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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1462 - (show annotations)
Mon Mar 10 16:28:54 2014 UTC (5 years, 9 months ago) by ph10
File size: 263734 byte(s)
Error occurred while calculating annotation data.
Fix typo that caused max/min quantifiers in ungreedy mode not to work.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.35-RC1 04-March-2014
5 ------------------------------
6
7 1. A new flag is set, when property checks are present in an XCLASS.
8 When this flag is not set, PCRE can perform certain optimizations
9 such as studying these XCLASS-es.
10
11 2. The auto-possessification of character sets were improved: a normal
12 and an extended character set can be compared now. Furthermore
13 the JIT compiler optimizes more character set checks.
14
15 3. Got rid of some compiler warnings for potentially uninitialized variables
16 that show up only when compiled with -O2.
17
18 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
19 of a match later then the end of the match. The pcretest program was not
20 handling the case sensibly - it was outputting from the start to the next
21 binary zero. It now reports this situation in a message, and outputs the
22 text from the end to the start.
23
24 5. Fast forward search is improved in JIT. Instead of the first three
25 characters, any three characters with fixed position can be searched.
26 Search order: first, last, middle.
27
28 6. Improve character range checks in JIT. Characters are read by an inprecise
29 function now, which returns with an unknown value if the character code is
30 above a certain treshold (e.g: 256). The only limitation is that the value
31 must be bigger than the treshold as well. This function is useful, when
32 the characters above the treshold are handled in the same way.
33
34 7. The macros whose names start with RAWUCHAR are placeholders for a future
35 mode in which only the bottom 21 bits of 32-bit data items are used. To
36 make this more memorable for those maintaining the code, the names have
37 been changed to start with UCHAR21, and an extensive comment has been added
38 to their definition.
39
40 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
41 to the export list in Makefile.am (they were accidentally omitted from the
42 8.34 tarball).
43
44 9. The informational output from pcretest used the phrase "starting byte set"
45 which is inappropriate for the 16-bit and 32-bit libraries. As the output
46 for "first char" and "need char" really means "non-UTF-char", I've changed
47 "byte" to "char", and slightly reworded the output. The documentation about
48 these values has also been (I hope) clarified.
49
50 10. Another JIT related optimization: use table jumps for selecting the correct
51 backtracking path, when more than four alternatives are present inside a
52 bracket.
53
54 11. Empty match is not possible, when the minimum length is greater than zero,
55 and there is no \K in the pattern. JIT should avoid empty match checks in
56 such cases.
57
58 12. In a caseless character class with UCP support, when a character with more
59 than one alternative case was not the first character of a range, not all
60 the alternative cases were added to the class. For example, s and \x{17f}
61 are both alternative cases for S: the class [RST] was handled correctly,
62 but [R-T] was not.
63
64 13. The configure.ac file always checked for pthread support when JIT was
65 enabled. This is not used in Windows, so I have put this test inside a
66 check for the presence of windows.h (which was already tested for).
67
68 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
69 The algorithm provides a way to skip certain starting offsets, and usually
70 faster than linear prefix searches.
71
72 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
73 as for 'fr_FR' and 'french'. For some reason, however, it then used the
74 Windows-specific input and output files, which have 'french' screwed in.
75 So this could never have worked. One of the problems with locales is that
76 they aren't always the same. I have now updated RunTest so that it checks
77 the output of the locale test (test 3) against three different output
78 files, and it allows the test to pass if any one of them matches. With luck
79 this should make the test pass on some versions of Solaris where it was
80 failing. Because of the uncertainty, the script did not used to stop if
81 test 3 failed; it now does. If further versions of a French locale ever
82 come to light, they can now easily be added.
83
84 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
85 there was a message during ./configure, but it did not stop. This now
86 provokes an error. The invalid example in README has been corrected.
87 If a value less than the minimum is given, the minimum value has always
88 been used, but now a warning is given.
89
90 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
91 was a bug in the test system, which is now fixed. Also, the list of various
92 configurations that are tested for each release did not have one with both
93 16/32 bits and --enable-bar-anycrlf. It now does.
94
95 18. pcretest was missing "-C bsr" for displaying the \R default setting.
96
97 19. Little endian PowerPC systems are supported now by the JIT compiler.
98
99 20. The fast forward newline mechanism could enter to an infinite loop on
100 certain invalid UTF-8 input. Although we don't support these cases
101 this issue can be fixed by a performance optimization.
102
103 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
104 not take account if existing stack usage. There is now a new global
105 variable called pcre_stack_guard that can be set to point to an external
106 function to check stack availability. It is called at the start of
107 processing every parenthesized group.
108
109 22. A typo in the code meant that in ungreedy mode the max/min qualifier
110 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
111 not match "ab".
112
113
114 Version 8.34 15-December-2013
115 -----------------------------
116
117 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
118 executable memory. Patch inspired by Carsten Klein.
119
120 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
121 this macro is never tested and has no effect, because the work to support
122 coverage involves only compiling and linking options and special targets in
123 the Makefile. The comment in config.h implied that defining the macro would
124 enable coverage support, which is totally false. There was also support for
125 setting this macro in the CMake files (my fault, I just copied it from
126 configure). SUPPORT_GCOV has now been removed.
127
128 3. Make a small performance improvement in strlen16() and strlen32() in
129 pcretest.
130
131 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
132 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
133 these cases). There was also one in pcretest.c.
134
135 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
136
137 6. In UTF mode, the code for checking whether a group could match an empty
138 string (which is used for indefinitely repeated groups to allow for
139 breaking an infinite loop) was broken when the group contained a repeated
140 negated single-character class with a character that occupied more than one
141 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
142 UTF-8 mode). The effect was undefined: the group might or might not be
143 deemed as matching an empty string, or the program might have crashed.
144
145 7. The code for checking whether a group could match an empty string was not
146 recognizing that \h, \H, \v, \V, and \R must match a character.
147
148 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
149 an empty string. If it can, pcretest shows this in its information output.
150
151 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
152 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
153 matched by pcre_exec() without using JIT:
154
155 (a) If the rest of the pattern did not match after a maximal run of
156 grapheme clusters, the code for backing up to try with fewer of them
157 did not always back up over a full grapheme when characters that do not
158 have the modifier quality were involved, e.g. Hangul syllables.
159
160 (b) If the match point in a subject started with modifier character, and
161 there was no match, the code could incorrectly back up beyond the match
162 point, and potentially beyond the first character in the subject,
163 leading to a segfault or an incorrect match result.
164
165 10. A conditional group with an assertion condition could lead to PCRE
166 recording an incorrect first data item for a match if no other first data
167 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
168 first data item, and therefore matched "ca" after "c" instead of at the
169 start.
170
171 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
172 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
173
174 12. The source of pcregrep now includes z/OS-specific code so that it can be
175 compiled for z/OS as part of the special z/OS distribution.
176
177 13. Added the -T and -TM options to pcretest.
178
179 14. The code in pcre_compile.c for creating the table of named capturing groups
180 has been refactored. Instead of creating the table dynamically during the
181 actual compiling pass, the information is remembered during the pre-compile
182 pass (on the stack unless there are more than 20 named groups, in which
183 case malloc() is used) and the whole table is created before the actual
184 compile happens. This has simplified the code (it is now nearly 150 lines
185 shorter) and prepared the way for better handling of references to groups
186 with duplicate names.
187
188 15. A back reference to a named subpattern when there is more than one of the
189 same name now checks them in the order in which they appear in the pattern.
190 The first one that is set is used for the reference. Previously only the
191 first one was inspected. This change makes PCRE more compatible with Perl.
192
193 16. Unicode character properties were updated from Unicode 6.3.0.
194
195 17. The compile-time code for auto-possessification has been refactored, based
196 on a patch by Zoltan Herczeg. It now happens after instead of during
197 compilation. The code is cleaner, and more cases are handled. The option
198 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
199 options in pcretest are provided to set it. It can also be set by
200 (*NO_AUTO_POSSESS) at the start of a pattern.
201
202 18. The character VT has been added to the default ("C" locale) set of
203 characters that match \s and are generally treated as white space,
204 following this same change in Perl 5.18. There is now no difference between
205 "Perl space" and "POSIX space". Whether VT is treated as white space in
206 other locales depends on the locale.
207
208 19. The code for checking named groups as conditions, either for being set or
209 for being recursed, has been refactored (this is related to 14 and 15
210 above). Processing unduplicated named groups should now be as fast at
211 numerical groups, and processing duplicated groups should be faster than
212 before.
213
214 20. Two patches to the CMake build system, by Alexander Barkov:
215
216 (1) Replace the "source" command by "." in CMakeLists.txt because
217 "source" is a bash-ism.
218
219 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
220 without these the CMake build does not work on Solaris.
221
222 21. Perl has changed its handling of \8 and \9. If there is no previously
223 encountered capturing group of those numbers, they are treated as the
224 literal characters 8 and 9 instead of a binary zero followed by the
225 literals. PCRE now does the same.
226
227 22. Following Perl, added \o{} to specify codepoints in octal, making it
228 possible to specify values greater than 0777 and also making them
229 unambiguous.
230
231 23. Perl now gives an error for missing closing braces after \x{... instead of
232 treating the string as literal. PCRE now does the same.
233
234 24. RunTest used to grumble if an inappropriate test was selected explicitly,
235 but just skip it when running all tests. This make it awkward to run ranges
236 of tests when one of them was inappropriate. Now it just skips any
237 inappropriate tests, as it always did when running all tests.
238
239 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
240 character types such as \d or \w, too many callouts were inserted, and the
241 data that they returned was rubbish.
242
243 26. In UCP mode, \s was not matching two of the characters that Perl matches,
244 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
245 were matched by \h. The code has now been refactored so that the lists of
246 the horizontal and vertical whitespace characters used for \h and \v (which
247 are defined only in one place) are now also used for \s.
248
249 27. Add JIT support for the 64 bit TileGX architecture.
250 Patch by Jiong Wang (Tilera Corporation).
251
252 28. Possessive quantifiers for classes (both explicit and automatically
253 generated) now use special opcodes instead of wrapping in ONCE brackets.
254
255 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
256 (because it's meaningless), this was not happening when PCRE_CASELESS was
257 set. Not wrong, but inefficient.
258
259 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
260 properties for \w, \d, etc) is present in a test regex. Otherwise if the
261 test contains no characters greater than 255, Perl doesn't realise it
262 should be using Unicode semantics.
263
264 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
265 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
266 does in Unicode mode.
267
268 32. Added the "forbid" facility to pcretest so that putting tests into the
269 wrong test files can sometimes be quickly detected.
270
271 33. There is now a limit (default 250) on the depth of nesting of parentheses.
272 This limit is imposed to control the amount of system stack used at compile
273 time. It can be changed at build time by --with-parens-nest-limit=xxx or
274 the equivalent in CMake.
275
276 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
277 errors. Perl warns for these when in warning mode, but PCRE has no facility
278 for giving warnings.
279
280 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
281 However, this was not working for (?!) because it is optimized to (*FAIL),
282 for which PCRE does not allow quantifiers. The optimization is now disabled
283 when a quantifier follows (?!). I can't see any use for this, but it makes
284 things uniform.
285
286 36. Perl no longer allows group names to start with digits, so I have made this
287 change also in PCRE. It simplifies the code a bit.
288
289 37. In extended mode, Perl ignores spaces before a + that indicates a
290 possessive quantifier. PCRE allowed a space before the quantifier, but not
291 before the possessive +. It now does.
292
293 38. The use of \K (reset reported match start) within a repeated possessive
294 group such as (a\Kb)*+ was not working.
295
296 40. Document that the same character tables must be used at compile time and
297 run time, and that the facility to pass tables to pcre_exec() and
298 pcre_dfa_exec() is for use only with saved/restored patterns.
299
300 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
301 features for Builds with MSVC:
302
303 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
304 builds). The .rc files can be used to set FileDescription and many other
305 attributes.
306
307 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
308 This allows higher-level build scripts which want .pdb files to avoid
309 hard-coding the exact files needed."
310
311 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
312 mean "start of word" and "end of word", respectively, as a transition aid.
313
314 43. A minimizing repeat of a class containing codepoints greater than 255 in
315 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
316 compiled to use the heap for recursion.
317
318 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
319 is configured.
320
321
322 Version 8.33 28-May-2013
323 ------------------------
324
325 1. Added 'U' to some constants that are compared to unsigned integers, to
326 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
327 variables that are added to signed variables, to ensure the result is
328 signed and can be negated.
329
330 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
331 CMake config files.
332
333 3. Revise the creation of config.h.generic so that all boolean macros are
334 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
335 overriding via -D on the command line possible.
336
337 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
338 to unsigned int is reported to make a quite noticeable speed difference in
339 a specific Windows environment. Testing on Linux did also appear to show
340 some benefit (and it is clearly not harmful). Also fixed the definition of
341 Xop which should be unsigned.
342
343 5. Related to (4), changing the definition of the intermediate variable cc
344 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
345 improvements.
346
347 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
348 unnecessary spaces.
349
350 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
351 and later.
352
353 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
354
355 9. Optimizing fast_forward_start_bits in JIT.
356
357 10. Adding support for callouts in JIT, and fixing some issues revealed
358 during this work. Namely:
359
360 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
361
362 (b) Minimum length was not checked before the matching is started.
363
364 11. The value of capture_last that is passed to callouts was incorrect in some
365 cases when there was a capture on one path that was subsequently abandoned
366 after a backtrack. Also, the capture_last value is now reset after a
367 recursion, since all captures are also reset in this case.
368
369 12. The interpreter no longer returns the "too many substrings" error in the
370 case when an overflowing capture is in a branch that is subsequently
371 abandoned after a backtrack.
372
373 13. In the pathological case when an offset vector of size 2 is used, pcretest
374 now prints out the matched string after a yield of 0 or 1.
375
376 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
377 Only supported by the JIT compiler at the moment.
378
379 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
380
381 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
382 offset of the starting point of the matching process, provided the offsets
383 vector is large enough.
384
385 17. The \A escape now records a lookbehind value of 1, though its execution
386 does not actually inspect the previous character. This is to ensure that,
387 in partial multi-segment matching, at least one character from the old
388 segment is retained when a new segment is processed. Otherwise, if there
389 are no lookbehinds in the pattern, \A might match incorrectly at the start
390 of a new segment.
391
392 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
393
394 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
395 gives some modest performance improvement in 8-bit mode.
396
397 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
398 be expressed in certain programming languages using Universal Character
399 Names.
400
401 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
402 which points out that "non characters" are not "characters that may not
403 appear in Unicode strings" but rather "characters that are reserved for
404 internal use and have only local meaning".
405
406 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
407 there was a conditional group that depended on an assertion, if the
408 assertion was false, the callout that immediately followed the alternation
409 in the condition was skipped when pcre_exec() was used for matching.
410
411 23. Allow an explicit callout to be inserted before an assertion that is the
412 condition for a conditional group, for compatibility with automatic
413 callouts, which always insert a callout at this point.
414
415 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
416 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
417
418 25. (*PRUNE) is now supported by the JIT compiler.
419
420 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
421
422 27. Fix the case where there are two or more SKIPs with arguments that may be
423 ignored.
424
425 28. (*SKIP) is now supported by the JIT compiler.
426
427 29. (*THEN) is now supported by the JIT compiler.
428
429 30. Update RunTest with additional test selector options.
430
431 31. The way PCRE handles backtracking verbs has been changed in two ways.
432
433 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
434 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
435 backtracking. In some cases this makes it more Perl-compatible, but Perl's
436 rather obscure rules do not always do the same thing.
437
438 (2) Previously, backtracking verbs were confined within assertions. This is
439 no longer the case for positive assertions, except for (*ACCEPT). Again,
440 this sometimes improves Perl compatibility, and sometimes does not.
441
442 32. A number of tests that were in test 2 because Perl did things differently
443 have been moved to test 1, because either Perl or PCRE has changed, and
444 these tests are now compatible.
445
446 32. Backtracking control verbs are now handled in the same way in JIT and
447 interpreter.
448
449 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
450 contained a forward subroutine reference caused a compile error.
451
452 34. Auto-detect and optimize limited repetitions in JIT.
453
454 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
455 blocking (*UTF) etc.
456
457 36. In the interpreter, maximizing pattern repetitions for characters and
458 character types now use tail recursion, which reduces stack usage.
459
460 37. The value of the max lookbehind was not correctly preserved if a compiled
461 and saved regex was reloaded on a host of different endianness.
462
463 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
464 of the compiled pattern block, expand the flags field from 16 to 32 bits
465 because it was almost full.
466
467 39. Try madvise first before posix_madvise.
468
469 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
470 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
471 original change locked out all matches of empty strings. This has been
472 changed so that one match of an empty string per line is recognized.
473 Subsequent searches on the same line (for colouring or for --only-matching,
474 for example) do not recognize empty strings.
475
476 41. Applied a user patch to fix a number of spelling mistakes in comments.
477
478 42. Data lines longer than 65536 caused pcretest to crash.
479
480 43. Clarified the data type for length and startoffset arguments for pcre_exec
481 and pcre_dfa_exec in the function-specific man pages, where they were
482 explicitly stated to be in bytes, never having been updated. I also added
483 some clarification to the pcreapi man page.
484
485 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
486 a segmentation fault.
487
488
489 Version 8.32 30-November-2012
490 -----------------------------
491
492 1. Improved JIT compiler optimizations for first character search and single
493 character iterators.
494
495 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
496 Patch by Daniel Richard G.
497
498 3. Single character iterator optimizations in the JIT compiler.
499
500 4. Improved JIT compiler optimizations for character ranges.
501
502 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
503 Reported by Giuseppe D'Angelo.
504
505 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
506 of a line, was being set incorrectly in cases where .* appeared inside
507 atomic brackets at the start of a pattern, or where there was a subsequent
508 *PRUNE or *SKIP.
509
510 7. Improved instruction cache flush for POWER/PowerPC.
511 Patch by Daniel Richard G.
512
513 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
514 grep:
515
516 (a) There is now no limit to the number of patterns to be matched.
517
518 (b) An error is given if a pattern is too long.
519
520 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
521 are now supported.
522
523 (d) --exclude-from and --include-from (multiple use) have been added.
524
525 (e) Exclusions and inclusions now apply to all files and directories, not
526 just to those obtained from scanning a directory recursively.
527
528 (f) Multiple uses of -f and --file-list are now supported.
529
530 (g) In a Windows environment, the default for -d has been changed from
531 "read" (the GNU grep default) to "skip", because otherwise the presence
532 of a directory in the file list provokes an error.
533
534 (h) The documentation has been revised and clarified in places.
535
536 9. Improve the matching speed of capturing brackets.
537
538 10. Changed the meaning of \X so that it now matches a Unicode extended
539 grapheme cluster.
540
541 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
542 out POSIX threads when JIT support is configured.
543
544 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
545
546 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
547 structure could go wrong in environments where size_t is not the same size
548 as int.
549
550 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
551 set.
552
553 15. The EBCDIC support had decayed; later updates to the code had included
554 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
555 general tidy up of EBCDIC-related issues, and the documentation was also
556 not quite right. There is now a test that can be run on ASCII systems to
557 check some of the EBCDIC-related things (but is it not a full test).
558
559 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
560 in a small tidy to the code.
561
562 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
563
564 18. If the --only-matching (-o) option in pcregrep is specified multiple
565 times, each one causes appropriate output. For example, -o1 -o2 outputs the
566 substrings matched by the 1st and 2nd capturing parentheses. A separating
567 string can be specified by --om-separator (default empty).
568
569 19. Improving the first n character searches.
570
571 20. Turn case lists for horizontal and vertical white space into macros so that
572 they are defined only once.
573
574 21. This set of changes together give more compatible Unicode case-folding
575 behaviour for characters that have more than one other case when UCP
576 support is available.
577
578 (a) The Unicode property table now has offsets into a new table of sets of
579 three or more characters that are case-equivalent. The MultiStage2.py
580 script that generates these tables (the pcre_ucd.c file) now scans
581 CaseFolding.txt instead of UnicodeData.txt for character case
582 information.
583
584 (b) The code for adding characters or ranges of characters to a character
585 class has been abstracted into a generalized function that also handles
586 case-independence. In UTF-mode with UCP support, this uses the new data
587 to handle characters with more than one other case.
588
589 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
590 whose other case is greater than 256 are now correctly matched
591 caselessly. Previously, the high codepoint matched the low one, but not
592 vice versa.
593
594 (d) The processing of \h, \H, \v, and \ in character classes now makes use
595 of the new class addition function, using character lists defined as
596 macros alongside the case definitions of 20 above.
597
598 (e) Caseless back references now work with characters that have more than
599 one other case.
600
601 (f) General caseless matching of characters with more than one other case
602 is supported.
603
604 22. Unicode character properties were updated from Unicode 6.2.0
605
606 23. Improved CMake support under Windows. Patch by Daniel Richard G.
607
608 24. Add support for 32-bit character strings, and UTF-32
609
610 25. Major JIT compiler update (code refactoring and bugfixing).
611 Experimental Sparc 32 support is added.
612
613 26. Applied a modified version of Daniel Richard G's patch to create
614 pcre.h.generic and config.h.generic by "make" instead of in the
615 PrepareRelease script.
616
617 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
618 pcre_compile.c when checking for a zero character.
619
620 28. Introducing a native interface for JIT. Through this interface, the compiled
621 machine code can be directly executed. The purpose of this interface is to
622 provide fast pattern matching, so several sanity checks are not performed.
623 However, feature tests are still performed. The new interface provides
624 1.4x speedup compared to the old one.
625
626 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
627 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
628 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
629
630 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
631 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
632 the "old" RFC 2279). Instead, it ended up passing a negative length to
633 pcre_exec().
634
635 31. Add support for GCC's visibility feature to hide internal functions.
636
637 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
638 "unknown -C option" after outputting 0 or 1.
639
640 33. There is now support for generating a code coverage report for the test
641 suite in environments where gcc is the compiler and lcov is installed. This
642 is mainly for the benefit of the developers.
643
644 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
645 unaddressable using valgrind annotations, allowing valgrind to detect
646 invalid memory accesses. This is mainly for the benefit of the developers.
647
648 25. (*UTF) can now be used to start a pattern in any of the three libraries.
649
650 26. Give configure error if --enable-cpp but no C++ compiler found.
651
652
653 Version 8.31 06-July-2012
654 -------------------------
655
656 1. Fixing a wrong JIT test case and some compiler warnings.
657
658 2. Removed a bashism from the RunTest script.
659
660 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
661 to unsigned type, result still unsigned" that was given by an MS compiler
662 on encountering the code "-sizeof(xxx)".
663
664 4. Partial matching support is added to the JIT compiler.
665
666 5. Fixed several bugs concerned with partial matching of items that consist
667 of more than one character:
668
669 (a) /^(..)\1/ did not partially match "aba" because checking references was
670 done on an "all or nothing" basis. This also applied to repeated
671 references.
672
673 (b) \R did not give a hard partial match if \r was found at the end of the
674 subject.
675
676 (c) \X did not give a hard partial match after matching one or more
677 characters at the end of the subject.
678
679 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
680 a partial match for the string "\r".
681
682 (e) When newline was set to CRLF, the metacharacter "." did not recognize
683 a partial match for a CR character at the end of the subject string.
684
685 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
686 running pcretest, the text "(JIT)" added to the output whenever JIT is
687 actually used to run the match.
688
689 7. Individual JIT compile options can be set in pcretest by following -s+[+]
690 or /S+[+] with a digit between 1 and 7.
691
692 8. OP_NOT now supports any UTF character not just single-byte ones.
693
694 9. (*MARK) control verb is now supported by the JIT compiler.
695
696 10. The command "./RunTest list" lists the available tests without actually
697 running any of them. (Because I keep forgetting what they all are.)
698
699 11. Add PCRE_INFO_MAXLOOKBEHIND.
700
701 12. Applied a (slightly modified) user-supplied patch that improves performance
702 when the heap is used for recursion (compiled with --disable-stack-for-
703 recursion). Instead of malloc and free for each heap frame each time a
704 logical recursion happens, frames are retained on a chain and re-used where
705 possible. This sometimes gives as much as 30% improvement.
706
707 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
708 call.
709
710 14. As documented, (*COMMIT) is now confined to within a positive assertion.
711
712 15. It is now possible to link pcretest with libedit as an alternative to
713 libreadline.
714
715 16. (*COMMIT) control verb is now supported by the JIT compiler.
716
717 17. The Unicode data tables have been updated to Unicode 6.1.0.
718
719 18. Added --file-list option to pcregrep.
720
721 19. Added binary file support to pcregrep, including the -a, --binary-files,
722 -I, and --text options.
723
724 20. The madvise function is renamed for posix_madvise for QNX compatibility
725 reasons. Fixed by Giuseppe D'Angelo.
726
727 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
728
729 22. Changed the option for creating symbolic links for 16-bit man pages from
730 -s to -sf so that re-installing does not cause issues.
731
732 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
733
734 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
735 matches in certain environments (the workspace was not being correctly
736 retained). Also added to pcre_dfa_exec() a simple plausibility check on
737 some of the workspace data at the beginning of a restart.
738
739 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
740 was not doing so when it should - probably a typo introduced by SVN 528
741 (change 8.10/14).
742
743 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
744 \w+ when the character tables indicated that \x{c4} was a word character.
745 There were several related cases, all because the tests for doing a table
746 lookup were testing for characters less than 127 instead of 255.
747
748 27. If a pattern contains capturing parentheses that are not used in a match,
749 their slots in the ovector are set to -1. For those that are higher than
750 any matched groups, this happens at the end of processing. In the case when
751 there were back references that the ovector was too small to contain
752 (causing temporary malloc'd memory to be used during matching), and the
753 highest capturing number was not used, memory off the end of the ovector
754 was incorrectly being set to -1. (It was using the size of the temporary
755 memory instead of the true size.)
756
757 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
758 ovector size, it uses memory at the end of the block that it has got.
759
760 29. Check for an overlong MARK name and give an error at compile time. The
761 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
762
763 30. JIT compiler update.
764
765 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
766 Rill for the patch.
767
768 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
769
770 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
771
772 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
773 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
774 was enabled.
775
776 35. Improve JIT code generation for greedy plus quantifier.
777
778 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
779 1 to "aa" instead of to an empty string. The bug affected repeated groups
780 that could potentially match an empty string.
781
782 37. Optimizing single character iterators in JIT.
783
784 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
785 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
786 codepoints that are too big for the mode are faulted, and in a UTF mode,
787 disallowed codepoints are also faulted.
788
789 39. If PCRE was compiled with UTF support, in three places in the DFA
790 matcher there was code that should only have been obeyed in UTF mode, but
791 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
792 processing when bytes with values greater than 127 were present. In 16-bit
793 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
794 both cases the values are those that cannot be the first data item in a UTF
795 character. The three items that might have provoked this were recursions,
796 possessively repeated groups, and atomic groups.
797
798 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
799 and pcregrep, because some OS require shared objects to be explicitly
800 passed to ld, causing the link step to fail if they are not.
801
802 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
803 mode, patterns that started with \h* or \R* might be incorrectly matched.
804
805
806 Version 8.30 04-February-2012
807 -----------------------------
808
809 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
810 name is defined in ctype.h.
811
812 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
813 only in quite long subpatterns.
814
815 3. Removed the function pcre_info(), which has been obsolete and deprecated
816 since it was replaced by pcre_fullinfo() in February 2000.
817
818 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
819 match a (*MARK), and the match failed at the start of the subject, a
820 reference to memory before the start of the subject could occur. This bug
821 was introduced by fix 17 of release 8.21.
822
823 5. A reference to an unset group with zero minimum repetition was giving
824 totally wrong answers (in non-JavaScript-compatibility mode). For example,
825 /(another)?(\1?)test/ matched against "hello world test". This bug was
826 introduced in release 8.13.
827
828 6. Add support for 16-bit character strings (a large amount of work involving
829 many changes and refactorings).
830
831 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
832 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
833 from a file.
834
835 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
836 rounding is not applied in this particular case).
837
838 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
839 if they appear, or are escaped, in patterns.
840
841 10. Get rid of a number of -Wunused-but-set-variable warnings.
842
843 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
844 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
845 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
846 also returns the mark "x". This bug applied to capturing parentheses,
847 non-capturing parentheses, and atomic parentheses. It also applied to some
848 assertions.
849
850 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
851 information out of configure.ac instead of relying on pcre.h.generic, which
852 is not stored in the repository.
853
854 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
855 -lreadline.
856
857 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
858
859 15. Applied Graycode's patch to put the top-level frame on the stack rather
860 than the heap when not using the stack for recursion. This gives a
861 performance improvement in many cases when recursion is not deep.
862
863 16. Experimental code added to "pcretest -C" to output the stack frame size.
864
865
866 Version 8.21 12-Dec-2011
867 ------------------------
868
869 1. Updating the JIT compiler.
870
871 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
872 are added as well.
873
874 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
875 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
876 calling _pcre_jit_exec. Some extra comments are added.
877
878 4. (*MARK) settings inside atomic groups that do not contain any capturing
879 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
880 was introduced by change 18 for 8.20.
881
882 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
883 ECMA-262 standard.
884
885 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
886 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
887 This bug was probably introduced by change 9 of 8.13.
888
889 7. While fixing 6 above, I noticed that a number of other items were being
890 incorrectly rejected as "not fixed length". This arose partly because newer
891 opcodes had not been added to the fixed-length checking code. I have (a)
892 corrected the bug and added tests for these items, and (b) arranged for an
893 error to occur if an unknown opcode is encountered while checking for fixed
894 length instead of just assuming "not fixed length". The items that were
895 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
896 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
897 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
898
899 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
900 being incorrectly compiled and would have given unpredicatble results.
901
902 9. A possessively repeated subpattern with minimum repeat count greater than
903 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
904 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
905 the first (A) could occur when it should not.
906
907 10. Add a cast and remove a redundant test from the code.
908
909 11. JIT should use pcre_malloc/pcre_free for allocation.
910
911 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
912 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
913 is anything other than /usr, -L is still shown).
914
915 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
916
917 14. Perl does not support \N without a following name in a [] class; PCRE now
918 also gives an error.
919
920 15. If a forward reference was repeated with an upper limit of around 2000,
921 it caused the error "internal error: overran compiling workspace". The
922 maximum number of forward references (including repeats) was limited by the
923 internal workspace, and dependent on the LINK_SIZE. The code has been
924 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
925 the default depends on LINK_SIZE. There is a new upper limit (for safety)
926 of around 200,000 forward references. While doing this, I also speeded up
927 the filling in of repeated forward references.
928
929 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
930 incorrectly expecting the subject to contain another "a" after the start.
931
932 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
933 in the match, the SKIP should be ignored. This was not happening; instead
934 the SKIP was being treated as NOMATCH. For patterns such as
935 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
936 tested.
937
938 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
939 now much more compatible with Perl, in particular in cases where the result
940 is a non-match for a non-anchored pattern. For example, if
941 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
942 "m", where previously it did not return a name. A side effect of this
943 change is that for partial matches, the last encountered mark name is
944 returned, as for non matches. A number of tests that were previously not
945 Perl-compatible have been moved into the Perl-compatible test files. The
946 refactoring has had the pleasing side effect of removing one argument from
947 the match() function, thus reducing its stack requirements.
948
949 19. If the /S+ option was used in pcretest to study a pattern using JIT,
950 subsequent uses of /S (without +) incorrectly behaved like /S+.
951
952 21. Retrieve executable code size support for the JIT compiler and fixing
953 some warnings.
954
955 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
956 not work when the shorter character appeared right at the end of the
957 subject string.
958
959 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
960 systems.
961
962 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
963 output it when the /M option is used in pcretest.
964
965 25. The CheckMan script was not being included in the distribution. Also, added
966 an explicit "perl" to run Perl scripts from the PrepareRelease script
967 because this is reportedly needed in Windows.
968
969 26. If study data was being save in a file and studying had not found a set of
970 "starts with" bytes for the pattern, the data written to the file (though
971 never used) was taken from uninitialized memory and so caused valgrind to
972 complain.
973
974 27. Updated RunTest.bat as provided by Sheri Pierce.
975
976 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
977
978 29. Computation of memory usage for the table of capturing group names was
979 giving an unnecessarily large value.
980
981
982 Version 8.20 21-Oct-2011
983 ------------------------
984
985 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
986 a POSIX class. After further experiments with Perl, which convinced me that
987 Perl has bugs and confusions, a closing square bracket is no longer allowed
988 in a POSIX name. This bug also affected patterns with classes that started
989 with full stops.
990
991 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
992 captured substring, but while checking the failing first alternative,
993 substring 1 is temporarily captured. If the output vector supplied to
994 pcre_exec() was not big enough for this capture, the yield of the function
995 was still zero ("insufficient space for captured substrings"). This cannot
996 be totally fixed without adding another stack variable, which seems a lot
997 of expense for a edge case. However, I have improved the situation in cases
998 such as /(a)(b)x|abc/ matched against "abc", where the return code
999 indicates that fewer than the maximum number of slots in the ovector have
1000 been set.
1001
1002 3. Related to (2) above: when there are more back references in a pattern than
1003 slots in the output vector, pcre_exec() uses temporary memory during
1004 matching, and copies in the captures as far as possible afterwards. It was
1005 using the entire output vector, but this conflicts with the specification
1006 that only 2/3 is used for passing back captured substrings. Now it uses
1007 only the first 2/3, for compatibility. This is, of course, another edge
1008 case.
1009
1010 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1011 main code base, and can be used by building with --enable-jit. When this is
1012 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1013 runtime --no-jit option is given.
1014
1015 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1016 ovector, the return from the function was zero, implying that there were
1017 other matches that did not fit. The correct "exactly full" value is now
1018 returned.
1019
1020 6. If a subpattern that was called recursively or as a subroutine contained
1021 (*PRUNE) or any other control that caused it to give a non-standard return,
1022 invalid errors such as "Error -26 (nested recursion at the same subject
1023 position)" or even infinite loops could occur.
1024
1025 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1026 computing the minimum length on reaching *ACCEPT, and so ended up with the
1027 wrong value of 1 rather than 0. Further investigation indicates that
1028 computing a minimum subject length in the presence of *ACCEPT is difficult
1029 (think back references, subroutine calls), and so I have changed the code
1030 so that no minimum is registered for a pattern that contains *ACCEPT.
1031
1032 8. If (*THEN) was present in the first (true) branch of a conditional group,
1033 it was not handled as intended. [But see 16 below.]
1034
1035 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1036 Sheri Pierce.
1037
1038 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1039 the first byte in a match must be "a".
1040
1041 11. Change 17 for 8.13 increased the recursion depth for patterns like
1042 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1043 pattern contains any instances of (*THEN). If it does not, the old
1044 optimizations are restored. It would be nice to do this on a per-group
1045 basis, but at the moment that is not feasible.
1046
1047 12. In some environments, the output of pcretest -C is CRLF terminated. This
1048 broke RunTest's code that checks for the link size. A single white space
1049 character after the value is now allowed for.
1050
1051 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1052 For "fr", it uses the Windows-specific input and output files.
1053
1054 14. If (*THEN) appeared in a group that was called recursively or as a
1055 subroutine, it did not work as intended. [But see next item.]
1056
1057 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1058 pattern fragments (but not containing any | characters). If A and B are
1059 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1060 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1061 D. In other words, Perl considers parentheses that do not contain any |
1062 characters to be part of a surrounding alternative, whereas PCRE was
1063 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1064 differently. PCRE now behaves in the same way as Perl, except in the case
1065 of subroutine/recursion calls such as (?1) which have in any case always
1066 been different (but PCRE had them first :-).
1067
1068 16. Related to 15 above: Perl does not treat the | in a conditional group as
1069 creating alternatives. Such a group is treated in the same way as an
1070 ordinary group without any | characters when processing (*THEN). PCRE has
1071 been changed to match Perl's behaviour.
1072
1073 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1074 RunGrepTest script failed.
1075
1076 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1077 inevitable for groups that contain captures, but it can lead to a lot of
1078 stack use in large patterns. The old behaviour has been restored for atomic
1079 groups that do not contain any capturing parentheses.
1080
1081 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1082 suppress the check for a minimum subject length at run time. (If it was
1083 given to pcre_exec() or pcre_dfa_exec() it did work.)
1084
1085 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1086 fail to work when decoding hex characters in data strings in EBCDIC
1087 environments.
1088
1089 21. It appears that in at least one Mac OS environment, the isxdigit() function
1090 is implemented as a macro that evaluates to its argument more than once,
1091 contravening the C 90 Standard (I haven't checked a later standard). There
1092 was an instance in pcretest which caused it to go wrong when processing
1093 \x{...} escapes in subject strings. The has been rewritten to avoid using
1094 things like p++ in the argument of isxdigit().
1095
1096
1097 Version 8.13 16-Aug-2011
1098 ------------------------
1099
1100 1. The Unicode data tables have been updated to Unicode 6.0.0.
1101
1102 2. Two minor typos in pcre_internal.h have been fixed.
1103
1104 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1105 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1106 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1107
1108 4. There were a number of related bugs in the code for matching backrefences
1109 caselessly in UTF-8 mode when codes for the characters concerned were
1110 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1111 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1112 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1113 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1114 2-byte code at the end of the subject (it thought there wasn't enough data
1115 left).
1116
1117 5. Comprehensive information about what went wrong is now returned by
1118 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1119 as the output vector has at least 2 elements. The offset of the start of
1120 the failing character and a reason code are placed in the vector.
1121
1122 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1123 now returned is for the first byte of the failing character, instead of the
1124 last byte inspected. This is an incompatible change, but I hope it is small
1125 enough not to be a problem. It makes the returned offset consistent with
1126 pcre_exec() and pcre_dfa_exec().
1127
1128 7. pcretest now gives a text phrase as well as the error number when
1129 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1130 failure, the offset and reason code are output.
1131
1132 8. When \R was used with a maximizing quantifier it failed to skip backwards
1133 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1134 back over a single character (\n). This seems wrong (because it treated the
1135 two characters as a single entity when going forwards), conflicts with the
1136 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1137 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1138 has been changed.
1139
1140 9. Some internal refactoring has changed the processing so that the handling
1141 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1142 time (the PCRE_DOTALL option was changed this way some time ago: version
1143 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1144 which was always a bit of a fudge. It also means that there is one less
1145 argument for the match() function, which reduces its stack requirements
1146 slightly. This change also fixes an incompatibility with Perl: the pattern
1147 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1148
1149 10. More internal refactoring has drastically reduced the number of recursive
1150 calls to match() for possessively repeated groups such as (abc)++ when
1151 using pcre_exec().
1152
1153 11. While implementing 10, a number of bugs in the handling of groups were
1154 discovered and fixed:
1155
1156 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1157 (a|)*(?1) gave a compile-time internal error.
1158 ((a|)+)+ did not notice that the outer group could match an empty string.
1159 (^a|^)+ was not marked as anchored.
1160 (.*a|.*)+ was not marked as matching at start or after a newline.
1161
1162 12. Yet more internal refactoring has removed another argument from the match()
1163 function. Special calls to this function are now indicated by setting a
1164 value in a variable in the "match data" data block.
1165
1166 13. Be more explicit in pcre_study() instead of relying on "default" for
1167 opcodes that mean there is no starting character; this means that when new
1168 ones are added and accidentally left out of pcre_study(), testing should
1169 pick them up.
1170
1171 14. The -s option of pcretest has been documented for ages as being an old
1172 synonym of -m (show memory usage). I have changed it to mean "force study
1173 for every regex", that is, assume /S for every regex. This is similar to -i
1174 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1175 using it. It makes it easier to run collections of tests with and without
1176 study enabled, and thereby test pcre_study() more easily. All the standard
1177 tests are now run with and without -s (but some patterns can be marked as
1178 "never study" - see 20 below).
1179
1180 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1181 restoration of the capturing data to the outer values was not happening
1182 correctly.
1183
1184 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1185 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1186 pattern had matched an empty string, and so incorrectly returned a no
1187 match.
1188
1189 17. There was optimizing code for the last branch of non-capturing parentheses,
1190 and also for the obeyed branch of a conditional subexpression, which used
1191 tail recursion to cut down on stack usage. Unfortunately, now that there is
1192 the possibility of (*THEN) occurring in these branches, tail recursion is
1193 no longer possible because the return has to be checked for (*THEN). These
1194 two optimizations have therefore been removed. [But see 8.20/11 above.]
1195
1196 18. If a pattern containing \R was studied, it was assumed that \R always
1197 matched two bytes, thus causing the minimum subject length to be
1198 incorrectly computed because \R can also match just one byte.
1199
1200 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1201 was incorrectly computed.
1202
1203 20. If /S is present twice on a test pattern in pcretest input, it now
1204 *disables* studying, thereby overriding the use of -s on the command line
1205 (see 14 above). This is necessary for one or two tests to keep the output
1206 identical in both cases.
1207
1208 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1209 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1210
1211 22. When an atomic group that contained a capturing parenthesis was
1212 successfully matched, but the branch in which it appeared failed, the
1213 capturing was not being forgotten if a higher numbered group was later
1214 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1215 group 1 to "a", when in fact it should be unset. This applied to multi-
1216 branched capturing and non-capturing groups, repeated or not, and also to
1217 positive assertions (capturing in negative assertions does not happen
1218 in PCRE) and also to nested atomic groups.
1219
1220 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1221 subject after a captured substring, to make it easier to tell which of a
1222 number of identical substrings has been captured.
1223
1224 24. The way atomic groups are processed by pcre_exec() has been changed so that
1225 if they are repeated, backtracking one repetition now resets captured
1226 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1227 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1228 "aaa". Previously, it would have been "a". As part of this code
1229 refactoring, the way recursive calls are handled has also been changed.
1230
1231 25. If an assertion condition captured any substrings, they were not passed
1232 back unless some other capturing happened later. For example, if
1233 (?(?=(a))a) was matched against "a", no capturing was returned.
1234
1235 26. When studying a pattern that contained subroutine calls or assertions,
1236 the code for finding the minimum length of a possible match was handling
1237 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1238 group 1 called group 2 while simultaneously a separate group 2 called group
1239 1). A stack overflow occurred in this case. I have fixed this by limiting
1240 the recursion depth to 10.
1241
1242 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1243 Fortmann. This supports explicit test numbers on the command line, and has
1244 argument validation and error reporting.
1245
1246 28. An instance of \X with an unlimited repeat could fail if at any point the
1247 first character it looked at was a mark character.
1248
1249 29. Some minor code refactoring concerning Unicode properties and scripts
1250 should reduce the stack requirement of match() slightly.
1251
1252 30. Added the '=' option to pcretest to check the setting of unused capturing
1253 slots at the end of the pattern, which are documented as being -1, but are
1254 not included in the return count.
1255
1256 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1257 compiled something random. Now it gives a compile-time error (as does
1258 Perl).
1259
1260 32. A *MARK encountered during the processing of a positive assertion is now
1261 recorded and passed back (compatible with Perl).
1262
1263 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1264 had alternative anchored branches, the search for a second match in a line
1265 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1266 matched the line "0102" twice. The same bug affected patterns that started
1267 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1268 twice.
1269
1270 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1271 does, and because of capturing effects, quantifying parenthesized
1272 assertions may at times be useful. Quantifiers are now allowed for
1273 parenthesized assertions.
1274
1275 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1276
1277 36. \g was being checked for fancy things in a character class, when it should
1278 just be a literal "g".
1279
1280 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1281 appearance of a nested POSIX class supersedes an apparent external class.
1282 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1283 unescaped square brackets may also appear as part of class names. For
1284 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1285 more like Perl. (But see 8.20/1 above.)
1286
1287 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1288 was because it thought it was \N{name}, which is not supported).
1289
1290 39. Add minix to OS list not supporting the -S option in pcretest.
1291
1292 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1293 cannot analyze patterns in sufficient detail to catch mutual recursions
1294 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1295 subgroup is called recursively as a subpattern for a second time at the
1296 same position in the subject string. In previous releases this might have
1297 been caught by the recursion limit, or it might have run out of stack.
1298
1299 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1300 happen only once. PCRE was, however incorrectly giving a compile time error
1301 "recursive call could loop indefinitely" because it cannot analyze the
1302 pattern in sufficient detail. The compile time test no longer happens when
1303 PCRE is compiling a conditional subpattern, but actual runaway loops are
1304 now caught at runtime (see 40 above).
1305
1306 42. It seems that Perl allows any characters other than a closing parenthesis
1307 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1308 has been changed to be the same.
1309
1310 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1311 as not to get warnings when autogen.sh is called. Also changed
1312 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1313
1314 44. To help people who use pcregrep to scan files containing exceedingly long
1315 lines, the following changes have been made:
1316
1317 (a) The default value of the buffer size parameter has been increased from
1318 8K to 20K. (The actual buffer used is three times this size.)
1319
1320 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1321 PCRE is built.
1322
1323 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1324 to be set at run time.
1325
1326 (d) Numerical values in pcregrep options can be followed by K or M, for
1327 example --buffer-size=50K.
1328
1329 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1330 given and the return code is set to 2.
1331
1332 45. Add a pointer to the latest mark to the callout data block.
1333
1334 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1335 partial match of an empty string instead of no match. This was specific to
1336 the use of ".".
1337
1338 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1339 complete match instead of a partial match. This bug was dependent on both
1340 the PCRE_UTF8 and PCRE_DOTALL options being set.
1341
1342 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1343 starting byte set, because \b was not being ignored.
1344
1345
1346 Version 8.12 15-Jan-2011
1347 ------------------------
1348
1349 1. Fixed some typos in the markup of the man pages, and wrote a script that
1350 checks for such things as part of the documentation building process.
1351
1352 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1353 --match-limit and --recursion-limit options (added for 8.11). In
1354 particular, this made one of the standard tests fail. (The integer value
1355 went into the wrong half of a long int.)
1356
1357 3. If the --colour option was given to pcregrep with -v (invert match), it
1358 did strange things, either producing crazy output, or crashing. It should,
1359 of course, ignore a request for colour when reporting lines that do not
1360 match.
1361
1362 4. Another pcregrep bug caused similar problems if --colour was specified with
1363 -M (multiline) and the pattern match finished with a line ending.
1364
1365 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1366 matched in multiline mode, the following line was shown as part of the
1367 match. This seems wrong, so I have changed it.
1368
1369 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1370 the check for further matches in the same line (so they could be coloured)
1371 to overrun the end of the current line. If another match was found, it was
1372 incorrectly shown (and then shown again when found in the next line).
1373
1374 7. If pcregrep was compiled under Windows, there was a reference to the
1375 function pcregrep_exit() before it was defined. I am assuming this was
1376 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1377 reported by a user. I've moved the definition above the reference.
1378
1379
1380 Version 8.11 10-Dec-2010
1381 ------------------------
1382
1383 1. (*THEN) was not working properly if there were untried alternatives prior
1384 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1385 backtracked to try for "b" instead of moving to the next alternative branch
1386 at the same level (in this case, to look for "c"). The Perl documentation
1387 is clear that when (*THEN) is backtracked onto, it goes to the "next
1388 alternative in the innermost enclosing group".
1389
1390 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1391 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1392 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1393 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1394 (*THEN).
1395
1396 3. If \s appeared in a character class, it removed the VT character from
1397 the class, even if it had been included by some previous item, for example
1398 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1399 of \s, but is part of the POSIX "space" class.)
1400
1401 4. A partial match never returns an empty string (because you can always
1402 match an empty string at the end of the subject); however the checking for
1403 an empty string was starting at the "start of match" point. This has been
1404 changed to the "earliest inspected character" point, because the returned
1405 data for a partial match starts at this character. This means that, for
1406 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1407 (previously it gave "no match").
1408
1409 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1410 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1411 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1412 has an implication that the given string is incomplete (because a partial
1413 match is preferred over a full match). For this reason, these items now
1414 give a partial match in this situation. [Aside: previously, the one case
1415 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1416 match rather than a full match, which was wrong by the old rules, but is
1417 now correct.]
1418
1419 6. There was a bug in the handling of #-introduced comments, recognized when
1420 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1421 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1422 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1423 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1424 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1425 places in pcre_compile().
1426
1427 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1428 comments when looking ahead for named forward references to subpatterns,
1429 the only newline sequence it recognized was NL. It now handles newlines
1430 according to the set newline convention.
1431
1432 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1433 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1434 cater for a lack of strerror(). These oversights have been fixed.
1435
1436 9. Added --match-limit and --recursion-limit to pcregrep.
1437
1438 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1439
1440 11. When the -o option was used, pcregrep was setting a return code of 1, even
1441 when matches were found, and --line-buffered was not being honoured.
1442
1443 12. Added an optional parentheses number to the -o and --only-matching options
1444 of pcregrep.
1445
1446 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1447 can match an empty string. The code to do it in pcretest and pcredemo
1448 needed fixing:
1449
1450 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1451 only one byte after an empty string match just before CRLF (this case
1452 just got forgotten; "any" and "anycrlf" were OK).
1453
1454 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1455 mode when an empty string match preceded an ASCII character followed by
1456 a non-ASCII character. (The code for advancing by one character rather
1457 than one byte was nonsense.)
1458
1459 (c) The pcredemo.c sample program did not have any code at all to handle
1460 the cases when CRLF is a valid newline sequence.
1461
1462 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1463 as a starting offset was within the subject string. There is now a new
1464 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1465 negative or greater than the length of the string. In order to test this,
1466 pcretest is extended to allow the setting of negative starting offsets.
1467
1468 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1469 starting offset points to the beginning of a UTF-8 character was
1470 unnecessarily clumsy. I tidied it up.
1471
1472 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1473 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1474
1475 17. Nobody had reported that the --include_dir option, which was added in
1476 release 7.7 should have been called --include-dir (hyphen, not underscore)
1477 for compatibility with GNU grep. I have changed it to --include-dir, but
1478 left --include_dir as an undocumented synonym, and the same for
1479 --exclude-dir, though that is not available in GNU grep, at least as of
1480 release 2.5.4.
1481
1482 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1483 characters from a string of bytes) have been redefined so as not to use
1484 loops, in order to improve performance in some environments. At the same
1485 time, I abstracted some of the common code into auxiliary macros to save
1486 repetition (this should not affect the compiled code).
1487
1488 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1489 compile-time error is now given if \c is not followed by an ASCII
1490 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1491 different, and any byte value is allowed.)
1492
1493 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1494 START_OPTIMIZE option, which is now allowed at compile time - but just
1495 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1496 to pcregrep and other applications that have no direct access to PCRE
1497 options. The new /Y option in pcretest sets this option when calling
1498 pcre_compile().
1499
1500 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1501 back references. Groups containing recursive back references were forced to
1502 be atomic by that change, but in the case of named groups, the amount of
1503 memory required was incorrectly computed, leading to "Failed: internal
1504 error: code overflow". This has been fixed.
1505
1506 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1507 pcretest.c, to avoid build problems in some Borland environments.
1508
1509
1510 Version 8.10 25-Jun-2010
1511 ------------------------
1512
1513 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1514 THEN.
1515
1516 2. (*ACCEPT) was not working when inside an atomic group.
1517
1518 3. Inside a character class, \B is treated as a literal by default, but
1519 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1520 causes the error). The code is unchanged, but I tidied the documentation.
1521
1522 4. Inside a character class, PCRE always treated \R and \X as literals,
1523 whereas Perl faults them if its -w option is set. I have changed PCRE so
1524 that it faults them when PCRE_EXTRA is set.
1525
1526 5. Added support for \N, which always matches any character other than
1527 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1528
1529 6. When compiling pcregrep with newer versions of gcc which may have
1530 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1531 declared with attribute warn_unused_result" were given. Just casting the
1532 result to (void) does not stop the warnings; a more elaborate fudge is
1533 needed. I've used a macro to implement this.
1534
1535 7. Minor change to pcretest.c to avoid a compiler warning.
1536
1537 8. Added four artifical Unicode properties to help with an option to make
1538 \s etc use properties (see next item). The new properties are: Xan
1539 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1540
1541 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1542 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1543 this option. Modified pcretest to add /W to test this facility. Added
1544 REG_UCP to make it available via the POSIX interface.
1545
1546 10. Added --line-buffered to pcregrep.
1547
1548 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1549 studied, and the match started with a letter with a code point greater than
1550 127 whose first byte was different to the first byte of the other case of
1551 the letter, the other case of this starting letter was not recognized
1552 (#976).
1553
1554 12. If a pattern that was studied started with a repeated Unicode property
1555 test, for example, \p{Nd}+, there was the theoretical possibility of
1556 setting up an incorrect bitmap of starting bytes, but fortunately it could
1557 not have actually happened in practice until change 8 above was made (it
1558 added property types that matched character-matching opcodes).
1559
1560 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1561 possible starting bytes for non-anchored patterns.
1562
1563 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1564 \R, and also a number of cases that involve Unicode properties, both
1565 explicit and implicit when PCRE_UCP is set.
1566
1567 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1568 input, it could crash or give wrong results if characters with values
1569 greater than 0xc0 were present in the subject string. (Detail: it assumed
1570 UTF-8 input when processing these items.)
1571
1572 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1573 size_t is 64-bit (#991).
1574
1575 17. Added a check for running out of memory when PCRE is compiled with
1576 --disable-stack-for-recursion (#990).
1577
1578 18. If the last data line in a file for pcretest does not have a newline on
1579 the end, a newline was missing in the output.
1580
1581 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1582 less than 128) in its various bitmaps. However, there is a facility for
1583 generating tables according to the current locale when PCRE is compiled. It
1584 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1585 characters, are recognized by isspace() and therefore were getting set in
1586 these tables, and indeed these tables seem to approximate to ISO 8859. This
1587 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1588 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1589 which of course cannot start UTF-8 characters. I have changed the code so
1590 that only real ASCII characters (less than 128) and the correct starting
1591 bytes for UTF-8 encodings are set for characters greater than 127 when in
1592 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1593 altogether.)
1594
1595 20. Added the /T option to pcretest so as to be able to run tests with non-
1596 standard character tables, thus making it possible to include the tests
1597 used for 19 above in the standard set of tests.
1598
1599 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1600 reference to a subpattern the other side of a comment that contains an
1601 opening parenthesis caused either an internal compiling error, or a
1602 reference to the wrong subpattern.
1603
1604
1605 Version 8.02 19-Mar-2010
1606 ------------------------
1607
1608 1. The Unicode data tables have been updated to Unicode 5.2.0.
1609
1610 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1611 configured.
1612
1613 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1614 original author of that file, following a query about its status.
1615
1616 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1617 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1618
1619 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1620 quantifier applied to a forward-referencing subroutine call, could compile
1621 incorrect code or give the error "internal error: previously-checked
1622 referenced subpattern not found".
1623
1624 6. Both MS Visual Studio and Symbian OS have problems with initializing
1625 variables to point to external functions. For these systems, therefore,
1626 pcre_malloc etc. are now initialized to local functions that call the
1627 relevant global functions.
1628
1629 7. There were two entries missing in the vectors called coptable and poptable
1630 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1631 I've fixed the data, and added a kludgy way of testing at compile time that
1632 the lengths are correct (equal to the number of opcodes).
1633
1634 8. Following on from 7, I added a similar kludge to check the length of the
1635 eint vector in pcreposix.c.
1636
1637 9. Error texts for pcre_compile() are held as one long string to avoid too
1638 much relocation at load time. To find a text, the string is searched,
1639 counting zeros. There was no check for running off the end of the string,
1640 which could happen if a new error number was added without updating the
1641 string.
1642
1643 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1644
1645 11. \K was not working if it appeared in an atomic group or in a group that
1646 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1647 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1648 the assertion is positive, but not if it is negative.
1649
1650 12. Change 11 fortuitously reduced the size of the stack frame used in the
1651 "match()" function of pcre_exec.c by one pointer. Forthcoming
1652 implementation of support for (*MARK) will need an extra pointer on the
1653 stack; I have reserved it now, so that the stack frame size does not
1654 decrease.
1655
1656 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1657 item in branch that calls a recursion is a subroutine call - as in the
1658 second branch in the above example - was incorrectly given the compile-
1659 time error "recursive call could loop indefinitely" because pcre_compile()
1660 was not correctly checking the subroutine for matching a non-empty string.
1661
1662 14. The checks for overrunning compiling workspace could trigger after an
1663 overrun had occurred. This is a "should never occur" error, but it can be
1664 triggered by pathological patterns such as hundreds of nested parentheses.
1665 The checks now trigger 100 bytes before the end of the workspace.
1666
1667 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1668
1669
1670 Version 8.01 19-Jan-2010
1671 ------------------------
1672
1673 1. If a pattern contained a conditional subpattern with only one branch (in
1674 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1675 computed the wrong minimum data length (which is of course zero for such
1676 subpatterns). This could cause incorrect "no match" results.
1677
1678 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1679 the pattern is reset in the first branch, pcre_compile() failed with
1680 "internal error: code overflow at offset...". This happened only when
1681 the reset was to the original external option setting. (An optimization
1682 abstracts leading options settings into an external setting, which was the
1683 cause of this.)
1684
1685 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1686 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1687 assertion pattern did not match (meaning that the assertion was true), it
1688 was incorrectly treated as false if the SKIP had been reached during the
1689 matching. This also applied to assertions used as conditions.
1690
1691 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1692 assertion subpattern, including such a pattern used as a condition,
1693 unpredictable results occurred, instead of the error return
1694 PCRE_ERROR_DFA_UITEM.
1695
1696 5. The C++ GlobalReplace function was not working like Perl for the special
1697 situation when an empty string is matched. It now does the fancy magic
1698 stuff that is necessary.
1699
1700 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1701 removed. (These were left over from very, very early versions of PCRE.)
1702
1703 7. Some cosmetic changes to the code to make life easier when compiling it
1704 as part of something else:
1705
1706 (a) Change DEBUG to PCRE_DEBUG.
1707
1708 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1709 called "current" as "current_branch", to prevent a collision with the
1710 Linux macro when compiled as a kernel module.
1711
1712 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1713 prevent a collision with the Linux macro when compiled as a kernel
1714 module.
1715
1716 8. In pcre_compile() there are some checks for integer overflows that used to
1717 cast potentially large values to (double). This has been changed to that
1718 when building, a check for int64_t is made, and if it is found, it is used
1719 instead, thus avoiding the use of floating point arithmetic. (There is no
1720 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1721 double.
1722
1723 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1724 2005 (difference between two addresses compared to an unsigned value).
1725
1726 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1727 custom one, because of the following reported problem in Windows:
1728
1729 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1730 under Win32.
1731 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1732 therefore missing the function definition.
1733 - The compiler thus generates a "C" signature for the test function.
1734 - The linker fails to find the "C" function.
1735 - PCRE fails to configure if asked to do so against libbz2.
1736
1737 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1738 messages were output:
1739
1740 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1741 rerunning libtoolize, to keep the correct libtool macros in-tree.
1742 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1743
1744 I have done both of these things.
1745
1746 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1747 most of the time, it *can* run out if it is given a pattern that contains a
1748 runaway infinite recursion. I updated the discussion in the pcrestack man
1749 page.
1750
1751 13. Now that we have gone to the x.xx style of version numbers, the minor
1752 version may start with zero. Using 08 or 09 is a bad idea because users
1753 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1754 interpreted as invalid octal numbers. I've updated the previous comment in
1755 configure.ac, and also added a check that gives an error if 08 or 09 are
1756 used.
1757
1758 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1759 causing partial matching to fail when the end of the subject matched \W
1760 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1761
1762 15. There were some discrepancies between the declarations in pcre_internal.h
1763 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1764 their definitions. The declarations used "const uschar *" and the
1765 definitions used USPTR. Even though USPTR is normally defined as "const
1766 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1767 reported that: "This difference in casting confuses some C++ compilers, for
1768 example, SunCC recognizes above declarations as different functions and
1769 generates broken code for hbpcre." I have changed the declarations to use
1770 USPTR.
1771
1772 16. GNU libtool is named differently on some systems. The autogen.sh script now
1773 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1774 (FreeBSD).
1775
1776 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1777 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1778 comment: "Figure out how to create a longlong from a string: strtoll and
1779 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1780 instance, but it only takes 2 args instead of 3!"
1781
1782 18. A subtle bug concerned with back references has been fixed by a change of
1783 specification, with a corresponding code fix. A pattern such as
1784 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1785 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1786 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1787 same bug. Such groups have to be quantified to be useful, or contained
1788 inside another quantified group. (If there's no repetition, the reference
1789 can never match.) The problem arises because, having left the group and
1790 moved on to the rest of the pattern, a later failure that backtracks into
1791 the group uses the captured value from the final iteration of the group
1792 rather than the correct earlier one. I have fixed this in PCRE by forcing
1793 any group that contains a reference to itself to be an atomic group; that
1794 is, there cannot be any backtracking into it once it has completed. This is
1795 similar to recursive and subroutine calls.
1796
1797
1798 Version 8.00 19-Oct-09
1799 ----------------------
1800
1801 1. The table for translating pcre_compile() error codes into POSIX error codes
1802 was out-of-date, and there was no check on the pcre_compile() error code
1803 being within the table. This could lead to an OK return being given in
1804 error.
1805
1806 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1807 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1808 in a Windows environment.
1809
1810 3. The pcregrep --count option prints the count for each file even when it is
1811 zero, as does GNU grep. However, pcregrep was also printing all files when
1812 --files-with-matches was added. Now, when both options are given, it prints
1813 counts only for those files that have at least one match. (GNU grep just
1814 prints the file name in this circumstance, but including the count seems
1815 more useful - otherwise, why use --count?) Also ensured that the
1816 combination -clh just lists non-zero counts, with no names.
1817
1818 4. The long form of the pcregrep -F option was incorrectly implemented as
1819 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1820 but it seems right to fix it, and I didn't think it was worth preserving
1821 the old behaviour.
1822
1823 5. The command line items --regex=pattern and --regexp=pattern were not
1824 recognized by pcregrep, which required --regex pattern or --regexp pattern
1825 (with a space rather than an '='). The man page documented the '=' forms,
1826 which are compatible with GNU grep; these now work.
1827
1828 6. No libpcreposix.pc file was created for pkg-config; there was just
1829 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1830
1831 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1832 when UCP support is not needed, by modifying the Python script that
1833 generates it from Unicode data files. This should not matter if the module
1834 is correctly used as a library, but I received one complaint about 50K of
1835 unwanted data. My guess is that the person linked everything into his
1836 program rather than using a library. Anyway, it does no harm.
1837
1838 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1839 was a minimum greater than 1 for a wide character in a possessive
1840 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1841 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1842 character. Chaos in the form of incorrect output or a compiling loop could
1843 result.
1844
1845 9. The restrictions on what a pattern can contain when partial matching is
1846 requested for pcre_exec() have been removed. All patterns can now be
1847 partially matched by this function. In addition, if there are at least two
1848 slots in the offset vector, the offset of the earliest inspected character
1849 for the match and the offset of the end of the subject are set in them when
1850 PCRE_ERROR_PARTIAL is returned.
1851
1852 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1853 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1854 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1855 and may be more useful for multi-segment matching.
1856
1857 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1858 used to be given if ever the end of the subject was reached; now it is
1859 given only if matching could not proceed because another character was
1860 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1861 string "Z", which now yields "no match" instead of "partial match". In the
1862 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1863 final character ended with (*FAIL).
1864
1865 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1866 if the pattern had a "must contain" character that was already found in the
1867 earlier partial match, unless partial matching was again requested. For
1868 example, with the pattern /dog.(body)?/, the "must contain" character is
1869 "g". If the first part-match was for the string "dog", restarting with
1870 "sbody" failed. This bug has been fixed.
1871
1872 13. The string returned by pcre_dfa_exec() after a partial match has been
1873 changed so that it starts at the first inspected character rather than the
1874 first character of the match. This makes a difference only if the pattern
1875 starts with a lookbehind assertion or \b or \B (\K is not supported by
1876 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1877 matching functions compatible, and I think it's the right thing to do.
1878
1879 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1880 so that the demonstration program is easily available in environments where
1881 PCRE has not been installed from source.
1882
1883 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1884 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1885 library.
1886
1887 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1888 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1889 is not the first non-POSIX option to be added. Clearly some people find
1890 these options useful.
1891
1892 17. If a caller to the POSIX matching function regexec() passes a non-zero
1893 value for nmatch with a NULL value for pmatch, the value of
1894 nmatch is forced to zero.
1895
1896 18. RunGrepTest did not have a test for the availability of the -u option of
1897 the diff command, as RunTest does. It now checks in the same way as
1898 RunTest, and also checks for the -b option.
1899
1900 19. If an odd number of negated classes containing just a single character
1901 interposed, within parentheses, between a forward reference to a named
1902 subpattern and the definition of the subpattern, compilation crashed with
1903 an internal error, complaining that it could not find the referenced
1904 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1905 [The bug was that it was starting one character too far in when skipping
1906 over the character class, thus treating the ] as data rather than
1907 terminating the class. This meant it could skip too much.]
1908
1909 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1910 /g option in pcretest when the pattern contains \K, which makes it possible
1911 to have an empty string match not at the start, even when the pattern is
1912 anchored. Updated pcretest and pcredemo to use this option.
1913
1914 21. If the maximum number of capturing subpatterns in a recursion was greater
1915 than the maximum at the outer level, the higher number was returned, but
1916 with unset values at the outer level. The correct (outer level) value is
1917 now given.
1918
1919 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1920 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1921 make it do so. The string so far is captured, making this feature
1922 compatible with Perl.
1923
1924 23. The tests have been re-organized, adding tests 11 and 12, to make it
1925 possible to check the Perl 5.10 features against Perl 5.10.
1926
1927 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1928 pattern matches a fixed length string. PCRE did not allow this; now it
1929 does. Neither allows recursion.
1930
1931 25. I finally figured out how to implement a request to provide the minimum
1932 length of subject string that was needed in order to match a given pattern.
1933 (It was back references and recursion that I had previously got hung up
1934 on.) This code has now been added to pcre_study(); it finds a lower bound
1935 to the length of subject needed. It is not necessarily the greatest lower
1936 bound, but using it to avoid searching strings that are too short does give
1937 some useful speed-ups. The value is available to calling programs via
1938 pcre_fullinfo().
1939
1940 26. While implementing 25, I discovered to my embarrassment that pcretest had
1941 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1942 study optimizations had never been tested with that matching function.
1943 Oops. What is worse, even when it was passed study data, there was a bug in
1944 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1945 were also very few tests of studied patterns with pcre_dfa_exec().
1946
1947 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1948 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1949 on the other side of the coin, they are no longer allowed to have different
1950 names, because these cannot be distinguished in PCRE, and this has caused
1951 confusion. (This is a difference from Perl.)
1952
1953 28. When duplicate subpattern names are present (necessarily with different
1954 numbers, as required by 27 above), and a test is made by name in a
1955 conditional pattern, either for a subpattern having been matched, or for
1956 recursion in such a pattern, all the associated numbered subpatterns are
1957 tested, and the overall condition is true if the condition is true for any
1958 one of them. This is the way Perl works, and is also more like the way
1959 testing by number works.
1960
1961
1962 Version 7.9 11-Apr-09
1963 ---------------------
1964
1965 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1966 (pcretest), all targets were linked against these libraries. This included
1967 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1968 libraries. This caused unwanted dependencies to be created. This problem
1969 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1970 pcretest is linked with readline.
1971
1972 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1973 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1974 moved outside it again, because FALSE and TRUE are already defined in AIX,
1975 but BOOL is not.
1976
1977 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1978 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1979
1980 4. The pcregrep documentation said spaces were inserted as well as colons (or
1981 hyphens) following file names and line numbers when outputting matching
1982 lines. This is not true; no spaces are inserted. I have also clarified the
1983 wording for the --colour (or --color) option.
1984
1985 5. In pcregrep, when --colour was used with -o, the list of matching strings
1986 was not coloured; this is different to GNU grep, so I have changed it to be
1987 the same.
1988
1989 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1990 each matching line was coloured. Now it goes on to look for further matches
1991 of any of the test patterns, which is the same behaviour as GNU grep.
1992
1993 7. A pattern that could match an empty string could cause pcregrep to loop; it
1994 doesn't make sense to accept an empty string match in pcregrep, so I have
1995 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1996 seems to be how GNU grep behaves. [But see later change 40 for release
1997 8.33.]
1998
1999 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2000 start or after a newline", because the conditional assertion was not being
2001 correctly handled. The rule now is that both the assertion and what follows
2002 in the first alternative must satisfy the test.
2003
2004 9. If auto-callout was enabled in a pattern with a conditional group whose
2005 condition was an assertion, PCRE could crash during matching, both with
2006 pcre_exec() and pcre_dfa_exec().
2007
2008 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2009 used for matching.
2010
2011 11. Unicode property support in character classes was not working for
2012 characters (bytes) greater than 127 when not in UTF-8 mode.
2013
2014 12. Added the -M command line option to pcretest.
2015
2016 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2017
2018 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2019
2020 16. Added comments and documentation about mis-use of no_arg in the C++
2021 wrapper.
2022
2023 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2024 from Martin Jerabek that uses macro names for all relevant character and
2025 string constants.
2026
2027 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2028 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2029 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2030 these, but not everybody uses configure.
2031
2032 19. A conditional group that had only one branch was not being correctly
2033 recognized as an item that could match an empty string. This meant that an
2034 enclosing group might also not be so recognized, causing infinite looping
2035 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2036 with the subject "ab", where knowledge that the repeated group can match
2037 nothing is needed in order to break the loop.
2038
2039 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2040 exec(), but without supplying a callout function, matching went wrong.
2041
2042 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2043 leak if the size of the offset vector was greater than 30. When the vector
2044 is smaller, the saved offsets during recursion go onto a local stack
2045 vector, but for larger vectors malloc() is used. It was failing to free
2046 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2047 error, in fact).
2048
2049 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2050 heapframe that is used only when UTF-8 support is enabled. This caused no
2051 problem, but was untidy.
2052
2053 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2054 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2055 included within another project.
2056
2057 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2058 slightly modified by me:
2059
2060 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2061 not building pcregrep.
2062
2063 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2064 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2065
2066 25. Forward references, both numeric and by name, in patterns that made use of
2067 duplicate group numbers, could behave incorrectly or give incorrect errors,
2068 because when scanning forward to find the reference group, PCRE was not
2069 taking into account the duplicate group numbers. A pattern such as
2070 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2071
2072 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2073 the feature of a custom pointer more persuasive (as requested by a user).
2074
2075 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2076 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2077 pre-defined.
2078
2079 28. Added support for (*UTF8) at the start of a pattern.
2080
2081 29. Arrange for flags added by the "release type" setting in CMake to be shown
2082 in the configuration summary.
2083
2084
2085 Version 7.8 05-Sep-08
2086 ---------------------
2087
2088 1. Replaced UCP searching code with optimized version as implemented for Ad
2089 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2090 stage table and inline lookup instead of a function, giving speed ups of 2
2091 to 5 times on some simple patterns that I tested. Permission was given to
2092 distribute the MultiStage2.py script that generates the tables (it's not in
2093 the tarball, but is in the Subversion repository).
2094
2095 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2096 scripts.
2097
2098 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2099 a group with a zero qualifier. The result of the study could be incorrect,
2100 or the function might crash, depending on the pattern.
2101
2102 4. Caseless matching was not working for non-ASCII characters in back
2103 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2104 It now works when Unicode Property Support is available.
2105
2106 5. In pcretest, an escape such as \x{de} in the data was always generating
2107 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2108 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2109 truncation.
2110
2111 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2112
2113 7. Added two (int) casts to pcregrep when printing the difference of two
2114 pointers, in case they are 64-bit values.
2115
2116 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2117 test 2 if it fails.
2118
2119 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2120 and a #define of that name to empty if it is not externally set. This is to
2121 allow users of MSVC to set it if necessary.
2122
2123 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2124 the convenience functions in the pcre_get.c source file.
2125
2126 11. An option change at the start of a pattern that had top-level alternatives
2127 could cause overwriting and/or a crash. This command provoked a crash in
2128 some environments:
2129
2130 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2131
2132 This potential security problem was recorded as CVE-2008-2371.
2133
2134 12. For a pattern where the match had to start at the beginning or immediately
2135 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2136 pcre_dfa_exec() could read past the end of the passed subject if there was
2137 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2138 pcretest so that it places the subject at the end of its malloc-ed buffer.
2139
2140 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2141 exec() might read past the end of the data buffer in UTF-8 mode.
2142
2143 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2144 the data contained the byte 0x85 as part of a UTF-8 character within its
2145 first line. This applied both to normal and DFA matching.
2146
2147 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2148 /^[^d]*?$/8 failed to match "abc".
2149
2150 16. Added a missing copyright notice to pcrecpp_internal.h.
2151
2152 17. Make it more clear in the documentation that values returned from
2153 pcre_exec() in ovector are byte offsets, not character counts.
2154
2155 18. Tidied a few places to stop certain compilers from issuing warnings.
2156
2157 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2158 supplied by Stefan Weber. I made a further small update for 7.8 because
2159 there is a change of source arrangements: the pcre_searchfuncs.c module is
2160 replaced by pcre_ucd.c.
2161
2162
2163 Version 7.7 07-May-08
2164 ---------------------
2165
2166 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2167 a string to a long long, pretend we don't even have a long long." This is
2168 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2169
2170 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2171 pre-7.6 versions, which defined a global no_arg variable instead of putting
2172 it in the RE class. (See also #8 below.)
2173
2174 3. Remove a line of dead code, identified by coverity and reported by Nuno
2175 Lopes.
2176
2177 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2178
2179 (1) The include/exclude patterns were being applied to the whole pathnames
2180 of files, instead of just to the final components.
2181
2182 (2) If there was more than one level of directory, the subdirectories were
2183 skipped unless they satisfied the include/exclude conditions. This is
2184 inconsistent with GNU grep (and could even be seen as contrary to the
2185 pcregrep specification - which I improved to make it absolutely clear).
2186 The action now is always to scan all levels of directory, and just
2187 apply the include/exclude patterns to regular files.
2188
2189 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2190 --exclude_dir in the tests to avoid scanning .svn directories.
2191
2192 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2193 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2194 doesn't support NULs in patterns.
2195
2196 7. Added some missing "const"s to declarations of static tables in
2197 pcre_compile.c and pcre_dfa_exec.c.
2198
2199 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2200 caused by fix #2 above. (Subsequently also a second patch to fix the
2201 first patch. And a third patch - this was a messy problem.)
2202
2203 9. Applied Craig's patch to remove the use of push_back().
2204
2205 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2206 matching function regexec().
2207
2208 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2209 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2210 references. PCRE supports relative numbers with this syntax (I don't think
2211 Oniguruma does).
2212
2213 12. Previously, a group with a zero repeat such as (...){0} was completely
2214 omitted from the compiled regex. However, this means that if the group
2215 was called as a subroutine from elsewhere in the pattern, things went wrong
2216 (an internal error was given). Such groups are now left in the compiled
2217 pattern, with a new opcode that causes them to be skipped at execution
2218 time.
2219
2220 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2221 to the way PCRE behaves:
2222
2223 (a) A lone ] character is dis-allowed (Perl treats it as data).
2224
2225 (b) A back reference to an unmatched subpattern matches an empty string
2226 (Perl fails the current match path).
2227
2228 (c) A data ] in a character class must be notated as \] because if the
2229 first data character in a class is ], it defines an empty class. (In
2230 Perl it is not possible to have an empty class.) The empty class []
2231 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2232 The negative empty class [^] matches any one character, independently
2233 of the DOTALL setting.
2234
2235 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2236 non-existent subpattern following a character class starting with ']' and
2237 containing () gave an internal compiling error instead of "reference to
2238 non-existent subpattern". Fortunately, when the pattern did exist, the
2239 compiled code was correct. (When scanning forwards to check for the
2240 existence of the subpattern, it was treating the data ']' as terminating
2241 the class, so got the count wrong. When actually compiling, the reference
2242 was subsequently set up correctly.)
2243
2244 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2245 it was being rejected as not supported by pcre_dfa_exec(), even though
2246 other assertions are supported. I have made pcre_dfa_exec() support
2247 (*FAIL).
2248
2249 16. The implementation of 13c above involved the invention of a new opcode,
2250 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2251 cannot be changed at match time, I realized I could make a small
2252 improvement to matching performance by compiling OP_ALLANY instead of
2253 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2254 on the OP_ANY path.
2255
2256 17. Compiling pcretest on Windows with readline support failed without the
2257 following two fixes: (1) Make the unistd.h include conditional on
2258 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2259
2260 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2261 ncurses library to be included for pcretest when ReadLine support is
2262 requested, but also to allow for it to be overridden. This patch came from
2263 Daniel Bergström.
2264
2265 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2266 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2267 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2268 spotting this.
2269
2270
2271 Version 7.6 28-Jan-08
2272 ---------------------
2273
2274 1. A character class containing a very large number of characters with
2275 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2276 overflow.
2277
2278 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2279 HAVE_LONG_LONG is not defined.
2280
2281 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2282 bring them up to date and include new features. This patch includes:
2283
2284 - Fixed PH's badly added libz and libbz2 support.
2285 - Fixed a problem with static linking.
2286 - Added pcredemo. [But later removed - see 7 below.]
2287 - Fixed dftables problem and added an option.
2288 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2289 HAVE_LONG_LONG.
2290 - Added readline support for pcretest.
2291 - Added an listing of the option settings after cmake has run.
2292
2293 4. A user submitted a patch to Makefile that makes it easy to create
2294 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2295 Makefile.am that cause it to include this special target, without
2296 affecting anything else. Note that the same mingw target plus all
2297 the other distribution libraries and programs are now supported
2298 when configuring with CMake (see 6 below) instead of with
2299 Configure/Make.
2300
2301 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2302 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2303 exported in the Windows port". It has not yet been confirmed that the patch
2304 solves the problem, but it does no harm.
2305
2306 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2307 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2308 with CMake, and also correct the comment about stack recursion.
2309
2310 7. Remove the automatic building of pcredemo from the ./configure system and
2311 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2312 of a program that users should build themselves after PCRE is installed, so
2313 building it automatically is not really right. What is more, it gave
2314 trouble in some build environments.
2315
2316 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2317
2318
2319 Version 7.5 10-Jan-08
2320 ---------------------
2321
2322 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2323 values in parens when parsing an RE using the C++ wrapper."
2324
2325 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2326 Characters greater than 255 were excluded from the class instead of being
2327 included.
2328
2329 3. The same bug as (2) above applied to negated POSIX classes such as
2330 [:^space:].
2331
2332 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2333 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2334 I have changed it.
2335
2336 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2337 first named subpattern) and a construct such as (?&a) would reference the
2338 first named subpattern whose name started with "a" (in other words, the
2339 length check was missing). Both these problems are fixed. "Subpattern name
2340 expected" is now given for (?&) (a zero-length name), and this patch also
2341 makes it give the same error for \k'' (previously it complained that that
2342 was a reference to a non-existent subpattern).
2343
2344 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2345 this is right because (?- can be followed by option settings as well as by
2346 digits. I have, however, made the messages clearer.
2347
2348 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2349 than the number used in the conditional) now cause a compile-time error.
2350 This is actually not compatible with Perl, which accepts such patterns, but
2351 treats the conditional as always being FALSE (as PCRE used to), but it
2352 seems to me that giving a diagnostic is better.
2353
2354 8. Change "alphameric" to the more common word "alphanumeric" in comments
2355 and messages.
2356
2357 9. Fix two occurrences of "backslash" in comments that should have been
2358 "backspace".
2359
2360 10. Remove two redundant lines of code that can never be obeyed (their function
2361 was moved elsewhere).
2362
2363 11. The program that makes PCRE's Unicode character property table had a bug
2364 which caused it to generate incorrect table entries for sequences of
2365 characters that have the same character type, but are in different scripts.
2366 It amalgamated them into a single range, with the script of the first of
2367 them. In other words, some characters were in the wrong script. There were
2368 thirteen such cases, affecting characters in the following ranges:
2369
2370 U+002b0 - U+002c1
2371 U+0060c - U+0060d
2372 U+0061e - U+00612
2373 U+0064b - U+0065e
2374 U+0074d - U+0076d
2375 U+01800 - U+01805
2376 U+01d00 - U+01d77
2377 U+01d9b - U+01dbf
2378 U+0200b - U+0200f
2379 U+030fc - U+030fe
2380 U+03260 - U+0327f
2381 U+0fb46 - U+0fbb1
2382 U+10450 - U+1049d
2383
2384 12. The -o option (show only the matching part of a line) for pcregrep was not
2385 compatible with GNU grep in that, if there was more than one match in a
2386 line, it showed only the first of them. It now behaves in the same way as
2387 GNU grep.
2388
2389 13. If the -o and -v options were combined for pcregrep, it printed a blank
2390 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2391 does the same. The return code can be used to tell if there were any
2392 non-matching lines.
2393
2394 14. Added --file-offsets and --line-offsets to pcregrep.
2395
2396 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2397 infinitely looping recursion. The bug was that positive lookaheads were not
2398 being skipped when checking for a possible empty match (negative lookaheads
2399 and both kinds of lookbehind were skipped).
2400
2401 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2402 inclusion of <windows.h> to before rather than after the definition of
2403 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2404
2405 17. Specifying a possessive quantifier with a specific limit for a Unicode
2406 character property caused pcre_compile() to compile bad code, which led at
2407 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2408 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2409 caused the error; without that there was no problem.
2410
2411 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2412
2413 19. Added --enable-pcretest-libreadline.
2414
2415 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2416 RE::GlobalReplace(). As a result, the number of replacements returned was
2417 double what it should be. I removed one of the increments, but Craig sent a
2418 later patch that removed the other one (the right fix) and added unit tests
2419 that check the return values (which was not done before).
2420
2421 21. Several CMake things:
2422
2423 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2424 the names libpcre and libpcreposix, not just pcre and pcreposix.
2425
2426 (2) The above change means that pcretest and pcregrep are now correctly
2427 linked with the newly-built libraries, not previously installed ones.
2428
2429 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2430
2431 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2432 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2433 UTF-8 newline character). The key issue is that the pattern starts .*;
2434 this means that the match must be either at the beginning, or after a
2435 newline. The bug was in the code for advancing after a failed match and
2436 checking that the new position followed a newline. It was not taking
2437 account of UTF-8 characters correctly.
2438
2439 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2440 character classes. PCRE was not treating the sequence [:...:] as a
2441 character class unless the ... were all letters. Perl, however, seems to
2442 allow any characters between [: and :], though of course it rejects as
2443 unknown any "names" that contain non-letters, because all the known class
2444 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2445 for example, whereas PCRE did not - it did not recognize a POSIX character
2446 class. This seemed a bit dangerous, so the code has been changed to be
2447 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2448 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2449 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2450 Perl does, and where it didn't before.
2451
2452 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2453 Windows environments %n is disabled by default.
2454
2455
2456 Version 7.4 21-Sep-07
2457 ---------------------
2458
2459 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2460 means that a class such as [\s] counted as "explicit reference to CR or
2461 LF". That isn't really right - the whole point of the change was to try to
2462 help when there was an actual mention of one of the two characters. So now
2463 the change happens only if \r or \n (or a literal CR or LF) character is
2464 encountered.
2465
2466 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2467 of both had grown to the point where there were only 3 bits left.
2468 Fortunately, there was spare space in the data structure, and so I have
2469 moved the internal flags into a new 16-bit field to free up more option
2470 bits.
2471
2472 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2473 but did not set the internal JCHANGED flag - either of these is enough to
2474 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2475 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2476 start sets both bits.
2477
2478 4. Added options (at build time, compile time, exec time) to change \R from
2479 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2480
2481 5. doc/pcresyntax.html was missing from the distribution.
2482
2483 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2484 compatibility, even though it is no longer used.
2485
2486 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2487 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2488 windows.h file is present (where different names are used). [This was
2489 reversed later after testing - see 16 below.]
2490
2491 8. Changed all #include <config.h> to #include "config.h". There were also
2492 some further <pcre.h> cases that I changed to "pcre.h".
2493
2494 9. When pcregrep was used with the --colour option, it missed the line ending
2495 sequence off the lines that it output.
2496
2497 10. It was pointed out to me that arrays of string pointers cause lots of
2498 relocations when a shared library is dynamically loaded. A technique of
2499 using a single long string with a table of offsets can drastically reduce
2500 these. I have refactored PCRE in four places to do this. The result is
2501 dramatic:
2502
2503 Originally: 290
2504 After changing UCP table: 187
2505 After changing error message table: 43
2506 After changing table of "verbs" 36
2507 After changing table of Posix names 22
2508
2509 Thanks to the folks working on Gregex for glib for this insight.
2510
2511 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2512 unicode-properties was also set.
2513
2514 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2515
2516 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2517 checked only for CRLF.
2518
2519 14. Added casts to pcretest.c to avoid compiler warnings.
2520
2521 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2522
2523 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2524 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2525 entirely. This removes changes made in 7 above.
2526
2527 17. The CMake files have been updated, and there is now more information about
2528 building with CMake in the NON-UNIX-USE document.
2529
2530
2531 Version 7.3 28-Aug-07
2532 ---------------------
2533
2534 1. In the rejigging of the build system that eventually resulted in 7.1, the
2535 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2536 brackets there is not right, since it causes compilers to look for an
2537 installed pcre.h, not the version that is in the source that is being
2538 compiled (which of course may be different). I have changed it back to:
2539
2540 #include "pcre.h"
2541
2542 I have a vague recollection that the change was concerned with compiling in
2543 different directories, but in the new build system, that is taken care of
2544 by the VPATH setting the Makefile.
2545
2546 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2547 when the subject happened to end in the byte 0x85 (e.g. if the last
2548 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2549 characters but of course it shouldn't be taken as a newline when it is part
2550 of another character. The bug was that, for an unlimited repeat of . in
2551 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2552 characters when looking for a newline.
2553
2554 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2555
2556 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2557 in debug output.
2558
2559 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2560 long printing in the pcrecpp unittest when running under MinGW.
2561
2562 6. ESC_K was left out of the EBCDIC table.
2563
2564 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2565 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2566 limit also applies to "virtual nesting" when a pattern is recursive, and in
2567 this case 1000 isn't so big. I have been able to remove this limit at the
2568 expense of backing off one optimization in certain circumstances. Normally,
2569 when pcre_exec() would call its internal match() function recursively and
2570 immediately return the result unconditionally, it uses a "tail recursion"
2571 feature to save stack. However, when a subpattern that can match an empty
2572 string has an unlimited repetition quantifier, it no longer makes this
2573 optimization. That gives it a stack frame in which to save the data for
2574 checking that an empty string has been matched. Previously this was taken
2575 from the 1000-entry workspace that had been reserved. So now there is no
2576 explicit limit, but more stack is used.
2577
2578 8. Applied Daniel's patches to solve problems with the import/export magic
2579 syntax that is required for Windows, and which was going wrong for the
2580 pcreposix and pcrecpp parts of the library. These were overlooked when this
2581 problem was solved for the main library.
2582
2583 9. There were some crude static tests to avoid integer overflow when computing
2584 the size of patterns that contain repeated groups with explicit upper
2585 limits. As the maximum quantifier is 65535, the maximum group length was
2586 set at 30,000 so that the product of these two numbers did not overflow a
2587 32-bit integer. However, it turns out that people want to use groups that
2588 are longer than 30,000 bytes (though not repeat them that many times).
2589 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2590 made it possible to implement the integer overflow checks in a much more
2591 dynamic way, which I have now done. The artificial limitation on group
2592 length has been removed - we now have only the limit on the total length of
2593 the compiled pattern, which depends on the LINK_SIZE setting.
2594
2595 10. Fixed a bug in the documentation for get/copy named substring when
2596 duplicate names are permitted. If none of the named substrings are set, the
2597 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2598 empty string.
2599
2600 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2601 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2602 because the ] is interpreted as the first data character and the
2603 terminating ] is not found. PCRE has been made compatible with Perl in this
2604 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2605 cause memory overwriting.
2606
2607 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2608 string has been matched (to stop an infinite loop). It was not recognizing
2609 a conditional subpattern that could match an empty string if that
2610 subpattern was within another subpattern. For example, it looped when
2611 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2612 condition was not nested. This bug has been fixed.
2613
2614 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2615 past the start of the subject in the presence of bytes with the top bit
2616 set, for example "\x8aBCD".
2617
2618 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2619 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2620
2621 14. Optimized (?!) to (*FAIL).
2622
2623 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2624 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2625 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2626 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2627 does: it's just the validity check that is more restrictive.
2628
2629 16. Inserted checks for integer overflows during escape sequence (backslash)
2630 processing, and also fixed erroneous offset values for syntax errors during
2631 backslash processing.
2632
2633 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2634 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2635
2636 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2637 caused an overrun.
2638
2639 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2640 something other than just ASCII characters) inside a group that had an
2641 unlimited repeat caused a loop at compile time (while checking to see
2642 whether the group could match an empty string).
2643
2644 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2645 [\P{Any}] did so. (Error in the code for printing property names.)
2646
2647 21. An orphan \E inside a character class could cause a crash.
2648
2649 22. A repeated capturing bracket such as (A)? could cause a wild memory
2650 reference during compilation.
2651
2652 23. There are several functions in pcre_compile() that scan along a compiled
2653 expression for various reasons (e.g. to see if it's fixed length for look
2654 behind). There were bugs in these functions when a repeated \p or \P was
2655 present in the pattern. These operators have additional parameters compared
2656 with \d, etc, and these were not being taken into account when moving along
2657 the compiled data. Specifically:
2658
2659 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2660 length.
2661
2662 (b) An item such as \pL+ within a repeated group could cause crashes or
2663 loops.
2664
2665 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2666 "reference to non-existent subpattern" error.
2667
2668 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2669
2670 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2671 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2672
2673 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2674
2675 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2676 character were causing crashes (broken optimization).
2677
2678 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2679 \p or \P) caused a compile-time loop.
2680
2681 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2682 break. For example, the unstudied pattern [\r\n]A does not match the string
2683 "\r\nA" because change 7.0/46 below moves the current point on by two
2684 characters after failing to match at the start. However, the pattern \nA
2685 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2686 the same is true. There doesn't seem any very clean way out of this, but
2687 what I have chosen to do makes the common cases work: PCRE now takes note
2688 of whether there can be an explicit match for \r or \n anywhere in the
2689 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2690 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2691 pattern has explicit CR or LF references.
2692
2693 29. Added (*CR) etc for changing newline setting at start of pattern.
2694
2695
2696 Version 7.2 19-Jun-07
2697 ---------------------
2698
2699 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2700 which is apparently normally available under Windows.
2701
2702 2. Re-jig the pcregrep tests with different newline settings in an attempt
2703 to make them independent of the local environment's newline setting.
2704
2705 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2706
2707 4. Some of the "internals" tests were previously cut out when the link size
2708 was not 2, because the output contained actual offsets. The recent new
2709 "Z" feature of pcretest means that these can be cut out, making the tests
2710 usable with all link sizes.
2711
2712 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2713 stack recursion. This gives a massive performance boost under BSD, but just
2714 a small improvement under Linux. However, it saves one field in the frame
2715 in all cases.
2716
2717 6. Added more features from the forthcoming Perl 5.10:
2718
2719 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2720 recursion call. It refers to the nth most recently opened parentheses.
2721
2722 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2723 to be opened parentheses.
2724
2725 (c) Conditions that refer to capturing parentheses can be specified
2726 relatively, for example, (?(-2)... or (?(+3)...
2727
2728 (d) \K resets the start of the current match so that everything before
2729 is not part of it.
2730
2731 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2732
2733 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2734 reference syntax.
2735
2736 (g) (?| introduces a group in which the numbering of parentheses in each
2737 alternative starts with the same number.
2738
2739 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2740
2741 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2742 PCRE_INFO_JCHANGED.
2743
2744 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2745 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2746 for detecting groups that can match an empty string.
2747
2748 9. A pattern with a very large number of alternatives (more than several
2749 hundred) was running out of internal workspace during the pre-compile
2750 phase, where pcre_compile() figures out how much memory will be needed. A
2751 bit of new cunning has reduced the workspace needed for groups with
2752 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2753 workspace instead of running out of the 4096 that are available.
2754
2755 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2756
2757 11. Applied patch from Google to remove an optimization that didn't quite work.
2758 The report of the bug said:
2759
2760 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2761 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2762 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2763
2764 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2765 it matched the wrong number of bytes.
2766
2767
2768 Version 7.1 24-Apr-07
2769 ---------------------
2770
2771 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2772 that is more "standard", making use of automake and other Autotools. There
2773 is some re-arrangement of the files and adjustment of comments consequent
2774 on this.
2775
2776 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2777 for recursive directory scanning broke on some systems because the files
2778 are not scanned in any specific order and on different systems the order
2779 was different. A call to "sort" has been inserted into RunGrepTest for the
2780 approprate test as a short-term fix. In the longer term there may be an
2781 alternative.
2782
2783 3. I had an email from Eric Raymond about problems translating some of PCRE's
2784 man pages to HTML (despite the fact that I distribute HTML pages, some
2785 people do their own conversions for various reasons). The problems
2786 concerned the use of low-level troff macros .br and .in. I have therefore
2787 removed all such uses from the man pages (some were redundant, some could
2788 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2789 HTML has been updated to handle .nf/.fi and to complain if it encounters
2790 .br or .in.
2791
2792 4. Updated comments in configure.ac that get placed in config.h.in and also
2793 arranged for config.h to be included in the distribution, with the name
2794 config.h.generic, for the benefit of those who have to compile without
2795 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2796
2797 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2798 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2799 makevp.bat for the current PCRE, using the additional files
2800 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2801
2802 6. A Windows user reported a minor discrepancy with test 2, which turned out
2803 to be caused by a trailing space on an input line that had got lost in his
2804 copy. The trailing space was an accident, so I've just removed it.
2805
2806 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2807 that is needed.
2808
2809 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2810 as "const" (a) because they are and (b) because it helps the PHP
2811 maintainers who have recently made a script to detect big data structures
2812 in the php code that should be moved to the .rodata section. I remembered
2813 to update Builducptable as well, so it won't revert if ucptable.h is ever
2814 re-created.
2815
2816 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2817 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2818 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2819 support is not required. This saves 1.5-2K of code, which is important in
2820 some applications.
2821
2822 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2823 so as not to refer to the tables, even though these functions will never be
2824 called when UTF-8 support is disabled. Otherwise there are problems with a
2825 shared library.
2826
2827 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2828
2829 (a) It was defining its arguments as char * instead of void *.
2830
2831 (b) It was assuming that all moves were upwards in memory; this was true
2832 a long time ago when I wrote it, but is no longer the case.
2833
2834 The emulated memove() is provided for those environments that have neither
2835 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2836 is clearly not the case, as these two bugs were recently reported.
2837
2838 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2839 and Detrail to create the HTML documentation, the .txt form of the man
2840 pages, and it removes trailing spaces from listed files. It also creates
2841 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2842 case, it wraps all the #defines with #ifndefs. This script should be run
2843 before "make dist".
2844
2845 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2846 with Unicode property support.
2847
2848 (a) For a maximizing quantifier, if the two different cases of the
2849 character were of different lengths in their UTF-8 codings (there are
2850 some cases like this - I found 11), and the matching function had to
2851 back up over a mixture of the two cases, it incorrectly assumed they
2852 were both the same length.
2853
2854 (b) When PCRE was configured to use the heap rather than the stack for
2855 recursion during matching, it was not correctly preserving the data for
2856 the other case of a UTF-8 character when checking ahead for a match
2857 while processing a minimizing repeat. If the check also involved
2858 matching a wide character, but failed, corruption could cause an
2859 erroneous result when trying to check for a repeat of the original
2860 character.
2861
2862 13. Some tidying changes to the testing mechanism:
2863
2864 (a) The RunTest script now detects the internal link size and whether there
2865 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2866 values substituted by "configure". (The RunGrepTest script already did
2867 this for UTF-8.) The configure.ac script no longer substitutes the
2868 relevant variables.
2869
2870 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2871 with length and offset values. This means that the output is different
2872 for different internal link sizes. Test 2 is skipped for link sizes
2873 other than 2 because of this, bypassing the problem. Unfortunately,
2874 there was also a test in test 3 (the locale tests) that used /B and
2875 failed for link sizes other than 2. Rather than cut the whole test out,
2876 I have added a new /Z option to pcretest that replaces the length and
2877 offset values with spaces. This is now used to make test 3 independent
2878 of link size. (Test 2 will be tidied up later.)
2879
2880 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2881 segmentation fault instead of returning the appropriate error message.
2882
2883 15. In multiline mode when the newline sequence was set to "any", the pattern
2884 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2885 This doesn't seem right; it now treats the CRLF combination as the line
2886 ending, and so does not match in that case. It's only a pattern such as ^$
2887 that would hit this one: something like ^ABC$ would have failed after \r
2888 and then tried again after \r\n.
2889
2890 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2891 in an attempt to make files that differ only in their line terminators
2892 compare equal. This works on Linux.
2893
2894 17. Under certain error circumstances pcregrep might try to free random memory
2895 as it exited. This is now fixed, thanks to valgrind.
2896
2897 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2898 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2899 was because its rules for how to advance for /g after matching an empty
2900 string at the end of a line did not allow for this case. They now check for
2901 it specially.
2902
2903 20. pcretest is supposed to handle patterns and data of any length, by
2904 extending its buffers when necessary. It was getting this wrong when the
2905 buffer for a data line had to be extended.
2906
2907 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2908 CRLF as a newline sequence.
2909
2910 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2911 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2912 I have nevertheless tidied it up.
2913
2914 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2915
2916 24. Added a man page for pcre-config.
2917
2918
2919 Version 7.0 19-Dec-06
2920 ---------------------
2921
2922 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2923 moving to gcc 4.1.1.
2924
2925 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2926 sys/time.h, which is documented as needed for this function. It doesn't
2927 seem to matter on Linux, but it showed up on some releases of OS X.
2928
2929 3. It seems that there are systems where bytes whose values are greater than
2930 127 match isprint() in the "C" locale. The "C" locale should be the
2931 default when a C program starts up. In most systems, only ASCII printing
2932 characters match isprint(). This difference caused the output from pcretest
2933 to vary, making some of the tests fail. I have changed pcretest so that:
2934
2935 (a) When it is outputting text in the compiled version of a pattern, bytes
2936 other than 32-126 are always shown as hex escapes.
2937
2938 (b) When it is outputting text that is a matched part of a subject string,
2939 it does the same, unless a different locale has been set for the match
2940 (using the /L modifier). In this case, it uses isprint() to decide.
2941
2942 4. Fixed a major bug that caused incorrect computation of the amount of memory
2943 required for a compiled pattern when options that changed within the
2944 pattern affected the logic of the preliminary scan that determines the
2945 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2946 that the computed length was too small. The symptoms of this bug were
2947 either the PCRE error "internal error: code overflow" from pcre_compile(),
2948 or a glibc crash with a message such as "pcretest: free(): invalid next
2949 size (fast)". Examples of patterns that provoked this bug (shown in
2950 pcretest format) are:
2951
2952 /(?-x: )/x
2953 /(?x)(?-x: \s*#\s*)/
2954 /((?i)[\x{c0}])/8
2955 /(?i:[\x{c0}])/8
2956
2957 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2958 is now done differently.
2959
2960 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2961 wrapper classes; (b) implement a new function in the C++ scanner that is
2962 more efficient than the old way of doing things because it avoids levels of
2963 recursion in the regex matching; (c) add a paragraph to the documentation
2964 for the FullMatch() function.
2965
2966 6. The escape sequence \n was being treated as whatever was defined as
2967 "newline". Not only was this contrary to the documentation, which states
2968 that \n is character 10 (hex 0A), but it also went horribly wrong when
2969 "newline" was defined as CRLF. This has been fixed.
2970
2971 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2972 was being set to -1 for the "end of line" case (supposedly a value that no
2973 character can have). Though this value is never used (the check for end of
2974 line is "zero bytes in current character"), it caused compiler complaints.
2975 I've changed it to 0xffffffff.
2976
2977 8. In pcre_version.c, the version string was being built by a sequence of
2978 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2979 string (as it is for production releases) called a macro with an empty
2980 argument. The C standard says the result of this is undefined. The gcc
2981 compiler treats it as an empty string (which was what was wanted) but it is
2982 reported that Visual C gives an error. The source has been hacked around to
2983 avoid this problem.
2984
2985 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2986 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2987 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2988 of them did).
2989
2990 10. Originally, pcretest opened its input and output without "b"; then I was
2991 told that "b" was needed in some environments, so it was added for release
2992 5.0 to both the input and output. (It makes no difference on Unix-like
2993 systems.) Later I was told that it is wrong for the input on Windows. I've
2994 now abstracted the modes into two macros, to make it easier to fiddle with
2995 them, and removed "b" from the input mode under Windows.
2996
2997 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2998
2999 12. Added -help and --help to pcretest as an official way of being reminded
3000 of the options.
3001
3002 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3003 and pcrecpp.cc because they annoy compilers at high warning levels.
3004
3005 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3006
3007 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3008 scripts are not C programs :-) and which was not noticed because it works
3009 on Linux.
3010
3011 16. pcretest is supposed to handle any length of pattern and data line (as one
3012 line or as a continued sequence of lines) by extending its input buffer if
3013 necessary. This feature was broken for very long pattern lines, leading to
3014 a string of junk being passed to pcre_compile() if the pattern was longer
3015 than about 50K.
3016
3017 17. I have done a major re-factoring of the way pcre_compile() computes the
3018 amount of memory needed for a compiled pattern. Previously, there was code
3019 that made a preliminary scan of the pattern in order to do this. That was
3020 OK when PCRE was new, but as the facilities have expanded, it has become
3021 harder and harder to keep it in step with the real compile phase, and there
3022 have been a number of bugs (see for example, 4 above). I have now found a
3023 cunning way of running the real compile function in a "fake" mode that
3024 enables it to compute how much memory it would need, while actually only
3025 ever using a few hundred bytes of working memory and without too many
3026 tests of the mode. This should make future maintenance and development
3027 easier. A side effect of this work is that the limit of 200 on the nesting
3028 depth of parentheses has been removed (though this was never a serious
3029 limitation, I suspect). However, there is a downside: pcre_compile() now
3030 runs more slowly than before (30% or more, depending on the pattern). I
3031 hope this isn't a big issue. There is no effect on runtime performance.
3032
3033 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3034 newline (only possible for the last line of a file) and it was a
3035 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3036
3037 19. Added additional timing features to pcretest. (1) The -tm option now times
3038 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3039 separate command line item, by a number that specifies the number of
3040 repeats to use when timing. The default is 50000; this gives better
3041 precision, but takes uncomfortably long for very large patterns.
3042
3043 20. Extended pcre_study() to be more clever in cases where a branch of a
3044 subpattern has no definite first character. For example, (a*|b*)[cd] would
3045 previously give no result from pcre_study(). Now it recognizes that the
3046 first character must be a, b, c, or d.
3047
3048 21. There was an incorrect error "recursive call could loop indefinitely" if
3049 a subpattern (or the entire pattern) that was being tested for matching an
3050 empty string contained only one non-empty item after a nested subpattern.
3051 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3052 incorrectly, because the \d was being skipped in the check.
3053
3054 22. The pcretest program now has a new pattern option /B and a command line
3055 option -b, which is equivalent to adding /B to every pattern. This causes
3056 it to show the compiled bytecode, without the additional information that
3057 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3058 is the same as /B/I).
3059
3060 23. A new optimization is now able automatically to treat some sequences such
3061 as a*b as a*+b. More specifically, if something simple (such as a character
3062 or a simple class like \d) has an unlimited quantifier, and is followed by
3063 something that cannot possibly match the quantified thing, the quantifier
3064 is automatically "possessified".
3065
3066 24. A recursive reference to a subpattern whose number was greater than 39
3067 went wrong under certain circumstances in UTF-8 mode. This bug could also
3068 have affected the operation of pcre_study().
3069
3070 25. Realized that a little bit of performance could be had by replacing
3071 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3072
3073 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3074
3075 27. Possessive quantifiers such as a++ were previously implemented by turning
3076 them into atomic groups such as ($>a+). Now they have their own opcodes,
3077 which improves performance. This includes the automatically created ones
3078 from 23 above.
3079
3080 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3081 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3082 the first matched character to be a colon. This applied both to named and
3083 numbered groups.
3084
3085 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3086
3087 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3088 building PCRE on Windows easier, so I have included it in the distribution.
3089
3090 31. There is now a check in pcretest against a ridiculously large number being
3091 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3092 loop, the loop is abandoned.
3093
3094 32. Forward references to subpatterns in conditions such as (?(2)...) where
3095 subpattern 2 is defined later cause pcre_compile() to search forwards in
3096 the pattern for the relevant set of parentheses. This search went wrong
3097 when there were unescaped parentheses in a character class, parentheses
3098 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3099
3100 33. "Subroutine" calls and backreferences were previously restricted to
3101 referencing subpatterns earlier in the regex. This restriction has now
3102 been removed.
3103
3104 34. Added a number of extra features that are going to be in Perl 5.10. On the
3105 whole, these are just syntactic alternatives for features that PCRE had
3106 previously implemented using the Python syntax or my own invention. The
3107 other formats are all retained for compatibility.
3108
3109 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3110 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3111 also .NET compatible.
3112
3113 (b) A recursion or subroutine call to a named group can now be defined as
3114 (?&name) as well as (?P>name).
3115
3116 (c) A backreference to a named group can now be defined as \k<name> or
3117 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3118 5.10, are also .NET compatible.
3119
3120 (d) A conditional reference to a named group can now use the syntax
3121 (?(<name>) or (?('name') as well as (?(name).
3122
3123 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3124 groups (named and numbered) that are never evaluated inline, but can be
3125 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3126 is always false. There may be only one alternative in such a group.
3127
3128 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3129 as the simple (?(R). The condition is true only if the most recent
3130 recursion is that of the given number or name. It does not search out
3131 through the entire recursion stack.
3132
3133 (g) The escape \gN or \g{N} has been added, where N is a positive or
3134 negative number, specifying an absolute or relative reference.
3135
3136 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3137 some "unreachable code" warnings.
3138
3139 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3140 things, this adds five new scripts.
3141
3142 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3143 There were also incompatibilities regarding the handling of \Q..\E inside
3144 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3145 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3146
3147 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3148 matches an empty string, and forcibly breaks the loop. There were bugs in
3149 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3150 against aaaa the result was just "a" rather than "aaaa", for example. Two
3151 separate and independent bugs (that affected different cases) have been
3152 fixed.
3153
3154 39. Refactored the code to abolish the use of different opcodes for small
3155 capturing bracket numbers. This is a tidy that I avoided doing when I
3156 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3157 The new approach is not only tidier, it makes it possible to reduce the
3158 memory needed to fix the previous bug (38).
3159
3160 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3161 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3162 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3163 mode.
3164
3165 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3166 report.
3167
3168 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3169 copy construction and assignment in the C++ wrapper.
3170
3171 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3172 couple of bugs that could have given wrong results in the "--newline=crlf"
3173 case.
3174
3175 44. Added a number of casts and did some reorganization of signed/unsigned int
3176 variables following suggestions from Dair Grant. Also renamed the variable
3177 "this" as "item" because it is a C++ keyword.
3178
3179 45. Arranged for dftables to add
3180
3181 #include "pcre_internal.h"
3182
3183 to pcre_chartables.c because without it, gcc 4.x may remove the array
3184 definition from the final binary if PCRE is built into a static library and
3185 dead code stripping is activated.
3186
3187 46. For an unanchored pattern, if a match attempt fails at the start of a
3188 newline sequence, and the newline setting is CRLF or ANY, and the next two
3189 characters are CRLF, advance by two characters instead of one.
3190
3191
3192 Version 6.7 04-Jul-06
3193 ---------------------
3194
3195 1. In order to handle tests when input lines are enormously long, pcretest has
3196 been re-factored so that it automatically extends its buffers when
3197 necessary. The code is crude, but this _is_ just a test program. The
3198 default size has been increased from 32K to 50K.
3199
3200 2. The code in pcre_study() was using the value of the re argument before
3201 testing it for NULL. (Of course, in any sensible call of the function, it
3202 won't be NULL.)
3203
3204 3. The memmove() emulation function in pcre_internal.h, which is used on
3205 systems that lack both memmove() and bcopy() - that is, hardly ever -
3206 was missing a "static" storage class specifier.
3207
3208 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3209 containing an extended class (one that cannot be represented by a bitmap
3210 because it contains high-valued characters or Unicode property items, e.g.
3211 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3212 pattern, but PCRE should not loop if you do not (it no longer does).
3213 [Detail: two cases were found: (a) a repeated subpattern containing an
3214 extended class; (b) a recursive reference to a subpattern that followed a
3215 previous extended class. It wasn't skipping over the extended class
3216 correctly when UTF-8 mode was not set.]
3217
3218 5. A negated single-character class was not being recognized as fixed-length
3219 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3220 compile error "lookbehind assertion is not fixed length".
3221
3222 6. The RunPerlTest auxiliary script was showing an unexpected difference
3223 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3224 write a Perl script that can interpret lines of an input file either as
3225 byte characters or as UTF-8, which is what "perltest" was being required to
3226 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3227 can't do is switch easily at run time between having the "use utf8;" pragma
3228 or not. In the end, I fudged it by using the RunPerlTest script to insert
3229 "use utf8;" explicitly for the UTF-8 tests.
3230
3231 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3232 the end of the subject string, contrary to the documentation and to what
3233 Perl does. This was true of both matching functions. Now it matches only at
3234 the start of the subject and immediately after *internal* newlines.
3235
3236 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3237 a pointer to an int instead of a pointer to an unsigned long int. This
3238 caused problems on 64-bit systems.
3239
3240 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3241 instance of the 'standard' template library not being so standard".
3242
3243 10. There was no check on the number of named subpatterns nor the maximum
3244 length of a subpattern name. The product of these values is used to compute
3245 the size of the memory block for a compiled pattern. By supplying a very
3246 long subpattern name and a large number of named subpatterns, the size
3247 computation could be caused to overflow. This is now prevented by limiting
3248 the length of names to 32 characters, and the number of named subpatterns
3249 to 10,000.
3250
3251 11. Subpatterns that are repeated with specific counts have to be replicated in
3252 the compiled pattern. The size of memory for this was computed from the
3253 length of the subpattern and the repeat count. The latter is limited to
3254 65535, but there was no limit on the former, meaning that integer overflow
3255 could in principle occur. The compiled length of a repeated subpattern is
3256 now limited to 30,000 bytes in order to prevent this.
3257
3258 12. Added the optional facility to have named substrings with the same name.
3259
3260 13. Added the ability to use a named substring as a condition, using the
3261 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3262 are numbers (not recommended). Forward references are permitted.
3263
3264 14. Added forward references in named backreferences (if you see what I mean).
3265
3266 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3267 pattern could run off the end of the subject. For example, the pattern
3268 "(?s)(.{1,5})"8 did this with the subject "ab".
3269
3270 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3271 PCRE_CASELESS was set when matching characters that were quantified with ?
3272 or *.
3273
3274 17. A character class other than a single negated character that had a minimum
3275 but no maximum quantifier - for example [ab]{6,} - was not handled
3276 correctly by pce_dfa_exec(). It would match only one character.
3277
3278 18. A valid (though odd) pattern that looked like a POSIX character
3279 class but used an invalid character after [ (for example [[,abc,]]) caused
3280 pcre_compile() to give the error "Failed: internal error: code overflow" or
3281 in some cases to crash with a glibc free() error. This could even happen if
3282 the pattern terminated after [[ but there just happened to be a sequence of
3283 letters, a binary zero, and a closing ] in the memory that followed.
3284
3285 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3286 over the years. Originally (before any Unicode support), just the bottom 8
3287 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3288 output from "man perlunicode" includes this:
3289
3290 The regular expression compiler produces polymorphic opcodes. That
3291 is, the pattern adapts to the data and automatically switches to
3292 the Unicode character scheme when presented with Unicode data--or
3293 instead uses a traditional byte scheme when presented with byte
3294 data.
3295
3296 Sadly, a wide octal escape does not cause a switch, and in a string with
3297 no other multibyte characters, these octal escapes are treated as before.
3298 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3299 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3300 Unicode string.
3301
3302 I have not perpetrated such confusion in PCRE. Up till now, it took just
3303 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3304 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3305 translate to the appropriate multibyte character.
3306
3307 29. Applied some refactoring to reduce the number of warnings from Microsoft
3308 and Borland compilers. This has included removing the fudge introduced
3309 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3310 a warning about an unused variable.
3311
3312 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3313 characters since release 4.0, because Perl (from release 5.004) does not.
3314 [Or at least, is documented not to: some releases seem to be in conflict
3315 with the documentation.] However, when a pattern was studied with
3316 pcre_study() and all its branches started with \s, PCRE still included VT
3317 as a possible starting character. Of course, this did no harm; it just
3318 caused an unnecessary match attempt.
3319
3320 22. Removed a now-redundant internal flag bit that recorded the fact that case
3321 dependency changed within the pattern. This was once needed for "required
3322 byte" processing, but is no longer used. This recovers a now-scarce options
3323 bit. Also moved the least significant internal flag bit to the most-
3324 significant bit of the word, which was not previously used (hangover from
3325 the days when it was an int rather than a uint) to free up another bit for
3326 the future.
3327
3328 23. Added support for CRLF line endings as well as CR and LF. As well as the
3329 default being selectable at build time, it can now be changed at runtime
3330 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3331 specify that it is scanning data with non-default line endings.
3332
3333 24. Changed the definition of CXXLINK to make it agree with the definition of
3334 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3335
3336 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3337 recursions. This makes a big different to stack usage for some patterns.
3338
3339 26. If a subpattern containing a named recursion or subroutine reference such
3340 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3341 the space required for the compiled pattern went wrong and gave too small a
3342 value. Depending on the environment, this could lead to "Failed: internal
3343 error: code overflow at offset 49" or "glibc detected double free or
3344 corruption" errors.
3345
3346 27. Applied patches from Google (a) to support the new newline modes and (b) to
3347 advance over multibyte UTF-8 characters in GlobalReplace.
3348
3349 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3350 difference for some implementation of PCRE in some Windows version.
3351
3352 29. Added some extra testing facilities to pcretest:
3353
3354 \q<number> in a data line sets the "match limit" value
3355 \Q<number> in a data line sets the "match recursion limt" value
3356 -S <number> sets the stack size, where <number> is in megabytes
3357
3358 The -S option isn't available for Windows.
3359
3360
3361 Version 6.6 06-Feb-06
3362 ---------------------
3363
3364 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3365 in pcreposix.h. I have copied the definition from pcre.h.
3366
3367 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3368 because pcre.h is no longer a built file.
3369
3370 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3371 not normally included in the compiled code.
3372
3373
3374 Version 6.5 01-Feb-06
3375 ---------------------
3376
3377 1. When using the partial match feature with pcre_dfa_exec(), it was not
3378 anchoring the second and subsequent partial matches at the new starting
3379 point. This could lead to incorrect results. For example, with the pattern
3380 /1234/, partially matching against "123" and then "a4" gave a match.
3381
3382 2. Changes to pcregrep:
3383
3384 (a) All non-match returns from pcre_exec() were being treated as failures
3385 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3386 error message is output. Some extra information is given for the
3387 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3388 probably the only errors that are likely to be caused by users (by
3389 specifying a regex that has nested indefinite repeats, for instance).
3390 If there are more than 20 of these errors, pcregrep is abandoned.
3391
3392 (b) A binary zero was treated as data while matching, but terminated the
3393 output line if it was written out. This has been fixed: binary zeroes
3394 are now no different to any other data bytes.
3395
3396 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3397 used to set a locale for matching. The --locale=xxxx long option has
3398 been added (no short equivalent) to specify a locale explicitly on the
3399 pcregrep command, overriding the environment variables.
3400
3401 (d) When -B was used with -n, some line numbers in the output were one less
3402 than they should have been.
3403
3404 (e) Added the -o (--only-matching) option.
3405
3406 (f) If -A or -C was used with -c (count only), some lines of context were
3407 accidentally printed for the final match.
3408
3409 (g) Added the -H (--with-filename) option.
3410
3411 (h) The combination of options -rh failed to suppress file names for files
3412 that were found from directory arguments.
3413
3414 (i) Added the -D (--devices) and -d (--directories) options.
3415
3416 (j) Added the -F (--fixed-strings) option.
3417
3418 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3419
3420 (l) Added the --colo(u)r option.
3421
3422 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3423 is not present by default.
3424
3425 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3426 items such as (?R) or (?1), when the recursion could match a number of
3427 alternatives. If it matched one of the alternatives, but subsequently,
3428 outside the recursion, there was a failure, the code tried to back up into
3429 the recursion. However, because of the way PCRE is implemented, this is not
3430 possible, and the result was an incorrect result from the match.
3431
3432 In order to prevent this happening, the specification of recursion has
3433 been changed so that all such subpatterns are automatically treated as
3434 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3435
3436 4. I had overlooked the fact that, in some locales, there are characters for
3437 which isalpha() is true but neither isupper() nor islower() are true. In
3438 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3439 and ordfeminine) are like this. This affected the treatment of \w and \W
3440 when they appeared in character classes, but not when they appeared outside
3441 a character class. The bit map for "word" characters is now created
3442 separately from the results of isalnum() instead of just taking it from the
3443 upper, lower, and digit maps. (Plus the underscore character, of course.)
3444
3445 5. The above bug also affected the handling of POSIX character classes such as
3446 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3447 permanent tables. Instead, the bit maps for such a class were previously
3448 created as the appropriate unions of the upper, lower, and digit bitmaps.
3449 Now they are created by subtraction from the [[:word:]] class, which has
3450 its own bitmap.
3451
3452 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3453 It is created by subtracting the vertical space characters (\x09, \x0a,
3454 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3455 subtraction was done in the overall bitmap for a character class, meaning
3456 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3457 be recognized. This bug has been fixed.
3458
3459 7. Patches from the folks at Google:
3460
3461 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3462 real life, but is still worth protecting against".
3463
3464 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3465 regular expressions".
3466
3467 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3468 have it.
3469
3470 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3471 "configure" and the latter not, in order to fix a problem somebody had
3472 with compiling the Arg class on HP-UX.
3473
3474 (e) Improve the error-handling of the C++ wrapper a little bit.
3475
3476 (f) New tests for checking recursion limiting.
3477
3478 8. The pcre_memmove() function, which is used only if the environment does not
3479 have a standard memmove() function (and is therefore rarely compiled),
3480 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3481 returning a result (though PCRE never actually uses the result).
3482
3483 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3484 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3485 returned instead of calling malloc() with an overflowing number that would
3486 most likely cause subsequent chaos.
3487
3488 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3489
3490 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3491 with this option is matched, the nmatch and pmatch options of regexec() are
3492 ignored.
3493
3494 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3495 provided in case anyone wants to the the POSIX interface with UTF-8
3496 strings.
3497
3498 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3499 C++ linking (needed for some HP-UX environments).
3500
3501 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3502 (unused parameter) and in the pcre_printint() function (omitted "default"
3503 switch label when the default is to do nothing).
3504
3505 15. Added some code to make it possible, when PCRE is compiled as a C++
3506 library, to replace subject pointers for pcre_exec() with a smart pointer
3507 class, thus making it possible to process discontinuous strings.
3508
3509 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3510 much the same function. They were added by different people who were trying
3511 to make PCRE easy to compile on non-Unix systems. It has been suggested
3512 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3513 for compiling on Windows systems. I have therefore replaced it with
3514 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3515 defaults to "extern" for C or "extern C" for C++, which works fine on
3516 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3517 SCOPE with something explicit in config.h. In addition:
3518
3519 (a) pcreposix.h still had just "extern" instead of either of these macros;
3520 I have replaced it with PCRE_DATA_SCOPE.
3521
3522 (b) Functions such as _pcre_xclass(), which are internal to the library,
3523 but external in the C sense, all had PCRE_EXPORT in their definitions.
3524 This is apparently wrong for the Windows case, so I have removed it.
3525 (It makes no difference on Unix-like systems.)
3526
3527 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3528 of recursive calls to match(). This is different to MATCH_LIMIT because
3529 that limits the total number of calls to match(), not all of which increase
3530 the depth of recursion. Limiting the recursion depth limits the amount of
3531 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3532 when PCRE is compiled, and changed at run time. A patch from Google adds
3533 this functionality to the C++ interface.
3534
3535 18. Changes to the handling of Unicode character properties:
3536
3537 (a) Updated the table to Unicode 4.1.0.
3538
3539 (b) Recognize characters that are not in the table as "Cn" (undefined).
3540
3541 (c) I revised the way the table is implemented to a much improved format
3542 which includes recognition of ranges. It now supports the ranges that
3543 are defined in UnicodeData.txt, and it also amalgamates other
3544 characters into ranges. This has reduced the number of entries in the
3545 table from around 16,000 to around 3,000, thus reducing its size
3546 considerably. I realized I did not need to use a tree structure after
3547 all - a binary chop search is just as efficient. Having reduced the
3548 number of entries, I extended their size from 6 bytes to 8 bytes to
3549 allow for more data.
3550
3551 (d) Added support for Unicode script names via properties such as \p{Han}.
3552
3553 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3554 matching that character.
3555
3556 20. When matching a repeated Unicode property with a minimum greater than zero,
3557 (for example \pL{2,}), PCRE could look past the end of the subject if it
3558 reached it while seeking the minimum number of characters. This could
3559 happen only if some of the characters were more than one byte long, because
3560 there is a check for at least the minimum number of bytes.
3561
3562 21. Refactored the implementation of \p and \P so as to be more general, to
3563 allow for more different types of property in future. This has changed the
3564 compiled form incompatibly. Anybody with saved compiled patterns that use
3565 \p or \P will have to recompile them.
3566
3567 22. Added "Any" and "L&" to the supported property types.
3568
3569 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3570 but give a compile time error if the value is greater than 0xff.
3571
3572 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3573 accidentally not being installed or uninstalled.
3574
3575 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3576 made were to insert the current release number. This seemed silly, because
3577 it made things harder for people building PCRE on systems that don't run
3578 "configure". I have turned pcre.h into a distributed file, no longer built
3579 by "configure", with the version identification directly included. There is
3580 no longer a pcre.h.in file.
3581
3582 However, this change necessitated a change to the pcre-config script as
3583 well. It is built from pcre-config.in, and one of the substitutions was the
3584 release number. I have updated configure.ac so that ./configure now finds
3585 the release number by grepping pcre.h.
3586
3587 26. Added the ability to run the tests under valgrind.
3588
3589
3590 Version 6.4 05-Sep-05
3591 ---------------------
3592
3593 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3594 "--" to be printed when multiple files were scanned, even when none of the
3595 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3596 consider it to be a bug, and have restored the previous behaviour.
3597
3598 2. A couple of code tidies to get rid of compiler warnings.
3599
3600 3. The pcretest program used to cheat by referring to symbols in the library
3601 whose names begin with _pcre_. These are internal symbols that are not
3602 really supposed to be visible externally, and in some environments it is
3603 possible to suppress them. The cheating is now confined to including
3604 certain files from the library's source, which is a bit cleaner.
3605
3606 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3607 file's purpose clearer.
3608
3609 5. Reorganized pcre_ucp_findchar().
3610
3611
3612 Version 6.3 15-Aug-05
3613 ---------------------
3614
3615 1. The file libpcre.pc.in did not have general read permission in the tarball.
3616
3617 2. There were some problems when building without C++ support:
3618
3619 (a) If C++ support was not built, "make install" and "make test" still
3620 tried to test it.
3621
3622 (b) There were problems when the value of CXX was explicitly set. Some
3623 changes have been made to try to fix these, and ...
3624
3625 (c) --disable-cpp can now be used to explicitly disable C++ support.
3626
3627 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3628 backslash in a target when C++ was disabled. This confuses some
3629 versions of "make", apparently. Using an intermediate variable solves
3630 this. (Same for CPP_LOBJ.)
3631
3632 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3633 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3634 necessary on certain architectures.
3635
3636 4. Added a setting of -export-symbols-regex to the link command to remove
3637 those symbols that are exported in the C sense, but actually are local
3638 within the library, and not documented. Their names all begin with
3639 "_pcre_". This is not a perfect job, because (a) we have to except some
3640 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3641 available (and never for static libraries). I have made a note to try to
3642 find a way round (a) in the future.
3643
3644
3645 Version 6.2 01-Aug-05
3646 ---------------------
3647
3648 1. There was no test for integer overflow of quantifier values. A construction
3649 such as {1111111111111111} would give undefined results. What is worse, if
3650 a minimum quantifier for a parenthesized subpattern overflowed and became
3651 negative, the calculation of the memory size went wrong. This could have
3652 led to memory overwriting.
3653
3654 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3655
3656 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3657 operating environments where this matters.
3658
3659 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3660 PCRE options from within the C++ wrapper.
3661
3662 5. Named capturing subpatterns were not being correctly counted when a pattern
3663 was compiled. This caused two problems: (a) If there were more than 100
3664 such subpatterns, the calculation of the memory needed for the whole
3665 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3666 back references of the form \12, where the number was greater than 9, were
3667 not recognized as back references, even though there were sufficient
3668 previous subpatterns.
3669
3670 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3671 versions of gcc, e.g. 2.95.4.
3672
3673
3674 Version 6.1 21-Jun-05
3675 ---------------------
3676
3677 1. There was one reference to the variable "posix" in pcretest.c that was not
3678 surrounded by "#if !defined NOPOSIX".
3679
3680 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3681 the cross-check on the old pcre_info() function, for the benefit of the
3682 cut-down version of PCRE that is currently imported into Exim.
3683
3684 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3685 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3686 patterns that start (?i) but I don't think that matters. The use of (?i) is
3687 just an example; this all applies to the other options as well.
3688
3689 4. Since libtool seems to echo the compile commands it is issuing, the output
3690 from "make" can be reduced a bit by putting "@" in front of each libtool
3691 compile command.
3692
3693 5. Patch from the folks at Google for configure.in to be a bit more thorough
3694 in checking for a suitable C++ installation before trying to compile the
3695 C++ stuff. This should fix a reported problem when a compiler was present,
3696 but no suitable headers.
3697
3698 6. The man pages all had just "PCRE" as their title. I have changed them to
3699 be the relevant file name. I have also arranged that these names are
3700 retained in the file doc/pcre.txt, which is a concatenation in text format
3701 of all the man pages except the little individual ones for each function.
3702
3703 7. The NON-UNIX-USE file had not been updated for the different set of source
3704 files that come with release 6. I also added a few comments about the C++
3705 wrapper.
3706
3707
3708 Version 6.0 07-Jun-05
3709 ---------------------
3710
3711 1. Some minor internal re-organization to help with my DFA experiments.
3712
3713 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3714 didn't matter for the library itself when fully configured, but did matter
3715 when compiling without UCP support, or within Exim, where the ucp files are
3716 not imported.
3717
3718 3. Refactoring of the library code to split up the various functions into
3719 different source modules. The addition of the new DFA matching code (see
3720 below) to a single monolithic source would have made it really too
3721 unwieldy, quite apart from causing all the code to be include in a
3722 statically linked application, when only some functions are used. This is
3723 relevant even without the DFA addition now that patterns can be compiled in
3724 one application and matched in another.
3725
3726 The downside of splitting up is that there have to be some external
3727 functions and data tables that are used internally in different modules of
3728 the library but which are not part of the API. These have all had their
3729 names changed to start with "_pcre_" so that they are unlikely to clash
3730 with other external names.
3731
3732 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3733 a different (DFA) algorithm. Although it is slower than the original
3734 function, it does have some advantages for certain types of matching
3735 problem.
3736
3737 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3738 including restarting after a partial match.
3739
3740 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3741 defined when compiling for Windows was sent to me. I have put it into the
3742 code, though I have no means of testing or verifying it.
3743
3744 7. Added the pcre_refcount() auxiliary function.
3745
3746 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3747 match before or at the first newline in the subject string. In pcretest,
3748 the /f option on a pattern can be used to set this.
3749
3750 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3751 would behave wrongly. This has been present in PCRE since release 4.0.
3752
3753 10. A number of changes to the pcregrep command:
3754
3755 (a) Refactored how -x works; insert ^(...)$ instead of setting
3756 PCRE_ANCHORED and checking the length, in preparation for adding
3757 something similar for -w.
3758
3759 (b) Added the -w (match as a word) option.
3760
3761 (c) Refactored the way lines are read and buffered so as to have more
3762 than one at a time available.
3763
3764 (d) Implemented a pcregrep test script.
3765
3766 (e) Added the -M (multiline match) option. This allows patterns to match
3767 over several lines of the subject. The buffering ensures that at least
3768 8K, or the rest of the document (whichever is the shorter) is available
3769 for matching (and similarly the previous 8K for lookbehind assertions).
3770
3771 (f) Changed the --help output so that it now says
3772
3773 -w, --word-regex(p)
3774
3775 instead of two lines, one with "regex" and the other with "regexp"
3776 because that confused at least one person since the short forms are the
3777 same. (This required a bit of code, as the output is generated
3778 automatically from a table. It wasn't just a text change.)
3779
3780 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3781 option but starts with a hyphen. Could be a pattern or a path name
3782 starting with a hyphen, for instance.
3783
3784 (h) "-" can be given as a file name to represent stdin.
3785
3786 (i) When file names are being printed, "(standard input)" is used for
3787 the standard input, for compatibility with GNU grep. Previously
3788 "<stdin>" was used.
3789
3790 (j) The option --label=xxx can be used to supply a name to be used for
3791 stdin when file names are being printed. There is no short form.
3792
3793 (k) Re-factored the options decoding logic because we are going to add
3794 two more options that take data. Such options can now be given in four
3795 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3796
3797 (l) Added the -A, -B, and -C options for requesting that lines of context
3798 around matches be printed.
3799
3800 (m) Added the -L option to print the names of files that do not contain
3801 any matching lines, that is, the complement of -l.
3802
3803 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3804 continue to scan other files.
3805
3806 (o) The -s option was incorrectly implemented. For compatibility with other
3807 greps, it now suppresses the error message for a non-existent or non-
3808 accessible file (but not the return code). There is a new option called
3809 -q that suppresses the output of matching lines, which was what -s was
3810 previously doing.
3811
3812 (p) Added --include and --exclude options to specify files for inclusion
3813 and exclusion when recursing.
3814
3815 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3816 Hopefully, it now does.
3817
3818 12. Missing cast in pcre_study().
3819
3820 13. Added an "uninstall" target to the makefile.
3821
3822 14. Replaced "extern" in the function prototypes in Makefile.in with
3823 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3824 world, but is set differently for Windows.
3825
3826 15. Added a second compiling function called pcre_compile2(). The only
3827 difference is that it has an extra argument, which is a pointer to an
3828 integer error code. When there is a compile-time failure, this is set
3829 non-zero, in addition to the error test pointer being set to point to an
3830 error message. The new argument may be NULL if no error number is required
3831 (but then you may as well call pcre_compile(), which is now just a
3832 wrapper). This facility is provided because some applications need a
3833 numeric error indication, but it has also enabled me to tidy up the way
3834 compile-time errors are handled in the POSIX wrapper.
3835
3836 16. Added VPATH=.libs to the makefile; this should help when building with one
3837 prefix path and installing with another. (Or so I'm told by someone who
3838 knows more about this stuff than I do.)
3839
3840 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3841 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3842 match everything, including newlines. This is not POSIX-compatible, but
3843 somebody wanted the feature. From pcretest it can be activated by using
3844 both the P and the s flags.
3845
3846 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3847
3848 19. libpcre.pc was being incorrectly installed as executable.
3849
3850 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3851 it now also looks for '\r' so that it will work unmodified on Windows.
3852
3853 21. Added Google's contributed C++ wrapper to the distribution.
3854
3855 22. Added some untidy missing memory free() calls in pcretest, to keep
3856 Electric Fence happy when testing.
3857
3858
3859
3860 Version 5.0 13-Sep-04
3861 ---------------------
3862
3863 1. Internal change: literal characters are no longer packed up into items
3864 containing multiple characters in a single byte-string. Each character
3865 is now matched using a separate opcode. However, there may be more than one
3866 byte in the character in UTF-8 mode.
3867
3868 2. The pcre_callout_block structure has two new fields: pattern_position and
3869 next_item_length. These contain the offset in the pattern to the next match
3870 item, and its length, respectively.
3871
3872 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3873 insertion of callouts before each pattern item. Added the /C option to
3874 pcretest to make use of this.
3875
3876 4. On the advice of a Windows user, the lines
3877
3878 #if defined(_WIN32) || defined(WIN32)
3879 _setmode( _fileno( stdout ), 0x8000 );
3880 #endif /* defined(_WIN32) || defined(WIN32) */
3881
3882 have been added to the source of pcretest. This apparently does useful
3883 magic in relation to line terminators.
3884
3885 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3886 for the benefit of those environments where the "b" makes a difference.
3887
3888 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3889 to know about it. I have put a hack into configure.in that adds in code
3890 to set GCC=yes if CC=icc. This seems to end up at a point in the
3891 generated configure script that is early enough to affect the setting of
3892 compiler options, which is what is needed, but I have no means of testing
3893 whether it really works. (The user who reported this had patched the
3894 generated configure script, which of course I cannot do.)
3895
3896 LATER: After change 22 below (new libtool files), the configure script
3897 seems to know about icc (and also ecc). Therefore, I have commented out
3898 this hack in configure.in.
3899
3900 7. Added support for pkg-config (2 patches were sent in).
3901
3902 8. Negated POSIX character classes that used a combination of internal tables
3903 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3904 [[:^ascii]]. Typically, they would match almost any characters. The other
3905 POSIX classes were not broken in this way.
3906
3907 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3908 to find the match, as PCRE was deluded into thinking that the match had to
3909 start at the start point or following a newline. The same bug applied to
3910 patterns with negative forward assertions or any backward assertions
3911 preceding ".*" at the start, unless the pattern required a fixed first
3912 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3913
3914 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3915 starting at the last subject character, bytes beyond the end of the subject
3916 string were read.
3917
3918 11. Renamed the variable "class" as "classbits" to make life easier for C++
3919 users. (Previously there was a macro definition, but it apparently wasn't
3920 enough.)
3921
3922 12. Added the new field "tables" to the extra data so that tables can be passed
3923 in at exec time, or the internal tables can be re-selected. This allows
3924 a compiled regex to be saved and re-used at a later time by a different
3925 program that might have everything at different addresses.
3926
3927 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3928 -R library as well as a -L library.
3929
3930 14. The debugging options of pcretest (-d on the command line or D on a
3931 pattern) showed incorrect output for anything following an extended class
3932 that contained multibyte characters and which was followed by a quantifier.
3933
3934 15. Added optional support for general category Unicode character properties
3935 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3936 support. It adds about 90K to the size of the library. The meanings of the
3937 inbuilt class escapes such as \d and \s have NOT been changed.
3938
3939 16. Updated pcredemo.c to include calls to free() to release the memory for the
3940 compiled pattern.
3941
3942 17. The generated file chartables.c was being created in the source directory
3943 instead of in the building directory. This caused the build to fail if the
3944 source directory was different from the building directory, and was
3945 read-only.
3946
3947 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3948 file. No doubt somebody will tell me if they don't make sense... Also added
3949 Dan Mooney's comments about building on OpenVMS.
3950
3951 19. Added support for partial matching via the PCRE_PARTIAL option for
3952 pcre_exec() and the \P data escape in pcretest.
3953
3954 20. Extended pcretest with 3 new pattern features:
3955
3956 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3957 write the compiled pattern to the file whose name is "rest-of-line".
3958 This is a straight binary dump of the data, with the saved pointer to
3959 the character tables forced to be NULL. The study data, if any, is
3960 written too. After writing, pcretest reads a new pattern.
3961
3962 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3963 compiled pattern from the given file. There must not be any
3964 occurrences of "<" in the file name (pretty unlikely); if there are,
3965 pcretest will instead treat the initial "<" as a pattern delimiter.
3966 After reading in the pattern, pcretest goes on to read data lines as
3967 usual.
3968
3969 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3970 and 16-bit fields in a compiled pattern, to simulate a pattern that
3971 was compiled on a host of opposite endianness.
3972
3973 21. The pcre-exec() function can now cope with patterns that were compiled on
3974 hosts of opposite endianness, with this restriction:
3975
3976 As for any compiled expression that is saved and used later, the tables
3977 pointer field cannot be preserved; the extra_data field in the arguments
3978 to pcre_exec() should be used to pass in a tables address if a value
3979 other than the default internal tables were used at compile time.
3980
3981 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3982 now diagnosed as an error. Previously, most of the time, a negative number
3983 would have been treated as zero, but if in addition "ovector" was passed as
3984 NULL, a crash could occur.
3985
3986 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3987 new versions from the libtool 1.5 distribution (the last one is a copy of
3988 a file called libtool.m4). This seems to have fixed the need to patch
3989 "configure" to support Darwin 1.3 (which I used to do). However, I still
3990 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3991 workstation).
3992
3993 24. Changed the PCRE licence to be the more standard "BSD" licence.
3994
3995
3996 Version 4.5 01-Dec-03
3997 ---------------------
3998
3999 1. There has been some re-arrangement of the code for the match() function so
4000 that it can be compiled in a version that does not call itself recursively.
4001 Instead, it keeps those local variables that need separate instances for
4002 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4003 needs to "recurse". Keeping track of where control must go is done by means
4004 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4005 hide most of the details from the main code, and operates only if
4006 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4007 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4008 operating.
4009
4010 To make it easier for callers to provide specially tailored get/free
4011 functions for this usage, two new functions, pcre_stack_malloc, and
4012 pcre_stack_free, are used. They are always called in strict stacking order,
4013 and the size of block requested is always the same.
4014
4015 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4016 PCRE has been compiled to use the stack or the heap for recursion. The
4017 -C option of pcretest uses this to show which version is compiled.
4018
4019 A new data escape \S, is added to pcretest; it causes the amounts of store
4020 obtained and freed by both kinds of malloc/free at match time to be added
4021 to the output.
4022
4023 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4024 what's available on my current Linux desktop machine.
4025
4026 3. When matching a UTF-8 string, the test for a valid string at the start has
4027 been extended. If start_offset is not zero, PCRE now checks that it points
4028 to a byte that is the start of a UTF-8 character. If not, it returns
4029 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4030 this is necessary because there may be backward assertions in the pattern.
4031 When matching the same subject several times, it may save resources to use
4032 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4033
4034 4. The code for checking the validity of UTF-8 strings has been tightened so
4035 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4036 containing "overlong sequences".
4037
4038 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4039 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4040 should have been "&", but it just so happened that all the cases this let
4041 through by mistake were picked up later in the function.
4042
4043 6. I had used a variable called "isblank" - this is a C99 function, causing
4044 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4045
4046 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4047 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4048 so the version is shown; (c) stop "make test" echoing "./RunTest".
4049
4050 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4051
4052 9. The prototype for memmove() for systems that don't have it was using
4053 size_t, but the inclusion of the header that defines size_t was later. I've
4054 moved the #includes for the C headers earlier to avoid this.
4055
4056 10. Added some adjustments to the code to make it easier to compiler on certain
4057 special systems:
4058
4059 (a) Some "const" qualifiers were missing.
4060 (b) Added the macro EXPORT before all exported functions; by default this
4061 is defined to be empty.
4062 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4063 that it reads its output file name as an argument instead of writing
4064 to the standard output and assuming this can be redirected.
4065
4066 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4067 class containing characters with values greater than 255, PCRE compilation
4068 went into a loop.
4069
4070 12. A recursive reference to a subpattern that was within another subpattern
4071 that had a minimum quantifier of zero caused PCRE to crash. For example,
4072 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4073 recursion. If the recursively-called subpattern itself had a zero repeat,
4074 that was OK.
4075
4076 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4077 buffer into which it was copied (for escape processing) was still set at
4078 1024, so long lines caused crashes.
4079
4080 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4081 "internal error: code overflow...". This applied to any character class
4082 that was followed by a possessive quantifier.
4083
4084 15. Modified the Makefile to add libpcre.la as a prerequisite for
4085 libpcreposix.la because I was told this is needed for a parallel build to
4086 work.
4087
4088 16. If a pattern that contained .* following optional items at the start was
4089 studied, the wrong optimizing data was generated, leading to matching
4090 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4091 matching string must start with a or b or c. The correct conclusion for
4092 this pattern is that a match can start with any character.
4093
4094
4095 Version 4.4 13-Aug-03
4096 ---------------------
4097
4098 1. In UTF-8 mode, a character class containing characters with values between
4099 127 and 255 was not handled correctly if the compiled pattern was studied.
4100 In fixing this, I have also improved the studying algorithm for such
4101 classes (slightly).
4102
4103 2. Three internal functions had redundant arguments passed to them. Removal
4104 might give a very teeny performance improvement.
4105
4106 3. Documentation bug: the value of the capture_top field in a callout is *one
4107 more than* the number of the hightest numbered captured substring.
4108
4109 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4110 in incorrectly linking with a previously installed version. They now link
4111 explicitly with libpcre.la.
4112
4113 5. configure.in no longer needs to recognize Cygwin specially.
4114
4115 6. A problem in pcre.in for Windows platforms is fixed.
4116
4117 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4118 pcretest, it used to include the size of the study block as part of its
4119 output. Unfortunately, the structure contains a field that has a different
4120 size on different hardware architectures. This meant that the tests that
4121 showed this size failed. As the block is currently always of a fixed size,
4122 this information isn't actually particularly useful in pcretest output, so
4123 I have just removed it.
4124
4125 8. Three pre-processor statements accidentally did not start in column 1.
4126 Sadly, there are *still* compilers around that complain, even though
4127 standard C has not required this for well over a decade. Sigh.
4128
4129 9. In pcretest, the code for checking callouts passed small integers in the
4130 callout_data field, which is a void * field. However, some picky compilers
4131 complained about the casts involved for this on 64-bit systems. Now
4132 pcretest passes the address of the small integer instead, which should get
4133 rid of the warnings.
4134
4135 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4136 both compile and run time, and gives an error if an invalid UTF-8 sequence
4137 is found. There is a option for disabling this check in cases where the
4138 string is known to be correct and/or the maximum performance is wanted.
4139
4140 11. In response to a bug report, I changed one line in Makefile.in from
4141
4142 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4143 to
4144 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4145
4146 to look similar to other lines, but I have no way of telling whether this
4147 is the right thing to do, as I do not use Windows. No doubt I'll get told
4148 if it's wrong...
4149
4150
4151 Version 4.3 21-May-03
4152 ---------------------
4153
4154 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4155 Makefile.
4156
4157 2. Some refactoring to improve the quality of the code:
4158
4159 (i) The utf8_table... variables are now declared "const".
4160
4161 (ii) The code for \cx, which used the "case flipping" table to upper case
4162 lower case letters, now just substracts 32. This is ASCII-specific,
4163 but the whole concept of \cx is ASCII-specific, so it seems
4164 reasonable.
4165
4166 (iii) PCRE was using its character types table to recognize decimal and
4167 hexadecimal digits in the pattern. This is silly, because it handles
4168 only 0-9, a-f, and A-F, but the character types table is locale-
4169 specific, which means strange things might happen. A private
4170 table is now used for this - though it costs 256 bytes, a table is
4171 much faster than multiple explicit tests. Of course, the standard
4172 character types table is still used for matching digits in subject
4173 strings against \d.
4174
4175 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4176 ending in _t are). So I've renamed it as ESC_tee.
4177
4178 3. The first argument for regexec() in the POSIX wrapper should have been
4179 defined as "const".
4180
4181 4. Changed pcretest to use malloc() for its buffers so that they can be
4182 Electric Fenced for debugging.
4183
4184 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4185 to read one or more bytes before the start of the subject string. Often this
4186 had no effect on PCRE's behaviour, but in some circumstances it could
4187 provoke a segmentation fault.
4188
4189 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4190 to try to read one or more bytes before the start of the subject string.
4191
4192 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4193 UTF-8 support could misbehave in various ways if the subject string
4194 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4195 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4196 back over UTF-8 characters.)
4197
4198
4199 Version 4.2 14-Apr-03
4200 ---------------------
4201
4202 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4203
4204 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4205 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4206 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4207 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4208 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4209 and BUILD_EXEEXT
4210 Note: automatic setting of the BUILD variables is not yet working
4211 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4212 compile-time but not at link-time
4213 [LINK]: use for linking executables only
4214 make different versions for Windows and non-Windows
4215 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4216 libraries
4217 [LINK_FOR_BUILD]: new variable
4218 [OBJEXT]: use throughout
4219 [EXEEXT]: use throughout
4220 <winshared>: new target
4221 <wininstall>: new target
4222 <dftables.o>: use native compiler
4223 <dftables>: use native linker
4224 <install>: handle Windows platform correctly
4225 <clean>: ditto
4226 <check>: ditto
4227 copy DLL to top builddir before testing
4228
4229 As part of these changes, -no-undefined was removed again. This was reported
4230 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4231 in any case.
4232
4233 3. Some tidies to get rid of compiler warnings:
4234
4235 . In the match_data structure, match_limit was an unsigned long int, whereas
4236 match_call_count was an int. I've made them both unsigned long ints.
4237
4238 . In pcretest the fact that a const uschar * doesn't automatically cast to
4239 a void * provoked a warning.
4240
4241 . Turning on some more compiler warnings threw up some "shadow" variables
4242 and a few more missing casts.
4243
4244 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4245 option, a class that contained a single character with a value between 128
4246 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4247
4248 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4249 option, a class that contained several characters, but with at least one
4250 whose value was between 128 and 255 caused PCRE to crash.
4251
4252
4253 Version 4.1 12-Mar-03
4254 ---------------------
4255
4256 1. Compiling with gcc -pedantic found a couple of places where casts were
4257 needed, and a string in dftables.c that was longer than standard compilers are
4258 required to support.
4259
4260 2. Compiling with Sun's compiler found a few more places where the code could
4261 be tidied up in order to avoid warnings.
4262
4263 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4264 first of these names is deprecated in the latest Autoconf in favour of the name
4265 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4266 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4267 analogy I have changed it to CFLAGS_FOR_BUILD.
4268
4269 4. Added -no-undefined to the linking command in the Makefile, because this is
4270 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4271 linking step for the pcreposix library.
4272
4273 5. PCRE was failing to diagnose the case of two named groups with the same
4274 name.
4275
4276 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4277 literal character that is needed in the subject for a match, and scans along to
4278 ensure that it is present before embarking on the full matching process. This
4279 saves time in cases of nested unlimited repeats that are never going to match.
4280 Problem: the scan can take a lot of time if the subject is very long (e.g.
4281 megabytes), thus penalizing straightforward matches. It is now done only if the
4282 amount of subject to be scanned is less than 1000 bytes.
4283
4284 7. A lesser problem with the same optimization is that it was recording the
4285 first character of an anchored pattern as "needed", thus provoking a search
4286 right along the subject, even when the first match of the pattern was going to
4287 fail. The "needed" character is now not set for anchored patterns, unless it
4288 follows something in the pattern that is of non-fixed length. Thus, it still
4289 fulfils its original purpose of finding quick non-matches in cases of nested
4290 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4291
4292
4293 Version 4.0 17-Feb-03
4294 ---------------------
4295
4296 1. If a comment in an extended regex that started immediately after a meta-item
4297 extended to the end of string, PCRE compiled incorrect data. This could lead to
4298 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4299
4300 2. Moved to autoconf 2.53 and libtool 1.4.2.
4301
4302 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4303 the special perltest8 script is no longer needed - all the tests can be run
4304 from a single perltest script.
4305
4306 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4307 by \s. It has now been removed in PCRE. This means it isn't recognized as
4308 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4309 class [:space:] *does* include VT, thereby creating a mess.
4310
4311 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4312 space and tab.
4313
4314 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4315 its new features into the main test script, reducing the number of scripts.
4316
4317 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4318 were backward compatible, and made the (?i) apply to the whole pattern, as if
4319 /i were given. Now it behaves more logically, and applies the option setting
4320 only to what follows. PCRE has been changed to follow suit. However, if it
4321 finds options settings right at the start of the pattern, it extracts them into
4322 the global options, as before. Thus, they show up in the info data.
4323
4324 8. Added support for the \Q...\E escape sequence. Characters in between are
4325 treated as literals. This is slightly different from Perl in that $ and @ are
4326 also handled as literals inside the quotes. In Perl, they will cause variable
4327 interpolation. Note the following examples:
4328
4329 Pattern PCRE matches Perl matches
4330
4331 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4332 \Qabc\$xyz\E abc\$xyz abc\$xyz
4333 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4334
4335 For compatibility with Perl, \Q...\E sequences are recognized inside character
4336 classes as well as outside them.
4337
4338 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4339 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4340 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4341 signed/unsigned warnings.
4342
4343 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4344 option for pcretest, so I've replaced it by a simple function that does just
4345 that job.
4346
4347 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4348 "pcregrep -".
4349
4350 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4351 Java package. This provides some syntactic sugar for simple cases of what my
4352 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4353 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4354 item, you can use this simplified notation. Note that only makes sense with
4355 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4356 greediness, whatever the setting of the PCRE_UNGREEDY option.
4357
4358 13. A change of greediness default within a pattern was not taking effect at
4359 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4360 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4361 was abstracted outside.
4362
4363 14. PCRE now supports the \G assertion. It is true when the current matching
4364 position is at the start point of the match. This differs from \A when the
4365 starting offset is non-zero. Used with the /g option of pcretest (or similar
4366 code), it works in the same way as it does for Perl's /g option. If all
4367 alternatives of a regex begin with \G, the expression is anchored to the start
4368 match position, and the "anchored" flag is set in the compiled expression.
4369
4370 15. Some bugs concerning the handling of certain option changes within patterns
4371 have been fixed. These applied to options other than (?ims). For example,
4372 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4373 been the other way round. Some of this was related to change 7 above.
4374
4375 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4376 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4377 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4378 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4379
4380 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4381 mode. Unlike ".", it always matches newline, whatever the setting of
4382 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4383 assertions. Perl allows it, but it doesn't (in general) work because it can't
4384 calculate the length of the lookbehind. At least, that's the case for Perl
4385 5.8.0 - I've been told they are going to document that it doesn't work in
4386 future.
4387
4388 18. Added an error diagnosis for escapes that PCRE does not support: these are
4389 \L, \l, \N, \P, \p, \U, \u, and \X.
4390
4391 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4392 reading past the end of the pattern in cases such as /[abcd/.
4393
4394 20. PCRE was getting more memory than necessary for patterns with classes that
4395 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4396
4397 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4398 compiling PCRE for use with Virtual Pascal.
4399
4400 22. Small fix to the Makefile to make it work properly if the build is done
4401 outside the source tree.
4402
4403 23. Added a new extension: a condition to go with recursion. If a conditional
4404 subpattern starts with (?(R) the "true" branch is used if recursion has
4405 happened, whereas the "false" branch is used only at the top level.
4406
4407 24. When there was a very long string of literal characters (over 255 bytes
4408 without UTF support, over 250 bytes with UTF support), the computation of how
4409 much memory was required could be incorrect, leading to segfaults or other
4410 strange effects.
4411
4412 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4413 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4414 there was a subsequent back reference to those brackets. This meant that, for
4415 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4416 possible to check for precisely this case. All we can do is abandon the
4417 optimization if .* occurs inside capturing brackets when there are any back
4418 references whatsoever. (See below for a better fix that came later.)
4419
4420 26. The handling of the optimization for finding the first character of a
4421 non-anchored pattern, and for finding a character that is required later in the
4422 match were failing in some cases. This didn't break the matching; it just
4423 failed to optimize when it could. The way this is done has been re-implemented.
4424
4425 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4426
4427 28. Added a new feature that provides some of the functionality that Perl
4428 provides with (?{...}). The facility is termed a "callout". The way it is done
4429 in PCRE is for the caller to provide an optional function, by setting
4430 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4431 global variable. By default it is unset, which disables all calling out. To get
4432 the function called, the regex must include (?C) at appropriate points. This
4433 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4434 This provides a means of identifying different callout points. When PCRE
4435 reaches such a point in the regex, if pcre_callout has been set, the external
4436 function is called. It is provided with data in a structure called
4437 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4438 matching continues; if it returns a non-zero value, the match at the current
4439 point fails. However, backtracking will occur if possible. [This was changed
4440 later and other features added - see item 49 below.]
4441
4442 29. pcretest is upgraded to test the callout functionality. It provides a
4443 callout function that displays information. By default, it shows the start of
4444 the match and the current position in the text. There are some new data escapes
4445 to vary what happens:
4446
4447 \C+ in addition, show current contents of captured substrings
4448 \C- do not supply a callout function
4449 \C!n return 1 when callout number n is reached
4450 \C!n!m return 1 when callout number n is reached for the mth time
4451
4452 30. If pcregrep was called with the -l option and just a single file name, it
4453 output "<stdin>" if a match was found, instead of the file name.
4454
4455 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4456 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4457 pcre_exec(). This saves a malloc/free per call. The default value of
4458 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4459 when configuring.
4460
4461 32. The default maximum size of a compiled pattern is 64K. There have been a
4462 few cases of people hitting this limit. The code now uses macros to handle the
4463 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4464 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4465 configuring. Tests 2 and 5 work only with 2-byte links because they output
4466 debugging information about compiled patterns.
4467
4468 33. Internal code re-arrangements:
4469
4470 (a) Moved the debugging function for printing out a compiled regex into
4471 its own source file (printint.c) and used #include to pull it into
4472 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4473 separate copies.
4474
4475 (b) Defined the list of op-code names for debugging as a macro in
4476 internal.h so that it is next to the definition of the opcodes.
4477
4478 (c) Defined a table of op-code lengths for simpler skipping along compiled
4479 code. This is again a macro in internal.h so that it is next to the
4480 definition of the opcodes.
4481
4482 34. Added support for recursive calls to individual subpatterns, along the
4483 lines of Robin Houston's patch (but implemented somewhat differently).
4484
4485 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4486 allow it to read and process whole directories in Win32. This code was
4487 contributed by Lionel Fourquaux; it has not been tested by me.
4488
4489 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4490 used to name a group. Names consist of alphanumerics and underscores, and must
4491 be unique. Back references use the syntax (?P=name) and recursive calls use
4492 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4493 numbers. The function pcre_fullinfo() can be used after compilation to extract
4494 a name/number map. There are three relevant calls:
4495
4496 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4497 PCRE_INFO_NAMECOUNT yields the number of entries
4498 PCRE_INFO_NAMETABLE yields a pointer to the map.
4499
4500 The map is a vector of fixed-size entries. The size of each entry depends on
4501 the length of the longest name used. The first two bytes of each entry are the
4502 group number, most significant byte first. There follows the corresponding
4503 name, zero terminated. The names are in alphabetical order.
4504
4505 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4506 case instead of 255. Making it the same both with and without UTF-8 support
4507 means that the same test output works with both.
4508
4509 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4510 calling malloc() with a zero argument.
4511
4512 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4513 optimization. I've improved things by keeping a bitmap of backreferences with
4514 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4515 fact referenced, the optimization can be applied. It is unlikely that a
4516 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4517 the match to follow \n) will appear inside brackets with a number greater than
4518 31, but if it does, any back reference > 31 suppresses the optimization.
4519
4520 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4521 of disabling numbered capturing parentheses. Any opening parenthesis that is
4522 not followed by ? behaves as if it were followed by ?: but named parentheses
4523 can still be used for capturing (and they will acquire numbers in the usual
4524 way).
4525
4526 41. Redesigned the return codes from the match() function into yes/no/error so
4527 that errors can be passed back from deep inside the nested calls. A malloc
4528 failure while inside a recursive subpattern call now causes the
4529 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4530
4531 42. It is now possible to set a limit on the number of times the match()
4532 function is called in a call to pcre_exec(). This facility makes it possible to
4533 limit the amount of recursion and backtracking, though not in a directly
4534 obvious way, because the match() function is used in a number of different
4535 circumstances. The count starts from zero for each position in the subject
4536 string (for non-anchored patterns). The default limit is, for compatibility, a
4537 large number, namely 10 000 000. You can change this in two ways:
4538
4539 (a) When configuring PCRE before making, you can use --with-match-limit=n
4540 to set a default value for the compiled library.
4541
4542 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4543 a different value is set. See 45 below.
4544
4545 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4546
4547 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4548 of things that can be changed at compile time. The first argument specifies
4549 what is wanted and the second points to where the information is to be placed.
4550 The current list of available information is:
4551
4552 PCRE_CONFIG_UTF8
4553
4554 The output is an integer that is set to one if UTF-8 support is available;
4555 otherwise it is set to zero.
4556
4557 PCRE_CONFIG_NEWLINE
4558
4559 The output is an integer that it set to the value of the code that is used for
4560 newline. It is either LF (10) or CR (13).
4561
4562 PCRE_CONFIG_LINK_SIZE
4563
4564 The output is an integer that contains the number of bytes used for internal
4565 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4566
4567 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4568
4569 The output is an integer that contains the threshold above which the POSIX
4570 interface uses malloc() for output vectors. See item 31 above.
4571
4572 PCRE_CONFIG_MATCH_LIMIT
4573
4574 The output is an unsigned integer that contains the default limit of the number
4575 of match() calls in a pcre_exec() execution. See 42 above.
4576
4577 44. pcretest has been upgraded by the addition of the -C option. This causes it
4578 to extract all the available output from the new pcre_config() function, and to
4579 output it. The program then exits immediately.
4580
4581 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4582 order to support additional features. One way would have been to define
4583 pcre_exec2() (for example) with extra arguments, but this would not have been
4584 extensible, and would also have required all calls to the original function to
4585 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4586 is used for passing in "extra" data from pcre_study().
4587
4588 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4589 contains the following fields:
4590
4591 flags a bitmap indicating which of the following fields are set
4592 study_data opaque data from pcre_study()
4593 match_limit a way of specifying a limit on match() calls for a specific
4594 call to pcre_exec()
4595 callout_data data for callouts (see 49 below)
4596
4597 The flag bits are also defined in pcre.h, and are
4598
4599 PCRE_EXTRA_STUDY_DATA
4600 PCRE_EXTRA_MATCH_LIMIT
4601 PCRE_EXTRA_CALLOUT_DATA
4602
4603 The pcre_study() function now returns one of these new pcre_extra blocks, with
4604 the actual study data pointed to by the study_data field, and the
4605 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4606 before. That is, this change is entirely upwards-compatible and requires no
4607 change to existing code.
4608
4609 If you want to pass in additional data to pcre_exec(), you can either place it
4610 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4611 block.
4612
4613 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4614 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4615 times with different match limits, until it finds the minimum value needed for
4616 pcre_exec() to complete. The value is then output. This can be instructive; for
4617 most simple matches the number is quite small, but for pathological cases it
4618 gets very large very quickly.
4619
4620 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4621 returns the size of the data block pointed to by the study_data field in a
4622 pcre_extra block, that is, the value that was passed as the argument to
4623 pcre_malloc() when PCRE was getting memory in which to place the information
4624 created by pcre_study(). The fourth argument should point to a size_t variable.
4625 pcretest has been extended so that this information is shown after a successful
4626 pcre_study() call when information about the compiled regex is being displayed.
4627
4628 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4629 because what follows is always an absolute path. (Later: it turns out that this
4630 is more than cosmetic for MinGW, because it doesn't like empty path
4631 components.)
4632
4633 49. Some changes have been made to the callout feature (see 28 above):
4634
4635 (i) A callout function now has three choices for what it returns:
4636
4637 0 => success, carry on matching
4638 > 0 => failure at this point, but backtrack if possible
4639 < 0 => serious error, return this value from pcre_exec()
4640
4641 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4642 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4643 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4644 use by callout functions. It will never be used by PCRE itself.
4645
4646 (ii) The pcre_extra structure (see 45 above) has a void * field called
4647 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4648 pcre_callout_block structure has a field of the same name. The contents of
4649 the field passed in the pcre_extra structure are passed to the callout
4650 function in the corresponding field in the callout block. This makes it
4651 easier to use the same callout-containing regex from multiple threads. For
4652 testing, the pcretest program has a new data escape
4653
4654 \C*n pass the number n (may be negative) as callout_data
4655
4656 If the callout function in pcretest receives a non-zero value as
4657 callout_data, it returns that value.
4658
4659 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4660 there were some redundant $(CFLAGS) in commands that are now specified as
4661 $(LINK), which already includes $(CFLAGS).
4662
4663 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4664 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4665 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4666 one-byte characters throughout. Note that case-insensitive matching applies
4667 only to characters whose values are less than 256. PCRE doesn't support the
4668 notion of cases for higher-valued characters.
4669
4670 (i) A character class whose characters are all within 0-255 is handled as
4671 a bit map, and the map is inverted for negative classes. Previously, a
4672 character > 255 always failed to match such a class; however it should
4673 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4674
4675 (ii) A negated character class with a single character < 255 is coded as
4676 "not this character" (OP_NOT). This wasn't working properly when the test
4677 character was multibyte, either singly or repeated.
4678
4679 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4680 mode, for example: \x{100}{2,3}.
4681
4682 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4683 singly or repeated) now correctly test multibyte characters. However,
4684 PCRE doesn't recognize any characters with values greater than 255 as
4685 digits, spaces, or word characters. Such characters always match \D, \S,
4686 and \W, and never match \d, \s, or \w.
4687
4688 (v) Classes may now contain characters and character ranges with values
4689 greater than 255. For example: [ab\x{100}-\x{400}].
4690
4691 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4692 PCRE in UTF-8 mode.
4693
4694 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4695 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4696 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4697 value.)
4698
4699 53. The single man page has become too large. I have therefore split it up into
4700 a number of separate man pages. These also give rise to individual HTML pages;
4701 these are now put in a separate directory, and there is an index.html page that
4702 lists them all. Some hyperlinking between the pages has been installed.
4703
4704 54. Added convenience functions for handling named capturing parentheses.
4705
4706 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4707 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4708 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4709 are faulted.
4710
4711 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4712 calling configure. These values are used when compiling the dftables.c program
4713 which is run to generate the source of the default character tables. They
4714 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4715 you will need to set these values.
4716
4717 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4718
4719
4720 Version 3.9 02-Jan-02
4721 ---------------------
4722
4723 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4724
4725 2. If --disable-static was given, the building process failed when trying to
4726 build pcretest and pcregrep. (For some reason it was using libtool to compile
4727 them, which is not right, as they aren't part of the library.)
4728
4729
4730 Version 3.8 18-Dec-01
4731 ---------------------
4732
4733 1. The experimental UTF-8 code was completely screwed up. It was packing the
4734 bytes in the wrong order. How dumb can you get?
4735
4736
4737 Version 3.7 29-Oct-01
4738 ---------------------
4739
4740 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4741 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4742 this didn't happen under Solaris 8, where I normally test things.
4743
4744 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4745 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4746 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4747 this fix an hour or so after the initial 3.7 release.)
4748
4749
4750 Version 3.6 23-Oct-01
4751 ---------------------
4752
4753 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4754 offsets passed as NULL with zero offset count.
4755
4756 2. The config.guess and config.sub files had not been updated when I moved to
4757 the latest autoconf.
4758
4759
4760 Version 3.5 15-Aug-01
4761 ---------------------
4762
4763 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4764 had been forgotten.
4765
4766 2. By using declared but undefined structures, we can avoid using "void"
4767 definitions in pcre.h while keeping the internal definitions of the structures
4768 private.
4769
4770 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4771 user point of view, this means that both static and shared libraries are built
4772 by default, but this can be individually controlled. More of the work of
4773 handling this static/shared cases is now inside libtool instead of PCRE's make
4774 file.
4775
4776 4. The pcretest utility is now installed along with pcregrep because it is
4777 useful for users (to test regexs) and by doing this, it automatically gets
4778 relinked by libtool. The documentation has been turned into a man page, so
4779 there are now .1, .txt, and .html versions in /doc.
4780
4781 5. Upgrades to pcregrep:
4782 (i) Added long-form option names like gnu grep.
4783 (ii) Added --help to list all options with an explanatory phrase.
4784 (iii) Added -r, --recursive to recurse into sub-directories.
4785 (iv) Added -f, --file to read patterns from a file.
4786
4787 6. pcre_exec() was referring to its "code" argument before testing that
4788 argument for NULL (and giving an error if it was NULL).
4789
4790 7. Upgraded Makefile.in to allow for compiling in a different directory from
4791 the source directory.
4792
4793 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4794 options bits, the pointer it was passed was to an int instead of to an unsigned
4795 long int. This mattered only on 64-bit systems.
4796
4797 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4798 generated) instead of pcre.in, which it its source. Also made the same change
4799 in several of the .c files.
4800
4801 10. A new release of gcc defines printf() as a macro, which broke pcretest
4802 because it had an ifdef in the middle of a string argument for printf(). Fixed
4803 by using separate calls to printf().
4804
4805 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4806 script, to force use of CR or LF instead of \n in the source. On non-Unix
4807 systems, the value can be set in config.h.
4808
4809 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4810 absolute limit. Changed the text of the error message to make this clear, and
4811 likewise updated the man page.
4812
4813 13. The limit of 99 on the number of capturing subpatterns has been removed.
4814 The new limit is 65535, which I hope will not be a "real" limit.
4815
4816
4817 Version 3.4 22-Aug-00
4818 ---------------------
4819
4820 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4821
4822 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4823
4824
4825 Version 3.3 01-Aug-00
4826 ---------------------
4827
4828 1. If an octal character was given, but the value was greater than \377, it
4829 was not getting masked to the least significant bits, as documented. This could
4830 lead to crashes in some systems.
4831
4832 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4833 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4834
4835 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4836 These just pass their arguments on to (pcre_free)(), but they are provided
4837 because some uses of PCRE bind it to non-C systems that can call its functions,
4838 but cannot call free() or pcre_free() directly.
4839
4840 4. Add "make test" as a synonym for "make check". Corrected some comments in
4841 the Makefile.
4842
4843 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4844 Makefile.
4845
4846 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4847 command called pgrep for grepping around the active processes.
4848
4849 7. Added the beginnings of support for UTF-8 character strings.
4850
4851 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4852 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4853 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4854 out for the ar command.)
4855
4856
4857 Version 3.2 12-May-00
4858 ---------------------
4859
4860 This is purely a bug fixing release.
4861
4862 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4863 of ZA. This was just one example of several cases that could provoke this bug,
4864 which was introduced by change 9 of version 2.00. The code for breaking
4865 infinite loops after an iteration that matches an empty string was't working
4866 correctly.
4867
4868 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4869 when matched against abbab (for example). After matching an empty string, it
4870 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4871 caused it to match further down the string than it should.
4872
4873 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4874 was there because it doesn't seem to be needed, and it causes trouble on some
4875 systems, as it is not a Standard C header. It has been removed.
4876
4877 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4878 were reported on the Macintosh. The changes were from
4879
4880 while ((c = *(++ptr)) != 0 && c != '\n');
4881 to
4882 while ((c = *(++ptr)) != 0 && c != '\n') ;
4883
4884 Totally extraordinary, but if that's what it takes...
4885
4886 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4887 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4888 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4889 assumes the way PCRE uses memmove() (always moving upwards).
4890
4891 6. PCRE is being used in one environment where strchr() is not available. There
4892 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4893 faster code anyway.
4894
4895
4896 Version 3.1 09-Feb-00
4897 ---------------------
4898
4899 The only change in this release is the fixing of some bugs in Makefile.in for
4900 the "install" target:
4901
4902 (1) It was failing to install pcreposix.h.
4903
4904 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4905
4906
4907 Version 3.0 01-Feb-00
4908 ---------------------
4909
4910 1. Add support for the /+ modifier to perltest (to output $` like it does in
4911 pcretest).
4912
4913 2. Add support for the /g modifier to perltest.
4914
4915 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4916 matches null strings.
4917
4918 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4919 pattern. Perl treats empty patterns specially - it reuses the most recent
4920 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4921 effect.
4922
4923 5. The POSIX interface was broken in that it was just handing over the POSIX
4924 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4925 required a bigger vector, with some working space on the end. This means that
4926 the POSIX wrapper now has to get and free some memory, and copy the results.
4927
4928 6. Added some simple autoconf support, placing the test data and the
4929 documentation in separate directories, re-organizing some of the
4930 information files, and making it build pcre-config (a GNU standard). Also added
4931 libtool support for building PCRE as a shared library, which is now the
4932 default.
4933
4934 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4935 09 are not valid octal constants. Single digits will be used for minor values
4936 less than 10.
4937
4938 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4939 existing programs that set these in the POSIX interface can use PCRE without
4940 modification.
4941
4942 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4943 return all that pcre_info() returns, plus additional data. The pcre_info()
4944 function is retained for compatibility, but is considered to be obsolete.
4945
4946 10. Added experimental recursion feature (?R) to handle one common case that
4947 Perl 5.6 will be able to do with (?p{...}).
4948
4949 11. Added support for POSIX character classes like [:alpha:], which Perl is
4950 adopting.
4951
4952
4953 Version 2.08 31-Aug-99
4954 ----------------------
4955
4956 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4957 trying to match at the startoffset position, but instead was moving forward to
4958 the next newline as if a previous match had failed.
4959
4960 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4961 and could get into a loop if a null string was matched other than at the start
4962 of the subject.
4963
4964 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4965 be distinguished at compile time, and for completeness also added PCRE_DATE.
4966
4967 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4968 in GnuWin32 environments.
4969
4970
4971 Version 2.07 29-Jul-99
4972 ----------------------
4973
4974 1. The documentation is now supplied in plain text form and HTML as well as in
4975 the form of man page sources.
4976
4977 2. C++ compilers don't like assigning (void *) values to other pointer types.
4978 In particular this affects malloc(). Although there is no problem in Standard
4979 C, I've put in casts to keep C++ compilers happy.
4980
4981 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4982 should be (const char *).
4983
4984 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4985 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4986 However, I haven't made this a standard facility. The documentation doesn't
4987 mention it, and the Makefile doesn't support it.
4988
4989 5. The Makefile now contains an "install" target, with editable destinations at
4990 the top of the file. The pcretest program is not installed.
4991
4992 6. pgrep -V now gives the PCRE version number and date.
4993
4994 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4995 causing the entire string to be ignored, instead of just the last character.
4996
4997 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4998 non-matching string, it can take a very, very long time, even for strings of
4999 quite modest length, because of the nested recursion. PCRE now does better in
5000 some of these cases. It does this by remembering the last required literal
5001 character in the pattern, and pre-searching the subject to ensure it is present
5002 before running the real match. In other words, it applies a heuristic to detect
5003 some types of certain failure quickly, and in the above example, if presented
5004 with a string that has no trailing " it gives "no match" very quickly.
5005
5006 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5007 other alternatives are tried instead.
5008
5009
5010 Version 2.06 09-Jun-99
5011 ----------------------
5012
5013 1. Change pcretest's output for amount of store used to show just the code
5014 space, because the remainder (the data block) varies in size between 32-bit and
5015 64-bit systems.
5016
5017 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5018 start matching at. This allows lookbehinds to work when searching for multiple
5019 occurrences in a string.
5020
5021 3. Added additional options to pcretest for testing multiple occurrences:
5022
5023 /+ outputs the rest of the string that follows a match
5024 /g loops for multiple occurrences, using the new startoffset argument
5025 /G loops for multiple occurrences by passing an incremented pointer
5026
5027 4. PCRE wasn't doing the "first character" optimization for patterns starting
5028 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5029 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5030 the letter 'x'. On long subject strings, this gives a significant speed-up.
5031
5032
5033 Version 2.05 21-Apr-99
5034 ----------------------
5035
5036 1. Changed the type of magic_number from int to long int so that it works
5037 properly on 16-bit systems.
5038
5039 2. Fixed a bug which caused patterns starting with .* not to work correctly
5040 when the subject string contained newline characters. PCRE was assuming
5041 anchoring for such patterns in all cases, which is not correct because .* will
5042 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5043 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5044 must be retried after every newline in the subject.
5045
5046
5047 Version 2.04 18-Feb-99
5048 ----------------------
5049
5050 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5051 computation of the store needed to hold the pattern was incorrect (too large).
5052 If such patterns were nested a few deep, this could multiply and become a real
5053 problem.
5054
5055 2. Added /M option to pcretest to show the memory requirement of a specific
5056 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5057
5058 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5059 compiled in such a way that the backtracking after subsequent failure was
5060 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5061 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5062
5063
5064 Version 2.03 02-Feb-99
5065 ----------------------
5066
5067 1. Fixed typo and small mistake in man page.
5068
5069 2. Added 4th condition (GPL supersedes if conflict) and created separate
5070 LICENCE file containing the conditions.
5071
5072 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5073 Perl, that is the internal \ allows the delimiter to be included in the
5074 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5075 the final delimiter, add \ to the end of the pattern (to test the error).
5076
5077 4. Added the convenience functions for extracting substrings after a successful
5078 match. Updated pcretest to make it able to test these functions.
5079
5080
5081 Version 2.02 14-Jan-99
5082 ----------------------
5083
5084 1. Initialized the working variables associated with each extraction so that
5085 their saving and restoring doesn't refer to uninitialized store.
5086
5087 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5088 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5089 fix the problem.
5090
5091 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5092 calls, and wasn't printing the correct value for compiling calls. Increased the
5093 default value of LOOPREPEAT, and the number of significant figures in the
5094 times.
5095
5096 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5097
5098 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5099 a building problem on Windows NT with a FAT file system.
5100
5101
5102 Version 2.01 21-Oct-98
5103 ----------------------
5104
5105 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5106 to character tables built by pcre_maketables() in the current locale. If NULL
5107 is passed, the default tables are used.
5108
5109
5110 Version 2.00 24-Sep-98
5111 ----------------------
5112
5113 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5114 it any more.
5115
5116 2. Allow quantification of (?>) groups, and make it work correctly.
5117
5118 3. The first character computation wasn't working for (?>) groups.
5119
5120 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5121 end of the subject) and add 5.005's \z, which really does match only at the
5122 very end of the subject.
5123
5124 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5125
5126 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5127 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5128 localized options. All options to pcre_study() were also removed.
5129
5130 7. Add other new features from 5.005:
5131
5132 $(?<= positive lookbehind
5133 $(?<! negative lookbehind
5134 (?imsx-imsx) added the unsetting capability
5135 such a setting is global if at outer level; local otherwise
5136 (?imsx-imsx:) non-capturing groups with option setting
5137 (?(cond)re|re) conditional pattern matching
5138
5139 A backreference to itself in a repeated group matches the previous
5140 captured string.
5141
5142 8. General tidying up of studying (both automatic and via "study")
5143 consequential on the addition of new assertions.
5144
5145 9. As in 5.005, unlimited repeated groups that could match an empty substring
5146 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5147 runtime if any iteration does actually match an empty substring.
5148
5149 10. Include the RunTest script in the distribution.
5150
5151 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5152 discrepancies, some of which were old and were also with respect to 5.004. They
5153 have now been fixed.
5154
5155
5156 Version 1.09 28-Apr-98
5157 ----------------------
5158
5159 1. A negated single character class followed by a quantifier with a minimum
5160 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5161 program crashes, or just wrong answers. This did not apply to negated classes
5162 containing more than one character, or to minima other than one.
5163
5164
5165 Version 1.08 27-Mar-98
5166 ----------------------
5167
5168 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5169
5170 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5171 latter must appear before anything that relies on it in the pattern.
5172
5173
5174 Version 1.07 16-Feb-98
5175 ----------------------
5176
5177 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5178 repeat of a potentially empty string).
5179
5180
5181 Version 1.06 23-Jan-98
5182 ----------------------
5183
5184 1. Added Markus Oberhumer's little patches for C++.
5185
5186 2. Literal strings longer than 255 characters were broken.
5187
5188
5189 Version 1.05 23-Dec-97
5190 ----------------------
5191
5192 1. Negated character classes containing more than one character were failing if
5193 PCRE_CASELESS was set at run time.
5194
5195
5196 Version 1.04 19-Dec-97
5197 ----------------------
5198
5199 1. Corrected the man page, where some "const" qualifiers had been omitted.
5200
5201 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5202 input syntax.
5203
5204 3. Fixed memory leak which occurred when a regex with back references was
5205 matched with an offsets vector that wasn't big enough. The temporary memory
5206 that is used in this case wasn't being freed if the match failed.
5207
5208 4. Tidied pcretest to ensure it frees memory that it gets.
5209
5210 5. Temporary memory was being obtained in the case where the passed offsets
5211 vector was exactly big enough.
5212
5213 6. Corrected definition of offsetof() from change 5 below.
5214
5215 7. I had screwed up change 6 below and broken the rules for the use of
5216 setjmp(). Now fixed.
5217
5218
5219 Version 1.03 18-Dec-97
5220 ----------------------
5221
5222 1. A erroneous regex with a missing opening parenthesis was correctly
5223 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5224 on some systems.
5225
5226 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5227 it was reported that one broken compiler failed on the former because "code" is
5228 also an independent variable.
5229
5230 3. The erroneous regex a[]b caused an array overrun reference.
5231
5232 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5233 fail on data ending with that character. (It was going on too far, and checking
5234 the next character, typically a binary zero.) This was specific to the
5235 optimized code for single-character negative classes.
5236
5237 5. Added a contributed patch from the TIN world which does the following:
5238
5239 + Add an undef for memmove, in case the the system defines a macro for it.
5240
5241 + Add a definition of offsetof(), in case there isn't one. (I don't know
5242 the reason behind this - offsetof() is part of the ANSI standard - but
5243 it does no harm).
5244
5245 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5246 most of the places where whitespace preceded '#'. I have given up and
5247 allowed the remaining 2 cases to be at the margin.
5248
5249 + Rename some variables in pcre to eliminate shadowing. This seems very
5250 pedantic, but does no harm, of course.
5251
5252 6. Moved the call to setjmp() into its own function, to get rid of warnings
5253 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5254
5255 7. Constructs such as \d{8,} were compiling into the equivalent of
5256 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5257 outcome, but in this particular case used more store than had been allocated,
5258 which caused the bug to be discovered because it threw up an internal error.
5259
5260 8. The debugging code in both pcre and pcretest for outputting the compiled
5261 form of a regex was going wrong in the case of back references followed by
5262 curly-bracketed repeats.
5263
5264
5265 Version 1.02 12-Dec-97
5266 ----------------------
5267
5268 1. Typos in pcre.3 and comments in the source fixed.
5269
5270 2. Applied a contributed patch to get rid of places where it used to remove
5271 'const' from variables, and fixed some signed/unsigned and uninitialized
5272 variable warnings.
5273
5274 3. Added the "runtest" target to Makefile.
5275
5276 4. Set default compiler flag to -O2 rather than just -O.
5277
5278
5279 Version 1.01 19-Nov-97
5280 ----------------------
5281
5282 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5283 like /([ab]*)*/, that is, for classes with more than one character in them.
5284
5285 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5286 as /((?>a*))*/ (a PCRE_EXTRA facility).
5287
5288
5289 Version 1.00 18-Nov-97
5290 ----------------------
5291
5292 1. Added compile-time macros to support systems such as SunOS4 which don't have
5293 memmove() or strerror() but have other things that can be used instead.
5294
5295 2. Arranged that "make clean" removes the executables.
5296
5297
5298 Version 0.99 27-Oct-97
5299 ----------------------
5300
5301 1. Fixed bug in code for optimizing classes with only one character. It was
5302 initializing a 32-byte map regardless, which could cause it to run off the end
5303 of the memory it had got.
5304
5305 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5306
5307
5308 Version 0.98 22-Oct-97
5309 ----------------------
5310
5311 1. Fixed bug in code for handling temporary memory usage when there are more
5312 back references than supplied space in the ovector. This could cause segfaults.
5313
5314
5315 Version 0.97 21-Oct-97
5316 ----------------------
5317
5318 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5319
5320 2. Optimized negated single characters not to use a bit map.
5321
5322 3. Brought error texts together as macro definitions; clarified some of them;
5323 fixed one that was wrong - it said "range out of order" when it meant "invalid
5324 escape sequence".
5325
5326 4. Changed some char * arguments to const char *.
5327
5328 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5329
5330 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5331 pcretest.
5332
5333
5334 Version 0.96 16-Oct-97
5335 ----------------------
5336
5337 1. Added a simple "pgrep" utility to the distribution.
5338
5339 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5340 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5341 where "ddd" means "one or more decimal digits".
5342
5343 3. Fixed serious bug. If a pattern had a back reference, but the call to
5344 pcre_exec() didn't supply a large enough ovector to record the related
5345 identifying subpattern, the match always failed. PCRE now remembers the number
5346 of the largest back reference, and gets some temporary memory in which to save
5347 the offsets during matching if necessary, in order to ensure that
5348 backreferences always work.
5349
5350 4. Increased the compatibility with Perl in a number of ways:
5351
5352 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5353 to request this handling. The option can be set at compile or exec time.
5354
5355 (b) $ matches before a terminating newline by default; an option
5356 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5357 mode). The option can be set at compile or exec time.
5358
5359 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5360 the same as Perl's. If the decimal number it represents is less than 10
5361 or there aren't that many previous left capturing parentheses, an octal
5362 escape is read. Inside a character class, it's always an octal escape,
5363 even if it is a single digit.
5364
5365 (d) An escaped but undefined alphabetic character is taken as a literal,
5366 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5367 escapes.
5368
5369 (e) {0} is now permitted. (The previous item is removed from the compiled
5370 pattern).
5371
5372 5. Changed all the names of code files so that the basic parts are no longer
5373 than 10 characters, and abolished the teeny "globals.c" file.
5374
5375 6. Changed the handling of character classes; they are now done with a 32-byte
5376 bit map always.
5377
5378 7. Added the -d and /D options to pcretest to make it possible to look at the
5379 internals of compilation without having to recompile pcre.
5380
5381
5382 Version 0.95 23-Sep-97
5383 ----------------------
5384
5385 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5386 \x20 at the start of a run of normal characters. These were being treated as
5387 real characters, instead of the source characters being re-checked.
5388
5389
5390 Version 0.94 18-Sep-97
5391 ----------------------
5392
5393 1. The functions are now thread-safe, with the caveat that the global variables
5394 containing pointers to malloc() and free() or alternative functions are the
5395 same for all threads.
5396
5397 2. Get pcre_study() to generate a bitmap of initial characters for non-
5398 anchored patterns when this is possible, and use it if passed to pcre_exec().
5399
5400
5401 Version 0.93 15-Sep-97
5402 ----------------------
5403
5404 1. /(b)|(:+)/ was computing an incorrect first character.
5405
5406 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5407 but not actually doing anything yet.
5408
5409 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5410 as Perl does (e.g. [-az] or [az-]).
5411
5412 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5413 all possible positions.
5414
5415 5. Split up into different modules to avoid including unneeded functions in a
5416 compiled binary. However, compile and exec are still in one module. The "study"
5417 function is split off.
5418
5419 6. The character tables are now in a separate module whose source is generated
5420 by an auxiliary program - but can then be edited by hand if required. There are
5421 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5422 toupper() in the code.
5423
5424 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5425 make them global. Abolish the function for setting them, as the caller can now
5426 set them directly.
5427
5428
5429 Version 0.92 11-Sep-97
5430 ----------------------
5431
5432 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5433 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5434
5435 2. Caseless matching was not working in character classes if the characters in
5436 the pattern were in upper case.
5437
5438 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5439
5440 4. Make PCRE_ANCHORED public and accept as a compile option.
5441
5442 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5443 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5444 pass them.
5445
5446 6. Give an error if bad option bits passed at compile or run time.
5447
5448 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5449 pcretest to cause it to pass that flag.
5450
5451 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5452 options, and the first character, if set.
5453
5454 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5455
5456
5457 Version 0.91 10-Sep-97
5458 ----------------------
5459
5460 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5461 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5462
5463 2. PCRE was looping on encountering an indefinitely repeated back reference to
5464 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5465 Perl does - treats the match as successful.
5466
5467 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5