appeared in a group that was called recursively or as a
817 subroutine, it did not work as intended. [But see next item.]
818
819 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
820 pattern fragments (but not containing any | characters). If A and B are
821 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
822 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
823 D. In other words, Perl considers parentheses that do not contain any |
824 characters to be part of a surrounding alternative, whereas PCRE was
825 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
826 differently. PCRE now behaves in the same way as Perl, except in the case
827 of subroutine/recursion calls such as (?1) which have in any case always
828 been different (but PCRE had them first :-).
829
830 16. Related to 15 above: Perl does not treat the | in a conditional group as
831 creating alternatives. Such a group is treated in the same way as an
832 ordinary group without any | characters when processing (*THEN). PCRE has
833 been changed to match Perl's behaviour.
834
835 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
836 RunGrepTest script failed.
837
838 18. Change 22 for version 13 caused atomic groups to use more stack. This is
839 inevitable for groups that contain captures, but it can lead to a lot of
840 stack use in large patterns. The old behaviour has been restored for atomic
841 groups that do not contain any capturing parentheses.
842
843 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
844 suppress the check for a minimum subject length at run time. (If it was
845 given to pcre_exec() or pcre_dfa_exec() it did work.)
846
847 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
848 fail to work when decoding hex characters in data strings in EBCDIC
849 environments.
850
851 21. It appears that in at least one Mac OS environment, the isxdigit() function
852 is implemented as a macro that evaluates to its argument more than once,
853 contravening the C 90 Standard (I haven't checked a later standard). There
854 was an instance in pcretest which caused it to go wrong when processing
855 \x{...} escapes in subject strings. The has been rewritten to avoid using
856 things like p++ in the argument of isxdigit().
857
858
859 Version 8.13 16-Aug-2011
860 ------------------------
861
862 1. The Unicode data tables have been updated to Unicode 6.0.0.
863
864 2. Two minor typos in pcre_internal.h have been fixed.
865
866 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
867 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
868 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
869
870 4. There were a number of related bugs in the code for matching backrefences
871 caselessly in UTF-8 mode when codes for the characters concerned were
872 different numbers of bytes. For example, U+023A and U+2C65 are an upper
873 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
874 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
875 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
876 2-byte code at the end of the subject (it thought there wasn't enough data
877 left).
878
879 5. Comprehensive information about what went wrong is now returned by
880 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
881 as the output vector has at least 2 elements. The offset of the start of
882 the failing character and a reason code are placed in the vector.
883
884 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
885 now returned is for the first byte of the failing character, instead of the
886 last byte inspected. This is an incompatible change, but I hope it is small
887 enough not to be a problem. It makes the returned offset consistent with
888 pcre_exec() and pcre_dfa_exec().
889
890 7. pcretest now gives a text phrase as well as the error number when
891 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
892 failure, the offset and reason code are output.
893
894 8. When \R was used with a maximizing quantifier it failed to skip backwards
895 over a \r\n pair if the subsequent match failed. Instead, it just skipped
896 back over a single character (\n). This seems wrong (because it treated the
897 two characters as a single entity when going forwards), conflicts with the
898 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
899 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
900 has been changed.
901
902 9. Some internal refactoring has changed the processing so that the handling
903 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
904 time (the PCRE_DOTALL option was changed this way some time ago: version
905 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
906 which was always a bit of a fudge. It also means that there is one less
907 argument for the match() function, which reduces its stack requirements
908 slightly. This change also fixes an incompatibility with Perl: the pattern
909 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
910
911 10. More internal refactoring has drastically reduced the number of recursive
912 calls to match() for possessively repeated groups such as (abc)++ when
913 using pcre_exec().
914
915 11. While implementing 10, a number of bugs in the handling of groups were
916 discovered and fixed:
917
918 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
919 (a|)*(?1) gave a compile-time internal error.
920 ((a|)+)+ did not notice that the outer group could match an empty string.
921 (^a|^)+ was not marked as anchored.
922 (.*a|.*)+ was not marked as matching at start or after a newline.
923
924 12. Yet more internal refactoring has removed another argument from the match()
925 function. Special calls to this function are now indicated by setting a
926 value in a variable in the "match data" data block.
927
928 13. Be more explicit in pcre_study() instead of relying on "default" for
929 opcodes that mean there is no starting character; this means that when new
930 ones are added and accidentally left out of pcre_study(), testing should
931 pick them up.
932
933 14. The -s option of pcretest has been documented for ages as being an old
934 synonym of -m (show memory usage). I have changed it to mean "force study
935 for every regex", that is, assume /S for every regex. This is similar to -i
936 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
937 using it. It makes it easier to run collections of tests with and without
938 study enabled, and thereby test pcre_study() more easily. All the standard
939 tests are now run with and without -s (but some patterns can be marked as
940 "never study" - see 20 below).
941
942 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
943 restoration of the capturing data to the outer values was not happening
944 correctly.
945
946 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
947 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
948 pattern had matched an empty string, and so incorrectly returned a no
949 match.
950
951 17. There was optimizing code for the last branch of non-capturing parentheses,
952 and also for the obeyed branch of a conditional subexpression, which used
953 tail recursion to cut down on stack usage. Unfortunately, now that there is
954 the possibility of (*THEN) occurring in these branches, tail recursion is
955 no longer possible because the return has to be checked for (*THEN). These
956 two optimizations have therefore been removed. [But see 8.20/11 above.]
957
958 18. If a pattern containing \R was studied, it was assumed that \R always
959 matched two bytes, thus causing the minimum subject length to be
960 incorrectly computed because \R can also match just one byte.
961
962 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
963 was incorrectly computed.
964
965 20. If /S is present twice on a test pattern in pcretest input, it now
966 *disables* studying, thereby overriding the use of -s on the command line
967 (see 14 above). This is necessary for one or two tests to keep the output
968 identical in both cases.
969
970 21. When (*ACCEPT) was used in an assertion that matched an empty string and
971 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
972
973 22. When an atomic group that contained a capturing parenthesis was
974 successfully matched, but the branch in which it appeared failed, the
975 capturing was not being forgotten if a higher numbered group was later
976 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
977 group 1 to "a", when in fact it should be unset. This applied to multi-
978 branched capturing and non-capturing groups, repeated or not, and also to
979 positive assertions (capturing in negative assertions does not happen
980 in PCRE) and also to nested atomic groups.
981
982 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
983 subject after a captured substring, to make it easier to tell which of a
984 number of identical substrings has been captured.
985
986 24. The way atomic groups are processed by pcre_exec() has been changed so that
987 if they are repeated, backtracking one repetition now resets captured
988 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
989 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
990 "aaa". Previously, it would have been "a". As part of this code
991 refactoring, the way recursive calls are handled has also been changed.
992
993 25. If an assertion condition captured any substrings, they were not passed
994 back unless some other capturing happened later. For example, if
995 (?(?=(a))a) was matched against "a", no capturing was returned.
996
997 26. When studying a pattern that contained subroutine calls or assertions,
998 the code for finding the minimum length of a possible match was handling
999 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1000 group 1 called group 2 while simultaneously a separate group 2 called group
1001 1). A stack overflow occurred in this case. I have fixed this by limiting
1002 the recursion depth to 10.
1003
1004 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1005 Fortmann. This supports explicit test numbers on the command line, and has
1006 argument validation and error reporting.
1007
1008 28. An instance of \X with an unlimited repeat could fail if at any point the
1009 first character it looked at was a mark character.
1010
1011 29. Some minor code refactoring concerning Unicode properties and scripts
1012 should reduce the stack requirement of match() slightly.
1013
1014 30. Added the '=' option to pcretest to check the setting of unused capturing
1015 slots at the end of the pattern, which are documented as being -1, but are
1016 not included in the return count.
1017
1018 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1019 compiled something random. Now it gives a compile-time error (as does
1020 Perl).
1021
1022 32. A *MARK encountered during the processing of a positive assertion is now
1023 recorded and passed back (compatible with Perl).
1024
1025 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1026 had alternative anchored branches, the search for a second match in a line
1027 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1028 matched the line "0102" twice. The same bug affected patterns that started
1029 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1030 twice.
1031
1032 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1033 does, and because of capturing effects, quantifying parenthesized
1034 assertions may at times be useful. Quantifiers are now allowed for
1035 parenthesized assertions.
1036
1037 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1038
1039 36. \g was being checked for fancy things in a character class, when it should
1040 just be a literal "g".
1041
1042 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1043 appearance of a nested POSIX class supersedes an apparent external class.
1044 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1045 unescaped square brackets may also appear as part of class names. For
1046 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1047 more like Perl. (But see 8.20/1 above.)
1048
1049 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1050 was because it thought it was \N{name}, which is not supported).
1051
1052 39. Add minix to OS list not supporting the -S option in pcretest.
1053
1054 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1055 cannot analyze patterns in sufficient detail to catch mutual recursions
1056 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1057 subgroup is called recursively as a subpattern for a second time at the
1058 same position in the subject string. In previous releases this might have
1059 been caught by the recursion limit, or it might have run out of stack.
1060
1061 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1062 happen only once. PCRE was, however incorrectly giving a compile time error
1063 "recursive call could loop indefinitely" because it cannot analyze the
1064 pattern in sufficient detail. The compile time test no longer happens when
1065 PCRE is compiling a conditional subpattern, but actual runaway loops are
1066 now caught at runtime (see 40 above).
1067
1068 42. It seems that Perl allows any characters other than a closing parenthesis
1069 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1070 has been changed to be the same.
1071
1072 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1073 as not to get warnings when autogen.sh is called. Also changed
1074 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1075
1076 44. To help people who use pcregrep to scan files containing exceedingly long
1077 lines, the following changes have been made:
1078
1079 (a) The default value of the buffer size parameter has been increased from
1080 8K to 20K. (The actual buffer used is three times this size.)
1081
1082 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1083 PCRE is built.
1084
1085 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1086 to be set at run time.
1087
1088 (d) Numerical values in pcregrep options can be followed by K or M, for
1089 example --buffer-size=50K.
1090
1091 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1092 given and the return code is set to 2.
1093
1094 45. Add a pointer to the latest mark to the callout data block.
1095
1096 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1097 partial match of an empty string instead of no match. This was specific to
1098 the use of ".".
1099
1100 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1101 complete match instead of a partial match. This bug was dependent on both
1102 the PCRE_UTF8 and PCRE_DOTALL options being set.
1103
1104 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1105 starting byte set, because \b was not being ignored.
1106
1107
1108 Version 8.12 15-Jan-2011
1109 ------------------------
1110
1111 1. Fixed some typos in the markup of the man pages, and wrote a script that
1112 checks for such things as part of the documentation building process.
1113
1114 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1115 --match-limit and --recursion-limit options (added for 8.11). In
1116 particular, this made one of the standard tests fail. (The integer value
1117 went into the wrong half of a long int.)
1118
1119 3. If the --colour option was given to pcregrep with -v (invert match), it
1120 did strange things, either producing crazy output, or crashing. It should,
1121 of course, ignore a request for colour when reporting lines that do not
1122 match.
1123
1124 4. Another pcregrep bug caused similar problems if --colour was specified with
1125 -M (multiline) and the pattern match finished with a line ending.
1126
1127 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1128 matched in multiline mode, the following line was shown as part of the
1129 match. This seems wrong, so I have changed it.
1130
1131 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1132 the check for further matches in the same line (so they could be coloured)
1133 to overrun the end of the current line. If another match was found, it was
1134 incorrectly shown (and then shown again when found in the next line).
1135
1136 7. If pcregrep was compiled under Windows, there was a reference to the
1137 function pcregrep_exit() before it was defined. I am assuming this was
1138 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1139 reported by a user. I've moved the definition above the reference.
1140
1141
1142 Version 8.11 10-Dec-2010
1143 ------------------------
1144
1145 1. (*THEN) was not working properly if there were untried alternatives prior
1146 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1147 backtracked to try for "b" instead of moving to the next alternative branch
1148 at the same level (in this case, to look for "c"). The Perl documentation
1149 is clear that when (*THEN) is backtracked onto, it goes to the "next
1150 alternative in the innermost enclosing group".
1151
1152 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1153 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1154 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1155 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1156 (*THEN).
1157
1158 3. If \s appeared in a character class, it removed the VT character from
1159 the class, even if it had been included by some previous item, for example
1160 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1161 of \s, but is part of the POSIX "space" class.)
1162
1163 4. A partial match never returns an empty string (because you can always
1164 match an empty string at the end of the subject); however the checking for
1165 an empty string was starting at the "start of match" point. This has been
1166 changed to the "earliest inspected character" point, because the returned
1167 data for a partial match starts at this character. This means that, for
1168 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1169 (previously it gave "no match").
1170
1171 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1172 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1173 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1174 has an implication that the given string is incomplete (because a partial
1175 match is preferred over a full match). For this reason, these items now
1176 give a partial match in this situation. [Aside: previously, the one case
1177 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1178 match rather than a full match, which was wrong by the old rules, but is
1179 now correct.]
1180
1181 6. There was a bug in the handling of #-introduced comments, recognized when
1182 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1183 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1184 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1185 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1186 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1187 places in pcre_compile().
1188
1189 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1190 comments when looking ahead for named forward references to subpatterns,
1191 the only newline sequence it recognized was NL. It now handles newlines
1192 according to the set newline convention.
1193
1194 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1195 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1196 cater for a lack of strerror(). These oversights have been fixed.
1197
1198 9. Added --match-limit and --recursion-limit to pcregrep.
1199
1200 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1201
1202 11. When the -o option was used, pcregrep was setting a return code of 1, even
1203 when matches were found, and --line-buffered was not being honoured.
1204
1205 12. Added an optional parentheses number to the -o and --only-matching options
1206 of pcregrep.
1207
1208 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1209 can match an empty string. The code to do it in pcretest and pcredemo
1210 needed fixing:
1211
1212 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1213 only one byte after an empty string match just before CRLF (this case
1214 just got forgotten; "any" and "anycrlf" were OK).
1215
1216 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1217 mode when an empty string match preceded an ASCII character followed by
1218 a non-ASCII character. (The code for advancing by one character rather
1219 than one byte was nonsense.)
1220
1221 (c) The pcredemo.c sample program did not have any code at all to handle
1222 the cases when CRLF is a valid newline sequence.
1223
1224 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1225 as a starting offset was within the subject string. There is now a new
1226 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1227 negative or greater than the length of the string. In order to test this,
1228 pcretest is extended to allow the setting of negative starting offsets.
1229
1230 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1231 starting offset points to the beginning of a UTF-8 character was
1232 unnecessarily clumsy. I tidied it up.
1233
1234 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1235 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1236
1237 17. Nobody had reported that the --include_dir option, which was added in
1238 release 7.7 should have been called --include-dir (hyphen, not underscore)
1239 for compatibility with GNU grep. I have changed it to --include-dir, but
1240 left --include_dir as an undocumented synonym, and the same for
1241 --exclude-dir, though that is not available in GNU grep, at least as of
1242 release 2.5.4.
1243
1244 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1245 characters from a string of bytes) have been redefined so as not to use
1246 loops, in order to improve performance in some environments. At the same
1247 time, I abstracted some of the common code into auxiliary macros to save
1248 repetition (this should not affect the compiled code).
1249
1250 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1251 compile-time error is now given if \c is not followed by an ASCII
1252 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1253 different, and any byte value is allowed.)
1254
1255 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1256 START_OPTIMIZE option, which is now allowed at compile time - but just
1257 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1258 to pcregrep and other applications that have no direct access to PCRE
1259 options. The new /Y option in pcretest sets this option when calling
1260 pcre_compile().
1261
1262 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1263 back references. Groups containing recursive back references were forced to
1264 be atomic by that change, but in the case of named groups, the amount of
1265 memory required was incorrectly computed, leading to "Failed: internal
1266 error: code overflow". This has been fixed.
1267
1268 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1269 pcretest.c, to avoid build problems in some Borland environments.
1270
1271
1272 Version 8.10 25-Jun-2010
1273 ------------------------
1274
1275 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1276 THEN.
1277
1278 2. (*ACCEPT) was not working when inside an atomic group.
1279
1280 3. Inside a character class, \B is treated as a literal by default, but
1281 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1282 causes the error). The code is unchanged, but I tidied the documentation.
1283
1284 4. Inside a character class, PCRE always treated \R and \X as literals,
1285 whereas Perl faults them if its -w option is set. I have changed PCRE so
1286 that it faults them when PCRE_EXTRA is set.
1287
1288 5. Added support for \N, which always matches any character other than
1289 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1290
1291 6. When compiling pcregrep with newer versions of gcc which may have
1292 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1293 declared with attribute warn_unused_result" were given. Just casting the
1294 result to (void) does not stop the warnings; a more elaborate fudge is
1295 needed. I've used a macro to implement this.
1296
1297 7. Minor change to pcretest.c to avoid a compiler warning.
1298
1299 8. Added four artifical Unicode properties to help with an option to make
1300 \s etc use properties (see next item). The new properties are: Xan
1301 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1302
1303 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1304 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1305 this option. Modified pcretest to add /W to test this facility. Added
1306 REG_UCP to make it available via the POSIX interface.
1307
1308 10. Added --line-buffered to pcregrep.
1309
1310 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1311 studied, and the match started with a letter with a code point greater than
1312 127 whose first byte was different to the first byte of the other case of
1313 the letter, the other case of this starting letter was not recognized
1314 (#976).
1315
1316 12. If a pattern that was studied started with a repeated Unicode property
1317 test, for example, \p{Nd}+, there was the theoretical possibility of
1318 setting up an incorrect bitmap of starting bytes, but fortunately it could
1319 not have actually happened in practice until change 8 above was made (it
1320 added property types that matched character-matching opcodes).
1321
1322 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1323 possible starting bytes for non-anchored patterns.
1324
1325 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1326 \R, and also a number of cases that involve Unicode properties, both
1327 explicit and implicit when PCRE_UCP is set.
1328
1329 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1330 input, it could crash or give wrong results if characters with values
1331 greater than 0xc0 were present in the subject string. (Detail: it assumed
1332 UTF-8 input when processing these items.)
1333
1334 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1335 size_t is 64-bit (#991).
1336
1337 17. Added a check for running out of memory when PCRE is compiled with
1338 --disable-stack-for-recursion (#990).
1339
1340 18. If the last data line in a file for pcretest does not have a newline on
1341 the end, a newline was missing in the output.
1342
1343 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1344 less than 128) in its various bitmaps. However, there is a facility for
1345 generating tables according to the current locale when PCRE is compiled. It
1346 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1347 characters, are recognized by isspace() and therefore were getting set in
1348 these tables, and indeed these tables seem to approximate to ISO 8859. This
1349 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1350 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1351 which of course cannot start UTF-8 characters. I have changed the code so
1352 that only real ASCII characters (less than 128) and the correct starting
1353 bytes for UTF-8 encodings are set for characters greater than 127 when in
1354 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1355 altogether.)
1356
1357 20. Added the /T option to pcretest so as to be able to run tests with non-
1358 standard character tables, thus making it possible to include the tests
1359 used for 19 above in the standard set of tests.
1360
1361 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1362 reference to a subpattern the other side of a comment that contains an
1363 opening parenthesis caused either an internal compiling error, or a
1364 reference to the wrong subpattern.
1365
1366
1367 Version 8.02 19-Mar-2010
1368 ------------------------
1369
1370 1. The Unicode data tables have been updated to Unicode 5.2.0.
1371
1372 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1373 configured.
1374
1375 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1376 original author of that file, following a query about its status.
1377
1378 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1379 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1380
1381 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1382 quantifier applied to a forward-referencing subroutine call, could compile
1383 incorrect code or give the error "internal error: previously-checked
1384 referenced subpattern not found".
1385
1386 6. Both MS Visual Studio and Symbian OS have problems with initializing
1387 variables to point to external functions. For these systems, therefore,
1388 pcre_malloc etc. are now initialized to local functions that call the
1389 relevant global functions.
1390
1391 7. There were two entries missing in the vectors called coptable and poptable
1392 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1393 I've fixed the data, and added a kludgy way of testing at compile time that
1394 the lengths are correct (equal to the number of opcodes).
1395
1396 8. Following on from 7, I added a similar kludge to check the length of the
1397 eint vector in pcreposix.c.
1398
1399 9. Error texts for pcre_compile() are held as one long string to avoid too
1400 much relocation at load time. To find a text, the string is searched,
1401 counting zeros. There was no check for running off the end of the string,
1402 which could happen if a new error number was added without updating the
1403 string.
1404
1405 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1406
1407 11. \K was not working if it appeared in an atomic group or in a group that
1408 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1409 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1410 the assertion is positive, but not if it is negative.
1411
1412 12. Change 11 fortuitously reduced the size of the stack frame used in the
1413 "match()" function of pcre_exec.c by one pointer. Forthcoming
1414 implementation of support for (*MARK) will need an extra pointer on the
1415 stack; I have reserved it now, so that the stack frame size does not
1416 decrease.
1417
1418 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1419 item in branch that calls a recursion is a subroutine call - as in the
1420 second branch in the above example - was incorrectly given the compile-
1421 time error "recursive call could loop indefinitely" because pcre_compile()
1422 was not correctly checking the subroutine for matching a non-empty string.
1423
1424 14. The checks for overrunning compiling workspace could trigger after an
1425 overrun had occurred. This is a "should never occur" error, but it can be
1426 triggered by pathological patterns such as hundreds of nested parentheses.
1427 The checks now trigger 100 bytes before the end of the workspace.
1428
1429 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1430
1431
1432 Version 8.01 19-Jan-2010
1433 ------------------------
1434
1435 1. If a pattern contained a conditional subpattern with only one branch (in
1436 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1437 computed the wrong minimum data length (which is of course zero for such
1438 subpatterns). This could cause incorrect "no match" results.
1439
1440 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1441 the pattern is reset in the first branch, pcre_compile() failed with
1442 "internal error: code overflow at offset...". This happened only when
1443 the reset was to the original external option setting. (An optimization
1444 abstracts leading options settings into an external setting, which was the
1445 cause of this.)
1446
1447 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1448 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1449 assertion pattern did not match (meaning that the assertion was true), it
1450 was incorrectly treated as false if the SKIP had been reached during the
1451 matching. This also applied to assertions used as conditions.
1452
1453 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1454 assertion subpattern, including such a pattern used as a condition,
1455 unpredictable results occurred, instead of the error return
1456 PCRE_ERROR_DFA_UITEM.
1457
1458 5. The C++ GlobalReplace function was not working like Perl for the special
1459 situation when an empty string is matched. It now does the fancy magic
1460 stuff that is necessary.
1461
1462 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1463 removed. (These were left over from very, very early versions of PCRE.)
1464
1465 7. Some cosmetic changes to the code to make life easier when compiling it
1466 as part of something else:
1467
1468 (a) Change DEBUG to PCRE_DEBUG.
1469
1470 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1471 called "current" as "current_branch", to prevent a collision with the
1472 Linux macro when compiled as a kernel module.
1473
1474 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1475 prevent a collision with the Linux macro when compiled as a kernel
1476 module.
1477
1478 8. In pcre_compile() there are some checks for integer overflows that used to
1479 cast potentially large values to (double). This has been changed to that
1480 when building, a check for int64_t is made, and if it is found, it is used
1481 instead, thus avoiding the use of floating point arithmetic. (There is no
1482 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1483 double.
1484
1485 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1486 2005 (difference between two addresses compared to an unsigned value).
1487
1488 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1489 custom one, because of the following reported problem in Windows:
1490
1491 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1492 under Win32.
1493 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1494 therefore missing the function definition.
1495 - The compiler thus generates a "C" signature for the test function.
1496 - The linker fails to find the "C" function.
1497 - PCRE fails to configure if asked to do so against libbz2.
1498
1499 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1500 messages were output:
1501
1502 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1503 rerunning libtoolize, to keep the correct libtool macros in-tree.
1504 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1505
1506 I have done both of these things.
1507
1508 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1509 most of the time, it *can* run out if it is given a pattern that contains a
1510 runaway infinite recursion. I updated the discussion in the pcrestack man
1511 page.
1512
1513 13. Now that we have gone to the x.xx style of version numbers, the minor
1514 version may start with zero. Using 08 or 09 is a bad idea because users
1515 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1516 interpreted as invalid octal numbers. I've updated the previous comment in
1517 configure.ac, and also added a check that gives an error if 08 or 09 are
1518 used.
1519
1520 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1521 causing partial matching to fail when the end of the subject matched \W
1522 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1523
1524 15. There were some discrepancies between the declarations in pcre_internal.h
1525 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1526 their definitions. The declarations used "const uschar *" and the
1527 definitions used USPTR. Even though USPTR is normally defined as "const
1528 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1529 reported that: "This difference in casting confuses some C++ compilers, for
1530 example, SunCC recognizes above declarations as different functions and
1531 generates broken code for hbpcre." I have changed the declarations to use
1532 USPTR.
1533
1534 16. GNU libtool is named differently on some systems. The autogen.sh script now
1535 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1536 (FreeBSD).
1537
1538 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1539 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1540 comment: "Figure out how to create a longlong from a string: strtoll and
1541 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1542 instance, but it only takes 2 args instead of 3!"
1543
1544 18. A subtle bug concerned with back references has been fixed by a change of
1545 specification, with a corresponding code fix. A pattern such as
1546 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1547 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1548 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1549 same bug. Such groups have to be quantified to be useful, or contained
1550 inside another quantified group. (If there's no repetition, the reference
1551 can never match.) The problem arises because, having left the group and
1552 moved on to the rest of the pattern, a later failure that backtracks into
1553 the group uses the captured value from the final iteration of the group
1554 rather than the correct earlier one. I have fixed this in PCRE by forcing
1555 any group that contains a reference to itself to be an atomic group; that
1556 is, there cannot be any backtracking into it once it has completed. This is
1557 similar to recursive and subroutine calls.
1558
1559
1560 Version 8.00 19-Oct-09
1561 ----------------------
1562
1563 1. The table for translating pcre_compile() error codes into POSIX error codes
1564 was out-of-date, and there was no check on the pcre_compile() error code
1565 being within the table. This could lead to an OK return being given in
1566 error.
1567
1568 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1569 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1570 in a Windows environment.
1571
1572 3. The pcregrep --count option prints the count for each file even when it is
1573 zero, as does GNU grep. However, pcregrep was also printing all files when
1574 --files-with-matches was added. Now, when both options are given, it prints
1575 counts only for those files that have at least one match. (GNU grep just
1576 prints the file name in this circumstance, but including the count seems
1577 more useful - otherwise, why use --count?) Also ensured that the
1578 combination -clh just lists non-zero counts, with no names.
1579
1580 4. The long form of the pcregrep -F option was incorrectly implemented as
1581 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1582 but it seems right to fix it, and I didn't think it was worth preserving
1583 the old behaviour.
1584
1585 5. The command line items --regex=pattern and --regexp=pattern were not
1586 recognized by pcregrep, which required --regex pattern or --regexp pattern
1587 (with a space rather than an '='). The man page documented the '=' forms,
1588 which are compatible with GNU grep; these now work.
1589
1590 6. No libpcreposix.pc file was created for pkg-config; there was just
1591 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1592
1593 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1594 when UCP support is not needed, by modifying the Python script that
1595 generates it from Unicode data files. This should not matter if the module
1596 is correctly used as a library, but I received one complaint about 50K of
1597 unwanted data. My guess is that the person linked everything into his
1598 program rather than using a library. Anyway, it does no harm.
1599
1600 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1601 was a minimum greater than 1 for a wide character in a possessive
1602 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1603 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1604 character. Chaos in the form of incorrect output or a compiling loop could
1605 result.
1606
1607 9. The restrictions on what a pattern can contain when partial matching is
1608 requested for pcre_exec() have been removed. All patterns can now be
1609 partially matched by this function. In addition, if there are at least two
1610 slots in the offset vector, the offset of the earliest inspected character
1611 for the match and the offset of the end of the subject are set in them when
1612 PCRE_ERROR_PARTIAL is returned.
1613
1614 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1615 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1616 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1617 and may be more useful for multi-segment matching.
1618
1619 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1620 used to be given if ever the end of the subject was reached; now it is
1621 given only if matching could not proceed because another character was
1622 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1623 string "Z", which now yields "no match" instead of "partial match". In the
1624 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1625 final character ended with (*FAIL).
1626
1627 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1628 if the pattern had a "must contain" character that was already found in the
1629 earlier partial match, unless partial matching was again requested. For
1630 example, with the pattern /dog.(body)?/, the "must contain" character is
1631 "g". If the first part-match was for the string "dog", restarting with
1632 "sbody" failed. This bug has been fixed.
1633
1634 13. The string returned by pcre_dfa_exec() after a partial match has been
1635 changed so that it starts at the first inspected character rather than the
1636 first character of the match. This makes a difference only if the pattern
1637 starts with a lookbehind assertion or \b or \B (\K is not supported by
1638 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1639 matching functions compatible, and I think it's the right thing to do.
1640
1641 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1642 so that the demonstration program is easily available in environments where
1643 PCRE has not been installed from source.
1644
1645 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1646 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1647 library.
1648
1649 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1650 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1651 is not the first non-POSIX option to be added. Clearly some people find
1652 these options useful.
1653
1654 17. If a caller to the POSIX matching function regexec() passes a non-zero
1655 value for nmatch with a NULL value for pmatch, the value of
1656 nmatch is forced to zero.
1657
1658 18. RunGrepTest did not have a test for the availability of the -u option of
1659 the diff command, as RunTest does. It now checks in the same way as
1660 RunTest, and also checks for the -b option.
1661
1662 19. If an odd number of negated classes containing just a single character
1663 interposed, within parentheses, between a forward reference to a named
1664 subpattern and the definition of the subpattern, compilation crashed with
1665 an internal error, complaining that it could not find the referenced
1666 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1667 [The bug was that it was starting one character too far in when skipping
1668 over the character class, thus treating the ] as data rather than
1669 terminating the class. This meant it could skip too much.]
1670
1671 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1672 /g option in pcretest when the pattern contains \K, which makes it possible
1673 to have an empty string match not at the start, even when the pattern is
1674 anchored. Updated pcretest and pcredemo to use this option.
1675
1676 21. If the maximum number of capturing subpatterns in a recursion was greater
1677 than the maximum at the outer level, the higher number was returned, but
1678 with unset values at the outer level. The correct (outer level) value is
1679 now given.
1680
1681 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1682 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1683 make it do so. The string so far is captured, making this feature
1684 compatible with Perl.
1685
1686 23. The tests have been re-organized, adding tests 11 and 12, to make it
1687 possible to check the Perl 5.10 features against Perl 5.10.
1688
1689 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1690 pattern matches a fixed length string. PCRE did not allow this; now it
1691 does. Neither allows recursion.
1692
1693 25. I finally figured out how to implement a request to provide the minimum
1694 length of subject string that was needed in order to match a given pattern.
1695 (It was back references and recursion that I had previously got hung up
1696 on.) This code has now been added to pcre_study(); it finds a lower bound
1697 to the length of subject needed. It is not necessarily the greatest lower
1698 bound, but using it to avoid searching strings that are too short does give
1699 some useful speed-ups. The value is available to calling programs via
1700 pcre_fullinfo().
1701
1702 26. While implementing 25, I discovered to my embarrassment that pcretest had
1703 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1704 study optimizations had never been tested with that matching function.
1705 Oops. What is worse, even when it was passed study data, there was a bug in
1706 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1707 were also very few tests of studied patterns with pcre_dfa_exec().
1708
1709 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1710 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1711 on the other side of the coin, they are no longer allowed to have different
1712 names, because these cannot be distinguished in PCRE, and this has caused
1713 confusion. (This is a difference from Perl.)
1714
1715 28. When duplicate subpattern names are present (necessarily with different
1716 numbers, as required by 27 above), and a test is made by name in a
1717 conditional pattern, either for a subpattern having been matched, or for
1718 recursion in such a pattern, all the associated numbered subpatterns are
1719 tested, and the overall condition is true if the condition is true for any
1720 one of them. This is the way Perl works, and is also more like the way
1721 testing by number works.
1722
1723
1724 Version 7.9 11-Apr-09
1725 ---------------------
1726
1727 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1728 (pcretest), all targets were linked against these libraries. This included
1729 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1730 libraries. This caused unwanted dependencies to be created. This problem
1731 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1732 pcretest is linked with readline.
1733
1734 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1735 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1736 moved outside it again, because FALSE and TRUE are already defined in AIX,
1737 but BOOL is not.
1738
1739 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1740 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1741
1742 4. The pcregrep documentation said spaces were inserted as well as colons (or
1743 hyphens) following file names and line numbers when outputting matching
1744 lines. This is not true; no spaces are inserted. I have also clarified the
1745 wording for the --colour (or --color) option.
1746
1747 5. In pcregrep, when --colour was used with -o, the list of matching strings
1748 was not coloured; this is different to GNU grep, so I have changed it to be
1749 the same.
1750
1751 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1752 each matching line was coloured. Now it goes on to look for further matches
1753 of any of the test patterns, which is the same behaviour as GNU grep.
1754
1755 7. A pattern that could match an empty string could cause pcregrep to loop; it
1756 doesn't make sense to accept an empty string match in pcregrep, so I have
1757 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1758 seems to be how GNU grep behaves. [But see later change 40 for release
1759 8.33.]
1760
1761 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1762 start or after a newline", because the conditional assertion was not being
1763 correctly handled. The rule now is that both the assertion and what follows
1764 in the first alternative must satisfy the test.
1765
1766 9. If auto-callout was enabled in a pattern with a conditional group whose
1767 condition was an assertion, PCRE could crash during matching, both with
1768 pcre_exec() and pcre_dfa_exec().
1769
1770 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1771 used for matching.
1772
1773 11. Unicode property support in character classes was not working for
1774 characters (bytes) greater than 127 when not in UTF-8 mode.
1775
1776 12. Added the -M command line option to pcretest.
1777
1778 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1779
1780 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1781
1782 16. Added comments and documentation about mis-use of no_arg in the C++
1783 wrapper.
1784
1785 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1786 from Martin Jerabek that uses macro names for all relevant character and
1787 string constants.
1788
1789 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1790 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1791 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1792 these, but not everybody uses configure.
1793
1794 19. A conditional group that had only one branch was not being correctly
1795 recognized as an item that could match an empty string. This meant that an
1796 enclosing group might also not be so recognized, causing infinite looping
1797 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1798 with the subject "ab", where knowledge that the repeated group can match
1799 nothing is needed in order to break the loop.
1800
1801 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1802 exec(), but without supplying a callout function, matching went wrong.
1803
1804 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1805 leak if the size of the offset vector was greater than 30. When the vector
1806 is smaller, the saved offsets during recursion go onto a local stack
1807 vector, but for larger vectors malloc() is used. It was failing to free
1808 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1809 error, in fact).
1810
1811 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1812 heapframe that is used only when UTF-8 support is enabled. This caused no
1813 problem, but was untidy.
1814
1815 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1816 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1817 included within another project.
1818
1819 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1820 slightly modified by me:
1821
1822 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1823 not building pcregrep.
1824
1825 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1826 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1827
1828 25. Forward references, both numeric and by name, in patterns that made use of
1829 duplicate group numbers, could behave incorrectly or give incorrect errors,
1830 because when scanning forward to find the reference group, PCRE was not
1831 taking into account the duplicate group numbers. A pattern such as
1832 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1833
1834 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1835 the feature of a custom pointer more persuasive (as requested by a user).
1836
1837 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1838 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1839 pre-defined.
1840
1841 28. Added support for (*UTF8) at the start of a pattern.
1842
1843 29. Arrange for flags added by the "release type" setting in CMake to be shown
1844 in the configuration summary.
1845
1846
1847 Version 7.8 05-Sep-08
1848 ---------------------
1849
1850 1. Replaced UCP searching code with optimized version as implemented for Ad
1851 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1852 stage table and inline lookup instead of a function, giving speed ups of 2
1853 to 5 times on some simple patterns that I tested. Permission was given to
1854 distribute the MultiStage2.py script that generates the tables (it's not in
1855 the tarball, but is in the Subversion repository).
1856
1857 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1858 scripts.
1859
1860 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1861 a group with a zero qualifier. The result of the study could be incorrect,
1862 or the function might crash, depending on the pattern.
1863
1864 4. Caseless matching was not working for non-ASCII characters in back
1865 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1866 It now works when Unicode Property Support is available.
1867
1868 5. In pcretest, an escape such as \x{de} in the data was always generating
1869 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1870 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1871 truncation.
1872
1873 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1874
1875 7. Added two (int) casts to pcregrep when printing the difference of two
1876 pointers, in case they are 64-bit values.
1877
1878 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1879 test 2 if it fails.
1880
1881 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1882 and a #define of that name to empty if it is not externally set. This is to
1883 allow users of MSVC to set it if necessary.
1884
1885 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1886 the convenience functions in the pcre_get.c source file.
1887
1888 11. An option change at the start of a pattern that had top-level alternatives
1889 could cause overwriting and/or a crash. This command provoked a crash in
1890 some environments:
1891
1892 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1893
1894 This potential security problem was recorded as CVE-2008-2371.
1895
1896 12. For a pattern where the match had to start at the beginning or immediately
1897 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1898 pcre_dfa_exec() could read past the end of the passed subject if there was
1899 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1900 pcretest so that it places the subject at the end of its malloc-ed buffer.
1901
1902 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1903 exec() might read past the end of the data buffer in UTF-8 mode.
1904
1905 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1906 the data contained the byte 0x85 as part of a UTF-8 character within its
1907 first line. This applied both to normal and DFA matching.
1908
1909 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1910 /^[^d]*?$/8 failed to match "abc".
1911
1912 16. Added a missing copyright notice to pcrecpp_internal.h.
1913
1914 17. Make it more clear in the documentation that values returned from
1915 pcre_exec() in ovector are byte offsets, not character counts.
1916
1917 18. Tidied a few places to stop certain compilers from issuing warnings.
1918
1919 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1920 supplied by Stefan Weber. I made a further small update for 7.8 because
1921 there is a change of source arrangements: the pcre_searchfuncs.c module is
1922 replaced by pcre_ucd.c.
1923
1924
1925 Version 7.7 07-May-08
1926 ---------------------
1927
1928 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1929 a string to a long long, pretend we don't even have a long long." This is
1930 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1931
1932 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1933 pre-7.6 versions, which defined a global no_arg variable instead of putting
1934 it in the RE class. (See also #8 below.)
1935
1936 3. Remove a line of dead code, identified by coverity and reported by Nuno
1937 Lopes.
1938
1939 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1940
1941 (1) The include/exclude patterns were being applied to the whole pathnames
1942 of files, instead of just to the final components.
1943
1944 (2) If there was more than one level of directory, the subdirectories were
1945 skipped unless they satisfied the include/exclude conditions. This is
1946 inconsistent with GNU grep (and could even be seen as contrary to the
1947 pcregrep specification - which I improved to make it absolutely clear).
1948 The action now is always to scan all levels of directory, and just
1949 apply the include/exclude patterns to regular files.
1950
1951 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1952 --exclude_dir in the tests to avoid scanning .svn directories.
1953
1954 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1955 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1956 doesn't support NULs in patterns.
1957
1958 7. Added some missing "const"s to declarations of static tables in
1959 pcre_compile.c and pcre_dfa_exec.c.
1960
1961 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1962 caused by fix #2 above. (Subsequently also a second patch to fix the
1963 first patch. And a third patch - this was a messy problem.)
1964
1965 9. Applied Craig's patch to remove the use of push_back().
1966
1967 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1968 matching function regexec().
1969
1970 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1971 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1972 references. PCRE supports relative numbers with this syntax (I don't think
1973 Oniguruma does).
1974
1975 12. Previously, a group with a zero repeat such as (...){0} was completely
1976 omitted from the compiled regex. However, this means that if the group
1977 was called as a subroutine from elsewhere in the pattern, things went wrong
1978 (an internal error was given). Such groups are now left in the compiled
1979 pattern, with a new opcode that causes them to be skipped at execution
1980 time.
1981
1982 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1983 to the way PCRE behaves:
1984
1985 (a) A lone ] character is dis-allowed (Perl treats it as data).
1986
1987 (b) A back reference to an unmatched subpattern matches an empty string
1988 (Perl fails the current match path).
1989
1990 (c) A data ] in a character class must be notated as \] because if the
1991 first data character in a class is ], it defines an empty class. (In
1992 Perl it is not possible to have an empty class.) The empty class []
1993 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1994 The negative empty class [^] matches any one character, independently
1995 of the DOTALL setting.
1996
1997 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1998 non-existent subpattern following a character class starting with ']' and
1999 containing () gave an internal compiling error instead of "reference to
2000 non-existent subpattern". Fortunately, when the pattern did exist, the
2001 compiled code was correct. (When scanning forwards to check for the
2002 existence of the subpattern, it was treating the data ']' as terminating
2003 the class, so got the count wrong. When actually compiling, the reference
2004 was subsequently set up correctly.)
2005
2006 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2007 it was being rejected as not supported by pcre_dfa_exec(), even though
2008 other assertions are supported. I have made pcre_dfa_exec() support
2009 (*FAIL).
2010
2011 16. The implementation of 13c above involved the invention of a new opcode,
2012 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2013 cannot be changed at match time, I realized I could make a small
2014 improvement to matching performance by compiling OP_ALLANY instead of
2015 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2016 on the OP_ANY path.
2017
2018 17. Compiling pcretest on Windows with readline support failed without the
2019 following two fixes: (1) Make the unistd.h include conditional on
2020 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2021
2022 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2023 ncurses library to be included for pcretest when ReadLine support is
2024 requested, but also to allow for it to be overridden. This patch came from
2025 Daniel Bergström.
2026
2027 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2028 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2029 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2030 spotting this.
2031
2032
2033 Version 7.6 28-Jan-08
2034 ---------------------
2035
2036 1. A character class containing a very large number of characters with
2037 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2038 overflow.
2039
2040 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2041 HAVE_LONG_LONG is not defined.
2042
2043 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2044 bring them up to date and include new features. This patch includes:
2045
2046 - Fixed PH's badly added libz and libbz2 support.
2047 - Fixed a problem with static linking.
2048 - Added pcredemo. [But later removed - see 7 below.]
2049 - Fixed dftables problem and added an option.
2050 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2051 HAVE_LONG_LONG.
2052 - Added readline support for pcretest.
2053 - Added an listing of the option settings after cmake has run.
2054
2055 4. A user submitted a patch to Makefile that makes it easy to create
2056 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2057 Makefile.am that cause it to include this special target, without
2058 affecting anything else. Note that the same mingw target plus all
2059 the other distribution libraries and programs are now supported
2060 when configuring with CMake (see 6 below) instead of with
2061 Configure/Make.
2062
2063 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2064 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2065 exported in the Windows port". It has not yet been confirmed that the patch
2066 solves the problem, but it does no harm.
2067
2068 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2069 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2070 with CMake, and also correct the comment about stack recursion.
2071
2072 7. Remove the automatic building of pcredemo from the ./configure system and
2073 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2074 of a program that users should build themselves after PCRE is installed, so
2075 building it automatically is not really right. What is more, it gave
2076 trouble in some build environments.
2077
2078 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2079
2080
2081 Version 7.5 10-Jan-08
2082 ---------------------
2083
2084 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2085 values in parens when parsing an RE using the C++ wrapper."
2086
2087 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2088 Characters greater than 255 were excluded from the class instead of being
2089 included.
2090
2091 3. The same bug as (2) above applied to negated POSIX classes such as
2092 [:^space:].
2093
2094 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2095 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2096 I have changed it.
2097
2098 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2099 first named subpattern) and a construct such as (?&a) would reference the
2100 first named subpattern whose name started with "a" (in other words, the
2101 length check was missing). Both these problems are fixed. "Subpattern name
2102 expected" is now given for (?&) (a zero-length name), and this patch also
2103 makes it give the same error for \k'' (previously it complained that that
2104 was a reference to a non-existent subpattern).
2105
2106 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2107 this is right because (?- can be followed by option settings as well as by
2108 digits. I have, however, made the messages clearer.
2109
2110 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2111 than the number used in the conditional) now cause a compile-time error.
2112 This is actually not compatible with Perl, which accepts such patterns, but
2113 treats the conditional as always being FALSE (as PCRE used to), but it
2114 seems to me that giving a diagnostic is better.
2115
2116 8. Change "alphameric" to the more common word "alphanumeric" in comments
2117 and messages.
2118
2119 9. Fix two occurrences of "backslash" in comments that should have been
2120 "backspace".
2121
2122 10. Remove two redundant lines of code that can never be obeyed (their function
2123 was moved elsewhere).
2124
2125 11. The program that makes PCRE's Unicode character property table had a bug
2126 which caused it to generate incorrect table entries for sequences of
2127 characters that have the same character type, but are in different scripts.
2128 It amalgamated them into a single range, with the script of the first of
2129 them. In other words, some characters were in the wrong script. There were
2130 thirteen such cases, affecting characters in the following ranges:
2131
2132 U+002b0 - U+002c1
2133 U+0060c - U+0060d
2134 U+0061e - U+00612
2135 U+0064b - U+0065e
2136 U+0074d - U+0076d
2137 U+01800 - U+01805
2138 U+01d00 - U+01d77
2139 U+01d9b - U+01dbf
2140 U+0200b - U+0200f
2141 U+030fc - U+030fe
2142 U+03260 - U+0327f
2143 U+0fb46 - U+0fbb1
2144 U+10450 - U+1049d
2145
2146 12. The -o option (show only the matching part of a line) for pcregrep was not
2147 compatible with GNU grep in that, if there was more than one match in a
2148 line, it showed only the first of them. It now behaves in the same way as
2149 GNU grep.
2150
2151 13. If the -o and -v options were combined for pcregrep, it printed a blank
2152 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2153 does the same. The return code can be used to tell if there were any
2154 non-matching lines.
2155
2156 14. Added --file-offsets and --line-offsets to pcregrep.
2157
2158 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2159 infinitely looping recursion. The bug was that positive lookaheads were not
2160 being skipped when checking for a possible empty match (negative lookaheads
2161 and both kinds of lookbehind were skipped).
2162
2163 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2164 inclusion of <windows.h> to before rather than after the definition of
2165 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2166
2167 17. Specifying a possessive quantifier with a specific limit for a Unicode
2168 character property caused pcre_compile() to compile bad code, which led at
2169 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2170 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2171 caused the error; without that there was no problem.
2172
2173 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2174
2175 19. Added --enable-pcretest-libreadline.
2176
2177 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2178 RE::GlobalReplace(). As a result, the number of replacements returned was
2179 double what it should be. I removed one of the increments, but Craig sent a
2180 later patch that removed the other one (the right fix) and added unit tests
2181 that check the return values (which was not done before).
2182
2183 21. Several CMake things:
2184
2185 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2186 the names libpcre and libpcreposix, not just pcre and pcreposix.
2187
2188 (2) The above change means that pcretest and pcregrep are now correctly
2189 linked with the newly-built libraries, not previously installed ones.
2190
2191 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2192
2193 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2194 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2195 UTF-8 newline character). The key issue is that the pattern starts .*;
2196 this means that the match must be either at the beginning, or after a
2197 newline. The bug was in the code for advancing after a failed match and
2198 checking that the new position followed a newline. It was not taking
2199 account of UTF-8 characters correctly.
2200
2201 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2202 character classes. PCRE was not treating the sequence [:...:] as a
2203 character class unless the ... were all letters. Perl, however, seems to
2204 allow any characters between [: and :], though of course it rejects as
2205 unknown any "names" that contain non-letters, because all the known class
2206 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2207 for example, whereas PCRE did not - it did not recognize a POSIX character
2208 class. This seemed a bit dangerous, so the code has been changed to be
2209 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2210 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2211 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2212 Perl does, and where it didn't before.
2213
2214 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2215 Windows environments %n is disabled by default.
2216
2217
2218 Version 7.4 21-Sep-07
2219 ---------------------
2220
2221 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2222 means that a class such as [\s] counted as "explicit reference to CR or
2223 LF". That isn't really right - the whole point of the change was to try to
2224 help when there was an actual mention of one of the two characters. So now
2225 the change happens only if \r or \n (or a literal CR or LF) character is
2226 encountered.
2227
2228 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2229 of both had grown to the point where there were only 3 bits left.
2230 Fortunately, there was spare space in the data structure, and so I have
2231 moved the internal flags into a new 16-bit field to free up more option
2232 bits.
2233
2234 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2235 but did not set the internal JCHANGED flag - either of these is enough to
2236 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2237 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2238 start sets both bits.
2239
2240 4. Added options (at build time, compile time, exec time) to change \R from
2241 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2242
2243 5. doc/pcresyntax.html was missing from the distribution.
2244
2245 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2246 compatibility, even though it is no longer used.
2247
2248 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2249 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2250 windows.h file is present (where different names are used). [This was
2251 reversed later after testing - see 16 below.]
2252
2253 8. Changed all #include <config.h> to #include "config.h". There were also
2254 some further <pcre.h> cases that I changed to "pcre.h".
2255
2256 9. When pcregrep was used with the --colour option, it missed the line ending
2257 sequence off the lines that it output.
2258
2259 10. It was pointed out to me that arrays of string pointers cause lots of
2260 relocations when a shared library is dynamically loaded. A technique of
2261 using a single long string with a table of offsets can drastically reduce
2262 these. I have refactored PCRE in four places to do this. The result is
2263 dramatic:
2264
2265 Originally: 290
2266 After changing UCP table: 187
2267 After changing error message table: 43
2268 After changing table of "verbs" 36
2269 After changing table of Posix names 22
2270
2271 Thanks to the folks working on Gregex for glib for this insight.
2272
2273 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2274 unicode-properties was also set.
2275
2276 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2277
2278 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2279 checked only for CRLF.
2280
2281 14. Added casts to pcretest.c to avoid compiler warnings.
2282
2283 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2284
2285 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2286 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2287 entirely. This removes changes made in 7 above.
2288
2289 17. The CMake files have been updated, and there is now more information about
2290 building with CMake in the NON-UNIX-USE document.
2291
2292
2293 Version 7.3 28-Aug-07
2294 ---------------------
2295
2296 1. In the rejigging of the build system that eventually resulted in 7.1, the
2297 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2298 brackets there is not right, since it causes compilers to look for an
2299 installed pcre.h, not the version that is in the source that is being
2300 compiled (which of course may be different). I have changed it back to:
2301
2302 #include "pcre.h"
2303
2304 I have a vague recollection that the change was concerned with compiling in
2305 different directories, but in the new build system, that is taken care of
2306 by the VPATH setting the Makefile.
2307
2308 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2309 when the subject happened to end in the byte 0x85 (e.g. if the last
2310 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2311 characters but of course it shouldn't be taken as a newline when it is part
2312 of another character. The bug was that, for an unlimited repeat of . in
2313 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2314 characters when looking for a newline.
2315
2316 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2317
2318 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2319 in debug output.
2320
2321 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2322 long printing in the pcrecpp unittest when running under MinGW.
2323
2324 6. ESC_K was left out of the EBCDIC table.
2325
2326 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2327 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2328 limit also applies to "virtual nesting" when a pattern is recursive, and in
2329 this case 1000 isn't so big. I have been able to remove this limit at the
2330 expense of backing off one optimization in certain circumstances. Normally,
2331 when pcre_exec() would call its internal match() function recursively and
2332 immediately return the result unconditionally, it uses a "tail recursion"
2333 feature to save stack. However, when a subpattern that can match an empty
2334 string has an unlimited repetition quantifier, it no longer makes this
2335 optimization. That gives it a stack frame in which to save the data for
2336 checking that an empty string has been matched. Previously this was taken
2337 from the 1000-entry workspace that had been reserved. So now there is no
2338 explicit limit, but more stack is used.
2339
2340 8. Applied Daniel's patches to solve problems with the import/export magic
2341 syntax that is required for Windows, and which was going wrong for the
2342 pcreposix and pcrecpp parts of the library. These were overlooked when this
2343 problem was solved for the main library.
2344
2345 9. There were some crude static tests to avoid integer overflow when computing
2346 the size of patterns that contain repeated groups with explicit upper
2347 limits. As the maximum quantifier is 65535, the maximum group length was
2348 set at 30,000 so that the product of these two numbers did not overflow a
2349 32-bit integer. However, it turns out that people want to use groups that
2350 are longer than 30,000 bytes (though not repeat them that many times).
2351 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2352 made it possible to implement the integer overflow checks in a much more
2353 dynamic way, which I have now done. The artificial limitation on group
2354 length has been removed - we now have only the limit on the total length of
2355 the compiled pattern, which depends on the LINK_SIZE setting.
2356
2357 10. Fixed a bug in the documentation for get/copy named substring when
2358 duplicate names are permitted. If none of the named substrings are set, the
2359 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2360 empty string.
2361
2362 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2363 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2364 because the ] is interpreted as the first data character and the
2365 terminating ] is not found. PCRE has been made compatible with Perl in this
2366 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2367 cause memory overwriting.
2368
2369 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2370 string has been matched (to stop an infinite loop). It was not recognizing
2371 a conditional subpattern that could match an empty string if that
2372 subpattern was within another subpattern. For example, it looped when
2373 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2374 condition was not nested. This bug has been fixed.
2375
2376 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2377 past the start of the subject in the presence of bytes with the top bit
2378 set, for example "\x8aBCD".
2379
2380 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2381 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2382
2383 14. Optimized (?!) to (*FAIL).
2384
2385 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2386 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2387 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2388 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2389 does: it's just the validity check that is more restrictive.
2390
2391 16. Inserted checks for integer overflows during escape sequence (backslash)
2392 processing, and also fixed erroneous offset values for syntax errors during
2393 backslash processing.
2394
2395 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2396 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2397
2398 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2399 caused an overrun.
2400
2401 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2402 something other than just ASCII characters) inside a group that had an
2403 unlimited repeat caused a loop at compile time (while checking to see
2404 whether the group could match an empty string).
2405
2406 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2407 [\P{Any}] did so. (Error in the code for printing property names.)
2408
2409 21. An orphan \E inside a character class could cause a crash.
2410
2411 22. A repeated capturing bracket such as (A)? could cause a wild memory
2412 reference during compilation.
2413
2414 23. There are several functions in pcre_compile() that scan along a compiled
2415 expression for various reasons (e.g. to see if it's fixed length for look
2416 behind). There were bugs in these functions when a repeated \p or \P was
2417 present in the pattern. These operators have additional parameters compared
2418 with \d, etc, and these were not being taken into account when moving along
2419 the compiled data. Specifically:
2420
2421 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2422 length.
2423
2424 (b) An item such as \pL+ within a repeated group could cause crashes or
2425 loops.
2426
2427 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2428 "reference to non-existent subpattern" error.
2429
2430 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2431
2432 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2433 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2434
2435 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2436
2437 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2438 character were causing crashes (broken optimization).
2439
2440 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2441 \p or \P) caused a compile-time loop.
2442
2443 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2444 break. For example, the unstudied pattern [\r\n]A does not match the string
2445 "\r\nA" because change 7.0/46 below moves the current point on by two
2446 characters after failing to match at the start. However, the pattern \nA
2447 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2448 the same is true. There doesn't seem any very clean way out of this, but
2449 what I have chosen to do makes the common cases work: PCRE now takes note
2450 of whether there can be an explicit match for \r or \n anywhere in the
2451 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2452 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2453 pattern has explicit CR or LF references.
2454
2455 29. Added (*CR) etc for changing newline setting at start of pattern.
2456
2457
2458 Version 7.2 19-Jun-07
2459 ---------------------
2460
2461 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2462 which is apparently normally available under Windows.
2463
2464 2. Re-jig the pcregrep tests with different newline settings in an attempt
2465 to make them independent of the local environment's newline setting.
2466
2467 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2468
2469 4. Some of the "internals" tests were previously cut out when the link size
2470 was not 2, because the output contained actual offsets. The recent new
2471 "Z" feature of pcretest means that these can be cut out, making the tests
2472 usable with all link sizes.
2473
2474 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2475 stack recursion. This gives a massive performance boost under BSD, but just
2476 a small improvement under Linux. However, it saves one field in the frame
2477 in all cases.
2478
2479 6. Added more features from the forthcoming Perl 5.10:
2480
2481 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2482 recursion call. It refers to the nth most recently opened parentheses.
2483
2484 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2485 to be opened parentheses.
2486
2487 (c) Conditions that refer to capturing parentheses can be specified
2488 relatively, for example, (?(-2)... or (?(+3)...
2489
2490 (d) \K resets the start of the current match so that everything before
2491 is not part of it.
2492
2493 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2494
2495 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2496 reference syntax.
2497
2498 (g) (?| introduces a group in which the numbering of parentheses in each
2499 alternative starts with the same number.
2500
2501 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2502
2503 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2504 PCRE_INFO_JCHANGED.
2505
2506 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2507 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2508 for detecting groups that can match an empty string.
2509
2510 9. A pattern with a very large number of alternatives (more than several
2511 hundred) was running out of internal workspace during the pre-compile
2512 phase, where pcre_compile() figures out how much memory will be needed. A
2513 bit of new cunning has reduced the workspace needed for groups with
2514 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2515 workspace instead of running out of the 4096 that are available.
2516
2517 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2518
2519 11. Applied patch from Google to remove an optimization that didn't quite work.
2520 The report of the bug said:
2521
2522 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2523 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2524 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2525
2526 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2527 it matched the wrong number of bytes.
2528
2529
2530 Version 7.1 24-Apr-07
2531 ---------------------
2532
2533 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2534 that is more "standard", making use of automake and other Autotools. There
2535 is some re-arrangement of the files and adjustment of comments consequent
2536 on this.
2537
2538 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2539 for recursive directory scanning broke on some systems because the files
2540 are not scanned in any specific order and on different systems the order
2541 was different. A call to "sort" has been inserted into RunGrepTest for the
2542 approprate test as a short-term fix. In the longer term there may be an
2543 alternative.
2544
2545 3. I had an email from Eric Raymond about problems translating some of PCRE's
2546 man pages to HTML (despite the fact that I distribute HTML pages, some
2547 people do their own conversions for various reasons). The problems
2548 concerned the use of low-level troff macros .br and .in. I have therefore
2549 removed all such uses from the man pages (some were redundant, some could
2550 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2551 HTML has been updated to handle .nf/.fi and to complain if it encounters
2552 .br or .in.
2553
2554 4. Updated comments in configure.ac that get placed in config.h.in and also
2555 arranged for config.h to be included in the distribution, with the name
2556 config.h.generic, for the benefit of those who have to compile without
2557 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2558
2559 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2560 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2561 makevp.bat for the current PCRE, using the additional files
2562 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2563
2564 6. A Windows user reported a minor discrepancy with test 2, which turned out
2565 to be caused by a trailing space on an input line that had got lost in his
2566 copy. The trailing space was an accident, so I've just removed it.
2567
2568 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2569 that is needed.
2570
2571 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2572 as "const" (a) because they are and (b) because it helps the PHP
2573 maintainers who have recently made a script to detect big data structures
2574 in the php code that should be moved to the .rodata section. I remembered
2575 to update Builducptable as well, so it won't revert if ucptable.h is ever
2576 re-created.
2577
2578 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2579 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2580 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2581 support is not required. This saves 1.5-2K of code, which is important in
2582 some applications.
2583
2584 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2585 so as not to refer to the tables, even though these functions will never be
2586 called when UTF-8 support is disabled. Otherwise there are problems with a
2587 shared library.
2588
2589 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2590
2591 (a) It was defining its arguments as char * instead of void *.
2592
2593 (b) It was assuming that all moves were upwards in memory; this was true
2594 a long time ago when I wrote it, but is no longer the case.
2595
2596 The emulated memove() is provided for those environments that have neither
2597 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2598 is clearly not the case, as these two bugs were recently reported.
2599
2600 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2601 and Detrail to create the HTML documentation, the .txt form of the man
2602 pages, and it removes trailing spaces from listed files. It also creates
2603 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2604 case, it wraps all the #defines with #ifndefs. This script should be run
2605 before "make dist".
2606
2607 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2608 with Unicode property support.
2609
2610 (a) For a maximizing quantifier, if the two different cases of the
2611 character were of different lengths in their UTF-8 codings (there are
2612 some cases like this - I found 11), and the matching function had to
2613 back up over a mixture of the two cases, it incorrectly assumed they
2614 were both the same length.
2615
2616 (b) When PCRE was configured to use the heap rather than the stack for
2617 recursion during matching, it was not correctly preserving the data for
2618 the other case of a UTF-8 character when checking ahead for a match
2619 while processing a minimizing repeat. If the check also involved
2620 matching a wide character, but failed, corruption could cause an
2621 erroneous result when trying to check for a repeat of the original
2622 character.
2623
2624 13. Some tidying changes to the testing mechanism:
2625
2626 (a) The RunTest script now detects the internal link size and whether there
2627 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2628 values substituted by "configure". (The RunGrepTest script already did
2629 this for UTF-8.) The configure.ac script no longer substitutes the
2630 relevant variables.
2631
2632 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2633 with length and offset values. This means that the output is different
2634 for different internal link sizes. Test 2 is skipped for link sizes
2635 other than 2 because of this, bypassing the problem. Unfortunately,
2636 there was also a test in test 3 (the locale tests) that used /B and
2637 failed for link sizes other than 2. Rather than cut the whole test out,
2638 I have added a new /Z option to pcretest that replaces the length and
2639 offset values with spaces. This is now used to make test 3 independent
2640 of link size. (Test 2 will be tidied up later.)
2641
2642 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2643 segmentation fault instead of returning the appropriate error message.
2644
2645 15. In multiline mode when the newline sequence was set to "any", the pattern
2646 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2647 This doesn't seem right; it now treats the CRLF combination as the line
2648 ending, and so does not match in that case. It's only a pattern such as ^$
2649 that would hit this one: something like ^ABC$ would have failed after \r
2650 and then tried again after \r\n.
2651
2652 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2653 in an attempt to make files that differ only in their line terminators
2654 compare equal. This works on Linux.
2655
2656 17. Under certain error circumstances pcregrep might try to free random memory
2657 as it exited. This is now fixed, thanks to valgrind.
2658
2659 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2660 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2661 was because its rules for how to advance for /g after matching an empty
2662 string at the end of a line did not allow for this case. They now check for
2663 it specially.
2664
2665 20. pcretest is supposed to handle patterns and data of any length, by
2666 extending its buffers when necessary. It was getting this wrong when the
2667 buffer for a data line had to be extended.
2668
2669 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2670 CRLF as a newline sequence.
2671
2672 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2673 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2674 I have nevertheless tidied it up.
2675
2676 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2677
2678 24. Added a man page for pcre-config.
2679
2680
2681 Version 7.0 19-Dec-06
2682 ---------------------
2683
2684 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2685 moving to gcc 4.1.1.
2686
2687 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2688 sys/time.h, which is documented as needed for this function. It doesn't
2689 seem to matter on Linux, but it showed up on some releases of OS X.
2690
2691 3. It seems that there are systems where bytes whose values are greater than
2692 127 match isprint() in the "C" locale. The "C" locale should be the
2693 default when a C program starts up. In most systems, only ASCII printing
2694 characters match isprint(). This difference caused the output from pcretest
2695 to vary, making some of the tests fail. I have changed pcretest so that:
2696
2697 (a) When it is outputting text in the compiled version of a pattern, bytes
2698 other than 32-126 are always shown as hex escapes.
2699
2700 (b) When it is outputting text that is a matched part of a subject string,
2701 it does the same, unless a different locale has been set for the match
2702 (using the /L modifier). In this case, it uses isprint() to decide.
2703
2704 4. Fixed a major bug that caused incorrect computation of the amount of memory
2705 required for a compiled pattern when options that changed within the
2706 pattern affected the logic of the preliminary scan that determines the
2707 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2708 that the computed length was too small. The symptoms of this bug were
2709 either the PCRE error "internal error: code overflow" from pcre_compile(),
2710 or a glibc crash with a message such as "pcretest: free(): invalid next
2711 size (fast)". Examples of patterns that provoked this bug (shown in
2712 pcretest format) are:
2713
2714 /(?-x: )/x
2715 /(?x)(?-x: \s*#\s*)/
2716 /((?i)[\x{c0}])/8
2717 /(?i:[\x{c0}])/8
2718
2719 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2720 is now done differently.
2721
2722 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2723 wrapper classes; (b) implement a new function in the C++ scanner that is
2724 more efficient than the old way of doing things because it avoids levels of
2725 recursion in the regex matching; (c) add a paragraph to the documentation
2726 for the FullMatch() function.
2727
2728 6. The escape sequence \n was being treated as whatever was defined as
2729 "newline". Not only was this contrary to the documentation, which states
2730 that \n is character 10 (hex 0A), but it also went horribly wrong when
2731 "newline" was defined as CRLF. This has been fixed.
2732
2733 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2734 was being set to -1 for the "end of line" case (supposedly a value that no
2735 character can have). Though this value is never used (the check for end of
2736 line is "zero bytes in current character"), it caused compiler complaints.
2737 I've changed it to 0xffffffff.
2738
2739 8. In pcre_version.c, the version string was being built by a sequence of
2740 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2741 string (as it is for production releases) called a macro with an empty
2742 argument. The C standard says the result of this is undefined. The gcc
2743 compiler treats it as an empty string (which was what was wanted) but it is
2744 reported that Visual C gives an error. The source has been hacked around to
2745 avoid this problem.
2746
2747 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2748 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2749 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2750 of them did).
2751
2752 10. Originally, pcretest opened its input and output without "b"; then I was
2753 told that "b" was needed in some environments, so it was added for release
2754 5.0 to both the input and output. (It makes no difference on Unix-like
2755 systems.) Later I was told that it is wrong for the input on Windows. I've
2756 now abstracted the modes into two macros, to make it easier to fiddle with
2757 them, and removed "b" from the input mode under Windows.
2758
2759 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2760
2761 12. Added -help and --help to pcretest as an official way of being reminded
2762 of the options.
2763
2764 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2765 and pcrecpp.cc because they annoy compilers at high warning levels.
2766
2767 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2768
2769 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2770 scripts are not C programs :-) and which was not noticed because it works
2771 on Linux.
2772
2773 16. pcretest is supposed to handle any length of pattern and data line (as one
2774 line or as a continued sequence of lines) by extending its input buffer if
2775 necessary. This feature was broken for very long pattern lines, leading to
2776 a string of junk being passed to pcre_compile() if the pattern was longer
2777 than about 50K.
2778
2779 17. I have done a major re-factoring of the way pcre_compile() computes the
2780 amount of memory needed for a compiled pattern. Previously, there was code
2781 that made a preliminary scan of the pattern in order to do this. That was
2782 OK when PCRE was new, but as the facilities have expanded, it has become
2783 harder and harder to keep it in step with the real compile phase, and there
2784 have been a number of bugs (see for example, 4 above). I have now found a
2785 cunning way of running the real compile function in a "fake" mode that
2786 enables it to compute how much memory it would need, while actually only
2787 ever using a few hundred bytes of working memory and without too many
2788 tests of the mode. This should make future maintenance and development
2789 easier. A side effect of this work is that the limit of 200 on the nesting
2790 depth of parentheses has been removed (though this was never a serious
2791 limitation, I suspect). However, there is a downside: pcre_compile() now
2792 runs more slowly than before (30% or more, depending on the pattern). I
2793 hope this isn't a big issue. There is no effect on runtime performance.
2794
2795 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2796 newline (only possible for the last line of a file) and it was a
2797 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2798
2799 19. Added additional timing features to pcretest. (1) The -tm option now times
2800 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2801 separate command line item, by a number that specifies the number of
2802 repeats to use when timing. The default is 50000; this gives better
2803 precision, but takes uncomfortably long for very large patterns.
2804
2805 20. Extended pcre_study() to be more clever in cases where a branch of a
2806 subpattern has no definite first character. For example, (a*|b*)[cd] would
2807 previously give no result from pcre_study(). Now it recognizes that the
2808 first character must be a, b, c, or d.
2809
2810 21. There was an incorrect error "recursive call could loop indefinitely" if
2811 a subpattern (or the entire pattern) that was being tested for matching an
2812 empty string contained only one non-empty item after a nested subpattern.
2813 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2814 incorrectly, because the \d was being skipped in the check.
2815
2816 22. The pcretest program now has a new pattern option /B and a command line
2817 option -b, which is equivalent to adding /B to every pattern. This causes
2818 it to show the compiled bytecode, without the additional information that
2819 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2820 is the same as /B/I).
2821
2822 23. A new optimization is now able automatically to treat some sequences such
2823 as a*b as a*+b. More specifically, if something simple (such as a character
2824 or a simple class like \d) has an unlimited quantifier, and is followed by
2825 something that cannot possibly match the quantified thing, the quantifier
2826 is automatically "possessified".
2827
2828 24. A recursive reference to a subpattern whose number was greater than 39
2829 went wrong under certain circumstances in UTF-8 mode. This bug could also
2830 have affected the operation of pcre_study().
2831
2832 25. Realized that a little bit of performance could be had by replacing
2833 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2834
2835 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2836
2837 27. Possessive quantifiers such as a++ were previously implemented by turning
2838 them into atomic groups such as ($>a+). Now they have their own opcodes,
2839 which improves performance. This includes the automatically created ones
2840 from 23 above.
2841
2842 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2843 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2844 the first matched character to be a colon. This applied both to named and
2845 numbered groups.
2846
2847 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2848
2849 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2850 building PCRE on Windows easier, so I have included it in the distribution.
2851
2852 31. There is now a check in pcretest against a ridiculously large number being
2853 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2854 loop, the loop is abandoned.
2855
2856 32. Forward references to subpatterns in conditions such as (?(2)...) where
2857 subpattern 2 is defined later cause pcre_compile() to search forwards in
2858 the pattern for the relevant set of parentheses. This search went wrong
2859 when there were unescaped parentheses in a character class, parentheses
2860 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2861
2862 33. "Subroutine" calls and backreferences were previously restricted to
2863 referencing subpatterns earlier in the regex. This restriction has now
2864 been removed.
2865
2866 34. Added a number of extra features that are going to be in Perl 5.10. On the
2867 whole, these are just syntactic alternatives for features that PCRE had
2868 previously implemented using the Python syntax or my own invention. The
2869 other formats are all retained for compatibility.
2870
2871 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2872 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2873 also .NET compatible.
2874
2875 (b) A recursion or subroutine call to a named group can now be defined as
2876 (?&name) as well as (?P>name).
2877
2878 (c) A backreference to a named group can now be defined as \k<name> or
2879 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2880 5.10, are also .NET compatible.
2881
2882 (d) A conditional reference to a named group can now use the syntax
2883 (?(<name>) or (?('name') as well as (?(name).
2884
2885 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2886 groups (named and numbered) that are never evaluated inline, but can be
2887 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2888 is always false. There may be only one alternative in such a group.
2889
2890 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2891 as the simple (?(R). The condition is true only if the most recent
2892 recursion is that of the given number or name. It does not search out
2893 through the entire recursion stack.
2894
2895 (g) The escape \gN or \g{N} has been added, where N is a positive or
2896 negative number, specifying an absolute or relative reference.
2897
2898 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2899 some "unreachable code" warnings.
2900
2901 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2902 things, this adds five new scripts.
2903
2904 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2905 There were also incompatibilities regarding the handling of \Q..\E inside
2906 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2907 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2908
2909 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2910 matches an empty string, and forcibly breaks the loop. There were bugs in
2911 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2912 against aaaa the result was just "a" rather than "aaaa", for example. Two
2913 separate and independent bugs (that affected different cases) have been
2914 fixed.
2915
2916 39. Refactored the code to abolish the use of different opcodes for small
2917 capturing bracket numbers. This is a tidy that I avoided doing when I
2918 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2919 The new approach is not only tidier, it makes it possible to reduce the
2920 memory needed to fix the previous bug (38).
2921
2922 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2923 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2924 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2925 mode.
2926
2927 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2928 report.
2929
2930 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2931 copy construction and assignment in the C++ wrapper.
2932
2933 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2934 couple of bugs that could have given wrong results in the "--newline=crlf"
2935 case.
2936
2937 44. Added a number of casts and did some reorganization of signed/unsigned int
2938 variables following suggestions from Dair Grant. Also renamed the variable
2939 "this" as "item" because it is a C++ keyword.
2940
2941 45. Arranged for dftables to add
2942
2943 #include "pcre_internal.h"
2944
2945 to pcre_chartables.c because without it, gcc 4.x may remove the array
2946 definition from the final binary if PCRE is built into a static library and
2947 dead code stripping is activated.
2948
2949 46. For an unanchored pattern, if a match attempt fails at the start of a
2950 newline sequence, and the newline setting is CRLF or ANY, and the next two
2951 characters are CRLF, advance by two characters instead of one.
2952
2953
2954 Version 6.7 04-Jul-06
2955 ---------------------
2956
2957 1. In order to handle tests when input lines are enormously long, pcretest has
2958 been re-factored so that it automatically extends its buffers when
2959 necessary. The code is crude, but this _is_ just a test program. The
2960 default size has been increased from 32K to 50K.
2961
2962 2. The code in pcre_study() was using the value of the re argument before
2963 testing it for NULL. (Of course, in any sensible call of the function, it
2964 won't be NULL.)
2965
2966 3. The memmove() emulation function in pcre_internal.h, which is used on
2967 systems that lack both memmove() and bcopy() - that is, hardly ever -
2968 was missing a "static" storage class specifier.
2969
2970 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2971 containing an extended class (one that cannot be represented by a bitmap
2972 because it contains high-valued characters or Unicode property items, e.g.
2973 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2974 pattern, but PCRE should not loop if you do not (it no longer does).
2975 [Detail: two cases were found: (a) a repeated subpattern containing an
2976 extended class; (b) a recursive reference to a subpattern that followed a
2977 previous extended class. It wasn't skipping over the extended class
2978 correctly when UTF-8 mode was not set.]
2979
2980 5. A negated single-character class was not being recognized as fixed-length
2981 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2982 compile error "lookbehind assertion is not fixed length".
2983
2984 6. The RunPerlTest auxiliary script was showing an unexpected difference
2985 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2986 write a Perl script that can interpret lines of an input file either as
2987 byte characters or as UTF-8, which is what "perltest" was being required to
2988 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2989 can't do is switch easily at run time between having the "use utf8;" pragma
2990 or not. In the end, I fudged it by using the RunPerlTest script to insert
2991 "use utf8;" explicitly for the UTF-8 tests.
2992
2993 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2994 the end of the subject string, contrary to the documentation and to what
2995 Perl does. This was true of both matching functions. Now it matches only at
2996 the start of the subject and immediately after *internal* newlines.
2997
2998 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2999 a pointer to an int instead of a pointer to an unsigned long int. This
3000 caused problems on 64-bit systems.
3001
3002 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3003 instance of the 'standard' template library not being so standard".
3004
3005 10. There was no check on the number of named subpatterns nor the maximum
3006 length of a subpattern name. The product of these values is used to compute
3007 the size of the memory block for a compiled pattern. By supplying a very
3008 long subpattern name and a large number of named subpatterns, the size
3009 computation could be caused to overflow. This is now prevented by limiting
3010 the length of names to 32 characters, and the number of named subpatterns
3011 to 10,000.
3012
3013 11. Subpatterns that are repeated with specific counts have to be replicated in
3014 the compiled pattern. The size of memory for this was computed from the
3015 length of the subpattern and the repeat count. The latter is limited to
3016 65535, but there was no limit on the former, meaning that integer overflow
3017 could in principle occur. The compiled length of a repeated subpattern is
3018 now limited to 30,000 bytes in order to prevent this.
3019
3020 12. Added the optional facility to have named substrings with the same name.
3021
3022 13. Added the ability to use a named substring as a condition, using the
3023 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3024 are numbers (not recommended). Forward references are permitted.
3025
3026 14. Added forward references in named backreferences (if you see what I mean).
3027
3028 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3029 pattern could run off the end of the subject. For example, the pattern
3030 "(?s)(.{1,5})"8 did this with the subject "ab".
3031
3032 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3033 PCRE_CASELESS was set when matching characters that were quantified with ?
3034 or *.
3035
3036 17. A character class other than a single negated character that had a minimum
3037 but no maximum quantifier - for example [ab]{6,} - was not handled
3038 correctly by pce_dfa_exec(). It would match only one character.
3039
3040 18. A valid (though odd) pattern that looked like a POSIX character
3041 class but used an invalid character after [ (for example [[,abc,]]) caused
3042 pcre_compile() to give the error "Failed: internal error: code overflow" or
3043 in some cases to crash with a glibc free() error. This could even happen if
3044 the pattern terminated after [[ but there just happened to be a sequence of
3045 letters, a binary zero, and a closing ] in the memory that followed.
3046
3047 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3048 over the years. Originally (before any Unicode support), just the bottom 8
3049 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3050 output from "man perlunicode" includes this:
3051
3052 The regular expression compiler produces polymorphic opcodes. That
3053 is, the pattern adapts to the data and automatically switches to
3054 the Unicode character scheme when presented with Unicode data--or
3055 instead uses a traditional byte scheme when presented with byte
3056 data.
3057
3058 Sadly, a wide octal escape does not cause a switch, and in a string with
3059 no other multibyte characters, these octal escapes are treated as before.
3060 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3061 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3062 Unicode string.
3063
3064 I have not perpetrated such confusion in PCRE. Up till now, it took just
3065 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3066 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3067 translate to the appropriate multibyte character.
3068
3069 29. Applied some refactoring to reduce the number of warnings from Microsoft
3070 and Borland compilers. This has included removing the fudge introduced
3071 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3072 a warning about an unused variable.
3073
3074 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3075 characters since release 4.0, because Perl (from release 5.004) does not.
3076 [Or at least, is documented not to: some releases seem to be in conflict
3077 with the documentation.] However, when a pattern was studied with
3078 pcre_study() and all its branches started with \s, PCRE still included VT
3079 as a possible starting character. Of course, this did no harm; it just
3080 caused an unnecessary match attempt.
3081
3082 22. Removed a now-redundant internal flag bit that recorded the fact that case
3083 dependency changed within the pattern. This was once needed for "required
3084 byte" processing, but is no longer used. This recovers a now-scarce options
3085 bit. Also moved the least significant internal flag bit to the most-
3086 significant bit of the word, which was not previously used (hangover from
3087 the days when it was an int rather than a uint) to free up another bit for
3088 the future.
3089
3090 23. Added support for CRLF line endings as well as CR and LF. As well as the
3091 default being selectable at build time, it can now be changed at runtime
3092 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3093 specify that it is scanning data with non-default line endings.
3094
3095 24. Changed the definition of CXXLINK to make it agree with the definition of
3096 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3097
3098 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3099 recursions. This makes a big different to stack usage for some patterns.
3100
3101 26. If a subpattern containing a named recursion or subroutine reference such
3102 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3103 the space required for the compiled pattern went wrong and gave too small a
3104 value. Depending on the environment, this could lead to "Failed: internal
3105 error: code overflow at offset 49" or "glibc detected double free or
3106 corruption" errors.
3107
3108 27. Applied patches from Google (a) to support the new newline modes and (b) to
3109 advance over multibyte UTF-8 characters in GlobalReplace.
3110
3111 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3112 difference for some implementation of PCRE in some Windows version.
3113
3114 29. Added some extra testing facilities to pcretest:
3115
3116 \q<number> in a data line sets the "match limit" value
3117 \Q<number> in a data line sets the "match recursion limt" value
3118 -S <number> sets the stack size, where <number> is in megabytes
3119
3120 The -S option isn't available for Windows.
3121
3122
3123 Version 6.6 06-Feb-06
3124 ---------------------
3125
3126 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3127 in pcreposix.h. I have copied the definition from pcre.h.
3128
3129 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3130 because pcre.h is no longer a built file.
3131
3132 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3133 not normally included in the compiled code.
3134
3135
3136 Version 6.5 01-Feb-06
3137 ---------------------
3138
3139 1. When using the partial match feature with pcre_dfa_exec(), it was not
3140 anchoring the second and subsequent partial matches at the new starting
3141 point. This could lead to incorrect results. For example, with the pattern
3142 /1234/, partially matching against "123" and then "a4" gave a match.
3143
3144 2. Changes to pcregrep:
3145
3146 (a) All non-match returns from pcre_exec() were being treated as failures
3147 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3148 error message is output. Some extra information is given for the
3149 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3150 probably the only errors that are likely to be caused by users (by
3151 specifying a regex that has nested indefinite repeats, for instance).
3152 If there are more than 20 of these errors, pcregrep is abandoned.
3153
3154 (b) A binary zero was treated as data while matching, but terminated the
3155 output line if it was written out. This has been fixed: binary zeroes
3156 are now no different to any other data bytes.
3157
3158 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3159 used to set a locale for matching. The --locale=xxxx long option has
3160 been added (no short equivalent) to specify a locale explicitly on the
3161 pcregrep command, overriding the environment variables.
3162
3163 (d) When -B was used with -n, some line numbers in the output were one less
3164 than they should have been.
3165
3166 (e) Added the -o (--only-matching) option.
3167
3168 (f) If -A or -C was used with -c (count only), some lines of context were
3169 accidentally printed for the final match.
3170
3171 (g) Added the -H (--with-filename) option.
3172
3173 (h) The combination of options -rh failed to suppress file names for files
3174 that were found from directory arguments.
3175
3176 (i) Added the -D (--devices) and -d (--directories) options.
3177
3178 (j) Added the -F (--fixed-strings) option.
3179
3180 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3181
3182 (l) Added the --colo(u)r option.
3183
3184 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3185 is not present by default.
3186
3187 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3188 items such as (?R) or (?1), when the recursion could match a number of
3189 alternatives. If it matched one of the alternatives, but subsequently,
3190 outside the recursion, there was a failure, the code tried to back up into
3191 the recursion. However, because of the way PCRE is implemented, this is not
3192 possible, and the result was an incorrect result from the match.
3193
3194 In order to prevent this happening, the specification of recursion has
3195 been changed so that all such subpatterns are automatically treated as
3196 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3197
3198 4. I had overlooked the fact that, in some locales, there are characters for
3199 which isalpha() is true but neither isupper() nor islower() are true. In
3200 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3201 and ordfeminine) are like this. This affected the treatment of \w and \W
3202 when they appeared in character classes, but not when they appeared outside
3203 a character class. The bit map for "word" characters is now created
3204 separately from the results of isalnum() instead of just taking it from the
3205 upper, lower, and digit maps. (Plus the underscore character, of course.)
3206
3207 5. The above bug also affected the handling of POSIX character classes such as
3208 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3209 permanent tables. Instead, the bit maps for such a class were previously
3210 created as the appropriate unions of the upper, lower, and digit bitmaps.
3211 Now they are created by subtraction from the [[:word:]] class, which has
3212 its own bitmap.
3213
3214 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3215 It is created by subtracting the vertical space characters (\x09, \x0a,
3216 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3217 subtraction was done in the overall bitmap for a character class, meaning
3218 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3219 be recognized. This bug has been fixed.
3220
3221 7. Patches from the folks at Google:
3222
3223 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3224 real life, but is still worth protecting against".
3225
3226 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3227 regular expressions".
3228
3229 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3230 have it.
3231
3232 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3233 "configure" and the latter not, in order to fix a problem somebody had
3234 with compiling the Arg class on HP-UX.
3235
3236 (e) Improve the error-handling of the C++ wrapper a little bit.
3237
3238 (f) New tests for checking recursion limiting.
3239
3240 8. The pcre_memmove() function, which is used only if the environment does not
3241 have a standard memmove() function (and is therefore rarely compiled),
3242 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3243 returning a result (though PCRE never actually uses the result).
3244
3245 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3246 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3247 returned instead of calling malloc() with an overflowing number that would
3248 most likely cause subsequent chaos.
3249
3250 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3251
3252 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3253 with this option is matched, the nmatch and pmatch options of regexec() are
3254 ignored.
3255
3256 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3257 provided in case anyone wants to the the POSIX interface with UTF-8
3258 strings.
3259
3260 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3261 C++ linking (needed for some HP-UX environments).
3262
3263 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3264 (unused parameter) and in the pcre_printint() function (omitted "default"
3265 switch label when the default is to do nothing).
3266
3267 15. Added some code to make it possible, when PCRE is compiled as a C++
3268 library, to replace subject pointers for pcre_exec() with a smart pointer
3269 class, thus making it possible to process discontinuous strings.
3270
3271 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3272 much the same function. They were added by different people who were trying
3273 to make PCRE easy to compile on non-Unix systems. It has been suggested
3274 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3275 for compiling on Windows systems. I have therefore replaced it with
3276 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3277 defaults to "extern" for C or "extern C" for C++, which works fine on
3278 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3279 SCOPE with something explicit in config.h. In addition:
3280
3281 (a) pcreposix.h still had just "extern" instead of either of these macros;
3282 I have replaced it with PCRE_DATA_SCOPE.
3283
3284 (b) Functions such as _pcre_xclass(), which are internal to the library,
3285 but external in the C sense, all had PCRE_EXPORT in their definitions.
3286 This is apparently wrong for the Windows case, so I have removed it.
3287 (It makes no difference on Unix-like systems.)
3288
3289 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3290 of recursive calls to match(). This is different to MATCH_LIMIT because
3291 that limits the total number of calls to match(), not all of which increase
3292 the depth of recursion. Limiting the recursion depth limits the amount of
3293 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3294 when PCRE is compiled, and changed at run time. A patch from Google adds
3295 this functionality to the C++ interface.
3296
3297 18. Changes to the handling of Unicode character properties:
3298
3299 (a) Updated the table to Unicode 4.1.0.
3300
3301 (b) Recognize characters that are not in the table as "Cn" (undefined).
3302
3303 (c) I revised the way the table is implemented to a much improved format
3304 which includes recognition of ranges. It now supports the ranges that
3305 are defined in UnicodeData.txt, and it also amalgamates other
3306 characters into ranges. This has reduced the number of entries in the
3307 table from around 16,000 to around 3,000, thus reducing its size
3308 considerably. I realized I did not need to use a tree structure after
3309 all - a binary chop search is just as efficient. Having reduced the
3310 number of entries, I extended their size from 6 bytes to 8 bytes to
3311 allow for more data.
3312
3313 (d) Added support for Unicode script names via properties such as \p{Han}.
3314
3315 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3316 matching that character.
3317
3318 20. When matching a repeated Unicode property with a minimum greater than zero,
3319 (for example \pL{2,}), PCRE could look past the end of the subject if it
3320 reached it while seeking the minimum number of characters. This could
3321 happen only if some of the characters were more than one byte long, because
3322 there is a check for at least the minimum number of bytes.
3323
3324 21. Refactored the implementation of \p and \P so as to be more general, to
3325 allow for more different types of property in future. This has changed the
3326 compiled form incompatibly. Anybody with saved compiled patterns that use
3327 \p or \P will have to recompile them.
3328
3329 22. Added "Any" and "L&" to the supported property types.
3330
3331 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3332 but give a compile time error if the value is greater than 0xff.
3333
3334 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3335 accidentally not being installed or uninstalled.
3336
3337 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3338 made were to insert the current release number. This seemed silly, because
3339 it made things harder for people building PCRE on systems that don't run
3340 "configure". I have turned pcre.h into a distributed file, no longer built
3341 by "configure", with the version identification directly included. There is
3342 no longer a pcre.h.in file.
3343
3344 However, this change necessitated a change to the pcre-config script as
3345 well. It is built from pcre-config.in, and one of the substitutions was the
3346 release number. I have updated configure.ac so that ./configure now finds
3347 the release number by grepping pcre.h.
3348
3349 26. Added the ability to run the tests under valgrind.
3350
3351
3352 Version 6.4 05-Sep-05
3353 ---------------------
3354
3355 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3356 "--" to be printed when multiple files were scanned, even when none of the
3357 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3358 consider it to be a bug, and have restored the previous behaviour.
3359
3360 2. A couple of code tidies to get rid of compiler warnings.
3361
3362 3. The pcretest program used to cheat by referring to symbols in the library
3363 whose names begin with _pcre_. These are internal symbols that are not
3364 really supposed to be visible externally, and in some environments it is
3365 possible to suppress them. The cheating is now confined to including
3366 certain files from the library's source, which is a bit cleaner.
3367
3368 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3369 file's purpose clearer.
3370
3371 5. Reorganized pcre_ucp_findchar().
3372
3373
3374 Version 6.3 15-Aug-05
3375 ---------------------
3376
3377 1. The file libpcre.pc.in did not have general read permission in the tarball.
3378
3379 2. There were some problems when building without C++ support:
3380
3381 (a) If C++ support was not built, "make install" and "make test" still
3382 tried to test it.
3383
3384 (b) There were problems when the value of CXX was explicitly set. Some
3385 changes have been made to try to fix these, and ...
3386
3387 (c) --disable-cpp can now be used to explicitly disable C++ support.
3388
3389 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3390 backslash in a target when C++ was disabled. This confuses some
3391 versions of "make", apparently. Using an intermediate variable solves
3392 this. (Same for CPP_LOBJ.)
3393
3394 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3395 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3396 necessary on certain architectures.
3397
3398 4. Added a setting of -export-symbols-regex to the link command to remove
3399 those symbols that are exported in the C sense, but actually are local
3400 within the library, and not documented. Their names all begin with
3401 "_pcre_". This is not a perfect job, because (a) we have to except some
3402 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3403 available (and never for static libraries). I have made a note to try to
3404 find a way round (a) in the future.
3405
3406
3407 Version 6.2 01-Aug-05
3408 ---------------------
3409
3410 1. There was no test for integer overflow of quantifier values. A construction
3411 such as {1111111111111111} would give undefined results. What is worse, if
3412 a minimum quantifier for a parenthesized subpattern overflowed and became
3413 negative, the calculation of the memory size went wrong. This could have
3414 led to memory overwriting.
3415
3416 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3417
3418 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3419 operating environments where this matters.
3420
3421 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3422 PCRE options from within the C++ wrapper.
3423
3424 5. Named capturing subpatterns were not being correctly counted when a pattern
3425 was compiled. This caused two problems: (a) If there were more than 100
3426 such subpatterns, the calculation of the memory needed for the whole
3427 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3428 back references of the form \12, where the number was greater than 9, were
3429 not recognized as back references, even though there were sufficient
3430 previous subpatterns.
3431
3432 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3433 versions of gcc, e.g. 2.95.4.
3434
3435
3436 Version 6.1 21-Jun-05
3437 ---------------------
3438
3439 1. There was one reference to the variable "posix" in pcretest.c that was not
3440 surrounded by "#if !defined NOPOSIX".
3441
3442 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3443 the cross-check on the old pcre_info() function, for the benefit of the
3444 cut-down version of PCRE that is currently imported into Exim.
3445
3446 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3447 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3448 patterns that start (?i) but I don't think that matters. The use of (?i) is
3449 just an example; this all applies to the other options as well.
3450
3451 4. Since libtool seems to echo the compile commands it is issuing, the output
3452 from "make" can be reduced a bit by putting "@" in front of each libtool
3453 compile command.
3454
3455 5. Patch from the folks at Google for configure.in to be a bit more thorough
3456 in checking for a suitable C++ installation before trying to compile the
3457 C++ stuff. This should fix a reported problem when a compiler was present,
3458 but no suitable headers.
3459
3460 6. The man pages all had just "PCRE" as their title. I have changed them to
3461 be the relevant file name. I have also arranged that these names are
3462 retained in the file doc/pcre.txt, which is a concatenation in text format
3463 of all the man pages except the little individual ones for each function.
3464
3465 7. The NON-UNIX-USE file had not been updated for the different set of source
3466 files that come with release 6. I also added a few comments about the C++
3467 wrapper.
3468
3469
3470 Version 6.0 07-Jun-05
3471 ---------------------
3472
3473 1. Some minor internal re-organization to help with my DFA experiments.
3474
3475 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3476 didn't matter for the library itself when fully configured, but did matter
3477 when compiling without UCP support, or within Exim, where the ucp files are
3478 not imported.
3479
3480 3. Refactoring of the library code to split up the various functions into
3481 different source modules. The addition of the new DFA matching code (see
3482 below) to a single monolithic source would have made it really too
3483 unwieldy, quite apart from causing all the code to be include in a
3484 statically linked application, when only some functions are used. This is
3485 relevant even without the DFA addition now that patterns can be compiled in
3486 one application and matched in another.
3487
3488 The downside of splitting up is that there have to be some external
3489 functions and data tables that are used internally in different modules of
3490 the library but which are not part of the API. These have all had their
3491 names changed to start with "_pcre_" so that they are unlikely to clash
3492 with other external names.
3493
3494 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3495 a different (DFA) algorithm. Although it is slower than the original
3496 function, it does have some advantages for certain types of matching
3497 problem.
3498
3499 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3500 including restarting after a partial match.
3501
3502 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3503 defined when compiling for Windows was sent to me. I have put it into the
3504 code, though I have no means of testing or verifying it.
3505
3506 7. Added the pcre_refcount() auxiliary function.
3507
3508 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3509 match before or at the first newline in the subject string. In pcretest,
3510 the /f option on a pattern can be used to set this.
3511
3512 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3513 would behave wrongly. This has been present in PCRE since release 4.0.
3514
3515 10. A number of changes to the pcregrep command:
3516
3517 (a) Refactored how -x works; insert ^(...)$ instead of setting
3518 PCRE_ANCHORED and checking the length, in preparation for adding
3519 something similar for -w.
3520
3521 (b) Added the -w (match as a word) option.
3522
3523 (c) Refactored the way lines are read and buffered so as to have more
3524 than one at a time available.
3525
3526 (d) Implemented a pcregrep test script.
3527
3528 (e) Added the -M (multiline match) option. This allows patterns to match
3529 over several lines of the subject. The buffering ensures that at least
3530 8K, or the rest of the document (whichever is the shorter) is available
3531 for matching (and similarly the previous 8K for lookbehind assertions).
3532
3533 (f) Changed the --help output so that it now says
3534
3535 -w, --word-regex(p)
3536
3537 instead of two lines, one with "regex" and the other with "regexp"
3538 because that confused at least one person since the short forms are the
3539 same. (This required a bit of code, as the output is generated
3540 automatically from a table. It wasn't just a text change.)
3541
3542 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3543 option but starts with a hyphen. Could be a pattern or a path name
3544 starting with a hyphen, for instance.
3545
3546 (h) "-" can be given as a file name to represent stdin.
3547
3548 (i) When file names are being printed, "(standard input)" is used for
3549 the standard input, for compatibility with GNU grep. Previously
3550 "<stdin>" was used.
3551
3552 (j) The option --label=xxx can be used to supply a name to be used for
3553 stdin when file names are being printed. There is no short form.
3554
3555 (k) Re-factored the options decoding logic because we are going to add
3556 two more options that take data. Such options can now be given in four
3557 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3558
3559 (l) Added the -A, -B, and -C options for requesting that lines of context
3560 around matches be printed.
3561
3562 (m) Added the -L option to print the names of files that do not contain
3563 any matching lines, that is, the complement of -l.
3564
3565 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3566 continue to scan other files.
3567
3568 (o) The -s option was incorrectly implemented. For compatibility with other
3569 greps, it now suppresses the error message for a non-existent or non-
3570 accessible file (but not the return code). There is a new option called
3571 -q that suppresses the output of matching lines, which was what -s was
3572 previously doing.
3573
3574 (p) Added --include and --exclude options to specify files for inclusion
3575 and exclusion when recursing.
3576
3577 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3578 Hopefully, it now does.
3579
3580 12. Missing cast in pcre_study().
3581
3582 13. Added an "uninstall" target to the makefile.
3583
3584 14. Replaced "extern" in the function prototypes in Makefile.in with
3585 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3586 world, but is set differently for Windows.
3587
3588 15. Added a second compiling function called pcre_compile2(). The only
3589 difference is that it has an extra argument, which is a pointer to an
3590 integer error code. When there is a compile-time failure, this is set
3591 non-zero, in addition to the error test pointer being set to point to an
3592 error message. The new argument may be NULL if no error number is required
3593 (but then you may as well call pcre_compile(), which is now just a
3594 wrapper). This facility is provided because some applications need a
3595 numeric error indication, but it has also enabled me to tidy up the way
3596 compile-time errors are handled in the POSIX wrapper.
3597
3598 16. Added VPATH=.libs to the makefile; this should help when building with one
3599 prefix path and installing with another. (Or so I'm told by someone who
3600 knows more about this stuff than I do.)
3601
3602 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3603 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3604 match everything, including newlines. This is not POSIX-compatible, but
3605 somebody wanted the feature. From pcretest it can be activated by using
3606 both the P and the s flags.
3607
3608 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3609
3610 19. libpcre.pc was being incorrectly installed as executable.
3611
3612 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3613 it now also looks for '\r' so that it will work unmodified on Windows.
3614
3615 21. Added Google's contributed C++ wrapper to the distribution.
3616
3617 22. Added some untidy missing memory free() calls in pcretest, to keep
3618 Electric Fence happy when testing.
3619
3620
3621
3622 Version 5.0 13-Sep-04
3623 ---------------------
3624
3625 1. Internal change: literal characters are no longer packed up into items
3626 containing multiple characters in a single byte-string. Each character
3627 is now matched using a separate opcode. However, there may be more than one
3628 byte in the character in UTF-8 mode.
3629
3630 2. The pcre_callout_block structure has two new fields: pattern_position and
3631 next_item_length. These contain the offset in the pattern to the next match
3632 item, and its length, respectively.
3633
3634 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3635 insertion of callouts before each pattern item. Added the /C option to
3636 pcretest to make use of this.
3637
3638 4. On the advice of a Windows user, the lines
3639
3640 #if defined(_WIN32) || defined(WIN32)
3641 _setmode( _fileno( stdout ), 0x8000 );
3642 #endif /* defined(_WIN32) || defined(WIN32) */
3643
3644 have been added to the source of pcretest. This apparently does useful
3645 magic in relation to line terminators.
3646
3647 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3648 for the benefit of those environments where the "b" makes a difference.
3649
3650 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3651 to know about it. I have put a hack into configure.in that adds in code
3652 to set GCC=yes if CC=icc. This seems to end up at a point in the
3653 generated configure script that is early enough to affect the setting of
3654 compiler options, which is what is needed, but I have no means of testing
3655 whether it really works. (The user who reported this had patched the
3656 generated configure script, which of course I cannot do.)
3657
3658 LATER: After change 22 below (new libtool files), the configure script
3659 seems to know about icc (and also ecc). Therefore, I have commented out
3660 this hack in configure.in.
3661
3662 7. Added support for pkg-config (2 patches were sent in).
3663
3664 8. Negated POSIX character classes that used a combination of internal tables
3665 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3666 [[:^ascii]]. Typically, they would match almost any characters. The other
3667 POSIX classes were not broken in this way.
3668
3669 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3670 to find the match, as PCRE was deluded into thinking that the match had to
3671 start at the start point or following a newline. The same bug applied to
3672 patterns with negative forward assertions or any backward assertions
3673 preceding ".*" at the start, unless the pattern required a fixed first
3674 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3675
3676 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3677 starting at the last subject character, bytes beyond the end of the subject
3678 string were read.
3679
3680 11. Renamed the variable "class" as "classbits" to make life easier for C++
3681 users. (Previously there was a macro definition, but it apparently wasn't
3682 enough.)
3683
3684 12. Added the new field "tables" to the extra data so that tables can be passed
3685 in at exec time, or the internal tables can be re-selected. This allows
3686 a compiled regex to be saved and re-used at a later time by a different
3687 program that might have everything at different addresses.
3688
3689 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3690 -R library as well as a -L library.
3691
3692 14. The debugging options of pcretest (-d on the command line or D on a
3693 pattern) showed incorrect output for anything following an extended class
3694 that contained multibyte characters and which was followed by a quantifier.
3695
3696 15. Added optional support for general category Unicode character properties
3697 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3698 support. It adds about 90K to the size of the library. The meanings of the
3699 inbuilt class escapes such as \d and \s have NOT been changed.
3700
3701 16. Updated pcredemo.c to include calls to free() to release the memory for the
3702 compiled pattern.
3703
3704 17. The generated file chartables.c was being created in the source directory
3705 instead of in the building directory. This caused the build to fail if the
3706 source directory was different from the building directory, and was
3707 read-only.
3708
3709 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3710 file. No doubt somebody will tell me if they don't make sense... Also added
3711 Dan Mooney's comments about building on OpenVMS.
3712
3713 19. Added support for partial matching via the PCRE_PARTIAL option for
3714 pcre_exec() and the \P data escape in pcretest.
3715
3716 20. Extended pcretest with 3 new pattern features:
3717
3718 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3719 write the compiled pattern to the file whose name is "rest-of-line".
3720 This is a straight binary dump of the data, with the saved pointer to
3721 the character tables forced to be NULL. The study data, if any, is
3722 written too. After writing, pcretest reads a new pattern.
3723
3724 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3725 compiled pattern from the given file. There must not be any
3726 occurrences of "<" in the file name (pretty unlikely); if there are,
3727 pcretest will instead treat the initial "<" as a pattern delimiter.
3728 After reading in the pattern, pcretest goes on to read data lines as
3729 usual.
3730
3731 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3732 and 16-bit fields in a compiled pattern, to simulate a pattern that
3733 was compiled on a host of opposite endianness.
3734
3735 21. The pcre-exec() function can now cope with patterns that were compiled on
3736 hosts of opposite endianness, with this restriction:
3737
3738 As for any compiled expression that is saved and used later, the tables
3739 pointer field cannot be preserved; the extra_data field in the arguments
3740 to pcre_exec() should be used to pass in a tables address if a value
3741 other than the default internal tables were used at compile time.
3742
3743 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3744 now diagnosed as an error. Previously, most of the time, a negative number
3745 would have been treated as zero, but if in addition "ovector" was passed as
3746 NULL, a crash could occur.
3747
3748 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3749 new versions from the libtool 1.5 distribution (the last one is a copy of
3750 a file called libtool.m4). This seems to have fixed the need to patch
3751 "configure" to support Darwin 1.3 (which I used to do). However, I still
3752 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3753 workstation).
3754
3755 24. Changed the PCRE licence to be the more standard "BSD" licence.
3756
3757
3758 Version 4.5 01-Dec-03
3759 ---------------------
3760
3761 1. There has been some re-arrangement of the code for the match() function so
3762 that it can be compiled in a version that does not call itself recursively.
3763 Instead, it keeps those local variables that need separate instances for
3764 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3765 needs to "recurse". Keeping track of where control must go is done by means
3766 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3767 hide most of the details from the main code, and operates only if
3768 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3769 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3770 operating.
3771
3772 To make it easier for callers to provide specially tailored get/free
3773 functions for this usage, two new functions, pcre_stack_malloc, and
3774 pcre_stack_free, are used. They are always called in strict stacking order,
3775 and the size of block requested is always the same.
3776
3777 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3778 PCRE has been compiled to use the stack or the heap for recursion. The
3779 -C option of pcretest uses this to show which version is compiled.
3780
3781 A new data escape \S, is added to pcretest; it causes the amounts of store
3782 obtained and freed by both kinds of malloc/free at match time to be added
3783 to the output.
3784
3785 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3786 what's available on my current Linux desktop machine.
3787
3788 3. When matching a UTF-8 string, the test for a valid string at the start has
3789 been extended. If start_offset is not zero, PCRE now checks that it points
3790 to a byte that is the start of a UTF-8 character. If not, it returns
3791 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3792 this is necessary because there may be backward assertions in the pattern.
3793 When matching the same subject several times, it may save resources to use
3794 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3795
3796 4. The code for checking the validity of UTF-8 strings has been tightened so
3797 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3798 containing "overlong sequences".
3799
3800 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3801 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3802 should have been "&", but it just so happened that all the cases this let
3803 through by mistake were picked up later in the function.
3804
3805 6. I had used a variable called "isblank" - this is a C99 function, causing
3806 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3807
3808 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3809 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3810 so the version is shown; (c) stop "make test" echoing "./RunTest".
3811
3812 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3813
3814 9. The prototype for memmove() for systems that don't have it was using
3815 size_t, but the inclusion of the header that defines size_t was later. I've
3816 moved the #includes for the C headers earlier to avoid this.
3817
3818 10. Added some adjustments to the code to make it easier to compiler on certain
3819 special systems:
3820
3821 (a) Some "const" qualifiers were missing.
3822 (b) Added the macro EXPORT before all exported functions; by default this
3823 is defined to be empty.
3824 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3825 that it reads its output file name as an argument instead of writing
3826 to the standard output and assuming this can be redirected.
3827
3828 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3829 class containing characters with values greater than 255, PCRE compilation
3830 went into a loop.
3831
3832 12. A recursive reference to a subpattern that was within another subpattern
3833 that had a minimum quantifier of zero caused PCRE to crash. For example,
3834 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3835 recursion. If the recursively-called subpattern itself had a zero repeat,
3836 that was OK.
3837
3838 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3839 buffer into which it was copied (for escape processing) was still set at
3840 1024, so long lines caused crashes.
3841
3842 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3843 "internal error: code overflow...". This applied to any character class
3844 that was followed by a possessive quantifier.
3845
3846 15. Modified the Makefile to add libpcre.la as a prerequisite for
3847 libpcreposix.la because I was told this is needed for a parallel build to
3848 work.
3849
3850 16. If a pattern that contained .* following optional items at the start was
3851 studied, the wrong optimizing data was generated, leading to matching
3852 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3853 matching string must start with a or b or c. The correct conclusion for
3854 this pattern is that a match can start with any character.
3855
3856
3857 Version 4.4 13-Aug-03
3858 ---------------------
3859
3860 1. In UTF-8 mode, a character class containing characters with values between
3861 127 and 255 was not handled correctly if the compiled pattern was studied.
3862 In fixing this, I have also improved the studying algorithm for such
3863 classes (slightly).
3864
3865 2. Three internal functions had redundant arguments passed to them. Removal
3866 might give a very teeny performance improvement.
3867
3868 3. Documentation bug: the value of the capture_top field in a callout is *one
3869 more than* the number of the hightest numbered captured substring.
3870
3871 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3872 in incorrectly linking with a previously installed version. They now link
3873 explicitly with libpcre.la.
3874
3875 5. configure.in no longer needs to recognize Cygwin specially.
3876
3877 6. A problem in pcre.in for Windows platforms is fixed.
3878
3879 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3880 pcretest, it used to include the size of the study block as part of its
3881 output. Unfortunately, the structure contains a field that has a different
3882 size on different hardware architectures. This meant that the tests that
3883 showed this size failed. As the block is currently always of a fixed size,
3884 this information isn't actually particularly useful in pcretest output, so
3885 I have just removed it.
3886
3887 8. Three pre-processor statements accidentally did not start in column 1.
3888 Sadly, there are *still* compilers around that complain, even though
3889 standard C has not required this for well over a decade. Sigh.
3890
3891 9. In pcretest, the code for checking callouts passed small integers in the
3892 callout_data field, which is a void * field. However, some picky compilers
3893 complained about the casts involved for this on 64-bit systems. Now
3894 pcretest passes the address of the small integer instead, which should get
3895 rid of the warnings.
3896
3897 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3898 both compile and run time, and gives an error if an invalid UTF-8 sequence
3899 is found. There is a option for disabling this check in cases where the
3900 string is known to be correct and/or the maximum performance is wanted.
3901
3902 11. In response to a bug report, I changed one line in Makefile.in from
3903
3904 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3905 to
3906 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3907
3908 to look similar to other lines, but I have no way of telling whether this
3909 is the right thing to do, as I do not use Windows. No doubt I'll get told
3910 if it's wrong...
3911
3912
3913 Version 4.3 21-May-03
3914 ---------------------
3915
3916 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3917 Makefile.
3918
3919 2. Some refactoring to improve the quality of the code:
3920
3921 (i) The utf8_table... variables are now declared "const".
3922
3923 (ii) The code for \cx, which used the "case flipping" table to upper case
3924 lower case letters, now just substracts 32. This is ASCII-specific,
3925 but the whole concept of \cx is ASCII-specific, so it seems
3926 reasonable.
3927
3928 (iii) PCRE was using its character types table to recognize decimal and
3929 hexadecimal digits in the pattern. This is silly, because it handles
3930 only 0-9, a-f, and A-F, but the character types table is locale-
3931 specific, which means strange things might happen. A private
3932 table is now used for this - though it costs 256 bytes, a table is
3933 much faster than multiple explicit tests. Of course, the standard
3934 character types table is still used for matching digits in subject
3935 strings against \d.
3936
3937 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3938 ending in _t are). So I've renamed it as ESC_tee.
3939
3940 3. The first argument for regexec() in the POSIX wrapper should have been
3941 defined as "const".
3942
3943 4. Changed pcretest to use malloc() for its buffers so that they can be
3944 Electric Fenced for debugging.
3945
3946 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3947 to read one or more bytes before the start of the subject string. Often this
3948 had no effect on PCRE's behaviour, but in some circumstances it could
3949 provoke a segmentation fault.
3950
3951 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3952 to try to read one or more bytes before the start of the subject string.
3953
3954 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3955 UTF-8 support could misbehave in various ways if the subject string
3956 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3957 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3958 back over UTF-8 characters.)
3959
3960
3961 Version 4.2 14-Apr-03
3962 ---------------------
3963
3964 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3965
3966 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3967 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3968 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3969 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3970 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3971 and BUILD_EXEEXT
3972 Note: automatic setting of the BUILD variables is not yet working
3973 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3974 compile-time but not at link-time
3975 [LINK]: use for linking executables only
3976 make different versions for Windows and non-Windows
3977 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3978 libraries
3979 [LINK_FOR_BUILD]: new variable
3980 [OBJEXT]: use throughout
3981 [EXEEXT]: use throughout
3982 <winshared>: new target
3983 <wininstall>: new target
3984 <dftables.o>: use native compiler
3985 <dftables>: use native linker
3986 <install>: handle Windows platform correctly
3987 <clean>: ditto
3988 <check>: ditto
3989 copy DLL to top builddir before testing
3990
3991 As part of these changes, -no-undefined was removed again. This was reported
3992 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3993 in any case.
3994
3995 3. Some tidies to get rid of compiler warnings:
3996
3997 . In the match_data structure, match_limit was an unsigned long int, whereas
3998 match_call_count was an int. I've made them both unsigned long ints.
3999
4000 . In pcretest the fact that a const uschar * doesn't automatically cast to
4001 a void * provoked a warning.
4002
4003 . Turning on some more compiler warnings threw up some "shadow" variables
4004 and a few more missing casts.
4005
4006 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4007 option, a class that contained a single character with a value between 128
4008 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4009
4010 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4011 option, a class that contained several characters, but with at least one
4012 whose value was between 128 and 255 caused PCRE to crash.
4013
4014
4015 Version 4.1 12-Mar-03
4016 ---------------------
4017
4018 1. Compiling with gcc -pedantic found a couple of places where casts were
4019 needed, and a string in dftables.c that was longer than standard compilers are
4020 required to support.
4021
4022 2. Compiling with Sun's compiler found a few more places where the code could
4023 be tidied up in order to avoid warnings.
4024
4025 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4026 first of these names is deprecated in the latest Autoconf in favour of the name
4027 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4028 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4029 analogy I have changed it to CFLAGS_FOR_BUILD.
4030
4031 4. Added -no-undefined to the linking command in the Makefile, because this is
4032 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4033 linking step for the pcreposix library.
4034
4035 5. PCRE was failing to diagnose the case of two named groups with the same
4036 name.
4037
4038 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4039 literal character that is needed in the subject for a match, and scans along to
4040 ensure that it is present before embarking on the full matching process. This
4041 saves time in cases of nested unlimited repeats that are never going to match.
4042 Problem: the scan can take a lot of time if the subject is very long (e.g.
4043 megabytes), thus penalizing straightforward matches. It is now done only if the
4044 amount of subject to be scanned is less than 1000 bytes.
4045
4046 7. A lesser problem with the same optimization is that it was recording the
4047 first character of an anchored pattern as "needed", thus provoking a search
4048 right along the subject, even when the first match of the pattern was going to
4049 fail. The "needed" character is now not set for anchored patterns, unless it
4050 follows something in the pattern that is of non-fixed length. Thus, it still
4051 fulfils its original purpose of finding quick non-matches in cases of nested
4052 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4053
4054
4055 Version 4.0 17-Feb-03
4056 ---------------------
4057
4058 1. If a comment in an extended regex that started immediately after a meta-item
4059 extended to the end of string, PCRE compiled incorrect data. This could lead to
4060 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4061
4062 2. Moved to autoconf 2.53 and libtool 1.4.2.
4063
4064 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4065 the special perltest8 script is no longer needed - all the tests can be run
4066 from a single perltest script.
4067
4068 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4069 by \s. It has now been removed in PCRE. This means it isn't recognized as
4070 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4071 class [:space:] *does* include VT, thereby creating a mess.
4072
4073 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4074 space and tab.
4075
4076 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4077 its new features into the main test script, reducing the number of scripts.
4078
4079 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4080 were backward compatible, and made the (?i) apply to the whole pattern, as if
4081 /i were given. Now it behaves more logically, and applies the option setting
4082 only to what follows. PCRE has been changed to follow suit. However, if it
4083 finds options settings right at the start of the pattern, it extracts them into
4084 the global options, as before. Thus, they show up in the info data.
4085
4086 8. Added support for the \Q...\E escape sequence. Characters in between are
4087 treated as literals. This is slightly different from Perl in that $ and @ are
4088 also handled as literals inside the quotes. In Perl, they will cause variable
4089 interpolation. Note the following examples:
4090
4091 Pattern PCRE matches Perl matches
4092
4093 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4094 \Qabc\$xyz\E abc\$xyz abc\$xyz
4095 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4096
4097 For compatibility with Perl, \Q...\E sequences are recognized inside character
4098 classes as well as outside them.
4099
4100 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4101 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4102 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4103 signed/unsigned warnings.
4104
4105 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4106 option for pcretest, so I've replaced it by a simple function that does just
4107 that job.
4108
4109 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4110 "pcregrep -".
4111
4112 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4113 Java package. This provides some syntactic sugar for simple cases of what my
4114 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4115 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4116 item, you can use this simplified notation. Note that only makes sense with
4117 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4118 greediness, whatever the setting of the PCRE_UNGREEDY option.
4119
4120 13. A change of greediness default within a pattern was not taking effect at
4121 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4122 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4123 was abstracted outside.
4124
4125 14. PCRE now supports the \G assertion. It is true when the current matching
4126 position is at the start point of the match. This differs from \A when the
4127 starting offset is non-zero. Used with the /g option of pcretest (or similar
4128 code), it works in the same way as it does for Perl's /g option. If all
4129 alternatives of a regex begin with \G, the expression is anchored to the start
4130 match position, and the "anchored" flag is set in the compiled expression.
4131
4132 15. Some bugs concerning the handling of certain option changes within patterns
4133 have been fixed. These applied to options other than (?ims). For example,
4134 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4135 been the other way round. Some of this was related to change 7 above.
4136
4137 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4138 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4139 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4140 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4141
4142 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4143 mode. Unlike ".", it always matches newline, whatever the setting of
4144 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4145 assertions. Perl allows it, but it doesn't (in general) work because it can't
4146 calculate the length of the lookbehind. At least, that's the case for Perl
4147 5.8.0 - I've been told they are going to document that it doesn't work in
4148 future.
4149
4150 18. Added an error diagnosis for escapes that PCRE does not support: these are
4151 \L, \l, \N, \P, \p, \U, \u, and \X.
4152
4153 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4154 reading past the end of the pattern in cases such as /[abcd/.
4155
4156 20. PCRE was getting more memory than necessary for patterns with classes that
4157 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4158
4159 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4160 compiling PCRE for use with Virtual Pascal.
4161
4162 22. Small fix to the Makefile to make it work properly if the build is done
4163 outside the source tree.
4164
4165 23. Added a new extension: a condition to go with recursion. If a conditional
4166 subpattern starts with (?(R) the "true" branch is used if recursion has
4167 happened, whereas the "false" branch is used only at the top level.
4168
4169 24. When there was a very long string of literal characters (over 255 bytes
4170 without UTF support, over 250 bytes with UTF support), the computation of how
4171 much memory was required could be incorrect, leading to segfaults or other
4172 strange effects.
4173
4174 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4175 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4176 there was a subsequent back reference to those brackets. This meant that, for
4177 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4178 possible to check for precisely this case. All we can do is abandon the
4179 optimization if .* occurs inside capturing brackets when there are any back
4180 references whatsoever. (See below for a better fix that came later.)
4181
4182 26. The handling of the optimization for finding the first character of a
4183 non-anchored pattern, and for finding a character that is required later in the
4184 match were failing in some cases. This didn't break the matching; it just
4185 failed to optimize when it could. The way this is done has been re-implemented.
4186
4187 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4188
4189 28. Added a new feature that provides some of the functionality that Perl
4190 provides with (?{...}). The facility is termed a "callout". The way it is done
4191 in PCRE is for the caller to provide an optional function, by setting
4192 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4193 global variable. By default it is unset, which disables all calling out. To get
4194 the function called, the regex must include (?C) at appropriate points. This
4195 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4196 This provides a means of identifying different callout points. When PCRE
4197 reaches such a point in the regex, if pcre_callout has been set, the external
4198 function is called. It is provided with data in a structure called
4199 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4200 matching continues; if it returns a non-zero value, the match at the current
4201 point fails. However, backtracking will occur if possible. [This was changed
4202 later and other features added - see item 49 below.]
4203
4204 29. pcretest is upgraded to test the callout functionality. It provides a
4205 callout function that displays information. By default, it shows the start of
4206 the match and the current position in the text. There are some new data escapes
4207 to vary what happens:
4208
4209 \C+ in addition, show current contents of captured substrings
4210 \C- do not supply a callout function
4211 \C!n return 1 when callout number n is reached
4212 \C!n!m return 1 when callout number n is reached for the mth time
4213
4214 30. If pcregrep was called with the -l option and just a single file name, it
4215 output "<stdin>" if a match was found, instead of the file name.
4216
4217 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4218 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4219 pcre_exec(). This saves a malloc/free per call. The default value of
4220 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4221 when configuring.
4222
4223 32. The default maximum size of a compiled pattern is 64K. There have been a
4224 few cases of people hitting this limit. The code now uses macros to handle the
4225 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4226 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4227 configuring. Tests 2 and 5 work only with 2-byte links because they output
4228 debugging information about compiled patterns.
4229
4230 33. Internal code re-arrangements:
4231
4232 (a) Moved the debugging function for printing out a compiled regex into
4233 its own source file (printint.c) and used #include to pull it into
4234 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4235 separate copies.
4236
4237 (b) Defined the list of op-code names for debugging as a macro in
4238 internal.h so that it is next to the definition of the opcodes.
4239
4240 (c) Defined a table of op-code lengths for simpler skipping along compiled
4241 code. This is again a macro in internal.h so that it is next to the
4242 definition of the opcodes.
4243
4244 34. Added support for recursive calls to individual subpatterns, along the
4245 lines of Robin Houston's patch (but implemented somewhat differently).
4246
4247 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4248 allow it to read and process whole directories in Win32. This code was
4249 contributed by Lionel Fourquaux; it has not been tested by me.
4250
4251 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4252 used to name a group. Names consist of alphanumerics and underscores, and must
4253 be unique. Back references use the syntax (?P=name) and recursive calls use
4254 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4255 numbers. The function pcre_fullinfo() can be used after compilation to extract
4256 a name/number map. There are three relevant calls:
4257
4258 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4259 PCRE_INFO_NAMECOUNT yields the number of entries
4260 PCRE_INFO_NAMETABLE yields a pointer to the map.
4261
4262 The map is a vector of fixed-size entries. The size of each entry depends on
4263 the length of the longest name used. The first two bytes of each entry are the
4264 group number, most significant byte first. There follows the corresponding
4265 name, zero terminated. The names are in alphabetical order.
4266
4267 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4268 case instead of 255. Making it the same both with and without UTF-8 support
4269 means that the same test output works with both.
4270
4271 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4272 calling malloc() with a zero argument.
4273
4274 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4275 optimization. I've improved things by keeping a bitmap of backreferences with
4276 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4277 fact referenced, the optimization can be applied. It is unlikely that a
4278 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4279 the match to follow \n) will appear inside brackets with a number greater than
4280 31, but if it does, any back reference > 31 suppresses the optimization.
4281
4282 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4283 of disabling numbered capturing parentheses. Any opening parenthesis that is
4284 not followed by ? behaves as if it were followed by ?: but named parentheses
4285 can still be used for capturing (and they will acquire numbers in the usual
4286 way).
4287
4288 41. Redesigned the return codes from the match() function into yes/no/error so
4289 that errors can be passed back from deep inside the nested calls. A malloc
4290 failure while inside a recursive subpattern call now causes the
4291 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4292
4293 42. It is now possible to set a limit on the number of times the match()
4294 function is called in a call to pcre_exec(). This facility makes it possible to
4295 limit the amount of recursion and backtracking, though not in a directly
4296 obvious way, because the match() function is used in a number of different
4297 circumstances. The count starts from zero for each position in the subject
4298 string (for non-anchored patterns). The default limit is, for compatibility, a
4299 large number, namely 10 000 000. You can change this in two ways:
4300
4301 (a) When configuring PCRE before making, you can use --with-match-limit=n
4302 to set a default value for the compiled library.
4303
4304 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4305 a different value is set. See 45 below.
4306
4307 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4308
4309 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4310 of things that can be changed at compile time. The first argument specifies
4311 what is wanted and the second points to where the information is to be placed.
4312 The current list of available information is:
4313
4314 PCRE_CONFIG_UTF8
4315
4316 The output is an integer that is set to one if UTF-8 support is available;
4317 otherwise it is set to zero.
4318
4319 PCRE_CONFIG_NEWLINE
4320
4321 The output is an integer that it set to the value of the code that is used for
4322 newline. It is either LF (10) or CR (13).
4323
4324 PCRE_CONFIG_LINK_SIZE
4325
4326 The output is an integer that contains the number of bytes used for internal
4327 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4328
4329 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4330
4331 The output is an integer that contains the threshold above which the POSIX
4332 interface uses malloc() for output vectors. See item 31 above.
4333
4334 PCRE_CONFIG_MATCH_LIMIT
4335
4336 The output is an unsigned integer that contains the default limit of the number
4337 of match() calls in a pcre_exec() execution. See 42 above.
4338
4339 44. pcretest has been upgraded by the addition of the -C option. This causes it
4340 to extract all the available output from the new pcre_config() function, and to
4341 output it. The program then exits immediately.
4342
4343 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4344 order to support additional features. One way would have been to define
4345 pcre_exec2() (for example) with extra arguments, but this would not have been
4346 extensible, and would also have required all calls to the original function to
4347 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4348 is used for passing in "extra" data from pcre_study().
4349
4350 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4351 contains the following fields:
4352
4353 flags a bitmap indicating which of the following fields are set
4354 study_data opaque data from pcre_study()
4355 match_limit a way of specifying a limit on match() calls for a specific
4356 call to pcre_exec()
4357 callout_data data for callouts (see 49 below)
4358
4359 The flag bits are also defined in pcre.h, and are
4360
4361 PCRE_EXTRA_STUDY_DATA
4362 PCRE_EXTRA_MATCH_LIMIT
4363 PCRE_EXTRA_CALLOUT_DATA
4364
4365 The pcre_study() function now returns one of these new pcre_extra blocks, with
4366 the actual study data pointed to by the study_data field, and the
4367 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4368 before. That is, this change is entirely upwards-compatible and requires no
4369 change to existing code.
4370
4371 If you want to pass in additional data to pcre_exec(), you can either place it
4372 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4373 block.
4374
4375 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4376 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4377 times with different match limits, until it finds the minimum value needed for
4378 pcre_exec() to complete. The value is then output. This can be instructive; for
4379 most simple matches the number is quite small, but for pathological cases it
4380 gets very large very quickly.
4381
4382 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4383 returns the size of the data block pointed to by the study_data field in a
4384 pcre_extra block, that is, the value that was passed as the argument to
4385 pcre_malloc() when PCRE was getting memory in which to place the information
4386 created by pcre_study(). The fourth argument should point to a size_t variable.
4387 pcretest has been extended so that this information is shown after a successful
4388 pcre_study() call when information about the compiled regex is being displayed.
4389
4390 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4391 because what follows is always an absolute path. (Later: it turns out that this
4392 is more than cosmetic for MinGW, because it doesn't like empty path
4393 components.)
4394
4395 49. Some changes have been made to the callout feature (see 28 above):
4396
4397 (i) A callout function now has three choices for what it returns:
4398
4399 0 => success, carry on matching
4400 > 0 => failure at this point, but backtrack if possible
4401 < 0 => serious error, return this value from pcre_exec()
4402
4403 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4404 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4405 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4406 use by callout functions. It will never be used by PCRE itself.
4407
4408 (ii) The pcre_extra structure (see 45 above) has a void * field called
4409 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4410 pcre_callout_block structure has a field of the same name. The contents of
4411 the field passed in the pcre_extra structure are passed to the callout
4412 function in the corresponding field in the callout block. This makes it
4413 easier to use the same callout-containing regex from multiple threads. For
4414 testing, the pcretest program has a new data escape
4415
4416 \C*n pass the number n (may be negative) as callout_data
4417
4418 If the callout function in pcretest receives a non-zero value as
4419 callout_data, it returns that value.
4420
4421 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4422 there were some redundant $(CFLAGS) in commands that are now specified as
4423 $(LINK), which already includes $(CFLAGS).
4424
4425 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4426 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4427 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4428 one-byte characters throughout. Note that case-insensitive matching applies
4429 only to characters whose values are less than 256. PCRE doesn't support the
4430 notion of cases for higher-valued characters.
4431
4432 (i) A character class whose characters are all within 0-255 is handled as
4433 a bit map, and the map is inverted for negative classes. Previously, a
4434 character > 255 always failed to match such a class; however it should
4435 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4436
4437 (ii) A negated character class with a single character < 255 is coded as
4438 "not this character" (OP_NOT). This wasn't working properly when the test
4439 character was multibyte, either singly or repeated.
4440
4441 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4442 mode, for example: \x{100}{2,3}.
4443
4444 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4445 singly or repeated) now correctly test multibyte characters. However,
4446 PCRE doesn't recognize any characters with values greater than 255 as
4447 digits, spaces, or word characters. Such characters always match \D, \S,
4448 and \W, and never match \d, \s, or \w.
4449
4450 (v) Classes may now contain characters and character ranges with values
4451 greater than 255. For example: [ab\x{100}-\x{400}].
4452
4453 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4454 PCRE in UTF-8 mode.
4455
4456 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4457 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4458 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4459 value.)
4460
4461 53. The single man page has become too large. I have therefore split it up into
4462 a number of separate man pages. These also give rise to individual HTML pages;
4463 these are now put in a separate directory, and there is an index.html page that
4464 lists them all. Some hyperlinking between the pages has been installed.
4465
4466 54. Added convenience functions for handling named capturing parentheses.
4467
4468 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4469 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4470 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4471 are faulted.
4472
4473 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4474 calling configure. These values are used when compiling the dftables.c program
4475 which is run to generate the source of the default character tables. They
4476 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4477 you will need to set these values.
4478
4479 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4480
4481
4482 Version 3.9 02-Jan-02
4483 ---------------------
4484
4485 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4486
4487 2. If --disable-static was given, the building process failed when trying to
4488 build pcretest and pcregrep. (For some reason it was using libtool to compile
4489 them, which is not right, as they aren't part of the library.)
4490
4491
4492 Version 3.8 18-Dec-01
4493 ---------------------
4494
4495 1. The experimental UTF-8 code was completely screwed up. It was packing the
4496 bytes in the wrong order. How dumb can you get?
4497
4498
4499 Version 3.7 29-Oct-01
4500 ---------------------
4501
4502 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4503 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4504 this didn't happen under Solaris 8, where I normally test things.
4505
4506 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4507 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4508 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4509 this fix an hour or so after the initial 3.7 release.)
4510
4511
4512 Version 3.6 23-Oct-01
4513 ---------------------
4514
4515 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4516 offsets passed as NULL with zero offset count.
4517
4518 2. The config.guess and config.sub files had not been updated when I moved to
4519 the latest autoconf.
4520
4521
4522 Version 3.5 15-Aug-01
4523 ---------------------
4524
4525 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4526 had been forgotten.
4527
4528 2. By using declared but undefined structures, we can avoid using "void"
4529 definitions in pcre.h while keeping the internal definitions of the structures
4530 private.
4531
4532 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4533 user point of view, this means that both static and shared libraries are built
4534 by default, but this can be individually controlled. More of the work of
4535 handling this static/shared cases is now inside libtool instead of PCRE's make
4536 file.
4537
4538 4. The pcretest utility is now installed along with pcregrep because it is
4539 useful for users (to test regexs) and by doing this, it automatically gets
4540 relinked by libtool. The documentation has been turned into a man page, so
4541 there are now .1, .txt, and .html versions in /doc.
4542
4543 5. Upgrades to pcregrep:
4544 (i) Added long-form option names like gnu grep.
4545 (ii) Added --help to list all options with an explanatory phrase.
4546 (iii) Added -r, --recursive to recurse into sub-directories.
4547 (iv) Added -f, --file to read patterns from a file.
4548
4549 6. pcre_exec() was referring to its "code" argument before testing that
4550 argument for NULL (and giving an error if it was NULL).
4551
4552 7. Upgraded Makefile.in to allow for compiling in a different directory from
4553 the source directory.
4554
4555 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4556 options bits, the pointer it was passed was to an int instead of to an unsigned
4557 long int. This mattered only on 64-bit systems.
4558
4559 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4560 generated) instead of pcre.in, which it its source. Also made the same change
4561 in several of the .c files.
4562
4563 10. A new release of gcc defines printf() as a macro, which broke pcretest
4564 because it had an ifdef in the middle of a string argument for printf(). Fixed
4565 by using separate calls to printf().
4566
4567 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4568 script, to force use of CR or LF instead of \n in the source. On non-Unix
4569 systems, the value can be set in config.h.
4570
4571 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4572 absolute limit. Changed the text of the error message to make this clear, and
4573 likewise updated the man page.
4574
4575 13. The limit of 99 on the number of capturing subpatterns has been removed.
4576 The new limit is 65535, which I hope will not be a "real" limit.
4577
4578
4579 Version 3.4 22-Aug-00
4580 ---------------------
4581
4582 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4583
4584 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4585
4586
4587 Version 3.3 01-Aug-00
4588 ---------------------
4589
4590 1. If an octal character was given, but the value was greater than \377, it
4591 was not getting masked to the least significant bits, as documented. This could
4592 lead to crashes in some systems.
4593
4594 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4595 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4596
4597 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4598 These just pass their arguments on to (pcre_free)(), but they are provided
4599 because some uses of PCRE bind it to non-C systems that can call its functions,
4600 but cannot call free() or pcre_free() directly.
4601
4602 4. Add "make test" as a synonym for "make check". Corrected some comments in
4603 the Makefile.
4604
4605 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4606 Makefile.
4607
4608 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4609 command called pgrep for grepping around the active processes.
4610
4611 7. Added the beginnings of support for UTF-8 character strings.
4612
4613 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4614 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4615 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4616 out for the ar command.)
4617
4618
4619 Version 3.2 12-May-00
4620 ---------------------
4621
4622 This is purely a bug fixing release.
4623
4624 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4625 of ZA. This was just one example of several cases that could provoke this bug,
4626 which was introduced by change 9 of version 2.00. The code for breaking
4627 infinite loops after an iteration that matches an empty string was't working
4628 correctly.
4629
4630 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4631 when matched against abbab (for example). After matching an empty string, it
4632 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4633 caused it to match further down the string than it should.
4634
4635 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4636 was there because it doesn't seem to be needed, and it causes trouble on some
4637 systems, as it is not a Standard C header. It has been removed.
4638
4639 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4640 were reported on the Macintosh. The changes were from
4641
4642 while ((c = *(++ptr)) != 0 && c != '\n');
4643 to
4644 while ((c = *(++ptr)) != 0 && c != '\n') ;
4645
4646 Totally extraordinary, but if that's what it takes...
4647
4648 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4649 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4650 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4651 assumes the way PCRE uses memmove() (always moving upwards).
4652
4653 6. PCRE is being used in one environment where strchr() is not available. There
4654 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4655 faster code anyway.
4656
4657
4658 Version 3.1 09-Feb-00
4659 ---------------------
4660
4661 The only change in this release is the fixing of some bugs in Makefile.in for
4662 the "install" target:
4663
4664 (1) It was failing to install pcreposix.h.
4665
4666 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4667
4668
4669 Version 3.0 01-Feb-00
4670 ---------------------
4671
4672 1. Add support for the /+ modifier to perltest (to output $` like it does in
4673 pcretest).
4674
4675 2. Add support for the /g modifier to perltest.
4676
4677 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4678 matches null strings.
4679
4680 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4681 pattern. Perl treats empty patterns specially - it reuses the most recent
4682 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4683 effect.
4684
4685 5. The POSIX interface was broken in that it was just handing over the POSIX
4686 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4687 required a bigger vector, with some working space on the end. This means that
4688 the POSIX wrapper now has to get and free some memory, and copy the results.
4689
4690 6. Added some simple autoconf support, placing the test data and the
4691 documentation in separate directories, re-organizing some of the
4692 information files, and making it build pcre-config (a GNU standard). Also added
4693 libtool support for building PCRE as a shared library, which is now the
4694 default.
4695
4696 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4697 09 are not valid octal constants. Single digits will be used for minor values
4698 less than 10.
4699
4700 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4701 existing programs that set these in the POSIX interface can use PCRE without
4702 modification.
4703
4704 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4705 return all that pcre_info() returns, plus additional data. The pcre_info()
4706 function is retained for compatibility, but is considered to be obsolete.
4707
4708 10. Added experimental recursion feature (?R) to handle one common case that
4709 Perl 5.6 will be able to do with (?p{...}).
4710
4711 11. Added support for POSIX character classes like [:alpha:], which Perl is
4712 adopting.
4713
4714
4715 Version 2.08 31-Aug-99
4716 ----------------------
4717
4718 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4719 trying to match at the startoffset position, but instead was moving forward to
4720 the next newline as if a previous match had failed.
4721
4722 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4723 and could get into a loop if a null string was matched other than at the start
4724 of the subject.
4725
4726 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4727 be distinguished at compile time, and for completeness also added PCRE_DATE.
4728
4729 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4730 in GnuWin32 environments.
4731
4732
4733 Version 2.07 29-Jul-99
4734 ----------------------
4735
4736 1. The documentation is now supplied in plain text form and HTML as well as in
4737 the form of man page sources.
4738
4739 2. C++ compilers don't like assigning (void *) values to other pointer types.
4740 In particular this affects malloc(). Although there is no problem in Standard
4741 C, I've put in casts to keep C++ compilers happy.
4742
4743 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4744 should be (const char *).
4745
4746 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4747 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4748 However, I haven't made this a standard facility. The documentation doesn't
4749 mention it, and the Makefile doesn't support it.
4750
4751 5. The Makefile now contains an "install" target, with editable destinations at
4752 the top of the file. The pcretest program is not installed.
4753
4754 6. pgrep -V now gives the PCRE version number and date.
4755
4756 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4757 causing the entire string to be ignored, instead of just the last character.
4758
4759 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4760 non-matching string, it can take a very, very long time, even for strings of
4761 quite modest length, because of the nested recursion. PCRE now does better in
4762 some of these cases. It does this by remembering the last required literal
4763 character in the pattern, and pre-searching the subject to ensure it is present
4764 before running the real match. In other words, it applies a heuristic to detect
4765 some types of certain failure quickly, and in the above example, if presented
4766 with a string that has no trailing " it gives "no match" very quickly.
4767
4768 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4769 other alternatives are tried instead.
4770
4771
4772 Version 2.06 09-Jun-99
4773 ----------------------
4774
4775 1. Change pcretest's output for amount of store used to show just the code
4776 space, because the remainder (the data block) varies in size between 32-bit and
4777 64-bit systems.
4778
4779 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4780 start matching at. This allows lookbehinds to work when searching for multiple
4781 occurrences in a string.
4782
4783 3. Added additional options to pcretest for testing multiple occurrences:
4784
4785 /+ outputs the rest of the string that follows a match
4786 /g loops for multiple occurrences, using the new startoffset argument
4787 /G loops for multiple occurrences by passing an incremented pointer
4788
4789 4. PCRE wasn't doing the "first character" optimization for patterns starting
4790 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4791 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4792 the letter 'x'. On long subject strings, this gives a significant speed-up.
4793
4794
4795 Version 2.05 21-Apr-99
4796 ----------------------
4797
4798 1. Changed the type of magic_number from int to long int so that it works
4799 properly on 16-bit systems.
4800
4801 2. Fixed a bug which caused patterns starting with .* not to work correctly
4802 when the subject string contained newline characters. PCRE was assuming
4803 anchoring for such patterns in all cases, which is not correct because .* will
4804 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4805 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4806 must be retried after every newline in the subject.
4807
4808
4809 Version 2.04 18-Feb-99
4810 ----------------------
4811
4812 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4813 computation of the store needed to hold the pattern was incorrect (too large).
4814 If such patterns were nested a few deep, this could multiply and become a real
4815 problem.
4816
4817 2. Added /M option to pcretest to show the memory requirement of a specific
4818 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4819
4820 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4821 compiled in such a way that the backtracking after subsequent failure was
4822 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4823 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4824
4825
4826 Version 2.03 02-Feb-99
4827 ----------------------
4828
4829 1. Fixed typo and small mistake in man page.
4830
4831 2. Added 4th condition (GPL supersedes if conflict) and created separate
4832 LICENCE file containing the conditions.
4833
4834 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4835 Perl, that is the internal \ allows the delimiter to be included in the
4836 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4837 the final delimiter, add \ to the end of the pattern (to test the error).
4838
4839 4. Added the convenience functions for extracting substrings after a successful
4840 match. Updated pcretest to make it able to test these functions.
4841
4842
4843 Version 2.02 14-Jan-99
4844 ----------------------
4845
4846 1. Initialized the working variables associated with each extraction so that
4847 their saving and restoring doesn't refer to uninitialized store.
4848
4849 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4850 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4851 fix the problem.
4852
4853 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4854 calls, and wasn't printing the correct value for compiling calls. Increased the
4855 default value of LOOPREPEAT, and the number of significant figures in the
4856 times.
4857
4858 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4859
4860 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4861 a building problem on Windows NT with a FAT file system.
4862
4863
4864 Version 2.01 21-Oct-98
4865 ----------------------
4866
4867 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4868 to character tables built by pcre_maketables() in the current locale. If NULL
4869 is passed, the default tables are used.
4870
4871
4872 Version 2.00 24-Sep-98
4873 ----------------------
4874
4875 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4876 it any more.
4877
4878 2. Allow quantification of (?>) groups, and make it work correctly.
4879
4880 3. The first character computation wasn't working for (?>) groups.
4881
4882 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4883 end of the subject) and add 5.005's \z, which really does match only at the
4884 very end of the subject.
4885
4886 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4887
4888 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4889 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4890 localized options. All options to pcre_study() were also removed.
4891
4892 7. Add other new features from 5.005:
4893
4894 $(?<= positive lookbehind
4895 $(?<! negative lookbehind
4896 (?imsx-imsx) added the unsetting capability
4897 such a setting is global if at outer level; local otherwise
4898 (?imsx-imsx:) non-capturing groups with option setting
4899 (?(cond)re|re) conditional pattern matching
4900
4901 A backreference to itself in a repeated group matches the previous
4902 captured string.
4903
4904 8. General tidying up of studying (both automatic and via "study")
4905 consequential on the addition of new assertions.
4906
4907 9. As in 5.005, unlimited repeated groups that could match an empty substring
4908 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4909 runtime if any iteration does actually match an empty substring.
4910
4911 10. Include the RunTest script in the distribution.
4912
4913 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4914 discrepancies, some of which were old and were also with respect to 5.004. They
4915 have now been fixed.
4916
4917
4918 Version 1.09 28-Apr-98
4919 ----------------------
4920
4921 1. A negated single character class followed by a quantifier with a minimum
4922 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4923 program crashes, or just wrong answers. This did not apply to negated classes
4924 containing more than one character, or to minima other than one.
4925
4926
4927 Version 1.08 27-Mar-98
4928 ----------------------
4929
4930 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4931
4932 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4933 latter must appear before anything that relies on it in the pattern.
4934
4935
4936 Version 1.07 16-Feb-98
4937 ----------------------
4938
4939 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4940 repeat of a potentially empty string).
4941
4942
4943 Version 1.06 23-Jan-98
4944 ----------------------
4945
4946 1. Added Markus Oberhumer's little patches for C++.
4947
4948 2. Literal strings longer than 255 characters were broken.
4949
4950
4951 Version 1.05 23-Dec-97
4952 ----------------------
4953
4954 1. Negated character classes containing more than one character were failing if
4955 PCRE_CASELESS was set at run time.
4956
4957
4958 Version 1.04 19-Dec-97
4959 ----------------------
4960
4961 1. Corrected the man page, where some "const" qualifiers had been omitted.
4962
4963 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4964 input syntax.
4965
4966 3. Fixed memory leak which occurred when a regex with back references was
4967 matched with an offsets vector that wasn't big enough. The temporary memory
4968 that is used in this case wasn't being freed if the match failed.
4969
4970 4. Tidied pcretest to ensure it frees memory that it gets.
4971
4972 5. Temporary memory was being obtained in the case where the passed offsets
4973 vector was exactly big enough.
4974
4975 6. Corrected definition of offsetof() from change 5 below.
4976
4977 7. I had screwed up change 6 below and broken the rules for the use of
4978 setjmp(). Now fixed.
4979
4980
4981 Version 1.03 18-Dec-97
4982 ----------------------
4983
4984 1. A erroneous regex with a missing opening parenthesis was correctly
4985 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4986 on some systems.
4987
4988 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4989 it was reported that one broken compiler failed on the former because "code" is
4990 also an independent variable.
4991
4992 3. The erroneous regex a[]b caused an array overrun reference.
4993
4994 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4995 fail on data ending with that character. (It was going on too far, and checking
4996 the next character, typically a binary zero.) This was specific to the
4997 optimized code for single-character negative classes.
4998
4999 5. Added a contributed patch from the TIN world which does the following:
5000
5001 + Add an undef for memmove, in case the the system defines a macro for it.
5002
5003 + Add a definition of offsetof(), in case there isn't one. (I don't know
5004 the reason behind this - offsetof() is part of the ANSI standard - but
5005 it does no harm).
5006
5007 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5008 most of the places where whitespace preceded '#'. I have given up and
5009 allowed the remaining 2 cases to be at the margin.
5010
5011 + Rename some variables in pcre to eliminate shadowing. This seems very
5012 pedantic, but does no harm, of course.
5013
5014 6. Moved the call to setjmp() into its own function, to get rid of warnings
5015 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5016
5017 7. Constructs such as \d{8,} were compiling into the equivalent of
5018 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5019 outcome, but in this particular case used more store than had been allocated,
5020 which caused the bug to be discovered because it threw up an internal error.
5021
5022 8. The debugging code in both pcre and pcretest for outputting the compiled
5023 form of a regex was going wrong in the case of back references followed by
5024 curly-bracketed repeats.
5025
5026
5027 Version 1.02 12-Dec-97
5028 ----------------------
5029
5030 1. Typos in pcre.3 and comments in the source fixed.
5031
5032 2. Applied a contributed patch to get rid of places where it used to remove
5033 'const' from variables, and fixed some signed/unsigned and uninitialized
5034 variable warnings.
5035
5036 3. Added the "runtest" target to Makefile.
5037
5038 4. Set default compiler flag to -O2 rather than just -O.
5039
5040
5041 Version 1.01 19-Nov-97
5042 ----------------------
5043
5044 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5045 like /([ab]*)*/, that is, for classes with more than one character in them.
5046
5047 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5048 as /((?>a*))*/ (a PCRE_EXTRA facility).
5049
5050
5051 Version 1.00 18-Nov-97
5052 ----------------------
5053
5054 1. Added compile-time macros to support systems such as SunOS4 which don't have
5055 memmove() or strerror() but have other things that can be used instead.
5056
5057 2. Arranged that "make clean" removes the executables.
5058
5059
5060 Version 0.99 27-Oct-97
5061 ----------------------
5062
5063 1. Fixed bug in code for optimizing classes with only one character. It was
5064 initializing a 32-byte map regardless, which could cause it to run off the end
5065 of the memory it had got.
5066
5067 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5068
5069
5070 Version 0.98 22-Oct-97
5071 ----------------------
5072
5073 1. Fixed bug in code for handling temporary memory usage when there are more
5074 back references than supplied space in the ovector. This could cause segfaults.
5075
5076
5077 Version 0.97 21-Oct-97
5078 ----------------------
5079
5080 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5081
5082 2. Optimized negated single characters not to use a bit map.
5083
5084 3. Brought error texts together as macro definitions; clarified some of them;
5085 fixed one that was wrong - it said "range out of order" when it meant "invalid
5086 escape sequence".
5087
5088 4. Changed some char * arguments to const char *.
5089
5090 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5091
5092 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5093 pcretest.
5094
5095
5096 Version 0.96 16-Oct-97
5097 ----------------------
5098
5099 1. Added a simple "pgrep" utility to the distribution.
5100
5101 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5102 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5103 where "ddd" means "one or more decimal digits".
5104
5105 3. Fixed serious bug. If a pattern had a back reference, but the call to
5106 pcre_exec() didn't supply a large enough ovector to record the related
5107 identifying subpattern, the match always failed. PCRE now remembers the number
5108 of the largest back reference, and gets some temporary memory in which to save
5109 the offsets during matching if necessary, in order to ensure that
5110 backreferences always work.
5111
5112 4. Increased the compatibility with Perl in a number of ways:
5113
5114 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5115 to request this handling. The option can be set at compile or exec time.
5116
5117 (b) $ matches before a terminating newline by default; an option
5118 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5119 mode). The option can be set at compile or exec time.
5120
5121 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5122 the same as Perl's. If the decimal number it represents is less than 10
5123 or there aren't that many previous left capturing parentheses, an octal
5124 escape is read. Inside a character class, it's always an octal escape,
5125 even if it is a single digit.
5126
5127 (d) An escaped but undefined alphabetic character is taken as a literal,
5128 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5129 escapes.
5130
5131 (e) {0} is now permitted. (The previous item is removed from the compiled
5132 pattern).
5133
5134 5. Changed all the names of code files so that the basic parts are no longer
5135 than 10 characters, and abolished the teeny "globals.c" file.
5136
5137 6. Changed the handling of character classes; they are now done with a 32-byte
5138 bit map always.
5139
5140 7. Added the -d and /D options to pcretest to make it possible to look at the
5141 internals of compilation without having to recompile pcre.
5142
5143
5144 Version 0.95 23-Sep-97
5145 ----------------------
5146
5147 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5148 \x20 at the start of a run of normal characters. These were being treated as
5149 real characters, instead of the source characters being re-checked.
5150
5151
5152 Version 0.94 18-Sep-97
5153 ----------------------
5154
5155 1. The functions are now thread-safe, with the caveat that the global variables
5156 containing pointers to malloc() and free() or alternative functions are the
5157 same for all threads.
5158
5159 2. Get pcre_study() to generate a bitmap of initial characters for non-
5160 anchored patterns when this is possible, and use it if passed to pcre_exec().
5161
5162
5163 Version 0.93 15-Sep-97
5164 ----------------------
5165
5166 1. /(b)|(:+)/ was computing an incorrect first character.
5167
5168 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5169 but not actually doing anything yet.
5170
5171 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5172 as Perl does (e.g. [-az] or [az-]).
5173
5174 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5175 all possible positions.
5176
5177 5. Split up into different modules to avoid including unneeded functions in a
5178 compiled binary. However, compile and exec are still in one module. The "study"
5179 function is split off.
5180
5181 6. The character tables are now in a separate module whose source is generated
5182 by an auxiliary program - but can then be edited by hand if required. There are
5183 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5184 toupper() in the code.
5185
5186 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5187 make them global. Abolish the function for setting them, as the caller can now
5188 set them directly.
5189
5190
5191 Version 0.92 11-Sep-97
5192 ----------------------
5193
5194 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5195 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5196
5197 2. Caseless matching was not working in character classes if the characters in
5198 the pattern were in upper case.
5199
5200 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5201
5202 4. Make PCRE_ANCHORED public and accept as a compile option.
5203
5204 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5205 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5206 pass them.
5207
5208 6. Give an error if bad option bits passed at compile or run time.
5209
5210 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5211 pcretest to cause it to pass that flag.
5212
5213 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5214 options, and the first character, if set.
5215
5216 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5217
5218
5219 Version 0.91 10-Sep-97
5220 ----------------------
5221
5222 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5223 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5224
5225 2. PCRE was looping on encountering an indefinitely repeated back reference to
5226 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5227 Perl does - treats the match as successful.
5228
5229 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5