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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1498 - (show annotations) (download)
Fri Aug 8 15:22:51 2014 UTC (2 weeks ago) by ph10
File size: 268458 byte(s)
Fix compile-time loop for recursive reference within a group with an indefinite 
repeat.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.36 xx-xxx-2014
5 ------------------------
6
7 1. Got rid of some compiler warnings in the C++ modules that were shown up by
8 -Wmissing-field-initializers and -Wunused-parameter.
9
10 2. The tests for quantifiers being too big (greater than 65535) were being
11 applied after reading the number, and stupidly assuming that integer
12 overflow would give a negative number. The tests are now applied as the
13 numbers are read.
14
15 3. Tidy code in pcre_exec.c where two branches that used to be different are
16 now the same.
17
18 4. The JIT compiler did not generate match limit checks for certain
19 bracketed expressions with quantifiers. This may lead to exponential
20 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
21 issue should be resolved now.
22
23 5. Fixed an issue, which occures when nested alternatives are optimized
24 with table jumps.
25
26 6. Inserted two casts and changed some ints to size_t in the light of some
27 reported 64-bit compiler warnings (Bugzilla 1477).
28
29 7. Fixed a bug concerned with zero-minimum possessive groups that could match
30 an empty string, which sometimes were behaving incorrectly in the
31 interpreter (though correctly in the JIT matcher). This pcretest input is
32 an example:
33
34 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
35 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
36
37 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
38 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
39 for an empty string was breaking the inner loop and carrying on at a lower
40 level, when possessive repeated groups should always return to a higher
41 level as they have no backtrack points in them. The empty string test now
42 occurs at the outer level.
43
44 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
45 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
46
47 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
48 doesn't).
49
50 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
51 to hold a pattern to be incorrectly computed (too small) when there were
52 named back references to duplicated names. This could cause "internal
53 error: code overflow" or "double free or corruption" or other memory
54 handling errors.
55
56 11. When named subpatterns had the same prefixes, back references could be
57 confused. For example, in this pattern:
58
59 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
60
61 the reference to 'Name' was incorrectly treated as a reference to a
62 duplicate name.
63
64 12. A pattern such as /^s?c/mi8 where the optional character has more than
65 one "other case" was incorrectly compiled such that it would only try to
66 match starting at "c".
67
68 13. When a pattern starting with \s was studied, VT was not included in the
69 list of possible starting characters; this should have been part of the
70 8.34/18 patch.
71
72 14. If a character class started [\Qx]... where x is any character, the class
73 was incorrectly terminated at the ].
74
75 15. If a pattern that started with a caseless match for a character with more
76 than one "other case" was studied, PCRE did not set up the starting code
77 unit bit map for the list of possible characters. Now it does. This is an
78 optimization improvement, not a bug fix.
79
80 16. The Unicode data tables have been updated to Unicode 7.0.0.
81
82 17. Fixed a number of memory leaks in pcregrep.
83
84 18. Avoid a compiler warning (from some compilers) for a function call with
85 a cast that removes "const" from an lvalue by using an intermediate
86 variable (to which the compiler does not object).
87
88 19. Incorrect code was compiled if a group that contained an internal recursive
89 back reference was optional (had quantifier with a minimum of zero). This
90 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
91 caused segmentation faults because of stack overflows at compile time.
92
93 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
94 group that is quantified with an indefinite repeat, caused a compile-time
95 loop which used up all the system stack and provoked a segmentation fault.
96 This was not the same bug as 19 above.
97
98
99 Version 8.35 04-April-2014
100 --------------------------
101
102 1. A new flag is set, when property checks are present in an XCLASS.
103 When this flag is not set, PCRE can perform certain optimizations
104 such as studying these XCLASS-es.
105
106 2. The auto-possessification of character sets were improved: a normal
107 and an extended character set can be compared now. Furthermore
108 the JIT compiler optimizes more character set checks.
109
110 3. Got rid of some compiler warnings for potentially uninitialized variables
111 that show up only when compiled with -O2.
112
113 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
114 of a match later then the end of the match. The pcretest program was not
115 handling the case sensibly - it was outputting from the start to the next
116 binary zero. It now reports this situation in a message, and outputs the
117 text from the end to the start.
118
119 5. Fast forward search is improved in JIT. Instead of the first three
120 characters, any three characters with fixed position can be searched.
121 Search order: first, last, middle.
122
123 6. Improve character range checks in JIT. Characters are read by an inprecise
124 function now, which returns with an unknown value if the character code is
125 above a certain threshold (e.g: 256). The only limitation is that the value
126 must be bigger than the threshold as well. This function is useful when
127 the characters above the threshold are handled in the same way.
128
129 7. The macros whose names start with RAWUCHAR are placeholders for a future
130 mode in which only the bottom 21 bits of 32-bit data items are used. To
131 make this more memorable for those maintaining the code, the names have
132 been changed to start with UCHAR21, and an extensive comment has been added
133 to their definition.
134
135 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
136 to the export list in Makefile.am (they were accidentally omitted from the
137 8.34 tarball).
138
139 9. The informational output from pcretest used the phrase "starting byte set"
140 which is inappropriate for the 16-bit and 32-bit libraries. As the output
141 for "first char" and "need char" really means "non-UTF-char", I've changed
142 "byte" to "char", and slightly reworded the output. The documentation about
143 these values has also been (I hope) clarified.
144
145 10. Another JIT related optimization: use table jumps for selecting the correct
146 backtracking path, when more than four alternatives are present inside a
147 bracket.
148
149 11. Empty match is not possible, when the minimum length is greater than zero,
150 and there is no \K in the pattern. JIT should avoid empty match checks in
151 such cases.
152
153 12. In a caseless character class with UCP support, when a character with more
154 than one alternative case was not the first character of a range, not all
155 the alternative cases were added to the class. For example, s and \x{17f}
156 are both alternative cases for S: the class [RST] was handled correctly,
157 but [R-T] was not.
158
159 13. The configure.ac file always checked for pthread support when JIT was
160 enabled. This is not used in Windows, so I have put this test inside a
161 check for the presence of windows.h (which was already tested for).
162
163 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
164 The algorithm provides a way to skip certain starting offsets, and usually
165 faster than linear prefix searches.
166
167 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
168 as for 'fr_FR' and 'french'. For some reason, however, it then used the
169 Windows-specific input and output files, which have 'french' screwed in.
170 So this could never have worked. One of the problems with locales is that
171 they aren't always the same. I have now updated RunTest so that it checks
172 the output of the locale test (test 3) against three different output
173 files, and it allows the test to pass if any one of them matches. With luck
174 this should make the test pass on some versions of Solaris where it was
175 failing. Because of the uncertainty, the script did not used to stop if
176 test 3 failed; it now does. If further versions of a French locale ever
177 come to light, they can now easily be added.
178
179 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
180 there was a message during ./configure, but it did not stop. This now
181 provokes an error. The invalid example in README has been corrected.
182 If a value less than the minimum is given, the minimum value has always
183 been used, but now a warning is given.
184
185 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
186 was a bug in the test system, which is now fixed. Also, the list of various
187 configurations that are tested for each release did not have one with both
188 16/32 bits and --enable-bar-anycrlf. It now does.
189
190 18. pcretest was missing "-C bsr" for displaying the \R default setting.
191
192 19. Little endian PowerPC systems are supported now by the JIT compiler.
193
194 20. The fast forward newline mechanism could enter to an infinite loop on
195 certain invalid UTF-8 input. Although we don't support these cases
196 this issue can be fixed by a performance optimization.
197
198 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
199 not take account if existing stack usage. There is now a new global
200 variable called pcre_stack_guard that can be set to point to an external
201 function to check stack availability. It is called at the start of
202 processing every parenthesized group.
203
204 22. A typo in the code meant that in ungreedy mode the max/min qualifier
205 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
206 not match "ab".
207
208 23. When UTF was disabled, the JIT program reported some incorrect compile
209 errors. These messages are silenced now.
210
211 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
212 compiler.
213
214 25. Change all the temporary files used in RunGrepTest to be different to those
215 used by RunTest so that the tests can be run simultaneously, for example by
216 "make -j check".
217
218
219 Version 8.34 15-December-2013
220 -----------------------------
221
222 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
223 executable memory. Patch inspired by Carsten Klein.
224
225 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
226 this macro is never tested and has no effect, because the work to support
227 coverage involves only compiling and linking options and special targets in
228 the Makefile. The comment in config.h implied that defining the macro would
229 enable coverage support, which is totally false. There was also support for
230 setting this macro in the CMake files (my fault, I just copied it from
231 configure). SUPPORT_GCOV has now been removed.
232
233 3. Make a small performance improvement in strlen16() and strlen32() in
234 pcretest.
235
236 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
237 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
238 these cases). There was also one in pcretest.c.
239
240 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
241
242 6. In UTF mode, the code for checking whether a group could match an empty
243 string (which is used for indefinitely repeated groups to allow for
244 breaking an infinite loop) was broken when the group contained a repeated
245 negated single-character class with a character that occupied more than one
246 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
247 UTF-8 mode). The effect was undefined: the group might or might not be
248 deemed as matching an empty string, or the program might have crashed.
249
250 7. The code for checking whether a group could match an empty string was not
251 recognizing that \h, \H, \v, \V, and \R must match a character.
252
253 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
254 an empty string. If it can, pcretest shows this in its information output.
255
256 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
257 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
258 matched by pcre_exec() without using JIT:
259
260 (a) If the rest of the pattern did not match after a maximal run of
261 grapheme clusters, the code for backing up to try with fewer of them
262 did not always back up over a full grapheme when characters that do not
263 have the modifier quality were involved, e.g. Hangul syllables.
264
265 (b) If the match point in a subject started with modifier character, and
266 there was no match, the code could incorrectly back up beyond the match
267 point, and potentially beyond the first character in the subject,
268 leading to a segfault or an incorrect match result.
269
270 10. A conditional group with an assertion condition could lead to PCRE
271 recording an incorrect first data item for a match if no other first data
272 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
273 first data item, and therefore matched "ca" after "c" instead of at the
274 start.
275
276 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
277 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
278
279 12. The source of pcregrep now includes z/OS-specific code so that it can be
280 compiled for z/OS as part of the special z/OS distribution.
281
282 13. Added the -T and -TM options to pcretest.
283
284 14. The code in pcre_compile.c for creating the table of named capturing groups
285 has been refactored. Instead of creating the table dynamically during the
286 actual compiling pass, the information is remembered during the pre-compile
287 pass (on the stack unless there are more than 20 named groups, in which
288 case malloc() is used) and the whole table is created before the actual
289 compile happens. This has simplified the code (it is now nearly 150 lines
290 shorter) and prepared the way for better handling of references to groups
291 with duplicate names.
292
293 15. A back reference to a named subpattern when there is more than one of the
294 same name now checks them in the order in which they appear in the pattern.
295 The first one that is set is used for the reference. Previously only the
296 first one was inspected. This change makes PCRE more compatible with Perl.
297
298 16. Unicode character properties were updated from Unicode 6.3.0.
299
300 17. The compile-time code for auto-possessification has been refactored, based
301 on a patch by Zoltan Herczeg. It now happens after instead of during
302 compilation. The code is cleaner, and more cases are handled. The option
303 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
304 options in pcretest are provided to set it. It can also be set by
305 (*NO_AUTO_POSSESS) at the start of a pattern.
306
307 18. The character VT has been added to the default ("C" locale) set of
308 characters that match \s and are generally treated as white space,
309 following this same change in Perl 5.18. There is now no difference between
310 "Perl space" and "POSIX space". Whether VT is treated as white space in
311 other locales depends on the locale.
312
313 19. The code for checking named groups as conditions, either for being set or
314 for being recursed, has been refactored (this is related to 14 and 15
315 above). Processing unduplicated named groups should now be as fast at
316 numerical groups, and processing duplicated groups should be faster than
317 before.
318
319 20. Two patches to the CMake build system, by Alexander Barkov:
320
321 (1) Replace the "source" command by "." in CMakeLists.txt because
322 "source" is a bash-ism.
323
324 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
325 without these the CMake build does not work on Solaris.
326
327 21. Perl has changed its handling of \8 and \9. If there is no previously
328 encountered capturing group of those numbers, they are treated as the
329 literal characters 8 and 9 instead of a binary zero followed by the
330 literals. PCRE now does the same.
331
332 22. Following Perl, added \o{} to specify codepoints in octal, making it
333 possible to specify values greater than 0777 and also making them
334 unambiguous.
335
336 23. Perl now gives an error for missing closing braces after \x{... instead of
337 treating the string as literal. PCRE now does the same.
338
339 24. RunTest used to grumble if an inappropriate test was selected explicitly,
340 but just skip it when running all tests. This make it awkward to run ranges
341 of tests when one of them was inappropriate. Now it just skips any
342 inappropriate tests, as it always did when running all tests.
343
344 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
345 character types such as \d or \w, too many callouts were inserted, and the
346 data that they returned was rubbish.
347
348 26. In UCP mode, \s was not matching two of the characters that Perl matches,
349 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
350 were matched by \h. The code has now been refactored so that the lists of
351 the horizontal and vertical whitespace characters used for \h and \v (which
352 are defined only in one place) are now also used for \s.
353
354 27. Add JIT support for the 64 bit TileGX architecture.
355 Patch by Jiong Wang (Tilera Corporation).
356
357 28. Possessive quantifiers for classes (both explicit and automatically
358 generated) now use special opcodes instead of wrapping in ONCE brackets.
359
360 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
361 (because it's meaningless), this was not happening when PCRE_CASELESS was
362 set. Not wrong, but inefficient.
363
364 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
365 properties for \w, \d, etc) is present in a test regex. Otherwise if the
366 test contains no characters greater than 255, Perl doesn't realise it
367 should be using Unicode semantics.
368
369 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
370 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
371 does in Unicode mode.
372
373 32. Added the "forbid" facility to pcretest so that putting tests into the
374 wrong test files can sometimes be quickly detected.
375
376 33. There is now a limit (default 250) on the depth of nesting of parentheses.
377 This limit is imposed to control the amount of system stack used at compile
378 time. It can be changed at build time by --with-parens-nest-limit=xxx or
379 the equivalent in CMake.
380
381 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
382 errors. Perl warns for these when in warning mode, but PCRE has no facility
383 for giving warnings.
384
385 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
386 However, this was not working for (?!) because it is optimized to (*FAIL),
387 for which PCRE does not allow quantifiers. The optimization is now disabled
388 when a quantifier follows (?!). I can't see any use for this, but it makes
389 things uniform.
390
391 36. Perl no longer allows group names to start with digits, so I have made this
392 change also in PCRE. It simplifies the code a bit.
393
394 37. In extended mode, Perl ignores spaces before a + that indicates a
395 possessive quantifier. PCRE allowed a space before the quantifier, but not
396 before the possessive +. It now does.
397
398 38. The use of \K (reset reported match start) within a repeated possessive
399 group such as (a\Kb)*+ was not working.
400
401 40. Document that the same character tables must be used at compile time and
402 run time, and that the facility to pass tables to pcre_exec() and
403 pcre_dfa_exec() is for use only with saved/restored patterns.
404
405 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
406 features for Builds with MSVC:
407
408 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
409 builds). The .rc files can be used to set FileDescription and many other
410 attributes.
411
412 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
413 This allows higher-level build scripts which want .pdb files to avoid
414 hard-coding the exact files needed."
415
416 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
417 mean "start of word" and "end of word", respectively, as a transition aid.
418
419 43. A minimizing repeat of a class containing codepoints greater than 255 in
420 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
421 compiled to use the heap for recursion.
422
423 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
424 is configured.
425
426
427 Version 8.33 28-May-2013
428 ------------------------
429
430 1. Added 'U' to some constants that are compared to unsigned integers, to
431 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
432 variables that are added to signed variables, to ensure the result is
433 signed and can be negated.
434
435 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
436 CMake config files.
437
438 3. Revise the creation of config.h.generic so that all boolean macros are
439 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
440 overriding via -D on the command line possible.
441
442 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
443 to unsigned int is reported to make a quite noticeable speed difference in
444 a specific Windows environment. Testing on Linux did also appear to show
445 some benefit (and it is clearly not harmful). Also fixed the definition of
446 Xop which should be unsigned.
447
448 5. Related to (4), changing the definition of the intermediate variable cc
449 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
450 improvements.
451
452 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
453 unnecessary spaces.
454
455 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
456 and later.
457
458 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
459
460 9. Optimizing fast_forward_start_bits in JIT.
461
462 10. Adding support for callouts in JIT, and fixing some issues revealed
463 during this work. Namely:
464
465 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
466
467 (b) Minimum length was not checked before the matching is started.
468
469 11. The value of capture_last that is passed to callouts was incorrect in some
470 cases when there was a capture on one path that was subsequently abandoned
471 after a backtrack. Also, the capture_last value is now reset after a
472 recursion, since all captures are also reset in this case.
473
474 12. The interpreter no longer returns the "too many substrings" error in the
475 case when an overflowing capture is in a branch that is subsequently
476 abandoned after a backtrack.
477
478 13. In the pathological case when an offset vector of size 2 is used, pcretest
479 now prints out the matched string after a yield of 0 or 1.
480
481 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
482 Only supported by the JIT compiler at the moment.
483
484 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
485
486 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
487 offset of the starting point of the matching process, provided the offsets
488 vector is large enough.
489
490 17. The \A escape now records a lookbehind value of 1, though its execution
491 does not actually inspect the previous character. This is to ensure that,
492 in partial multi-segment matching, at least one character from the old
493 segment is retained when a new segment is processed. Otherwise, if there
494 are no lookbehinds in the pattern, \A might match incorrectly at the start
495 of a new segment.
496
497 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
498
499 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
500 gives some modest performance improvement in 8-bit mode.
501
502 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
503 be expressed in certain programming languages using Universal Character
504 Names.
505
506 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
507 which points out that "non characters" are not "characters that may not
508 appear in Unicode strings" but rather "characters that are reserved for
509 internal use and have only local meaning".
510
511 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
512 there was a conditional group that depended on an assertion, if the
513 assertion was false, the callout that immediately followed the alternation
514 in the condition was skipped when pcre_exec() was used for matching.
515
516 23. Allow an explicit callout to be inserted before an assertion that is the
517 condition for a conditional group, for compatibility with automatic
518 callouts, which always insert a callout at this point.
519
520 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
521 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
522
523 25. (*PRUNE) is now supported by the JIT compiler.
524
525 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
526
527 27. Fix the case where there are two or more SKIPs with arguments that may be
528 ignored.
529
530 28. (*SKIP) is now supported by the JIT compiler.
531
532 29. (*THEN) is now supported by the JIT compiler.
533
534 30. Update RunTest with additional test selector options.
535
536 31. The way PCRE handles backtracking verbs has been changed in two ways.
537
538 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
539 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
540 backtracking. In some cases this makes it more Perl-compatible, but Perl's
541 rather obscure rules do not always do the same thing.
542
543 (2) Previously, backtracking verbs were confined within assertions. This is
544 no longer the case for positive assertions, except for (*ACCEPT). Again,
545 this sometimes improves Perl compatibility, and sometimes does not.
546
547 32. A number of tests that were in test 2 because Perl did things differently
548 have been moved to test 1, because either Perl or PCRE has changed, and
549 these tests are now compatible.
550
551 32. Backtracking control verbs are now handled in the same way in JIT and
552 interpreter.
553
554 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
555 contained a forward subroutine reference caused a compile error.
556
557 34. Auto-detect and optimize limited repetitions in JIT.
558
559 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
560 blocking (*UTF) etc.
561
562 36. In the interpreter, maximizing pattern repetitions for characters and
563 character types now use tail recursion, which reduces stack usage.
564
565 37. The value of the max lookbehind was not correctly preserved if a compiled
566 and saved regex was reloaded on a host of different endianness.
567
568 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
569 of the compiled pattern block, expand the flags field from 16 to 32 bits
570 because it was almost full.
571
572 39. Try madvise first before posix_madvise.
573
574 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
575 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
576 original change locked out all matches of empty strings. This has been
577 changed so that one match of an empty string per line is recognized.
578 Subsequent searches on the same line (for colouring or for --only-matching,
579 for example) do not recognize empty strings.
580
581 41. Applied a user patch to fix a number of spelling mistakes in comments.
582
583 42. Data lines longer than 65536 caused pcretest to crash.
584
585 43. Clarified the data type for length and startoffset arguments for pcre_exec
586 and pcre_dfa_exec in the function-specific man pages, where they were
587 explicitly stated to be in bytes, never having been updated. I also added
588 some clarification to the pcreapi man page.
589
590 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
591 a segmentation fault.
592
593
594 Version 8.32 30-November-2012
595 -----------------------------
596
597 1. Improved JIT compiler optimizations for first character search and single
598 character iterators.
599
600 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
601 Patch by Daniel Richard G.
602
603 3. Single character iterator optimizations in the JIT compiler.
604
605 4. Improved JIT compiler optimizations for character ranges.
606
607 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
608 Reported by Giuseppe D'Angelo.
609
610 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
611 of a line, was being set incorrectly in cases where .* appeared inside
612 atomic brackets at the start of a pattern, or where there was a subsequent
613 *PRUNE or *SKIP.
614
615 7. Improved instruction cache flush for POWER/PowerPC.
616 Patch by Daniel Richard G.
617
618 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
619 grep:
620
621 (a) There is now no limit to the number of patterns to be matched.
622
623 (b) An error is given if a pattern is too long.
624
625 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
626 are now supported.
627
628 (d) --exclude-from and --include-from (multiple use) have been added.
629
630 (e) Exclusions and inclusions now apply to all files and directories, not
631 just to those obtained from scanning a directory recursively.
632
633 (f) Multiple uses of -f and --file-list are now supported.
634
635 (g) In a Windows environment, the default for -d has been changed from
636 "read" (the GNU grep default) to "skip", because otherwise the presence
637 of a directory in the file list provokes an error.
638
639 (h) The documentation has been revised and clarified in places.
640
641 9. Improve the matching speed of capturing brackets.
642
643 10. Changed the meaning of \X so that it now matches a Unicode extended
644 grapheme cluster.
645
646 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
647 out POSIX threads when JIT support is configured.
648
649 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
650
651 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
652 structure could go wrong in environments where size_t is not the same size
653 as int.
654
655 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
656 set.
657
658 15. The EBCDIC support had decayed; later updates to the code had included
659 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
660 general tidy up of EBCDIC-related issues, and the documentation was also
661 not quite right. There is now a test that can be run on ASCII systems to
662 check some of the EBCDIC-related things (but is it not a full test).
663
664 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
665 in a small tidy to the code.
666
667 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
668
669 18. If the --only-matching (-o) option in pcregrep is specified multiple
670 times, each one causes appropriate output. For example, -o1 -o2 outputs the
671 substrings matched by the 1st and 2nd capturing parentheses. A separating
672 string can be specified by --om-separator (default empty).
673
674 19. Improving the first n character searches.
675
676 20. Turn case lists for horizontal and vertical white space into macros so that
677 they are defined only once.
678
679 21. This set of changes together give more compatible Unicode case-folding
680 behaviour for characters that have more than one other case when UCP
681 support is available.
682
683 (a) The Unicode property table now has offsets into a new table of sets of
684 three or more characters that are case-equivalent. The MultiStage2.py
685 script that generates these tables (the pcre_ucd.c file) now scans
686 CaseFolding.txt instead of UnicodeData.txt for character case
687 information.
688
689 (b) The code for adding characters or ranges of characters to a character
690 class has been abstracted into a generalized function that also handles
691 case-independence. In UTF-mode with UCP support, this uses the new data
692 to handle characters with more than one other case.
693
694 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
695 whose other case is greater than 256 are now correctly matched
696 caselessly. Previously, the high codepoint matched the low one, but not
697 vice versa.
698
699 (d) The processing of \h, \H, \v, and \ in character classes now makes use
700 of the new class addition function, using character lists defined as
701 macros alongside the case definitions of 20 above.
702
703 (e) Caseless back references now work with characters that have more than
704 one other case.
705
706 (f) General caseless matching of characters with more than one other case
707 is supported.
708
709 22. Unicode character properties were updated from Unicode 6.2.0
710
711 23. Improved CMake support under Windows. Patch by Daniel Richard G.
712
713 24. Add support for 32-bit character strings, and UTF-32
714
715 25. Major JIT compiler update (code refactoring and bugfixing).
716 Experimental Sparc 32 support is added.
717
718 26. Applied a modified version of Daniel Richard G's patch to create
719 pcre.h.generic and config.h.generic by "make" instead of in the
720 PrepareRelease script.
721
722 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
723 pcre_compile.c when checking for a zero character.
724
725 28. Introducing a native interface for JIT. Through this interface, the compiled
726 machine code can be directly executed. The purpose of this interface is to
727 provide fast pattern matching, so several sanity checks are not performed.
728 However, feature tests are still performed. The new interface provides
729 1.4x speedup compared to the old one.
730
731 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
732 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
733 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
734
735 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
736 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
737 the "old" RFC 2279). Instead, it ended up passing a negative length to
738 pcre_exec().
739
740 31. Add support for GCC's visibility feature to hide internal functions.
741
742 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
743 "unknown -C option" after outputting 0 or 1.
744
745 33. There is now support for generating a code coverage report for the test
746 suite in environments where gcc is the compiler and lcov is installed. This
747 is mainly for the benefit of the developers.
748
749 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
750 unaddressable using valgrind annotations, allowing valgrind to detect
751 invalid memory accesses. This is mainly for the benefit of the developers.
752
753 25. (*UTF) can now be used to start a pattern in any of the three libraries.
754
755 26. Give configure error if --enable-cpp but no C++ compiler found.
756
757
758 Version 8.31 06-July-2012
759 -------------------------
760
761 1. Fixing a wrong JIT test case and some compiler warnings.
762
763 2. Removed a bashism from the RunTest script.
764
765 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
766 to unsigned type, result still unsigned" that was given by an MS compiler
767 on encountering the code "-sizeof(xxx)".
768
769 4. Partial matching support is added to the JIT compiler.
770
771 5. Fixed several bugs concerned with partial matching of items that consist
772 of more than one character:
773
774 (a) /^(..)\1/ did not partially match "aba" because checking references was
775 done on an "all or nothing" basis. This also applied to repeated
776 references.
777
778 (b) \R did not give a hard partial match if \r was found at the end of the
779 subject.
780
781 (c) \X did not give a hard partial match after matching one or more
782 characters at the end of the subject.
783
784 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
785 a partial match for the string "\r".
786
787 (e) When newline was set to CRLF, the metacharacter "." did not recognize
788 a partial match for a CR character at the end of the subject string.
789
790 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
791 running pcretest, the text "(JIT)" added to the output whenever JIT is
792 actually used to run the match.
793
794 7. Individual JIT compile options can be set in pcretest by following -s+[+]
795 or /S+[+] with a digit between 1 and 7.
796
797 8. OP_NOT now supports any UTF character not just single-byte ones.
798
799 9. (*MARK) control verb is now supported by the JIT compiler.
800
801 10. The command "./RunTest list" lists the available tests without actually
802 running any of them. (Because I keep forgetting what they all are.)
803
804 11. Add PCRE_INFO_MAXLOOKBEHIND.
805
806 12. Applied a (slightly modified) user-supplied patch that improves performance
807 when the heap is used for recursion (compiled with --disable-stack-for-
808 recursion). Instead of malloc and free for each heap frame each time a
809 logical recursion happens, frames are retained on a chain and re-used where
810 possible. This sometimes gives as much as 30% improvement.
811
812 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
813 call.
814
815 14. As documented, (*COMMIT) is now confined to within a positive assertion.
816
817 15. It is now possible to link pcretest with libedit as an alternative to
818 libreadline.
819
820 16. (*COMMIT) control verb is now supported by the JIT compiler.
821
822 17. The Unicode data tables have been updated to Unicode 6.1.0.
823
824 18. Added --file-list option to pcregrep.
825
826 19. Added binary file support to pcregrep, including the -a, --binary-files,
827 -I, and --text options.
828
829 20. The madvise function is renamed for posix_madvise for QNX compatibility
830 reasons. Fixed by Giuseppe D'Angelo.
831
832 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
833
834 22. Changed the option for creating symbolic links for 16-bit man pages from
835 -s to -sf so that re-installing does not cause issues.
836
837 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
838
839 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
840 matches in certain environments (the workspace was not being correctly
841 retained). Also added to pcre_dfa_exec() a simple plausibility check on
842 some of the workspace data at the beginning of a restart.
843
844 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
845 was not doing so when it should - probably a typo introduced by SVN 528
846 (change 8.10/14).
847
848 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
849 \w+ when the character tables indicated that \x{c4} was a word character.
850 There were several related cases, all because the tests for doing a table
851 lookup were testing for characters less than 127 instead of 255.
852
853 27. If a pattern contains capturing parentheses that are not used in a match,
854 their slots in the ovector are set to -1. For those that are higher than
855 any matched groups, this happens at the end of processing. In the case when
856 there were back references that the ovector was too small to contain
857 (causing temporary malloc'd memory to be used during matching), and the
858 highest capturing number was not used, memory off the end of the ovector
859 was incorrectly being set to -1. (It was using the size of the temporary
860 memory instead of the true size.)
861
862 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
863 ovector size, it uses memory at the end of the block that it has got.
864
865 29. Check for an overlong MARK name and give an error at compile time. The
866 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
867
868 30. JIT compiler update.
869
870 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
871 Rill for the patch.
872
873 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
874
875 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
876
877 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
878 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
879 was enabled.
880
881 35. Improve JIT code generation for greedy plus quantifier.
882
883 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
884 1 to "aa" instead of to an empty string. The bug affected repeated groups
885 that could potentially match an empty string.
886
887 37. Optimizing single character iterators in JIT.
888
889 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
890 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
891 codepoints that are too big for the mode are faulted, and in a UTF mode,
892 disallowed codepoints are also faulted.
893
894 39. If PCRE was compiled with UTF support, in three places in the DFA
895 matcher there was code that should only have been obeyed in UTF mode, but
896 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
897 processing when bytes with values greater than 127 were present. In 16-bit
898 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
899 both cases the values are those that cannot be the first data item in a UTF
900 character. The three items that might have provoked this were recursions,
901 possessively repeated groups, and atomic groups.
902
903 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
904 and pcregrep, because some OS require shared objects to be explicitly
905 passed to ld, causing the link step to fail if they are not.
906
907 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
908 mode, patterns that started with \h* or \R* might be incorrectly matched.
909
910
911 Version 8.30 04-February-2012
912 -----------------------------
913
914 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
915 name is defined in ctype.h.
916
917 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
918 only in quite long subpatterns.
919
920 3. Removed the function pcre_info(), which has been obsolete and deprecated
921 since it was replaced by pcre_fullinfo() in February 2000.
922
923 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
924 match a (*MARK), and the match failed at the start of the subject, a
925 reference to memory before the start of the subject could occur. This bug
926 was introduced by fix 17 of release 8.21.
927
928 5. A reference to an unset group with zero minimum repetition was giving
929 totally wrong answers (in non-JavaScript-compatibility mode). For example,
930 /(another)?(\1?)test/ matched against "hello world test". This bug was
931 introduced in release 8.13.
932
933 6. Add support for 16-bit character strings (a large amount of work involving
934 many changes and refactorings).
935
936 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
937 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
938 from a file.
939
940 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
941 rounding is not applied in this particular case).
942
943 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
944 if they appear, or are escaped, in patterns.
945
946 10. Get rid of a number of -Wunused-but-set-variable warnings.
947
948 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
949 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
950 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
951 also returns the mark "x". This bug applied to capturing parentheses,
952 non-capturing parentheses, and atomic parentheses. It also applied to some
953 assertions.
954
955 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
956 information out of configure.ac instead of relying on pcre.h.generic, which
957 is not stored in the repository.
958
959 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
960 -lreadline.
961
962 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
963
964 15. Applied Graycode's patch to put the top-level frame on the stack rather
965 than the heap when not using the stack for recursion. This gives a
966 performance improvement in many cases when recursion is not deep.
967
968 16. Experimental code added to "pcretest -C" to output the stack frame size.
969
970
971 Version 8.21 12-Dec-2011
972 ------------------------
973
974 1. Updating the JIT compiler.
975
976 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
977 are added as well.
978
979 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
980 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
981 calling _pcre_jit_exec. Some extra comments are added.
982
983 4. (*MARK) settings inside atomic groups that do not contain any capturing
984 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
985 was introduced by change 18 for 8.20.
986
987 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
988 ECMA-262 standard.
989
990 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
991 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
992 This bug was probably introduced by change 9 of 8.13.
993
994 7. While fixing 6 above, I noticed that a number of other items were being
995 incorrectly rejected as "not fixed length". This arose partly because newer
996 opcodes had not been added to the fixed-length checking code. I have (a)
997 corrected the bug and added tests for these items, and (b) arranged for an
998 error to occur if an unknown opcode is encountered while checking for fixed
999 length instead of just assuming "not fixed length". The items that were
1000 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1001 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1002 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1003
1004 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1005 being incorrectly compiled and would have given unpredicatble results.
1006
1007 9. A possessively repeated subpattern with minimum repeat count greater than
1008 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1009 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1010 the first (A) could occur when it should not.
1011
1012 10. Add a cast and remove a redundant test from the code.
1013
1014 11. JIT should use pcre_malloc/pcre_free for allocation.
1015
1016 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1017 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1018 is anything other than /usr, -L is still shown).
1019
1020 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1021
1022 14. Perl does not support \N without a following name in a [] class; PCRE now
1023 also gives an error.
1024
1025 15. If a forward reference was repeated with an upper limit of around 2000,
1026 it caused the error "internal error: overran compiling workspace". The
1027 maximum number of forward references (including repeats) was limited by the
1028 internal workspace, and dependent on the LINK_SIZE. The code has been
1029 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1030 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1031 of around 200,000 forward references. While doing this, I also speeded up
1032 the filling in of repeated forward references.
1033
1034 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1035 incorrectly expecting the subject to contain another "a" after the start.
1036
1037 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1038 in the match, the SKIP should be ignored. This was not happening; instead
1039 the SKIP was being treated as NOMATCH. For patterns such as
1040 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1041 tested.
1042
1043 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1044 now much more compatible with Perl, in particular in cases where the result
1045 is a non-match for a non-anchored pattern. For example, if
1046 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1047 "m", where previously it did not return a name. A side effect of this
1048 change is that for partial matches, the last encountered mark name is
1049 returned, as for non matches. A number of tests that were previously not
1050 Perl-compatible have been moved into the Perl-compatible test files. The
1051 refactoring has had the pleasing side effect of removing one argument from
1052 the match() function, thus reducing its stack requirements.
1053
1054 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1055 subsequent uses of /S (without +) incorrectly behaved like /S+.
1056
1057 21. Retrieve executable code size support for the JIT compiler and fixing
1058 some warnings.
1059
1060 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1061 not work when the shorter character appeared right at the end of the
1062 subject string.
1063
1064 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1065 systems.
1066
1067 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1068 output it when the /M option is used in pcretest.
1069
1070 25. The CheckMan script was not being included in the distribution. Also, added
1071 an explicit "perl" to run Perl scripts from the PrepareRelease script
1072 because this is reportedly needed in Windows.
1073
1074 26. If study data was being save in a file and studying had not found a set of
1075 "starts with" bytes for the pattern, the data written to the file (though
1076 never used) was taken from uninitialized memory and so caused valgrind to
1077 complain.
1078
1079 27. Updated RunTest.bat as provided by Sheri Pierce.
1080
1081 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1082
1083 29. Computation of memory usage for the table of capturing group names was
1084 giving an unnecessarily large value.
1085
1086
1087 Version 8.20 21-Oct-2011
1088 ------------------------
1089
1090 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1091 a POSIX class. After further experiments with Perl, which convinced me that
1092 Perl has bugs and confusions, a closing square bracket is no longer allowed
1093 in a POSIX name. This bug also affected patterns with classes that started
1094 with full stops.
1095
1096 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1097 captured substring, but while checking the failing first alternative,
1098 substring 1 is temporarily captured. If the output vector supplied to
1099 pcre_exec() was not big enough for this capture, the yield of the function
1100 was still zero ("insufficient space for captured substrings"). This cannot
1101 be totally fixed without adding another stack variable, which seems a lot
1102 of expense for a edge case. However, I have improved the situation in cases
1103 such as /(a)(b)x|abc/ matched against "abc", where the return code
1104 indicates that fewer than the maximum number of slots in the ovector have
1105 been set.
1106
1107 3. Related to (2) above: when there are more back references in a pattern than
1108 slots in the output vector, pcre_exec() uses temporary memory during
1109 matching, and copies in the captures as far as possible afterwards. It was
1110 using the entire output vector, but this conflicts with the specification
1111 that only 2/3 is used for passing back captured substrings. Now it uses
1112 only the first 2/3, for compatibility. This is, of course, another edge
1113 case.
1114
1115 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1116 main code base, and can be used by building with --enable-jit. When this is
1117 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1118 runtime --no-jit option is given.
1119
1120 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1121 ovector, the return from the function was zero, implying that there were
1122 other matches that did not fit. The correct "exactly full" value is now
1123 returned.
1124
1125 6. If a subpattern that was called recursively or as a subroutine contained
1126 (*PRUNE) or any other control that caused it to give a non-standard return,
1127 invalid errors such as "Error -26 (nested recursion at the same subject
1128 position)" or even infinite loops could occur.
1129
1130 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1131 computing the minimum length on reaching *ACCEPT, and so ended up with the
1132 wrong value of 1 rather than 0. Further investigation indicates that
1133 computing a minimum subject length in the presence of *ACCEPT is difficult
1134 (think back references, subroutine calls), and so I have changed the code
1135 so that no minimum is registered for a pattern that contains *ACCEPT.
1136
1137 8. If (*THEN) was present in the first (true) branch of a conditional group,
1138 it was not handled as intended. [But see 16 below.]
1139
1140 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1141 Sheri Pierce.
1142
1143 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1144 the first byte in a match must be "a".
1145
1146 11. Change 17 for 8.13 increased the recursion depth for patterns like
1147 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1148 pattern contains any instances of (*THEN). If it does not, the old
1149 optimizations are restored. It would be nice to do this on a per-group
1150 basis, but at the moment that is not feasible.
1151
1152 12. In some environments, the output of pcretest -C is CRLF terminated. This
1153 broke RunTest's code that checks for the link size. A single white space
1154 character after the value is now allowed for.
1155
1156 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1157 For "fr", it uses the Windows-specific input and output files.
1158
1159 14. If (*THEN) appeared in a group that was called recursively or as a
1160 subroutine, it did not work as intended. [But see next item.]
1161
1162 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1163 pattern fragments (but not containing any | characters). If A and B are
1164 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1165 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1166 D. In other words, Perl considers parentheses that do not contain any |
1167 characters to be part of a surrounding alternative, whereas PCRE was
1168 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1169 differently. PCRE now behaves in the same way as Perl, except in the case
1170 of subroutine/recursion calls such as (?1) which have in any case always
1171 been different (but PCRE had them first :-).
1172
1173 16. Related to 15 above: Perl does not treat the | in a conditional group as
1174 creating alternatives. Such a group is treated in the same way as an
1175 ordinary group without any | characters when processing (*THEN). PCRE has
1176 been changed to match Perl's behaviour.
1177
1178 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1179 RunGrepTest script failed.
1180
1181 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1182 inevitable for groups that contain captures, but it can lead to a lot of
1183 stack use in large patterns. The old behaviour has been restored for atomic
1184 groups that do not contain any capturing parentheses.
1185
1186 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1187 suppress the check for a minimum subject length at run time. (If it was
1188 given to pcre_exec() or pcre_dfa_exec() it did work.)
1189
1190 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1191 fail to work when decoding hex characters in data strings in EBCDIC
1192 environments.
1193
1194 21. It appears that in at least one Mac OS environment, the isxdigit() function
1195 is implemented as a macro that evaluates to its argument more than once,
1196 contravening the C 90 Standard (I haven't checked a later standard). There
1197 was an instance in pcretest which caused it to go wrong when processing
1198 \x{...} escapes in subject strings. The has been rewritten to avoid using
1199 things like p++ in the argument of isxdigit().
1200
1201
1202 Version 8.13 16-Aug-2011
1203 ------------------------
1204
1205 1. The Unicode data tables have been updated to Unicode 6.0.0.
1206
1207 2. Two minor typos in pcre_internal.h have been fixed.
1208
1209 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1210 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1211 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1212
1213 4. There were a number of related bugs in the code for matching backrefences
1214 caselessly in UTF-8 mode when codes for the characters concerned were
1215 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1216 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1217 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1218 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1219 2-byte code at the end of the subject (it thought there wasn't enough data
1220 left).
1221
1222 5. Comprehensive information about what went wrong is now returned by
1223 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1224 as the output vector has at least 2 elements. The offset of the start of
1225 the failing character and a reason code are placed in the vector.
1226
1227 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1228 now returned is for the first byte of the failing character, instead of the
1229 last byte inspected. This is an incompatible change, but I hope it is small
1230 enough not to be a problem. It makes the returned offset consistent with
1231 pcre_exec() and pcre_dfa_exec().
1232
1233 7. pcretest now gives a text phrase as well as the error number when
1234 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1235 failure, the offset and reason code are output.
1236
1237 8. When \R was used with a maximizing quantifier it failed to skip backwards
1238 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1239 back over a single character (\n). This seems wrong (because it treated the
1240 two characters as a single entity when going forwards), conflicts with the
1241 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1242 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1243 has been changed.
1244
1245 9. Some internal refactoring has changed the processing so that the handling
1246 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1247 time (the PCRE_DOTALL option was changed this way some time ago: version
1248 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1249 which was always a bit of a fudge. It also means that there is one less
1250 argument for the match() function, which reduces its stack requirements
1251 slightly. This change also fixes an incompatibility with Perl: the pattern
1252 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1253
1254 10. More internal refactoring has drastically reduced the number of recursive
1255 calls to match() for possessively repeated groups such as (abc)++ when
1256 using pcre_exec().
1257
1258 11. While implementing 10, a number of bugs in the handling of groups were
1259 discovered and fixed:
1260
1261 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1262 (a|)*(?1) gave a compile-time internal error.
1263 ((a|)+)+ did not notice that the outer group could match an empty string.
1264 (^a|^)+ was not marked as anchored.
1265 (.*a|.*)+ was not marked as matching at start or after a newline.
1266
1267 12. Yet more internal refactoring has removed another argument from the match()
1268 function. Special calls to this function are now indicated by setting a
1269 value in a variable in the "match data" data block.
1270
1271 13. Be more explicit in pcre_study() instead of relying on "default" for
1272 opcodes that mean there is no starting character; this means that when new
1273 ones are added and accidentally left out of pcre_study(), testing should
1274 pick them up.
1275
1276 14. The -s option of pcretest has been documented for ages as being an old
1277 synonym of -m (show memory usage). I have changed it to mean "force study
1278 for every regex", that is, assume /S for every regex. This is similar to -i
1279 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1280 using it. It makes it easier to run collections of tests with and without
1281 study enabled, and thereby test pcre_study() more easily. All the standard
1282 tests are now run with and without -s (but some patterns can be marked as
1283 "never study" - see 20 below).
1284
1285 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1286 restoration of the capturing data to the outer values was not happening
1287 correctly.
1288
1289 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1290 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1291 pattern had matched an empty string, and so incorrectly returned a no
1292 match.
1293
1294 17. There was optimizing code for the last branch of non-capturing parentheses,
1295 and also for the obeyed branch of a conditional subexpression, which used
1296 tail recursion to cut down on stack usage. Unfortunately, now that there is
1297 the possibility of (*THEN) occurring in these branches, tail recursion is
1298 no longer possible because the return has to be checked for (*THEN). These
1299 two optimizations have therefore been removed. [But see 8.20/11 above.]
1300
1301 18. If a pattern containing \R was studied, it was assumed that \R always
1302 matched two bytes, thus causing the minimum subject length to be
1303 incorrectly computed because \R can also match just one byte.
1304
1305 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1306 was incorrectly computed.
1307
1308 20. If /S is present twice on a test pattern in pcretest input, it now
1309 *disables* studying, thereby overriding the use of -s on the command line
1310 (see 14 above). This is necessary for one or two tests to keep the output
1311 identical in both cases.
1312
1313 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1314 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1315
1316 22. When an atomic group that contained a capturing parenthesis was
1317 successfully matched, but the branch in which it appeared failed, the
1318 capturing was not being forgotten if a higher numbered group was later
1319 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1320 group 1 to "a", when in fact it should be unset. This applied to multi-
1321 branched capturing and non-capturing groups, repeated or not, and also to
1322 positive assertions (capturing in negative assertions does not happen
1323 in PCRE) and also to nested atomic groups.
1324
1325 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1326 subject after a captured substring, to make it easier to tell which of a
1327 number of identical substrings has been captured.
1328
1329 24. The way atomic groups are processed by pcre_exec() has been changed so that
1330 if they are repeated, backtracking one repetition now resets captured
1331 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1332 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1333 "aaa". Previously, it would have been "a". As part of this code
1334 refactoring, the way recursive calls are handled has also been changed.
1335
1336 25. If an assertion condition captured any substrings, they were not passed
1337 back unless some other capturing happened later. For example, if
1338 (?(?=(a))a) was matched against "a", no capturing was returned.
1339
1340 26. When studying a pattern that contained subroutine calls or assertions,
1341 the code for finding the minimum length of a possible match was handling
1342 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1343 group 1 called group 2 while simultaneously a separate group 2 called group
1344 1). A stack overflow occurred in this case. I have fixed this by limiting
1345 the recursion depth to 10.
1346
1347 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1348 Fortmann. This supports explicit test numbers on the command line, and has
1349 argument validation and error reporting.
1350
1351 28. An instance of \X with an unlimited repeat could fail if at any point the
1352 first character it looked at was a mark character.
1353
1354 29. Some minor code refactoring concerning Unicode properties and scripts
1355 should reduce the stack requirement of match() slightly.
1356
1357 30. Added the '=' option to pcretest to check the setting of unused capturing
1358 slots at the end of the pattern, which are documented as being -1, but are
1359 not included in the return count.
1360
1361 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1362 compiled something random. Now it gives a compile-time error (as does
1363 Perl).
1364
1365 32. A *MARK encountered during the processing of a positive assertion is now
1366 recorded and passed back (compatible with Perl).
1367
1368 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1369 had alternative anchored branches, the search for a second match in a line
1370 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1371 matched the line "0102" twice. The same bug affected patterns that started
1372 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1373 twice.
1374
1375 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1376 does, and because of capturing effects, quantifying parenthesized
1377 assertions may at times be useful. Quantifiers are now allowed for
1378 parenthesized assertions.
1379
1380 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1381
1382 36. \g was being checked for fancy things in a character class, when it should
1383 just be a literal "g".
1384
1385 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1386 appearance of a nested POSIX class supersedes an apparent external class.
1387 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1388 unescaped square brackets may also appear as part of class names. For
1389 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1390 more like Perl. (But see 8.20/1 above.)
1391
1392 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1393 was because it thought it was \N{name}, which is not supported).
1394
1395 39. Add minix to OS list not supporting the -S option in pcretest.
1396
1397 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1398 cannot analyze patterns in sufficient detail to catch mutual recursions
1399 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1400 subgroup is called recursively as a subpattern for a second time at the
1401 same position in the subject string. In previous releases this might have
1402 been caught by the recursion limit, or it might have run out of stack.
1403
1404 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1405 happen only once. PCRE was, however incorrectly giving a compile time error
1406 "recursive call could loop indefinitely" because it cannot analyze the
1407 pattern in sufficient detail. The compile time test no longer happens when
1408 PCRE is compiling a conditional subpattern, but actual runaway loops are
1409 now caught at runtime (see 40 above).
1410
1411 42. It seems that Perl allows any characters other than a closing parenthesis
1412 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1413 has been changed to be the same.
1414
1415 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1416 as not to get warnings when autogen.sh is called. Also changed
1417 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1418
1419 44. To help people who use pcregrep to scan files containing exceedingly long
1420 lines, the following changes have been made:
1421
1422 (a) The default value of the buffer size parameter has been increased from
1423 8K to 20K. (The actual buffer used is three times this size.)
1424
1425 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1426 PCRE is built.
1427
1428 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1429 to be set at run time.
1430
1431 (d) Numerical values in pcregrep options can be followed by K or M, for
1432 example --buffer-size=50K.
1433
1434 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1435 given and the return code is set to 2.
1436
1437 45. Add a pointer to the latest mark to the callout data block.
1438
1439 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1440 partial match of an empty string instead of no match. This was specific to
1441 the use of ".".
1442
1443 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1444 complete match instead of a partial match. This bug was dependent on both
1445 the PCRE_UTF8 and PCRE_DOTALL options being set.
1446
1447 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1448 starting byte set, because \b was not being ignored.
1449
1450
1451 Version 8.12 15-Jan-2011
1452 ------------------------
1453
1454 1. Fixed some typos in the markup of the man pages, and wrote a script that
1455 checks for such things as part of the documentation building process.
1456
1457 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1458 --match-limit and --recursion-limit options (added for 8.11). In
1459 particular, this made one of the standard tests fail. (The integer value
1460 went into the wrong half of a long int.)
1461
1462 3. If the --colour option was given to pcregrep with -v (invert match), it
1463 did strange things, either producing crazy output, or crashing. It should,
1464 of course, ignore a request for colour when reporting lines that do not
1465 match.
1466
1467 4. Another pcregrep bug caused similar problems if --colour was specified with
1468 -M (multiline) and the pattern match finished with a line ending.
1469
1470 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1471 matched in multiline mode, the following line was shown as part of the
1472 match. This seems wrong, so I have changed it.
1473
1474 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1475 the check for further matches in the same line (so they could be coloured)
1476 to overrun the end of the current line. If another match was found, it was
1477 incorrectly shown (and then shown again when found in the next line).
1478
1479 7. If pcregrep was compiled under Windows, there was a reference to the
1480 function pcregrep_exit() before it was defined. I am assuming this was
1481 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1482 reported by a user. I've moved the definition above the reference.
1483
1484
1485 Version 8.11 10-Dec-2010
1486 ------------------------
1487
1488 1. (*THEN) was not working properly if there were untried alternatives prior
1489 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1490 backtracked to try for "b" instead of moving to the next alternative branch
1491 at the same level (in this case, to look for "c"). The Perl documentation
1492 is clear that when (*THEN) is backtracked onto, it goes to the "next
1493 alternative in the innermost enclosing group".
1494
1495 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1496 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1497 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1498 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1499 (*THEN).
1500
1501 3. If \s appeared in a character class, it removed the VT character from
1502 the class, even if it had been included by some previous item, for example
1503 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1504 of \s, but is part of the POSIX "space" class.)
1505
1506 4. A partial match never returns an empty string (because you can always
1507 match an empty string at the end of the subject); however the checking for
1508 an empty string was starting at the "start of match" point. This has been
1509 changed to the "earliest inspected character" point, because the returned
1510 data for a partial match starts at this character. This means that, for
1511 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1512 (previously it gave "no match").
1513
1514 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1515 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1516 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1517 has an implication that the given string is incomplete (because a partial
1518 match is preferred over a full match). For this reason, these items now
1519 give a partial match in this situation. [Aside: previously, the one case
1520 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1521 match rather than a full match, which was wrong by the old rules, but is
1522 now correct.]
1523
1524 6. There was a bug in the handling of #-introduced comments, recognized when
1525 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1526 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1527 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1528 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1529 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1530 places in pcre_compile().
1531
1532 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1533 comments when looking ahead for named forward references to subpatterns,
1534 the only newline sequence it recognized was NL. It now handles newlines
1535 according to the set newline convention.
1536
1537 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1538 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1539 cater for a lack of strerror(). These oversights have been fixed.
1540
1541 9. Added --match-limit and --recursion-limit to pcregrep.
1542
1543 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1544
1545 11. When the -o option was used, pcregrep was setting a return code of 1, even
1546 when matches were found, and --line-buffered was not being honoured.
1547
1548 12. Added an optional parentheses number to the -o and --only-matching options
1549 of pcregrep.
1550
1551 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1552 can match an empty string. The code to do it in pcretest and pcredemo
1553 needed fixing:
1554
1555 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1556 only one byte after an empty string match just before CRLF (this case
1557 just got forgotten; "any" and "anycrlf" were OK).
1558
1559 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1560 mode when an empty string match preceded an ASCII character followed by
1561 a non-ASCII character. (The code for advancing by one character rather
1562 than one byte was nonsense.)
1563
1564 (c) The pcredemo.c sample program did not have any code at all to handle
1565 the cases when CRLF is a valid newline sequence.
1566
1567 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1568 as a starting offset was within the subject string. There is now a new
1569 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1570 negative or greater than the length of the string. In order to test this,
1571 pcretest is extended to allow the setting of negative starting offsets.
1572
1573 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1574 starting offset points to the beginning of a UTF-8 character was
1575 unnecessarily clumsy. I tidied it up.
1576
1577 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1578 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1579
1580 17. Nobody had reported that the --include_dir option, which was added in
1581 release 7.7 should have been called --include-dir (hyphen, not underscore)
1582 for compatibility with GNU grep. I have changed it to --include-dir, but
1583 left --include_dir as an undocumented synonym, and the same for
1584 --exclude-dir, though that is not available in GNU grep, at least as of
1585 release 2.5.4.
1586
1587 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1588 characters from a string of bytes) have been redefined so as not to use
1589 loops, in order to improve performance in some environments. At the same
1590 time, I abstracted some of the common code into auxiliary macros to save
1591 repetition (this should not affect the compiled code).
1592
1593 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1594 compile-time error is now given if \c is not followed by an ASCII
1595 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1596 different, and any byte value is allowed.)
1597
1598 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1599 START_OPTIMIZE option, which is now allowed at compile time - but just
1600 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1601 to pcregrep and other applications that have no direct access to PCRE
1602 options. The new /Y option in pcretest sets this option when calling
1603 pcre_compile().
1604
1605 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1606 back references. Groups containing recursive back references were forced to
1607 be atomic by that change, but in the case of named groups, the amount of
1608 memory required was incorrectly computed, leading to "Failed: internal
1609 error: code overflow". This has been fixed.
1610
1611 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1612 pcretest.c, to avoid build problems in some Borland environments.
1613
1614
1615 Version 8.10 25-Jun-2010
1616 ------------------------
1617
1618 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1619 THEN.
1620
1621 2. (*ACCEPT) was not working when inside an atomic group.
1622
1623 3. Inside a character class, \B is treated as a literal by default, but
1624 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1625 causes the error). The code is unchanged, but I tidied the documentation.
1626
1627 4. Inside a character class, PCRE always treated \R and \X as literals,
1628 whereas Perl faults them if its -w option is set. I have changed PCRE so
1629 that it faults them when PCRE_EXTRA is set.
1630
1631 5. Added support for \N, which always matches any character other than
1632 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1633
1634 6. When compiling pcregrep with newer versions of gcc which may have
1635 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1636 declared with attribute warn_unused_result" were given. Just casting the
1637 result to (void) does not stop the warnings; a more elaborate fudge is
1638 needed. I've used a macro to implement this.
1639
1640 7. Minor change to pcretest.c to avoid a compiler warning.
1641
1642 8. Added four artifical Unicode properties to help with an option to make
1643 \s etc use properties (see next item). The new properties are: Xan
1644 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1645
1646 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1647 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1648 this option. Modified pcretest to add /W to test this facility. Added
1649 REG_UCP to make it available via the POSIX interface.
1650
1651 10. Added --line-buffered to pcregrep.
1652
1653 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1654 studied, and the match started with a letter with a code point greater than
1655 127 whose first byte was different to the first byte of the other case of
1656 the letter, the other case of this starting letter was not recognized
1657 (#976).
1658
1659 12. If a pattern that was studied started with a repeated Unicode property
1660 test, for example, \p{Nd}+, there was the theoretical possibility of
1661 setting up an incorrect bitmap of starting bytes, but fortunately it could
1662 not have actually happened in practice until change 8 above was made (it
1663 added property types that matched character-matching opcodes).
1664
1665 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1666 possible starting bytes for non-anchored patterns.
1667
1668 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1669 \R, and also a number of cases that involve Unicode properties, both
1670 explicit and implicit when PCRE_UCP is set.
1671
1672 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1673 input, it could crash or give wrong results if characters with values
1674 greater than 0xc0 were present in the subject string. (Detail: it assumed
1675 UTF-8 input when processing these items.)
1676
1677 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1678 size_t is 64-bit (#991).
1679
1680 17. Added a check for running out of memory when PCRE is compiled with
1681 --disable-stack-for-recursion (#990).
1682
1683 18. If the last data line in a file for pcretest does not have a newline on
1684 the end, a newline was missing in the output.
1685
1686 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1687 less than 128) in its various bitmaps. However, there is a facility for
1688 generating tables according to the current locale when PCRE is compiled. It
1689 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1690 characters, are recognized by isspace() and therefore were getting set in
1691 these tables, and indeed these tables seem to approximate to ISO 8859. This
1692 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1693 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1694 which of course cannot start UTF-8 characters. I have changed the code so
1695 that only real ASCII characters (less than 128) and the correct starting
1696 bytes for UTF-8 encodings are set for characters greater than 127 when in
1697 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1698 altogether.)
1699
1700 20. Added the /T option to pcretest so as to be able to run tests with non-
1701 standard character tables, thus making it possible to include the tests
1702 used for 19 above in the standard set of tests.
1703
1704 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1705 reference to a subpattern the other side of a comment that contains an
1706 opening parenthesis caused either an internal compiling error, or a
1707 reference to the wrong subpattern.
1708
1709
1710 Version 8.02 19-Mar-2010
1711 ------------------------
1712
1713 1. The Unicode data tables have been updated to Unicode 5.2.0.
1714
1715 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1716 configured.
1717
1718 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1719 original author of that file, following a query about its status.
1720
1721 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1722 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1723
1724 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1725 quantifier applied to a forward-referencing subroutine call, could compile
1726 incorrect code or give the error "internal error: previously-checked
1727 referenced subpattern not found".
1728
1729 6. Both MS Visual Studio and Symbian OS have problems with initializing
1730 variables to point to external functions. For these systems, therefore,
1731 pcre_malloc etc. are now initialized to local functions that call the
1732 relevant global functions.
1733
1734 7. There were two entries missing in the vectors called coptable and poptable
1735 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1736 I've fixed the data, and added a kludgy way of testing at compile time that
1737 the lengths are correct (equal to the number of opcodes).
1738
1739 8. Following on from 7, I added a similar kludge to check the length of the
1740 eint vector in pcreposix.c.
1741
1742 9. Error texts for pcre_compile() are held as one long string to avoid too
1743 much relocation at load time. To find a text, the string is searched,
1744 counting zeros. There was no check for running off the end of the string,
1745 which could happen if a new error number was added without updating the
1746 string.
1747
1748 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1749
1750 11. \K was not working if it appeared in an atomic group or in a group that
1751 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1752 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1753 the assertion is positive, but not if it is negative.
1754
1755 12. Change 11 fortuitously reduced the size of the stack frame used in the
1756 "match()" function of pcre_exec.c by one pointer. Forthcoming
1757 implementation of support for (*MARK) will need an extra pointer on the
1758 stack; I have reserved it now, so that the stack frame size does not
1759 decrease.
1760
1761 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1762 item in branch that calls a recursion is a subroutine call - as in the
1763 second branch in the above example - was incorrectly given the compile-
1764 time error "recursive call could loop indefinitely" because pcre_compile()
1765 was not correctly checking the subroutine for matching a non-empty string.
1766
1767 14. The checks for overrunning compiling workspace could trigger after an
1768 overrun had occurred. This is a "should never occur" error, but it can be
1769 triggered by pathological patterns such as hundreds of nested parentheses.
1770 The checks now trigger 100 bytes before the end of the workspace.
1771
1772 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1773
1774
1775 Version 8.01 19-Jan-2010
1776 ------------------------
1777
1778 1. If a pattern contained a conditional subpattern with only one branch (in
1779 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1780 computed the wrong minimum data length (which is of course zero for such
1781 subpatterns). This could cause incorrect "no match" results.
1782
1783 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1784 the pattern is reset in the first branch, pcre_compile() failed with
1785 "internal error: code overflow at offset...". This happened only when
1786 the reset was to the original external option setting. (An optimization
1787 abstracts leading options settings into an external setting, which was the
1788 cause of this.)
1789
1790 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1791 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1792 assertion pattern did not match (meaning that the assertion was true), it
1793 was incorrectly treated as false if the SKIP had been reached during the
1794 matching. This also applied to assertions used as conditions.
1795
1796 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1797 assertion subpattern, including such a pattern used as a condition,
1798 unpredictable results occurred, instead of the error return
1799 PCRE_ERROR_DFA_UITEM.
1800
1801 5. The C++ GlobalReplace function was not working like Perl for the special
1802 situation when an empty string is matched. It now does the fancy magic
1803 stuff that is necessary.
1804
1805 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1806 removed. (These were left over from very, very early versions of PCRE.)
1807
1808 7. Some cosmetic changes to the code to make life easier when compiling it
1809 as part of something else:
1810
1811 (a) Change DEBUG to PCRE_DEBUG.
1812
1813 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1814 called "current" as "current_branch", to prevent a collision with the
1815 Linux macro when compiled as a kernel module.
1816
1817 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1818 prevent a collision with the Linux macro when compiled as a kernel
1819 module.
1820
1821 8. In pcre_compile() there are some checks for integer overflows that used to
1822 cast potentially large values to (double). This has been changed to that
1823 when building, a check for int64_t is made, and if it is found, it is used
1824 instead, thus avoiding the use of floating point arithmetic. (There is no
1825 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1826 double.
1827
1828 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1829 2005 (difference between two addresses compared to an unsigned value).
1830
1831 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1832 custom one, because of the following reported problem in Windows:
1833
1834 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1835 under Win32.
1836 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1837 therefore missing the function definition.
1838 - The compiler thus generates a "C" signature for the test function.
1839 - The linker fails to find the "C" function.
1840 - PCRE fails to configure if asked to do so against libbz2.
1841
1842 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1843 messages were output:
1844
1845 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1846 rerunning libtoolize, to keep the correct libtool macros in-tree.
1847 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1848
1849 I have done both of these things.
1850
1851 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1852 most of the time, it *can* run out if it is given a pattern that contains a
1853 runaway infinite recursion. I updated the discussion in the pcrestack man
1854 page.
1855
1856 13. Now that we have gone to the x.xx style of version numbers, the minor
1857 version may start with zero. Using 08 or 09 is a bad idea because users
1858 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1859 interpreted as invalid octal numbers. I've updated the previous comment in
1860 configure.ac, and also added a check that gives an error if 08 or 09 are
1861 used.
1862
1863 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1864 causing partial matching to fail when the end of the subject matched \W
1865 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1866
1867 15. There were some discrepancies between the declarations in pcre_internal.h
1868 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1869 their definitions. The declarations used "const uschar *" and the
1870 definitions used USPTR. Even though USPTR is normally defined as "const
1871 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1872 reported that: "This difference in casting confuses some C++ compilers, for
1873 example, SunCC recognizes above declarations as different functions and
1874 generates broken code for hbpcre." I have changed the declarations to use
1875 USPTR.
1876
1877 16. GNU libtool is named differently on some systems. The autogen.sh script now
1878 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1879 (FreeBSD).
1880
1881 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1882 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1883 comment: "Figure out how to create a longlong from a string: strtoll and
1884 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1885 instance, but it only takes 2 args instead of 3!"
1886
1887 18. A subtle bug concerned with back references has been fixed by a change of
1888 specification, with a corresponding code fix. A pattern such as
1889 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1890 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1891 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1892 same bug. Such groups have to be quantified to be useful, or contained
1893 inside another quantified group. (If there's no repetition, the reference
1894 can never match.) The problem arises because, having left the group and
1895 moved on to the rest of the pattern, a later failure that backtracks into
1896 the group uses the captured value from the final iteration of the group
1897 rather than the correct earlier one. I have fixed this in PCRE by forcing
1898 any group that contains a reference to itself to be an atomic group; that
1899 is, there cannot be any backtracking into it once it has completed. This is
1900 similar to recursive and subroutine calls.
1901
1902
1903 Version 8.00 19-Oct-09
1904 ----------------------
1905
1906 1. The table for translating pcre_compile() error codes into POSIX error codes
1907 was out-of-date, and there was no check on the pcre_compile() error code
1908 being within the table. This could lead to an OK return being given in
1909 error.
1910
1911 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1912 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1913 in a Windows environment.
1914
1915 3. The pcregrep --count option prints the count for each file even when it is
1916 zero, as does GNU grep. However, pcregrep was also printing all files when
1917 --files-with-matches was added. Now, when both options are given, it prints
1918 counts only for those files that have at least one match. (GNU grep just
1919 prints the file name in this circumstance, but including the count seems
1920 more useful - otherwise, why use --count?) Also ensured that the
1921 combination -clh just lists non-zero counts, with no names.
1922
1923 4. The long form of the pcregrep -F option was incorrectly implemented as
1924 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1925 but it seems right to fix it, and I didn't think it was worth preserving
1926 the old behaviour.
1927
1928 5. The command line items --regex=pattern and --regexp=pattern were not
1929 recognized by pcregrep, which required --regex pattern or --regexp pattern
1930 (with a space rather than an '='). The man page documented the '=' forms,
1931 which are compatible with GNU grep; these now work.
1932
1933 6. No libpcreposix.pc file was created for pkg-config; there was just
1934 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1935
1936 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1937 when UCP support is not needed, by modifying the Python script that
1938 generates it from Unicode data files. This should not matter if the module
1939 is correctly used as a library, but I received one complaint about 50K of
1940 unwanted data. My guess is that the person linked everything into his
1941 program rather than using a library. Anyway, it does no harm.
1942
1943 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1944 was a minimum greater than 1 for a wide character in a possessive
1945 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1946 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1947 character. Chaos in the form of incorrect output or a compiling loop could
1948 result.
1949
1950 9. The restrictions on what a pattern can contain when partial matching is
1951 requested for pcre_exec() have been removed. All patterns can now be
1952 partially matched by this function. In addition, if there are at least two
1953 slots in the offset vector, the offset of the earliest inspected character
1954 for the match and the offset of the end of the subject are set in them when
1955 PCRE_ERROR_PARTIAL is returned.
1956
1957 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1958 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1959 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1960 and may be more useful for multi-segment matching.
1961
1962 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1963 used to be given if ever the end of the subject was reached; now it is
1964 given only if matching could not proceed because another character was
1965 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1966 string "Z", which now yields "no match" instead of "partial match". In the
1967 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1968 final character ended with (*FAIL).
1969
1970 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1971 if the pattern had a "must contain" character that was already found in the
1972 earlier partial match, unless partial matching was again requested. For
1973 example, with the pattern /dog.(body)?/, the "must contain" character is
1974 "g". If the first part-match was for the string "dog", restarting with
1975 "sbody" failed. This bug has been fixed.
1976
1977 13. The string returned by pcre_dfa_exec() after a partial match has been
1978 changed so that it starts at the first inspected character rather than the
1979 first character of the match. This makes a difference only if the pattern
1980 starts with a lookbehind assertion or \b or \B (\K is not supported by
1981 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1982 matching functions compatible, and I think it's the right thing to do.
1983
1984 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1985 so that the demonstration program is easily available in environments where
1986 PCRE has not been installed from source.
1987
1988 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1989 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1990 library.
1991
1992 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1993 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1994 is not the first non-POSIX option to be added. Clearly some people find
1995 these options useful.
1996
1997 17. If a caller to the POSIX matching function regexec() passes a non-zero
1998 value for nmatch with a NULL value for pmatch, the value of
1999 nmatch is forced to zero.
2000
2001 18. RunGrepTest did not have a test for the availability of the -u option of
2002 the diff command, as RunTest does. It now checks in the same way as
2003 RunTest, and also checks for the -b option.
2004
2005 19. If an odd number of negated classes containing just a single character
2006 interposed, within parentheses, between a forward reference to a named
2007 subpattern and the definition of the subpattern, compilation crashed with
2008 an internal error, complaining that it could not find the referenced
2009 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2010 [The bug was that it was starting one character too far in when skipping
2011 over the character class, thus treating the ] as data rather than
2012 terminating the class. This meant it could skip too much.]
2013
2014 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2015 /g option in pcretest when the pattern contains \K, which makes it possible
2016 to have an empty string match not at the start, even when the pattern is
2017 anchored. Updated pcretest and pcredemo to use this option.
2018
2019 21. If the maximum number of capturing subpatterns in a recursion was greater
2020 than the maximum at the outer level, the higher number was returned, but
2021 with unset values at the outer level. The correct (outer level) value is
2022 now given.
2023
2024 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2025 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2026 make it do so. The string so far is captured, making this feature
2027 compatible with Perl.
2028
2029 23. The tests have been re-organized, adding tests 11 and 12, to make it
2030 possible to check the Perl 5.10 features against Perl 5.10.
2031
2032 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2033 pattern matches a fixed length string. PCRE did not allow this; now it
2034 does. Neither allows recursion.
2035
2036 25. I finally figured out how to implement a request to provide the minimum
2037 length of subject string that was needed in order to match a given pattern.
2038 (It was back references and recursion that I had previously got hung up
2039 on.) This code has now been added to pcre_study(); it finds a lower bound
2040 to the length of subject needed. It is not necessarily the greatest lower
2041 bound, but using it to avoid searching strings that are too short does give
2042 some useful speed-ups. The value is available to calling programs via
2043 pcre_fullinfo().
2044
2045 26. While implementing 25, I discovered to my embarrassment that pcretest had
2046 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2047 study optimizations had never been tested with that matching function.
2048 Oops. What is worse, even when it was passed study data, there was a bug in
2049 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2050 were also very few tests of studied patterns with pcre_dfa_exec().
2051
2052 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2053 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2054 on the other side of the coin, they are no longer allowed to have different
2055 names, because these cannot be distinguished in PCRE, and this has caused
2056 confusion. (This is a difference from Perl.)
2057
2058 28. When duplicate subpattern names are present (necessarily with different
2059 numbers, as required by 27 above), and a test is made by name in a
2060 conditional pattern, either for a subpattern having been matched, or for
2061 recursion in such a pattern, all the associated numbered subpatterns are
2062 tested, and the overall condition is true if the condition is true for any
2063 one of them. This is the way Perl works, and is also more like the way
2064 testing by number works.
2065
2066
2067 Version 7.9 11-Apr-09
2068 ---------------------
2069
2070 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2071 (pcretest), all targets were linked against these libraries. This included
2072 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2073 libraries. This caused unwanted dependencies to be created. This problem
2074 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2075 pcretest is linked with readline.
2076
2077 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2078 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2079 moved outside it again, because FALSE and TRUE are already defined in AIX,
2080 but BOOL is not.
2081
2082 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2083 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2084
2085 4. The pcregrep documentation said spaces were inserted as well as colons (or
2086 hyphens) following file names and line numbers when outputting matching
2087 lines. This is not true; no spaces are inserted. I have also clarified the
2088 wording for the --colour (or --color) option.
2089
2090 5. In pcregrep, when --colour was used with -o, the list of matching strings
2091 was not coloured; this is different to GNU grep, so I have changed it to be
2092 the same.
2093
2094 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2095 each matching line was coloured. Now it goes on to look for further matches
2096 of any of the test patterns, which is the same behaviour as GNU grep.
2097
2098 7. A pattern that could match an empty string could cause pcregrep to loop; it
2099 doesn't make sense to accept an empty string match in pcregrep, so I have
2100 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2101 seems to be how GNU grep behaves. [But see later change 40 for release
2102 8.33.]
2103
2104 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2105 start or after a newline", because the conditional assertion was not being
2106 correctly handled. The rule now is that both the assertion and what follows
2107 in the first alternative must satisfy the test.
2108
2109 9. If auto-callout was enabled in a pattern with a conditional group whose
2110 condition was an assertion, PCRE could crash during matching, both with
2111 pcre_exec() and pcre_dfa_exec().
2112
2113 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2114 used for matching.
2115
2116 11. Unicode property support in character classes was not working for
2117 characters (bytes) greater than 127 when not in UTF-8 mode.
2118
2119 12. Added the -M command line option to pcretest.
2120
2121 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2122
2123 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2124
2125 16. Added comments and documentation about mis-use of no_arg in the C++
2126 wrapper.
2127
2128 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2129 from Martin Jerabek that uses macro names for all relevant character and
2130 string constants.
2131
2132 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2133 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2134 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2135 these, but not everybody uses configure.
2136
2137 19. A conditional group that had only one branch was not being correctly
2138 recognized as an item that could match an empty string. This meant that an
2139 enclosing group might also not be so recognized, causing infinite looping
2140 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2141 with the subject "ab", where knowledge that the repeated group can match
2142 nothing is needed in order to break the loop.
2143
2144 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2145 exec(), but without supplying a callout function, matching went wrong.
2146
2147 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2148 leak if the size of the offset vector was greater than 30. When the vector
2149 is smaller, the saved offsets during recursion go onto a local stack
2150 vector, but for larger vectors malloc() is used. It was failing to free
2151 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2152 error, in fact).
2153
2154 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2155 heapframe that is used only when UTF-8 support is enabled. This caused no
2156 problem, but was untidy.
2157
2158 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2159 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2160 included within another project.
2161
2162 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2163 slightly modified by me:
2164
2165 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2166 not building pcregrep.
2167
2168 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2169 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2170
2171 25. Forward references, both numeric and by name, in patterns that made use of
2172 duplicate group numbers, could behave incorrectly or give incorrect errors,
2173 because when scanning forward to find the reference group, PCRE was not
2174 taking into account the duplicate group numbers. A pattern such as
2175 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2176
2177 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2178 the feature of a custom pointer more persuasive (as requested by a user).
2179
2180 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2181 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2182 pre-defined.
2183
2184 28. Added support for (*UTF8) at the start of a pattern.
2185
2186 29. Arrange for flags added by the "release type" setting in CMake to be shown
2187 in the configuration summary.
2188
2189
2190 Version 7.8 05-Sep-08
2191 ---------------------
2192
2193 1. Replaced UCP searching code with optimized version as implemented for Ad
2194 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2195 stage table and inline lookup instead of a function, giving speed ups of 2
2196 to 5 times on some simple patterns that I tested. Permission was given to
2197 distribute the MultiStage2.py script that generates the tables (it's not in
2198 the tarball, but is in the Subversion repository).
2199
2200 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2201 scripts.
2202
2203 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2204 a group with a zero qualifier. The result of the study could be incorrect,
2205 or the function might crash, depending on the pattern.
2206
2207 4. Caseless matching was not working for non-ASCII characters in back
2208 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2209 It now works when Unicode Property Support is available.
2210
2211 5. In pcretest, an escape such as \x{de} in the data was always generating
2212 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2213 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2214 truncation.
2215
2216 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2217
2218 7. Added two (int) casts to pcregrep when printing the difference of two
2219 pointers, in case they are 64-bit values.
2220
2221 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2222 test 2 if it fails.
2223
2224 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2225 and a #define of that name to empty if it is not externally set. This is to
2226 allow users of MSVC to set it if necessary.
2227
2228 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2229 the convenience functions in the pcre_get.c source file.
2230
2231 11. An option change at the start of a pattern that had top-level alternatives
2232 could cause overwriting and/or a crash. This command provoked a crash in
2233 some environments:
2234
2235 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2236
2237 This potential security problem was recorded as CVE-2008-2371.
2238
2239 12. For a pattern where the match had to start at the beginning or immediately
2240 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2241 pcre_dfa_exec() could read past the end of the passed subject if there was
2242 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2243 pcretest so that it places the subject at the end of its malloc-ed buffer.
2244
2245 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2246 exec() might read past the end of the data buffer in UTF-8 mode.
2247
2248 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2249 the data contained the byte 0x85 as part of a UTF-8 character within its
2250 first line. This applied both to normal and DFA matching.
2251
2252 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2253 /^[^d]*?$/8 failed to match "abc".
2254
2255 16. Added a missing copyright notice to pcrecpp_internal.h.
2256
2257 17. Make it more clear in the documentation that values returned from
2258 pcre_exec() in ovector are byte offsets, not character counts.
2259
2260 18. Tidied a few places to stop certain compilers from issuing warnings.
2261
2262 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2263 supplied by Stefan Weber. I made a further small update for 7.8 because
2264 there is a change of source arrangements: the pcre_searchfuncs.c module is
2265 replaced by pcre_ucd.c.
2266
2267
2268 Version 7.7 07-May-08
2269 ---------------------
2270
2271 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2272 a string to a long long, pretend we don't even have a long long." This is
2273 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2274
2275 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2276 pre-7.6 versions, which defined a global no_arg variable instead of putting
2277 it in the RE class. (See also #8 below.)
2278
2279 3. Remove a line of dead code, identified by coverity and reported by Nuno
2280 Lopes.
2281
2282 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2283
2284 (1) The include/exclude patterns were being applied to the whole pathnames
2285 of files, instead of just to the final components.
2286
2287 (2) If there was more than one level of directory, the subdirectories were
2288 skipped unless they satisfied the include/exclude conditions. This is
2289 inconsistent with GNU grep (and could even be seen as contrary to the
2290 pcregrep specification - which I improved to make it absolutely clear).
2291 The action now is always to scan all levels of directory, and just
2292 apply the include/exclude patterns to regular files.
2293
2294 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2295 --exclude_dir in the tests to avoid scanning .svn directories.
2296
2297 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2298 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2299 doesn't support NULs in patterns.
2300
2301 7. Added some missing "const"s to declarations of static tables in
2302 pcre_compile.c and pcre_dfa_exec.c.
2303
2304 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2305 caused by fix #2 above. (Subsequently also a second patch to fix the
2306 first patch. And a third patch - this was a messy problem.)
2307
2308 9. Applied Craig's patch to remove the use of push_back().
2309
2310 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2311 matching function regexec().
2312
2313 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2314 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2315 references. PCRE supports relative numbers with this syntax (I don't think
2316 Oniguruma does).
2317
2318 12. Previously, a group with a zero repeat such as (...){0} was completely
2319 omitted from the compiled regex. However, this means that if the group
2320 was called as a subroutine from elsewhere in the pattern, things went wrong
2321 (an internal error was given). Such groups are now left in the compiled
2322 pattern, with a new opcode that causes them to be skipped at execution
2323 time.
2324
2325 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2326 to the way PCRE behaves:
2327
2328 (a) A lone ] character is dis-allowed (Perl treats it as data).
2329
2330 (b) A back reference to an unmatched subpattern matches an empty string
2331 (Perl fails the current match path).
2332
2333 (c) A data ] in a character class must be notated as \] because if the
2334 first data character in a class is ], it defines an empty class. (In
2335 Perl it is not possible to have an empty class.) The empty class []
2336 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2337 The negative empty class [^] matches any one character, independently
2338 of the DOTALL setting.
2339
2340 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2341 non-existent subpattern following a character class starting with ']' and
2342 containing () gave an internal compiling error instead of "reference to
2343 non-existent subpattern". Fortunately, when the pattern did exist, the
2344 compiled code was correct. (When scanning forwards to check for the
2345 existence of the subpattern, it was treating the data ']' as terminating
2346 the class, so got the count wrong. When actually compiling, the reference
2347 was subsequently set up correctly.)
2348
2349 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2350 it was being rejected as not supported by pcre_dfa_exec(), even though
2351 other assertions are supported. I have made pcre_dfa_exec() support
2352 (*FAIL).
2353
2354 16. The implementation of 13c above involved the invention of a new opcode,
2355 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2356 cannot be changed at match time, I realized I could make a small
2357 improvement to matching performance by compiling OP_ALLANY instead of
2358 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2359 on the OP_ANY path.
2360
2361 17. Compiling pcretest on Windows with readline support failed without the
2362 following two fixes: (1) Make the unistd.h include conditional on
2363 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2364
2365 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2366 ncurses library to be included for pcretest when ReadLine support is
2367 requested, but also to allow for it to be overridden. This patch came from
2368 Daniel Bergström.
2369
2370 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2371 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2372 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2373 spotting this.
2374
2375
2376 Version 7.6 28-Jan-08
2377 ---------------------
2378
2379 1. A character class containing a very large number of characters with
2380 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2381 overflow.
2382
2383 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2384 HAVE_LONG_LONG is not defined.
2385
2386 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2387 bring them up to date and include new features. This patch includes:
2388
2389 - Fixed PH's badly added libz and libbz2 support.
2390 - Fixed a problem with static linking.
2391 - Added pcredemo. [But later removed - see 7 below.]
2392 - Fixed dftables problem and added an option.
2393 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2394 HAVE_LONG_LONG.
2395 - Added readline support for pcretest.
2396 - Added an listing of the option settings after cmake has run.
2397
2398 4. A user submitted a patch to Makefile that makes it easy to create
2399 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2400 Makefile.am that cause it to include this special target, without
2401 affecting anything else. Note that the same mingw target plus all
2402 the other distribution libraries and programs are now supported
2403 when configuring with CMake (see 6 below) instead of with
2404 Configure/Make.
2405
2406 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2407 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2408 exported in the Windows port". It has not yet been confirmed that the patch
2409 solves the problem, but it does no harm.
2410
2411 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2412 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2413 with CMake, and also correct the comment about stack recursion.
2414
2415 7. Remove the automatic building of pcredemo from the ./configure system and
2416 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2417 of a program that users should build themselves after PCRE is installed, so
2418 building it automatically is not really right. What is more, it gave
2419 trouble in some build environments.
2420
2421 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2422
2423
2424 Version 7.5 10-Jan-08
2425 ---------------------
2426
2427 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2428 values in parens when parsing an RE using the C++ wrapper."
2429
2430 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2431 Characters greater than 255 were excluded from the class instead of being
2432 included.
2433
2434 3. The same bug as (2) above applied to negated POSIX classes such as
2435 [:^space:].
2436
2437 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2438 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2439 I have changed it.
2440
2441 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2442 first named subpattern) and a construct such as (?&a) would reference the
2443 first named subpattern whose name started with "a" (in other words, the
2444 length check was missing). Both these problems are fixed. "Subpattern name
2445 expected" is now given for (?&) (a zero-length name), and this patch also
2446 makes it give the same error for \k'' (previously it complained that that
2447 was a reference to a non-existent subpattern).
2448
2449 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2450 this is right because (?- can be followed by option settings as well as by
2451 digits. I have, however, made the messages clearer.
2452
2453 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2454 than the number used in the conditional) now cause a compile-time error.
2455 This is actually not compatible with Perl, which accepts such patterns, but
2456 treats the conditional as always being FALSE (as PCRE used to), but it
2457 seems to me that giving a diagnostic is better.
2458
2459 8. Change "alphameric" to the more common word "alphanumeric" in comments
2460 and messages.
2461
2462 9. Fix two occurrences of "backslash" in comments that should have been
2463 "backspace".
2464
2465 10. Remove two redundant lines of code that can never be obeyed (their function
2466 was moved elsewhere).
2467
2468 11. The program that makes PCRE's Unicode character property table had a bug
2469 which caused it to generate incorrect table entries for sequences of
2470 characters that have the same character type, but are in different scripts.
2471 It amalgamated them into a single range, with the script of the first of
2472 them. In other words, some characters were in the wrong script. There were
2473 thirteen such cases, affecting characters in the following ranges:
2474
2475 U+002b0 - U+002c1
2476 U+0060c - U+0060d
2477 U+0061e - U+00612
2478 U+0064b - U+0065e
2479 U+0074d - U+0076d
2480 U+01800 - U+01805
2481 U+01d00 - U+01d77
2482 U+01d9b - U+01dbf
2483 U+0200b - U+0200f
2484 U+030fc - U+030fe
2485 U+03260 - U+0327f
2486 U+0fb46 - U+0fbb1
2487 U+10450 - U+1049d
2488
2489 12. The -o option (show only the matching part of a line) for pcregrep was not
2490 compatible with GNU grep in that, if there was more than one match in a
2491 line, it showed only the first of them. It now behaves in the same way as
2492 GNU grep.
2493
2494 13. If the -o and -v options were combined for pcregrep, it printed a blank
2495 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2496 does the same. The return code can be used to tell if there were any
2497 non-matching lines.
2498
2499 14. Added --file-offsets and --line-offsets to pcregrep.
2500
2501 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2502 infinitely looping recursion. The bug was that positive lookaheads were not
2503 being skipped when checking for a possible empty match (negative lookaheads
2504 and both kinds of lookbehind were skipped).
2505
2506 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2507 inclusion of <windows.h> to before rather than after the definition of
2508 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2509
2510 17. Specifying a possessive quantifier with a specific limit for a Unicode
2511 character property caused pcre_compile() to compile bad code, which led at
2512 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2513 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2514 caused the error; without that there was no problem.
2515
2516 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2517
2518 19. Added --enable-pcretest-libreadline.
2519
2520 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2521 RE::GlobalReplace(). As a result, the number of replacements returned was
2522 double what it should be. I removed one of the increments, but Craig sent a
2523 later patch that removed the other one (the right fix) and added unit tests
2524 that check the return values (which was not done before).
2525
2526 21. Several CMake things:
2527
2528 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2529 the names libpcre and libpcreposix, not just pcre and pcreposix.
2530
2531 (2) The above change means that pcretest and pcregrep are now correctly
2532 linked with the newly-built libraries, not previously installed ones.
2533
2534 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2535
2536 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2537 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2538 UTF-8 newline character). The key issue is that the pattern starts .*;
2539 this means that the match must be either at the beginning, or after a
2540 newline. The bug was in the code for advancing after a failed match and
2541 checking that the new position followed a newline. It was not taking
2542 account of UTF-8 characters correctly.
2543
2544 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2545 character classes. PCRE was not treating the sequence [:...:] as a
2546 character class unless the ... were all letters. Perl, however, seems to
2547 allow any characters between [: and :], though of course it rejects as
2548 unknown any "names" that contain non-letters, because all the known class
2549 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2550 for example, whereas PCRE did not - it did not recognize a POSIX character
2551 class. This seemed a bit dangerous, so the code has been changed to be
2552 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2553 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2554 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2555 Perl does, and where it didn't before.
2556
2557 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2558 Windows environments %n is disabled by default.
2559
2560
2561 Version 7.4 21-Sep-07
2562 ---------------------
2563
2564 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2565 means that a class such as [\s] counted as "explicit reference to CR or
2566 LF". That isn't really right - the whole point of the change was to try to
2567 help when there was an actual mention of one of the two characters. So now
2568 the change happens only if \r or \n (or a literal CR or LF) character is
2569 encountered.
2570
2571 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2572 of both had grown to the point where there were only 3 bits left.
2573 Fortunately, there was spare space in the data structure, and so I have
2574 moved the internal flags into a new 16-bit field to free up more option
2575 bits.
2576
2577 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2578 but did not set the internal JCHANGED flag - either of these is enough to
2579 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2580 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2581 start sets both bits.
2582
2583 4. Added options (at build time, compile time, exec time) to change \R from
2584 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2585
2586 5. doc/pcresyntax.html was missing from the distribution.
2587
2588 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2589 compatibility, even though it is no longer used.
2590
2591 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2592 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2593 windows.h file is present (where different names are used). [This was
2594 reversed later after testing - see 16 below.]
2595
2596 8. Changed all #include <config.h> to #include "config.h". There were also
2597 some further <pcre.h> cases that I changed to "pcre.h".
2598
2599 9. When pcregrep was used with the --colour option, it missed the line ending
2600 sequence off the lines that it output.
2601
2602 10. It was pointed out to me that arrays of string pointers cause lots of
2603 relocations when a shared library is dynamically loaded. A technique of
2604 using a single long string with a table of offsets can drastically reduce
2605 these. I have refactored PCRE in four places to do this. The result is
2606 dramatic:
2607
2608 Originally: 290
2609 After changing UCP table: 187
2610 After changing error message table: 43
2611 After changing table of "verbs" 36
2612 After changing table of Posix names 22
2613
2614 Thanks to the folks working on Gregex for glib for this insight.
2615
2616 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2617 unicode-properties was also set.
2618
2619 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2620
2621 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2622 checked only for CRLF.
2623
2624 14. Added casts to pcretest.c to avoid compiler warnings.
2625
2626 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2627
2628 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2629 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2630 entirely. This removes changes made in 7 above.
2631
2632 17. The CMake files have been updated, and there is now more information about
2633 building with CMake in the NON-UNIX-USE document.
2634
2635
2636 Version 7.3 28-Aug-07
2637 ---------------------
2638
2639 1. In the rejigging of the build system that eventually resulted in 7.1, the
2640 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2641 brackets there is not right, since it causes compilers to look for an
2642 installed pcre.h, not the version that is in the source that is being
2643 compiled (which of course may be different). I have changed it back to:
2644
2645 #include "pcre.h"
2646
2647 I have a vague recollection that the change was concerned with compiling in
2648 different directories, but in the new build system, that is taken care of
2649 by the VPATH setting the Makefile.
2650
2651 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2652 when the subject happened to end in the byte 0x85 (e.g. if the last
2653 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2654 characters but of course it shouldn't be taken as a newline when it is part
2655 of another character. The bug was that, for an unlimited repeat of . in
2656 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2657 characters when looking for a newline.
2658
2659 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2660
2661 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2662 in debug output.
2663
2664 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2665 long printing in the pcrecpp unittest when running under MinGW.
2666
2667 6. ESC_K was left out of the EBCDIC table.
2668
2669 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2670 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2671 limit also applies to "virtual nesting" when a pattern is recursive, and in
2672 this case 1000 isn't so big. I have been able to remove this limit at the
2673 expense of backing off one optimization in certain circumstances. Normally,
2674 when pcre_exec() would call its internal match() function recursively and
2675 immediately return the result unconditionally, it uses a "tail recursion"
2676 feature to save stack. However, when a subpattern that can match an empty
2677 string has an unlimited repetition quantifier, it no longer makes this
2678 optimization. That gives it a stack frame in which to save the data for
2679 checking that an empty string has been matched. Previously this was taken
2680 from the 1000-entry workspace that had been reserved. So now there is no
2681 explicit limit, but more stack is used.
2682
2683 8. Applied Daniel's patches to solve problems with the import/export magic
2684 syntax that is required for Windows, and which was going wrong for the
2685 pcreposix and pcrecpp parts of the library. These were overlooked when this
2686 problem was solved for the main library.
2687
2688 9. There were some crude static tests to avoid integer overflow when computing
2689 the size of patterns that contain repeated groups with explicit upper
2690 limits. As the maximum quantifier is 65535, the maximum group length was
2691 set at 30,000 so that the product of these two numbers did not overflow a
2692 32-bit integer. However, it turns out that people want to use groups that
2693 are longer than 30,000 bytes (though not repeat them that many times).
2694 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2695 made it possible to implement the integer overflow checks in a much more
2696 dynamic way, which I have now done. The artificial limitation on group
2697 length has been removed - we now have only the limit on the total length of
2698 the compiled pattern, which depends on the LINK_SIZE setting.
2699
2700 10. Fixed a bug in the documentation for get/copy named substring when
2701 duplicate names are permitted. If none of the named substrings are set, the
2702 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2703 empty string.
2704
2705 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2706 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2707 because the ] is interpreted as the first data character and the
2708 terminating ] is not found. PCRE has been made compatible with Perl in this
2709 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2710 cause memory overwriting.
2711
2712 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2713 string has been matched (to stop an infinite loop). It was not recognizing
2714 a conditional subpattern that could match an empty string if that
2715 subpattern was within another subpattern. For example, it looped when
2716 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2717 condition was not nested. This bug has been fixed.
2718
2719 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2720 past the start of the subject in the presence of bytes with the top bit
2721 set, for example "\x8aBCD".
2722
2723 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2724 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2725
2726 14. Optimized (?!) to (*FAIL).
2727
2728 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2729 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2730 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2731 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2732 does: it's just the validity check that is more restrictive.
2733
2734 16. Inserted checks for integer overflows during escape sequence (backslash)
2735 processing, and also fixed erroneous offset values for syntax errors during
2736 backslash processing.
2737
2738 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2739 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2740
2741 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2742 caused an overrun.
2743
2744 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2745 something other than just ASCII characters) inside a group that had an
2746 unlimited repeat caused a loop at compile time (while checking to see
2747 whether the group could match an empty string).
2748
2749 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2750 [\P{Any}] did so. (Error in the code for printing property names.)
2751
2752 21. An orphan \E inside a character class could cause a crash.
2753
2754 22. A repeated capturing bracket such as (A)? could cause a wild memory
2755 reference during compilation.
2756
2757 23. There are several functions in pcre_compile() that scan along a compiled
2758 expression for various reasons (e.g. to see if it's fixed length for look
2759 behind). There were bugs in these functions when a repeated \p or \P was
2760 present in the pattern. These operators have additional parameters compared
2761 with \d, etc, and these were not being taken into account when moving along
2762 the compiled data. Specifically:
2763
2764 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2765 length.
2766
2767 (b) An item such as \pL+ within a repeated group could cause crashes or
2768 loops.
2769
2770 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2771 "reference to non-existent subpattern" error.
2772
2773 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2774
2775 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2776 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2777
2778 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2779
2780 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2781 character were causing crashes (broken optimization).
2782
2783 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2784 \p or \P) caused a compile-time loop.
2785
2786 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2787 break. For example, the unstudied pattern [\r\n]A does not match the string
2788 "\r\nA" because change 7.0/46 below moves the current point on by two
2789 characters after failing to match at the start. However, the pattern \nA
2790 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2791 the same is true. There doesn't seem any very clean way out of this, but
2792 what I have chosen to do makes the common cases work: PCRE now takes note
2793 of whether there can be an explicit match for \r or \n anywhere in the
2794 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2795 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2796 pattern has explicit CR or LF references.
2797
2798 29. Added (*CR) etc for changing newline setting at start of pattern.
2799
2800
2801 Version 7.2 19-Jun-07
2802 ---------------------
2803
2804 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2805 which is apparently normally available under Windows.
2806
2807 2. Re-jig the pcregrep tests with different newline settings in an attempt
2808 to make them independent of the local environment's newline setting.
2809
2810 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2811
2812 4. Some of the "internals" tests were previously cut out when the link size
2813 was not 2, because the output contained actual offsets. The recent new
2814 "Z" feature of pcretest means that these can be cut out, making the tests
2815 usable with all link sizes.
2816
2817 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2818 stack recursion. This gives a massive performance boost under BSD, but just
2819 a small improvement under Linux. However, it saves one field in the frame
2820 in all cases.
2821
2822 6. Added more features from the forthcoming Perl 5.10:
2823
2824 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2825 recursion call. It refers to the nth most recently opened parentheses.
2826
2827 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2828 to be opened parentheses.
2829
2830 (c) Conditions that refer to capturing parentheses can be specified
2831 relatively, for example, (?(-2)... or (?(+3)...
2832
2833 (d) \K resets the start of the current match so that everything before
2834 is not part of it.
2835
2836 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2837
2838 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2839 reference syntax.
2840
2841 (g) (?| introduces a group in which the numbering of parentheses in each
2842 alternative starts with the same number.
2843
2844 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2845
2846 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2847 PCRE_INFO_JCHANGED.
2848
2849 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2850 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2851 for detecting groups that can match an empty string.
2852
2853 9. A pattern with a very large number of alternatives (more than several
2854 hundred) was running out of internal workspace during the pre-compile
2855 phase, where pcre_compile() figures out how much memory will be needed. A
2856 bit of new cunning has reduced the workspace needed for groups with
2857 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2858 workspace instead of running out of the 4096 that are available.
2859
2860 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2861
2862 11. Applied patch from Google to remove an optimization that didn't quite work.
2863 The report of the bug said:
2864
2865 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2866 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2867 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2868
2869 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2870 it matched the wrong number of bytes.
2871
2872
2873 Version 7.1 24-Apr-07
2874 ---------------------
2875
2876 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2877 that is more "standard", making use of automake and other Autotools. There
2878 is some re-arrangement of the files and adjustment of comments consequent
2879 on this.
2880
2881 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2882 for recursive directory scanning broke on some systems because the files
2883 are not scanned in any specific order and on different systems the order
2884 was different. A call to "sort" has been inserted into RunGrepTest for the
2885 approprate test as a short-term fix. In the longer term there may be an
2886 alternative.
2887
2888 3. I had an email from Eric Raymond about problems translating some of PCRE's
2889 man pages to HTML (despite the fact that I distribute HTML pages, some
2890 people do their own conversions for various reasons). The problems
2891 concerned the use of low-level troff macros .br and .in. I have therefore
2892 removed all such uses from the man pages (some were redundant, some could
2893 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2894 HTML has been updated to handle .nf/.fi and to complain if it encounters
2895 .br or .in.
2896
2897 4. Updated comments in configure.ac that get placed in config.h.in and also
2898 arranged for config.h to be included in the distribution, with the name
2899 config.h.generic, for the benefit of those who have to compile without
2900 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2901
2902 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2903 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2904 makevp.bat for the current PCRE, using the additional files
2905 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2906
2907 6. A Windows user reported a minor discrepancy with test 2, which turned out
2908 to be caused by a trailing space on an input line that had got lost in his
2909 copy. The trailing space was an accident, so I've just removed it.
2910
2911 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2912 that is needed.
2913
2914 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2915 as "const" (a) because they are and (b) because it helps the PHP
2916 maintainers who have recently made a script to detect big data structures
2917 in the php code that should be moved to the .rodata section. I remembered
2918 to update Builducptable as well, so it won't revert if ucptable.h is ever
2919 re-created.
2920
2921 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2922 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2923 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2924 support is not required. This saves 1.5-2K of code, which is important in
2925 some applications.
2926
2927 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2928 so as not to refer to the tables, even though these functions will never be
2929 called when UTF-8 support is disabled. Otherwise there are problems with a
2930 shared library.
2931
2932 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2933
2934 (a) It was defining its arguments as char * instead of void *.
2935
2936 (b) It was assuming that all moves were upwards in memory; this was true
2937 a long time ago when I wrote it, but is no longer the case.
2938
2939 The emulated memove() is provided for those environments that have neither
2940 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2941 is clearly not the case, as these two bugs were recently reported.
2942
2943 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2944 and Detrail to create the HTML documentation, the .txt form of the man
2945 pages, and it removes trailing spaces from listed files. It also creates
2946 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2947 case, it wraps all the #defines with #ifndefs. This script should be run
2948 before "make dist".
2949
2950 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2951 with Unicode property support.
2952
2953 (a) For a maximizing quantifier, if the two different cases of the
2954 character were of different lengths in their UTF-8 codings (there are
2955 some cases like this - I found 11), and the matching function had to
2956 back up over a mixture of the two cases, it incorrectly assumed they
2957 were both the same length.
2958
2959 (b) When PCRE was configured to use the heap rather than the stack for
2960 recursion during matching, it was not correctly preserving the data for
2961 the other case of a UTF-8 character when checking ahead for a match
2962 while processing a minimizing repeat. If the check also involved
2963 matching a wide character, but failed, corruption could cause an
2964 erroneous result when trying to check for a repeat of the original
2965 character.
2966
2967 13. Some tidying changes to the testing mechanism:
2968
2969 (a) The RunTest script now detects the internal link size and whether there
2970 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2971 values substituted by "configure". (The RunGrepTest script already did
2972 this for UTF-8.) The configure.ac script no longer substitutes the
2973 relevant variables.
2974
2975 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2976 with length and offset values. This means that the output is different
2977 for different internal link sizes. Test 2 is skipped for link sizes
2978 other than 2 because of this, bypassing the problem. Unfortunately,
2979 there was also a test in test 3 (the locale tests) that used /B and
2980 failed for link sizes other than 2. Rather than cut the whole test out,
2981 I have added a new /Z option to pcretest that replaces the length and
2982 offset values with spaces. This is now used to make test 3 independent
2983 of link size. (Test 2 will be tidied up later.)
2984
2985 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2986 segmentation fault instead of returning the appropriate error message.
2987
2988 15. In multiline mode when the newline sequence was set to "any", the pattern
2989 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2990 This doesn't seem right; it now treats the CRLF combination as the line
2991 ending, and so does not match in that case. It's only a pattern such as ^$
2992 that would hit this one: something like ^ABC$ would have failed after \r
2993 and then tried again after \r\n.
2994
2995 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2996 in an attempt to make files that differ only in their line terminators
2997 compare equal. This works on Linux.
2998
2999 17. Under certain error circumstances pcregrep might try to free random memory
3000 as it exited. This is now fixed, thanks to valgrind.
3001
3002 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3003 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3004 was because its rules for how to advance for /g after matching an empty
3005 string at the end of a line did not allow for this case. They now check for
3006 it specially.
3007
3008 20. pcretest is supposed to handle patterns and data of any length, by
3009 extending its buffers when necessary. It was getting this wrong when the
3010 buffer for a data line had to be extended.
3011
3012 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3013 CRLF as a newline sequence.
3014
3015 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3016 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3017 I have nevertheless tidied it up.
3018
3019 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3020
3021 24. Added a man page for pcre-config.
3022
3023
3024 Version 7.0 19-Dec-06
3025 ---------------------
3026
3027 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3028 moving to gcc 4.1.1.
3029
3030 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3031 sys/time.h, which is documented as needed for this function. It doesn't
3032 seem to matter on Linux, but it showed up on some releases of OS X.
3033
3034 3. It seems that there are systems where bytes whose values are greater than
3035 127 match isprint() in the "C" locale. The "C" locale should be the
3036 default when a C program starts up. In most systems, only ASCII printing
3037 characters match isprint(). This difference caused the output from pcretest
3038 to vary, making some of the tests fail. I have changed pcretest so that:
3039
3040 (a) When it is outputting text in the compiled version of a pattern, bytes
3041 other than 32-126 are always shown as hex escapes.
3042
3043 (b) When it is outputting text that is a matched part of a subject string,
3044 it does the same, unless a different locale has been set for the match
3045 (using the /L modifier). In this case, it uses isprint() to decide.
3046
3047 4. Fixed a major bug that caused incorrect computation of the amount of memory
3048 required for a compiled pattern when options that changed within the
3049 pattern affected the logic of the preliminary scan that determines the
3050 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3051 that the computed length was too small. The symptoms of this bug were
3052 either the PCRE error "internal error: code overflow" from pcre_compile(),
3053 or a glibc crash with a message such as "pcretest: free(): invalid next
3054 size (fast)". Examples of patterns that provoked this bug (shown in
3055 pcretest format) are:
3056
3057 /(?-x: )/x
3058 /(?x)(?-x: \s*#\s*)/
3059 /((?i)[\x{c0}])/8
3060 /(?i:[\x{c0}])/8
3061
3062 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3063 is now done differently.
3064
3065 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3066 wrapper classes; (b) implement a new function in the C++ scanner that is
3067 more efficient than the old way of doing things because it avoids levels of
3068 recursion in the regex matching; (c) add a paragraph to the documentation
3069 for the FullMatch() function.
3070
3071 6. The escape sequence \n was being treated as whatever was defined as
3072 "newline". Not only was this contrary to the documentation, which states
3073 that \n is character 10 (hex 0A), but it also went horribly wrong when
3074 "newline" was defined as CRLF. This has been fixed.
3075
3076 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3077 was being set to -1 for the "end of line" case (supposedly a value that no
3078 character can have). Though this value is never used (the check for end of
3079 line is "zero bytes in current character"), it caused compiler complaints.
3080 I've changed it to 0xffffffff.
3081
3082 8. In pcre_version.c, the version string was being built by a sequence of
3083 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3084 string (as it is for production releases) called a macro with an empty
3085 argument. The C standard says the result of this is undefined. The gcc
3086 compiler treats it as an empty string (which was what was wanted) but it is
3087 reported that Visual C gives an error. The source has been hacked around to
3088 avoid this problem.
3089
3090 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3091 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3092 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3093 of them did).
3094
3095 10. Originally, pcretest opened its input and output without "b"; then I was
3096 told that "b" was needed in some environments, so it was added for release
3097 5.0 to both the input and output. (It makes no difference on Unix-like
3098 systems.) Later I was told that it is wrong for the input on Windows. I've
3099 now abstracted the modes into two macros, to make it easier to fiddle with
3100 them, and removed "b" from the input mode under Windows.
3101
3102 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3103
3104 12. Added -help and --help to pcretest as an official way of being reminded
3105 of the options.
3106
3107 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3108 and pcrecpp.cc because they annoy compilers at high warning levels.
3109
3110 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3111
3112 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3113 scripts are not C programs :-) and which was not noticed because it works
3114 on Linux.
3115
3116 16. pcretest is supposed to handle any length of pattern and data line (as one
3117 line or as a continued sequence of lines) by extending its input buffer if
3118 necessary. This feature was broken for very long pattern lines, leading to
3119 a string of junk being passed to pcre_compile() if the pattern was longer
3120 than about 50K.
3121
3122 17. I have done a major re-factoring of the way pcre_compile() computes the
3123 amount of memory needed for a compiled pattern. Previously, there was code
3124 that made a preliminary scan of the pattern in order to do this. That was
3125 OK when PCRE was new, but as the facilities have expanded, it has become
3126 harder and harder to keep it in step with the real compile phase, and there
3127 have been a number of bugs (see for example, 4 above). I have now found a
3128 cunning way of running the real compile function in a "fake" mode that
3129 enables it to compute how much memory it would need, while actually only
3130 ever using a few hundred bytes of working memory and without too many
3131 tests of the mode. This should make future maintenance and development
3132 easier. A side effect of this work is that the limit of 200 on the nesting
3133 depth of parentheses has been removed (though this was never a serious
3134 limitation, I suspect). However, there is a downside: pcre_compile() now
3135 runs more slowly than before (30% or more, depending on the pattern). I
3136 hope this isn't a big issue. There is no effect on runtime performance.
3137
3138 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3139 newline (only possible for the last line of a file) and it was a
3140 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3141
3142 19. Added additional timing features to pcretest. (1) The -tm option now times
3143 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3144 separate command line item, by a number that specifies the number of
3145 repeats to use when timing. The default is 50000; this gives better
3146 precision, but takes uncomfortably long for very large patterns.
3147
3148 20. Extended pcre_study() to be more clever in cases where a branch of a
3149 subpattern has no definite first character. For example, (a*|b*)[cd] would
3150 previously give no result from pcre_study(). Now it recognizes that the
3151 first character must be a, b, c, or d.
3152
3153 21. There was an incorrect error "recursive call could loop indefinitely" if
3154 a subpattern (or the entire pattern) that was being tested for matching an
3155 empty string contained only one non-empty item after a nested subpattern.
3156 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3157 incorrectly, because the \d was being skipped in the check.
3158
3159 22. The pcretest program now has a new pattern option /B and a command line
3160 option -b, which is equivalent to adding /B to every pattern. This causes
3161 it to show the compiled bytecode, without the additional information that
3162 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3163 is the same as /B/I).
3164
3165 23. A new optimization is now able automatically to treat some sequences such
3166 as a*b as a*+b. More specifically, if something simple (such as a character
3167 or a simple class like \d) has an unlimited quantifier, and is followed by
3168 something that cannot possibly match the quantified thing, the quantifier
3169 is automatically "possessified".
3170
3171 24. A recursive reference to a subpattern whose number was greater than 39
3172 went wrong under certain circumstances in UTF-8 mode. This bug could also
3173 have affected the operation of pcre_study().
3174
3175 25. Realized that a little bit of performance could be had by replacing
3176 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3177
3178 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3179
3180 27. Possessive quantifiers such as a++ were previously implemented by turning
3181 them into atomic groups such as ($>a+). Now they have their own opcodes,
3182 which improves performance. This includes the automatically created ones
3183 from 23 above.
3184
3185 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3186 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3187 the first matched character to be a colon. This applied both to named and
3188 numbered groups.
3189
3190 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3191
3192 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3193 building PCRE on Windows easier, so I have included it in the distribution.
3194
3195 31. There is now a check in pcretest against a ridiculously large number being
3196 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3197 loop, the loop is abandoned.
3198
3199 32. Forward references to subpatterns in conditions such as (?(2)...) where
3200 subpattern 2 is defined later cause pcre_compile() to search forwards in
3201 the pattern for the relevant set of parentheses. This search went wrong
3202 when there were unescaped parentheses in a character class, parentheses
3203 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3204
3205 33. "Subroutine" calls and backreferences were previously restricted to
3206 referencing subpatterns earlier in the regex. This restriction has now
3207 been removed.
3208
3209 34. Added a number of extra features that are going to be in Perl 5.10. On the
3210 whole, these are just syntactic alternatives for features that PCRE had
3211 previously implemented using the Python syntax or my own invention. The
3212 other formats are all retained for compatibility.
3213
3214 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3215 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3216 also .NET compatible.
3217
3218 (b) A recursion or subroutine call to a named group can now be defined as
3219 (?&name) as well as (?P>name).
3220
3221 (c) A backreference to a named group can now be defined as \k<name> or
3222 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3223 5.10, are also .NET compatible.
3224
3225 (d) A conditional reference to a named group can now use the syntax
3226 (?(<name>) or (?('name') as well as (?(name).
3227
3228 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3229 groups (named and numbered) that are never evaluated inline, but can be
3230 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3231 is always false. There may be only one alternative in such a group.
3232
3233 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3234 as the simple (?(R). The condition is true only if the most recent
3235 recursion is that of the given number or name. It does not search out
3236 through the entire recursion stack.
3237
3238 (g) The escape \gN or \g{N} has been added, where N is a positive or
3239 negative number, specifying an absolute or relative reference.
3240
3241 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3242 some "unreachable code" warnings.
3243
3244 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3245 things, this adds five new scripts.
3246
3247 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3248 There were also incompatibilities regarding the handling of \Q..\E inside
3249 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3250 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3251
3252 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3253 matches an empty string, and forcibly breaks the loop. There were bugs in
3254 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3255 against aaaa the result was just "a" rather than "aaaa", for example. Two
3256 separate and independent bugs (that affected different cases) have been
3257 fixed.
3258
3259 39. Refactored the code to abolish the use of different opcodes for small
3260 capturing bracket numbers. This is a tidy that I avoided doing when I
3261 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3262 The new approach is not only tidier, it makes it possible to reduce the
3263 memory needed to fix the previous bug (38).
3264
3265 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3266 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3267 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3268 mode.
3269
3270 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3271 report.
3272
3273 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3274 copy construction and assignment in the C++ wrapper.
3275
3276 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3277 couple of bugs that could have given wrong results in the "--newline=crlf"
3278 case.
3279
3280 44. Added a number of casts and did some reorganization of signed/unsigned int
3281 variables following suggestions from Dair Grant. Also renamed the variable
3282 "this" as "item" because it is a C++ keyword.
3283
3284 45. Arranged for dftables to add
3285
3286 #include "pcre_internal.h"
3287
3288 to pcre_chartables.c because without it, gcc 4.x may remove the array
3289 definition from the final binary if PCRE is built into a static library and
3290 dead code stripping is activated.
3291
3292 46. For an unanchored pattern, if a match attempt fails at the start of a
3293 newline sequence, and the newline setting is CRLF or ANY, and the next two
3294 characters are CRLF, advance by two characters instead of one.
3295
3296
3297 Version 6.7 04-Jul-06
3298 ---------------------
3299
3300 1. In order to handle tests when input lines are enormously long, pcretest has
3301 been re-factored so that it automatically extends its buffers when
3302 necessary. The code is crude, but this _is_ just a test program. The
3303 default size has been increased from 32K to 50K.
3304
3305 2. The code in pcre_study() was using the value of the re argument before
3306 testing it for NULL. (Of course, in any sensible call of the function, it
3307 won't be NULL.)
3308
3309 3. The memmove() emulation function in pcre_internal.h, which is used on
3310 systems that lack both memmove() and bcopy() - that is, hardly ever -
3311 was missing a "static" storage class specifier.
3312
3313 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3314 containing an extended class (one that cannot be represented by a bitmap
3315 because it contains high-valued characters or Unicode property items, e.g.
3316 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3317 pattern, but PCRE should not loop if you do not (it no longer does).
3318 [Detail: two cases were found: (a) a repeated subpattern containing an
3319 extended class; (b) a recursive reference to a subpattern that followed a
3320 previous extended class. It wasn't skipping over the extended class
3321 correctly when UTF-8 mode was not set.]
3322
3323 5. A negated single-character class was not being recognized as fixed-length
3324 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3325 compile error "lookbehind assertion is not fixed length".
3326
3327 6. The RunPerlTest auxiliary script was showing an unexpected difference
3328 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3329 write a Perl script that can interpret lines of an input file either as
3330 byte characters or as UTF-8, which is what "perltest" was being required to
3331 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3332 can't do is switch easily at run time between having the "use utf8;" pragma
3333 or not. In the end, I fudged it by using the RunPerlTest script to insert
3334 "use utf8;" explicitly for the UTF-8 tests.
3335
3336 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3337 the end of the subject string, contrary to the documentation and to what
3338 Perl does. This was true of both matching functions. Now it matches only at
3339 the start of the subject and immediately after *internal* newlines.
3340
3341 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3342 a pointer to an int instead of a pointer to an unsigned long int. This
3343 caused problems on 64-bit systems.
3344
3345 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3346 instance of the 'standard' template library not being so standard".
3347
3348 10. There was no check on the number of named subpatterns nor the maximum
3349 length of a subpattern name. The product of these values is used to compute
3350 the size of the memory block for a compiled pattern. By supplying a very
3351 long subpattern name and a large number of named subpatterns, the size
3352 computation could be caused to overflow. This is now prevented by limiting
3353 the length of names to 32 characters, and the number of named subpatterns
3354 to 10,000.
3355
3356 11. Subpatterns that are repeated with specific counts have to be replicated in
3357 the compiled pattern. The size of memory for this was computed from the
3358 length of the subpattern and the repeat count. The latter is limited to
3359 65535, but there was no limit on the former, meaning that integer overflow
3360 could in principle occur. The compiled length of a repeated subpattern is
3361 now limited to 30,000 bytes in order to prevent this.
3362
3363 12. Added the optional facility to have named substrings with the same name.
3364
3365 13. Added the ability to use a named substring as a condition, using the
3366 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3367 are numbers (not recommended). Forward references are permitted.
3368
3369 14. Added forward references in named backreferences (if you see what I mean).
3370
3371 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3372 pattern could run off the end of the subject. For example, the pattern
3373 "(?s)(.{1,5})"8 did this with the subject "ab".
3374
3375 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3376 PCRE_CASELESS was set when matching characters that were quantified with ?
3377 or *.
3378
3379 17. A character class other than a single negated character that had a minimum
3380 but no maximum quantifier - for example [ab]{6,} - was not handled
3381 correctly by pce_dfa_exec(). It would match only one character.
3382
3383 18. A valid (though odd) pattern that looked like a POSIX character
3384 class but used an invalid character after [ (for example [[,abc,]]) caused
3385 pcre_compile() to give the error "Failed: internal error: code overflow" or
3386 in some cases to crash with a glibc free() error. This could even happen if
3387 the pattern terminated after [[ but there just happened to be a sequence of
3388 letters, a binary zero, and a closing ] in the memory that followed.
3389
3390 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3391 over the years. Originally (before any Unicode support), just the bottom 8
3392 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3393 output from "man perlunicode" includes this:
3394
3395 The regular expression compiler produces polymorphic opcodes. That
3396 is, the pattern adapts to the data and automatically switches to
3397 the Unicode character scheme when presented with Unicode data--or
3398 instead uses a traditional byte scheme when presented with byte
3399 data.
3400
3401 Sadly, a wide octal escape does not cause a switch, and in a string with
3402 no other multibyte characters, these octal escapes are treated as before.
3403 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3404 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3405 Unicode string.
3406
3407 I have not perpetrated such confusion in PCRE. Up till now, it took just
3408 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3409 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3410 translate to the appropriate multibyte character.
3411
3412 29. Applied some refactoring to reduce the number of warnings from Microsoft
3413 and Borland compilers. This has included removing the fudge introduced
3414 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3415 a warning about an unused variable.
3416
3417 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3418 characters since release 4.0, because Perl (from release 5.004) does not.
3419 [Or at least, is documented not to: some releases seem to be in conflict
3420 with the documentation.] However, when a pattern was studied with
3421 pcre_study() and all its branches started with \s, PCRE still included VT
3422 as a possible starting character. Of course, this did no harm; it just
3423 caused an unnecessary match attempt.
3424
3425 22. Removed a now-redundant internal flag bit that recorded the fact that case
3426 dependency changed within the pattern. This was once needed for "required
3427 byte" processing, but is no longer used. This recovers a now-scarce options
3428 bit. Also moved the least significant internal flag bit to the most-
3429 significant bit of the word, which was not previously used (hangover from
3430 the days when it was an int rather than a uint) to free up another bit for
3431 the future.
3432
3433 23. Added support for CRLF line endings as well as CR and LF. As well as the
3434 default being selectable at build time, it can now be changed at runtime
3435 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3436 specify that it is scanning data with non-default line endings.
3437
3438 24. Changed the definition of CXXLINK to make it agree with the definition of
3439 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3440
3441 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3442 recursions. This makes a big different to stack usage for some patterns.
3443
3444 26. If a subpattern containing a named recursion or subroutine reference such
3445 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3446 the space required for the compiled pattern went wrong and gave too small a
3447 value. Depending on the environment, this could lead to "Failed: internal
3448 error: code overflow at offset 49" or "glibc detected double free or
3449 corruption" errors.
3450
3451 27. Applied patches from Google (a) to support the new newline modes and (b) to
3452 advance over multibyte UTF-8 characters in GlobalReplace.
3453
3454 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3455 difference for some implementation of PCRE in some Windows version.
3456
3457 29. Added some extra testing facilities to pcretest:
3458
3459 \q<number> in a data line sets the "match limit" value
3460 \Q<number> in a data line sets the "match recursion limt" value
3461 -S <number> sets the stack size, where <number> is in megabytes
3462
3463 The -S option isn't available for Windows.
3464
3465
3466 Version 6.6 06-Feb-06
3467 ---------------------
3468
3469 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3470 in pcreposix.h. I have copied the definition from pcre.h.
3471
3472 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3473 because pcre.h is no longer a built file.
3474
3475 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3476 not normally included in the compiled code.
3477
3478
3479 Version 6.5 01-Feb-06
3480 ---------------------
3481
3482 1. When using the partial match feature with pcre_dfa_exec(), it was not
3483 anchoring the second and subsequent partial matches at the new starting
3484 point. This could lead to incorrect results. For example, with the pattern
3485 /1234/, partially matching against "123" and then "a4" gave a match.
3486
3487 2. Changes to pcregrep:
3488
3489 (a) All non-match returns from pcre_exec() were being treated as failures
3490 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3491 error message is output. Some extra information is given for the
3492 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3493 probably the only errors that are likely to be caused by users (by
3494 specifying a regex that has nested indefinite repeats, for instance).
3495 If there are more than 20 of these errors, pcregrep is abandoned.
3496
3497 (b) A binary zero was treated as data while matching, but terminated the
3498 output line if it was written out. This has been fixed: binary zeroes
3499 are now no different to any other data bytes.
3500
3501 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3502 used to set a locale for matching. The --locale=xxxx long option has
3503 been added (no short equivalent) to specify a locale explicitly on the
3504 pcregrep command, overriding the environment variables.
3505
3506 (d) When -B was used with -n, some line numbers in the output were one less
3507 than they should have been.
3508
3509 (e) Added the -o (--only-matching) option.
3510
3511 (f) If -A or -C was used with -c (count only), some lines of context were
3512 accidentally printed for the final match.
3513
3514 (g) Added the -H (--with-filename) option.
3515
3516 (h) The combination of options -rh failed to suppress file names for files
3517 that were found from directory arguments.
3518
3519 (i) Added the -D (--devices) and -d (--directories) options.
3520
3521 (j) Added the -F (--fixed-strings) option.
3522
3523 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3524
3525 (l) Added the --colo(u)r option.
3526
3527 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3528 is not present by default.
3529
3530 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3531 items such as (?R) or (?1), when the recursion could match a number of
3532 alternatives. If it matched one of the alternatives, but subsequently,
3533 outside the recursion, there was a failure, the code tried to back up into
3534 the recursion. However, because of the way PCRE is implemented, this is not
3535 possible, and the result was an incorrect result from the match.
3536
3537 In order to prevent this happening, the specification of recursion has
3538 been changed so that all such subpatterns are automatically treated as
3539 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3540
3541 4. I had overlooked the fact that, in some locales, there are characters for
3542 which isalpha() is true but neither isupper() nor islower() are true. In
3543 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3544 and ordfeminine) are like this. This affected the treatment of \w and \W
3545 when they appeared in character classes, but not when they appeared outside
3546 a character class. The bit map for "word" characters is now created
3547 separately from the results of isalnum() instead of just taking it from the
3548 upper, lower, and digit maps. (Plus the underscore character, of course.)
3549
3550 5. The above bug also affected the handling of POSIX character classes such as
3551 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3552 permanent tables. Instead, the bit maps for such a class were previously
3553 created as the appropriate unions of the upper, lower, and digit bitmaps.
3554 Now they are created by subtraction from the [[:word:]] class, which has
3555 its own bitmap.
3556
3557 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3558 It is created by subtracting the vertical space characters (\x09, \x0a,
3559 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3560 subtraction was done in the overall bitmap for a character class, meaning
3561 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3562 be recognized. This bug has been fixed.
3563
3564 7. Patches from the folks at Google:
3565
3566 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3567 real life, but is still worth protecting against".
3568
3569 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3570 regular expressions".
3571
3572 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3573 have it.
3574
3575 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3576 "configure" and the latter not, in order to fix a problem somebody had
3577 with compiling the Arg class on HP-UX.
3578
3579 (e) Improve the error-handling of the C++ wrapper a little bit.
3580
3581 (f) New tests for checking recursion limiting.
3582
3583 8. The pcre_memmove() function, which is used only if the environment does not
3584 have a standard memmove() function (and is therefore rarely compiled),
3585 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3586 returning a result (though PCRE never actually uses the result).
3587
3588 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3589 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3590 returned instead of calling malloc() with an overflowing number that would
3591 most likely cause subsequent chaos.
3592
3593 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3594
3595 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3596 with this option is matched, the nmatch and pmatch options of regexec() are
3597 ignored.
3598
3599 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3600 provided in case anyone wants to the the POSIX interface with UTF-8
3601 strings.
3602
3603 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3604 C++ linking (needed for some HP-UX environments).
3605
3606 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3607 (unused parameter) and in the pcre_printint() function (omitted "default"
3608 switch label when the default is to do nothing).
3609
3610 15. Added some code to make it possible, when PCRE is compiled as a C++
3611 library, to replace subject pointers for pcre_exec() with a smart pointer
3612 class, thus making it possible to process discontinuous strings.
3613
3614 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3615 much the same function. They were added by different people who were trying
3616 to make PCRE easy to compile on non-Unix systems. It has been suggested
3617 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3618 for compiling on Windows systems. I have therefore replaced it with
3619 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3620 defaults to "extern" for C or "extern C" for C++, which works fine on
3621 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3622 SCOPE with something explicit in config.h. In addition:
3623
3624 (a) pcreposix.h still had just "extern" instead of either of these macros;
3625 I have replaced it with PCRE_DATA_SCOPE.
3626
3627 (b) Functions such as _pcre_xclass(), which are internal to the library,
3628 but external in the C sense, all had PCRE_EXPORT in their definitions.
3629 This is apparently wrong for the Windows case, so I have removed it.
3630 (It makes no difference on Unix-like systems.)
3631
3632 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3633 of recursive calls to match(). This is different to MATCH_LIMIT because
3634 that limits the total number of calls to match(), not all of which increase
3635 the depth of recursion. Limiting the recursion depth limits the amount of
3636 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3637 when PCRE is compiled, and changed at run time. A patch from Google adds
3638 this functionality to the C++ interface.
3639
3640 18. Changes to the handling of Unicode character properties:
3641
3642 (a) Updated the table to Unicode 4.1.0.
3643
3644 (b) Recognize characters that are not in the table as "Cn" (undefined).
3645
3646 (c) I revised the way the table is implemented to a much improved format
3647 which includes recognition of ranges. It now supports the ranges that
3648 are defined in UnicodeData.txt, and it also amalgamates other
3649 characters into ranges. This has reduced the number of entries in the
3650 table from around 16,000 to around 3,000, thus reducing its size
3651 considerably. I realized I did not need to use a tree structure after
3652 all - a binary chop search is just as efficient. Having reduced the
3653 number of entries, I extended their size from 6 bytes to 8 bytes to
3654 allow for more data.
3655
3656 (d) Added support for Unicode script names via properties such as \p{Han}.
3657
3658 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3659 matching that character.
3660
3661 20. When matching a repeated Unicode property with a minimum greater than zero,
3662 (for example \pL{2,}), PCRE could look past the end of the subject if it
3663 reached it while seeking the minimum number of characters. This could
3664 happen only if some of the characters were more than one byte long, because
3665 there is a check for at least the minimum number of bytes.
3666
3667 21. Refactored the implementation of \p and \P so as to be more general, to
3668 allow for more different types of property in future. This has changed the
3669 compiled form incompatibly. Anybody with saved compiled patterns that use
3670 \p or \P will have to recompile them.
3671
3672 22. Added "Any" and "L&" to the supported property types.
3673
3674 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3675 but give a compile time error if the value is greater than 0xff.
3676
3677 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3678 accidentally not being installed or uninstalled.
3679
3680 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3681 made were to insert the current release number. This seemed silly, because
3682 it made things harder for people building PCRE on systems that don't run
3683 "configure". I have turned pcre.h into a distributed file, no longer built
3684 by "configure", with the version identification directly included. There is
3685 no longer a pcre.h.in file.
3686
3687 However, this change necessitated a change to the pcre-config script as
3688 well. It is built from pcre-config.in, and one of the substitutions was the
3689 release number. I have updated configure.ac so that ./configure now finds
3690 the release number by grepping pcre.h.
3691
3692 26. Added the ability to run the tests under valgrind.
3693
3694
3695 Version 6.4 05-Sep-05
3696 ---------------------
3697
3698 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3699 "--" to be printed when multiple files were scanned, even when none of the
3700 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3701 consider it to be a bug, and have restored the previous behaviour.
3702
3703 2. A couple of code tidies to get rid of compiler warnings.
3704
3705 3. The pcretest program used to cheat by referring to symbols in the library
3706 whose names begin with _pcre_. These are internal symbols that are not
3707 really supposed to be visible externally, and in some environments it is
3708 possible to suppress them. The cheating is now confined to including
3709 certain files from the library's source, which is a bit cleaner.
3710
3711 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3712 file's purpose clearer.
3713
3714 5. Reorganized pcre_ucp_findchar().
3715
3716
3717 Version 6.3 15-Aug-05
3718 ---------------------
3719
3720 1. The file libpcre.pc.in did not have general read permission in the tarball.
3721
3722 2. There were some problems when building without C++ support:
3723
3724 (a) If C++ support was not built, "make install" and "make test" still
3725 tried to test it.
3726
3727 (b) There were problems when the value of CXX was explicitly set. Some
3728 changes have been made to try to fix these, and ...
3729
3730 (c) --disable-cpp can now be used to explicitly disable C++ support.
3731
3732 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3733 backslash in a target when C++ was disabled. This confuses some
3734 versions of "make", apparently. Using an intermediate variable solves
3735 this. (Same for CPP_LOBJ.)
3736
3737 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3738 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3739 necessary on certain architectures.
3740
3741 4. Added a setting of -export-symbols-regex to the link command to remove
3742 those symbols that are exported in the C sense, but actually are local
3743 within the library, and not documented. Their names all begin with
3744 "_pcre_". This is not a perfect job, because (a) we have to except some
3745 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3746 available (and never for static libraries). I have made a note to try to
3747 find a way round (a) in the future.
3748
3749
3750 Version 6.2 01-Aug-05
3751 ---------------------
3752
3753 1. There was no test for integer overflow of quantifier values. A construction
3754 such as {1111111111111111} would give undefined results. What is worse, if
3755 a minimum quantifier for a parenthesized subpattern overflowed and became
3756 negative, the calculation of the memory size went wrong. This could have
3757 led to memory overwriting.
3758
3759 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3760
3761 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3762 operating environments where this matters.
3763
3764 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3765 PCRE options from within the C++ wrapper.
3766
3767 5. Named capturing subpatterns were not being correctly counted when a pattern
3768 was compiled. This caused two problems: (a) If there were more than 100
3769 such subpatterns, the calculation of the memory needed for the whole
3770 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3771 back references of the form \12, where the number was greater than 9, were
3772 not recognized as back references, even though there were sufficient
3773 previous subpatterns.
3774
3775 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3776 versions of gcc, e.g. 2.95.4.
3777
3778
3779 Version 6.1 21-Jun-05
3780 ---------------------
3781
3782 1. There was one reference to the variable "posix" in pcretest.c that was not
3783 surrounded by "#if !defined NOPOSIX".
3784
3785 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3786 the cross-check on the old pcre_info() function, for the benefit of the
3787 cut-down version of PCRE that is currently imported into Exim.
3788
3789 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3790 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3791 patterns that start (?i) but I don't think that matters. The use of (?i) is
3792 just an example; this all applies to the other options as well.
3793
3794 4. Since libtool seems to echo the compile commands it is issuing, the output
3795 from "make" can be reduced a bit by putting "@" in front of each libtool
3796 compile command.
3797
3798 5. Patch from the folks at Google for configure.in to be a bit more thorough
3799 in checking for a suitable C++ installation before trying to compile the
3800 C++ stuff. This should fix a reported problem when a compiler was present,
3801 but no suitable headers.
3802
3803 6. The man pages all had just "PCRE" as their title. I have changed them to
3804 be the relevant file name. I have also arranged that these names are
3805 retained in the file doc/pcre.txt, which is a concatenation in text format
3806 of all the man pages except the little individual ones for each function.
3807
3808 7. The NON-UNIX-USE file had not been updated for the different set of source
3809 files that come with release 6. I also added a few comments about the C++
3810 wrapper.
3811
3812
3813 Version 6.0 07-Jun-05
3814 ---------------------
3815
3816 1. Some minor internal re-organization to help with my DFA experiments.
3817
3818 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3819 didn't matter for the library itself when fully configured, but did matter
3820 when compiling without UCP support, or within Exim, where the ucp files are
3821 not imported.
3822
3823 3. Refactoring of the library code to split up the various functions into
3824 different source modules. The addition of the new DFA matching code (see
3825 below) to a single monolithic source would have made it really too
3826 unwieldy, quite apart from causing all the code to be include in a
3827 statically linked application, when only some functions are used. This is
3828 relevant even without the DFA addition now that patterns can be compiled in
3829 one application and matched in another.
3830
3831 The downside of splitting up is that there have to be some external
3832 functions and data tables that are used internally in different modules of
3833 the library but which are not part of the API. These have all had their
3834 names changed to start with "_pcre_" so that they are unlikely to clash
3835 with other external names.
3836
3837 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3838 a different (DFA) algorithm. Although it is slower than the original
3839 function, it does have some advantages for certain types of matching
3840 problem.
3841
3842 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3843 including restarting after a partial match.
3844
3845 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3846 defined when compiling for Windows was sent to me. I have put it into the
3847 code, though I have no means of testing or verifying it.
3848
3849 7. Added the pcre_refcount() auxiliary function.
3850
3851 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3852 match before or at the first newline in the subject string. In pcretest,
3853 the /f option on a pattern can be used to set this.
3854
3855 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3856 would behave wrongly. This has been present in PCRE since release 4.0.
3857
3858 10. A number of changes to the pcregrep command:
3859
3860 (a) Refactored how -x works; insert ^(...)$ instead of setting
3861 PCRE_ANCHORED and checking the length, in preparation for adding
3862 something similar for -w.
3863
3864 (b) Added the -w (match as a word) option.
3865
3866 (c) Refactored the way lines are read and buffered so as to have more
3867 than one at a time available.
3868
3869 (d) Implemented a pcregrep test script.
3870
3871 (e) Added the -M (multiline match) option. This allows patterns to match
3872 over several lines of the subject. The buffering ensures that at least
3873 8K, or the rest of the document (whichever is the shorter) is available
3874 for matching (and similarly the previous 8K for lookbehind assertions).
3875
3876 (f) Changed the --help output so that it now says
3877
3878 -w, --word-regex(p)
3879
3880 instead of two lines, one with "regex" and the other with "regexp"
3881 because that confused at least one person since the short forms are the
3882 same. (This required a bit of code, as the output is generated
3883 automatically from a table. It wasn't just a text change.)
3884
3885 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3886 option but starts with a hyphen. Could be a pattern or a path name
3887 starting with a hyphen, for instance.
3888
3889 (h) "-" can be given as a file name to represent stdin.
3890
3891 (i) When file names are being printed, "(standard input)" is used for
3892 the standard input, for compatibility with GNU grep. Previously
3893 "<stdin>" was used.
3894
3895 (j) The option --label=xxx can be used to supply a name to be used for
3896 stdin when file names are being printed. There is no short form.
3897
3898 (k) Re-factored the options decoding logic because we are going to add
3899 two more options that take data. Such options can now be given in four
3900 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3901
3902 (l) Added the -A, -B, and -C options for requesting that lines of context
3903 around matches be printed.
3904
3905 (m) Added the -L option to print the names of files that do not contain
3906 any matching lines, that is, the complement of -l.
3907
3908 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3909 continue to scan other files.
3910
3911 (o) The -s option was incorrectly implemented. For compatibility with other
3912 greps, it now suppresses the error message for a non-existent or non-
3913 accessible file (but not the return code). There is a new option called
3914 -q that suppresses the output of matching lines, which was what -s was
3915 previously doing.
3916
3917 (p) Added --include and --exclude options to specify files for inclusion
3918 and exclusion when recursing.
3919
3920 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3921 Hopefully, it now does.
3922
3923 12. Missing cast in pcre_study().
3924
3925 13. Added an "uninstall" target to the makefile.
3926
3927 14. Replaced "extern" in the function prototypes in Makefile.in with
3928 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3929 world, but is set differently for Windows.
3930
3931 15. Added a second compiling function called pcre_compile2(). The only
3932 difference is that it has an extra argument, which is a pointer to an
3933 integer error code. When there is a compile-time failure, this is set
3934 non-zero, in addition to the error test pointer being set to point to an
3935 error message. The new argument may be NULL if no error number is required
3936 (but then you may as well call pcre_compile(), which is now just a
3937 wrapper). This facility is provided because some applications need a
3938 numeric error indication, but it has also enabled me to tidy up the way
3939 compile-time errors are handled in the POSIX wrapper.
3940
3941 16. Added VPATH=.libs to the makefile; this should help when building with one
3942 prefix path and installing with another. (Or so I'm told by someone who
3943 knows more about this stuff than I do.)
3944
3945 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3946 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3947 match everything, including newlines. This is not POSIX-compatible, but
3948 somebody wanted the feature. From pcretest it can be activated by using
3949 both the P and the s flags.
3950
3951 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3952
3953 19. libpcre.pc was being incorrectly installed as executable.
3954
3955 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3956 it now also looks for '\r' so that it will work unmodified on Windows.
3957
3958 21. Added Google's contributed C++ wrapper to the distribution.
3959
3960 22. Added some untidy missing memory free() calls in pcretest, to keep
3961 Electric Fence happy when testing.
3962
3963
3964
3965 Version 5.0 13-Sep-04
3966 ---------------------
3967
3968 1. Internal change: literal characters are no longer packed up into items
3969 containing multiple characters in a single byte-string. Each character
3970 is now matched using a separate opcode. However, there may be more than one
3971 byte in the character in UTF-8 mode.
3972
3973 2. The pcre_callout_block structure has two new fields: pattern_position and
3974 next_item_length. These contain the offset in the pattern to the next match
3975 item, and its length, respectively.
3976
3977 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3978 insertion of callouts before each pattern item. Added the /C option to
3979 pcretest to make use of this.
3980
3981 4. On the advice of a Windows user, the lines
3982
3983 #if defined(_WIN32) || defined(WIN32)
3984 _setmode( _fileno( stdout ), 0x8000 );
3985 #endif /* defined(_WIN32) || defined(WIN32) */
3986
3987 have been added to the source of pcretest. This apparently does useful
3988 magic in relation to line terminators.
3989
3990 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3991 for the benefit of those environments where the "b" makes a difference.
3992
3993 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3994 to know about it. I have put a hack into configure.in that adds in code
3995 to set GCC=yes if CC=icc. This seems to end up at a point in the
3996 generated configure script that is early enough to affect the setting of
3997 compiler options, which is what is needed, but I have no means of testing
3998 whether it really works. (The user who reported this had patched the
3999 generated configure script, which of course I cannot do.)
4000
4001 LATER: After change 22 below (new libtool files), the configure script
4002 seems to know about icc (and also ecc). Therefore, I have commented out
4003 this hack in configure.in.
4004
4005 7. Added support for pkg-config (2 patches were sent in).
4006
4007 8. Negated POSIX character classes that used a combination of internal tables
4008 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4009 [[:^ascii]]. Typically, they would match almost any characters. The other
4010 POSIX classes were not broken in this way.
4011
4012 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4013 to find the match, as PCRE was deluded into thinking that the match had to
4014 start at the start point or following a newline. The same bug applied to
4015 patterns with negative forward assertions or any backward assertions
4016 preceding ".*" at the start, unless the pattern required a fixed first
4017 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4018
4019 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4020 starting at the last subject character, bytes beyond the end of the subject
4021 string were read.
4022
4023 11. Renamed the variable "class" as "classbits" to make life easier for C++
4024 users. (Previously there was a macro definition, but it apparently wasn't
4025 enough.)
4026
4027 12. Added the new field "tables" to the extra data so that tables can be passed
4028 in at exec time, or the internal tables can be re-selected. This allows
4029 a compiled regex to be saved and re-used at a later time by a different
4030 program that might have everything at different addresses.
4031
4032 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4033 -R library as well as a -L library.
4034
4035 14. The debugging options of pcretest (-d on the command line or D on a
4036 pattern) showed incorrect output for anything following an extended class
4037 that contained multibyte characters and which was followed by a quantifier.
4038
4039 15. Added optional support for general category Unicode character properties
4040 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4041 support. It adds about 90K to the size of the library. The meanings of the
4042 inbuilt class escapes such as \d and \s have NOT been changed.
4043
4044 16. Updated pcredemo.c to include calls to free() to release the memory for the
4045 compiled pattern.
4046
4047 17. The generated file chartables.c was being created in the source directory
4048 instead of in the building directory. This caused the build to fail if the
4049 source directory was different from the building directory, and was
4050 read-only.
4051
4052 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4053 file. No doubt somebody will tell me if they don't make sense... Also added
4054 Dan Mooney's comments about building on OpenVMS.
4055
4056 19. Added support for partial matching via the PCRE_PARTIAL option for
4057 pcre_exec() and the \P data escape in pcretest.
4058
4059 20. Extended pcretest with 3 new pattern features:
4060
4061 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4062 write the compiled pattern to the file whose name is "rest-of-line".
4063 This is a straight binary dump of the data, with the saved pointer to
4064 the character tables forced to be NULL. The study data, if any, is
4065 written too. After writing, pcretest reads a new pattern.
4066
4067 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4068 compiled pattern from the given file. There must not be any
4069 occurrences of "<" in the file name (pretty unlikely); if there are,
4070 pcretest will instead treat the initial "<" as a pattern delimiter.
4071 After reading in the pattern, pcretest goes on to read data lines as
4072 usual.
4073
4074 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4075 and 16-bit fields in a compiled pattern, to simulate a pattern that
4076 was compiled on a host of opposite endianness.
4077
4078 21. The pcre-exec() function can now cope with patterns that were compiled on
4079 hosts of opposite endianness, with this restriction:
4080
4081 As for any compiled expression that is saved and used later, the tables
4082 pointer field cannot be preserved; the extra_data field in the arguments
4083 to pcre_exec() should be used to pass in a tables address if a value
4084 other than the default internal tables were used at compile time.
4085
4086 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4087 now diagnosed as an error. Previously, most of the time, a negative number
4088 would have been treated as zero, but if in addition "ovector" was passed as
4089 NULL, a crash could occur.
4090
4091 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4092 new versions from the libtool 1.5 distribution (the last one is a copy of
4093 a file called libtool.m4). This seems to have fixed the need to patch
4094 "configure" to support Darwin 1.3 (which I used to do). However, I still
4095 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4096 workstation).
4097
4098 24. Changed the PCRE licence to be the more standard "BSD" licence.
4099
4100
4101 Version 4.5 01-Dec-03
4102 ---------------------
4103
4104 1. There has been some re-arrangement of the code for the match() function so
4105 that it can be compiled in a version that does not call itself recursively.
4106 Instead, it keeps those local variables that need separate instances for
4107 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4108 needs to "recurse". Keeping track of where control must go is done by means
4109 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4110 hide most of the details from the main code, and operates only if
4111 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4112 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4113 operating.
4114
4115 To make it easier for callers to provide specially tailored get/free
4116 functions for this usage, two new functions, pcre_stack_malloc, and
4117 pcre_stack_free, are used. They are always called in strict stacking order,
4118 and the size of block requested is always the same.
4119
4120 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4121 PCRE has been compiled to use the stack or the heap for recursion. The
4122 -C option of pcretest uses this to show which version is compiled.
4123
4124 A new data escape \S, is added to pcretest; it causes the amounts of store
4125 obtained and freed by both kinds of malloc/free at match time to be added
4126 to the output.
4127
4128 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4129 what's available on my current Linux desktop machine.
4130
4131 3. When matching a UTF-8 string, the test for a valid string at the start has
4132 been extended. If start_offset is not zero, PCRE now checks that it points
4133 to a byte that is the start of a UTF-8 character. If not, it returns
4134 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4135 this is necessary because there may be backward assertions in the pattern.
4136 When matching the same subject several times, it may save resources to use
4137 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4138
4139 4. The code for checking the validity of UTF-8 strings has been tightened so
4140 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4141 containing "overlong sequences".
4142
4143 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4144 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4145 should have been "&", but it just so happened that all the cases this let
4146 through by mistake were picked up later in the function.
4147
4148 6. I had used a variable called "isblank" - this is a C99 function, causing
4149 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4150
4151 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4152 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4153 so the version is shown; (c) stop "make test" echoing "./RunTest".
4154
4155 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4156
4157 9. The prototype for memmove() for systems that don't have it was using
4158 size_t, but the inclusion of the header that defines size_t was later. I've
4159 moved the #includes for the C headers earlier to avoid this.
4160
4161 10. Added some adjustments to the code to make it easier to compiler on certain
4162 special systems:
4163
4164 (a) Some "const" qualifiers were missing.
4165 (b) Added the macro EXPORT before all exported functions; by default this
4166 is defined to be empty.
4167 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4168 that it reads its output file name as an argument instead of writing
4169 to the standard output and assuming this can be redirected.
4170
4171 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4172 class containing characters with values greater than 255, PCRE compilation
4173 went into a loop.
4174
4175 12. A recursive reference to a subpattern that was within another subpattern
4176 that had a minimum quantifier of zero caused PCRE to crash. For example,
4177 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4178 recursion. If the recursively-called subpattern itself had a zero repeat,
4179 that was OK.
4180
4181 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4182 buffer into which it was copied (for escape processing) was still set at
4183 1024, so long lines caused crashes.
4184
4185 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4186 "internal error: code overflow...". This applied to any character class
4187 that was followed by a possessive quantifier.
4188
4189 15. Modified the Makefile to add libpcre.la as a prerequisite for
4190 libpcreposix.la because I was told this is needed for a parallel build to
4191 work.
4192
4193 16. If a pattern that contained .* following optional items at the start was
4194 studied, the wrong optimizing data was generated, leading to matching
4195 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4196 matching string must start with a or b or c. The correct conclusion for
4197 this pattern is that a match can start with any character.
4198
4199
4200 Version 4.4 13-Aug-03
4201 ---------------------
4202
4203 1. In UTF-8 mode, a character class containing characters with values between
4204 127 and 255 was not handled correctly if the compiled pattern was studied.
4205 In fixing this, I have also improved the studying algorithm for such
4206 classes (slightly).
4207
4208 2. Three internal functions had redundant arguments passed to them. Removal
4209 might give a very teeny performance improvement.
4210
4211 3. Documentation bug: the value of the capture_top field in a callout is *one
4212 more than* the number of the hightest numbered captured substring.
4213
4214 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4215 in incorrectly linking with a previously installed version. They now link
4216 explicitly with libpcre.la.
4217
4218 5. configure.in no longer needs to recognize Cygwin specially.
4219
4220 6. A problem in pcre.in for Windows platforms is fixed.
4221
4222 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4223 pcretest, it used to include the size of the study block as part of its
4224 output. Unfortunately, the structure contains a field that has a different
4225 size on different hardware architectures. This meant that the tests that
4226 showed this size failed. As the block is currently always of a fixed size,
4227 this information isn't actually particularly useful in pcretest output, so
4228 I have just removed it.
4229
4230 8. Three pre-processor statements accidentally did not start in column 1.
4231 Sadly, there are *still* compilers around that complain, even though
4232 standard C has not required this for well over a decade. Sigh.
4233
4234 9. In pcretest, the code for checking callouts passed small integers in the
4235 callout_data field, which is a void * field. However, some picky compilers
4236 complained about the casts involved for this on 64-bit systems. Now
4237 pcretest passes the address of the small integer instead, which should get
4238 rid of the warnings.
4239
4240 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4241 both compile and run time, and gives an error if an invalid UTF-8 sequence
4242 is found. There is a option for disabling this check in cases where the
4243 string is known to be correct and/or the maximum performance is wanted.
4244
4245 11. In response to a bug report, I changed one line in Makefile.in from
4246
4247 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4248 to
4249 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4250
4251 to look similar to other lines, but I have no way of telling whether this
4252 is the right thing to do, as I do not use Windows. No doubt I'll get told
4253 if it's wrong...
4254
4255
4256 Version 4.3 21-May-03
4257 ---------------------
4258
4259 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4260 Makefile.
4261
4262 2. Some refactoring to improve the quality of the code:
4263
4264 (i) The utf8_table... variables are now declared "const".
4265
4266 (ii) The code for \cx, which used the "case flipping" table to upper case
4267 lower case letters, now just substracts 32. This is ASCII-specific,
4268 but the whole concept of \cx is ASCII-specific, so it seems
4269 reasonable.
4270
4271 (iii) PCRE was using its character types table to recognize decimal and
4272 hexadecimal digits in the pattern. This is silly, because it handles
4273 only 0-9, a-f, and A-F, but the character types table is locale-
4274 specific, which means strange things might happen. A private
4275 table is now used for this - though it costs 256 bytes, a table is
4276 much faster than multiple explicit tests. Of course, the standard
4277 character types table is still used for matching digits in subject
4278 strings against \d.
4279
4280 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4281 ending in _t are). So I've renamed it as ESC_tee.
4282
4283 3. The first argument for regexec() in the POSIX wrapper should have been
4284 defined as "const".
4285
4286 4. Changed pcretest to use malloc() for its buffers so that they can be
4287 Electric Fenced for debugging.
4288
4289 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4290 to read one or more bytes before the start of the subject string. Often this
4291 had no effect on PCRE's behaviour, but in some circumstances it could
4292 provoke a segmentation fault.
4293
4294 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4295 to try to read one or more bytes before the start of the subject string.
4296
4297 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4298 UTF-8 support could misbehave in various ways if the subject string
4299 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4300 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4301 back over UTF-8 characters.)
4302
4303
4304 Version 4.2 14-Apr-03
4305 ---------------------
4306
4307 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4308
4309 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4310 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4311 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4312 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4313 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4314 and BUILD_EXEEXT
4315 Note: automatic setting of the BUILD variables is not yet working
4316 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4317 compile-time but not at link-time
4318 [LINK]: use for linking executables only
4319 make different versions for Windows and non-Windows
4320 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4321 libraries
4322 [LINK_FOR_BUILD]: new variable
4323 [OBJEXT]: use throughout
4324 [EXEEXT]: use throughout
4325 <winshared>: new target
4326 <wininstall>: new target
4327 <dftables.o>: use native compiler
4328 <dftables>: use native linker
4329 <install>: handle Windows platform correctly
4330 <clean>: ditto
4331 <check>: ditto
4332 copy DLL to top builddir before testing
4333
4334 As part of these changes, -no-undefined was removed again. This was reported
4335 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4336 in any case.
4337
4338 3. Some tidies to get rid of compiler warnings:
4339
4340 . In the match_data structure, match_limit was an unsigned long int, whereas
4341 match_call_count was an int. I've made them both unsigned long ints.
4342
4343 . In pcretest the fact that a const uschar * doesn't automatically cast to
4344 a void * provoked a warning.
4345
4346 . Turning on some more compiler warnings threw up some "shadow" variables
4347 and a few more missing casts.
4348
4349 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4350 option, a class that contained a single character with a value between 128
4351 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4352
4353 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4354 option, a class that contained several characters, but with at least one
4355 whose value was between 128 and 255 caused PCRE to crash.
4356
4357
4358 Version 4.1 12-Mar-03
4359 ---------------------
4360
4361 1. Compiling with gcc -pedantic found a couple of places where casts were
4362 needed, and a string in dftables.c that was longer than standard compilers are
4363 required to support.
4364
4365 2. Compiling with Sun's compiler found a few more places where the code could
4366 be tidied up in order to avoid warnings.
4367
4368 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4369 first of these names is deprecated in the latest Autoconf in favour of the name
4370 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4371 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4372 analogy I have changed it to CFLAGS_FOR_BUILD.
4373
4374 4. Added -no-undefined to the linking command in the Makefile, because this is
4375 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4376 linking step for the pcreposix library.
4377
4378 5. PCRE was failing to diagnose the case of two named groups with the same
4379 name.
4380
4381 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4382 literal character that is needed in the subject for a match, and scans along to
4383 ensure that it is present before embarking on the full matching process. This
4384 saves time in cases of nested unlimited repeats that are never going to match.
4385 Problem: the scan can take a lot of time if the subject is very long (e.g.
4386 megabytes), thus penalizing straightforward matches. It is now done only if the
4387 amount of subject to be scanned is less than 1000 bytes.
4388
4389 7. A lesser problem with the same optimization is that it was recording the
4390 first character of an anchored pattern as "needed", thus provoking a search
4391 right along the subject, even when the first match of the pattern was going to
4392 fail. The "needed" character is now not set for anchored patterns, unless it
4393 follows something in the pattern that is of non-fixed length. Thus, it still
4394 fulfils its original purpose of finding quick non-matches in cases of nested
4395 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4396
4397
4398 Version 4.0 17-Feb-03
4399 ---------------------
4400
4401 1. If a comment in an extended regex that started immediately after a meta-item
4402 extended to the end of string, PCRE compiled incorrect data. This could lead to
4403 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4404
4405 2. Moved to autoconf 2.53 and libtool 1.4.2.
4406
4407 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4408 the special perltest8 script is no longer needed - all the tests can be run
4409 from a single perltest script.
4410
4411 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4412 by \s. It has now been removed in PCRE. This means it isn't recognized as
4413 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4414 class [:space:] *does* include VT, thereby creating a mess.
4415
4416 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4417 space and tab.
4418
4419 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4420 its new features into the main test script, reducing the number of scripts.
4421
4422 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4423 were backward compatible, and made the (?i) apply to the whole pattern, as if
4424 /i were given. Now it behaves more logically, and applies the option setting
4425 only to what follows. PCRE has been changed to follow suit. However, if it
4426 finds options settings right at the start of the pattern, it extracts them into
4427 the global options, as before. Thus, they show up in the info data.
4428
4429 8. Added support for the \Q...\E escape sequence. Characters in between are
4430 treated as literals. This is slightly different from Perl in that $ and @ are
4431 also handled as literals inside the quotes. In Perl, they will cause variable
4432 interpolation. Note the following examples:
4433
4434 Pattern PCRE matches Perl matches
4435
4436 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4437 \Qabc\$xyz\E abc\$xyz abc\$xyz
4438 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4439
4440 For compatibility with Perl, \Q...\E sequences are recognized inside character
4441 classes as well as outside them.
4442
4443 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4444 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4445 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4446 signed/unsigned warnings.
4447
4448 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4449 option for pcretest, so I've replaced it by a simple function that does just
4450 that job.
4451
4452 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4453 "pcregrep -".
4454
4455 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4456 Java package. This provides some syntactic sugar for simple cases of what my
4457 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4458 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4459 item, you can use this simplified notation. Note that only makes sense with
4460 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4461 greediness, whatever the setting of the PCRE_UNGREEDY option.
4462
4463 13. A change of greediness default within a pattern was not taking effect at
4464 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4465 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4466 was abstracted outside.
4467
4468 14. PCRE now supports the \G assertion. It is true when the current matching
4469 position is at the start point of the match. This differs from \A when the
4470 starting offset is non-zero. Used with the /g option of pcretest (or similar
4471 code), it works in the same way as it does for Perl's /g option. If all
4472 alternatives of a regex begin with \G, the expression is anchored to the start
4473 match position, and the "anchored" flag is set in the compiled expression.
4474
4475 15. Some bugs concerning the handling of certain option changes within patterns
4476 have been fixed. These applied to options other than (?ims). For example,
4477 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4478 been the other way round. Some of this was related to change 7 above.
4479
4480 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4481 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4482 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4483 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4484
4485 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4486 mode. Unlike ".", it always matches newline, whatever the setting of
4487 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4488 assertions. Perl allows it, but it doesn't (in general) work because it can't
4489 calculate the length of the lookbehind. At least, that's the case for Perl
4490 5.8.0 - I've been told they are going to document that it doesn't work in
4491 future.
4492
4493 18. Added an error diagnosis for escapes that PCRE does not support: these are
4494 \L, \l, \N, \P, \p, \U, \u, and \X.
4495
4496 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4497 reading past the end of the pattern in cases such as /[abcd/.
4498
4499 20. PCRE was getting more memory than necessary for patterns with classes that
4500 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4501
4502 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4503 compiling PCRE for use with Virtual Pascal.
4504
4505 22. Small fix to the Makefile to make it work properly if the build is done
4506 outside the source tree.
4507
4508 23. Added a new extension: a condition to go with recursion. If a conditional
4509 subpattern starts with (?(R) the "true" branch is used if recursion has
4510 happened, whereas the "false" branch is used only at the top level.
4511
4512 24. When there was a very long string of literal characters (over 255 bytes
4513 without UTF support, over 250 bytes with UTF support), the computation of how
4514 much memory was required could be incorrect, leading to segfaults or other
4515 strange effects.
4516
4517 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4518 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4519 there was a subsequent back reference to those brackets. This meant that, for
4520 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4521 possible to check for precisely this case. All we can do is abandon the
4522 optimization if .* occurs inside capturing brackets when there are any back
4523 references whatsoever. (See below for a better fix that came later.)
4524
4525 26. The handling of the optimization for finding the first character of a
4526 non-anchored pattern, and for finding a character that is required later in the
4527 match were failing in some cases. This didn't break the matching; it just
4528 failed to optimize when it could. The way this is done has been re-implemented.
4529
4530 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4531
4532 28. Added a new feature that provides some of the functionality that Perl
4533 provides with (?{...}). The facility is termed a "callout". The way it is done
4534 in PCRE is for the caller to provide an optional function, by setting
4535 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4536 global variable. By default it is unset, which disables all calling out. To get
4537 the function called, the regex must include (?C) at appropriate points. This
4538 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4539 This provides a means of identifying different callout points. When PCRE
4540 reaches such a point in the regex, if pcre_callout has been set, the external
4541 function is called. It is provided with data in a structure called
4542 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4543 matching continues; if it returns a non-zero value, the match at the current
4544 point fails. However, backtracking will occur if possible. [This was changed
4545 later and other features added - see item 49 below.]
4546
4547 29. pcretest is upgraded to test the callout functionality. It provides a
4548 callout function that displays information. By default, it shows the start of
4549 the match and the current position in the text. There are some new data escapes
4550 to vary what happens:
4551
4552 \C+ in addition, show current contents of captured substrings
4553 \C- do not supply a callout function
4554 \C!n return 1 when callout number n is reached
4555 \C!n!m return 1 when callout number n is reached for the mth time
4556
4557 30. If pcregrep was called with the -l option and just a single file name, it
4558 output "<stdin>" if a match was found, instead of the file name.
4559
4560 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4561 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4562 pcre_exec(). This saves a malloc/free per call. The default value of
4563 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4564 when configuring.
4565
4566 32. The default maximum size of a compiled pattern is 64K. There have been a
4567 few cases of people hitting this limit. The code now uses macros to handle the
4568 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4569 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4570 configuring. Tests 2 and 5 work only with 2-byte links because they output
4571 debugging information about compiled patterns.
4572
4573 33. Internal code re-arrangements:
4574
4575 (a) Moved the debugging function for printing out a compiled regex into
4576 its own source file (printint.c) and used #include to pull it into
4577 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4578 separate copies.
4579
4580 (b) Defined the list of op-code names for debugging as a macro in
4581 internal.h so that it is next to the definition of the opcodes.
4582
4583 (c) Defined a table of op-code lengths for simpler skipping along compiled
4584 code. This is again a macro in internal.h so that it is next to the
4585 definition of the opcodes.
4586
4587 34. Added support for recursive calls to individual subpatterns, along the
4588 lines of Robin Houston's patch (but implemented somewhat differently).
4589
4590 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4591 allow it to read and process whole directories in Win32. This code was
4592 contributed by Lionel Fourquaux; it has not been tested by me.
4593
4594 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4595 used to name a group. Names consist of alphanumerics and underscores, and must
4596 be unique. Back references use the syntax (?P=name) and recursive calls use
4597 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4598 numbers. The function pcre_fullinfo() can be used after compilation to extract
4599 a name/number map. There are three relevant calls:
4600
4601 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4602 PCRE_INFO_NAMECOUNT yields the number of entries
4603 PCRE_INFO_NAMETABLE yields a pointer to the map.
4604
4605 The map is a vector of fixed-size entries. The size of each entry depends on
4606 the length of the longest name used. The first two bytes of each entry are the
4607 group number, most significant byte first. There follows the corresponding
4608 name, zero terminated. The names are in alphabetical order.
4609
4610 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4611 case instead of 255. Making it the same both with and without UTF-8 support
4612 means that the same test output works with both.
4613
4614 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4615 calling malloc() with a zero argument.
4616
4617 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4618 optimization. I've improved things by keeping a bitmap of backreferences with
4619 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4620 fact referenced, the optimization can be applied. It is unlikely that a
4621 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4622 the match to follow \n) will appear inside brackets with a number greater than
4623 31, but if it does, any back reference > 31 suppresses the optimization.
4624
4625 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4626 of disabling numbered capturing parentheses. Any opening parenthesis that is
4627 not followed by ? behaves as if it were followed by ?: but named parentheses
4628 can still be used for capturing (and they will acquire numbers in the usual
4629 way).
4630
4631 41. Redesigned the return codes from the match() function into yes/no/error so
4632 that errors can be passed back from deep inside the nested calls. A malloc
4633 failure while inside a recursive subpattern call now causes the
4634 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4635
4636 42. It is now possible to set a limit on the number of times the match()
4637 function is called in a call to pcre_exec(). This facility makes it possible to
4638 limit the amount of recursion and backtracking, though not in a directly
4639 obvious way, because the match() function is used in a number of different
4640 circumstances. The count starts from zero for each position in the subject
4641 string (for non-anchored patterns). The default limit is, for compatibility, a
4642 large number, namely 10 000 000. You can change this in two ways:
4643
4644 (a) When configuring PCRE before making, you can use --with-match-limit=n
4645 to set a default value for the compiled library.
4646
4647 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4648 a different value is set. See 45 below.
4649
4650 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4651
4652 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4653 of things that can be changed at compile time. The first argument specifies
4654 what is wanted and the second points to where the information is to be placed.
4655 The current list of available information is:
4656
4657 PCRE_CONFIG_UTF8
4658
4659 The output is an integer that is set to one if UTF-8 support is available;
4660 otherwise it is set to zero.
4661
4662 PCRE_CONFIG_NEWLINE
4663
4664 The output is an integer that it set to the value of the code that is used for
4665 newline. It is either LF (10) or CR (13).
4666
4667 PCRE_CONFIG_LINK_SIZE
4668
4669 The output is an integer that contains the number of bytes used for internal
4670 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4671
4672 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4673
4674 The output is an integer that contains the threshold above which the POSIX
4675 interface uses malloc() for output vectors. See item 31 above.
4676
4677 PCRE_CONFIG_MATCH_LIMIT
4678
4679 The output is an unsigned integer that contains the default limit of the number
4680 of match() calls in a pcre_exec() execution. See 42 above.
4681
4682 44. pcretest has been upgraded by the addition of the -C option. This causes it
4683 to extract all the available output from the new pcre_config() function, and to
4684 output it. The program then exits immediately.
4685
4686 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4687 order to support additional features. One way would have been to define
4688 pcre_exec2() (for example) with extra arguments, but this would not have been
4689 extensible, and would also have required all calls to the original function to
4690 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4691 is used for passing in "extra" data from pcre_study().
4692
4693 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4694 contains the following fields:
4695
4696 flags a bitmap indicating which of the following fields are set
4697 study_data opaque data from pcre_study()
4698 match_limit a way of specifying a limit on match() calls for a specific
4699 call to pcre_exec()
4700 callout_data data for callouts (see 49 below)
4701
4702 The flag bits are also defined in pcre.h, and are
4703
4704 PCRE_EXTRA_STUDY_DATA
4705 PCRE_EXTRA_MATCH_LIMIT
4706 PCRE_EXTRA_CALLOUT_DATA
4707
4708 The pcre_study() function now returns one of these new pcre_extra blocks, with
4709 the actual study data pointed to by the study_data field, and the
4710 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4711 before. That is, this change is entirely upwards-compatible and requires no
4712 change to existing code.
4713
4714 If you want to pass in additional data to pcre_exec(), you can either place it
4715 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4716 block.
4717
4718 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4719 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4720 times with different match limits, until it finds the minimum value needed for
4721 pcre_exec() to complete. The value is then output. This can be instructive; for
4722 most simple matches the number is quite small, but for pathological cases it
4723 gets very large very quickly.
4724
4725 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4726 returns the size of the data block pointed to by the study_data field in a
4727 pcre_extra block, that is, the value that was passed as the argument to
4728 pcre_malloc() when PCRE was getting memory in which to place the information
4729 created by pcre_study(). The fourth argument should point to a size_t variable.
4730 pcretest has been extended so that this information is shown after a successful
4731 pcre_study() call when information about the compiled regex is being displayed.
4732
4733 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4734 because what follows is always an absolute path. (Later: it turns out that this
4735 is more than cosmetic for MinGW, because it doesn't like empty path
4736 components.)
4737
4738 49. Some changes have been made to the callout feature (see 28 above):
4739
4740 (i) A callout function now has three choices for what it returns:
4741
4742 0 => success, carry on matching
4743 > 0 => failure at this point, but backtrack if possible
4744 < 0 => serious error, return this value from pcre_exec()
4745
4746 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4747 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4748 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4749 use by callout functions. It will never be used by PCRE itself.
4750
4751 (ii) The pcre_extra structure (see 45 above) has a void * field called
4752 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4753 pcre_callout_block structure has a field of the same name. The contents of
4754 the field passed in the pcre_extra structure are passed to the callout
4755 function in the corresponding field in the callout block. This makes it
4756 easier to use the same callout-containing regex from multiple threads. For
4757 testing, the pcretest program has a new data escape
4758
4759 \C*n pass the number n (may be negative) as callout_data
4760
4761 If the callout function in pcretest receives a non-zero value as
4762 callout_data, it returns that value.
4763
4764 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4765 there were some redundant $(CFLAGS) in commands that are now specified as
4766 $(LINK), which already includes $(CFLAGS).
4767
4768 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4769 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4770 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4771 one-byte characters throughout. Note that case-insensitive matching applies
4772 only to characters whose values are less than 256. PCRE doesn't support the
4773 notion of cases for higher-valued characters.
4774
4775 (i) A character class whose characters are all within 0-255 is handled as
4776 a bit map, and the map is inverted for negative classes. Previously, a
4777 character > 255 always failed to match such a class; however it should
4778 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4779
4780 (ii) A negated character class with a single character < 255 is coded as
4781 "not this character" (OP_NOT). This wasn't working properly when the test
4782 character was multibyte, either singly or repeated.
4783
4784 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4785 mode, for example: \x{100}{2,3}.
4786
4787 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4788 singly or repeated) now correctly test multibyte characters. However,
4789 PCRE doesn't recognize any characters with values greater than 255 as
4790 digits, spaces, or word characters. Such characters always match \D, \S,
4791 and \W, and never match \d, \s, or \w.
4792
4793 (v) Classes may now contain characters and character ranges with values
4794 greater than 255. For example: [ab\x{100}-\x{400}].
4795
4796 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4797 PCRE in UTF-8 mode.
4798
4799 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4800 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4801 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4802 value.)
4803
4804 53. The single man page has become too large. I have therefore split it up into
4805 a number of separate man pages. These also give rise to individual HTML pages;
4806 these are now put in a separate directory, and there is an index.html page that
4807 lists them all. Some hyperlinking between the pages has been installed.
4808
4809 54. Added convenience functions for handling named capturing parentheses.
4810
4811 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4812 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4813 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4814 are faulted.
4815
4816 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4817 calling configure. These values are used when compiling the dftables.c program
4818 which is run to generate the source of the default character tables. They
4819 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4820 you will need to set these values.
4821
4822 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4823
4824
4825 Version 3.9 02-Jan-02
4826 ---------------------
4827
4828 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4829
4830 2. If --disable-static was given, the building process failed when trying to
4831 build pcretest and pcregrep. (For some reason it was using libtool to compile
4832 them, which is not right, as they aren't part of the library.)
4833
4834
4835 Version 3.8 18-Dec-01
4836 ---------------------
4837
4838 1. The experimental UTF-8 code was completely screwed up. It was packing the
4839 bytes in the wrong order. How dumb can you get?
4840
4841
4842 Version 3.7 29-Oct-01
4843 ---------------------
4844
4845 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4846 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4847 this didn't happen under Solaris 8, where I normally test things.
4848
4849 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4850 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4851 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4852 this fix an hour or so after the initial 3.7 release.)
4853
4854
4855 Version 3.6 23-Oct-01
4856 ---------------------
4857
4858 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4859 offsets passed as NULL with zero offset count.
4860
4861 2. The config.guess and config.sub files had not been updated when I moved to
4862 the latest autoconf.
4863
4864
4865 Version 3.5 15-Aug-01
4866 ---------------------
4867
4868 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4869 had been forgotten.
4870
4871 2. By using declared but undefined structures, we can avoid using "void"
4872 definitions in pcre.h while keeping the internal definitions of the structures
4873 private.
4874
4875 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4876 user point of view, this means that both static and shared libraries are built
4877 by default, but this can be individually controlled. More of the work of
4878 handling this static/shared cases is now inside libtool instead of PCRE's make
4879 file.
4880
4881 4. The pcretest utility is now installed along with pcregrep because it is
4882 useful for users (to test regexs) and by doing this, it automatically gets
4883 relinked by libtool. The documentation has been turned into a man page, so
4884 there are now .1, .txt, and .html versions in /doc.
4885
4886 5. Upgrades to pcregrep:
4887 (i) Added long-form option names like gnu grep.
4888 (ii) Added --help to list all options with an explanatory phrase.
4889 (iii) Added -r, --recursive to recurse into sub-directories.
4890 (iv) Added -f, --file to read patterns from a file.
4891
4892 6. pcre_exec() was referring to its "code" argument before testing that
4893 argument for NULL (and giving an error if it was NULL).
4894
4895 7. Upgraded Makefile.in to allow for compiling in a different directory from
4896 the source directory.
4897
4898 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4899 options bits, the pointer it was passed was to an int instead of to an unsigned
4900 long int. This mattered only on 64-bit systems.
4901
4902 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4903 generated) instead of pcre.in, which it its source. Also made the same change
4904 in several of the .c files.
4905
4906 10. A new release of gcc defines printf() as a macro, which broke pcretest
4907 because it had an ifdef in the middle of a string argument for printf(). Fixed
4908 by using separate calls to printf().
4909
4910 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4911 script, to force use of CR or LF instead of \n in the source. On non-Unix
4912 systems, the value can be set in config.h.
4913
4914 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4915 absolute limit. Changed the text of the error message to make this clear, and
4916 likewise updated the man page.
4917
4918 13. The limit of 99 on the number of capturing subpatterns has been removed.
4919 The new limit is 65535, which I hope will not be a "real" limit.
4920
4921
4922 Version 3.4 22-Aug-00
4923 ---------------------
4924
4925 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4926
4927 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4928
4929
4930 Version 3.3 01-Aug-00
4931 ---------------------
4932
4933 1. If an octal character was given, but the value was greater than \377, it
4934 was not getting masked to the least significant bits, as documented. This could
4935 lead to crashes in some systems.
4936
4937 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4938 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4939
4940 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4941 These just pass their arguments on to (pcre_free)(), but they are provided
4942 because some uses of PCRE bind it to non-C systems that can call its functions,
4943 but cannot call free() or pcre_free() directly.
4944
4945 4. Add "make test" as a synonym for "make check". Corrected some comments in
4946 the Makefile.
4947
4948 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4949 Makefile.
4950
4951 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4952 command called pgrep for grepping around the active processes.
4953
4954 7. Added the beginnings of support for UTF-8 character strings.
4955
4956 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4957 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4958 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4959 out for the ar command.)
4960
4961
4962 Version 3.2 12-May-00
4963 ---------------------
4964
4965 This is purely a bug fixing release.
4966
4967 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4968 of ZA. This was just one example of several cases that could provoke this bug,
4969 which was introduced by change 9 of version 2.00. The code for breaking
4970 infinite loops after an iteration that matches an empty string was't working
4971 correctly.
4972
4973 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4974 when matched against abbab (for example). After matching an empty string, it
4975 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4976 caused it to match further down the string than it should.
4977
4978 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4979 was there because it doesn't seem to be needed, and it causes trouble on some
4980 systems, as it is not a Standard C header. It has been removed.
4981
4982 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4983 were reported on the Macintosh. The changes were from
4984
4985 while ((c = *(++ptr)) != 0 && c != '\n');
4986 to
4987 while ((c = *(++ptr)) != 0 && c != '\n') ;
4988
4989 Totally extraordinary, but if that's what it takes...
4990
4991 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4992 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4993 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4994 assumes the way PCRE uses memmove() (always moving upwards).
4995
4996 6. PCRE is being used in one environment where strchr() is not available. There
4997 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4998 faster code anyway.
4999
5000
5001 Version 3.1 09-Feb-00
5002 ---------------------
5003
5004 The only change in this release is the fixing of some bugs in Makefile.in for
5005 the "install" target:
5006
5007 (1) It was failing to install pcreposix.h.
5008
5009 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5010
5011
5012 Version 3.0 01-Feb-00
5013 ---------------------
5014
5015 1. Add support for the /+ modifier to perltest (to output $` like it does in
5016 pcretest).
5017
5018 2. Add support for the /g modifier to perltest.
5019
5020 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5021 matches null strings.
5022
5023 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5024 pattern. Perl treats empty patterns specially - it reuses the most recent
5025 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5026 effect.
5027
5028 5. The POSIX interface was broken in that it was just handing over the POSIX
5029 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5030 required a bigger vector, with some working space on the end. This means that
5031 the POSIX wrapper now has to get and free some memory, and copy the results.
5032
5033 6. Added some simple autoconf support, placing the test data and the
5034 documentation in separate directories, re-organizing some of the
5035 information files, and making it build pcre-config (a GNU standard). Also added
5036 libtool support for building PCRE as a shared library, which is now the
5037 default.
5038
5039 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5040 09 are not valid octal constants. Single digits will be used for minor values
5041 less than 10.
5042
5043 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5044 existing programs that set these in the POSIX interface can use PCRE without
5045 modification.
5046
5047 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5048 return all that pcre_info() returns, plus additional data. The pcre_info()
5049 function is retained for compatibility, but is considered to be obsolete.
5050
5051 10. Added experimental recursion feature (?R) to handle one common case that
5052 Perl 5.6 will be able to do with (?p{...}).
5053
5054 11. Added support for POSIX character classes like [:alpha:], which Perl is
5055 adopting.
5056
5057
5058 Version 2.08 31-Aug-99
5059 ----------------------
5060
5061 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5062 trying to match at the startoffset position, but instead was moving forward to
5063 the next newline as if a previous match had failed.
5064
5065 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5066 and could get into a loop if a null string was matched other than at the start
5067 of the subject.
5068
5069 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5070 be distinguished at compile time, and for completeness also added PCRE_DATE.
5071
5072 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5073 in GnuWin32 environments.
5074
5075
5076 Version 2.07 29-Jul-99
5077 ----------------------
5078
5079 1. The documentation is now supplied in plain text form and HTML as well as in
5080 the form of man page sources.
5081
5082 2. C++ compilers don't like assigning (void *) values to other pointer types.
5083 In particular this affects malloc(). Although there is no problem in Standard
5084 C, I've put in casts to keep C++ compilers happy.
5085
5086 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5087 should be (const char *).
5088
5089 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5090 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5091 However, I haven't made this a standard facility. The documentation doesn't
5092 mention it, and the Makefile doesn't support it.
5093
5094 5. The Makefile now contains an "install" target, with editable destinations at
5095 the top of the file. The pcretest program is not installed.
5096
5097 6. pgrep -V now gives the PCRE version number and date.
5098
5099 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5100 causing the entire string to be ignored, instead of just the last character.
5101
5102 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5103 non-matching string, it can take a very, very long time, even for strings of
5104 quite modest length, because of the nested recursion. PCRE now does better in
5105 some of these cases. It does this by remembering the last required literal
5106 character in the pattern, and pre-searching the subject to ensure it is present
5107 before running the real match. In other words, it applies a heuristic to detect
5108 some types of certain failure quickly, and in the above example, if presented
5109 with a string that has no trailing " it gives "no match" very quickly.
5110
5111 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5112 other alternatives are tried instead.
5113
5114
5115 Version 2.06 09-Jun-99
5116 ----------------------
5117
5118 1. Change pcretest's output for amount of store used to show just the code
5119 space, because the remainder (the data block) varies in size between 32-bit and
5120 64-bit systems.
5121
5122 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5123 start matching at. This allows lookbehinds to work when searching for multiple
5124 occurrences in a string.
5125
5126 3. Added additional options to pcretest for testing multiple occurrences:
5127
5128 /+ outputs the rest of the string that follows a match
5129 /g loops for multiple occurrences, using the new startoffset argument
5130 /G loops for multiple occurrences by passing an incremented pointer
5131
5132 4. PCRE wasn't doing the "first character" optimization for patterns starting
5133 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5134 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5135 the letter 'x'. On long subject strings, this gives a significant speed-up.
5136
5137
5138 Version 2.05 21-Apr-99
5139 ----------------------
5140
5141 1. Changed the type of magic_number from int to long int so that it works
5142 properly on 16-bit systems.
5143
5144 2. Fixed a bug which caused patterns starting with .* not to work correctly
5145 when the subject string contained newline characters. PCRE was assuming
5146 anchoring for such patterns in all cases, which is not correct because .* will
5147 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5148 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5149 must be retried after every newline in the subject.
5150
5151
5152 Version 2.04 18-Feb-99
5153 ----------------------
5154
5155 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5156 computation of the store needed to hold the pattern was incorrect (too large).
5157 If such patterns were nested a few deep, this could multiply and become a real
5158 problem.
5159
5160 2. Added /M option to pcretest to show the memory requirement of a specific
5161 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5162
5163 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5164 compiled in such a way that the backtracking after subsequent failure was
5165 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5166 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5167
5168
5169 Version 2.03 02-Feb-99
5170 ----------------------
5171
5172 1. Fixed typo and small mistake in man page.
5173
5174 2. Added 4th condition (GPL supersedes if conflict) and created separate
5175 LICENCE file containing the conditions.
5176
5177 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5178 Perl, that is the internal \ allows the delimiter to be included in the
5179 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5180 the final delimiter, add \ to the end of the pattern (to test the error).
5181
5182 4. Added the convenience functions for extracting substrings after a successful
5183 match. Updated pcretest to make it able to test these functions.
5184
5185
5186 Version 2.02 14-Jan-99
5187 ----------------------
5188
5189 1. Initialized the working variables associated with each extraction so that
5190 their saving and restoring doesn't refer to uninitialized store.
5191
5192 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5193 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5194 fix the problem.
5195
5196 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5197 calls, and wasn't printing the correct value for compiling calls. Increased the
5198 default value of LOOPREPEAT, and the number of significant figures in the
5199 times.
5200
5201 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5202
5203 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5204 a building problem on Windows NT with a FAT file system.
5205
5206
5207 Version 2.01 21-Oct-98
5208 ----------------------
5209
5210 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5211 to character tables built by pcre_maketables() in the current locale. If NULL
5212 is passed, the default tables are used.
5213
5214
5215 Version 2.00 24-Sep-98
5216 ----------------------
5217
5218 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5219 it any more.
5220
5221 2. Allow quantification of (?>) groups, and make it work correctly.
5222
5223 3. The first character computation wasn't working for (?>) groups.
5224
5225 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5226 end of the subject) and add 5.005's \z, which really does match only at the
5227 very end of the subject.
5228
5229 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5230
5231 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5232 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5233 localized options. All options to pcre_study() were also removed.
5234
5235 7. Add other new features from 5.005:
5236
5237 $(?<= positive lookbehind
5238 $(?<! negative lookbehind
5239 (?imsx-imsx) added the unsetting capability
5240 such a setting is global if at outer level; local otherwise
5241 (?imsx-imsx:) non-capturing groups with option setting
5242 (?(cond)re|re) conditional pattern matching
5243
5244 A backreference to itself in a repeated group matches the previous
5245 captured string.
5246
5247 8. General tidying up of studying (both automatic and via "study")
5248 consequential on the addition of new assertions.
5249
5250 9. As in 5.005, unlimited repeated groups that could match an empty substring
5251 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5252 runtime if any iteration does actually match an empty substring.
5253
5254 10. Include the RunTest script in the distribution.
5255
5256 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5257 discrepancies, some of which were old and were also with respect to 5.004. They
5258 have now been fixed.
5259
5260
5261 Version 1.09 28-Apr-98
5262 ----------------------
5263
5264 1. A negated single character class followed by a quantifier with a minimum
5265 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5266 program crashes, or just wrong answers. This did not apply to negated classes
5267 containing more than one character, or to minima other than one.
5268
5269
5270 Version 1.08 27-Mar-98
5271 ----------------------
5272
5273 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5274
5275 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5276 latter must appear before anything that relies on it in the pattern.
5277
5278
5279 Version 1.07 16-Feb-98
5280 ----------------------
5281
5282 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5283 repeat of a potentially empty string).
5284
5285
5286 Version 1.06 23-Jan-98
5287 ----------------------
5288
5289 1. Added Markus Oberhumer's little patches for C++.
5290
5291 2. Literal strings longer than 255 characters were broken.
5292
5293
5294 Version 1.05 23-Dec-97
5295 ----------------------
5296
5297 1. Negated character classes containing more than one character were failing if
5298 PCRE_CASELESS was set at run time.
5299
5300
5301 Version 1.04 19-Dec-97
5302 ----------------------
5303
5304 1. Corrected the man page, where some "const" qualifiers had been omitted.
5305
5306 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5307 input syntax.
5308
5309 3. Fixed memory leak which occurred when a regex with back references was
5310 matched with an offsets vector that wasn't big enough. The temporary memory
5311 that is used in this case wasn't being freed if the match failed.
5312
5313 4. Tidied pcretest to ensure it frees memory that it gets.
5314
5315 5. Temporary memory was being obtained in the case where the passed offsets
5316 vector was exactly big enough.
5317
5318 6. Corrected definition of offsetof() from change 5 below.
5319
5320 7. I had screwed up change 6 below and broken the rules for the use of
5321 setjmp(). Now fixed.
5322
5323
5324 Version 1.03 18-Dec-97
5325 ----------------------
5326
5327 1. A erroneous regex with a missing opening parenthesis was correctly
5328 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5329 on some systems.
5330
5331 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5332 it was reported that one broken compiler failed on the former because "code" is
5333 also an independent variable.
5334
5335 3. The erroneous regex a[]b caused an array overrun reference.
5336
5337 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5338 fail on data ending with that character. (It was going on too far, and checking
5339 the next character, typically a binary zero.) This was specific to the
5340 optimized code for single-character negative classes.
5341
5342 5. Added a contributed patch from the TIN world which does the following:
5343
5344 + Add an undef for memmove, in case the the system defines a macro for it.
5345
5346 + Add a definition of offsetof(), in case there isn't one. (I don't know
5347 the reason behind this - offsetof() is part of the ANSI standard - but
5348 it does no harm).
5349
5350 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5351 most of the places where whitespace preceded '#'. I have given up and
5352 allowed the remaining 2 cases to be at the margin.
5353
5354 + Rename some variables in pcre to eliminate shadowing. This seems very
5355 pedantic, but does no harm, of course.
5356
5357 6. Moved the call to setjmp() into its own function, to get rid of warnings
5358 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5359
5360 7. Constructs such as \d{8,} were compiling into the equivalent of
5361 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5362 outcome, but in this particular case used more store than had been allocated,
5363 which caused the bug to be discovered because it threw up an internal error.
5364
5365 8. The debugging code in both pcre and pcretest for outputting the compiled
5366 form of a regex was going wrong in the case of back references followed by
5367 curly-bracketed repeats.
5368
5369
5370 Version 1.02 12-Dec-97
5371 ----------------------
5372
5373 1. Typos in pcre.3 and comments in the source fixed.
5374
5375 2. Applied a contributed patch to get rid of places where it used to remove
5376 'const' from variables, and fixed some signed/unsigned and uninitialized
5377 variable warnings.
5378
5379 3. Added the "runtest" target to Makefile.
5380
5381 4. Set default compiler flag to -O2 rather than just -O.
5382
5383
5384 Version 1.01 19-Nov-97
5385 ----------------------
5386
5387 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5388 like /([ab]*)*/, that is, for classes with more than one character in them.
5389
5390 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5391 as /((?>a*))*/ (a PCRE_EXTRA facility).
5392
5393
5394 Version 1.00 18-Nov-97
5395 ----------------------
5396
5397 1. Added compile-time macros to support systems such as SunOS4 which don't have
5398 memmove() or strerror() but have other things that can be used instead.
5399
5400 2. Arranged that "make clean" removes the executables.
5401
5402
5403 Version 0.99 27-Oct-97
5404 ----------------------
5405
5406 1. Fixed bug in code for optimizing classes with only one character. It was
5407 initializing a 32-byte map regardless, which could cause it to run off the end
5408 of the memory it had got.
5409
5410 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5411
5412
5413 Version 0.98 22-Oct-97
5414 ----------------------
5415
5416 1. Fixed bug in code for handling temporary memory usage when there are more
5417 back references than supplied space in the ovector. This could cause segfaults.
5418
5419
5420 Version 0.97 21-Oct-97
5421 ----------------------
5422
5423 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5424
5425 2. Optimized negated single characters not to use a bit map.
5426
5427 3. Brought error texts together as macro definitions; clarified some of them;
5428 fixed one that was wrong - it said "range out of order" when it meant "invalid
5429 escape sequence".
5430
5431 4. Changed some char * arguments to const char *.
5432
5433 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5434
5435 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5436 pcretest.
5437
5438
5439 Version 0.96 16-Oct-97
5440 ----------------------
5441
5442 1. Added a simple "pgrep" utility to the distribution.
5443
5444 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5445 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5446 where "ddd" means "one or more decimal digits".
5447
5448 3. Fixed serious bug. If a pattern had a back reference, but the call to
5449 pcre_exec() didn't supply a large enough ovector to record the related
5450 identifying subpattern, the match always failed. PCRE now remembers the number
5451 of the largest back reference, and gets some temporary memory in which to save
5452 the offsets during matching if necessary, in order to ensure that
5453 backreferences always work.
5454
5455 4. Increased the compatibility with Perl in a number of ways:
5456
5457 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5458 to request this handling. The option can be set at compile or exec time.
5459
5460 (b) $ matches before a terminating newline by default; an option
5461 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5462 mode). The option can be set at compile or exec time.
5463
5464 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5465 the same as Perl's. If the decimal number it represents is less than 10
5466 or there aren't that many previous left capturing parentheses, an octal
5467 escape is read. Inside a character class, it's always an octal escape,
5468 even if it is a single digit.
5469
5470 (d) An escaped but undefined alphabetic character is taken as a literal,
5471 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5472 escapes.
5473
5474 (e) {0} is now permitted. (The previous item is removed from the compiled
5475 pattern).
5476
5477 5. Changed all the names of code files so that the basic parts are no longer
5478 than 10 characters, and abolished the teeny "globals.c" file.
5479
5480 6. Changed the handling of character classes; they are now done with a 32-byte
5481 bit map always.
5482
5483 7. Added the -d and /D options to pcretest to make it possible to look at the
5484 internals of compilation without having to recompile pcre.
5485
5486
5487 Version 0.95 23-Sep-97
5488 ----------------------
5489
5490 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5491 \x20 at the start of a run of normal characters. These were being treated as
5492 real characters, instead of the source characters being re-checked.
5493
5494
5495 Version 0.94 18-Sep-97
5496 ----------------------
5497
5498 1. The functions are now thread-safe, with the caveat that the global variables
5499 containing pointers to malloc() and free() or alternative functions are the
5500 same for all threads.
5501
5502 2. Get pcre_study() to generate a bitmap of initial characters for non-
5503 anchored patterns when this is possible, and use it if passed to pcre_exec().
5504
5505
5506 Version 0.93 15-Sep-97
5507 ----------------------
5508
5509 1. /(b)|(:+)/ was computing an incorrect first character.
5510
5511 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5512 but not actually doing anything yet.
5513
5514 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5515 as Perl does (e.g. [-az] or [az-]).
5516
5517 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5518 all possible positions.
5519
5520 5. Split up into different modules to avoid including unneeded functions in a
5521 compiled binary. However, compile and exec are still in one module. The "study"
5522 function is split off.
5523
5524 6. The character tables are now in a separate module whose source is generated
5525 by an auxiliary program - but can then be edited by hand if required. There are
5526 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5527 toupper() in the code.
5528
5529 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5530 make them global. Abolish the function for setting them, as the caller can now
5531 set them directly.
5532
5533
5534 Version 0.92 11-Sep-97
5535 ----------------------
5536
5537 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5538 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5539
5540 2. Caseless matching was not working in character classes if the characters in
5541 the pattern were in upper case.
5542
5543 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5544
5545 4. Make PCRE_ANCHORED public and accept as a compile option.
5546
5547 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5548 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5549 pass them.
5550
5551 6. Give an error if bad option bits passed at compile or run time.
5552
5553 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5554 pcretest to cause it to pass that flag.
5555
5556 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5557 options, and the first character, if set.
5558
5559 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5560
5561
5562 Version 0.91 10-Sep-97
5563 ----------------------
5564
5565 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5566 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5567
5568 2. PCRE was looping on encountering an indefinitely repeated back reference to
5569 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5570 Perl does - treats the match as successful.
5571
5572 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12