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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1257 - (show annotations)
Fri Feb 22 20:20:30 2013 UTC (6 years, 7 months ago) by ph10
File size: 243227 byte(s)
Error occurred while calculating annotation data.
Change some pcre_uchar variables to pcre_uint32 in pcre_exec.c, for better 
performance.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.33 xx-xxxx-201x
5 -------------------------
6
7 1. Added 'U' to some constants that are compared to unsigned integers, to
8 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
9 variables that are added to signed variables, to ensure the result is
10 signed and can be negated.
11
12 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
13 CMake config files.
14
15 3. Revise the creation of config.h.generic so that all boolean macros are
16 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
17 overriding via -D on the command line possible.
18
19 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
20 to unsigned int is reported to make a quite noticeable speed difference in
21 a specific Windows environment. Testing on Linux did also appear to show
22 some benefit (and it is clearly not harmful). Also fixed the definition of
23 Xop which should be unsigned.
24
25 5. Related to (4), changing the definition of the intermediate variable cc
26 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
27 improvements.
28
29 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
30 unnecessary spaces.
31
32 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
33 and later.
34
35 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
36
37 9. Optimizing fast_forward_start_bits in JIT.
38
39 10. Adding experimental support for callouts in JIT, and fixing some
40 issues revealed during this work. Namely:
41
42 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
43
44 (b) Minimum length was not checked before the matching is started.
45
46 11. The value of capture_last that is passed to callouts was incorrect in some
47 cases when there was a capture on one path that was subsequently abandoned
48 after a backtrack. Also, the capture_last value is now reset after a
49 recursion, since all captures are also reset in this case.
50
51 12. The interpreter no longer returns the "too many substrings" error in the
52 case when an overflowing capture is in a branch that is subsequently
53 abandoned after a backtrack.
54
55 13. In the pathological case when an offset vector of size 2 is used, pcretest
56 now prints out the matched string after a yield of 0 or 1.
57
58 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
59 Only supported by the JIT compiler at the moment.
60
61 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
62
63 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
64 offset of the starting point of the matching process, provided the offsets
65 vector is large enough.
66
67 17. The \A escape now records a lookbehind value of 1, though its execution
68 does not actually inspect the previous character. This is to ensure that,
69 in partial multi-segment matching, at least one character from the old
70 segment is retained when a new segment is processed. Otherwise, if there
71 are no lookbehinds in the pattern, \A might match incorrectly at the start
72 of a new segment.
73
74 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
75
76 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
77 gives some modest performance improvement in 8-bit mode.
78
79
80 Version 8.32 30-November-2012
81 -----------------------------
82
83 1. Improved JIT compiler optimizations for first character search and single
84 character iterators.
85
86 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
87 Patch by Daniel Richard G.
88
89 3. Single character iterator optimizations in the JIT compiler.
90
91 4. Improved JIT compiler optimizations for character ranges.
92
93 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
94 Reported by Giuseppe D'Angelo.
95
96 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
97 of a line, was being set incorrectly in cases where .* appeared inside
98 atomic brackets at the start of a pattern, or where there was a subsequent
99 *PRUNE or *SKIP.
100
101 7. Improved instruction cache flush for POWER/PowerPC.
102 Patch by Daniel Richard G.
103
104 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
105 grep:
106
107 (a) There is now no limit to the number of patterns to be matched.
108
109 (b) An error is given if a pattern is too long.
110
111 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
112 are now supported.
113
114 (d) --exclude-from and --include-from (multiple use) have been added.
115
116 (e) Exclusions and inclusions now apply to all files and directories, not
117 just to those obtained from scanning a directory recursively.
118
119 (f) Multiple uses of -f and --file-list are now supported.
120
121 (g) In a Windows environment, the default for -d has been changed from
122 "read" (the GNU grep default) to "skip", because otherwise the presence
123 of a directory in the file list provokes an error.
124
125 (h) The documentation has been revised and clarified in places.
126
127 9. Improve the matching speed of capturing brackets.
128
129 10. Changed the meaning of \X so that it now matches a Unicode extended
130 grapheme cluster.
131
132 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
133 out POSIX threads when JIT support is configured.
134
135 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
136
137 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
138 structure could go wrong in environments where size_t is not the same size
139 as int.
140
141 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
142 set.
143
144 15. The EBCDIC support had decayed; later updates to the code had included
145 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
146 general tidy up of EBCDIC-related issues, and the documentation was also
147 not quite right. There is now a test that can be run on ASCII systems to
148 check some of the EBCDIC-related things (but is it not a full test).
149
150 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
151 in a small tidy to the code.
152
153 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
154
155 18. If the --only-matching (-o) option in pcregrep is specified multiple
156 times, each one causes appropriate output. For example, -o1 -o2 outputs the
157 substrings matched by the 1st and 2nd capturing parentheses. A separating
158 string can be specified by --om-separator (default empty).
159
160 19. Improving the first n character searches.
161
162 20. Turn case lists for horizontal and vertical white space into macros so that
163 they are defined only once.
164
165 21. This set of changes together give more compatible Unicode case-folding
166 behaviour for characters that have more than one other case when UCP
167 support is available.
168
169 (a) The Unicode property table now has offsets into a new table of sets of
170 three or more characters that are case-equivalent. The MultiStage2.py
171 script that generates these tables (the pcre_ucd.c file) now scans
172 CaseFolding.txt instead of UnicodeData.txt for character case
173 information.
174
175 (b) The code for adding characters or ranges of characters to a character
176 class has been abstracted into a generalized function that also handles
177 case-independence. In UTF-mode with UCP support, this uses the new data
178 to handle characters with more than one other case.
179
180 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
181 whose other case is greater than 256 are now correctly matched
182 caselessly. Previously, the high codepoint matched the low one, but not
183 vice versa.
184
185 (d) The processing of \h, \H, \v, and \ in character classes now makes use
186 of the new class addition function, using character lists defined as
187 macros alongside the case definitions of 20 above.
188
189 (e) Caseless back references now work with characters that have more than
190 one other case.
191
192 (f) General caseless matching of characters with more than one other case
193 is supported.
194
195 22. Unicode character properties were updated from Unicode 6.2.0
196
197 23. Improved CMake support under Windows. Patch by Daniel Richard G.
198
199 24. Add support for 32-bit character strings, and UTF-32
200
201 25. Major JIT compiler update (code refactoring and bugfixing).
202 Experimental Sparc 32 support is added.
203
204 26. Applied a modified version of Daniel Richard G's patch to create
205 pcre.h.generic and config.h.generic by "make" instead of in the
206 PrepareRelease script.
207
208 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
209 pcre_compile.c when checking for a zero character.
210
211 28. Introducing a native interface for JIT. Through this interface, the compiled
212 machine code can be directly executed. The purpose of this interface is to
213 provide fast pattern matching, so several sanity checks are not performed.
214 However, feature tests are still performed. The new interface provides
215 1.4x speedup compared to the old one.
216
217 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
218 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
219 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
220
221 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
222 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
223 the "old" RFC 2279). Instead, it ended up passing a negative length to
224 pcre_exec().
225
226 31. Add support for GCC's visibility feature to hide internal functions.
227
228 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
229 "unknown -C option" after outputting 0 or 1.
230
231 33. There is now support for generating a code coverage report for the test
232 suite in environments where gcc is the compiler and lcov is installed. This
233 is mainly for the benefit of the developers.
234
235 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
236 unaddressable using valgrind annotations, allowing valgrind to detect
237 invalid memory accesses. This is mainly for the benefit of the developers.
238
239 25. (*UTF) can now be used to start a pattern in any of the three libraries.
240
241 26. Give configure error if --enable-cpp but no C++ compiler found.
242
243
244 Version 8.31 06-July-2012
245 -------------------------
246
247 1. Fixing a wrong JIT test case and some compiler warnings.
248
249 2. Removed a bashism from the RunTest script.
250
251 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
252 to unsigned type, result still unsigned" that was given by an MS compiler
253 on encountering the code "-sizeof(xxx)".
254
255 4. Partial matching support is added to the JIT compiler.
256
257 5. Fixed several bugs concerned with partial matching of items that consist
258 of more than one character:
259
260 (a) /^(..)\1/ did not partially match "aba" because checking references was
261 done on an "all or nothing" basis. This also applied to repeated
262 references.
263
264 (b) \R did not give a hard partial match if \r was found at the end of the
265 subject.
266
267 (c) \X did not give a hard partial match after matching one or more
268 characters at the end of the subject.
269
270 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
271 a partial match for the string "\r".
272
273 (e) When newline was set to CRLF, the metacharacter "." did not recognize
274 a partial match for a CR character at the end of the subject string.
275
276 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
277 running pcretest, the text "(JIT)" added to the output whenever JIT is
278 actually used to run the match.
279
280 7. Individual JIT compile options can be set in pcretest by following -s+[+]
281 or /S+[+] with a digit between 1 and 7.
282
283 8. OP_NOT now supports any UTF character not just single-byte ones.
284
285 9. (*MARK) control verb is now supported by the JIT compiler.
286
287 10. The command "./RunTest list" lists the available tests without actually
288 running any of them. (Because I keep forgetting what they all are.)
289
290 11. Add PCRE_INFO_MAXLOOKBEHIND.
291
292 12. Applied a (slightly modified) user-supplied patch that improves performance
293 when the heap is used for recursion (compiled with --disable-stack-for-
294 recursion). Instead of malloc and free for each heap frame each time a
295 logical recursion happens, frames are retained on a chain and re-used where
296 possible. This sometimes gives as much as 30% improvement.
297
298 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
299 call.
300
301 14. As documented, (*COMMIT) is now confined to within a positive assertion.
302
303 15. It is now possible to link pcretest with libedit as an alternative to
304 libreadline.
305
306 16. (*COMMIT) control verb is now supported by the JIT compiler.
307
308 17. The Unicode data tables have been updated to Unicode 6.1.0.
309
310 18. Added --file-list option to pcregrep.
311
312 19. Added binary file support to pcregrep, including the -a, --binary-files,
313 -I, and --text options.
314
315 20. The madvise function is renamed for posix_madvise for QNX compatibility
316 reasons. Fixed by Giuseppe D'Angelo.
317
318 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
319
320 22. Changed the option for creating symbolic links for 16-bit man pages from
321 -s to -sf so that re-installing does not cause issues.
322
323 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
324
325 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
326 matches in certain environments (the workspace was not being correctly
327 retained). Also added to pcre_dfa_exec() a simple plausibility check on
328 some of the workspace data at the beginning of a restart.
329
330 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
331 was not doing so when it should - probably a typo introduced by SVN 528
332 (change 8.10/14).
333
334 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
335 \w+ when the character tables indicated that \x{c4} was a word character.
336 There were several related cases, all because the tests for doing a table
337 lookup were testing for characters less than 127 instead of 255.
338
339 27. If a pattern contains capturing parentheses that are not used in a match,
340 their slots in the ovector are set to -1. For those that are higher than
341 any matched groups, this happens at the end of processing. In the case when
342 there were back references that the ovector was too small to contain
343 (causing temporary malloc'd memory to be used during matching), and the
344 highest capturing number was not used, memory off the end of the ovector
345 was incorrectly being set to -1. (It was using the size of the temporary
346 memory instead of the true size.)
347
348 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
349 ovector size, it uses memory at the end of the block that it has got.
350
351 29. Check for an overlong MARK name and give an error at compile time. The
352 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
353
354 30. JIT compiler update.
355
356 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
357 Rill for the patch.
358
359 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
360
361 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
362
363 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
364 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
365 was enabled.
366
367 35. Improve JIT code generation for greedy plus quantifier.
368
369 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
370 1 to "aa" instead of to an empty string. The bug affected repeated groups
371 that could potentially match an empty string.
372
373 37. Optimizing single character iterators in JIT.
374
375 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
376 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
377 codepoints that are too big for the mode are faulted, and in a UTF mode,
378 disallowed codepoints are also faulted.
379
380 39. If PCRE was compiled with UTF support, in three places in the DFA
381 matcher there was code that should only have been obeyed in UTF mode, but
382 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
383 processing when bytes with values greater than 127 were present. In 16-bit
384 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
385 both cases the values are those that cannot be the first data item in a UTF
386 character. The three items that might have provoked this were recursions,
387 possessively repeated groups, and atomic groups.
388
389 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
390 and pcregrep, because some OS require shared objects to be explicitly
391 passed to ld, causing the link step to fail if they are not.
392
393 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
394 mode, patterns that started with \h* or \R* might be incorrectly matched.
395
396
397 Version 8.30 04-February-2012
398 -----------------------------
399
400 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
401 name is defined in ctype.h.
402
403 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
404 only in quite long subpatterns.
405
406 3. Removed the function pcre_info(), which has been obsolete and deprecated
407 since it was replaced by pcre_fullinfo() in February 2000.
408
409 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
410 match a (*MARK), and the match failed at the start of the subject, a
411 reference to memory before the start of the subject could occur. This bug
412 was introduced by fix 17 of release 8.21.
413
414 5. A reference to an unset group with zero minimum repetition was giving
415 totally wrong answers (in non-JavaScript-compatibility mode). For example,
416 /(another)?(\1?)test/ matched against "hello world test". This bug was
417 introduced in release 8.13.
418
419 6. Add support for 16-bit character strings (a large amount of work involving
420 many changes and refactorings).
421
422 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
423 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
424 from a file.
425
426 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
427 rounding is not applied in this particular case).
428
429 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
430 if they appear, or are escaped, in patterns.
431
432 10. Get rid of a number of -Wunused-but-set-variable warnings.
433
434 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
435 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
436 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
437 also returns the mark "x". This bug applied to capturing parentheses,
438 non-capturing parentheses, and atomic parentheses. It also applied to some
439 assertions.
440
441 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
442 information out of configure.ac instead of relying on pcre.h.generic, which
443 is not stored in the repository.
444
445 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
446 -lreadline.
447
448 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
449
450 15. Applied Graycode's patch to put the top-level frame on the stack rather
451 than the heap when not using the stack for recursion. This gives a
452 performance improvement in many cases when recursion is not deep.
453
454 16. Experimental code added to "pcretest -C" to output the stack frame size.
455
456
457 Version 8.21 12-Dec-2011
458 ------------------------
459
460 1. Updating the JIT compiler.
461
462 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
463 are added as well.
464
465 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
466 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
467 calling _pcre_jit_exec. Some extra comments are added.
468
469 4. (*MARK) settings inside atomic groups that do not contain any capturing
470 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
471 was introduced by change 18 for 8.20.
472
473 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
474 ECMA-262 standard.
475
476 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
477 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
478 This bug was probably introduced by change 9 of 8.13.
479
480 7. While fixing 6 above, I noticed that a number of other items were being
481 incorrectly rejected as "not fixed length". This arose partly because newer
482 opcodes had not been added to the fixed-length checking code. I have (a)
483 corrected the bug and added tests for these items, and (b) arranged for an
484 error to occur if an unknown opcode is encountered while checking for fixed
485 length instead of just assuming "not fixed length". The items that were
486 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
487 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
488 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
489
490 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
491 being incorrectly compiled and would have given unpredicatble results.
492
493 9. A possessively repeated subpattern with minimum repeat count greater than
494 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
495 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
496 the first (A) could occur when it should not.
497
498 10. Add a cast and remove a redundant test from the code.
499
500 11. JIT should use pcre_malloc/pcre_free for allocation.
501
502 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
503 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
504 is anything other than /usr, -L is still shown).
505
506 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
507
508 14. Perl does not support \N without a following name in a [] class; PCRE now
509 also gives an error.
510
511 15. If a forward reference was repeated with an upper limit of around 2000,
512 it caused the error "internal error: overran compiling workspace". The
513 maximum number of forward references (including repeats) was limited by the
514 internal workspace, and dependent on the LINK_SIZE. The code has been
515 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
516 the default depends on LINK_SIZE. There is a new upper limit (for safety)
517 of around 200,000 forward references. While doing this, I also speeded up
518 the filling in of repeated forward references.
519
520 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
521 incorrectly expecting the subject to contain another "a" after the start.
522
523 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
524 in the match, the SKIP should be ignored. This was not happening; instead
525 the SKIP was being treated as NOMATCH. For patterns such as
526 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
527 tested.
528
529 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
530 now much more compatible with Perl, in particular in cases where the result
531 is a non-match for a non-anchored pattern. For example, if
532 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
533 "m", where previously it did not return a name. A side effect of this
534 change is that for partial matches, the last encountered mark name is
535 returned, as for non matches. A number of tests that were previously not
536 Perl-compatible have been moved into the Perl-compatible test files. The
537 refactoring has had the pleasing side effect of removing one argument from
538 the match() function, thus reducing its stack requirements.
539
540 19. If the /S+ option was used in pcretest to study a pattern using JIT,
541 subsequent uses of /S (without +) incorrectly behaved like /S+.
542
543 21. Retrieve executable code size support for the JIT compiler and fixing
544 some warnings.
545
546 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
547 not work when the shorter character appeared right at the end of the
548 subject string.
549
550 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
551 systems.
552
553 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
554 output it when the /M option is used in pcretest.
555
556 25. The CheckMan script was not being included in the distribution. Also, added
557 an explicit "perl" to run Perl scripts from the PrepareRelease script
558 because this is reportedly needed in Windows.
559
560 26. If study data was being save in a file and studying had not found a set of
561 "starts with" bytes for the pattern, the data written to the file (though
562 never used) was taken from uninitialized memory and so caused valgrind to
563 complain.
564
565 27. Updated RunTest.bat as provided by Sheri Pierce.
566
567 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
568
569 29. Computation of memory usage for the table of capturing group names was
570 giving an unnecessarily large value.
571
572
573 Version 8.20 21-Oct-2011
574 ------------------------
575
576 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
577 a POSIX class. After further experiments with Perl, which convinced me that
578 Perl has bugs and confusions, a closing square bracket is no longer allowed
579 in a POSIX name. This bug also affected patterns with classes that started
580 with full stops.
581
582 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
583 captured substring, but while checking the failing first alternative,
584 substring 1 is temporarily captured. If the output vector supplied to
585 pcre_exec() was not big enough for this capture, the yield of the function
586 was still zero ("insufficient space for captured substrings"). This cannot
587 be totally fixed without adding another stack variable, which seems a lot
588 of expense for a edge case. However, I have improved the situation in cases
589 such as /(a)(b)x|abc/ matched against "abc", where the return code
590 indicates that fewer than the maximum number of slots in the ovector have
591 been set.
592
593 3. Related to (2) above: when there are more back references in a pattern than
594 slots in the output vector, pcre_exec() uses temporary memory during
595 matching, and copies in the captures as far as possible afterwards. It was
596 using the entire output vector, but this conflicts with the specification
597 that only 2/3 is used for passing back captured substrings. Now it uses
598 only the first 2/3, for compatibility. This is, of course, another edge
599 case.
600
601 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
602 main code base, and can be used by building with --enable-jit. When this is
603 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
604 runtime --no-jit option is given.
605
606 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
607 ovector, the return from the function was zero, implying that there were
608 other matches that did not fit. The correct "exactly full" value is now
609 returned.
610
611 6. If a subpattern that was called recursively or as a subroutine contained
612 (*PRUNE) or any other control that caused it to give a non-standard return,
613 invalid errors such as "Error -26 (nested recursion at the same subject
614 position)" or even infinite loops could occur.
615
616 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
617 computing the minimum length on reaching *ACCEPT, and so ended up with the
618 wrong value of 1 rather than 0. Further investigation indicates that
619 computing a minimum subject length in the presence of *ACCEPT is difficult
620 (think back references, subroutine calls), and so I have changed the code
621 so that no minimum is registered for a pattern that contains *ACCEPT.
622
623 8. If (*THEN) was present in the first (true) branch of a conditional group,
624 it was not handled as intended. [But see 16 below.]
625
626 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
627 Sheri Pierce.
628
629 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
630 the first byte in a match must be "a".
631
632 11. Change 17 for 8.13 increased the recursion depth for patterns like
633 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
634 pattern contains any instances of (*THEN). If it does not, the old
635 optimizations are restored. It would be nice to do this on a per-group
636 basis, but at the moment that is not feasible.
637
638 12. In some environments, the output of pcretest -C is CRLF terminated. This
639 broke RunTest's code that checks for the link size. A single white space
640 character after the value is now allowed for.
641
642 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
643 For "fr", it uses the Windows-specific input and output files.
644
645 14. If (*THEN) appeared in a group that was called recursively or as a
646 subroutine, it did not work as intended. [But see next item.]
647
648 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
649 pattern fragments (but not containing any | characters). If A and B are
650 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
651 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
652 D. In other words, Perl considers parentheses that do not contain any |
653 characters to be part of a surrounding alternative, whereas PCRE was
654 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
655 differently. PCRE now behaves in the same way as Perl, except in the case
656 of subroutine/recursion calls such as (?1) which have in any case always
657 been different (but PCRE had them first :-).
658
659 16. Related to 15 above: Perl does not treat the | in a conditional group as
660 creating alternatives. Such a group is treated in the same way as an
661 ordinary group without any | characters when processing (*THEN). PCRE has
662 been changed to match Perl's behaviour.
663
664 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
665 RunGrepTest script failed.
666
667 18. Change 22 for version 13 caused atomic groups to use more stack. This is
668 inevitable for groups that contain captures, but it can lead to a lot of
669 stack use in large patterns. The old behaviour has been restored for atomic
670 groups that do not contain any capturing parentheses.
671
672 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
673 suppress the check for a minimum subject length at run time. (If it was
674 given to pcre_exec() or pcre_dfa_exec() it did work.)
675
676 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
677 fail to work when decoding hex characters in data strings in EBCDIC
678 environments.
679
680 21. It appears that in at least one Mac OS environment, the isxdigit() function
681 is implemented as a macro that evaluates to its argument more than once,
682 contravening the C 90 Standard (I haven't checked a later standard). There
683 was an instance in pcretest which caused it to go wrong when processing
684 \x{...} escapes in subject strings. The has been rewritten to avoid using
685 things like p++ in the argument of isxdigit().
686
687
688 Version 8.13 16-Aug-2011
689 ------------------------
690
691 1. The Unicode data tables have been updated to Unicode 6.0.0.
692
693 2. Two minor typos in pcre_internal.h have been fixed.
694
695 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
696 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
697 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
698
699 4. There were a number of related bugs in the code for matching backrefences
700 caselessly in UTF-8 mode when codes for the characters concerned were
701 different numbers of bytes. For example, U+023A and U+2C65 are an upper
702 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
703 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
704 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
705 2-byte code at the end of the subject (it thought there wasn't enough data
706 left).
707
708 5. Comprehensive information about what went wrong is now returned by
709 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
710 as the output vector has at least 2 elements. The offset of the start of
711 the failing character and a reason code are placed in the vector.
712
713 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
714 now returned is for the first byte of the failing character, instead of the
715 last byte inspected. This is an incompatible change, but I hope it is small
716 enough not to be a problem. It makes the returned offset consistent with
717 pcre_exec() and pcre_dfa_exec().
718
719 7. pcretest now gives a text phrase as well as the error number when
720 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
721 failure, the offset and reason code are output.
722
723 8. When \R was used with a maximizing quantifier it failed to skip backwards
724 over a \r\n pair if the subsequent match failed. Instead, it just skipped
725 back over a single character (\n). This seems wrong (because it treated the
726 two characters as a single entity when going forwards), conflicts with the
727 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
728 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
729 has been changed.
730
731 9. Some internal refactoring has changed the processing so that the handling
732 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
733 time (the PCRE_DOTALL option was changed this way some time ago: version
734 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
735 which was always a bit of a fudge. It also means that there is one less
736 argument for the match() function, which reduces its stack requirements
737 slightly. This change also fixes an incompatibility with Perl: the pattern
738 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
739
740 10. More internal refactoring has drastically reduced the number of recursive
741 calls to match() for possessively repeated groups such as (abc)++ when
742 using pcre_exec().
743
744 11. While implementing 10, a number of bugs in the handling of groups were
745 discovered and fixed:
746
747 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
748 (a|)*(?1) gave a compile-time internal error.
749 ((a|)+)+ did not notice that the outer group could match an empty string.
750 (^a|^)+ was not marked as anchored.
751 (.*a|.*)+ was not marked as matching at start or after a newline.
752
753 12. Yet more internal refactoring has removed another argument from the match()
754 function. Special calls to this function are now indicated by setting a
755 value in a variable in the "match data" data block.
756
757 13. Be more explicit in pcre_study() instead of relying on "default" for
758 opcodes that mean there is no starting character; this means that when new
759 ones are added and accidentally left out of pcre_study(), testing should
760 pick them up.
761
762 14. The -s option of pcretest has been documented for ages as being an old
763 synonym of -m (show memory usage). I have changed it to mean "force study
764 for every regex", that is, assume /S for every regex. This is similar to -i
765 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
766 using it. It makes it easier to run collections of tests with and without
767 study enabled, and thereby test pcre_study() more easily. All the standard
768 tests are now run with and without -s (but some patterns can be marked as
769 "never study" - see 20 below).
770
771 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
772 restoration of the capturing data to the outer values was not happening
773 correctly.
774
775 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
776 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
777 pattern had matched an empty string, and so incorrectly returned a no
778 match.
779
780 17. There was optimizing code for the last branch of non-capturing parentheses,
781 and also for the obeyed branch of a conditional subexpression, which used
782 tail recursion to cut down on stack usage. Unfortunately, now that there is
783 the possibility of (*THEN) occurring in these branches, tail recursion is
784 no longer possible because the return has to be checked for (*THEN). These
785 two optimizations have therefore been removed. [But see 8.20/11 above.]
786
787 18. If a pattern containing \R was studied, it was assumed that \R always
788 matched two bytes, thus causing the minimum subject length to be
789 incorrectly computed because \R can also match just one byte.
790
791 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
792 was incorrectly computed.
793
794 20. If /S is present twice on a test pattern in pcretest input, it now
795 *disables* studying, thereby overriding the use of -s on the command line
796 (see 14 above). This is necessary for one or two tests to keep the output
797 identical in both cases.
798
799 21. When (*ACCEPT) was used in an assertion that matched an empty string and
800 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
801
802 22. When an atomic group that contained a capturing parenthesis was
803 successfully matched, but the branch in which it appeared failed, the
804 capturing was not being forgotten if a higher numbered group was later
805 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
806 group 1 to "a", when in fact it should be unset. This applied to multi-
807 branched capturing and non-capturing groups, repeated or not, and also to
808 positive assertions (capturing in negative assertions does not happen
809 in PCRE) and also to nested atomic groups.
810
811 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
812 subject after a captured substring, to make it easier to tell which of a
813 number of identical substrings has been captured.
814
815 24. The way atomic groups are processed by pcre_exec() has been changed so that
816 if they are repeated, backtracking one repetition now resets captured
817 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
818 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
819 "aaa". Previously, it would have been "a". As part of this code
820 refactoring, the way recursive calls are handled has also been changed.
821
822 25. If an assertion condition captured any substrings, they were not passed
823 back unless some other capturing happened later. For example, if
824 (?(?=(a))a) was matched against "a", no capturing was returned.
825
826 26. When studying a pattern that contained subroutine calls or assertions,
827 the code for finding the minimum length of a possible match was handling
828 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
829 group 1 called group 2 while simultaneously a separate group 2 called group
830 1). A stack overflow occurred in this case. I have fixed this by limiting
831 the recursion depth to 10.
832
833 27. Updated RunTest.bat in the distribution to the version supplied by Tom
834 Fortmann. This supports explicit test numbers on the command line, and has
835 argument validation and error reporting.
836
837 28. An instance of \X with an unlimited repeat could fail if at any point the
838 first character it looked at was a mark character.
839
840 29. Some minor code refactoring concerning Unicode properties and scripts
841 should reduce the stack requirement of match() slightly.
842
843 30. Added the '=' option to pcretest to check the setting of unused capturing
844 slots at the end of the pattern, which are documented as being -1, but are
845 not included in the return count.
846
847 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
848 compiled something random. Now it gives a compile-time error (as does
849 Perl).
850
851 32. A *MARK encountered during the processing of a positive assertion is now
852 recorded and passed back (compatible with Perl).
853
854 33. If --only-matching or --colour was set on a pcregrep call whose pattern
855 had alternative anchored branches, the search for a second match in a line
856 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
857 matched the line "0102" twice. The same bug affected patterns that started
858 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
859 twice.
860
861 34. Previously, PCRE did not allow quantification of assertions. However, Perl
862 does, and because of capturing effects, quantifying parenthesized
863 assertions may at times be useful. Quantifiers are now allowed for
864 parenthesized assertions.
865
866 35. A minor code tidy in pcre_compile() when checking options for \R usage.
867
868 36. \g was being checked for fancy things in a character class, when it should
869 just be a literal "g".
870
871 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
872 appearance of a nested POSIX class supersedes an apparent external class.
873 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
874 unescaped square brackets may also appear as part of class names. For
875 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
876 more like Perl. (But see 8.20/1 above.)
877
878 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
879 was because it thought it was \N{name}, which is not supported).
880
881 39. Add minix to OS list not supporting the -S option in pcretest.
882
883 40. PCRE tries to detect cases of infinite recursion at compile time, but it
884 cannot analyze patterns in sufficient detail to catch mutual recursions
885 such as ((?1))((?2)). There is now a runtime test that gives an error if a
886 subgroup is called recursively as a subpattern for a second time at the
887 same position in the subject string. In previous releases this might have
888 been caught by the recursion limit, or it might have run out of stack.
889
890 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
891 happen only once. PCRE was, however incorrectly giving a compile time error
892 "recursive call could loop indefinitely" because it cannot analyze the
893 pattern in sufficient detail. The compile time test no longer happens when
894 PCRE is compiling a conditional subpattern, but actual runaway loops are
895 now caught at runtime (see 40 above).
896
897 42. It seems that Perl allows any characters other than a closing parenthesis
898 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
899 has been changed to be the same.
900
901 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
902 as not to get warnings when autogen.sh is called. Also changed
903 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
904
905 44. To help people who use pcregrep to scan files containing exceedingly long
906 lines, the following changes have been made:
907
908 (a) The default value of the buffer size parameter has been increased from
909 8K to 20K. (The actual buffer used is three times this size.)
910
911 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
912 PCRE is built.
913
914 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
915 to be set at run time.
916
917 (d) Numerical values in pcregrep options can be followed by K or M, for
918 example --buffer-size=50K.
919
920 (e) If a line being scanned overflows pcregrep's buffer, an error is now
921 given and the return code is set to 2.
922
923 45. Add a pointer to the latest mark to the callout data block.
924
925 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
926 partial match of an empty string instead of no match. This was specific to
927 the use of ".".
928
929 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
930 complete match instead of a partial match. This bug was dependent on both
931 the PCRE_UTF8 and PCRE_DOTALL options being set.
932
933 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
934 starting byte set, because \b was not being ignored.
935
936
937 Version 8.12 15-Jan-2011
938 ------------------------
939
940 1. Fixed some typos in the markup of the man pages, and wrote a script that
941 checks for such things as part of the documentation building process.
942
943 2. On a big-endian 64-bit system, pcregrep did not correctly process the
944 --match-limit and --recursion-limit options (added for 8.11). In
945 particular, this made one of the standard tests fail. (The integer value
946 went into the wrong half of a long int.)
947
948 3. If the --colour option was given to pcregrep with -v (invert match), it
949 did strange things, either producing crazy output, or crashing. It should,
950 of course, ignore a request for colour when reporting lines that do not
951 match.
952
953 4. Another pcregrep bug caused similar problems if --colour was specified with
954 -M (multiline) and the pattern match finished with a line ending.
955
956 5. In pcregrep, when a pattern that ended with a literal newline sequence was
957 matched in multiline mode, the following line was shown as part of the
958 match. This seems wrong, so I have changed it.
959
960 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
961 the check for further matches in the same line (so they could be coloured)
962 to overrun the end of the current line. If another match was found, it was
963 incorrectly shown (and then shown again when found in the next line).
964
965 7. If pcregrep was compiled under Windows, there was a reference to the
966 function pcregrep_exit() before it was defined. I am assuming this was
967 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
968 reported by a user. I've moved the definition above the reference.
969
970
971 Version 8.11 10-Dec-2010
972 ------------------------
973
974 1. (*THEN) was not working properly if there were untried alternatives prior
975 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
976 backtracked to try for "b" instead of moving to the next alternative branch
977 at the same level (in this case, to look for "c"). The Perl documentation
978 is clear that when (*THEN) is backtracked onto, it goes to the "next
979 alternative in the innermost enclosing group".
980
981 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
982 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
983 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
984 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
985 (*THEN).
986
987 3. If \s appeared in a character class, it removed the VT character from
988 the class, even if it had been included by some previous item, for example
989 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
990 of \s, but is part of the POSIX "space" class.)
991
992 4. A partial match never returns an empty string (because you can always
993 match an empty string at the end of the subject); however the checking for
994 an empty string was starting at the "start of match" point. This has been
995 changed to the "earliest inspected character" point, because the returned
996 data for a partial match starts at this character. This means that, for
997 example, /(?<=abc)def/ gives a partial match for the subject "abc"
998 (previously it gave "no match").
999
1000 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1001 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1002 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1003 has an implication that the given string is incomplete (because a partial
1004 match is preferred over a full match). For this reason, these items now
1005 give a partial match in this situation. [Aside: previously, the one case
1006 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1007 match rather than a full match, which was wrong by the old rules, but is
1008 now correct.]
1009
1010 6. There was a bug in the handling of #-introduced comments, recognized when
1011 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1012 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1013 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1014 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1015 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1016 places in pcre_compile().
1017
1018 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1019 comments when looking ahead for named forward references to subpatterns,
1020 the only newline sequence it recognized was NL. It now handles newlines
1021 according to the set newline convention.
1022
1023 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1024 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1025 cater for a lack of strerror(). These oversights have been fixed.
1026
1027 9. Added --match-limit and --recursion-limit to pcregrep.
1028
1029 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1030
1031 11. When the -o option was used, pcregrep was setting a return code of 1, even
1032 when matches were found, and --line-buffered was not being honoured.
1033
1034 12. Added an optional parentheses number to the -o and --only-matching options
1035 of pcregrep.
1036
1037 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1038 can match an empty string. The code to do it in pcretest and pcredemo
1039 needed fixing:
1040
1041 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1042 only one byte after an empty string match just before CRLF (this case
1043 just got forgotten; "any" and "anycrlf" were OK).
1044
1045 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1046 mode when an empty string match preceded an ASCII character followed by
1047 a non-ASCII character. (The code for advancing by one character rather
1048 than one byte was nonsense.)
1049
1050 (c) The pcredemo.c sample program did not have any code at all to handle
1051 the cases when CRLF is a valid newline sequence.
1052
1053 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1054 as a starting offset was within the subject string. There is now a new
1055 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1056 negative or greater than the length of the string. In order to test this,
1057 pcretest is extended to allow the setting of negative starting offsets.
1058
1059 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1060 starting offset points to the beginning of a UTF-8 character was
1061 unnecessarily clumsy. I tidied it up.
1062
1063 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1064 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1065
1066 17. Nobody had reported that the --include_dir option, which was added in
1067 release 7.7 should have been called --include-dir (hyphen, not underscore)
1068 for compatibility with GNU grep. I have changed it to --include-dir, but
1069 left --include_dir as an undocumented synonym, and the same for
1070 --exclude-dir, though that is not available in GNU grep, at least as of
1071 release 2.5.4.
1072
1073 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1074 characters from a string of bytes) have been redefined so as not to use
1075 loops, in order to improve performance in some environments. At the same
1076 time, I abstracted some of the common code into auxiliary macros to save
1077 repetition (this should not affect the compiled code).
1078
1079 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1080 compile-time error is now given if \c is not followed by an ASCII
1081 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1082 different, and any byte value is allowed.)
1083
1084 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1085 START_OPTIMIZE option, which is now allowed at compile time - but just
1086 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1087 to pcregrep and other applications that have no direct access to PCRE
1088 options. The new /Y option in pcretest sets this option when calling
1089 pcre_compile().
1090
1091 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1092 back references. Groups containing recursive back references were forced to
1093 be atomic by that change, but in the case of named groups, the amount of
1094 memory required was incorrectly computed, leading to "Failed: internal
1095 error: code overflow". This has been fixed.
1096
1097 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1098 pcretest.c, to avoid build problems in some Borland environments.
1099
1100
1101 Version 8.10 25-Jun-2010
1102 ------------------------
1103
1104 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1105 THEN.
1106
1107 2. (*ACCEPT) was not working when inside an atomic group.
1108
1109 3. Inside a character class, \B is treated as a literal by default, but
1110 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1111 causes the error). The code is unchanged, but I tidied the documentation.
1112
1113 4. Inside a character class, PCRE always treated \R and \X as literals,
1114 whereas Perl faults them if its -w option is set. I have changed PCRE so
1115 that it faults them when PCRE_EXTRA is set.
1116
1117 5. Added support for \N, which always matches any character other than
1118 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1119
1120 6. When compiling pcregrep with newer versions of gcc which may have
1121 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1122 declared with attribute warn_unused_result" were given. Just casting the
1123 result to (void) does not stop the warnings; a more elaborate fudge is
1124 needed. I've used a macro to implement this.
1125
1126 7. Minor change to pcretest.c to avoid a compiler warning.
1127
1128 8. Added four artifical Unicode properties to help with an option to make
1129 \s etc use properties (see next item). The new properties are: Xan
1130 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1131
1132 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1133 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1134 this option. Modified pcretest to add /W to test this facility. Added
1135 REG_UCP to make it available via the POSIX interface.
1136
1137 10. Added --line-buffered to pcregrep.
1138
1139 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1140 studied, and the match started with a letter with a code point greater than
1141 127 whose first byte was different to the first byte of the other case of
1142 the letter, the other case of this starting letter was not recognized
1143 (#976).
1144
1145 12. If a pattern that was studied started with a repeated Unicode property
1146 test, for example, \p{Nd}+, there was the theoretical possibility of
1147 setting up an incorrect bitmap of starting bytes, but fortunately it could
1148 not have actually happened in practice until change 8 above was made (it
1149 added property types that matched character-matching opcodes).
1150
1151 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1152 possible starting bytes for non-anchored patterns.
1153
1154 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1155 \R, and also a number of cases that involve Unicode properties, both
1156 explicit and implicit when PCRE_UCP is set.
1157
1158 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1159 input, it could crash or give wrong results if characters with values
1160 greater than 0xc0 were present in the subject string. (Detail: it assumed
1161 UTF-8 input when processing these items.)
1162
1163 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1164 size_t is 64-bit (#991).
1165
1166 17. Added a check for running out of memory when PCRE is compiled with
1167 --disable-stack-for-recursion (#990).
1168
1169 18. If the last data line in a file for pcretest does not have a newline on
1170 the end, a newline was missing in the output.
1171
1172 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1173 less than 128) in its various bitmaps. However, there is a facility for
1174 generating tables according to the current locale when PCRE is compiled. It
1175 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1176 characters, are recognized by isspace() and therefore were getting set in
1177 these tables, and indeed these tables seem to approximate to ISO 8859. This
1178 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1179 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1180 which of course cannot start UTF-8 characters. I have changed the code so
1181 that only real ASCII characters (less than 128) and the correct starting
1182 bytes for UTF-8 encodings are set for characters greater than 127 when in
1183 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1184 altogether.)
1185
1186 20. Added the /T option to pcretest so as to be able to run tests with non-
1187 standard character tables, thus making it possible to include the tests
1188 used for 19 above in the standard set of tests.
1189
1190 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1191 reference to a subpattern the other side of a comment that contains an
1192 opening parenthesis caused either an internal compiling error, or a
1193 reference to the wrong subpattern.
1194
1195
1196 Version 8.02 19-Mar-2010
1197 ------------------------
1198
1199 1. The Unicode data tables have been updated to Unicode 5.2.0.
1200
1201 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1202 configured.
1203
1204 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1205 original author of that file, following a query about its status.
1206
1207 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1208 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1209
1210 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1211 quantifier applied to a forward-referencing subroutine call, could compile
1212 incorrect code or give the error "internal error: previously-checked
1213 referenced subpattern not found".
1214
1215 6. Both MS Visual Studio and Symbian OS have problems with initializing
1216 variables to point to external functions. For these systems, therefore,
1217 pcre_malloc etc. are now initialized to local functions that call the
1218 relevant global functions.
1219
1220 7. There were two entries missing in the vectors called coptable and poptable
1221 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1222 I've fixed the data, and added a kludgy way of testing at compile time that
1223 the lengths are correct (equal to the number of opcodes).
1224
1225 8. Following on from 7, I added a similar kludge to check the length of the
1226 eint vector in pcreposix.c.
1227
1228 9. Error texts for pcre_compile() are held as one long string to avoid too
1229 much relocation at load time. To find a text, the string is searched,
1230 counting zeros. There was no check for running off the end of the string,
1231 which could happen if a new error number was added without updating the
1232 string.
1233
1234 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1235
1236 11. \K was not working if it appeared in an atomic group or in a group that
1237 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1238 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1239 the assertion is positive, but not if it is negative.
1240
1241 12. Change 11 fortuitously reduced the size of the stack frame used in the
1242 "match()" function of pcre_exec.c by one pointer. Forthcoming
1243 implementation of support for (*MARK) will need an extra pointer on the
1244 stack; I have reserved it now, so that the stack frame size does not
1245 decrease.
1246
1247 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1248 item in branch that calls a recursion is a subroutine call - as in the
1249 second branch in the above example - was incorrectly given the compile-
1250 time error "recursive call could loop indefinitely" because pcre_compile()
1251 was not correctly checking the subroutine for matching a non-empty string.
1252
1253 14. The checks for overrunning compiling workspace could trigger after an
1254 overrun had occurred. This is a "should never occur" error, but it can be
1255 triggered by pathological patterns such as hundreds of nested parentheses.
1256 The checks now trigger 100 bytes before the end of the workspace.
1257
1258 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1259
1260
1261 Version 8.01 19-Jan-2010
1262 ------------------------
1263
1264 1. If a pattern contained a conditional subpattern with only one branch (in
1265 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1266 computed the wrong minimum data length (which is of course zero for such
1267 subpatterns). This could cause incorrect "no match" results.
1268
1269 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1270 the pattern is reset in the first branch, pcre_compile() failed with
1271 "internal error: code overflow at offset...". This happened only when
1272 the reset was to the original external option setting. (An optimization
1273 abstracts leading options settings into an external setting, which was the
1274 cause of this.)
1275
1276 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1277 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1278 assertion pattern did not match (meaning that the assertion was true), it
1279 was incorrectly treated as false if the SKIP had been reached during the
1280 matching. This also applied to assertions used as conditions.
1281
1282 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1283 assertion subpattern, including such a pattern used as a condition,
1284 unpredictable results occurred, instead of the error return
1285 PCRE_ERROR_DFA_UITEM.
1286
1287 5. The C++ GlobalReplace function was not working like Perl for the special
1288 situation when an empty string is matched. It now does the fancy magic
1289 stuff that is necessary.
1290
1291 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1292 removed. (These were left over from very, very early versions of PCRE.)
1293
1294 7. Some cosmetic changes to the code to make life easier when compiling it
1295 as part of something else:
1296
1297 (a) Change DEBUG to PCRE_DEBUG.
1298
1299 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1300 called "current" as "current_branch", to prevent a collision with the
1301 Linux macro when compiled as a kernel module.
1302
1303 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1304 prevent a collision with the Linux macro when compiled as a kernel
1305 module.
1306
1307 8. In pcre_compile() there are some checks for integer overflows that used to
1308 cast potentially large values to (double). This has been changed to that
1309 when building, a check for int64_t is made, and if it is found, it is used
1310 instead, thus avoiding the use of floating point arithmetic. (There is no
1311 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1312 double.
1313
1314 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1315 2005 (difference between two addresses compared to an unsigned value).
1316
1317 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1318 custom one, because of the following reported problem in Windows:
1319
1320 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1321 under Win32.
1322 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1323 therefore missing the function definition.
1324 - The compiler thus generates a "C" signature for the test function.
1325 - The linker fails to find the "C" function.
1326 - PCRE fails to configure if asked to do so against libbz2.
1327
1328 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1329 messages were output:
1330
1331 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1332 rerunning libtoolize, to keep the correct libtool macros in-tree.
1333 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1334
1335 I have done both of these things.
1336
1337 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1338 most of the time, it *can* run out if it is given a pattern that contains a
1339 runaway infinite recursion. I updated the discussion in the pcrestack man
1340 page.
1341
1342 13. Now that we have gone to the x.xx style of version numbers, the minor
1343 version may start with zero. Using 08 or 09 is a bad idea because users
1344 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1345 interpreted as invalid octal numbers. I've updated the previous comment in
1346 configure.ac, and also added a check that gives an error if 08 or 09 are
1347 used.
1348
1349 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1350 causing partial matching to fail when the end of the subject matched \W
1351 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1352
1353 15. There were some discrepancies between the declarations in pcre_internal.h
1354 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1355 their definitions. The declarations used "const uschar *" and the
1356 definitions used USPTR. Even though USPTR is normally defined as "const
1357 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1358 reported that: "This difference in casting confuses some C++ compilers, for
1359 example, SunCC recognizes above declarations as different functions and
1360 generates broken code for hbpcre." I have changed the declarations to use
1361 USPTR.
1362
1363 16. GNU libtool is named differently on some systems. The autogen.sh script now
1364 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1365 (FreeBSD).
1366
1367 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1368 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1369 comment: "Figure out how to create a longlong from a string: strtoll and
1370 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1371 instance, but it only takes 2 args instead of 3!"
1372
1373 18. A subtle bug concerned with back references has been fixed by a change of
1374 specification, with a corresponding code fix. A pattern such as
1375 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1376 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1377 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1378 same bug. Such groups have to be quantified to be useful, or contained
1379 inside another quantified group. (If there's no repetition, the reference
1380 can never match.) The problem arises because, having left the group and
1381 moved on to the rest of the pattern, a later failure that backtracks into
1382 the group uses the captured value from the final iteration of the group
1383 rather than the correct earlier one. I have fixed this in PCRE by forcing
1384 any group that contains a reference to itself to be an atomic group; that
1385 is, there cannot be any backtracking into it once it has completed. This is
1386 similar to recursive and subroutine calls.
1387
1388
1389 Version 8.00 19-Oct-09
1390 ----------------------
1391
1392 1. The table for translating pcre_compile() error codes into POSIX error codes
1393 was out-of-date, and there was no check on the pcre_compile() error code
1394 being within the table. This could lead to an OK return being given in
1395 error.
1396
1397 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1398 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1399 in a Windows environment.
1400
1401 3. The pcregrep --count option prints the count for each file even when it is
1402 zero, as does GNU grep. However, pcregrep was also printing all files when
1403 --files-with-matches was added. Now, when both options are given, it prints
1404 counts only for those files that have at least one match. (GNU grep just
1405 prints the file name in this circumstance, but including the count seems
1406 more useful - otherwise, why use --count?) Also ensured that the
1407 combination -clh just lists non-zero counts, with no names.
1408
1409 4. The long form of the pcregrep -F option was incorrectly implemented as
1410 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1411 but it seems right to fix it, and I didn't think it was worth preserving
1412 the old behaviour.
1413
1414 5. The command line items --regex=pattern and --regexp=pattern were not
1415 recognized by pcregrep, which required --regex pattern or --regexp pattern
1416 (with a space rather than an '='). The man page documented the '=' forms,
1417 which are compatible with GNU grep; these now work.
1418
1419 6. No libpcreposix.pc file was created for pkg-config; there was just
1420 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1421
1422 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1423 when UCP support is not needed, by modifying the Python script that
1424 generates it from Unicode data files. This should not matter if the module
1425 is correctly used as a library, but I received one complaint about 50K of
1426 unwanted data. My guess is that the person linked everything into his
1427 program rather than using a library. Anyway, it does no harm.
1428
1429 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1430 was a minimum greater than 1 for a wide character in a possessive
1431 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1432 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1433 character. Chaos in the form of incorrect output or a compiling loop could
1434 result.
1435
1436 9. The restrictions on what a pattern can contain when partial matching is
1437 requested for pcre_exec() have been removed. All patterns can now be
1438 partially matched by this function. In addition, if there are at least two
1439 slots in the offset vector, the offset of the earliest inspected character
1440 for the match and the offset of the end of the subject are set in them when
1441 PCRE_ERROR_PARTIAL is returned.
1442
1443 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1444 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1445 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1446 and may be more useful for multi-segment matching.
1447
1448 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1449 used to be given if ever the end of the subject was reached; now it is
1450 given only if matching could not proceed because another character was
1451 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1452 string "Z", which now yields "no match" instead of "partial match". In the
1453 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1454 final character ended with (*FAIL).
1455
1456 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1457 if the pattern had a "must contain" character that was already found in the
1458 earlier partial match, unless partial matching was again requested. For
1459 example, with the pattern /dog.(body)?/, the "must contain" character is
1460 "g". If the first part-match was for the string "dog", restarting with
1461 "sbody" failed. This bug has been fixed.
1462
1463 13. The string returned by pcre_dfa_exec() after a partial match has been
1464 changed so that it starts at the first inspected character rather than the
1465 first character of the match. This makes a difference only if the pattern
1466 starts with a lookbehind assertion or \b or \B (\K is not supported by
1467 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1468 matching functions compatible, and I think it's the right thing to do.
1469
1470 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1471 so that the demonstration program is easily available in environments where
1472 PCRE has not been installed from source.
1473
1474 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1475 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1476 library.
1477
1478 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1479 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1480 is not the first non-POSIX option to be added. Clearly some people find
1481 these options useful.
1482
1483 17. If a caller to the POSIX matching function regexec() passes a non-zero
1484 value for nmatch with a NULL value for pmatch, the value of
1485 nmatch is forced to zero.
1486
1487 18. RunGrepTest did not have a test for the availability of the -u option of
1488 the diff command, as RunTest does. It now checks in the same way as
1489 RunTest, and also checks for the -b option.
1490
1491 19. If an odd number of negated classes containing just a single character
1492 interposed, within parentheses, between a forward reference to a named
1493 subpattern and the definition of the subpattern, compilation crashed with
1494 an internal error, complaining that it could not find the referenced
1495 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1496 [The bug was that it was starting one character too far in when skipping
1497 over the character class, thus treating the ] as data rather than
1498 terminating the class. This meant it could skip too much.]
1499
1500 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1501 /g option in pcretest when the pattern contains \K, which makes it possible
1502 to have an empty string match not at the start, even when the pattern is
1503 anchored. Updated pcretest and pcredemo to use this option.
1504
1505 21. If the maximum number of capturing subpatterns in a recursion was greater
1506 than the maximum at the outer level, the higher number was returned, but
1507 with unset values at the outer level. The correct (outer level) value is
1508 now given.
1509
1510 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1511 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1512 make it do so. The string so far is captured, making this feature
1513 compatible with Perl.
1514
1515 23. The tests have been re-organized, adding tests 11 and 12, to make it
1516 possible to check the Perl 5.10 features against Perl 5.10.
1517
1518 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1519 pattern matches a fixed length string. PCRE did not allow this; now it
1520 does. Neither allows recursion.
1521
1522 25. I finally figured out how to implement a request to provide the minimum
1523 length of subject string that was needed in order to match a given pattern.
1524 (It was back references and recursion that I had previously got hung up
1525 on.) This code has now been added to pcre_study(); it finds a lower bound
1526 to the length of subject needed. It is not necessarily the greatest lower
1527 bound, but using it to avoid searching strings that are too short does give
1528 some useful speed-ups. The value is available to calling programs via
1529 pcre_fullinfo().
1530
1531 26. While implementing 25, I discovered to my embarrassment that pcretest had
1532 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1533 study optimizations had never been tested with that matching function.
1534 Oops. What is worse, even when it was passed study data, there was a bug in
1535 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1536 were also very few tests of studied patterns with pcre_dfa_exec().
1537
1538 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1539 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1540 on the other side of the coin, they are no longer allowed to have different
1541 names, because these cannot be distinguished in PCRE, and this has caused
1542 confusion. (This is a difference from Perl.)
1543
1544 28. When duplicate subpattern names are present (necessarily with different
1545 numbers, as required by 27 above), and a test is made by name in a
1546 conditional pattern, either for a subpattern having been matched, or for
1547 recursion in such a pattern, all the associated numbered subpatterns are
1548 tested, and the overall condition is true if the condition is true for any
1549 one of them. This is the way Perl works, and is also more like the way
1550 testing by number works.
1551
1552
1553 Version 7.9 11-Apr-09
1554 ---------------------
1555
1556 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1557 (pcretest), all targets were linked against these libraries. This included
1558 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1559 libraries. This caused unwanted dependencies to be created. This problem
1560 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1561 pcretest is linked with readline.
1562
1563 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1564 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1565 moved outside it again, because FALSE and TRUE are already defined in AIX,
1566 but BOOL is not.
1567
1568 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1569 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1570
1571 4. The pcregrep documentation said spaces were inserted as well as colons (or
1572 hyphens) following file names and line numbers when outputting matching
1573 lines. This is not true; no spaces are inserted. I have also clarified the
1574 wording for the --colour (or --color) option.
1575
1576 5. In pcregrep, when --colour was used with -o, the list of matching strings
1577 was not coloured; this is different to GNU grep, so I have changed it to be
1578 the same.
1579
1580 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1581 each matching line was coloured. Now it goes on to look for further matches
1582 of any of the test patterns, which is the same behaviour as GNU grep.
1583
1584 7. A pattern that could match an empty string could cause pcregrep to loop; it
1585 doesn't make sense to accept an empty string match in pcregrep, so I have
1586 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1587 seems to be how GNU grep behaves.
1588
1589 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1590 start or after a newline", because the conditional assertion was not being
1591 correctly handled. The rule now is that both the assertion and what follows
1592 in the first alternative must satisfy the test.
1593
1594 9. If auto-callout was enabled in a pattern with a conditional group whose
1595 condition was an assertion, PCRE could crash during matching, both with
1596 pcre_exec() and pcre_dfa_exec().
1597
1598 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1599 used for matching.
1600
1601 11. Unicode property support in character classes was not working for
1602 characters (bytes) greater than 127 when not in UTF-8 mode.
1603
1604 12. Added the -M command line option to pcretest.
1605
1606 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1607
1608 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1609
1610 16. Added comments and documentation about mis-use of no_arg in the C++
1611 wrapper.
1612
1613 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1614 from Martin Jerabek that uses macro names for all relevant character and
1615 string constants.
1616
1617 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1618 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1619 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1620 these, but not everybody uses configure.
1621
1622 19. A conditional group that had only one branch was not being correctly
1623 recognized as an item that could match an empty string. This meant that an
1624 enclosing group might also not be so recognized, causing infinite looping
1625 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1626 with the subject "ab", where knowledge that the repeated group can match
1627 nothing is needed in order to break the loop.
1628
1629 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1630 exec(), but without supplying a callout function, matching went wrong.
1631
1632 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1633 leak if the size of the offset vector was greater than 30. When the vector
1634 is smaller, the saved offsets during recursion go onto a local stack
1635 vector, but for larger vectors malloc() is used. It was failing to free
1636 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1637 error, in fact).
1638
1639 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1640 heapframe that is used only when UTF-8 support is enabled. This caused no
1641 problem, but was untidy.
1642
1643 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1644 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1645 included within another project.
1646
1647 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1648 slightly modified by me:
1649
1650 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1651 not building pcregrep.
1652
1653 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1654 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1655
1656 25. Forward references, both numeric and by name, in patterns that made use of
1657 duplicate group numbers, could behave incorrectly or give incorrect errors,
1658 because when scanning forward to find the reference group, PCRE was not
1659 taking into account the duplicate group numbers. A pattern such as
1660 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1661
1662 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1663 the feature of a custom pointer more persuasive (as requested by a user).
1664
1665 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1666 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1667 pre-defined.
1668
1669 28. Added support for (*UTF8) at the start of a pattern.
1670
1671 29. Arrange for flags added by the "release type" setting in CMake to be shown
1672 in the configuration summary.
1673
1674
1675 Version 7.8 05-Sep-08
1676 ---------------------
1677
1678 1. Replaced UCP searching code with optimized version as implemented for Ad
1679 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1680 stage table and inline lookup instead of a function, giving speed ups of 2
1681 to 5 times on some simple patterns that I tested. Permission was given to
1682 distribute the MultiStage2.py script that generates the tables (it's not in
1683 the tarball, but is in the Subversion repository).
1684
1685 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1686 scripts.
1687
1688 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1689 a group with a zero qualifier. The result of the study could be incorrect,
1690 or the function might crash, depending on the pattern.
1691
1692 4. Caseless matching was not working for non-ASCII characters in back
1693 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1694 It now works when Unicode Property Support is available.
1695
1696 5. In pcretest, an escape such as \x{de} in the data was always generating
1697 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1698 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1699 truncation.
1700
1701 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1702
1703 7. Added two (int) casts to pcregrep when printing the difference of two
1704 pointers, in case they are 64-bit values.
1705
1706 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1707 test 2 if it fails.
1708
1709 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1710 and a #define of that name to empty if it is not externally set. This is to
1711 allow users of MSVC to set it if necessary.
1712
1713 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1714 the convenience functions in the pcre_get.c source file.
1715
1716 11. An option change at the start of a pattern that had top-level alternatives
1717 could cause overwriting and/or a crash. This command provoked a crash in
1718 some environments:
1719
1720 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1721
1722 This potential security problem was recorded as CVE-2008-2371.
1723
1724 12. For a pattern where the match had to start at the beginning or immediately
1725 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1726 pcre_dfa_exec() could read past the end of the passed subject if there was
1727 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1728 pcretest so that it places the subject at the end of its malloc-ed buffer.
1729
1730 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1731 exec() might read past the end of the data buffer in UTF-8 mode.
1732
1733 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1734 the data contained the byte 0x85 as part of a UTF-8 character within its
1735 first line. This applied both to normal and DFA matching.
1736
1737 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1738 /^[^d]*?$/8 failed to match "abc".
1739
1740 16. Added a missing copyright notice to pcrecpp_internal.h.
1741
1742 17. Make it more clear in the documentation that values returned from
1743 pcre_exec() in ovector are byte offsets, not character counts.
1744
1745 18. Tidied a few places to stop certain compilers from issuing warnings.
1746
1747 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1748 supplied by Stefan Weber. I made a further small update for 7.8 because
1749 there is a change of source arrangements: the pcre_searchfuncs.c module is
1750 replaced by pcre_ucd.c.
1751
1752
1753 Version 7.7 07-May-08
1754 ---------------------
1755
1756 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1757 a string to a long long, pretend we don't even have a long long." This is
1758 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1759
1760 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1761 pre-7.6 versions, which defined a global no_arg variable instead of putting
1762 it in the RE class. (See also #8 below.)
1763
1764 3. Remove a line of dead code, identified by coverity and reported by Nuno
1765 Lopes.
1766
1767 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1768
1769 (1) The include/exclude patterns were being applied to the whole pathnames
1770 of files, instead of just to the final components.
1771
1772 (2) If there was more than one level of directory, the subdirectories were
1773 skipped unless they satisfied the include/exclude conditions. This is
1774 inconsistent with GNU grep (and could even be seen as contrary to the
1775 pcregrep specification - which I improved to make it absolutely clear).
1776 The action now is always to scan all levels of directory, and just
1777 apply the include/exclude patterns to regular files.
1778
1779 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1780 --exclude_dir in the tests to avoid scanning .svn directories.
1781
1782 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1783 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1784 doesn't support NULs in patterns.
1785
1786 7. Added some missing "const"s to declarations of static tables in
1787 pcre_compile.c and pcre_dfa_exec.c.
1788
1789 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1790 caused by fix #2 above. (Subsequently also a second patch to fix the
1791 first patch. And a third patch - this was a messy problem.)
1792
1793 9. Applied Craig's patch to remove the use of push_back().
1794
1795 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1796 matching function regexec().
1797
1798 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1799 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1800 references. PCRE supports relative numbers with this syntax (I don't think
1801 Oniguruma does).
1802
1803 12. Previously, a group with a zero repeat such as (...){0} was completely
1804 omitted from the compiled regex. However, this means that if the group
1805 was called as a subroutine from elsewhere in the pattern, things went wrong
1806 (an internal error was given). Such groups are now left in the compiled
1807 pattern, with a new opcode that causes them to be skipped at execution
1808 time.
1809
1810 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1811 to the way PCRE behaves:
1812
1813 (a) A lone ] character is dis-allowed (Perl treats it as data).
1814
1815 (b) A back reference to an unmatched subpattern matches an empty string
1816 (Perl fails the current match path).
1817
1818 (c) A data ] in a character class must be notated as \] because if the
1819 first data character in a class is ], it defines an empty class. (In
1820 Perl it is not possible to have an empty class.) The empty class []
1821 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1822 The negative empty class [^] matches any one character, independently
1823 of the DOTALL setting.
1824
1825 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1826 non-existent subpattern following a character class starting with ']' and
1827 containing () gave an internal compiling error instead of "reference to
1828 non-existent subpattern". Fortunately, when the pattern did exist, the
1829 compiled code was correct. (When scanning forwards to check for the
1830 existencd of the subpattern, it was treating the data ']' as terminating
1831 the class, so got the count wrong. When actually compiling, the reference
1832 was subsequently set up correctly.)
1833
1834 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1835 it was being rejected as not supported by pcre_dfa_exec(), even though
1836 other assertions are supported. I have made pcre_dfa_exec() support
1837 (*FAIL).
1838
1839 16. The implementation of 13c above involved the invention of a new opcode,
1840 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1841 cannot be changed at match time, I realized I could make a small
1842 improvement to matching performance by compiling OP_ALLANY instead of
1843 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1844 on the OP_ANY path.
1845
1846 17. Compiling pcretest on Windows with readline support failed without the
1847 following two fixes: (1) Make the unistd.h include conditional on
1848 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1849
1850 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1851 ncurses library to be included for pcretest when ReadLine support is
1852 requested, but also to allow for it to be overridden. This patch came from
1853 Daniel Bergström.
1854
1855 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1856 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1857 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1858 spotting this.
1859
1860
1861 Version 7.6 28-Jan-08
1862 ---------------------
1863
1864 1. A character class containing a very large number of characters with
1865 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1866 overflow.
1867
1868 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1869 HAVE_LONG_LONG is not defined.
1870
1871 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1872 bring them up to date and include new features. This patch includes:
1873
1874 - Fixed PH's badly added libz and libbz2 support.
1875 - Fixed a problem with static linking.
1876 - Added pcredemo. [But later removed - see 7 below.]
1877 - Fixed dftables problem and added an option.
1878 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1879 HAVE_LONG_LONG.
1880 - Added readline support for pcretest.
1881 - Added an listing of the option settings after cmake has run.
1882
1883 4. A user submitted a patch to Makefile that makes it easy to create
1884 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1885 Makefile.am that cause it to include this special target, without
1886 affecting anything else. Note that the same mingw target plus all
1887 the other distribution libraries and programs are now supported
1888 when configuring with CMake (see 6 below) instead of with
1889 Configure/Make.
1890
1891 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1892 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1893 exported in the Windows port". It has not yet been confirmed that the patch
1894 solves the problem, but it does no harm.
1895
1896 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1897 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1898 with CMake, and also correct the comment about stack recursion.
1899
1900 7. Remove the automatic building of pcredemo from the ./configure system and
1901 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1902 of a program that users should build themselves after PCRE is installed, so
1903 building it automatically is not really right. What is more, it gave
1904 trouble in some build environments.
1905
1906 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1907
1908
1909 Version 7.5 10-Jan-08
1910 ---------------------
1911
1912 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1913 values in parens when parsing an RE using the C++ wrapper."
1914
1915 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1916 Characters greater than 255 were excluded from the class instead of being
1917 included.
1918
1919 3. The same bug as (2) above applied to negated POSIX classes such as
1920 [:^space:].
1921
1922 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1923 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1924 I have changed it.
1925
1926 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1927 first named subpattern) and a construct such as (?&a) would reference the
1928 first named subpattern whose name started with "a" (in other words, the
1929 length check was missing). Both these problems are fixed. "Subpattern name
1930 expected" is now given for (?&) (a zero-length name), and this patch also
1931 makes it give the same error for \k'' (previously it complained that that
1932 was a reference to a non-existent subpattern).
1933
1934 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1935 this is right because (?- can be followed by option settings as well as by
1936 digits. I have, however, made the messages clearer.
1937
1938 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1939 than the number used in the conditional) now cause a compile-time error.
1940 This is actually not compatible with Perl, which accepts such patterns, but
1941 treats the conditional as always being FALSE (as PCRE used to), but it
1942 seems to me that giving a diagnostic is better.
1943
1944 8. Change "alphameric" to the more common word "alphanumeric" in comments
1945 and messages.
1946
1947 9. Fix two occurrences of "backslash" in comments that should have been
1948 "backspace".
1949
1950 10. Remove two redundant lines of code that can never be obeyed (their function
1951 was moved elsewhere).
1952
1953 11. The program that makes PCRE's Unicode character property table had a bug
1954 which caused it to generate incorrect table entries for sequences of
1955 characters that have the same character type, but are in different scripts.
1956 It amalgamated them into a single range, with the script of the first of
1957 them. In other words, some characters were in the wrong script. There were
1958 thirteen such cases, affecting characters in the following ranges:
1959
1960 U+002b0 - U+002c1
1961 U+0060c - U+0060d
1962 U+0061e - U+00612
1963 U+0064b - U+0065e
1964 U+0074d - U+0076d
1965 U+01800 - U+01805
1966 U+01d00 - U+01d77
1967 U+01d9b - U+01dbf
1968 U+0200b - U+0200f
1969 U+030fc - U+030fe
1970 U+03260 - U+0327f
1971 U+0fb46 - U+0fbb1
1972 U+10450 - U+1049d
1973
1974 12. The -o option (show only the matching part of a line) for pcregrep was not
1975 compatible with GNU grep in that, if there was more than one match in a
1976 line, it showed only the first of them. It now behaves in the same way as
1977 GNU grep.
1978
1979 13. If the -o and -v options were combined for pcregrep, it printed a blank
1980 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1981 does the same. The return code can be used to tell if there were any
1982 non-matching lines.
1983
1984 14. Added --file-offsets and --line-offsets to pcregrep.
1985
1986 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1987 infinitely looping recursion. The bug was that positive lookaheads were not
1988 being skipped when checking for a possible empty match (negative lookaheads
1989 and both kinds of lookbehind were skipped).
1990
1991 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1992 inclusion of <windows.h> to before rather than after the definition of
1993 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1994
1995 17. Specifying a possessive quantifier with a specific limit for a Unicode
1996 character property caused pcre_compile() to compile bad code, which led at
1997 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1998 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1999 caused the error; without that there was no problem.
2000
2001 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2002
2003 19. Added --enable-pcretest-libreadline.
2004
2005 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2006 RE::GlobalReplace(). As a result, the number of replacements returned was
2007 double what it should be. I removed one of the increments, but Craig sent a
2008 later patch that removed the other one (the right fix) and added unit tests
2009 that check the return values (which was not done before).
2010
2011 21. Several CMake things:
2012
2013 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2014 the names libpcre and libpcreposix, not just pcre and pcreposix.
2015
2016 (2) The above change means that pcretest and pcregrep are now correctly
2017 linked with the newly-built libraries, not previously installed ones.
2018
2019 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2020
2021 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2022 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2023 UTF-8 newline character). The key issue is that the pattern starts .*;
2024 this means that the match must be either at the beginning, or after a
2025 newline. The bug was in the code for advancing after a failed match and
2026 checking that the new position followed a newline. It was not taking
2027 account of UTF-8 characters correctly.
2028
2029 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2030 character classes. PCRE was not treating the sequence [:...:] as a
2031 character class unless the ... were all letters. Perl, however, seems to
2032 allow any characters between [: and :], though of course it rejects as
2033 unknown any "names" that contain non-letters, because all the known class
2034 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2035 for example, whereas PCRE did not - it did not recognize a POSIX character
2036 class. This seemed a bit dangerous, so the code has been changed to be
2037 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2038 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2039 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2040 Perl does, and where it didn't before.
2041
2042 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2043 Windows environments %n is disabled by default.
2044
2045
2046 Version 7.4 21-Sep-07
2047 ---------------------
2048
2049 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2050 means that a class such as [\s] counted as "explicit reference to CR or
2051 LF". That isn't really right - the whole point of the change was to try to
2052 help when there was an actual mention of one of the two characters. So now
2053 the change happens only if \r or \n (or a literal CR or LF) character is
2054 encountered.
2055
2056 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2057 of both had grown to the point where there were only 3 bits left.
2058 Fortunately, there was spare space in the data structure, and so I have
2059 moved the internal flags into a new 16-bit field to free up more option
2060 bits.
2061
2062 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2063 but did not set the internal JCHANGED flag - either of these is enough to
2064 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2065 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2066 start sets both bits.
2067
2068 4. Added options (at build time, compile time, exec time) to change \R from
2069 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2070
2071 5. doc/pcresyntax.html was missing from the distribution.
2072
2073 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2074 compatibility, even though it is no longer used.
2075
2076 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2077 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2078 windows.h file is present (where different names are used). [This was
2079 reversed later after testing - see 16 below.]
2080
2081 8. Changed all #include <config.h> to #include "config.h". There were also
2082 some further <pcre.h> cases that I changed to "pcre.h".
2083
2084 9. When pcregrep was used with the --colour option, it missed the line ending
2085 sequence off the lines that it output.
2086
2087 10. It was pointed out to me that arrays of string pointers cause lots of
2088 relocations when a shared library is dynamically loaded. A technique of
2089 using a single long string with a table of offsets can drastically reduce
2090 these. I have refactored PCRE in four places to do this. The result is
2091 dramatic:
2092
2093 Originally: 290
2094 After changing UCP table: 187
2095 After changing error message table: 43
2096 After changing table of "verbs" 36
2097 After changing table of Posix names 22
2098
2099 Thanks to the folks working on Gregex for glib for this insight.
2100
2101 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2102 unicode-properties was also set.
2103
2104 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2105
2106 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2107 checked only for CRLF.
2108
2109 14. Added casts to pcretest.c to avoid compiler warnings.
2110
2111 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2112
2113 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2114 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2115 entirely. This removes changes made in 7 above.
2116
2117 17. The CMake files have been updated, and there is now more information about
2118 building with CMake in the NON-UNIX-USE document.
2119
2120
2121 Version 7.3 28-Aug-07
2122 ---------------------
2123
2124 1. In the rejigging of the build system that eventually resulted in 7.1, the
2125 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2126 brackets there is not right, since it causes compilers to look for an
2127 installed pcre.h, not the version that is in the source that is being
2128 compiled (which of course may be different). I have changed it back to:
2129
2130 #include "pcre.h"
2131
2132 I have a vague recollection that the change was concerned with compiling in
2133 different directories, but in the new build system, that is taken care of
2134 by the VPATH setting the Makefile.
2135
2136 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2137 when the subject happened to end in the byte 0x85 (e.g. if the last
2138 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2139 characters but of course it shouldn't be taken as a newline when it is part
2140 of another character. The bug was that, for an unlimited repeat of . in
2141 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2142 characters when looking for a newline.
2143
2144 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2145
2146 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2147 in debug output.
2148
2149 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2150 long printing in the pcrecpp unittest when running under MinGW.
2151
2152 6. ESC_K was left out of the EBCDIC table.
2153
2154 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2155 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2156 limit also applies to "virtual nesting" when a pattern is recursive, and in
2157 this case 1000 isn't so big. I have been able to remove this limit at the
2158 expense of backing off one optimization in certain circumstances. Normally,
2159 when pcre_exec() would call its internal match() function recursively and
2160 immediately return the result unconditionally, it uses a "tail recursion"
2161 feature to save stack. However, when a subpattern that can match an empty
2162 string has an unlimited repetition quantifier, it no longer makes this
2163 optimization. That gives it a stack frame in which to save the data for
2164 checking that an empty string has been matched. Previously this was taken
2165 from the 1000-entry workspace that had been reserved. So now there is no
2166 explicit limit, but more stack is used.
2167
2168 8. Applied Daniel's patches to solve problems with the import/export magic
2169 syntax that is required for Windows, and which was going wrong for the
2170 pcreposix and pcrecpp parts of the library. These were overlooked when this
2171 problem was solved for the main library.
2172
2173 9. There were some crude static tests to avoid integer overflow when computing
2174 the size of patterns that contain repeated groups with explicit upper
2175 limits. As the maximum quantifier is 65535, the maximum group length was
2176 set at 30,000 so that the product of these two numbers did not overflow a
2177 32-bit integer. However, it turns out that people want to use groups that
2178 are longer than 30,000 bytes (though not repeat them that many times).
2179 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2180 made it possible to implement the integer overflow checks in a much more
2181 dynamic way, which I have now done. The artificial limitation on group
2182 length has been removed - we now have only the limit on the total length of
2183 the compiled pattern, which depends on the LINK_SIZE setting.
2184
2185 10. Fixed a bug in the documentation for get/copy named substring when
2186 duplicate names are permitted. If none of the named substrings are set, the
2187 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2188 empty string.
2189
2190 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2191 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2192 because the ] is interpreted as the first data character and the
2193 terminating ] is not found. PCRE has been made compatible with Perl in this
2194 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2195 cause memory overwriting.
2196
2197 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2198 string has been matched (to stop an infinite loop). It was not recognizing
2199 a conditional subpattern that could match an empty string if that
2200 subpattern was within another subpattern. For example, it looped when
2201 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2202 condition was not nested. This bug has been fixed.
2203
2204 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2205 past the start of the subject in the presence of bytes with the top bit
2206 set, for example "\x8aBCD".
2207
2208 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2209 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2210
2211 14. Optimized (?!) to (*FAIL).
2212
2213 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2214 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2215 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2216 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2217 does: it's just the validity check that is more restrictive.
2218
2219 16. Inserted checks for integer overflows during escape sequence (backslash)
2220 processing, and also fixed erroneous offset values for syntax errors during
2221 backslash processing.
2222
2223 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2224 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2225
2226 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2227 caused an overrun.
2228
2229 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2230 something other than just ASCII characters) inside a group that had an
2231 unlimited repeat caused a loop at compile time (while checking to see
2232 whether the group could match an empty string).
2233
2234 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2235 [\P{Any}] did so. (Error in the code for printing property names.)
2236
2237 21. An orphan \E inside a character class could cause a crash.
2238
2239 22. A repeated capturing bracket such as (A)? could cause a wild memory
2240 reference during compilation.
2241
2242 23. There are several functions in pcre_compile() that scan along a compiled
2243 expression for various reasons (e.g. to see if it's fixed length for look
2244 behind). There were bugs in these functions when a repeated \p or \P was
2245 present in the pattern. These operators have additional parameters compared
2246 with \d, etc, and these were not being taken into account when moving along
2247 the compiled data. Specifically:
2248
2249 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2250 length.
2251
2252 (b) An item such as \pL+ within a repeated group could cause crashes or
2253 loops.
2254
2255 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2256 "reference to non-existent subpattern" error.
2257
2258 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2259
2260 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2261 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2262
2263 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2264
2265 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2266 character were causing crashes (broken optimization).
2267
2268 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2269 \p or \P) caused a compile-time loop.
2270
2271 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2272 break. For example, the unstudied pattern [\r\n]A does not match the string
2273 "\r\nA" because change 7.0/46 below moves the current point on by two
2274 characters after failing to match at the start. However, the pattern \nA
2275 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2276 the same is true. There doesn't seem any very clean way out of this, but
2277 what I have chosen to do makes the common cases work: PCRE now takes note
2278 of whether there can be an explicit match for \r or \n anywhere in the
2279 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2280 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2281 pattern has explicit CR or LF references.
2282
2283 29. Added (*CR) etc for changing newline setting at start of pattern.
2284
2285
2286 Version 7.2 19-Jun-07
2287 ---------------------
2288
2289 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2290 which is apparently normally available under Windows.
2291
2292 2. Re-jig the pcregrep tests with different newline settings in an attempt
2293 to make them independent of the local environment's newline setting.
2294
2295 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2296
2297 4. Some of the "internals" tests were previously cut out when the link size
2298 was not 2, because the output contained actual offsets. The recent new
2299 "Z" feature of pcretest means that these can be cut out, making the tests
2300 usable with all link sizes.
2301
2302 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2303 stack recursion. This gives a massive performance boost under BSD, but just
2304 a small improvement under Linux. However, it saves one field in the frame
2305 in all cases.
2306
2307 6. Added more features from the forthcoming Perl 5.10:
2308
2309 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2310 recursion call. It refers to the nth most recently opened parentheses.
2311
2312 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2313 to be opened parentheses.
2314
2315 (c) Conditions that refer to capturing parentheses can be specified
2316 relatively, for example, (?(-2)... or (?(+3)...
2317
2318 (d) \K resets the start of the current match so that everything before
2319 is not part of it.
2320
2321 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2322
2323 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2324 reference syntax.
2325
2326 (g) (?| introduces a group in which the numbering of parentheses in each
2327 alternative starts with the same number.
2328
2329 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2330
2331 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2332 PCRE_INFO_JCHANGED.
2333
2334 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2335 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2336 for detecting groups that can match an empty string.
2337
2338 9. A pattern with a very large number of alternatives (more than several
2339 hundred) was running out of internal workspace during the pre-compile
2340 phase, where pcre_compile() figures out how much memory will be needed. A
2341 bit of new cunning has reduced the workspace needed for groups with
2342 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2343 workspace instead of running out of the 4096 that are available.
2344
2345 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2346
2347 11. Applied patch from Google to remove an optimization that didn't quite work.
2348 The report of the bug said:
2349
2350 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2351 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2352 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2353
2354 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2355 it matched the wrong number of bytes.
2356
2357
2358 Version 7.1 24-Apr-07
2359 ---------------------
2360
2361 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2362 that is more "standard", making use of automake and other Autotools. There
2363 is some re-arrangement of the files and adjustment of comments consequent
2364 on this.
2365
2366 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2367 for recursive directory scanning broke on some systems because the files
2368 are not scanned in any specific order and on different systems the order
2369 was different. A call to "sort" has been inserted into RunGrepTest for the
2370 approprate test as a short-term fix. In the longer term there may be an
2371 alternative.
2372
2373 3. I had an email from Eric Raymond about problems translating some of PCRE's
2374 man pages to HTML (despite the fact that I distribute HTML pages, some
2375 people do their own conversions for various reasons). The problems
2376 concerned the use of low-level troff macros .br and .in. I have therefore
2377 removed all such uses from the man pages (some were redundant, some could
2378 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2379 HTML has been updated to handle .nf/.fi and to complain if it encounters
2380 .br or .in.
2381
2382 4. Updated comments in configure.ac that get placed in config.h.in and also
2383 arranged for config.h to be included in the distribution, with the name
2384 config.h.generic, for the benefit of those who have to compile without
2385 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2386
2387 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2388 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2389 makevp.bat for the current PCRE, using the additional files
2390 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2391
2392 6. A Windows user reported a minor discrepancy with test 2, which turned out
2393 to be caused by a trailing space on an input line that had got lost in his
2394 copy. The trailing space was an accident, so I've just removed it.
2395
2396 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2397 that is needed.
2398
2399 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2400 as "const" (a) because they are and (b) because it helps the PHP
2401 maintainers who have recently made a script to detect big data structures
2402 in the php code that should be moved to the .rodata section. I remembered
2403 to update Builducptable as well, so it won't revert if ucptable.h is ever
2404 re-created.
2405
2406 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2407 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2408 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2409 support is not required. This saves 1.5-2K of code, which is important in
2410 some applications.
2411
2412 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2413 so as not to refer to the tables, even though these functions will never be
2414 called when UTF-8 support is disabled. Otherwise there are problems with a
2415 shared library.
2416
2417 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2418
2419 (a) It was defining its arguments as char * instead of void *.
2420
2421 (b) It was assuming that all moves were upwards in memory; this was true
2422 a long time ago when I wrote it, but is no longer the case.
2423
2424 The emulated memove() is provided for those environments that have neither
2425 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2426 is clearly not the case, as these two bugs were recently reported.
2427
2428 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2429 and Detrail to create the HTML documentation, the .txt form of the man
2430 pages, and it removes trailing spaces from listed files. It also creates
2431 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2432 case, it wraps all the #defines with #ifndefs. This script should be run
2433 before "make dist".
2434
2435 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2436 with Unicode property support.
2437
2438 (a) For a maximizing quantifier, if the two different cases of the
2439 character were of different lengths in their UTF-8 codings (there are
2440 some cases like this - I found 11), and the matching function had to
2441 back up over a mixture of the two cases, it incorrectly assumed they
2442 were both the same length.
2443
2444 (b) When PCRE was configured to use the heap rather than the stack for
2445 recursion during matching, it was not correctly preserving the data for
2446 the other case of a UTF-8 character when checking ahead for a match
2447 while processing a minimizing repeat. If the check also involved
2448 matching a wide character, but failed, corruption could cause an
2449 erroneous result when trying to check for a repeat of the original
2450 character.
2451
2452 13. Some tidying changes to the testing mechanism:
2453
2454 (a) The RunTest script now detects the internal link size and whether there
2455 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2456 values substituted by "configure". (The RunGrepTest script already did
2457 this for UTF-8.) The configure.ac script no longer substitutes the
2458 relevant variables.
2459
2460 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2461 with length and offset values. This means that the output is different
2462 for different internal link sizes. Test 2 is skipped for link sizes
2463 other than 2 because of this, bypassing the problem. Unfortunately,
2464 there was also a test in test 3 (the locale tests) that used /B and
2465 failed for link sizes other than 2. Rather than cut the whole test out,
2466 I have added a new /Z option to pcretest that replaces the length and
2467 offset values with spaces. This is now used to make test 3 independent
2468 of link size. (Test 2 will be tidied up later.)
2469
2470 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2471 segmentation fault instead of returning the appropriate error message.
2472
2473 15. In multiline mode when the newline sequence was set to "any", the pattern
2474 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2475 This doesn't seem right; it now treats the CRLF combination as the line
2476 ending, and so does not match in that case. It's only a pattern such as ^$
2477 that would hit this one: something like ^ABC$ would have failed after \r
2478 and then tried again after \r\n.
2479
2480 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2481 in an attempt to make files that differ only in their line terminators
2482 compare equal. This works on Linux.
2483
2484 17. Under certain error circumstances pcregrep might try to free random memory
2485 as it exited. This is now fixed, thanks to valgrind.
2486
2487 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2488 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2489 was because its rules for how to advance for /g after matching an empty
2490 string at the end of a line did not allow for this case. They now check for
2491 it specially.
2492
2493 20. pcretest is supposed to handle patterns and data of any length, by
2494 extending its buffers when necessary. It was getting this wrong when the
2495 buffer for a data line had to be extended.
2496
2497 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2498 CRLF as a newline sequence.
2499
2500 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2501 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2502 I have nevertheless tidied it up.
2503
2504 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2505
2506 24. Added a man page for pcre-config.
2507
2508
2509 Version 7.0 19-Dec-06
2510 ---------------------
2511
2512 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2513 moving to gcc 4.1.1.
2514
2515 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2516 sys/time.h, which is documented as needed for this function. It doesn't
2517 seem to matter on Linux, but it showed up on some releases of OS X.
2518
2519 3. It seems that there are systems where bytes whose values are greater than
2520 127 match isprint() in the "C" locale. The "C" locale should be the
2521 default when a C program starts up. In most systems, only ASCII printing
2522 characters match isprint(). This difference caused the output from pcretest
2523 to vary, making some of the tests fail. I have changed pcretest so that:
2524
2525 (a) When it is outputting text in the compiled version of a pattern, bytes
2526 other than 32-126 are always shown as hex escapes.
2527
2528 (b) When it is outputting text that is a matched part of a subject string,
2529 it does the same, unless a different locale has been set for the match
2530 (using the /L modifier). In this case, it uses isprint() to decide.
2531
2532 4. Fixed a major bug that caused incorrect computation of the amount of memory
2533 required for a compiled pattern when options that changed within the
2534 pattern affected the logic of the preliminary scan that determines the
2535 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2536 that the computed length was too small. The symptoms of this bug were
2537 either the PCRE error "internal error: code overflow" from pcre_compile(),
2538 or a glibc crash with a message such as "pcretest: free(): invalid next
2539 size (fast)". Examples of patterns that provoked this bug (shown in
2540 pcretest format) are:
2541
2542 /(?-x: )/x
2543 /(?x)(?-x: \s*#\s*)/
2544 /((?i)[\x{c0}])/8
2545 /(?i:[\x{c0}])/8
2546
2547 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2548 is now done differently.
2549
2550 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2551 wrapper classes; (b) implement a new function in the C++ scanner that is
2552 more efficient than the old way of doing things because it avoids levels of
2553 recursion in the regex matching; (c) add a paragraph to the documentation
2554 for the FullMatch() function.
2555
2556 6. The escape sequence \n was being treated as whatever was defined as
2557 "newline". Not only was this contrary to the documentation, which states
2558 that \n is character 10 (hex 0A), but it also went horribly wrong when
2559 "newline" was defined as CRLF. This has been fixed.
2560
2561 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2562 was being set to -1 for the "end of line" case (supposedly a value that no
2563 character can have). Though this value is never used (the check for end of
2564 line is "zero bytes in current character"), it caused compiler complaints.
2565 I've changed it to 0xffffffff.
2566
2567 8. In pcre_version.c, the version string was being built by a sequence of
2568 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2569 string (as it is for production releases) called a macro with an empty
2570 argument. The C standard says the result of this is undefined. The gcc
2571 compiler treats it as an empty string (which was what was wanted) but it is
2572 reported that Visual C gives an error. The source has been hacked around to
2573 avoid this problem.
2574
2575 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2576 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2577 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2578 of them did).
2579
2580 10. Originally, pcretest opened its input and output without "b"; then I was
2581 told that "b" was needed in some environments, so it was added for release
2582 5.0 to both the input and output. (It makes no difference on Unix-like
2583 systems.) Later I was told that it is wrong for the input on Windows. I've
2584 now abstracted the modes into two macros, to make it easier to fiddle with
2585 them, and removed "b" from the input mode under Windows.
2586
2587 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2588
2589 12. Added -help and --help to pcretest as an official way of being reminded
2590 of the options.
2591
2592 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2593 and pcrecpp.cc because they annoy compilers at high warning levels.
2594
2595 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2596
2597 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2598 scripts are not C programs :-) and which was not noticed because it works
2599 on Linux.
2600
2601 16. pcretest is supposed to handle any length of pattern and data line (as one
2602 line or as a continued sequence of lines) by extending its input buffer if
2603 necessary. This feature was broken for very long pattern lines, leading to
2604 a string of junk being passed to pcre_compile() if the pattern was longer
2605 than about 50K.
2606
2607 17. I have done a major re-factoring of the way pcre_compile() computes the
2608 amount of memory needed for a compiled pattern. Previously, there was code
2609 that made a preliminary scan of the pattern in order to do this. That was
2610 OK when PCRE was new, but as the facilities have expanded, it has become
2611 harder and harder to keep it in step with the real compile phase, and there
2612 have been a number of bugs (see for example, 4 above). I have now found a
2613 cunning way of running the real compile function in a "fake" mode that
2614 enables it to compute how much memory it would need, while actually only
2615 ever using a few hundred bytes of working memory and without too many
2616 tests of the mode. This should make future maintenance and development
2617 easier. A side effect of this work is that the limit of 200 on the nesting
2618 depth of parentheses has been removed (though this was never a serious
2619 limitation, I suspect). However, there is a downside: pcre_compile() now
2620 runs more slowly than before (30% or more, depending on the pattern). I
2621 hope this isn't a big issue. There is no effect on runtime performance.
2622
2623 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2624 newline (only possible for the last line of a file) and it was a
2625 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2626
2627 19. Added additional timing features to pcretest. (1) The -tm option now times
2628 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2629 separate command line item, by a number that specifies the number of
2630 repeats to use when timing. The default is 50000; this gives better
2631 precision, but takes uncomfortably long for very large patterns.
2632
2633 20. Extended pcre_study() to be more clever in cases where a branch of a
2634 subpattern has no definite first character. For example, (a*|b*)[cd] would
2635 previously give no result from pcre_study(). Now it recognizes that the
2636 first character must be a, b, c, or d.
2637
2638 21. There was an incorrect error "recursive call could loop indefinitely" if
2639 a subpattern (or the entire pattern) that was being tested for matching an
2640 empty string contained only one non-empty item after a nested subpattern.
2641 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2642 incorrectly, because the \d was being skipped in the check.
2643
2644 22. The pcretest program now has a new pattern option /B and a command line
2645 option -b, which is equivalent to adding /B to every pattern. This causes
2646 it to show the compiled bytecode, without the additional information that
2647 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2648 is the same as /B/I).
2649
2650 23. A new optimization is now able automatically to treat some sequences such
2651 as a*b as a*+b. More specifically, if something simple (such as a character
2652 or a simple class like \d) has an unlimited quantifier, and is followed by
2653 something that cannot possibly match the quantified thing, the quantifier
2654 is automatically "possessified".
2655
2656 24. A recursive reference to a subpattern whose number was greater than 39
2657 went wrong under certain circumstances in UTF-8 mode. This bug could also
2658 have affected the operation of pcre_study().
2659
2660 25. Realized that a little bit of performance could be had by replacing
2661 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2662
2663 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2664
2665 27. Possessive quantifiers such as a++ were previously implemented by turning
2666 them into atomic groups such as ($>a+). Now they have their own opcodes,
2667 which improves performance. This includes the automatically created ones
2668 from 23 above.
2669
2670 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2671 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2672 the first matched character to be a colon. This applied both to named and
2673 numbered groups.
2674
2675 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2676
2677 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2678 building PCRE on Windows easier, so I have included it in the distribution.
2679
2680 31. There is now a check in pcretest against a ridiculously large number being
2681 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2682 loop, the loop is abandoned.
2683
2684 32. Forward references to subpatterns in conditions such as (?(2)...) where
2685 subpattern 2 is defined later cause pcre_compile() to search forwards in
2686 the pattern for the relevant set of parentheses. This search went wrong
2687 when there were unescaped parentheses in a character class, parentheses
2688 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2689
2690 33. "Subroutine" calls and backreferences were previously restricted to
2691 referencing subpatterns earlier in the regex. This restriction has now
2692 been removed.
2693
2694 34. Added a number of extra features that are going to be in Perl 5.10. On the
2695 whole, these are just syntactic alternatives for features that PCRE had
2696 previously implemented using the Python syntax or my own invention. The
2697 other formats are all retained for compatibility.
2698
2699 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2700 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2701 also .NET compatible.
2702
2703 (b) A recursion or subroutine call to a named group can now be defined as
2704 (?&name) as well as (?P>name).
2705
2706 (c) A backreference to a named group can now be defined as \k<name> or
2707 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2708 5.10, are also .NET compatible.
2709
2710 (d) A conditional reference to a named group can now use the syntax
2711 (?(<name>) or (?('name') as well as (?(name).
2712
2713 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2714 groups (named and numbered) that are never evaluated inline, but can be
2715 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2716 is always false. There may be only one alternative in such a group.
2717
2718 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2719 as the simple (?(R). The condition is true only if the most recent
2720 recursion is that of the given number or name. It does not search out
2721 through the entire recursion stack.
2722
2723 (g) The escape \gN or \g{N} has been added, where N is a positive or
2724 negative number, specifying an absolute or relative reference.
2725
2726 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2727 some "unreachable code" warnings.
2728
2729 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2730 things, this adds five new scripts.
2731
2732 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2733 There were also incompatibilities regarding the handling of \Q..\E inside
2734 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2735 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2736
2737 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2738 matches an empty string, and forcibly breaks the loop. There were bugs in
2739 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2740 against aaaa the result was just "a" rather than "aaaa", for example. Two
2741 separate and independent bugs (that affected different cases) have been
2742 fixed.
2743
2744 39. Refactored the code to abolish the use of different opcodes for small
2745 capturing bracket numbers. This is a tidy that I avoided doing when I
2746 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2747 The new approach is not only tidier, it makes it possible to reduce the
2748 memory needed to fix the previous bug (38).
2749
2750 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2751 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2752 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2753 mode.
2754
2755 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2756 report.
2757
2758 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2759 copy construction and assignment in the C++ wrapper.
2760
2761 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2762 couple of bugs that could have given wrong results in the "--newline=crlf"
2763 case.
2764
2765 44. Added a number of casts and did some reorganization of signed/unsigned int
2766 variables following suggestions from Dair Grant. Also renamed the variable
2767 "this" as "item" because it is a C++ keyword.
2768
2769 45. Arranged for dftables to add
2770
2771 #include "pcre_internal.h"
2772
2773 to pcre_chartables.c because without it, gcc 4.x may remove the array
2774 definition from the final binary if PCRE is built into a static library and
2775 dead code stripping is activated.
2776
2777 46. For an unanchored pattern, if a match attempt fails at the start of a
2778 newline sequence, and the newline setting is CRLF or ANY, and the next two
2779 characters are CRLF, advance by two characters instead of one.
2780
2781
2782 Version 6.7 04-Jul-06
2783 ---------------------
2784
2785 1. In order to handle tests when input lines are enormously long, pcretest has
2786 been re-factored so that it automatically extends its buffers when
2787 necessary. The code is crude, but this _is_ just a test program. The
2788 default size has been increased from 32K to 50K.
2789
2790 2. The code in pcre_study() was using the value of the re argument before
2791 testing it for NULL. (Of course, in any sensible call of the function, it
2792 won't be NULL.)
2793
2794 3. The memmove() emulation function in pcre_internal.h, which is used on
2795 systems that lack both memmove() and bcopy() - that is, hardly ever -
2796 was missing a "static" storage class specifier.
2797
2798 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2799 containing an extended class (one that cannot be represented by a bitmap
2800 because it contains high-valued characters or Unicode property items, e.g.
2801 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2802 pattern, but PCRE should not loop if you do not (it no longer does).
2803 [Detail: two cases were found: (a) a repeated subpattern containing an
2804 extended class; (b) a recursive reference to a subpattern that followed a
2805 previous extended class. It wasn't skipping over the extended class
2806 correctly when UTF-8 mode was not set.]
2807
2808 5. A negated single-character class was not being recognized as fixed-length
2809 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2810 compile error "lookbehind assertion is not fixed length".
2811
2812 6. The RunPerlTest auxiliary script was showing an unexpected difference
2813 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2814 write a Perl script that can interpret lines of an input file either as
2815 byte characters or as UTF-8, which is what "perltest" was being required to
2816 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2817 can't do is switch easily at run time between having the "use utf8;" pragma
2818 or not. In the end, I fudged it by using the RunPerlTest script to insert
2819 "use utf8;" explicitly for the UTF-8 tests.
2820
2821 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2822 the end of the subject string, contrary to the documentation and to what
2823 Perl does. This was true of both matching functions. Now it matches only at
2824 the start of the subject and immediately after *internal* newlines.
2825
2826 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2827 a pointer to an int instead of a pointer to an unsigned long int. This
2828 caused problems on 64-bit systems.
2829
2830 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2831 instance of the 'standard' template library not being so standard".
2832
2833 10. There was no check on the number of named subpatterns nor the maximum
2834 length of a subpattern name. The product of these values is used to compute
2835 the size of the memory block for a compiled pattern. By supplying a very
2836 long subpattern name and a large number of named subpatterns, the size
2837 computation could be caused to overflow. This is now prevented by limiting
2838 the length of names to 32 characters, and the number of named subpatterns
2839 to 10,000.
2840
2841 11. Subpatterns that are repeated with specific counts have to be replicated in
2842 the compiled pattern. The size of memory for this was computed from the
2843 length of the subpattern and the repeat count. The latter is limited to
2844 65535, but there was no limit on the former, meaning that integer overflow
2845 could in principle occur. The compiled length of a repeated subpattern is
2846 now limited to 30,000 bytes in order to prevent this.
2847
2848 12. Added the optional facility to have named substrings with the same name.
2849
2850 13. Added the ability to use a named substring as a condition, using the
2851 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2852 are numbers (not recommended). Forward references are permitted.
2853
2854 14. Added forward references in named backreferences (if you see what I mean).
2855
2856 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2857 pattern could run off the end of the subject. For example, the pattern
2858 "(?s)(.{1,5})"8 did this with the subject "ab".
2859
2860 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2861 PCRE_CASELESS was set when matching characters that were quantified with ?
2862 or *.
2863
2864 17. A character class other than a single negated character that had a minimum
2865 but no maximum quantifier - for example [ab]{6,} - was not handled
2866 correctly by pce_dfa_exec(). It would match only one character.
2867
2868 18. A valid (though odd) pattern that looked like a POSIX character
2869 class but used an invalid character after [ (for example [[,abc,]]) caused
2870 pcre_compile() to give the error "Failed: internal error: code overflow" or
2871 in some cases to crash with a glibc free() error. This could even happen if
2872 the pattern terminated after [[ but there just happened to be a sequence of
2873 letters, a binary zero, and a closing ] in the memory that followed.
2874
2875 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2876 over the years. Originally (before any Unicode support), just the bottom 8
2877 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2878 output from "man perlunicode" includes this:
2879
2880 The regular expression compiler produces polymorphic opcodes. That
2881 is, the pattern adapts to the data and automatically switches to
2882 the Unicode character scheme when presented with Unicode data--or
2883 instead uses a traditional byte scheme when presented with byte
2884 data.
2885
2886 Sadly, a wide octal escape does not cause a switch, and in a string with
2887 no other multibyte characters, these octal escapes are treated as before.
2888 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2889 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2890 Unicode string.
2891
2892 I have not perpetrated such confusion in PCRE. Up till now, it took just
2893 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2894 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2895 translate to the appropriate multibyte character.
2896
2897 29. Applied some refactoring to reduce the number of warnings from Microsoft
2898 and Borland compilers. This has included removing the fudge introduced
2899 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2900 a warning about an unused variable.
2901
2902 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2903 characters since release 4.0, because Perl (from release 5.004) does not.
2904 [Or at least, is documented not to: some releases seem to be in conflict
2905 with the documentation.] However, when a pattern was studied with
2906 pcre_study() and all its branches started with \s, PCRE still included VT
2907 as a possible starting character. Of course, this did no harm; it just
2908 caused an unnecessary match attempt.
2909
2910 22. Removed a now-redundant internal flag bit that recorded the fact that case
2911 dependency changed within the pattern. This was once needed for "required
2912 byte" processing, but is no longer used. This recovers a now-scarce options
2913 bit. Also moved the least significant internal flag bit to the most-
2914 significant bit of the word, which was not previously used (hangover from
2915 the days when it was an int rather than a uint) to free up another bit for
2916 the future.
2917
2918 23. Added support for CRLF line endings as well as CR and LF. As well as the
2919 default being selectable at build time, it can now be changed at runtime
2920 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2921 specify that it is scanning data with non-default line endings.
2922
2923 24. Changed the definition of CXXLINK to make it agree with the definition of
2924 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2925
2926 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2927 recursions. This makes a big different to stack usage for some patterns.
2928
2929 26. If a subpattern containing a named recursion or subroutine reference such
2930 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2931 the space required for the compiled pattern went wrong and gave too small a
2932 value. Depending on the environment, this could lead to "Failed: internal
2933 error: code overflow at offset 49" or "glibc detected double free or
2934 corruption" errors.
2935
2936 27. Applied patches from Google (a) to support the new newline modes and (b) to
2937 advance over multibyte UTF-8 characters in GlobalReplace.
2938
2939 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2940 difference for some implementation of PCRE in some Windows version.
2941
2942 29. Added some extra testing facilities to pcretest:
2943
2944 \q<number> in a data line sets the "match limit" value
2945 \Q<number> in a data line sets the "match recursion limt" value
2946 -S <number> sets the stack size, where <number> is in megabytes
2947
2948 The -S option isn't available for Windows.
2949
2950
2951 Version 6.6 06-Feb-06
2952 ---------------------
2953
2954 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2955 in pcreposix.h. I have copied the definition from pcre.h.
2956
2957 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2958 because pcre.h is no longer a built file.
2959
2960 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2961 not normally included in the compiled code.
2962
2963
2964 Version 6.5 01-Feb-06
2965 ---------------------
2966
2967 1. When using the partial match feature with pcre_dfa_exec(), it was not
2968 anchoring the second and subsequent partial matches at the new starting
2969 point. This could lead to incorrect results. For example, with the pattern
2970 /1234/, partially matching against "123" and then "a4" gave a match.
2971
2972 2. Changes to pcregrep:
2973
2974 (a) All non-match returns from pcre_exec() were being treated as failures
2975 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2976 error message is output. Some extra information is given for the
2977 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2978 probably the only errors that are likely to be caused by users (by
2979 specifying a regex that has nested indefinite repeats, for instance).
2980 If there are more than 20 of these errors, pcregrep is abandoned.
2981
2982 (b) A binary zero was treated as data while matching, but terminated the
2983 output line if it was written out. This has been fixed: binary zeroes
2984 are now no different to any other data bytes.
2985
2986 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2987 used to set a locale for matching. The --locale=xxxx long option has
2988 been added (no short equivalent) to specify a locale explicitly on the
2989 pcregrep command, overriding the environment variables.
2990
2991 (d) When -B was used with -n, some line numbers in the output were one less
2992 than they should have been.
2993
2994 (e) Added the -o (--only-matching) option.
2995
2996 (f) If -A or -C was used with -c (count only), some lines of context were
2997 accidentally printed for the final match.
2998
2999 (g) Added the -H (--with-filename) option.
3000
3001 (h) The combination of options -rh failed to suppress file names for files
3002 that were found from directory arguments.
3003
3004 (i) Added the -D (--devices) and -d (--directories) options.
3005
3006 (j) Added the -F (--fixed-strings) option.
3007
3008 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3009
3010 (l) Added the --colo(u)r option.
3011
3012 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3013 is not present by default.
3014
3015 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3016 items such as (?R) or (?1), when the recursion could match a number of
3017 alternatives. If it matched one of the alternatives, but subsequently,
3018 outside the recursion, there was a failure, the code tried to back up into
3019 the recursion. However, because of the way PCRE is implemented, this is not
3020 possible, and the result was an incorrect result from the match.
3021
3022 In order to prevent this happening, the specification of recursion has
3023 been changed so that all such subpatterns are automatically treated as
3024 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3025
3026 4. I had overlooked the fact that, in some locales, there are characters for
3027 which isalpha() is true but neither isupper() nor islower() are true. In
3028 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3029 and ordfeminine) are like this. This affected the treatment of \w and \W
3030 when they appeared in character classes, but not when they appeared outside
3031 a character class. The bit map for "word" characters is now created
3032 separately from the results of isalnum() instead of just taking it from the
3033 upper, lower, and digit maps. (Plus the underscore character, of course.)
3034
3035 5. The above bug also affected the handling of POSIX character classes such as
3036 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3037 permanent tables. Instead, the bit maps for such a class were previously
3038 created as the appropriate unions of the upper, lower, and digit bitmaps.
3039 Now they are created by subtraction from the [[:word:]] class, which has
3040 its own bitmap.
3041
3042 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3043 It is created by subtracting the vertical space characters (\x09, \x0a,
3044 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3045 subtraction was done in the overall bitmap for a character class, meaning
3046 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3047 be recognized. This bug has been fixed.
3048
3049 7. Patches from the folks at Google:
3050
3051 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3052 real life, but is still worth protecting against".
3053
3054 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3055 regular expressions".
3056
3057 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3058 have it.
3059
3060 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3061 "configure" and the latter not, in order to fix a problem somebody had
3062 with compiling the Arg class on HP-UX.
3063
3064 (e) Improve the error-handling of the C++ wrapper a little bit.
3065
3066 (f) New tests for checking recursion limiting.
3067
3068 8. The pcre_memmove() function, which is used only if the environment does not
3069 have a standard memmove() function (and is therefore rarely compiled),
3070 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3071 returning a result (though PCRE never actually uses the result).
3072
3073 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3074 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3075 returned instead of calling malloc() with an overflowing number that would
3076 most likely cause subsequent chaos.
3077
3078 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3079
3080 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3081 with this option is matched, the nmatch and pmatch options of regexec() are
3082 ignored.
3083
3084 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3085 provided in case anyone wants to the the POSIX interface with UTF-8
3086 strings.
3087
3088 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3089 C++ linking (needed for some HP-UX environments).
3090
3091 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3092 (unused parameter) and in the pcre_printint() function (omitted "default"
3093 switch label when the default is to do nothing).
3094
3095 15. Added some code to make it possible, when PCRE is compiled as a C++
3096 library, to replace subject pointers for pcre_exec() with a smart pointer
3097 class, thus making it possible to process discontinuous strings.
3098
3099 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3100 much the same function. They were added by different people who were trying
3101 to make PCRE easy to compile on non-Unix systems. It has been suggested
3102 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3103 for compiling on Windows systems. I have therefore replaced it with
3104 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3105 defaults to "extern" for C or "extern C" for C++, which works fine on
3106 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3107 SCOPE with something explicit in config.h. In addition:
3108
3109 (a) pcreposix.h still had just "extern" instead of either of these macros;
3110 I have replaced it with PCRE_DATA_SCOPE.
3111
3112 (b) Functions such as _pcre_xclass(), which are internal to the library,
3113 but external in the C sense, all had PCRE_EXPORT in their definitions.
3114 This is apparently wrong for the Windows case, so I have removed it.
3115 (It makes no difference on Unix-like systems.)
3116
3117 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3118 of recursive calls to match(). This is different to MATCH_LIMIT because
3119 that limits the total number of calls to match(), not all of which increase
3120 the depth of recursion. Limiting the recursion depth limits the amount of
3121 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3122 when PCRE is compiled, and changed at run time. A patch from Google adds
3123 this functionality to the C++ interface.
3124
3125 18. Changes to the handling of Unicode character properties:
3126
3127 (a) Updated the table to Unicode 4.1.0.
3128
3129 (b) Recognize characters that are not in the table as "Cn" (undefined).
3130
3131 (c) I revised the way the table is implemented to a much improved format
3132 which includes recognition of ranges. It now supports the ranges that
3133 are defined in UnicodeData.txt, and it also amalgamates other
3134 characters into ranges. This has reduced the number of entries in the
3135 table from around 16,000 to around 3,000, thus reducing its size
3136 considerably. I realized I did not need to use a tree structure after
3137 all - a binary chop search is just as efficient. Having reduced the
3138 number of entries, I extended their size from 6 bytes to 8 bytes to
3139 allow for more data.
3140
3141 (d) Added support for Unicode script names via properties such as \p{Han}.
3142
3143 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3144 matching that character.
3145
3146 20. When matching a repeated Unicode property with a minimum greater than zero,
3147 (for example \pL{2,}), PCRE could look past the end of the subject if it
3148 reached it while seeking the minimum number of characters. This could
3149 happen only if some of the characters were more than one byte long, because
3150 there is a check for at least the minimum number of bytes.
3151
3152 21. Refactored the implementation of \p and \P so as to be more general, to
3153 allow for more different types of property in future. This has changed the
3154 compiled form incompatibly. Anybody with saved compiled patterns that use
3155 \p or \P will have to recompile them.
3156
3157 22. Added "Any" and "L&" to the supported property types.
3158
3159 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3160 but give a compile time error if the value is greater than 0xff.
3161
3162 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3163 accidentally not being installed or uninstalled.
3164
3165 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3166 made were to insert the current release number. This seemed silly, because
3167 it made things harder for people building PCRE on systems that don't run
3168 "configure". I have turned pcre.h into a distributed file, no longer built
3169 by "configure", with the version identification directly included. There is
3170 no longer a pcre.h.in file.
3171
3172 However, this change necessitated a change to the pcre-config script as
3173 well. It is built from pcre-config.in, and one of the substitutions was the
3174 release number. I have updated configure.ac so that ./configure now finds
3175 the release number by grepping pcre.h.
3176
3177 26. Added the ability to run the tests under valgrind.
3178
3179
3180 Version 6.4 05-Sep-05
3181 ---------------------
3182
3183 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3184 "--" to be printed when multiple files were scanned, even when none of the
3185 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3186 consider it to be a bug, and have restored the previous behaviour.
3187
3188 2. A couple of code tidies to get rid of compiler warnings.
3189
3190 3. The pcretest program used to cheat by referring to symbols in the library
3191 whose names begin with _pcre_. These are internal symbols that are not
3192 really supposed to be visible externally, and in some environments it is
3193 possible to suppress them. The cheating is now confined to including
3194 certain files from the library's source, which is a bit cleaner.
3195
3196 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3197 file's purpose clearer.
3198
3199 5. Reorganized pcre_ucp_findchar().
3200
3201
3202 Version 6.3 15-Aug-05
3203 ---------------------
3204
3205 1. The file libpcre.pc.in did not have general read permission in the tarball.
3206
3207 2. There were some problems when building without C++ support:
3208
3209 (a) If C++ support was not built, "make install" and "make test" still
3210 tried to test it.
3211
3212 (b) There were problems when the value of CXX was explicitly set. Some
3213 changes have been made to try to fix these, and ...
3214
3215 (c) --disable-cpp can now be used to explicitly disable C++ support.
3216
3217 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3218 backslash in a target when C++ was disabled. This confuses some
3219 versions of "make", apparently. Using an intermediate variable solves
3220 this. (Same for CPP_LOBJ.)
3221
3222 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3223 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3224 necessary on certain architectures.
3225
3226 4. Added a setting of -export-symbols-regex to the link command to remove
3227 those symbols that are exported in the C sense, but actually are local
3228 within the library, and not documented. Their names all begin with
3229 "_pcre_". This is not a perfect job, because (a) we have to except some
3230 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3231 available (and never for static libraries). I have made a note to try to
3232 find a way round (a) in the future.
3233
3234
3235 Version 6.2 01-Aug-05
3236 ---------------------
3237
3238 1. There was no test for integer overflow of quantifier values. A construction
3239 such as {1111111111111111} would give undefined results. What is worse, if
3240 a minimum quantifier for a parenthesized subpattern overflowed and became
3241 negative, the calculation of the memory size went wrong. This could have
3242 led to memory overwriting.
3243
3244 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3245
3246 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3247 operating environments where this matters.
3248
3249 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3250 PCRE options from within the C++ wrapper.
3251
3252 5. Named capturing subpatterns were not being correctly counted when a pattern
3253 was compiled. This caused two problems: (a) If there were more than 100
3254 such subpatterns, the calculation of the memory needed for the whole
3255 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3256 back references of the form \12, where the number was greater than 9, were
3257 not recognized as back references, even though there were sufficient
3258 previous subpatterns.
3259
3260 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3261 versions of gcc, e.g. 2.95.4.
3262
3263
3264 Version 6.1 21-Jun-05
3265 ---------------------
3266
3267 1. There was one reference to the variable "posix" in pcretest.c that was not
3268 surrounded by "#if !defined NOPOSIX".
3269
3270 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3271 the cross-check on the old pcre_info() function, for the benefit of the
3272 cut-down version of PCRE that is currently imported into Exim.
3273
3274 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3275 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3276 patterns that start (?i) but I don't think that matters. The use of (?i) is
3277 just an example; this all applies to the other options as well.
3278
3279 4. Since libtool seems to echo the compile commands it is issuing, the output
3280 from "make" can be reduced a bit by putting "@" in front of each libtool
3281 compile command.
3282
3283 5. Patch from the folks at Google for configure.in to be a bit more thorough
3284 in checking for a suitable C++ installation before trying to compile the
3285 C++ stuff. This should fix a reported problem when a compiler was present,
3286 but no suitable headers.
3287
3288 6. The man pages all had just "PCRE" as their title. I have changed them to
3289 be the relevant file name. I have also arranged that these names are
3290 retained in the file doc/pcre.txt, which is a concatenation in text format
3291 of all the man pages except the little individual ones for each function.
3292
3293 7. The NON-UNIX-USE file had not been updated for the different set of source
3294 files that come with release 6. I also added a few comments about the C++
3295 wrapper.
3296
3297
3298 Version 6.0 07-Jun-05
3299 ---------------------
3300
3301 1. Some minor internal re-organization to help with my DFA experiments.
3302
3303 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3304 didn't matter for the library itself when fully configured, but did matter
3305 when compiling without UCP support, or within Exim, where the ucp files are
3306 not imported.
3307
3308 3. Refactoring of the library code to split up the various functions into
3309 different source modules. The addition of the new DFA matching code (see
3310 below) to a single monolithic source would have made it really too
3311 unwieldy, quite apart from causing all the code to be include in a
3312 statically linked application, when only some functions are used. This is
3313 relevant even without the DFA addition now that patterns can be compiled in
3314 one application and matched in another.
3315
3316 The downside of splitting up is that there have to be some external
3317 functions and data tables that are used internally in different modules of
3318 the library but which are not part of the API. These have all had their
3319 names changed to start with "_pcre_" so that they are unlikely to clash
3320 with other external names.
3321
3322 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3323 a different (DFA) algorithm. Although it is slower than the original
3324 function, it does have some advantages for certain types of matching
3325 problem.
3326
3327 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3328 including restarting after a partial match.
3329
3330 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3331 defined when compiling for Windows was sent to me. I have put it into the
3332 code, though I have no means of testing or verifying it.
3333
3334 7. Added the pcre_refcount() auxiliary function.
3335
3336 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3337 match before or at the first newline in the subject string. In pcretest,
3338 the /f option on a pattern can be used to set this.
3339
3340 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3341 would behave wrongly. This has been present in PCRE since release 4.0.
3342
3343 10. A number of changes to the pcregrep command:
3344
3345 (a) Refactored how -x works; insert ^(...)$ instead of setting
3346 PCRE_ANCHORED and checking the length, in preparation for adding
3347 something similar for -w.
3348
3349 (b) Added the -w (match as a word) option.
3350
3351 (c) Refactored the way lines are read and buffered so as to have more
3352 than one at a time available.
3353
3354 (d) Implemented a pcregrep test script.
3355
3356 (e) Added the -M (multiline match) option. This allows patterns to match
3357 over several lines of the subject. The buffering ensures that at least
3358 8K, or the rest of the document (whichever is the shorter) is available
3359 for matching (and similarly the previous 8K for lookbehind assertions).
3360
3361 (f) Changed the --help output so that it now says
3362
3363 -w, --word-regex(p)
3364
3365 instead of two lines, one with "regex" and the other with "regexp"
3366 because that confused at least one person since the short forms are the
3367 same. (This required a bit of code, as the output is generated
3368 automatically from a table. It wasn't just a text change.)
3369
3370 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3371 option but starts with a hyphen. Could be a pattern or a path name
3372 starting with a hyphen, for instance.
3373
3374 (h) "-" can be given as a file name to represent stdin.
3375
3376 (i) When file names are being printed, "(standard input)" is used for
3377 the standard input, for compatibility with GNU grep. Previously
3378 "<stdin>" was used.
3379
3380 (j) The option --label=xxx can be used to supply a name to be used for
3381 stdin when file names are being printed. There is no short form.
3382
3383 (k) Re-factored the options decoding logic because we are going to add
3384 two more options that take data. Such options can now be given in four
3385 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3386
3387 (l) Added the -A, -B, and -C options for requesting that lines of context
3388 around matches be printed.
3389
3390 (m) Added the -L option to print the names of files that do not contain
3391 any matching lines, that is, the complement of -l.
3392
3393 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3394 continue to scan other files.
3395
3396 (o) The -s option was incorrectly implemented. For compatibility with other
3397 greps, it now suppresses the error message for a non-existent or non-
3398 accessible file (but not the return code). There is a new option called
3399 -q that suppresses the output of matching lines, which was what -s was
3400 previously doing.
3401
3402 (p) Added --include and --exclude options to specify files for inclusion
3403 and exclusion when recursing.
3404
3405 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3406 Hopefully, it now does.
3407
3408 12. Missing cast in pcre_study().
3409
3410 13. Added an "uninstall" target to the makefile.
3411
3412 14. Replaced "extern" in the function prototypes in Makefile.in with
3413 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3414 world, but is set differently for Windows.
3415
3416 15. Added a second compiling function called pcre_compile2(). The only
3417 difference is that it has an extra argument, which is a pointer to an
3418 integer error code. When there is a compile-time failure, this is set
3419 non-zero, in addition to the error test pointer being set to point to an
3420 error message. The new argument may be NULL if no error number is required
3421 (but then you may as well call pcre_compile(), which is now just a
3422 wrapper). This facility is provided because some applications need a
3423 numeric error indication, but it has also enabled me to tidy up the way
3424 compile-time errors are handled in the POSIX wrapper.
3425
3426 16. Added VPATH=.libs to the makefile; this should help when building with one
3427 prefix path and installing with another. (Or so I'm told by someone who
3428 knows more about this stuff than I do.)
3429
3430 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3431 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3432 match everything, including newlines. This is not POSIX-compatible, but
3433 somebody wanted the feature. From pcretest it can be activated by using
3434 both the P and the s flags.
3435
3436 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3437
3438 19. libpcre.pc was being incorrectly installed as executable.
3439
3440 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3441 it now also looks for '\r' so that it will work unmodified on Windows.
3442
3443 21. Added Google's contributed C++ wrapper to the distribution.
3444
3445 22. Added some untidy missing memory free() calls in pcretest, to keep
3446 Electric Fence happy when testing.
3447
3448
3449
3450 Version 5.0 13-Sep-04
3451 ---------------------
3452
3453 1. Internal change: literal characters are no longer packed up into items
3454 containing multiple characters in a single byte-string. Each character
3455 is now matched using a separate opcode. However, there may be more than one
3456 byte in the character in UTF-8 mode.
3457
3458 2. The pcre_callout_block structure has two new fields: pattern_position and
3459 next_item_length. These contain the offset in the pattern to the next match
3460 item, and its length, respectively.
3461
3462 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3463 insertion of callouts before each pattern item. Added the /C option to
3464 pcretest to make use of this.
3465
3466 4. On the advice of a Windows user, the lines
3467
3468 #if defined(_WIN32) || defined(WIN32)
3469 _setmode( _fileno( stdout ), 0x8000 );
3470 #endif /* defined(_WIN32) || defined(WIN32) */
3471
3472 have been added to the source of pcretest. This apparently does useful
3473 magic in relation to line terminators.
3474
3475 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3476 for the benefit of those environments where the "b" makes a difference.
3477
3478 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3479 to know about it. I have put a hack into configure.in that adds in code
3480 to set GCC=yes if CC=icc. This seems to end up at a point in the
3481 generated configure script that is early enough to affect the setting of
3482 compiler options, which is what is needed, but I have no means of testing
3483 whether it really works. (The user who reported this had patched the
3484 generated configure script, which of course I cannot do.)
3485
3486 LATER: After change 22 below (new libtool files), the configure script
3487 seems to know about icc (and also ecc). Therefore, I have commented out
3488 this hack in configure.in.
3489
3490 7. Added support for pkg-config (2 patches were sent in).
3491
3492 8. Negated POSIX character classes that used a combination of internal tables
3493 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3494 [[:^ascii]]. Typically, they would match almost any characters. The other
3495 POSIX classes were not broken in this way.
3496
3497 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3498 to find the match, as PCRE was deluded into thinking that the match had to
3499 start at the start point or following a newline. The same bug applied to
3500 patterns with negative forward assertions or any backward assertions
3501 preceding ".*" at the start, unless the pattern required a fixed first
3502 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3503
3504 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3505 starting at the last subject character, bytes beyond the end of the subject
3506 string were read.
3507
3508 11. Renamed the variable "class" as "classbits" to make life easier for C++
3509 users. (Previously there was a macro definition, but it apparently wasn't
3510 enough.)
3511
3512 12. Added the new field "tables" to the extra data so that tables can be passed
3513 in at exec time, or the internal tables can be re-selected. This allows
3514 a compiled regex to be saved and re-used at a later time by a different
3515 program that might have everything at different addresses.
3516
3517 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3518 -R library as well as a -L library.
3519
3520 14. The debugging options of pcretest (-d on the command line or D on a
3521 pattern) showed incorrect output for anything following an extended class
3522 that contained multibyte characters and which was followed by a quantifier.
3523
3524 15. Added optional support for general category Unicode character properties
3525 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3526 support. It adds about 90K to the size of the library. The meanings of the
3527 inbuilt class escapes such as \d and \s have NOT been changed.
3528
3529 16. Updated pcredemo.c to include calls to free() to release the memory for the
3530 compiled pattern.
3531
3532 17. The generated file chartables.c was being created in the source directory
3533 instead of in the building directory. This caused the build to fail if the
3534 source directory was different from the building directory, and was
3535 read-only.
3536
3537 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3538 file. No doubt somebody will tell me if they don't make sense... Also added
3539 Dan Mooney's comments about building on OpenVMS.
3540
3541 19. Added support for partial matching via the PCRE_PARTIAL option for
3542 pcre_exec() and the \P data escape in pcretest.
3543
3544 20. Extended pcretest with 3 new pattern features:
3545
3546 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3547 write the compiled pattern to the file whose name is "rest-of-line".
3548 This is a straight binary dump of the data, with the saved pointer to
3549 the character tables forced to be NULL. The study data, if any, is
3550 written too. After writing, pcretest reads a new pattern.
3551
3552 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3553 compiled pattern from the given file. There must not be any
3554 occurrences of "<" in the file name (pretty unlikely); if there are,
3555 pcretest will instead treat the initial "<" as a pattern delimiter.
3556 After reading in the pattern, pcretest goes on to read data lines as
3557 usual.
3558
3559 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3560 and 16-bit fields in a compiled pattern, to simulate a pattern that
3561 was compiled on a host of opposite endianness.
3562
3563 21. The pcre-exec() function can now cope with patterns that were compiled on
3564 hosts of opposite endianness, with this restriction:
3565
3566 As for any compiled expression that is saved and used later, the tables
3567 pointer field cannot be preserved; the extra_data field in the arguments
3568 to pcre_exec() should be used to pass in a tables address if a value
3569 other than the default internal tables were used at compile time.
3570
3571 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3572 now diagnosed as an error. Previously, most of the time, a negative number
3573 would have been treated as zero, but if in addition "ovector" was passed as
3574 NULL, a crash could occur.
3575
3576 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3577 new versions from the libtool 1.5 distribution (the last one is a copy of
3578 a file called libtool.m4). This seems to have fixed the need to patch
3579 "configure" to support Darwin 1.3 (which I used to do). However, I still
3580 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3581 workstation).
3582
3583 24. Changed the PCRE licence to be the more standard "BSD" licence.
3584
3585
3586 Version 4.5 01-Dec-03
3587 ---------------------
3588
3589 1. There has been some re-arrangement of the code for the match() function so
3590 that it can be compiled in a version that does not call itself recursively.
3591 Instead, it keeps those local variables that need separate instances for
3592 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3593 needs to "recurse". Keeping track of where control must go is done by means
3594 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3595 hide most of the details from the main code, and operates only if
3596 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3597 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3598 operating.
3599
3600 To make it easier for callers to provide specially tailored get/free
3601 functions for this usage, two new functions, pcre_stack_malloc, and
3602 pcre_stack_free, are used. They are always called in strict stacking order,
3603 and the size of block requested is always the same.
3604
3605 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3606 PCRE has been compiled to use the stack or the heap for recursion. The
3607 -C option of pcretest uses this to show which version is compiled.
3608
3609 A new data escape \S, is added to pcretest; it causes the amounts of store
3610 obtained and freed by both kinds of malloc/free at match time to be added
3611 to the output.
3612
3613 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3614 what's available on my current Linux desktop machine.
3615
3616 3. When matching a UTF-8 string, the test for a valid string at the start has
3617 been extended. If start_offset is not zero, PCRE now checks that it points
3618 to a byte that is the start of a UTF-8 character. If not, it returns
3619 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3620 this is necessary because there may be backward assertions in the pattern.
3621 When matching the same subject several times, it may save resources to use
3622 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3623
3624 4. The code for checking the validity of UTF-8 strings has been tightened so
3625 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3626 containing "overlong sequences".
3627
3628 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3629 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3630 should have been "&", but it just so happened that all the cases this let
3631 through by mistake were picked up later in the function.
3632
3633 6. I had used a variable called "isblank" - this is a C99 function, causing
3634 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3635
3636 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3637 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3638 so the version is shown; (c) stop "make test" echoing "./RunTest".
3639
3640 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3641
3642 9. The prototype for memmove() for systems that don't have it was using
3643 size_t, but the inclusion of the header that defines size_t was later. I've
3644 moved the #includes for the C headers earlier to avoid this.
3645
3646 10. Added some adjustments to the code to make it easier to compiler on certain
3647 special systems:
3648
3649 (a) Some "const" qualifiers were missing.
3650 (b) Added the macro EXPORT before all exported functions; by default this
3651 is defined to be empty.
3652 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3653 that it reads its output file name as an argument instead of writing
3654 to the standard output and assuming this can be redirected.
3655
3656 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3657 class containing characters with values greater than 255, PCRE compilation
3658 went into a loop.
3659
3660 12. A recursive reference to a subpattern that was within another subpattern
3661 that had a minimum quantifier of zero caused PCRE to crash. For example,
3662 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3663 recursion. If the recursively-called subpattern itself had a zero repeat,
3664 that was OK.
3665
3666 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3667 buffer into which it was copied (for escape processing) was still set at
3668 1024, so long lines caused crashes.
3669
3670 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3671 "internal error: code overflow...". This applied to any character class
3672 that was followed by a possessive quantifier.
3673
3674 15. Modified the Makefile to add libpcre.la as a prerequisite for
3675 libpcreposix.la because I was told this is needed for a parallel build to
3676 work.
3677
3678 16. If a pattern that contained .* following optional items at the start was
3679 studied, the wrong optimizing data was generated, leading to matching
3680 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3681 matching string must start with a or b or c. The correct conclusion for
3682 this pattern is that a match can start with any character.
3683
3684
3685 Version 4.4 13-Aug-03
3686 ---------------------
3687
3688 1. In UTF-8 mode, a character class containing characters with values between
3689 127 and 255 was not handled correctly if the compiled pattern was studied.
3690 In fixing this, I have also improved the studying algorithm for such
3691 classes (slightly).
3692
3693 2. Three internal functions had redundant arguments passed to them. Removal
3694 might give a very teeny performance improvement.
3695
3696 3. Documentation bug: the value of the capture_top field in a callout is *one
3697 more than* the number of the hightest numbered captured substring.
3698
3699 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3700 in incorrectly linking with a previously installed version. They now link
3701 explicitly with libpcre.la.
3702
3703 5. configure.in no longer needs to recognize Cygwin specially.
3704
3705 6. A problem in pcre.in for Windows platforms is fixed.
3706
3707 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3708 pcretest, it used to include the size of the study block as part of its
3709 output. Unfortunately, the structure contains a field that has a different
3710 size on different hardware architectures. This meant that the tests that
3711 showed this size failed. As the block is currently always of a fixed size,
3712 this information isn't actually particularly useful in pcretest output, so
3713 I have just removed it.
3714
3715 8. Three pre-processor statements accidentally did not start in column 1.
3716 Sadly, there are *still* compilers around that complain, even though
3717 standard C has not required this for well over a decade. Sigh.
3718
3719 9. In pcretest, the code for checking callouts passed small integers in the
3720 callout_data field, which is a void * field. However, some picky compilers
3721 complained about the casts involved for this on 64-bit systems. Now
3722 pcretest passes the address of the small integer instead, which should get
3723 rid of the warnings.
3724
3725 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3726 both compile and run time, and gives an error if an invalid UTF-8 sequence
3727 is found. There is a option for disabling this check in cases where the
3728 string is known to be correct and/or the maximum performance is wanted.
3729
3730 11. In response to a bug report, I changed one line in Makefile.in from
3731
3732 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3733 to
3734 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3735
3736 to look similar to other lines, but I have no way of telling whether this
3737 is the right thing to do, as I do not use Windows. No doubt I'll get told
3738 if it's wrong...
3739
3740
3741 Version 4.3 21-May-03
3742 ---------------------
3743
3744 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3745 Makefile.
3746
3747 2. Some refactoring to improve the quality of the code:
3748
3749 (i) The utf8_table... variables are now declared "const".
3750
3751 (ii) The code for \cx, which used the "case flipping" table to upper case
3752 lower case letters, now just substracts 32. This is ASCII-specific,
3753 but the whole concept of \cx is ASCII-specific, so it seems
3754 reasonable.
3755
3756 (iii) PCRE was using its character types table to recognize decimal and
3757 hexadecimal digits in the pattern. This is silly, because it handles
3758 only 0-9, a-f, and A-F, but the character types table is locale-
3759 specific, which means strange things might happen. A private
3760 table is now used for this - though it costs 256 bytes, a table is
3761 much faster than multiple explicit tests. Of course, the standard
3762 character types table is still used for matching digits in subject
3763 strings against \d.
3764
3765 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3766 ending in _t are). So I've renamed it as ESC_tee.
3767
3768 3. The first argument for regexec() in the POSIX wrapper should have been
3769 defined as "const".
3770
3771 4. Changed pcretest to use malloc() for its buffers so that they can be
3772 Electric Fenced for debugging.
3773
3774 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3775 to read one or more bytes before the start of the subject string. Often this
3776 had no effect on PCRE's behaviour, but in some circumstances it could
3777 provoke a segmentation fault.
3778
3779 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3780 to try to read one or more bytes before the start of the subject string.
3781
3782 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3783 UTF-8 support could misbehave in various ways if the subject string
3784 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3785 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3786 back over UTF-8 characters.)
3787
3788
3789 Version 4.2 14-Apr-03
3790 ---------------------
3791
3792 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3793
3794 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3795 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3796 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3797 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3798 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3799 and BUILD_EXEEXT
3800 Note: automatic setting of the BUILD variables is not yet working
3801 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3802 compile-time but not at link-time
3803 [LINK]: use for linking executables only
3804 make different versions for Windows and non-Windows
3805 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3806 libraries
3807 [LINK_FOR_BUILD]: new variable
3808 [OBJEXT]: use throughout
3809 [EXEEXT]: use throughout
3810 <winshared>: new target
3811 <wininstall>: new target
3812 <dftables.o>: use native compiler
3813 <dftables>: use native linker
3814 <install>: handle Windows platform correctly
3815 <clean>: ditto
3816 <check>: ditto
3817 copy DLL to top builddir before testing
3818
3819 As part of these changes, -no-undefined was removed again. This was reported
3820 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3821 in any case.
3822
3823 3. Some tidies to get rid of compiler warnings:
3824
3825 . In the match_data structure, match_limit was an unsigned long int, whereas
3826 match_call_count was an int. I've made them both unsigned long ints.
3827
3828 . In pcretest the fact that a const uschar * doesn't automatically cast to
3829 a void * provoked a warning.
3830
3831 . Turning on some more compiler warnings threw up some "shadow" variables
3832 and a few more missing casts.
3833
3834 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3835 option, a class that contained a single character with a value between 128
3836 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3837
3838 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3839 option, a class that contained several characters, but with at least one
3840 whose value was between 128 and 255 caused PCRE to crash.
3841
3842
3843 Version 4.1 12-Mar-03
3844 ---------------------
3845
3846 1. Compiling with gcc -pedantic found a couple of places where casts were
3847 needed, and a string in dftables.c that was longer than standard compilers are
3848 required to support.
3849
3850 2. Compiling with Sun's compiler found a few more places where the code could
3851 be tidied up in order to avoid warnings.
3852
3853 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3854 first of these names is deprecated in the latest Autoconf in favour of the name
3855 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3856 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3857 analogy I have changed it to CFLAGS_FOR_BUILD.
3858
3859 4. Added -no-undefined to the linking command in the Makefile, because this is
3860 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3861 linking step for the pcreposix library.
3862
3863 5. PCRE was failing to diagnose the case of two named groups with the same
3864 name.
3865
3866 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3867 literal character that is needed in the subject for a match, and scans along to
3868 ensure that it is present before embarking on the full matching process. This
3869 saves time in cases of nested unlimited repeats that are never going to match.
3870 Problem: the scan can take a lot of time if the subject is very long (e.g.
3871 megabytes), thus penalizing straightforward matches. It is now done only if the
3872 amount of subject to be scanned is less than 1000 bytes.
3873
3874 7. A lesser problem with the same optimization is that it was recording the
3875 first character of an anchored pattern as "needed", thus provoking a search
3876 right along the subject, even when the first match of the pattern was going to
3877 fail. The "needed" character is now not set for anchored patterns, unless it
3878 follows something in the pattern that is of non-fixed length. Thus, it still
3879 fulfils its original purpose of finding quick non-matches in cases of nested
3880 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3881
3882
3883 Version 4.0 17-Feb-03
3884 ---------------------
3885
3886 1. If a comment in an extended regex that started immediately after a meta-item
3887 extended to the end of string, PCRE compiled incorrect data. This could lead to
3888 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3889
3890 2. Moved to autoconf 2.53 and libtool 1.4.2.
3891
3892 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3893 the special perltest8 script is no longer needed - all the tests can be run
3894 from a single perltest script.
3895
3896 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3897 by \s. It has now been removed in PCRE. This means it isn't recognized as
3898 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3899 class [:space:] *does* include VT, thereby creating a mess.
3900
3901 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3902 space and tab.
3903
3904 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3905 its new features into the main test script, reducing the number of scripts.
3906
3907 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3908 were backward compatible, and made the (?i) apply to the whole pattern, as if
3909 /i were given. Now it behaves more logically, and applies the option setting
3910 only to what follows. PCRE has been changed to follow suit. However, if it
3911 finds options settings right at the start of the pattern, it extracts them into
3912 the global options, as before. Thus, they show up in the info data.
3913
3914 8. Added support for the \Q...\E escape sequence. Characters in between are
3915 treated as literals. This is slightly different from Perl in that $ and @ are
3916 also handled as literals inside the quotes. In Perl, they will cause variable
3917 interpolation. Note the following examples:
3918
3919 Pattern PCRE matches Perl matches
3920
3921 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3922 \Qabc\$xyz\E abc\$xyz abc\$xyz
3923 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3924
3925 For compatibility with Perl, \Q...\E sequences are recognized inside character
3926 classes as well as outside them.
3927
3928 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3929 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3930 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3931 signed/unsigned warnings.
3932
3933 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3934 option for pcretest, so I've replaced it by a simple function that does just
3935 that job.
3936
3937 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3938 "pcregrep -".
3939
3940 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3941 Java package. This provides some syntactic sugar for simple cases of what my
3942 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3943 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3944 item, you can use this simplified notation. Note that only makes sense with
3945 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3946 greediness, whatever the setting of the PCRE_UNGREEDY option.
3947
3948 13. A change of greediness default within a pattern was not taking effect at
3949 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3950 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3951 was abstracted outside.
3952
3953 14. PCRE now supports the \G assertion. It is true when the current matching
3954 position is at the start point of the match. This differs from \A when the
3955 starting offset is non-zero. Used with the /g option of pcretest (or similar
3956 code), it works in the same way as it does for Perl's /g option. If all
3957 alternatives of a regex begin with \G, the expression is anchored to the start
3958 match position, and the "anchored" flag is set in the compiled expression.
3959
3960 15. Some bugs concerning the handling of certain option changes within patterns
3961 have been fixed. These applied to options other than (?ims). For example,
3962 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3963 been the other way round. Some of this was related to change 7 above.
3964
3965 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3966 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3967 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3968 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3969
3970 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3971 mode. Unlike ".", it always matches newline, whatever the setting of
3972 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3973 assertions. Perl allows it, but it doesn't (in general) work because it can't
3974 calculate the length of the lookbehind. At least, that's the case for Perl
3975 5.8.0 - I've been told they are going to document that it doesn't work in
3976 future.
3977
3978 18. Added an error diagnosis for escapes that PCRE does not support: these are
3979 \L, \l, \N, \P, \p, \U, \u, and \X.
3980
3981 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3982 reading past the end of the pattern in cases such as /[abcd/.
3983
3984 20. PCRE was getting more memory than necessary for patterns with classes that
3985 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3986
3987 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3988 compiling PCRE for use with Virtual Pascal.
3989
3990 22. Small fix to the Makefile to make it work properly if the build is done
3991 outside the source tree.
3992
3993 23. Added a new extension: a condition to go with recursion. If a conditional
3994 subpattern starts with (?(R) the "true" branch is used if recursion has
3995 happened, whereas the "false" branch is used only at the top level.
3996
3997 24. When there was a very long string of literal characters (over 255 bytes
3998 without UTF support, over 250 bytes with UTF support), the computation of how
3999 much memory was required could be incorrect, leading to segfaults or other
4000 strange effects.
4001
4002 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4003 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4004 there was a subsequent back reference to those brackets. This meant that, for
4005 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4006 possible to check for precisely this case. All we can do is abandon the
4007 optimization if .* occurs inside capturing brackets when there are any back
4008 references whatsoever. (See below for a better fix that came later.)
4009
4010 26. The handling of the optimization for finding the first character of a
4011 non-anchored pattern, and for finding a character that is required later in the
4012 match were failing in some cases. This didn't break the matching; it just
4013 failed to optimize when it could. The way this is done has been re-implemented.
4014
4015 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4016
4017 28. Added a new feature that provides some of the functionality that Perl
4018 provides with (?{...}). The facility is termed a "callout". The way it is done
4019 in PCRE is for the caller to provide an optional function, by setting
4020 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4021 global variable. By default it is unset, which disables all calling out. To get
4022 the function called, the regex must include (?C) at appropriate points. This
4023 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4024 This provides a means of identifying different callout points. When PCRE
4025 reaches such a point in the regex, if pcre_callout has been set, the external
4026 function is called. It is provided with data in a structure called
4027 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4028 matching continues; if it returns a non-zero value, the match at the current
4029 point fails. However, backtracking will occur if possible. [This was changed
4030 later and other features added - see item 49 below.]
4031
4032 29. pcretest is upgraded to test the callout functionality. It provides a
4033 callout function that displays information. By default, it shows the start of
4034 the match and the current position in the text. There are some new data escapes
4035 to vary what happens:
4036
4037 \C+ in addition, show current contents of captured substrings
4038 \C- do not supply a callout function
4039 \C!n return 1 when callout number n is reached
4040 \C!n!m return 1 when callout number n is reached for the mth time
4041
4042 30. If pcregrep was called with the -l option and just a single file name, it
4043 output "<stdin>" if a match was found, instead of the file name.
4044
4045 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4046 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4047 pcre_exec(). This saves a malloc/free per call. The default value of
4048 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4049 when configuring.
4050
4051 32. The default maximum size of a compiled pattern is 64K. There have been a
4052 few cases of people hitting this limit. The code now uses macros to handle the
4053 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4054 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4055 configuring. Tests 2 and 5 work only with 2-byte links because they output
4056 debugging information about compiled patterns.
4057
4058 33. Internal code re-arrangements:
4059
4060 (a) Moved the debugging function for printing out a compiled regex into
4061 its own source file (printint.c) and used #include to pull it into
4062 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4063 separate copies.
4064
4065 (b) Defined the list of op-code names for debugging as a macro in
4066 internal.h so that it is next to the definition of the opcodes.
4067
4068 (c) Defined a table of op-code lengths for simpler skipping along compiled
4069 code. This is again a macro in internal.h so that it is next to the
4070 definition of the opcodes.
4071
4072 34. Added support for recursive calls to individual subpatterns, along the
4073 lines of Robin Houston's patch (but implemented somewhat differently).
4074
4075 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4076 allow it to read and process whole directories in Win32. This code was
4077 contributed by Lionel Fourquaux; it has not been tested by me.
4078
4079 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4080 used to name a group. Names consist of alphanumerics and underscores, and must
4081 be unique. Back references use the syntax (?P=name) and recursive calls use
4082 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4083 numbers. The function pcre_fullinfo() can be used after compilation to extract
4084 a name/number map. There are three relevant calls:
4085
4086 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4087 PCRE_INFO_NAMECOUNT yields the number of entries
4088 PCRE_INFO_NAMETABLE yields a pointer to the map.
4089
4090 The map is a vector of fixed-size entries. The size of each entry depends on
4091 the length of the longest name used. The first two bytes of each entry are the
4092 group number, most significant byte first. There follows the corresponding
4093 name, zero terminated. The names are in alphabetical order.
4094
4095 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4096 case instead of 255. Making it the same both with and without UTF-8 support
4097 means that the same test output works with both.
4098
4099 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4100 calling malloc() with a zero argument.
4101
4102 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4103 optimization. I've improved things by keeping a bitmap of backreferences with
4104 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4105 fact referenced, the optimization can be applied. It is unlikely that a
4106 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4107 the match to follow \n) will appear inside brackets with a number greater than
4108 31, but if it does, any back reference > 31 suppresses the optimization.
4109
4110 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4111 of disabling numbered capturing parentheses. Any opening parenthesis that is
4112 not followed by ? behaves as if it were followed by ?: but named parentheses
4113 can still be used for capturing (and they will acquire numbers in the usual
4114 way).
4115
4116 41. Redesigned the return codes from the match() function into yes/no/error so
4117 that errors can be passed back from deep inside the nested calls. A malloc
4118 failure while inside a recursive subpattern call now causes the
4119 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4120
4121 42. It is now possible to set a limit on the number of times the match()
4122 function is called in a call to pcre_exec(). This facility makes it possible to
4123 limit the amount of recursion and backtracking, though not in a directly
4124 obvious way, because the match() function is used in a number of different
4125 circumstances. The count starts from zero for each position in the subject
4126 string (for non-anchored patterns). The default limit is, for compatibility, a
4127 large number, namely 10 000 000. You can change this in two ways:
4128
4129 (a) When configuring PCRE before making, you can use --with-match-limit=n
4130 to set a default value for the compiled library.
4131
4132 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4133 a different value is set. See 45 below.
4134
4135 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4136
4137 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4138 of things that can be changed at compile time. The first argument specifies
4139 what is wanted and the second points to where the information is to be placed.
4140 The current list of available information is:
4141
4142 PCRE_CONFIG_UTF8
4143
4144 The output is an integer that is set to one if UTF-8 support is available;
4145 otherwise it is set to zero.
4146
4147 PCRE_CONFIG_NEWLINE
4148
4149 The output is an integer that it set to the value of the code that is used for
4150 newline. It is either LF (10) or CR (13).
4151
4152 PCRE_CONFIG_LINK_SIZE
4153
4154 The output is an integer that contains the number of bytes used for internal
4155 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4156
4157 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4158
4159 The output is an integer that contains the threshold above which the POSIX
4160 interface uses malloc() for output vectors. See item 31 above.
4161
4162 PCRE_CONFIG_MATCH_LIMIT
4163
4164 The output is an unsigned integer that contains the default limit of the number
4165 of match() calls in a pcre_exec() execution. See 42 above.
4166
4167 44. pcretest has been upgraded by the addition of the -C option. This causes it
4168 to extract all the available output from the new pcre_config() function, and to
4169 output it. The program then exits immediately.
4170
4171 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4172 order to support additional features. One way would have been to define
4173 pcre_exec2() (for example) with extra arguments, but this would not have been
4174 extensible, and would also have required all calls to the original function to
4175 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4176 is used for passing in "extra" data from pcre_study().
4177
4178 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4179 contains the following fields:
4180
4181 flags a bitmap indicating which of the following fields are set
4182 study_data opaque data from pcre_study()
4183 match_limit a way of specifying a limit on match() calls for a specific
4184 call to pcre_exec()
4185 callout_data data for callouts (see 49 below)
4186
4187 The flag bits are also defined in pcre.h, and are
4188
4189 PCRE_EXTRA_STUDY_DATA
4190 PCRE_EXTRA_MATCH_LIMIT
4191 PCRE_EXTRA_CALLOUT_DATA
4192
4193 The pcre_study() function now returns one of these new pcre_extra blocks, with
4194 the actual study data pointed to by the study_data field, and the
4195 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4196 before. That is, this change is entirely upwards-compatible and requires no
4197 change to existing code.
4198
4199 If you want to pass in additional data to pcre_exec(), you can either place it
4200 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4201 block.
4202
4203 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4204 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4205 times with different match limits, until it finds the minimum value needed for
4206 pcre_exec() to complete. The value is then output. This can be instructive; for
4207 most simple matches the number is quite small, but for pathological cases it
4208 gets very large very quickly.
4209
4210 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4211 returns the size of the data block pointed to by the study_data field in a
4212 pcre_extra block, that is, the value that was passed as the argument to
4213 pcre_malloc() when PCRE was getting memory in which to place the information
4214 created by pcre_study(). The fourth argument should point to a size_t variable.
4215 pcretest has been extended so that this information is shown after a successful
4216 pcre_study() call when information about the compiled regex is being displayed.
4217
4218 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4219 because what follows is always an absolute path. (Later: it turns out that this
4220 is more than cosmetic for MinGW, because it doesn't like empty path
4221 components.)
4222
4223 49. Some changes have been made to the callout feature (see 28 above):
4224
4225 (i) A callout function now has three choices for what it returns:
4226
4227 0 => success, carry on matching
4228 > 0 => failure at this point, but backtrack if possible
4229 < 0 => serious error, return this value from pcre_exec()
4230
4231 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4232 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4233 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4234 use by callout functions. It will never be used by PCRE itself.
4235
4236 (ii) The pcre_extra structure (see 45 above) has a void * field called
4237 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4238 pcre_callout_block structure has a field of the same name. The contents of
4239 the field passed in the pcre_extra structure are passed to the callout
4240 function in the corresponding field in the callout block. This makes it
4241 easier to use the same callout-containing regex from multiple threads. For
4242 testing, the pcretest program has a new data escape
4243
4244 \C*n pass the number n (may be negative) as callout_data
4245
4246 If the callout function in pcretest receives a non-zero value as
4247 callout_data, it returns that value.
4248
4249 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4250 there were some redundant $(CFLAGS) in commands that are now specified as
4251 $(LINK), which already includes $(CFLAGS).
4252
4253 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4254 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4255 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4256 one-byte characters throughout. Note that case-insensitive matching applies
4257 only to characters whose values are less than 256. PCRE doesn't support the
4258 notion of cases for higher-valued characters.
4259
4260 (i) A character class whose characters are all within 0-255 is handled as
4261 a bit map, and the map is inverted for negative classes. Previously, a
4262 character > 255 always failed to match such a class; however it should
4263 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4264
4265 (ii) A negated character class with a single character < 255 is coded as
4266 "not this character" (OP_NOT). This wasn't working properly when the test
4267 character was multibyte, either singly or repeated.
4268
4269 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4270 mode, for example: \x{100}{2,3}.
4271
4272 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4273 singly or repeated) now correctly test multibyte characters. However,
4274 PCRE doesn't recognize any characters with values greater than 255 as
4275 digits, spaces, or word characters. Such characters always match \D, \S,
4276 and \W, and never match \d, \s, or \w.
4277
4278 (v) Classes may now contain characters and character ranges with values
4279 greater than 255. For example: [ab\x{100}-\x{400}].
4280
4281 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4282 PCRE in UTF-8 mode.
4283
4284 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4285 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4286 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4287 value.)
4288
4289 53. The single man page has become too large. I have therefore split it up into
4290 a number of separate man pages. These also give rise to individual HTML pages;
4291 these are now put in a separate directory, and there is an index.html page that
4292 lists them all. Some hyperlinking between the pages has been installed.
4293
4294 54. Added convenience functions for handling named capturing parentheses.
4295
4296 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4297 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4298 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4299 are faulted.
4300
4301 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4302 calling configure. These values are used when compiling the dftables.c program
4303 which is run to generate the source of the default character tables. They
4304 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4305 you will need to set these values.
4306
4307 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4308
4309
4310 Version 3.9 02-Jan-02
4311 ---------------------
4312
4313 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4314
4315 2. If --disable-static was given, the building process failed when trying to
4316 build pcretest and pcregrep. (For some reason it was using libtool to compile
4317 them, which is not right, as they aren't part of the library.)
4318
4319
4320 Version 3.8 18-Dec-01
4321 ---------------------
4322
4323 1. The experimental UTF-8 code was completely screwed up. It was packing the
4324 bytes in the wrong order. How dumb can you get?
4325
4326
4327 Version 3.7 29-Oct-01
4328 ---------------------
4329
4330 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4331 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4332 this didn't happen under Solaris 8, where I normally test things.
4333
4334 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4335 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4336 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4337 this fix an hour or so after the initial 3.7 release.)
4338
4339
4340 Version 3.6 23-Oct-01
4341 ---------------------
4342
4343 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4344 offsets passed as NULL with zero offset count.
4345
4346 2. The config.guess and config.sub files had not been updated when I moved to
4347 the latest autoconf.
4348
4349
4350 Version 3.5 15-Aug-01
4351 ---------------------
4352
4353 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4354 had been forgotten.
4355
4356 2. By using declared but undefined structures, we can avoid using "void"
4357 definitions in pcre.h while keeping the internal definitions of the structures
4358 private.
4359
4360 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4361 user point of view, this means that both static and shared libraries are built
4362 by default, but this can be individually controlled. More of the work of
4363 handling this static/shared cases is now inside libtool instead of PCRE's make
4364 file.
4365
4366 4. The pcretest utility is now installed along with pcregrep because it is
4367 useful for users (to test regexs) and by doing this, it automatically gets
4368 relinked by libtool. The documentation has been turned into a man page, so
4369 there are now .1, .txt, and .html versions in /doc.
4370
4371 5. Upgrades to pcregrep:
4372 (i) Added long-form option names like gnu grep.
4373 (ii) Added --help to list all options with an explanatory phrase.
4374 (iii) Added -r, --recursive to recurse into sub-directories.
4375 (iv) Added -f, --file to read patterns from a file.
4376
4377 6. pcre_exec() was referring to its "code" argument before testing that
4378 argument for NULL (and giving an error if it was NULL).
4379
4380 7. Upgraded Makefile.in to allow for compiling in a different directory from
4381 the source directory.
4382
4383 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4384 options bits, the pointer it was passed was to an int instead of to an unsigned
4385 long int. This mattered only on 64-bit systems.
4386
4387 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4388 generated) instead of pcre.in, which it its source. Also made the same change
4389 in several of the .c files.
4390
4391 10. A new release of gcc defines printf() as a macro, which broke pcretest
4392 because it had an ifdef in the middle of a string argument for printf(). Fixed
4393 by using separate calls to printf().
4394
4395 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4396 script, to force use of CR or LF instead of \n in the source. On non-Unix
4397 systems, the value can be set in config.h.
4398
4399 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4400 absolute limit. Changed the text of the error message to make this clear, and
4401 likewise updated the man page.
4402
4403 13. The limit of 99 on the number of capturing subpatterns has been removed.
4404 The new limit is 65535, which I hope will not be a "real" limit.
4405
4406
4407 Version 3.4 22-Aug-00
4408 ---------------------
4409
4410 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4411
4412 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4413
4414
4415 Version 3.3 01-Aug-00
4416 ---------------------
4417
4418 1. If an octal character was given, but the value was greater than \377, it
4419 was not getting masked to the least significant bits, as documented. This could
4420 lead to crashes in some systems.
4421
4422 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4423 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4424
4425 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4426 These just pass their arguments on to (pcre_free)(), but they are provided
4427 because some uses of PCRE bind it to non-C systems that can call its functions,
4428 but cannot call free() or pcre_free() directly.
4429
4430 4. Add "make test" as a synonym for "make check". Corrected some comments in
4431 the Makefile.
4432
4433 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4434 Makefile.
4435
4436 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4437 command called pgrep for grepping around the active processes.
4438
4439 7. Added the beginnings of support for UTF-8 character strings.
4440
4441 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4442 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4443 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4444 out for the ar command.)
4445
4446
4447 Version 3.2 12-May-00
4448 ---------------------
4449
4450 This is purely a bug fixing release.
4451
4452 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4453 of ZA. This was just one example of several cases that could provoke this bug,
4454 which was introduced by change 9 of version 2.00. The code for breaking
4455 infinite loops after an iteration that matches an empty string was't working
4456 correctly.
4457
4458 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4459 when matched against abbab (for example). After matching an empty string, it
4460 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4461 caused it to match further down the string than it should.
4462
4463 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4464 was there because it doesn't seem to be needed, and it causes trouble on some
4465 systems, as it is not a Standard C header. It has been removed.
4466
4467 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4468 were reported on the Macintosh. The changes were from
4469
4470 while ((c = *(++ptr)) != 0 && c != '\n');
4471 to
4472 while ((c = *(++ptr)) != 0 && c != '\n') ;
4473
4474 Totally extraordinary, but if that's what it takes...
4475
4476 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4477 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4478 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4479 assumes the way PCRE uses memmove() (always moving upwards).
4480
4481 6. PCRE is being used in one environment where strchr() is not available. There
4482 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4483 faster code anyway.
4484
4485
4486 Version 3.1 09-Feb-00
4487 ---------------------
4488
4489 The only change in this release is the fixing of some bugs in Makefile.in for
4490 the "install" target:
4491
4492 (1) It was failing to install pcreposix.h.
4493
4494 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4495
4496
4497 Version 3.0 01-Feb-00
4498 ---------------------
4499
4500 1. Add support for the /+ modifier to perltest (to output $` like it does in
4501 pcretest).
4502
4503 2. Add support for the /g modifier to perltest.
4504
4505 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4506 matches null strings.
4507
4508 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4509 pattern. Perl treats empty patterns specially - it reuses the most recent
4510 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4511 effect.
4512
4513 5. The POSIX interface was broken in that it was just handing over the POSIX
4514 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4515 required a bigger vector, with some working space on the end. This means that
4516 the POSIX wrapper now has to get and free some memory, and copy the results.
4517
4518 6. Added some simple autoconf support, placing the test data and the
4519 documentation in separate directories, re-organizing some of the
4520 information files, and making it build pcre-config (a GNU standard). Also added
4521 libtool support for building PCRE as a shared library, which is now the
4522 default.
4523
4524 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4525 09 are not valid octal constants. Single digits will be used for minor values
4526 less than 10.
4527
4528 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4529 existing programs that set these in the POSIX interface can use PCRE without
4530 modification.
4531
4532 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4533 return all that pcre_info() returns, plus additional data. The pcre_info()
4534 function is retained for compatibility, but is considered to be obsolete.
4535
4536 10. Added experimental recursion feature (?R) to handle one common case that
4537 Perl 5.6 will be able to do with (?p{...}).
4538
4539 11. Added support for POSIX character classes like [:alpha:], which Perl is
4540 adopting.
4541
4542
4543 Version 2.08 31-Aug-99
4544 ----------------------
4545
4546 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4547 trying to match at the startoffset position, but instead was moving forward to
4548 the next newline as if a previous match had failed.
4549
4550 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4551 and could get into a loop if a null string was matched other than at the start
4552 of the subject.
4553
4554 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4555 be distinguished at compile time, and for completeness also added PCRE_DATE.
4556
4557 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4558 in GnuWin32 environments.
4559
4560
4561 Version 2.07 29-Jul-99
4562 ----------------------
4563
4564 1. The documentation is now supplied in plain text form and HTML as well as in
4565 the form of man page sources.
4566
4567 2. C++ compilers don't like assigning (void *) values to other pointer types.
4568 In particular this affects malloc(). Although there is no problem in Standard
4569 C, I've put in casts to keep C++ compilers happy.
4570
4571 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4572 should be (const char *).
4573
4574 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4575 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4576 However, I haven't made this a standard facility. The documentation doesn't
4577 mention it, and the Makefile doesn't support it.
4578
4579 5. The Makefile now contains an "install" target, with editable destinations at
4580 the top of the file. The pcretest program is not installed.
4581
4582 6. pgrep -V now gives the PCRE version number and date.
4583
4584 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4585 causing the entire string to be ignored, instead of just the last character.
4586
4587 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4588 non-matching string, it can take a very, very long time, even for strings of
4589 quite modest length, because of the nested recursion. PCRE now does better in
4590 some of these cases. It does this by remembering the last required literal
4591 character in the pattern, and pre-searching the subject to ensure it is present
4592 before running the real match. In other words, it applies a heuristic to detect
4593 some types of certain failure quickly, and in the above example, if presented
4594 with a string that has no trailing " it gives "no match" very quickly.
4595
4596 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4597 other alternatives are tried instead.
4598
4599
4600 Version 2.06 09-Jun-99
4601 ----------------------
4602
4603 1. Change pcretest's output for amount of store used to show just the code
4604 space, because the remainder (the data block) varies in size between 32-bit and
4605 64-bit systems.
4606
4607 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4608 start matching at. This allows lookbehinds to work when searching for multiple
4609 occurrences in a string.
4610
4611 3. Added additional options to pcretest for testing multiple occurrences:
4612
4613 /+ outputs the rest of the string that follows a match
4614 /g loops for multiple occurrences, using the new startoffset argument
4615 /G loops for multiple occurrences by passing an incremented pointer
4616
4617 4. PCRE wasn't doing the "first character" optimization for patterns starting
4618 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4619 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4620 the letter 'x'. On long subject strings, this gives a significant speed-up.
4621
4622
4623 Version 2.05 21-Apr-99
4624 ----------------------
4625
4626 1. Changed the type of magic_number from int to long int so that it works
4627 properly on 16-bit systems.
4628
4629 2. Fixed a bug which caused patterns starting with .* not to work correctly
4630 when the subject string contained newline characters. PCRE was assuming
4631 anchoring for such patterns in all cases, which is not correct because .* will
4632 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4633 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4634 must be retried after every newline in the subject.
4635
4636
4637 Version 2.04 18-Feb-99
4638 ----------------------
4639
4640 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4641 computation of the store needed to hold the pattern was incorrect (too large).
4642 If such patterns were nested a few deep, this could multiply and become a real
4643 problem.
4644
4645 2. Added /M option to pcretest to show the memory requirement of a specific
4646 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4647
4648 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4649 compiled in such a way that the backtracking after subsequent failure was
4650 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4651 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4652
4653
4654 Version 2.03 02-Feb-99
4655 ----------------------
4656
4657 1. Fixed typo and small mistake in man page.
4658
4659 2. Added 4th condition (GPL supersedes if conflict) and created separate
4660 LICENCE file containing the conditions.
4661
4662 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4663 Perl, that is the internal \ allows the delimiter to be included in the
4664 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4665 the final delimiter, add \ to the end of the pattern (to test the error).
4666
4667 4. Added the convenience functions for extracting substrings after a successful
4668 match. Updated pcretest to make it able to test these functions.
4669
4670
4671 Version 2.02 14-Jan-99
4672 ----------------------
4673
4674 1. Initialized the working variables associated with each extraction so that
4675 their saving and restoring doesn't refer to uninitialized store.
4676
4677 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4678 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4679 fix the problem.
4680
4681 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4682 calls, and wasn't printing the correct value for compiling calls. Increased the
4683 default value of LOOPREPEAT, and the number of significant figures in the
4684 times.
4685
4686 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4687
4688 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4689 a building problem on Windows NT with a FAT file system.
4690
4691
4692 Version 2.01 21-Oct-98
4693 ----------------------
4694
4695 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4696 to character tables built by pcre_maketables() in the current locale. If NULL
4697 is passed, the default tables are used.
4698
4699
4700 Version 2.00 24-Sep-98
4701 ----------------------
4702
4703 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4704 it any more.
4705
4706 2. Allow quantification of (?>) groups, and make it work correctly.
4707
4708 3. The first character computation wasn't working for (?>) groups.
4709
4710 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4711 end of the subject) and add 5.005's \z, which really does match only at the
4712 very end of the subject.
4713
4714 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4715
4716 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4717 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4718 localized options. All options to pcre_study() were also removed.
4719
4720 7. Add other new features from 5.005:
4721
4722 $(?<= positive lookbehind
4723 $(?<! negative lookbehind
4724 (?imsx-imsx) added the unsetting capability
4725 such a setting is global if at outer level; local otherwise
4726 (?imsx-imsx:) non-capturing groups with option setting
4727 (?(cond)re|re) conditional pattern matching
4728
4729 A backreference to itself in a repeated group matches the previous
4730 captured string.
4731
4732 8. General tidying up of studying (both automatic and via "study")
4733 consequential on the addition of new assertions.
4734
4735 9. As in 5.005, unlimited repeated groups that could match an empty substring
4736 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4737 runtime if any iteration does actually match an empty substring.
4738
4739 10. Include the RunTest script in the distribution.
4740
4741 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4742 discrepancies, some of which were old and were also with respect to 5.004. They
4743 have now been fixed.
4744
4745
4746 Version 1.09 28-Apr-98
4747 ----------------------
4748
4749 1. A negated single character class followed by a quantifier with a minimum
4750 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4751 program crashes, or just wrong answers. This did not apply to negated classes
4752 containing more than one character, or to minima other than one.
4753
4754
4755 Version 1.08 27-Mar-98
4756 ----------------------
4757
4758 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4759
4760 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4761 latter must appear before anything that relies on it in the pattern.
4762
4763
4764 Version 1.07 16-Feb-98
4765 ----------------------
4766
4767 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4768 repeat of a potentially empty string).
4769
4770
4771 Version 1.06 23-Jan-98
4772 ----------------------
4773
4774 1. Added Markus Oberhumer's little patches for C++.
4775
4776 2. Literal strings longer than 255 characters were broken.
4777
4778
4779 Version 1.05 23-Dec-97
4780 ----------------------
4781
4782 1. Negated character classes containing more than one character were failing if
4783 PCRE_CASELESS was set at run time.
4784
4785
4786 Version 1.04 19-Dec-97
4787 ----------------------
4788
4789 1. Corrected the man page, where some "const" qualifiers had been omitted.
4790
4791 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4792 input syntax.
4793
4794 3. Fixed memory leak which occurred when a regex with back references was
4795 matched with an offsets vector that wasn't big enough. The temporary memory
4796 that is used in this case wasn't being freed if the match failed.
4797
4798 4. Tidied pcretest to ensure it frees memory that it gets.
4799
4800 5. Temporary memory was being obtained in the case where the passed offsets
4801 vector was exactly big enough.
4802
4803 6. Corrected definition of offsetof() from change 5 below.
4804
4805 7. I had screwed up change 6 below and broken the rules for the use of
4806 setjmp(). Now fixed.
4807
4808
4809 Version 1.03 18-Dec-97
4810 ----------------------
4811
4812 1. A erroneous regex with a missing opening parenthesis was correctly
4813 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4814 on some systems.
4815
4816 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4817 it was reported that one broken compiler failed on the former because "code" is
4818 also an independent variable.
4819
4820 3. The erroneous regex a[]b caused an array overrun reference.
4821
4822 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4823 fail on data ending with that character. (It was going on too far, and checking
4824 the next character, typically a binary zero.) This was specific to the
4825 optimized code for single-character negative classes.
4826
4827 5. Added a contributed patch from the TIN world which does the following:
4828
4829 + Add an undef for memmove, in case the the system defines a macro for it.
4830
4831 + Add a definition of offsetof(), in case there isn't one. (I don't know
4832 the reason behind this - offsetof() is part of the ANSI standard - but
4833 it does no harm).
4834
4835 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4836 most of the places where whitespace preceded '#'. I have given up and
4837 allowed the remaining 2 cases to be at the margin.
4838
4839 + Rename some variables in pcre to eliminate shadowing. This seems very
4840 pedantic, but does no harm, of course.
4841
4842 6. Moved the call to setjmp() into its own function, to get rid of warnings
4843 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4844
4845 7. Constructs such as \d{8,} were compiling into the equivalent of
4846 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4847 outcome, but in this particular case used more store than had been allocated,
4848 which caused the bug to be discovered because it threw up an internal error.
4849
4850 8. The debugging code in both pcre and pcretest for outputting the compiled
4851 form of a regex was going wrong in the case of back references followed by
4852 curly-bracketed repeats.
4853
4854
4855 Version 1.02 12-Dec-97
4856 ----------------------
4857
4858 1. Typos in pcre.3 and comments in the source fixed.
4859
4860 2. Applied a contributed patch to get rid of places where it used to remove
4861 'const' from variables, and fixed some signed/unsigned and uninitialized
4862 variable warnings.
4863
4864 3. Added the "runtest" target to Makefile.
4865
4866 4. Set default compiler flag to -O2 rather than just -O.
4867
4868
4869 Version 1.01 19-Nov-97
4870 ----------------------
4871
4872 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4873 like /([ab]*)*/, that is, for classes with more than one character in them.
4874
4875 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4876 as /((?>a*))*/ (a PCRE_EXTRA facility).
4877
4878
4879 Version 1.00 18-Nov-97
4880 ----------------------
4881
4882 1. Added compile-time macros to support systems such as SunOS4 which don't have
4883 memmove() or strerror() but have other things that can be used instead.
4884
4885 2. Arranged that "make clean" removes the executables.
4886
4887
4888 Version 0.99 27-Oct-97
4889 ----------------------
4890
4891 1. Fixed bug in code for optimizing classes with only one character. It was
4892 initializing a 32-byte map regardless, which could cause it to run off the end
4893 of the memory it had got.
4894
4895 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4896
4897
4898 Version 0.98 22-Oct-97
4899 ----------------------
4900
4901 1. Fixed bug in code for handling temporary memory usage when there are more
4902 back references than supplied space in the ovector. This could cause segfaults.
4903
4904
4905 Version 0.97 21-Oct-97
4906 ----------------------
4907
4908 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4909
4910 2. Optimized negated single characters not to use a bit map.
4911
4912 3. Brought error texts together as macro definitions; clarified some of them;
4913 fixed one that was wrong - it said "range out of order" when it meant "invalid
4914 escape sequence".
4915
4916 4. Changed some char * arguments to const char *.
4917
4918 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4919
4920 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4921 pcretest.
4922
4923
4924 Version 0.96 16-Oct-97
4925 ----------------------
4926
4927 1. Added a simple "pgrep" utility to the distribution.
4928
4929 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4930 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4931 where "ddd" means "one or more decimal digits".
4932
4933 3. Fixed serious bug. If a pattern had a back reference, but the call to
4934 pcre_exec() didn't supply a large enough ovector to record the related
4935 identifying subpattern, the match always failed. PCRE now remembers the number
4936 of the largest back reference, and gets some temporary memory in which to save
4937 the offsets during matching if necessary, in order to ensure that
4938 backreferences always work.
4939
4940 4. Increased the compatibility with Perl in a number of ways:
4941
4942 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4943 to request this handling. The option can be set at compile or exec time.
4944
4945 (b) $ matches before a terminating newline by default; an option
4946 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4947 mode). The option can be set at compile or exec time.
4948
4949 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4950 the same as Perl's. If the decimal number it represents is less than 10
4951 or there aren't that many previous left capturing parentheses, an octal
4952 escape is read. Inside a character class, it's always an octal escape,
4953 even if it is a single digit.
4954
4955 (d) An escaped but undefined alphabetic character is taken as a literal,
4956 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4957 escapes.
4958
4959 (e) {0} is now permitted. (The previous item is removed from the compiled
4960 pattern).
4961
4962 5. Changed all the names of code files so that the basic parts are no longer
4963 than 10 characters, and abolished the teeny "globals.c" file.
4964
4965 6. Changed the handling of character classes; they are now done with a 32-byte
4966 bit map always.
4967
4968 7. Added the -d and /D options to pcretest to make it possible to look at the
4969 internals of compilation without having to recompile pcre.
4970
4971
4972 Version 0.95 23-Sep-97
4973 ----------------------
4974
4975 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4976 \x20 at the start of a run of normal characters. These were being treated as
4977 real characters, instead of the source characters being re-checked.
4978
4979
4980 Version 0.94 18-Sep-97
4981 ----------------------
4982
4983 1. The functions are now thread-safe, with the caveat that the global variables
4984 containing pointers to malloc() and free() or alternative functions are the
4985 same for all threads.
4986
4987 2. Get pcre_study() to generate a bitmap of initial characters for non-
4988 anchored patterns when this is possible, and use it if passed to pcre_exec().
4989
4990
4991 Version 0.93 15-Sep-97
4992 ----------------------
4993
4994 1. /(b)|(:+)/ was computing an incorrect first character.
4995
4996 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4997 but not actually doing anything yet.
4998
4999 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5000 as Perl does (e.g. [-az] or [az-]).
5001
5002 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5003 all possible positions.
5004
5005 5. Split up into different modules to avoid including unneeded functions in a
5006 compiled binary. However, compile and exec are still in one module. The "study"
5007 function is split off.
5008
5009 6. The character tables are now in a separate module whose source is generated
5010 by an auxiliary program - but can then be edited by hand if required. There are
5011 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5012 toupper() in the code.
5013
5014 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5015 make them global. Abolish the function for setting them, as the caller can now
5016 set them directly.
5017
5018
5019 Version 0.92 11-Sep-97
5020 ----------------------
5021
5022 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5023 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5024
5025 2. Caseless matching was not working in character classes if the characters in
5026 the pattern were in upper case.
5027
5028 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5029
5030 4. Make PCRE_ANCHORED public and accept as a compile option.
5031
5032 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5033 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5034 pass them.
5035
5036 6. Give an error if bad option bits passed at compile or run time.
5037
5038 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5039 pcretest to cause it to pass that flag.
5040
5041 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5042 options, and the first character, if set.
5043
5044 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5045
5046
5047 Version 0.91 10-Sep-97
5048 ----------------------
5049
5050 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5051 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5052
5053 2. PCRE was looping on encountering an indefinitely repeated back reference to
5054 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5055 Perl does - treats the match as successful.
5056
5057 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5