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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1504 - (show annotations) (download)
Fri Sep 26 09:21:46 2014 UTC (5 weeks ago) by ph10
File size: 268478 byte(s)
Final tidies for 8.36

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.36 26-September-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 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
99 Patch by Mike Frysinger.
100
101
102 Version 8.35 04-April-2014
103 --------------------------
104
105 1. A new flag is set, when property checks are present in an XCLASS.
106 When this flag is not set, PCRE can perform certain optimizations
107 such as studying these XCLASS-es.
108
109 2. The auto-possessification of character sets were improved: a normal
110 and an extended character set can be compared now. Furthermore
111 the JIT compiler optimizes more character set checks.
112
113 3. Got rid of some compiler warnings for potentially uninitialized variables
114 that show up only when compiled with -O2.
115
116 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
117 of a match later then the end of the match. The pcretest program was not
118 handling the case sensibly - it was outputting from the start to the next
119 binary zero. It now reports this situation in a message, and outputs the
120 text from the end to the start.
121
122 5. Fast forward search is improved in JIT. Instead of the first three
123 characters, any three characters with fixed position can be searched.
124 Search order: first, last, middle.
125
126 6. Improve character range checks in JIT. Characters are read by an inprecise
127 function now, which returns with an unknown value if the character code is
128 above a certain threshold (e.g: 256). The only limitation is that the value
129 must be bigger than the threshold as well. This function is useful when
130 the characters above the threshold are handled in the same way.
131
132 7. The macros whose names start with RAWUCHAR are placeholders for a future
133 mode in which only the bottom 21 bits of 32-bit data items are used. To
134 make this more memorable for those maintaining the code, the names have
135 been changed to start with UCHAR21, and an extensive comment has been added
136 to their definition.
137
138 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
139 to the export list in Makefile.am (they were accidentally omitted from the
140 8.34 tarball).
141
142 9. The informational output from pcretest used the phrase "starting byte set"
143 which is inappropriate for the 16-bit and 32-bit libraries. As the output
144 for "first char" and "need char" really means "non-UTF-char", I've changed
145 "byte" to "char", and slightly reworded the output. The documentation about
146 these values has also been (I hope) clarified.
147
148 10. Another JIT related optimization: use table jumps for selecting the correct
149 backtracking path, when more than four alternatives are present inside a
150 bracket.
151
152 11. Empty match is not possible, when the minimum length is greater than zero,
153 and there is no \K in the pattern. JIT should avoid empty match checks in
154 such cases.
155
156 12. In a caseless character class with UCP support, when a character with more
157 than one alternative case was not the first character of a range, not all
158 the alternative cases were added to the class. For example, s and \x{17f}
159 are both alternative cases for S: the class [RST] was handled correctly,
160 but [R-T] was not.
161
162 13. The configure.ac file always checked for pthread support when JIT was
163 enabled. This is not used in Windows, so I have put this test inside a
164 check for the presence of windows.h (which was already tested for).
165
166 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
167 The algorithm provides a way to skip certain starting offsets, and usually
168 faster than linear prefix searches.
169
170 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
171 as for 'fr_FR' and 'french'. For some reason, however, it then used the
172 Windows-specific input and output files, which have 'french' screwed in.
173 So this could never have worked. One of the problems with locales is that
174 they aren't always the same. I have now updated RunTest so that it checks
175 the output of the locale test (test 3) against three different output
176 files, and it allows the test to pass if any one of them matches. With luck
177 this should make the test pass on some versions of Solaris where it was
178 failing. Because of the uncertainty, the script did not used to stop if
179 test 3 failed; it now does. If further versions of a French locale ever
180 come to light, they can now easily be added.
181
182 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
183 there was a message during ./configure, but it did not stop. This now
184 provokes an error. The invalid example in README has been corrected.
185 If a value less than the minimum is given, the minimum value has always
186 been used, but now a warning is given.
187
188 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
189 was a bug in the test system, which is now fixed. Also, the list of various
190 configurations that are tested for each release did not have one with both
191 16/32 bits and --enable-bar-anycrlf. It now does.
192
193 18. pcretest was missing "-C bsr" for displaying the \R default setting.
194
195 19. Little endian PowerPC systems are supported now by the JIT compiler.
196
197 20. The fast forward newline mechanism could enter to an infinite loop on
198 certain invalid UTF-8 input. Although we don't support these cases
199 this issue can be fixed by a performance optimization.
200
201 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
202 not take account if existing stack usage. There is now a new global
203 variable called pcre_stack_guard that can be set to point to an external
204 function to check stack availability. It is called at the start of
205 processing every parenthesized group.
206
207 22. A typo in the code meant that in ungreedy mode the max/min qualifier
208 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
209 not match "ab".
210
211 23. When UTF was disabled, the JIT program reported some incorrect compile
212 errors. These messages are silenced now.
213
214 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
215 compiler.
216
217 25. Change all the temporary files used in RunGrepTest to be different to those
218 used by RunTest so that the tests can be run simultaneously, for example by
219 "make -j check".
220
221
222 Version 8.34 15-December-2013
223 -----------------------------
224
225 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
226 executable memory. Patch inspired by Carsten Klein.
227
228 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
229 this macro is never tested and has no effect, because the work to support
230 coverage involves only compiling and linking options and special targets in
231 the Makefile. The comment in config.h implied that defining the macro would
232 enable coverage support, which is totally false. There was also support for
233 setting this macro in the CMake files (my fault, I just copied it from
234 configure). SUPPORT_GCOV has now been removed.
235
236 3. Make a small performance improvement in strlen16() and strlen32() in
237 pcretest.
238
239 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
240 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
241 these cases). There was also one in pcretest.c.
242
243 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
244
245 6. In UTF mode, the code for checking whether a group could match an empty
246 string (which is used for indefinitely repeated groups to allow for
247 breaking an infinite loop) was broken when the group contained a repeated
248 negated single-character class with a character that occupied more than one
249 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
250 UTF-8 mode). The effect was undefined: the group might or might not be
251 deemed as matching an empty string, or the program might have crashed.
252
253 7. The code for checking whether a group could match an empty string was not
254 recognizing that \h, \H, \v, \V, and \R must match a character.
255
256 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
257 an empty string. If it can, pcretest shows this in its information output.
258
259 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
260 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
261 matched by pcre_exec() without using JIT:
262
263 (a) If the rest of the pattern did not match after a maximal run of
264 grapheme clusters, the code for backing up to try with fewer of them
265 did not always back up over a full grapheme when characters that do not
266 have the modifier quality were involved, e.g. Hangul syllables.
267
268 (b) If the match point in a subject started with modifier character, and
269 there was no match, the code could incorrectly back up beyond the match
270 point, and potentially beyond the first character in the subject,
271 leading to a segfault or an incorrect match result.
272
273 10. A conditional group with an assertion condition could lead to PCRE
274 recording an incorrect first data item for a match if no other first data
275 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
276 first data item, and therefore matched "ca" after "c" instead of at the
277 start.
278
279 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
280 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
281
282 12. The source of pcregrep now includes z/OS-specific code so that it can be
283 compiled for z/OS as part of the special z/OS distribution.
284
285 13. Added the -T and -TM options to pcretest.
286
287 14. The code in pcre_compile.c for creating the table of named capturing groups
288 has been refactored. Instead of creating the table dynamically during the
289 actual compiling pass, the information is remembered during the pre-compile
290 pass (on the stack unless there are more than 20 named groups, in which
291 case malloc() is used) and the whole table is created before the actual
292 compile happens. This has simplified the code (it is now nearly 150 lines
293 shorter) and prepared the way for better handling of references to groups
294 with duplicate names.
295
296 15. A back reference to a named subpattern when there is more than one of the
297 same name now checks them in the order in which they appear in the pattern.
298 The first one that is set is used for the reference. Previously only the
299 first one was inspected. This change makes PCRE more compatible with Perl.
300
301 16. Unicode character properties were updated from Unicode 6.3.0.
302
303 17. The compile-time code for auto-possessification has been refactored, based
304 on a patch by Zoltan Herczeg. It now happens after instead of during
305 compilation. The code is cleaner, and more cases are handled. The option
306 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
307 options in pcretest are provided to set it. It can also be set by
308 (*NO_AUTO_POSSESS) at the start of a pattern.
309
310 18. The character VT has been added to the default ("C" locale) set of
311 characters that match \s and are generally treated as white space,
312 following this same change in Perl 5.18. There is now no difference between
313 "Perl space" and "POSIX space". Whether VT is treated as white space in
314 other locales depends on the locale.
315
316 19. The code for checking named groups as conditions, either for being set or
317 for being recursed, has been refactored (this is related to 14 and 15
318 above). Processing unduplicated named groups should now be as fast at
319 numerical groups, and processing duplicated groups should be faster than
320 before.
321
322 20. Two patches to the CMake build system, by Alexander Barkov:
323
324 (1) Replace the "source" command by "." in CMakeLists.txt because
325 "source" is a bash-ism.
326
327 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
328 without these the CMake build does not work on Solaris.
329
330 21. Perl has changed its handling of \8 and \9. If there is no previously
331 encountered capturing group of those numbers, they are treated as the
332 literal characters 8 and 9 instead of a binary zero followed by the
333 literals. PCRE now does the same.
334
335 22. Following Perl, added \o{} to specify codepoints in octal, making it
336 possible to specify values greater than 0777 and also making them
337 unambiguous.
338
339 23. Perl now gives an error for missing closing braces after \x{... instead of
340 treating the string as literal. PCRE now does the same.
341
342 24. RunTest used to grumble if an inappropriate test was selected explicitly,
343 but just skip it when running all tests. This make it awkward to run ranges
344 of tests when one of them was inappropriate. Now it just skips any
345 inappropriate tests, as it always did when running all tests.
346
347 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
348 character types such as \d or \w, too many callouts were inserted, and the
349 data that they returned was rubbish.
350
351 26. In UCP mode, \s was not matching two of the characters that Perl matches,
352 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
353 were matched by \h. The code has now been refactored so that the lists of
354 the horizontal and vertical whitespace characters used for \h and \v (which
355 are defined only in one place) are now also used for \s.
356
357 27. Add JIT support for the 64 bit TileGX architecture.
358 Patch by Jiong Wang (Tilera Corporation).
359
360 28. Possessive quantifiers for classes (both explicit and automatically
361 generated) now use special opcodes instead of wrapping in ONCE brackets.
362
363 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
364 (because it's meaningless), this was not happening when PCRE_CASELESS was
365 set. Not wrong, but inefficient.
366
367 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
368 properties for \w, \d, etc) is present in a test regex. Otherwise if the
369 test contains no characters greater than 255, Perl doesn't realise it
370 should be using Unicode semantics.
371
372 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
373 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
374 does in Unicode mode.
375
376 32. Added the "forbid" facility to pcretest so that putting tests into the
377 wrong test files can sometimes be quickly detected.
378
379 33. There is now a limit (default 250) on the depth of nesting of parentheses.
380 This limit is imposed to control the amount of system stack used at compile
381 time. It can be changed at build time by --with-parens-nest-limit=xxx or
382 the equivalent in CMake.
383
384 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
385 errors. Perl warns for these when in warning mode, but PCRE has no facility
386 for giving warnings.
387
388 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
389 However, this was not working for (?!) because it is optimized to (*FAIL),
390 for which PCRE does not allow quantifiers. The optimization is now disabled
391 when a quantifier follows (?!). I can't see any use for this, but it makes
392 things uniform.
393
394 36. Perl no longer allows group names to start with digits, so I have made this
395 change also in PCRE. It simplifies the code a bit.
396
397 37. In extended mode, Perl ignores spaces before a + that indicates a
398 possessive quantifier. PCRE allowed a space before the quantifier, but not
399 before the possessive +. It now does.
400
401 38. The use of \K (reset reported match start) within a repeated possessive
402 group such as (a\Kb)*+ was not working.
403
404 40. Document that the same character tables must be used at compile time and
405 run time, and that the facility to pass tables to pcre_exec() and
406 pcre_dfa_exec() is for use only with saved/restored patterns.
407
408 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
409 features for Builds with MSVC:
410
411 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
412 builds). The .rc files can be used to set FileDescription and many other
413 attributes.
414
415 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
416 This allows higher-level build scripts which want .pdb files to avoid
417 hard-coding the exact files needed."
418
419 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
420 mean "start of word" and "end of word", respectively, as a transition aid.
421
422 43. A minimizing repeat of a class containing codepoints greater than 255 in
423 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
424 compiled to use the heap for recursion.
425
426 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
427 is configured.
428
429
430 Version 8.33 28-May-2013
431 ------------------------
432
433 1. Added 'U' to some constants that are compared to unsigned integers, to
434 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
435 variables that are added to signed variables, to ensure the result is
436 signed and can be negated.
437
438 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
439 CMake config files.
440
441 3. Revise the creation of config.h.generic so that all boolean macros are
442 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
443 overriding via -D on the command line possible.
444
445 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
446 to unsigned int is reported to make a quite noticeable speed difference in
447 a specific Windows environment. Testing on Linux did also appear to show
448 some benefit (and it is clearly not harmful). Also fixed the definition of
449 Xop which should be unsigned.
450
451 5. Related to (4), changing the definition of the intermediate variable cc
452 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
453 improvements.
454
455 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
456 unnecessary spaces.
457
458 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
459 and later.
460
461 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
462
463 9. Optimizing fast_forward_start_bits in JIT.
464
465 10. Adding support for callouts in JIT, and fixing some issues revealed
466 during this work. Namely:
467
468 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
469
470 (b) Minimum length was not checked before the matching is started.
471
472 11. The value of capture_last that is passed to callouts was incorrect in some
473 cases when there was a capture on one path that was subsequently abandoned
474 after a backtrack. Also, the capture_last value is now reset after a
475 recursion, since all captures are also reset in this case.
476
477 12. The interpreter no longer returns the "too many substrings" error in the
478 case when an overflowing capture is in a branch that is subsequently
479 abandoned after a backtrack.
480
481 13. In the pathological case when an offset vector of size 2 is used, pcretest
482 now prints out the matched string after a yield of 0 or 1.
483
484 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
485 Only supported by the JIT compiler at the moment.
486
487 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
488
489 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
490 offset of the starting point of the matching process, provided the offsets
491 vector is large enough.
492
493 17. The \A escape now records a lookbehind value of 1, though its execution
494 does not actually inspect the previous character. This is to ensure that,
495 in partial multi-segment matching, at least one character from the old
496 segment is retained when a new segment is processed. Otherwise, if there
497 are no lookbehinds in the pattern, \A might match incorrectly at the start
498 of a new segment.
499
500 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
501
502 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
503 gives some modest performance improvement in 8-bit mode.
504
505 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
506 be expressed in certain programming languages using Universal Character
507 Names.
508
509 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
510 which points out that "non characters" are not "characters that may not
511 appear in Unicode strings" but rather "characters that are reserved for
512 internal use and have only local meaning".
513
514 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
515 there was a conditional group that depended on an assertion, if the
516 assertion was false, the callout that immediately followed the alternation
517 in the condition was skipped when pcre_exec() was used for matching.
518
519 23. Allow an explicit callout to be inserted before an assertion that is the
520 condition for a conditional group, for compatibility with automatic
521 callouts, which always insert a callout at this point.
522
523 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
524 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
525
526 25. (*PRUNE) is now supported by the JIT compiler.
527
528 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
529
530 27. Fix the case where there are two or more SKIPs with arguments that may be
531 ignored.
532
533 28. (*SKIP) is now supported by the JIT compiler.
534
535 29. (*THEN) is now supported by the JIT compiler.
536
537 30. Update RunTest with additional test selector options.
538
539 31. The way PCRE handles backtracking verbs has been changed in two ways.
540
541 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
542 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
543 backtracking. In some cases this makes it more Perl-compatible, but Perl's
544 rather obscure rules do not always do the same thing.
545
546 (2) Previously, backtracking verbs were confined within assertions. This is
547 no longer the case for positive assertions, except for (*ACCEPT). Again,
548 this sometimes improves Perl compatibility, and sometimes does not.
549
550 32. A number of tests that were in test 2 because Perl did things differently
551 have been moved to test 1, because either Perl or PCRE has changed, and
552 these tests are now compatible.
553
554 32. Backtracking control verbs are now handled in the same way in JIT and
555 interpreter.
556
557 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
558 contained a forward subroutine reference caused a compile error.
559
560 34. Auto-detect and optimize limited repetitions in JIT.
561
562 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
563 blocking (*UTF) etc.
564
565 36. In the interpreter, maximizing pattern repetitions for characters and
566 character types now use tail recursion, which reduces stack usage.
567
568 37. The value of the max lookbehind was not correctly preserved if a compiled
569 and saved regex was reloaded on a host of different endianness.
570
571 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
572 of the compiled pattern block, expand the flags field from 16 to 32 bits
573 because it was almost full.
574
575 39. Try madvise first before posix_madvise.
576
577 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
578 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
579 original change locked out all matches of empty strings. This has been
580 changed so that one match of an empty string per line is recognized.
581 Subsequent searches on the same line (for colouring or for --only-matching,
582 for example) do not recognize empty strings.
583
584 41. Applied a user patch to fix a number of spelling mistakes in comments.
585
586 42. Data lines longer than 65536 caused pcretest to crash.
587
588 43. Clarified the data type for length and startoffset arguments for pcre_exec
589 and pcre_dfa_exec in the function-specific man pages, where they were
590 explicitly stated to be in bytes, never having been updated. I also added
591 some clarification to the pcreapi man page.
592
593 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
594 a segmentation fault.
595
596
597 Version 8.32 30-November-2012
598 -----------------------------
599
600 1. Improved JIT compiler optimizations for first character search and single
601 character iterators.
602
603 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
604 Patch by Daniel Richard G.
605
606 3. Single character iterator optimizations in the JIT compiler.
607
608 4. Improved JIT compiler optimizations for character ranges.
609
610 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
611 Reported by Giuseppe D'Angelo.
612
613 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
614 of a line, was being set incorrectly in cases where .* appeared inside
615 atomic brackets at the start of a pattern, or where there was a subsequent
616 *PRUNE or *SKIP.
617
618 7. Improved instruction cache flush for POWER/PowerPC.
619 Patch by Daniel Richard G.
620
621 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
622 grep:
623
624 (a) There is now no limit to the number of patterns to be matched.
625
626 (b) An error is given if a pattern is too long.
627
628 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
629 are now supported.
630
631 (d) --exclude-from and --include-from (multiple use) have been added.
632
633 (e) Exclusions and inclusions now apply to all files and directories, not
634 just to those obtained from scanning a directory recursively.
635
636 (f) Multiple uses of -f and --file-list are now supported.
637
638 (g) In a Windows environment, the default for -d has been changed from
639 "read" (the GNU grep default) to "skip", because otherwise the presence
640 of a directory in the file list provokes an error.
641
642 (h) The documentation has been revised and clarified in places.
643
644 9. Improve the matching speed of capturing brackets.
645
646 10. Changed the meaning of \X so that it now matches a Unicode extended
647 grapheme cluster.
648
649 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
650 out POSIX threads when JIT support is configured.
651
652 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
653
654 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
655 structure could go wrong in environments where size_t is not the same size
656 as int.
657
658 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
659 set.
660
661 15. The EBCDIC support had decayed; later updates to the code had included
662 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
663 general tidy up of EBCDIC-related issues, and the documentation was also
664 not quite right. There is now a test that can be run on ASCII systems to
665 check some of the EBCDIC-related things (but is it not a full test).
666
667 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
668 in a small tidy to the code.
669
670 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
671
672 18. If the --only-matching (-o) option in pcregrep is specified multiple
673 times, each one causes appropriate output. For example, -o1 -o2 outputs the
674 substrings matched by the 1st and 2nd capturing parentheses. A separating
675 string can be specified by --om-separator (default empty).
676
677 19. Improving the first n character searches.
678
679 20. Turn case lists for horizontal and vertical white space into macros so that
680 they are defined only once.
681
682 21. This set of changes together give more compatible Unicode case-folding
683 behaviour for characters that have more than one other case when UCP
684 support is available.
685
686 (a) The Unicode property table now has offsets into a new table of sets of
687 three or more characters that are case-equivalent. The MultiStage2.py
688 script that generates these tables (the pcre_ucd.c file) now scans
689 CaseFolding.txt instead of UnicodeData.txt for character case
690 information.
691
692 (b) The code for adding characters or ranges of characters to a character
693 class has been abstracted into a generalized function that also handles
694 case-independence. In UTF-mode with UCP support, this uses the new data
695 to handle characters with more than one other case.
696
697 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
698 whose other case is greater than 256 are now correctly matched
699 caselessly. Previously, the high codepoint matched the low one, but not
700 vice versa.
701
702 (d) The processing of \h, \H, \v, and \ in character classes now makes use
703 of the new class addition function, using character lists defined as
704 macros alongside the case definitions of 20 above.
705
706 (e) Caseless back references now work with characters that have more than
707 one other case.
708
709 (f) General caseless matching of characters with more than one other case
710 is supported.
711
712 22. Unicode character properties were updated from Unicode 6.2.0
713
714 23. Improved CMake support under Windows. Patch by Daniel Richard G.
715
716 24. Add support for 32-bit character strings, and UTF-32
717
718 25. Major JIT compiler update (code refactoring and bugfixing).
719 Experimental Sparc 32 support is added.
720
721 26. Applied a modified version of Daniel Richard G's patch to create
722 pcre.h.generic and config.h.generic by "make" instead of in the
723 PrepareRelease script.
724
725 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
726 pcre_compile.c when checking for a zero character.
727
728 28. Introducing a native interface for JIT. Through this interface, the compiled
729 machine code can be directly executed. The purpose of this interface is to
730 provide fast pattern matching, so several sanity checks are not performed.
731 However, feature tests are still performed. The new interface provides
732 1.4x speedup compared to the old one.
733
734 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
735 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
736 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
737
738 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
739 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
740 the "old" RFC 2279). Instead, it ended up passing a negative length to
741 pcre_exec().
742
743 31. Add support for GCC's visibility feature to hide internal functions.
744
745 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
746 "unknown -C option" after outputting 0 or 1.
747
748 33. There is now support for generating a code coverage report for the test
749 suite in environments where gcc is the compiler and lcov is installed. This
750 is mainly for the benefit of the developers.
751
752 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
753 unaddressable using valgrind annotations, allowing valgrind to detect
754 invalid memory accesses. This is mainly for the benefit of the developers.
755
756 25. (*UTF) can now be used to start a pattern in any of the three libraries.
757
758 26. Give configure error if --enable-cpp but no C++ compiler found.
759
760
761 Version 8.31 06-July-2012
762 -------------------------
763
764 1. Fixing a wrong JIT test case and some compiler warnings.
765
766 2. Removed a bashism from the RunTest script.
767
768 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
769 to unsigned type, result still unsigned" that was given by an MS compiler
770 on encountering the code "-sizeof(xxx)".
771
772 4. Partial matching support is added to the JIT compiler.
773
774 5. Fixed several bugs concerned with partial matching of items that consist
775 of more than one character:
776
777 (a) /^(..)\1/ did not partially match "aba" because checking references was
778 done on an "all or nothing" basis. This also applied to repeated
779 references.
780
781 (b) \R did not give a hard partial match if \r was found at the end of the
782 subject.
783
784 (c) \X did not give a hard partial match after matching one or more
785 characters at the end of the subject.
786
787 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
788 a partial match for the string "\r".
789
790 (e) When newline was set to CRLF, the metacharacter "." did not recognize
791 a partial match for a CR character at the end of the subject string.
792
793 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
794 running pcretest, the text "(JIT)" added to the output whenever JIT is
795 actually used to run the match.
796
797 7. Individual JIT compile options can be set in pcretest by following -s+[+]
798 or /S+[+] with a digit between 1 and 7.
799
800 8. OP_NOT now supports any UTF character not just single-byte ones.
801
802 9. (*MARK) control verb is now supported by the JIT compiler.
803
804 10. The command "./RunTest list" lists the available tests without actually
805 running any of them. (Because I keep forgetting what they all are.)
806
807 11. Add PCRE_INFO_MAXLOOKBEHIND.
808
809 12. Applied a (slightly modified) user-supplied patch that improves performance
810 when the heap is used for recursion (compiled with --disable-stack-for-
811 recursion). Instead of malloc and free for each heap frame each time a
812 logical recursion happens, frames are retained on a chain and re-used where
813 possible. This sometimes gives as much as 30% improvement.
814
815 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
816 call.
817
818 14. As documented, (*COMMIT) is now confined to within a positive assertion.
819
820 15. It is now possible to link pcretest with libedit as an alternative to
821 libreadline.
822
823 16. (*COMMIT) control verb is now supported by the JIT compiler.
824
825 17. The Unicode data tables have been updated to Unicode 6.1.0.
826
827 18. Added --file-list option to pcregrep.
828
829 19. Added binary file support to pcregrep, including the -a, --binary-files,
830 -I, and --text options.
831
832 20. The madvise function is renamed for posix_madvise for QNX compatibility
833 reasons. Fixed by Giuseppe D'Angelo.
834
835 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
836
837 22. Changed the option for creating symbolic links for 16-bit man pages from
838 -s to -sf so that re-installing does not cause issues.
839
840 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
841
842 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
843 matches in certain environments (the workspace was not being correctly
844 retained). Also added to pcre_dfa_exec() a simple plausibility check on
845 some of the workspace data at the beginning of a restart.
846
847 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
848 was not doing so when it should - probably a typo introduced by SVN 528
849 (change 8.10/14).
850
851 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
852 \w+ when the character tables indicated that \x{c4} was a word character.
853 There were several related cases, all because the tests for doing a table
854 lookup were testing for characters less than 127 instead of 255.
855
856 27. If a pattern contains capturing parentheses that are not used in a match,
857 their slots in the ovector are set to -1. For those that are higher than
858 any matched groups, this happens at the end of processing. In the case when
859 there were back references that the ovector was too small to contain
860 (causing temporary malloc'd memory to be used during matching), and the
861 highest capturing number was not used, memory off the end of the ovector
862 was incorrectly being set to -1. (It was using the size of the temporary
863 memory instead of the true size.)
864
865 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
866 ovector size, it uses memory at the end of the block that it has got.
867
868 29. Check for an overlong MARK name and give an error at compile time. The
869 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
870
871 30. JIT compiler update.
872
873 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
874 Rill for the patch.
875
876 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
877
878 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
879
880 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
881 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
882 was enabled.
883
884 35. Improve JIT code generation for greedy plus quantifier.
885
886 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
887 1 to "aa" instead of to an empty string. The bug affected repeated groups
888 that could potentially match an empty string.
889
890 37. Optimizing single character iterators in JIT.
891
892 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
893 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
894 codepoints that are too big for the mode are faulted, and in a UTF mode,
895 disallowed codepoints are also faulted.
896
897 39. If PCRE was compiled with UTF support, in three places in the DFA
898 matcher there was code that should only have been obeyed in UTF mode, but
899 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
900 processing when bytes with values greater than 127 were present. In 16-bit
901 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
902 both cases the values are those that cannot be the first data item in a UTF
903 character. The three items that might have provoked this were recursions,
904 possessively repeated groups, and atomic groups.
905
906 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
907 and pcregrep, because some OS require shared objects to be explicitly
908 passed to ld, causing the link step to fail if they are not.
909
910 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
911 mode, patterns that started with \h* or \R* might be incorrectly matched.
912
913
914 Version 8.30 04-February-2012
915 -----------------------------
916
917 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
918 name is defined in ctype.h.
919
920 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
921 only in quite long subpatterns.
922
923 3. Removed the function pcre_info(), which has been obsolete and deprecated
924 since it was replaced by pcre_fullinfo() in February 2000.
925
926 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
927 match a (*MARK), and the match failed at the start of the subject, a
928 reference to memory before the start of the subject could occur. This bug
929 was introduced by fix 17 of release 8.21.
930
931 5. A reference to an unset group with zero minimum repetition was giving
932 totally wrong answers (in non-JavaScript-compatibility mode). For example,
933 /(another)?(\1?)test/ matched against "hello world test". This bug was
934 introduced in release 8.13.
935
936 6. Add support for 16-bit character strings (a large amount of work involving
937 many changes and refactorings).
938
939 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
940 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
941 from a file.
942
943 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
944 rounding is not applied in this particular case).
945
946 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
947 if they appear, or are escaped, in patterns.
948
949 10. Get rid of a number of -Wunused-but-set-variable warnings.
950
951 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
952 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
953 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
954 also returns the mark "x". This bug applied to capturing parentheses,
955 non-capturing parentheses, and atomic parentheses. It also applied to some
956 assertions.
957
958 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
959 information out of configure.ac instead of relying on pcre.h.generic, which
960 is not stored in the repository.
961
962 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
963 -lreadline.
964
965 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
966
967 15. Applied Graycode's patch to put the top-level frame on the stack rather
968 than the heap when not using the stack for recursion. This gives a
969 performance improvement in many cases when recursion is not deep.
970
971 16. Experimental code added to "pcretest -C" to output the stack frame size.
972
973
974 Version 8.21 12-Dec-2011
975 ------------------------
976
977 1. Updating the JIT compiler.
978
979 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
980 are added as well.
981
982 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
983 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
984 calling _pcre_jit_exec. Some extra comments are added.
985
986 4. (*MARK) settings inside atomic groups that do not contain any capturing
987 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
988 was introduced by change 18 for 8.20.
989
990 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
991 ECMA-262 standard.
992
993 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
994 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
995 This bug was probably introduced by change 9 of 8.13.
996
997 7. While fixing 6 above, I noticed that a number of other items were being
998 incorrectly rejected as "not fixed length". This arose partly because newer
999 opcodes had not been added to the fixed-length checking code. I have (a)
1000 corrected the bug and added tests for these items, and (b) arranged for an
1001 error to occur if an unknown opcode is encountered while checking for fixed
1002 length instead of just assuming "not fixed length". The items that were
1003 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1004 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1005 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1006
1007 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1008 being incorrectly compiled and would have given unpredicatble results.
1009
1010 9. A possessively repeated subpattern with minimum repeat count greater than
1011 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1012 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1013 the first (A) could occur when it should not.
1014
1015 10. Add a cast and remove a redundant test from the code.
1016
1017 11. JIT should use pcre_malloc/pcre_free for allocation.
1018
1019 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1020 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1021 is anything other than /usr, -L is still shown).
1022
1023 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1024
1025 14. Perl does not support \N without a following name in a [] class; PCRE now
1026 also gives an error.
1027
1028 15. If a forward reference was repeated with an upper limit of around 2000,
1029 it caused the error "internal error: overran compiling workspace". The
1030 maximum number of forward references (including repeats) was limited by the
1031 internal workspace, and dependent on the LINK_SIZE. The code has been
1032 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1033 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1034 of around 200,000 forward references. While doing this, I also speeded up
1035 the filling in of repeated forward references.
1036
1037 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1038 incorrectly expecting the subject to contain another "a" after the start.
1039
1040 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1041 in the match, the SKIP should be ignored. This was not happening; instead
1042 the SKIP was being treated as NOMATCH. For patterns such as
1043 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1044 tested.
1045
1046 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1047 now much more compatible with Perl, in particular in cases where the result
1048 is a non-match for a non-anchored pattern. For example, if
1049 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1050 "m", where previously it did not return a name. A side effect of this
1051 change is that for partial matches, the last encountered mark name is
1052 returned, as for non matches. A number of tests that were previously not
1053 Perl-compatible have been moved into the Perl-compatible test files. The
1054 refactoring has had the pleasing side effect of removing one argument from
1055 the match() function, thus reducing its stack requirements.
1056
1057 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1058 subsequent uses of /S (without +) incorrectly behaved like /S+.
1059
1060 21. Retrieve executable code size support for the JIT compiler and fixing
1061 some warnings.
1062
1063 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1064 not work when the shorter character appeared right at the end of the
1065 subject string.
1066
1067 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1068 systems.
1069
1070 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1071 output it when the /M option is used in pcretest.
1072
1073 25. The CheckMan script was not being included in the distribution. Also, added
1074 an explicit "perl" to run Perl scripts from the PrepareRelease script
1075 because this is reportedly needed in Windows.
1076
1077 26. If study data was being save in a file and studying had not found a set of
1078 "starts with" bytes for the pattern, the data written to the file (though
1079 never used) was taken from uninitialized memory and so caused valgrind to
1080 complain.
1081
1082 27. Updated RunTest.bat as provided by Sheri Pierce.
1083
1084 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1085
1086 29. Computation of memory usage for the table of capturing group names was
1087 giving an unnecessarily large value.
1088
1089
1090 Version 8.20 21-Oct-2011
1091 ------------------------
1092
1093 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1094 a POSIX class. After further experiments with Perl, which convinced me that
1095 Perl has bugs and confusions, a closing square bracket is no longer allowed
1096 in a POSIX name. This bug also affected patterns with classes that started
1097 with full stops.
1098
1099 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1100 captured substring, but while checking the failing first alternative,
1101 substring 1 is temporarily captured. If the output vector supplied to
1102 pcre_exec() was not big enough for this capture, the yield of the function
1103 was still zero ("insufficient space for captured substrings"). This cannot
1104 be totally fixed without adding another stack variable, which seems a lot
1105 of expense for a edge case. However, I have improved the situation in cases
1106 such as /(a)(b)x|abc/ matched against "abc", where the return code
1107 indicates that fewer than the maximum number of slots in the ovector have
1108 been set.
1109
1110 3. Related to (2) above: when there are more back references in a pattern than
1111 slots in the output vector, pcre_exec() uses temporary memory during
1112 matching, and copies in the captures as far as possible afterwards. It was
1113 using the entire output vector, but this conflicts with the specification
1114 that only 2/3 is used for passing back captured substrings. Now it uses
1115 only the first 2/3, for compatibility. This is, of course, another edge
1116 case.
1117
1118 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1119 main code base, and can be used by building with --enable-jit. When this is
1120 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1121 runtime --no-jit option is given.
1122
1123 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1124 ovector, the return from the function was zero, implying that there were
1125 other matches that did not fit. The correct "exactly full" value is now
1126 returned.
1127
1128 6. If a subpattern that was called recursively or as a subroutine contained
1129 (*PRUNE) or any other control that caused it to give a non-standard return,
1130 invalid errors such as "Error -26 (nested recursion at the same subject
1131 position)" or even infinite loops could occur.
1132
1133 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1134 computing the minimum length on reaching *ACCEPT, and so ended up with the
1135 wrong value of 1 rather than 0. Further investigation indicates that
1136 computing a minimum subject length in the presence of *ACCEPT is difficult
1137 (think back references, subroutine calls), and so I have changed the code
1138 so that no minimum is registered for a pattern that contains *ACCEPT.
1139
1140 8. If (*THEN) was present in the first (true) branch of a conditional group,
1141 it was not handled as intended. [But see 16 below.]
1142
1143 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1144 Sheri Pierce.
1145
1146 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1147 the first byte in a match must be "a".
1148
1149 11. Change 17 for 8.13 increased the recursion depth for patterns like
1150 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1151 pattern contains any instances of (*THEN). If it does not, the old
1152 optimizations are restored. It would be nice to do this on a per-group
1153 basis, but at the moment that is not feasible.
1154
1155 12. In some environments, the output of pcretest -C is CRLF terminated. This
1156 broke RunTest's code that checks for the link size. A single white space
1157 character after the value is now allowed for.
1158
1159 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1160 For "fr", it uses the Windows-specific input and output files.
1161
1162 14. If (*THEN) appeared in a group that was called recursively or as a
1163 subroutine, it did not work as intended. [But see next item.]
1164
1165 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1166 pattern fragments (but not containing any | characters). If A and B are
1167 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1168 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1169 D. In other words, Perl considers parentheses that do not contain any |
1170 characters to be part of a surrounding alternative, whereas PCRE was
1171 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1172 differently. PCRE now behaves in the same way as Perl, except in the case
1173 of subroutine/recursion calls such as (?1) which have in any case always
1174 been different (but PCRE had them first :-).
1175
1176 16. Related to 15 above: Perl does not treat the | in a conditional group as
1177 creating alternatives. Such a group is treated in the same way as an
1178 ordinary group without any | characters when processing (*THEN). PCRE has
1179 been changed to match Perl's behaviour.
1180
1181 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1182 RunGrepTest script failed.
1183
1184 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1185 inevitable for groups that contain captures, but it can lead to a lot of
1186 stack use in large patterns. The old behaviour has been restored for atomic
1187 groups that do not contain any capturing parentheses.
1188
1189 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1190 suppress the check for a minimum subject length at run time. (If it was
1191 given to pcre_exec() or pcre_dfa_exec() it did work.)
1192
1193 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1194 fail to work when decoding hex characters in data strings in EBCDIC
1195 environments.
1196
1197 21. It appears that in at least one Mac OS environment, the isxdigit() function
1198 is implemented as a macro that evaluates to its argument more than once,
1199 contravening the C 90 Standard (I haven't checked a later standard). There
1200 was an instance in pcretest which caused it to go wrong when processing
1201 \x{...} escapes in subject strings. The has been rewritten to avoid using
1202 things like p++ in the argument of isxdigit().
1203
1204
1205 Version 8.13 16-Aug-2011
1206 ------------------------
1207
1208 1. The Unicode data tables have been updated to Unicode 6.0.0.
1209
1210 2. Two minor typos in pcre_internal.h have been fixed.
1211
1212 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1213 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1214 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1215
1216 4. There were a number of related bugs in the code for matching backrefences
1217 caselessly in UTF-8 mode when codes for the characters concerned were
1218 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1219 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1220 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1221 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1222 2-byte code at the end of the subject (it thought there wasn't enough data
1223 left).
1224
1225 5. Comprehensive information about what went wrong is now returned by
1226 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1227 as the output vector has at least 2 elements. The offset of the start of
1228 the failing character and a reason code are placed in the vector.
1229
1230 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1231 now returned is for the first byte of the failing character, instead of the
1232 last byte inspected. This is an incompatible change, but I hope it is small
1233 enough not to be a problem. It makes the returned offset consistent with
1234 pcre_exec() and pcre_dfa_exec().
1235
1236 7. pcretest now gives a text phrase as well as the error number when
1237 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1238 failure, the offset and reason code are output.
1239
1240 8. When \R was used with a maximizing quantifier it failed to skip backwards
1241 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1242 back over a single character (\n). This seems wrong (because it treated the
1243 two characters as a single entity when going forwards), conflicts with the
1244 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1245 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1246 has been changed.
1247
1248 9. Some internal refactoring has changed the processing so that the handling
1249 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1250 time (the PCRE_DOTALL option was changed this way some time ago: version
1251 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1252 which was always a bit of a fudge. It also means that there is one less
1253 argument for the match() function, which reduces its stack requirements
1254 slightly. This change also fixes an incompatibility with Perl: the pattern
1255 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1256
1257 10. More internal refactoring has drastically reduced the number of recursive
1258 calls to match() for possessively repeated groups such as (abc)++ when
1259 using pcre_exec().
1260
1261 11. While implementing 10, a number of bugs in the handling of groups were
1262 discovered and fixed:
1263
1264 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1265 (a|)*(?1) gave a compile-time internal error.
1266 ((a|)+)+ did not notice that the outer group could match an empty string.
1267 (^a|^)+ was not marked as anchored.
1268 (.*a|.*)+ was not marked as matching at start or after a newline.
1269
1270 12. Yet more internal refactoring has removed another argument from the match()
1271 function. Special calls to this function are now indicated by setting a
1272 value in a variable in the "match data" data block.
1273
1274 13. Be more explicit in pcre_study() instead of relying on "default" for
1275 opcodes that mean there is no starting character; this means that when new
1276 ones are added and accidentally left out of pcre_study(), testing should
1277 pick them up.
1278
1279 14. The -s option of pcretest has been documented for ages as being an old
1280 synonym of -m (show memory usage). I have changed it to mean "force study
1281 for every regex", that is, assume /S for every regex. This is similar to -i
1282 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1283 using it. It makes it easier to run collections of tests with and without
1284 study enabled, and thereby test pcre_study() more easily. All the standard
1285 tests are now run with and without -s (but some patterns can be marked as
1286 "never study" - see 20 below).
1287
1288 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1289 restoration of the capturing data to the outer values was not happening
1290 correctly.
1291
1292 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1293 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1294 pattern had matched an empty string, and so incorrectly returned a no
1295 match.
1296
1297 17. There was optimizing code for the last branch of non-capturing parentheses,
1298 and also for the obeyed branch of a conditional subexpression, which used
1299 tail recursion to cut down on stack usage. Unfortunately, now that there is
1300 the possibility of (*THEN) occurring in these branches, tail recursion is
1301 no longer possible because the return has to be checked for (*THEN). These
1302 two optimizations have therefore been removed. [But see 8.20/11 above.]
1303
1304 18. If a pattern containing \R was studied, it was assumed that \R always
1305 matched two bytes, thus causing the minimum subject length to be
1306 incorrectly computed because \R can also match just one byte.
1307
1308 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1309 was incorrectly computed.
1310
1311 20. If /S is present twice on a test pattern in pcretest input, it now
1312 *disables* studying, thereby overriding the use of -s on the command line
1313 (see 14 above). This is necessary for one or two tests to keep the output
1314 identical in both cases.
1315
1316 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1317 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1318
1319 22. When an atomic group that contained a capturing parenthesis was
1320 successfully matched, but the branch in which it appeared failed, the
1321 capturing was not being forgotten if a higher numbered group was later
1322 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1323 group 1 to "a", when in fact it should be unset. This applied to multi-
1324 branched capturing and non-capturing groups, repeated or not, and also to
1325 positive assertions (capturing in negative assertions does not happen
1326 in PCRE) and also to nested atomic groups.
1327
1328 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1329 subject after a captured substring, to make it easier to tell which of a
1330 number of identical substrings has been captured.
1331
1332 24. The way atomic groups are processed by pcre_exec() has been changed so that
1333 if they are repeated, backtracking one repetition now resets captured
1334 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1335 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1336 "aaa". Previously, it would have been "a". As part of this code
1337 refactoring, the way recursive calls are handled has also been changed.
1338
1339 25. If an assertion condition captured any substrings, they were not passed
1340 back unless some other capturing happened later. For example, if
1341 (?(?=(a))a) was matched against "a", no capturing was returned.
1342
1343 26. When studying a pattern that contained subroutine calls or assertions,
1344 the code for finding the minimum length of a possible match was handling
1345 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1346 group 1 called group 2 while simultaneously a separate group 2 called group
1347 1). A stack overflow occurred in this case. I have fixed this by limiting
1348 the recursion depth to 10.
1349
1350 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1351 Fortmann. This supports explicit test numbers on the command line, and has
1352 argument validation and error reporting.
1353
1354 28. An instance of \X with an unlimited repeat could fail if at any point the
1355 first character it looked at was a mark character.
1356
1357 29. Some minor code refactoring concerning Unicode properties and scripts
1358 should reduce the stack requirement of match() slightly.
1359
1360 30. Added the '=' option to pcretest to check the setting of unused capturing
1361 slots at the end of the pattern, which are documented as being -1, but are
1362 not included in the return count.
1363
1364 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1365 compiled something random. Now it gives a compile-time error (as does
1366 Perl).
1367
1368 32. A *MARK encountered during the processing of a positive assertion is now
1369 recorded and passed back (compatible with Perl).
1370
1371 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1372 had alternative anchored branches, the search for a second match in a line
1373 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1374 matched the line "0102" twice. The same bug affected patterns that started
1375 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1376 twice.
1377
1378 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1379 does, and because of capturing effects, quantifying parenthesized
1380 assertions may at times be useful. Quantifiers are now allowed for
1381 parenthesized assertions.
1382
1383 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1384
1385 36. \g was being checked for fancy things in a character class, when it should
1386 just be a literal "g".
1387
1388 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1389 appearance of a nested POSIX class supersedes an apparent external class.
1390 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1391 unescaped square brackets may also appear as part of class names. For
1392 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1393 more like Perl. (But see 8.20/1 above.)
1394
1395 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1396 was because it thought it was \N{name}, which is not supported).
1397
1398 39. Add minix to OS list not supporting the -S option in pcretest.
1399
1400 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1401 cannot analyze patterns in sufficient detail to catch mutual recursions
1402 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1403 subgroup is called recursively as a subpattern for a second time at the
1404 same position in the subject string. In previous releases this might have
1405 been caught by the recursion limit, or it might have run out of stack.
1406
1407 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1408 happen only once. PCRE was, however incorrectly giving a compile time error
1409 "recursive call could loop indefinitely" because it cannot analyze the
1410 pattern in sufficient detail. The compile time test no longer happens when
1411 PCRE is compiling a conditional subpattern, but actual runaway loops are
1412 now caught at runtime (see 40 above).
1413
1414 42. It seems that Perl allows any characters other than a closing parenthesis
1415 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1416 has been changed to be the same.
1417
1418 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1419 as not to get warnings when autogen.sh is called. Also changed
1420 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1421
1422 44. To help people who use pcregrep to scan files containing exceedingly long
1423 lines, the following changes have been made:
1424
1425 (a) The default value of the buffer size parameter has been increased from
1426 8K to 20K. (The actual buffer used is three times this size.)
1427
1428 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1429 PCRE is built.
1430
1431 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1432 to be set at run time.
1433
1434 (d) Numerical values in pcregrep options can be followed by K or M, for
1435 example --buffer-size=50K.
1436
1437 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1438 given and the return code is set to 2.
1439
1440 45. Add a pointer to the latest mark to the callout data block.
1441
1442 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1443 partial match of an empty string instead of no match. This was specific to
1444 the use of ".".
1445
1446 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1447 complete match instead of a partial match. This bug was dependent on both
1448 the PCRE_UTF8 and PCRE_DOTALL options being set.
1449
1450 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1451 starting byte set, because \b was not being ignored.
1452
1453
1454 Version 8.12 15-Jan-2011
1455 ------------------------
1456
1457 1. Fixed some typos in the markup of the man pages, and wrote a script that
1458 checks for such things as part of the documentation building process.
1459
1460 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1461 --match-limit and --recursion-limit options (added for 8.11). In
1462 particular, this made one of the standard tests fail. (The integer value
1463 went into the wrong half of a long int.)
1464
1465 3. If the --colour option was given to pcregrep with -v (invert match), it
1466 did strange things, either producing crazy output, or crashing. It should,
1467 of course, ignore a request for colour when reporting lines that do not
1468 match.
1469
1470 4. Another pcregrep bug caused similar problems if --colour was specified with
1471 -M (multiline) and the pattern match finished with a line ending.
1472
1473 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1474 matched in multiline mode, the following line was shown as part of the
1475 match. This seems wrong, so I have changed it.
1476
1477 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1478 the check for further matches in the same line (so they could be coloured)
1479 to overrun the end of the current line. If another match was found, it was
1480 incorrectly shown (and then shown again when found in the next line).
1481
1482 7. If pcregrep was compiled under Windows, there was a reference to the
1483 function pcregrep_exit() before it was defined. I am assuming this was
1484 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1485 reported by a user. I've moved the definition above the reference.
1486
1487
1488 Version 8.11 10-Dec-2010
1489 ------------------------
1490
1491 1. (*THEN) was not working properly if there were untried alternatives prior
1492 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1493 backtracked to try for "b" instead of moving to the next alternative branch
1494 at the same level (in this case, to look for "c"). The Perl documentation
1495 is clear that when (*THEN) is backtracked onto, it goes to the "next
1496 alternative in the innermost enclosing group".
1497
1498 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1499 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1500 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1501 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1502 (*THEN).
1503
1504 3. If \s appeared in a character class, it removed the VT character from
1505 the class, even if it had been included by some previous item, for example
1506 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1507 of \s, but is part of the POSIX "space" class.)
1508
1509 4. A partial match never returns an empty string (because you can always
1510 match an empty string at the end of the subject); however the checking for
1511 an empty string was starting at the "start of match" point. This has been
1512 changed to the "earliest inspected character" point, because the returned
1513 data for a partial match starts at this character. This means that, for
1514 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1515 (previously it gave "no match").
1516
1517 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1518 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1519 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1520 has an implication that the given string is incomplete (because a partial
1521 match is preferred over a full match). For this reason, these items now
1522 give a partial match in this situation. [Aside: previously, the one case
1523 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1524 match rather than a full match, which was wrong by the old rules, but is
1525 now correct.]
1526
1527 6. There was a bug in the handling of #-introduced comments, recognized when
1528 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1529 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1530 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1531 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1532 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1533 places in pcre_compile().
1534
1535 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1536 comments when looking ahead for named forward references to subpatterns,
1537 the only newline sequence it recognized was NL. It now handles newlines
1538 according to the set newline convention.
1539
1540 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1541 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1542 cater for a lack of strerror(). These oversights have been fixed.
1543
1544 9. Added --match-limit and --recursion-limit to pcregrep.
1545
1546 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1547
1548 11. When the -o option was used, pcregrep was setting a return code of 1, even
1549 when matches were found, and --line-buffered was not being honoured.
1550
1551 12. Added an optional parentheses number to the -o and --only-matching options
1552 of pcregrep.
1553
1554 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1555 can match an empty string. The code to do it in pcretest and pcredemo
1556 needed fixing:
1557
1558 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1559 only one byte after an empty string match just before CRLF (this case
1560 just got forgotten; "any" and "anycrlf" were OK).
1561
1562 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1563 mode when an empty string match preceded an ASCII character followed by
1564 a non-ASCII character. (The code for advancing by one character rather
1565 than one byte was nonsense.)
1566
1567 (c) The pcredemo.c sample program did not have any code at all to handle
1568 the cases when CRLF is a valid newline sequence.
1569
1570 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1571 as a starting offset was within the subject string. There is now a new
1572 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1573 negative or greater than the length of the string. In order to test this,
1574 pcretest is extended to allow the setting of negative starting offsets.
1575
1576 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1577 starting offset points to the beginning of a UTF-8 character was
1578 unnecessarily clumsy. I tidied it up.
1579
1580 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1581 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1582
1583 17. Nobody had reported that the --include_dir option, which was added in
1584 release 7.7 should have been called --include-dir (hyphen, not underscore)
1585 for compatibility with GNU grep. I have changed it to --include-dir, but
1586 left --include_dir as an undocumented synonym, and the same for
1587 --exclude-dir, though that is not available in GNU grep, at least as of
1588 release 2.5.4.
1589
1590 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1591 characters from a string of bytes) have been redefined so as not to use
1592 loops, in order to improve performance in some environments. At the same
1593 time, I abstracted some of the common code into auxiliary macros to save
1594 repetition (this should not affect the compiled code).
1595
1596 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1597 compile-time error is now given if \c is not followed by an ASCII
1598 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1599 different, and any byte value is allowed.)
1600
1601 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1602 START_OPTIMIZE option, which is now allowed at compile time - but just
1603 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1604 to pcregrep and other applications that have no direct access to PCRE
1605 options. The new /Y option in pcretest sets this option when calling
1606 pcre_compile().
1607
1608 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1609 back references. Groups containing recursive back references were forced to
1610 be atomic by that change, but in the case of named groups, the amount of
1611 memory required was incorrectly computed, leading to "Failed: internal
1612 error: code overflow". This has been fixed.
1613
1614 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1615 pcretest.c, to avoid build problems in some Borland environments.
1616
1617
1618 Version 8.10 25-Jun-2010
1619 ------------------------
1620
1621 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1622 THEN.
1623
1624 2. (*ACCEPT) was not working when inside an atomic group.
1625
1626 3. Inside a character class, \B is treated as a literal by default, but
1627 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1628 causes the error). The code is unchanged, but I tidied the documentation.
1629
1630 4. Inside a character class, PCRE always treated \R and \X as literals,
1631 whereas Perl faults them if its -w option is set. I have changed PCRE so
1632 that it faults them when PCRE_EXTRA is set.
1633
1634 5. Added support for \N, which always matches any character other than
1635 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1636
1637 6. When compiling pcregrep with newer versions of gcc which may have
1638 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1639 declared with attribute warn_unused_result" were given. Just casting the
1640 result to (void) does not stop the warnings; a more elaborate fudge is
1641 needed. I've used a macro to implement this.
1642
1643 7. Minor change to pcretest.c to avoid a compiler warning.
1644
1645 8. Added four artifical Unicode properties to help with an option to make
1646 \s etc use properties (see next item). The new properties are: Xan
1647 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1648
1649 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1650 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1651 this option. Modified pcretest to add /W to test this facility. Added
1652 REG_UCP to make it available via the POSIX interface.
1653
1654 10. Added --line-buffered to pcregrep.
1655
1656 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1657 studied, and the match started with a letter with a code point greater than
1658 127 whose first byte was different to the first byte of the other case of
1659 the letter, the other case of this starting letter was not recognized
1660 (#976).
1661
1662 12. If a pattern that was studied started with a repeated Unicode property
1663 test, for example, \p{Nd}+, there was the theoretical possibility of
1664 setting up an incorrect bitmap of starting bytes, but fortunately it could
1665 not have actually happened in practice until change 8 above was made (it
1666 added property types that matched character-matching opcodes).
1667
1668 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1669 possible starting bytes for non-anchored patterns.
1670
1671 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1672 \R, and also a number of cases that involve Unicode properties, both
1673 explicit and implicit when PCRE_UCP is set.
1674
1675 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1676 input, it could crash or give wrong results if characters with values
1677 greater than 0xc0 were present in the subject string. (Detail: it assumed
1678 UTF-8 input when processing these items.)
1679
1680 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1681 size_t is 64-bit (#991).
1682
1683 17. Added a check for running out of memory when PCRE is compiled with
1684 --disable-stack-for-recursion (#990).
1685
1686 18. If the last data line in a file for pcretest does not have a newline on
1687 the end, a newline was missing in the output.
1688
1689 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1690 less than 128) in its various bitmaps. However, there is a facility for
1691 generating tables according to the current locale when PCRE is compiled. It
1692 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1693 characters, are recognized by isspace() and therefore were getting set in
1694 these tables, and indeed these tables seem to approximate to ISO 8859. This
1695 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1696 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1697 which of course cannot start UTF-8 characters. I have changed the code so
1698 that only real ASCII characters (less than 128) and the correct starting
1699 bytes for UTF-8 encodings are set for characters greater than 127 when in
1700 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1701 altogether.)
1702
1703 20. Added the /T option to pcretest so as to be able to run tests with non-
1704 standard character tables, thus making it possible to include the tests
1705 used for 19 above in the standard set of tests.
1706
1707 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1708 reference to a subpattern the other side of a comment that contains an
1709 opening parenthesis caused either an internal compiling error, or a
1710 reference to the wrong subpattern.
1711
1712
1713 Version 8.02 19-Mar-2010
1714 ------------------------
1715
1716 1. The Unicode data tables have been updated to Unicode 5.2.0.
1717
1718 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1719 configured.
1720
1721 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1722 original author of that file, following a query about its status.
1723
1724 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1725 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1726
1727 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1728 quantifier applied to a forward-referencing subroutine call, could compile
1729 incorrect code or give the error "internal error: previously-checked
1730 referenced subpattern not found".
1731
1732 6. Both MS Visual Studio and Symbian OS have problems with initializing
1733 variables to point to external functions. For these systems, therefore,
1734 pcre_malloc etc. are now initialized to local functions that call the
1735 relevant global functions.
1736
1737 7. There were two entries missing in the vectors called coptable and poptable
1738 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1739 I've fixed the data, and added a kludgy way of testing at compile time that
1740 the lengths are correct (equal to the number of opcodes).
1741
1742 8. Following on from 7, I added a similar kludge to check the length of the
1743 eint vector in pcreposix.c.
1744
1745 9. Error texts for pcre_compile() are held as one long string to avoid too
1746 much relocation at load time. To find a text, the string is searched,
1747 counting zeros. There was no check for running off the end of the string,
1748 which could happen if a new error number was added without updating the
1749 string.
1750
1751 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1752
1753 11. \K was not working if it appeared in an atomic group or in a group that
1754 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1755 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1756 the assertion is positive, but not if it is negative.
1757
1758 12. Change 11 fortuitously reduced the size of the stack frame used in the
1759 "match()" function of pcre_exec.c by one pointer. Forthcoming
1760 implementation of support for (*MARK) will need an extra pointer on the
1761 stack; I have reserved it now, so that the stack frame size does not
1762 decrease.
1763
1764 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1765 item in branch that calls a recursion is a subroutine call - as in the
1766 second branch in the above example - was incorrectly given the compile-
1767 time error "recursive call could loop indefinitely" because pcre_compile()
1768 was not correctly checking the subroutine for matching a non-empty string.
1769
1770 14. The checks for overrunning compiling workspace could trigger after an
1771 overrun had occurred. This is a "should never occur" error, but it can be
1772 triggered by pathological patterns such as hundreds of nested parentheses.
1773 The checks now trigger 100 bytes before the end of the workspace.
1774
1775 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1776
1777
1778 Version 8.01 19-Jan-2010
1779 ------------------------
1780
1781 1. If a pattern contained a conditional subpattern with only one branch (in
1782 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1783 computed the wrong minimum data length (which is of course zero for such
1784 subpatterns). This could cause incorrect "no match" results.
1785
1786 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1787 the pattern is reset in the first branch, pcre_compile() failed with
1788 "internal error: code overflow at offset...". This happened only when
1789 the reset was to the original external option setting. (An optimization
1790 abstracts leading options settings into an external setting, which was the
1791 cause of this.)
1792
1793 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1794 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1795 assertion pattern did not match (meaning that the assertion was true), it
1796 was incorrectly treated as false if the SKIP had been reached during the
1797 matching. This also applied to assertions used as conditions.
1798
1799 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1800 assertion subpattern, including such a pattern used as a condition,
1801 unpredictable results occurred, instead of the error return
1802 PCRE_ERROR_DFA_UITEM.
1803
1804 5. The C++ GlobalReplace function was not working like Perl for the special
1805 situation when an empty string is matched. It now does the fancy magic
1806 stuff that is necessary.
1807
1808 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1809 removed. (These were left over from very, very early versions of PCRE.)
1810
1811 7. Some cosmetic changes to the code to make life easier when compiling it
1812 as part of something else:
1813
1814 (a) Change DEBUG to PCRE_DEBUG.
1815
1816 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1817 called "current" as "current_branch", to prevent a collision with the
1818 Linux macro when compiled as a kernel module.
1819
1820 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1821 prevent a collision with the Linux macro when compiled as a kernel
1822 module.
1823
1824 8. In pcre_compile() there are some checks for integer overflows that used to
1825 cast potentially large values to (double). This has been changed to that
1826 when building, a check for int64_t is made, and if it is found, it is used
1827 instead, thus avoiding the use of floating point arithmetic. (There is no
1828 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1829 double.
1830
1831 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1832 2005 (difference between two addresses compared to an unsigned value).
1833
1834 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1835 custom one, because of the following reported problem in Windows:
1836
1837 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1838 under Win32.
1839 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1840 therefore missing the function definition.
1841 - The compiler thus generates a "C" signature for the test function.
1842 - The linker fails to find the "C" function.
1843 - PCRE fails to configure if asked to do so against libbz2.
1844
1845 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1846 messages were output:
1847
1848 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1849 rerunning libtoolize, to keep the correct libtool macros in-tree.
1850 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1851
1852 I have done both of these things.
1853
1854 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1855 most of the time, it *can* run out if it is given a pattern that contains a
1856 runaway infinite recursion. I updated the discussion in the pcrestack man
1857 page.
1858
1859 13. Now that we have gone to the x.xx style of version numbers, the minor
1860 version may start with zero. Using 08 or 09 is a bad idea because users
1861 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1862 interpreted as invalid octal numbers. I've updated the previous comment in
1863 configure.ac, and also added a check that gives an error if 08 or 09 are
1864 used.
1865
1866 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1867 causing partial matching to fail when the end of the subject matched \W
1868 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1869
1870 15. There were some discrepancies between the declarations in pcre_internal.h
1871 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1872 their definitions. The declarations used "const uschar *" and the
1873 definitions used USPTR. Even though USPTR is normally defined as "const
1874 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1875 reported that: "This difference in casting confuses some C++ compilers, for
1876 example, SunCC recognizes above declarations as different functions and
1877 generates broken code for hbpcre." I have changed the declarations to use
1878 USPTR.
1879
1880 16. GNU libtool is named differently on some systems. The autogen.sh script now
1881 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1882 (FreeBSD).
1883
1884 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1885 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1886 comment: "Figure out how to create a longlong from a string: strtoll and
1887 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1888 instance, but it only takes 2 args instead of 3!"
1889
1890 18. A subtle bug concerned with back references has been fixed by a change of
1891 specification, with a corresponding code fix. A pattern such as
1892 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1893 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1894 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1895 same bug. Such groups have to be quantified to be useful, or contained
1896 inside another quantified group. (If there's no repetition, the reference
1897 can never match.) The problem arises because, having left the group and
1898 moved on to the rest of the pattern, a later failure that backtracks into
1899 the group uses the captured value from the final iteration of the group
1900 rather than the correct earlier one. I have fixed this in PCRE by forcing
1901 any group that contains a reference to itself to be an atomic group; that
1902 is, there cannot be any backtracking into it once it has completed. This is
1903 similar to recursive and subroutine calls.
1904
1905
1906 Version 8.00 19-Oct-09
1907 ----------------------
1908
1909 1. The table for translating pcre_compile() error codes into POSIX error codes
1910 was out-of-date, and there was no check on the pcre_compile() error code
1911 being within the table. This could lead to an OK return being given in
1912 error.
1913
1914 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1915 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1916 in a Windows environment.
1917
1918 3. The pcregrep --count option prints the count for each file even when it is
1919 zero, as does GNU grep. However, pcregrep was also printing all files when
1920 --files-with-matches was added. Now, when both options are given, it prints
1921 counts only for those files that have at least one match. (GNU grep just
1922 prints the file name in this circumstance, but including the count seems
1923 more useful - otherwise, why use --count?) Also ensured that the
1924 combination -clh just lists non-zero counts, with no names.
1925
1926 4. The long form of the pcregrep -F option was incorrectly implemented as
1927 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1928 but it seems right to fix it, and I didn't think it was worth preserving
1929 the old behaviour.
1930
1931 5. The command line items --regex=pattern and --regexp=pattern were not
1932 recognized by pcregrep, which required --regex pattern or --regexp pattern
1933 (with a space rather than an '='). The man page documented the '=' forms,
1934 which are compatible with GNU grep; these now work.
1935
1936 6. No libpcreposix.pc file was created for pkg-config; there was just
1937 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1938
1939 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1940 when UCP support is not needed, by modifying the Python script that
1941 generates it from Unicode data files. This should not matter if the module
1942 is correctly used as a library, but I received one complaint about 50K of
1943 unwanted data. My guess is that the person linked everything into his
1944 program rather than using a library. Anyway, it does no harm.
1945
1946 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1947 was a minimum greater than 1 for a wide character in a possessive
1948 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1949 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1950 character. Chaos in the form of incorrect output or a compiling loop could
1951 result.
1952
1953 9. The restrictions on what a pattern can contain when partial matching is
1954 requested for pcre_exec() have been removed. All patterns can now be
1955 partially matched by this function. In addition, if there are at least two
1956 slots in the offset vector, the offset of the earliest inspected character
1957 for the match and the offset of the end of the subject are set in them when
1958 PCRE_ERROR_PARTIAL is returned.
1959
1960 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1961 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1962 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1963 and may be more useful for multi-segment matching.
1964
1965 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1966 used to be given if ever the end of the subject was reached; now it is
1967 given only if matching could not proceed because another character was
1968 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1969 string "Z", which now yields "no match" instead of "partial match". In the
1970 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1971 final character ended with (*FAIL).
1972
1973 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1974 if the pattern had a "must contain" character that was already found in the
1975 earlier partial match, unless partial matching was again requested. For
1976 example, with the pattern /dog.(body)?/, the "must contain" character is
1977 "g". If the first part-match was for the string "dog", restarting with
1978 "sbody" failed. This bug has been fixed.
1979
1980 13. The string returned by pcre_dfa_exec() after a partial match has been
1981 changed so that it starts at the first inspected character rather than the
1982 first character of the match. This makes a difference only if the pattern
1983 starts with a lookbehind assertion or \b or \B (\K is not supported by
1984 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1985 matching functions compatible, and I think it's the right thing to do.
1986
1987 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1988 so that the demonstration program is easily available in environments where
1989 PCRE has not been installed from source.
1990
1991 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1992 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1993 library.
1994
1995 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1996 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1997 is not the first non-POSIX option to be added. Clearly some people find
1998 these options useful.
1999
2000 17. If a caller to the POSIX matching function regexec() passes a non-zero
2001 value for nmatch with a NULL value for pmatch, the value of
2002 nmatch is forced to zero.
2003
2004 18. RunGrepTest did not have a test for the availability of the -u option of
2005 the diff command, as RunTest does. It now checks in the same way as
2006 RunTest, and also checks for the -b option.
2007
2008 19. If an odd number of negated classes containing just a single character
2009 interposed, within parentheses, between a forward reference to a named
2010 subpattern and the definition of the subpattern, compilation crashed with
2011 an internal error, complaining that it could not find the referenced
2012 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2013 [The bug was that it was starting one character too far in when skipping
2014 over the character class, thus treating the ] as data rather than
2015 terminating the class. This meant it could skip too much.]
2016
2017 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2018 /g option in pcretest when the pattern contains \K, which makes it possible
2019 to have an empty string match not at the start, even when the pattern is
2020 anchored. Updated pcretest and pcredemo to use this option.
2021
2022 21. If the maximum number of capturing subpatterns in a recursion was greater
2023 than the maximum at the outer level, the higher number was returned, but
2024 with unset values at the outer level. The correct (outer level) value is
2025 now given.
2026
2027 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2028 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2029 make it do so. The string so far is captured, making this feature
2030 compatible with Perl.
2031
2032 23. The tests have been re-organized, adding tests 11 and 12, to make it
2033 possible to check the Perl 5.10 features against Perl 5.10.
2034
2035 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2036 pattern matches a fixed length string. PCRE did not allow this; now it
2037 does. Neither allows recursion.
2038
2039 25. I finally figured out how to implement a request to provide the minimum
2040 length of subject string that was needed in order to match a given pattern.
2041 (It was back references and recursion that I had previously got hung up
2042 on.) This code has now been added to pcre_study(); it finds a lower bound
2043 to the length of subject needed. It is not necessarily the greatest lower
2044 bound, but using it to avoid searching strings that are too short does give
2045 some useful speed-ups. The value is available to calling programs via
2046 pcre_fullinfo().
2047
2048 26. While implementing 25, I discovered to my embarrassment that pcretest had
2049 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2050 study optimizations had never been tested with that matching function.
2051 Oops. What is worse, even when it was passed study data, there was a bug in
2052 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2053 were also very few tests of studied patterns with pcre_dfa_exec().
2054
2055 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2056 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2057 on the other side of the coin, they are no longer allowed to have different
2058 names, because these cannot be distinguished in PCRE, and this has caused
2059 confusion. (This is a difference from Perl.)
2060
2061 28. When duplicate subpattern names are present (necessarily with different
2062 numbers, as required by 27 above), and a test is made by name in a
2063 conditional pattern, either for a subpattern having been matched, or for
2064 recursion in such a pattern, all the associated numbered subpatterns are
2065 tested, and the overall condition is true if the condition is true for any
2066 one of them. This is the way Perl works, and is also more like the way
2067 testing by number works.
2068
2069
2070 Version 7.9 11-Apr-09
2071 ---------------------
2072
2073 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2074 (pcretest), all targets were linked against these libraries. This included
2075 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2076 libraries. This caused unwanted dependencies to be created. This problem
2077 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2078 pcretest is linked with readline.
2079
2080 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2081 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2082 moved outside it again, because FALSE and TRUE are already defined in AIX,
2083 but BOOL is not.
2084
2085 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2086 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2087
2088 4. The pcregrep documentation said spaces were inserted as well as colons (or
2089 hyphens) following file names and line numbers when outputting matching
2090 lines. This is not true; no spaces are inserted. I have also clarified the
2091 wording for the --colour (or --color) option.
2092
2093 5. In pcregrep, when --colour was used with -o, the list of matching strings
2094 was not coloured; this is different to GNU grep, so I have changed it to be
2095 the same.
2096
2097 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2098 each matching line was coloured. Now it goes on to look for further matches
2099 of any of the test patterns, which is the same behaviour as GNU grep.
2100
2101 7. A pattern that could match an empty string could cause pcregrep to loop; it
2102 doesn't make sense to accept an empty string match in pcregrep, so I have
2103 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2104 seems to be how GNU grep behaves. [But see later change 40 for release
2105 8.33.]
2106
2107 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2108 start or after a newline", because the conditional assertion was not being
2109 correctly handled. The rule now is that both the assertion and what follows
2110 in the first alternative must satisfy the test.
2111
2112 9. If auto-callout was enabled in a pattern with a conditional group whose
2113 condition was an assertion, PCRE could crash during matching, both with
2114 pcre_exec() and pcre_dfa_exec().
2115
2116 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2117 used for matching.
2118
2119 11. Unicode property support in character classes was not working for
2120 characters (bytes) greater than 127 when not in UTF-8 mode.
2121
2122 12. Added the -M command line option to pcretest.
2123
2124 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2125
2126 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2127
2128 16. Added comments and documentation about mis-use of no_arg in the C++
2129 wrapper.
2130
2131 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2132 from Martin Jerabek that uses macro names for all relevant character and
2133 string constants.
2134
2135 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2136 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2137 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2138 these, but not everybody uses configure.
2139
2140 19. A conditional group that had only one branch was not being correctly
2141 recognized as an item that could match an empty string. This meant that an
2142 enclosing group might also not be so recognized, causing infinite looping
2143 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2144 with the subject "ab", where knowledge that the repeated group can match
2145 nothing is needed in order to break the loop.
2146
2147 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2148 exec(), but without supplying a callout function, matching went wrong.
2149
2150 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2151 leak if the size of the offset vector was greater than 30. When the vector
2152 is smaller, the saved offsets during recursion go onto a local stack
2153 vector, but for larger vectors malloc() is used. It was failing to free
2154 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2155 error, in fact).
2156
2157 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2158 heapframe that is used only when UTF-8 support is enabled. This caused no
2159 problem, but was untidy.
2160
2161 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2162 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2163 included within another project.
2164
2165 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2166 slightly modified by me:
2167
2168 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2169 not building pcregrep.
2170
2171 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2172 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2173
2174 25. Forward references, both numeric and by name, in patterns that made use of
2175 duplicate group numbers, could behave incorrectly or give incorrect errors,
2176 because when scanning forward to find the reference group, PCRE was not
2177 taking into account the duplicate group numbers. A pattern such as
2178 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2179
2180 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2181 the feature of a custom pointer more persuasive (as requested by a user).
2182
2183 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2184 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2185 pre-defined.
2186
2187 28. Added support for (*UTF8) at the start of a pattern.
2188
2189 29. Arrange for flags added by the "release type" setting in CMake to be shown
2190 in the configuration summary.
2191
2192
2193 Version 7.8 05-Sep-08
2194 ---------------------
2195
2196 1. Replaced UCP searching code with optimized version as implemented for Ad
2197 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2198 stage table and inline lookup instead of a function, giving speed ups of 2
2199 to 5 times on some simple patterns that I tested. Permission was given to
2200 distribute the MultiStage2.py script that generates the tables (it's not in
2201 the tarball, but is in the Subversion repository).
2202
2203 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2204 scripts.
2205
2206 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2207 a group with a zero qualifier. The result of the study could be incorrect,
2208 or the function might crash, depending on the pattern.
2209
2210 4. Caseless matching was not working for non-ASCII characters in back
2211 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2212 It now works when Unicode Property Support is available.
2213
2214 5. In pcretest, an escape such as \x{de} in the data was always generating
2215 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2216 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2217 truncation.
2218
2219 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2220
2221 7. Added two (int) casts to pcregrep when printing the difference of two
2222 pointers, in case they are 64-bit values.
2223
2224 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2225 test 2 if it fails.
2226
2227 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2228 and a #define of that name to empty if it is not externally set. This is to
2229 allow users of MSVC to set it if necessary.
2230
2231 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2232 the convenience functions in the pcre_get.c source file.
2233
2234 11. An option change at the start of a pattern that had top-level alternatives
2235 could cause overwriting and/or a crash. This command provoked a crash in
2236 some environments:
2237
2238 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2239
2240 This potential security problem was recorded as CVE-2008-2371.
2241
2242 12. For a pattern where the match had to start at the beginning or immediately
2243 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2244 pcre_dfa_exec() could read past the end of the passed subject if there was
2245 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2246 pcretest so that it places the subject at the end of its malloc-ed buffer.
2247
2248 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2249 exec() might read past the end of the data buffer in UTF-8 mode.
2250
2251 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2252 the data contained the byte 0x85 as part of a UTF-8 character within its
2253 first line. This applied both to normal and DFA matching.
2254
2255 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2256 /^[^d]*?$/8 failed to match "abc".
2257
2258 16. Added a missing copyright notice to pcrecpp_internal.h.
2259
2260 17. Make it more clear in the documentation that values returned from
2261 pcre_exec() in ovector are byte offsets, not character counts.
2262
2263 18. Tidied a few places to stop certain compilers from issuing warnings.
2264
2265 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2266 supplied by Stefan Weber. I made a further small update for 7.8 because
2267 there is a change of source arrangements: the pcre_searchfuncs.c module is
2268 replaced by pcre_ucd.c.
2269
2270
2271 Version 7.7 07-May-08
2272 ---------------------
2273
2274 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2275 a string to a long long, pretend we don't even have a long long." This is
2276 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2277
2278 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2279 pre-7.6 versions, which defined a global no_arg variable instead of putting
2280 it in the RE class. (See also #8 below.)
2281
2282 3. Remove a line of dead code, identified by coverity and reported by Nuno
2283 Lopes.
2284
2285 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2286
2287 (1) The include/exclude patterns were being applied to the whole pathnames
2288 of files, instead of just to the final components.
2289
2290 (2) If there was more than one level of directory, the subdirectories were
2291 skipped unless they satisfied the include/exclude conditions. This is
2292 inconsistent with GNU grep (and could even be seen as contrary to the
2293 pcregrep specification - which I improved to make it absolutely clear).
2294 The action now is always to scan all levels of directory, and just
2295 apply the include/exclude patterns to regular files.
2296
2297 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2298 --exclude_dir in the tests to avoid scanning .svn directories.
2299
2300 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2301 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2302 doesn't support NULs in patterns.
2303
2304 7. Added some missing "const"s to declarations of static tables in
2305 pcre_compile.c and pcre_dfa_exec.c.
2306
2307 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2308 caused by fix #2 above. (Subsequently also a second patch to fix the
2309 first patch. And a third patch - this was a messy problem.)
2310
2311 9. Applied Craig's patch to remove the use of push_back().
2312
2313 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2314 matching function regexec().
2315
2316 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2317 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2318 references. PCRE supports relative numbers with this syntax (I don't think
2319 Oniguruma does).
2320
2321 12. Previously, a group with a zero repeat such as (...){0} was completely
2322 omitted from the compiled regex. However, this means that if the group
2323 was called as a subroutine from elsewhere in the pattern, things went wrong
2324 (an internal error was given). Such groups are now left in the compiled
2325 pattern, with a new opcode that causes them to be skipped at execution
2326 time.
2327
2328 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2329 to the way PCRE behaves:
2330
2331 (a) A lone ] character is dis-allowed (Perl treats it as data).
2332
2333 (b) A back reference to an unmatched subpattern matches an empty string
2334 (Perl fails the current match path).
2335
2336 (c) A data ] in a character class must be notated as \] because if the
2337 first data character in a class is ], it defines an empty class. (In
2338 Perl it is not possible to have an empty class.) The empty class []
2339 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2340 The negative empty class [^] matches any one character, independently
2341 of the DOTALL setting.
2342
2343 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2344 non-existent subpattern following a character class starting with ']' and
2345 containing () gave an internal compiling error instead of "reference to
2346 non-existent subpattern". Fortunately, when the pattern did exist, the
2347 compiled code was correct. (When scanning forwards to check for the
2348 existence of the subpattern, it was treating the data ']' as terminating
2349 the class, so got the count wrong. When actually compiling, the reference
2350 was subsequently set up correctly.)
2351
2352 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2353 it was being rejected as not supported by pcre_dfa_exec(), even though
2354 other assertions are supported. I have made pcre_dfa_exec() support
2355 (*FAIL).
2356
2357 16. The implementation of 13c above involved the invention of a new opcode,
2358 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2359 cannot be changed at match time, I realized I could make a small
2360 improvement to matching performance by compiling OP_ALLANY instead of
2361 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2362 on the OP_ANY path.
2363
2364 17. Compiling pcretest on Windows with readline support failed without the
2365 following two fixes: (1) Make the unistd.h include conditional on
2366 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2367
2368 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2369 ncurses library to be included for pcretest when ReadLine support is
2370 requested, but also to allow for it to be overridden. This patch came from
2371 Daniel Bergström.
2372
2373 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2374 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2375 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2376 spotting this.
2377
2378
2379 Version 7.6 28-Jan-08
2380 ---------------------
2381
2382 1. A character class containing a very large number of characters with
2383 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2384 overflow.
2385
2386 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2387 HAVE_LONG_LONG is not defined.
2388
2389 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2390 bring them up to date and include new features. This patch includes:
2391
2392 - Fixed PH's badly added libz and libbz2 support.
2393 - Fixed a problem with static linking.
2394 - Added pcredemo. [But later removed - see 7 below.]
2395 - Fixed dftables problem and added an option.
2396 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2397 HAVE_LONG_LONG.
2398 - Added readline support for pcretest.
2399 - Added an listing of the option settings after cmake has run.
2400
2401 4. A user submitted a patch to Makefile that makes it easy to create
2402 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2403 Makefile.am that cause it to include this special target, without
2404 affecting anything else. Note that the same mingw target plus all
2405 the other distribution libraries and programs are now supported
2406 when configuring with CMake (see 6 below) instead of with
2407 Configure/Make.
2408
2409 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2410 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2411 exported in the Windows port". It has not yet been confirmed that the patch
2412 solves the problem, but it does no harm.
2413
2414 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2415 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2416 with CMake, and also correct the comment about stack recursion.
2417
2418 7. Remove the automatic building of pcredemo from the ./configure system and
2419 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2420 of a program that users should build themselves after PCRE is installed, so
2421 building it automatically is not really right. What is more, it gave
2422 trouble in some build environments.
2423
2424 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2425
2426
2427 Version 7.5 10-Jan-08
2428 ---------------------
2429
2430 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2431 values in parens when parsing an RE using the C++ wrapper."
2432
2433 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2434 Characters greater than 255 were excluded from the class instead of being
2435 included.
2436
2437 3. The same bug as (2) above applied to negated POSIX classes such as
2438 [:^space:].
2439
2440 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2441 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2442 I have changed it.
2443
2444 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2445 first named subpattern) and a construct such as (?&a) would reference the
2446 first named subpattern whose name started with "a" (in other words, the
2447 length check was missing). Both these problems are fixed. "Subpattern name
2448 expected" is now given for (?&) (a zero-length name), and this patch also
2449 makes it give the same error for \k'' (previously it complained that that
2450 was a reference to a non-existent subpattern).
2451
2452 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2453 this is right because (?- can be followed by option settings as well as by
2454 digits. I have, however, made the messages clearer.
2455
2456 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2457 than the number used in the conditional) now cause a compile-time error.
2458 This is actually not compatible with Perl, which accepts such patterns, but
2459 treats the conditional as always being FALSE (as PCRE used to), but it
2460 seems to me that giving a diagnostic is better.
2461
2462 8. Change "alphameric" to the more common word "alphanumeric" in comments
2463 and messages.
2464
2465 9. Fix two occurrences of "backslash" in comments that should have been
2466 "backspace".
2467
2468 10. Remove two redundant lines of code that can never be obeyed (their function
2469 was moved elsewhere).
2470
2471 11. The program that makes PCRE's Unicode character property table had a bug
2472 which caused it to generate incorrect table entries for sequences of
2473 characters that have the same character type, but are in different scripts.
2474 It amalgamated them into a single range, with the script of the first of
2475 them. In other words, some characters were in the wrong script. There were
2476 thirteen such cases, affecting characters in the following ranges:
2477
2478 U+002b0 - U+002c1
2479 U+0060c - U+0060d
2480 U+0061e - U+00612
2481 U+0064b - U+0065e
2482 U+0074d - U+0076d
2483 U+01800 - U+01805
2484 U+01d00 - U+01d77
2485 U+01d9b - U+01dbf
2486 U+0200b - U+0200f
2487 U+030fc - U+030fe
2488 U+03260 - U+0327f
2489 U+0fb46 - U+0fbb1
2490 U+10450 - U+1049d
2491
2492 12. The -o option (show only the matching part of a line) for pcregrep was not
2493 compatible with GNU grep in that, if there was more than one match in a
2494 line, it showed only the first of them. It now behaves in the same way as
2495 GNU grep.
2496
2497 13. If the -o and -v options were combined for pcregrep, it printed a blank
2498 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2499 does the same. The return code can be used to tell if there were any
2500 non-matching lines.
2501
2502 14. Added --file-offsets and --line-offsets to pcregrep.
2503
2504 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2505 infinitely looping recursion. The bug was that positive lookaheads were not
2506 being skipped when checking for a possible empty match (negative lookaheads
2507 and both kinds of lookbehind were skipped).
2508
2509 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2510 inclusion of <windows.h> to before rather than after the definition of
2511 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2512
2513 17. Specifying a possessive quantifier with a specific limit for a Unicode
2514 character property caused pcre_compile() to compile bad code, which led at
2515 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2516 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2517 caused the error; without that there was no problem.
2518
2519 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2520
2521 19. Added --enable-pcretest-libreadline.
2522
2523 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2524 RE::GlobalReplace(). As a result, the number of replacements returned was
2525 double what it should be. I removed one of the increments, but Craig sent a
2526 later patch that removed the other one (the right fix) and added unit tests
2527 that check the return values (which was not done before).
2528
2529 21. Several CMake things:
2530
2531 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2532 the names libpcre and libpcreposix, not just pcre and pcreposix.
2533
2534 (2) The above change means that pcretest and pcregrep are now correctly
2535 linked with the newly-built libraries, not previously installed ones.
2536
2537 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2538
2539 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2540 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2541 UTF-8 newline character). The key issue is that the pattern starts .*;
2542 this means that the match must be either at the beginning, or after a
2543 newline. The bug was in the code for advancing after a failed match and
2544 checking that the new position followed a newline. It was not taking
2545 account of UTF-8 characters correctly.
2546
2547 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2548 character classes. PCRE was not treating the sequence [:...:] as a
2549 character class unless the ... were all letters. Perl, however, seems to
2550 allow any characters between [: and :], though of course it rejects as
2551 unknown any "names" that contain non-letters, because all the known class
2552 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2553 for example, whereas PCRE did not - it did not recognize a POSIX character
2554 class. This seemed a bit dangerous, so the code has been changed to be
2555 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2556 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2557 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2558 Perl does, and where it didn't before.
2559
2560 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2561 Windows environments %n is disabled by default.
2562
2563
2564 Version 7.4 21-Sep-07
2565 ---------------------
2566
2567 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2568 means that a class such as [\s] counted as "explicit reference to CR or
2569 LF". That isn't really right - the whole point of the change was to try to
2570 help when there was an actual mention of one of the two characters. So now
2571 the change happens only if \r or \n (or a literal CR or LF) character is
2572 encountered.
2573
2574 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2575 of both had grown to the point where there were only 3 bits left.
2576 Fortunately, there was spare space in the data structure, and so I have
2577 moved the internal flags into a new 16-bit field to free up more option
2578 bits.
2579
2580 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2581 but did not set the internal JCHANGED flag - either of these is enough to
2582 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2583 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2584 start sets both bits.
2585
2586 4. Added options (at build time, compile time, exec time) to change \R from
2587 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2588
2589 5. doc/pcresyntax.html was missing from the distribution.
2590
2591 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2592 compatibility, even though it is no longer used.
2593
2594 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2595 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2596 windows.h file is present (where different names are used). [This was
2597 reversed later after testing - see 16 below.]
2598
2599 8. Changed all #include <config.h> to #include "config.h". There were also
2600 some further <pcre.h> cases that I changed to "pcre.h".
2601
2602 9. When pcregrep was used with the --colour option, it missed the line ending
2603 sequence off the lines that it output.
2604
2605 10. It was pointed out to me that arrays of string pointers cause lots of
2606 relocations when a shared library is dynamically loaded. A technique of
2607 using a single long string with a table of offsets can drastically reduce
2608 these. I have refactored PCRE in four places to do this. The result is
2609 dramatic:
2610
2611 Originally: 290
2612 After changing UCP table: 187
2613 After changing error message table: 43
2614 After changing table of "verbs" 36
2615 After changing table of Posix names 22
2616
2617 Thanks to the folks working on Gregex for glib for this insight.
2618
2619 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2620 unicode-properties was also set.
2621
2622 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2623
2624 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2625 checked only for CRLF.
2626
2627 14. Added casts to pcretest.c to avoid compiler warnings.
2628
2629 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2630
2631 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2632 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2633 entirely. This removes changes made in 7 above.
2634
2635 17. The CMake files have been updated, and there is now more information about
2636 building with CMake in the NON-UNIX-USE document.
2637
2638
2639 Version 7.3 28-Aug-07
2640 ---------------------
2641
2642 1. In the rejigging of the build system that eventually resulted in 7.1, the
2643 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2644 brackets there is not right, since it causes compilers to look for an
2645 installed pcre.h, not the version that is in the source that is being
2646 compiled (which of course may be different). I have changed it back to:
2647
2648 #include "pcre.h"
2649
2650 I have a vague recollection that the change was concerned with compiling in
2651 different directories, but in the new build system, that is taken care of
2652 by the VPATH setting the Makefile.
2653
2654 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2655 when the subject happened to end in the byte 0x85 (e.g. if the last
2656 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2657 characters but of course it shouldn't be taken as a newline when it is part
2658 of another character. The bug was that, for an unlimited repeat of . in
2659 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2660 characters when looking for a newline.
2661
2662 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2663
2664 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2665 in debug output.
2666
2667 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2668 long printing in the pcrecpp unittest when running under MinGW.
2669
2670 6. ESC_K was left out of the EBCDIC table.
2671
2672 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2673 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2674 limit also applies to "virtual nesting" when a pattern is recursive, and in
2675 this case 1000 isn't so big. I have been able to remove this limit at the
2676 expense of backing off one optimization in certain circumstances. Normally,
2677 when pcre_exec() would call its internal match() function recursively and
2678 immediately return the result unconditionally, it uses a "tail recursion"
2679 feature to save stack. However, when a subpattern that can match an empty
2680 string has an unlimited repetition quantifier, it no longer makes this
2681 optimization. That gives it a stack frame in which to save the data for
2682 checking that an empty string has been matched. Previously this was taken
2683 from the 1000-entry workspace that had been reserved. So now there is no
2684 explicit limit, but more stack is used.
2685
2686 8. Applied Daniel's patches to solve problems with the import/export magic
2687 syntax that is required for Windows, and which was going wrong for the
2688 pcreposix and pcrecpp parts of the library. These were overlooked when this
2689 problem was solved for the main library.
2690
2691 9. There were some crude static tests to avoid integer overflow when computing
2692 the size of patterns that contain repeated groups with explicit upper
2693 limits. As the maximum quantifier is 65535, the maximum group length was
2694 set at 30,000 so that the product of these two numbers did not overflow a
2695 32-bit integer. However, it turns out that people want to use groups that
2696 are longer than 30,000 bytes (though not repeat them that many times).
2697 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2698 made it possible to implement the integer overflow checks in a much more
2699 dynamic way, which I have now done. The artificial limitation on group
2700 length has been removed - we now have only the limit on the total length of
2701 the compiled pattern, which depends on the LINK_SIZE setting.
2702
2703 10. Fixed a bug in the documentation for get/copy named substring when
2704 duplicate names are permitted. If none of the named substrings are set, the
2705 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2706 empty string.
2707
2708 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2709 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2710 because the ] is interpreted as the first data character and the
2711 terminating ] is not found. PCRE has been made compatible with Perl in this
2712 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2713 cause memory overwriting.
2714
2715 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2716 string has been matched (to stop an infinite loop). It was not recognizing
2717 a conditional subpattern that could match an empty string if that
2718 subpattern was within another subpattern. For example, it looped when
2719 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2720 condition was not nested. This bug has been fixed.
2721
2722 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2723 past the start of the subject in the presence of bytes with the top bit
2724 set, for example "\x8aBCD".
2725
2726 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2727 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2728
2729 14. Optimized (?!) to (*FAIL).
2730
2731 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2732 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2733 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2734 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2735 does: it's just the validity check that is more restrictive.
2736
2737 16. Inserted checks for integer overflows during escape sequence (backslash)
2738 processing, and also fixed erroneous offset values for syntax errors during
2739 backslash processing.
2740
2741 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2742 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2743
2744 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2745 caused an overrun.
2746
2747 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2748 something other than just ASCII characters) inside a group that had an
2749 unlimited repeat caused a loop at compile time (while checking to see
2750 whether the group could match an empty string).
2751
2752 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2753 [\P{Any}] did so. (Error in the code for printing property names.)
2754
2755 21. An orphan \E inside a character class could cause a crash.
2756
2757 22. A repeated capturing bracket such as (A)? could cause a wild memory
2758 reference during compilation.
2759
2760 23. There are several functions in pcre_compile() that scan along a compiled
2761 expression for various reasons (e.g. to see if it's fixed length for look
2762 behind). There were bugs in these functions when a repeated \p or \P was
2763 present in the pattern. These operators have additional parameters compared
2764 with \d, etc, and these were not being taken into account when moving along
2765 the compiled data. Specifically:
2766
2767 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2768 length.
2769
2770 (b) An item such as \pL+ within a repeated group could cause crashes or
2771 loops.
2772
2773 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2774 "reference to non-existent subpattern" error.
2775
2776 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2777
2778 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2779 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2780
2781 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2782
2783 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2784 character were causing crashes (broken optimization).
2785
2786 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2787 \p or \P) caused a compile-time loop.
2788
2789 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2790 break. For example, the unstudied pattern [\r\n]A does not match the string
2791 "\r\nA" because change 7.0/46 below moves the current point on by two
2792 characters after failing to match at the start. However, the pattern \nA
2793 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2794 the same is true. There doesn't seem any very clean way out of this, but
2795 what I have chosen to do makes the common cases work: PCRE now takes note
2796 of whether there can be an explicit match for \r or \n anywhere in the
2797 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2798 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2799 pattern has explicit CR or LF references.
2800
2801 29. Added (*CR) etc for changing newline setting at start of pattern.
2802
2803
2804 Version 7.2 19-Jun-07
2805 ---------------------
2806
2807 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2808 which is apparently normally available under Windows.
2809
2810 2. Re-jig the pcregrep tests with different newline settings in an attempt
2811 to make them independent of the local environment's newline setting.
2812
2813 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2814
2815 4. Some of the "internals" tests were previously cut out when the link size
2816 was not 2, because the output contained actual offsets. The recent new
2817 "Z" feature of pcretest means that these can be cut out, making the tests
2818 usable with all link sizes.
2819
2820 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2821 stack recursion. This gives a massive performance boost under BSD, but just
2822 a small improvement under Linux. However, it saves one field in the frame
2823 in all cases.
2824
2825 6. Added more features from the forthcoming Perl 5.10:
2826
2827 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2828 recursion call. It refers to the nth most recently opened parentheses.
2829
2830 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2831 to be opened parentheses.
2832
2833 (c) Conditions that refer to capturing parentheses can be specified
2834 relatively, for example, (?(-2)... or (?(+3)...
2835
2836 (d) \K resets the start of the current match so that everything before
2837 is not part of it.
2838
2839 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2840
2841 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2842 reference syntax.
2843
2844 (g) (?| introduces a group in which the numbering of parentheses in each
2845 alternative starts with the same number.
2846
2847 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2848
2849 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2850 PCRE_INFO_JCHANGED.
2851
2852 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2853 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2854 for detecting groups that can match an empty string.
2855
2856 9. A pattern with a very large number of alternatives (more than several
2857 hundred) was running out of internal workspace during the pre-compile
2858 phase, where pcre_compile() figures out how much memory will be needed. A
2859 bit of new cunning has reduced the workspace needed for groups with
2860 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2861 workspace instead of running out of the 4096 that are available.
2862
2863 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2864
2865 11. Applied patch from Google to remove an optimization that didn't quite work.
2866 The report of the bug said:
2867
2868 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2869 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2870 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2871
2872 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2873 it matched the wrong number of bytes.
2874
2875
2876 Version 7.1 24-Apr-07
2877 ---------------------
2878
2879 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2880 that is more "standard", making use of automake and other Autotools. There
2881 is some re-arrangement of the files and adjustment of comments consequent
2882 on this.
2883
2884 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2885 for recursive directory scanning broke on some systems because the files
2886 are not scanned in any specific order and on different systems the order
2887 was different. A call to "sort" has been inserted into RunGrepTest for the
2888 approprate test as a short-term fix. In the longer term there may be an
2889 alternative.
2890
2891 3. I had an email from Eric Raymond about problems translating some of PCRE's
2892 man pages to HTML (despite the fact that I distribute HTML pages, some
2893 people do their own conversions for various reasons). The problems
2894 concerned the use of low-level troff macros .br and .in. I have therefore
2895 removed all such uses from the man pages (some were redundant, some could
2896 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2897 HTML has been updated to handle .nf/.fi and to complain if it encounters
2898 .br or .in.
2899
2900 4. Updated comments in configure.ac that get placed in config.h.in and also
2901 arranged for config.h to be included in the distribution, with the name
2902 config.h.generic, for the benefit of those who have to compile without
2903 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2904
2905 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2906 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2907 makevp.bat for the current PCRE, using the additional files
2908 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2909
2910 6. A Windows user reported a minor discrepancy with test 2, which turned out
2911 to be caused by a trailing space on an input line that had got lost in his
2912 copy. The trailing space was an accident, so I've just removed it.
2913
2914 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2915 that is needed.
2916
2917 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2918 as "const" (a) because they are and (b) because it helps the PHP
2919 maintainers who have recently made a script to detect big data structures
2920 in the php code that should be moved to the .rodata section. I remembered
2921 to update Builducptable as well, so it won't revert if ucptable.h is ever
2922 re-created.
2923
2924 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2925 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2926 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2927 support is not required. This saves 1.5-2K of code, which is important in
2928 some applications.
2929
2930 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2931 so as not to refer to the tables, even though these functions will never be
2932 called when UTF-8 support is disabled. Otherwise there are problems with a
2933 shared library.
2934
2935 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2936
2937 (a) It was defining its arguments as char * instead of void *.
2938
2939 (b) It was assuming that all moves were upwards in memory; this was true
2940 a long time ago when I wrote it, but is no longer the case.
2941
2942 The emulated memove() is provided for those environments that have neither
2943 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2944 is clearly not the case, as these two bugs were recently reported.
2945
2946 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2947 and Detrail to create the HTML documentation, the .txt form of the man
2948 pages, and it removes trailing spaces from listed files. It also creates
2949 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2950 case, it wraps all the #defines with #ifndefs. This script should be run
2951 before "make dist".
2952
2953 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2954 with Unicode property support.
2955
2956 (a) For a maximizing quantifier, if the two different cases of the
2957 character were of different lengths in their UTF-8 codings (there are
2958 some cases like this - I found 11), and the matching function had to
2959 back up over a mixture of the two cases, it incorrectly assumed they
2960 were both the same length.
2961
2962 (b) When PCRE was configured to use the heap rather than the stack for
2963 recursion during matching, it was not correctly preserving the data for
2964 the other case of a UTF-8 character when checking ahead for a match
2965 while processing a minimizing repeat. If the check also involved
2966 matching a wide character, but failed, corruption could cause an
2967 erroneous result when trying to check for a repeat of the original
2968 character.
2969
2970 13. Some tidying changes to the testing mechanism:
2971
2972 (a) The RunTest script now detects the internal link size and whether there
2973 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2974 values substituted by "configure". (The RunGrepTest script already did
2975 this for UTF-8.) The configure.ac script no longer substitutes the
2976 relevant variables.
2977
2978 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2979 with length and offset values. This means that the output is different
2980 for different internal link sizes. Test 2 is skipped for link sizes
2981 other than 2 because of this, bypassing the problem. Unfortunately,
2982 there was also a test in test 3 (the locale tests) that used /B and
2983 failed for link sizes other than 2. Rather than cut the whole test out,
2984 I have added a new /Z option to pcretest that replaces the length and
2985 offset values with spaces. This is now used to make test 3 independent
2986 of link size. (Test 2 will be tidied up later.)
2987
2988 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2989 segmentation fault instead of returning the appropriate error message.
2990
2991 15. In multiline mode when the newline sequence was set to "any", the pattern
2992 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2993 This doesn't seem right; it now treats the CRLF combination as the line
2994 ending, and so does not match in that case. It's only a pattern such as ^$
2995 that would hit this one: something like ^ABC$ would have failed after \r
2996 and then tried again after \r\n.
2997
2998 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2999 in an attempt to make files that differ only in their line terminators
3000 compare equal. This works on Linux.
3001
3002 17. Under certain error circumstances pcregrep might try to free random memory
3003 as it exited. This is now fixed, thanks to valgrind.
3004
3005 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3006 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3007 was because its rules for how to advance for /g after matching an empty
3008 string at the end of a line did not allow for this case. They now check for
3009 it specially.
3010
3011 20. pcretest is supposed to handle patterns and data of any length, by
3012 extending its buffers when necessary. It was getting this wrong when the
3013 buffer for a data line had to be extended.
3014
3015 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3016 CRLF as a newline sequence.
3017
3018 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3019 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3020 I have nevertheless tidied it up.
3021
3022 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3023
3024 24. Added a man page for pcre-config.
3025
3026
3027 Version 7.0 19-Dec-06
3028 ---------------------
3029
3030 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3031 moving to gcc 4.1.1.
3032
3033 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3034 sys/time.h, which is documented as needed for this function. It doesn't
3035 seem to matter on Linux, but it showed up on some releases of OS X.
3036
3037 3. It seems that there are systems where bytes whose values are greater than
3038 127 match isprint() in the "C" locale. The "C" locale should be the
3039 default when a C program starts up. In most systems, only ASCII printing
3040 characters match isprint(). This difference caused the output from pcretest
3041 to vary, making some of the tests fail. I have changed pcretest so that:
3042
3043 (a) When it is outputting text in the compiled version of a pattern, bytes
3044 other than 32-126 are always shown as hex escapes.
3045
3046 (b) When it is outputting text that is a matched part of a subject string,
3047 it does the same, unless a different locale has been set for the match
3048 (using the /L modifier). In this case, it uses isprint() to decide.
3049
3050 4. Fixed a major bug that caused incorrect computation of the amount of memory
3051 required for a compiled pattern when options that changed within the
3052 pattern affected the logic of the preliminary scan that determines the
3053 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3054 that the computed length was too small. The symptoms of this bug were
3055 either the PCRE error "internal error: code overflow" from pcre_compile(),
3056 or a glibc crash with a message such as "pcretest: free(): invalid next
3057 size (fast)". Examples of patterns that provoked this bug (shown in
3058 pcretest format) are:
3059
3060 /(?-x: )/x
3061 /(?x)(?-x: \s*#\s*)/
3062 /((?i)[\x{c0}])/8
3063 /(?i:[\x{c0}])/8
3064
3065 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3066 is now done differently.
3067
3068 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3069 wrapper classes; (b) implement a new function in the C++ scanner that is
3070 more efficient than the old way of doing things because it avoids levels of
3071 recursion in the regex matching; (c) add a paragraph to the documentation
3072 for the FullMatch() function.
3073
3074 6. The escape sequence \n was being treated as whatever was defined as
3075 "newline". Not only was this contrary to the documentation, which states
3076 that \n is character 10 (hex 0A), but it also went horribly wrong when
3077 "newline" was defined as CRLF. This has been fixed.
3078
3079 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3080 was being set to -1 for the "end of line" case (supposedly a value that no
3081 character can have). Though this value is never used (the check for end of
3082 line is "zero bytes in current character"), it caused compiler complaints.
3083 I've changed it to 0xffffffff.
3084
3085 8. In pcre_version.c, the version string was being built by a sequence of
3086 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3087 string (as it is for production releases) called a macro with an empty
3088 argument. The C standard says the result of this is undefined. The gcc
3089 compiler treats it as an empty string (which was what was wanted) but it is
3090 reported that Visual C gives an error. The source has been hacked around to
3091 avoid this problem.
3092
3093 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3094 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3095 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3096 of them did).
3097
3098 10. Originally, pcretest opened its input and output without "b"; then I was
3099 told that "b" was needed in some environments, so it was added for release
3100 5.0 to both the input and output. (It makes no difference on Unix-like
3101 systems.) Later I was told that it is wrong for the input on Windows. I've
3102 now abstracted the modes into two macros, to make it easier to fiddle with
3103 them, and removed "b" from the input mode under Windows.
3104
3105 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3106
3107 12. Added -help and --help to pcretest as an official way of being reminded
3108 of the options.
3109
3110 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3111 and pcrecpp.cc because they annoy compilers at high warning levels.
3112
3113 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3114
3115 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3116 scripts are not C programs :-) and which was not noticed because it works
3117 on Linux.
3118
3119 16. pcretest is supposed to handle any length of pattern and data line (as one
3120 line or as a continued sequence of lines) by extending its input buffer if
3121 necessary. This feature was broken for very long pattern lines, leading to
3122 a string of junk being passed to pcre_compile() if the pattern was longer
3123 than about 50K.
3124
3125 17. I have done a major re-factoring of the way pcre_compile() computes the
3126 amount of memory needed for a compiled pattern. Previously, there was code
3127 that made a preliminary scan of the pattern in order to do this. That was
3128 OK when PCRE was new, but as the facilities have expanded, it has become
3129 harder and harder to keep it in step with the real compile phase, and there
3130 have been a number of bugs (see for example, 4 above). I have now found a
3131 cunning way of running the real compile function in a "fake" mode that
3132 enables it to compute how much memory it would need, while actually only
3133 ever using a few hundred bytes of working memory and without too many
3134 tests of the mode. This should make future maintenance and development
3135 easier. A side effect of this work is that the limit of 200 on the nesting
3136 depth of parentheses has been removed (though this was never a serious
3137 limitation, I suspect). However, there is a downside: pcre_compile() now
3138 runs more slowly than before (30% or more, depending on the pattern). I
3139 hope this isn't a big issue. There is no effect on runtime performance.
3140
3141 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3142 newline (only possible for the last line of a file) and it was a
3143 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3144
3145 19. Added additional timing features to pcretest. (1) The -tm option now times
3146 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3147 separate command line item, by a number that specifies the number of
3148 repeats to use when timing. The default is 50000; this gives better
3149 precision, but takes uncomfortably long for very large patterns.
3150
3151 20. Extended pcre_study() to be more clever in cases where a branch of a
3152 subpattern has no definite first character. For example, (a*|b*)[cd] would
3153 previously give no result from pcre_study(). Now it recognizes that the
3154 first character must be a, b, c, or d.
3155
3156 21. There was an incorrect error "recursive call could loop indefinitely" if
3157 a subpattern (or the entire pattern) that was being tested for matching an
3158 empty string contained only one non-empty item after a nested subpattern.
3159 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3160 incorrectly, because the \d was being skipped in the check.
3161
3162 22. The pcretest program now has a new pattern option /B and a command line
3163 option -b, which is equivalent to adding /B to every pattern. This causes
3164 it to show the compiled bytecode, without the additional information that
3165 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3166 is the same as /B/I).
3167
3168 23. A new optimization is now able automatically to treat some sequences such
3169 as a*b as a*+b. More specifically, if something simple (such as a character
3170 or a simple class like \d) has an unlimited quantifier, and is followed by
3171 something that cannot possibly match the quantified thing, the quantifier
3172 is automatically "possessified".
3173
3174 24. A recursive reference to a subpattern whose number was greater than 39
3175 went wrong under certain circumstances in UTF-8 mode. This bug could also
3176 have affected the operation of pcre_study().
3177
3178 25. Realized that a little bit of performance could be had by replacing
3179 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3180
3181 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3182
3183 27. Possessive quantifiers such as a++ were previously implemented by turning
3184 them into atomic groups such as ($>a+). Now they have their own opcodes,
3185 which improves performance. This includes the automatically created ones
3186 from 23 above.
3187
3188 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3189 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3190 the first matched character to be a colon. This applied both to named and
3191 numbered groups.
3192
3193 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3194
3195 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3196 building PCRE on Windows easier, so I have included it in the distribution.
3197
3198 31. There is now a check in pcretest against a ridiculously large number being
3199 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3200 loop, the loop is abandoned.
3201
3202 32. Forward references to subpatterns in conditions such as (?(2)...) where
3203 subpattern 2 is defined later cause pcre_compile() to search forwards in
3204 the pattern for the relevant set of parentheses. This search went wrong
3205 when there were unescaped parentheses in a character class, parentheses
3206 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3207
3208 33. "Subroutine" calls and backreferences were previously restricted to
3209 referencing subpatterns earlier in the regex. This restriction has now
3210 been removed.
3211
3212 34. Added a number of extra features that are going to be in Perl 5.10. On the
3213 whole, these are just syntactic alternatives for features that PCRE had
3214 previously implemented using the Python syntax or my own invention. The
3215 other formats are all retained for compatibility.
3216
3217 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3218 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3219 also .NET compatible.
3220
3221 (b) A recursion or subroutine call to a named group can now be defined as
3222 (?&name) as well as (?P>name).
3223
3224 (c) A backreference to a named group can now be defined as \k<name> or
3225 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3226 5.10, are also .NET compatible.
3227
3228 (d) A conditional reference to a named group can now use the syntax
3229 (?(<name>) or (?('name') as well as (?(name).
3230
3231 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3232 groups (named and numbered) that are never evaluated inline, but can be
3233 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3234 is always false. There may be only one alternative in such a group.
3235
3236 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3237 as the simple (?(R). The condition is true only if the most recent
3238 recursion is that of the given number or name. It does not search out
3239 through the entire recursion stack.
3240
3241 (g) The escape \gN or \g{N} has been added, where N is a positive or
3242 negative number, specifying an absolute or relative reference.
3243
3244 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3245 some "unreachable code" warnings.
3246
3247 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3248 things, this adds five new scripts.
3249
3250 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3251 There were also incompatibilities regarding the handling of \Q..\E inside
3252 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3253 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3254
3255 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3256 matches an empty string, and forcibly breaks the loop. There were bugs in
3257 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3258 against aaaa the result was just "a" rather than "aaaa", for example. Two
3259 separate and independent bugs (that affected different cases) have been
3260 fixed.
3261
3262 39. Refactored the code to abolish the use of different opcodes for small
3263 capturing bracket numbers. This is a tidy that I avoided doing when I
3264 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3265 The new approach is not only tidier, it makes it possible to reduce the
3266 memory needed to fix the previous bug (38).
3267
3268 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3269 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3270 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3271 mode.
3272
3273 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3274 report.
3275
3276 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3277 copy construction and assignment in the C++ wrapper.
3278
3279 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3280 couple of bugs that could have given wrong results in the "--newline=crlf"
3281 case.
3282
3283 44. Added a number of casts and did some reorganization of signed/unsigned int
3284 variables following suggestions from Dair Grant. Also renamed the variable
3285 "this" as "item" because it is a C++ keyword.
3286
3287 45. Arranged for dftables to add
3288
3289 #include "pcre_internal.h"
3290
3291 to pcre_chartables.c because without it, gcc 4.x may remove the array
3292 definition from the final binary if PCRE is built into a static library and
3293 dead code stripping is activated.
3294
3295 46. For an unanchored pattern, if a match attempt fails at the start of a
3296 newline sequence, and the newline setting is CRLF or ANY, and the next two
3297 characters are CRLF, advance by two characters instead of one.
3298
3299
3300 Version 6.7 04-Jul-06
3301 ---------------------
3302
3303 1. In order to handle tests when input lines are enormously long, pcretest has
3304 been re-factored so that it automatically extends its buffers when
3305 necessary. The code is crude, but this _is_ just a test program. The
3306 default size has been increased from 32K to 50K.
3307
3308 2. The code in pcre_study() was using the value of the re argument before
3309 testing it for NULL. (Of course, in any sensible call of the function, it
3310 won't be NULL.)
3311
3312 3. The memmove() emulation function in pcre_internal.h, which is used on
3313 systems that lack both memmove() and bcopy() - that is, hardly ever -
3314 was missing a "static" storage class specifier.
3315
3316 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3317 containing an extended class (one that cannot be represented by a bitmap
3318 because it contains high-valued characters or Unicode property items, e.g.
3319 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3320 pattern, but PCRE should not loop if you do not (it no longer does).
3321 [Detail: two cases were found: (a) a repeated subpattern containing an
3322 extended class; (b) a recursive reference to a subpattern that followed a
3323 previous extended class. It wasn't skipping over the extended class
3324 correctly when UTF-8 mode was not set.]
3325
3326 5. A negated single-character class was not being recognized as fixed-length
3327 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3328 compile error "lookbehind assertion is not fixed length".
3329
3330 6. The RunPerlTest auxiliary script was showing an unexpected difference
3331 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3332 write a Perl script that can interpret lines of an input file either as
3333 byte characters or as UTF-8, which is what "perltest" was being required to
3334 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3335 can't do is switch easily at run time between having the "use utf8;" pragma
3336 or not. In the end, I fudged it by using the RunPerlTest script to insert
3337 "use utf8;" explicitly for the UTF-8 tests.
3338
3339 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3340 the end of the subject string, contrary to the documentation and to what
3341 Perl does. This was true of both matching functions. Now it matches only at
3342 the start of the subject and immediately after *internal* newlines.
3343
3344 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3345 a pointer to an int instead of a pointer to an unsigned long int. This
3346 caused problems on 64-bit systems.
3347
3348 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3349 instance of the 'standard' template library not being so standard".
3350
3351 10. There was no check on the number of named subpatterns nor the maximum
3352 length of a subpattern name. The product of these values is used to compute
3353 the size of the memory block for a compiled pattern. By supplying a very
3354 long subpattern name and a large number of named subpatterns, the size
3355 computation could be caused to overflow. This is now prevented by limiting
3356 the length of names to 32 characters, and the number of named subpatterns
3357 to 10,000.
3358
3359 11. Subpatterns that are repeated with specific counts have to be replicated in
3360 the compiled pattern. The size of memory for this was computed from the
3361 length of the subpattern and the repeat count. The latter is limited to
3362 65535, but there was no limit on the former, meaning that integer overflow
3363 could in principle occur. The compiled length of a repeated subpattern is
3364 now limited to 30,000 bytes in order to prevent this.
3365
3366 12. Added the optional facility to have named substrings with the same name.
3367
3368 13. Added the ability to use a named substring as a condition, using the
3369 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3370 are numbers (not recommended). Forward references are permitted.
3371
3372 14. Added forward references in named backreferences (if you see what I mean).
3373
3374 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3375 pattern could run off the end of the subject. For example, the pattern
3376 "(?s)(.{1,5})"8 did this with the subject "ab".
3377
3378 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3379 PCRE_CASELESS was set when matching characters that were quantified with ?
3380 or *.
3381
3382 17. A character class other than a single negated character that had a minimum
3383 but no maximum quantifier - for example [ab]{6,} - was not handled
3384 correctly by pce_dfa_exec(). It would match only one character.
3385
3386 18. A valid (though odd) pattern that looked like a POSIX character
3387 class but used an invalid character after [ (for example [[,abc,]]) caused
3388 pcre_compile() to give the error "Failed: internal error: code overflow" or
3389 in some cases to crash with a glibc free() error. This could even happen if
3390 the pattern terminated after [[ but there just happened to be a sequence of
3391 letters, a binary zero, and a closing ] in the memory that followed.
3392
3393 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3394 over the years. Originally (before any Unicode support), just the bottom 8
3395 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3396 output from "man perlunicode" includes this:
3397
3398 The regular expression compiler produces polymorphic opcodes. That
3399 is, the pattern adapts to the data and automatically switches to
3400 the Unicode character scheme when presented with Unicode data--or
3401 instead uses a traditional byte scheme when presented with byte
3402 data.
3403
3404 Sadly, a wide octal escape does not cause a switch, and in a string with
3405 no other multibyte characters, these octal escapes are treated as before.
3406 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3407 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3408 Unicode string.
3409
3410 I have not perpetrated such confusion in PCRE. Up till now, it took just
3411 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3412 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3413 translate to the appropriate multibyte character.
3414
3415 29. Applied some refactoring to reduce the number of warnings from Microsoft
3416 and Borland compilers. This has included removing the fudge introduced
3417 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3418 a warning about an unused variable.
3419
3420 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3421 characters since release 4.0, because Perl (from release 5.004) does not.
3422 [Or at least, is documented not to: some releases seem to be in conflict
3423 with the documentation.] However, when a pattern was studied with
3424 pcre_study() and all its branches started with \s, PCRE still included VT
3425 as a possible starting character. Of course, this did no harm; it just
3426 caused an unnecessary match attempt.
3427
3428 22. Removed a now-redundant internal flag bit that recorded the fact that case
3429 dependency changed within the pattern. This was once needed for "required
3430 byte" processing, but is no longer used. This recovers a now-scarce options
3431 bit. Also moved the least significant internal flag bit to the most-
3432 significant bit of the word, which was not previously used (hangover from
3433 the days when it was an int rather than a uint) to free up another bit for
3434 the future.
3435
3436 23. Added support for CRLF line endings as well as CR and LF. As well as the
3437 default being selectable at build time, it can now be changed at runtime
3438 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3439 specify that it is scanning data with non-default line endings.
3440
3441 24. Changed the definition of CXXLINK to make it agree with the definition of
3442 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3443
3444 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3445 recursions. This makes a big different to stack usage for some patterns.
3446
3447 26. If a subpattern containing a named recursion or subroutine reference such
3448 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3449 the space required for the compiled pattern went wrong and gave too small a
3450 value. Depending on the environment, this could lead to "Failed: internal
3451 error: code overflow at offset 49" or "glibc detected double free or
3452 corruption" errors.
3453
3454 27. Applied patches from Google (a) to support the new newline modes and (b) to
3455 advance over multibyte UTF-8 characters in GlobalReplace.
3456
3457 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3458 difference for some implementation of PCRE in some Windows version.
3459
3460 29. Added some extra testing facilities to pcretest:
3461
3462 \q<number> in a data line sets the "match limit" value
3463 \Q<number> in a data line sets the "match recursion limt" value
3464 -S <number> sets the stack size, where <number> is in megabytes
3465
3466 The -S option isn't available for Windows.
3467
3468
3469 Version 6.6 06-Feb-06
3470 ---------------------
3471
3472 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3473 in pcreposix.h. I have copied the definition from pcre.h.
3474
3475 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3476 because pcre.h is no longer a built file.
3477
3478 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3479 not normally included in the compiled code.
3480
3481
3482 Version 6.5 01-Feb-06
3483 ---------------------
3484
3485 1. When using the partial match feature with pcre_dfa_exec(), it was not
3486 anchoring the second and subsequent partial matches at the new starting
3487 point. This could lead to incorrect results. For example, with the pattern
3488 /1234/, partially matching against "123" and then "a4" gave a match.
3489
3490 2. Changes to pcregrep:
3491
3492 (a) All non-match returns from pcre_exec() were being treated as failures
3493 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3494 error message is output. Some extra information is given for the
3495 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3496 probably the only errors that are likely to be caused by users (by
3497 specifying a regex that has nested indefinite repeats, for instance).
3498 If there are more than 20 of these errors, pcregrep is abandoned.
3499
3500 (b) A binary zero was treated as data while matching, but terminated the
3501 output line if it was written out. This has been fixed: binary zeroes
3502 are now no different to any other data bytes.
3503
3504 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3505 used to set a locale for matching. The --locale=xxxx long option has
3506 been added (no short equivalent) to specify a locale explicitly on the
3507 pcregrep command, overriding the environment variables.
3508
3509 (d) When -B was used with -n, some line numbers in the output were one less
3510 than they should have been.
3511
3512 (e) Added the -o (--only-matching) option.
3513
3514 (f) If -A or -C was used with -c (count only), some lines of context were
3515 accidentally printed for the final match.
3516
3517 (g) Added the -H (--with-filename) option.
3518
3519 (h) The combination of options -rh failed to suppress file names for files
3520 that were found from directory arguments.
3521
3522 (i) Added the -D (--devices) and -d (--directories) options.
3523
3524 (j) Added the -F (--fixed-strings) option.
3525
3526 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3527
3528 (l) Added the --colo(u)r option.
3529
3530 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3531 is not present by default.
3532
3533 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3534 items such as (?R) or (?1), when the recursion could match a number of
3535 alternatives. If it matched one of the alternatives, but subsequently,
3536 outside the recursion, there was a failure, the code tried to back up into
3537 the recursion. However, because of the way PCRE is implemented, this is not
3538 possible, and the result was an incorrect result from the match.
3539
3540 In order to prevent this happening, the specification of recursion has
3541 been changed so that all such subpatterns are automatically treated as
3542 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3543
3544 4. I had overlooked the fact that, in some locales, there are characters for
3545 which isalpha() is true but neither isupper() nor islower() are true. In
3546 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3547 and ordfeminine) are like this. This affected the treatment of \w and \W
3548 when they appeared in character classes, but not when they appeared outside
3549 a character class. The bit map for "word" characters is now created
3550 separately from the results of isalnum() instead of just taking it from the
3551 upper, lower, and digit maps. (Plus the underscore character, of course.)
3552
3553 5. The above bug also affected the handling of POSIX character classes such as
3554 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3555 permanent tables. Instead, the bit maps for such a class were previously
3556 created as the appropriate unions of the upper, lower, and digit bitmaps.
3557 Now they are created by subtraction from the [[:word:]] class, which has
3558 its own bitmap.
3559
3560 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3561 It is created by subtracting the vertical space characters (\x09, \x0a,
3562 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3563 subtraction was done in the overall bitmap for a character class, meaning
3564 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3565 be recognized. This bug has been fixed.
3566
3567 7. Patches from the folks at Google:
3568
3569 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3570 real life, but is still worth protecting against".
3571
3572 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3573 regular expressions".
3574
3575 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3576 have it.
3577
3578 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3579 "configure" and the latter not, in order to fix a problem somebody had
3580 with compiling the Arg class on HP-UX.
3581
3582 (e) Improve the error-handling of the C++ wrapper a little bit.
3583
3584 (f) New tests for checking recursion limiting.
3585
3586 8. The pcre_memmove() function, which is used only if the environment does not
3587 have a standard memmove() function (and is therefore rarely compiled),
3588 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3589 returning a result (though PCRE never actually uses the result).
3590
3591 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3592 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3593 returned instead of calling malloc() with an overflowing number that would
3594 most likely cause subsequent chaos.
3595
3596 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3597
3598 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3599 with this option is matched, the nmatch and pmatch options of regexec() are
3600 ignored.
3601
3602 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3603 provided in case anyone wants to the the POSIX interface with UTF-8
3604 strings.
3605
3606 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3607 C++ linking (needed for some HP-UX environments).
3608
3609 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3610 (unused parameter) and in the pcre_printint() function (omitted "default"
3611 switch label when the default is to do nothing).
3612
3613 15. Added some code to make it possible, when PCRE is compiled as a C++
3614 library, to replace subject pointers for pcre_exec() with a smart pointer
3615 class, thus making it possible to process discontinuous strings.
3616
3617 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3618 much the same function. They were added by different people who were trying
3619 to make PCRE easy to compile on non-Unix systems. It has been suggested
3620 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3621 for compiling on Windows systems. I have therefore replaced it with
3622 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3623 defaults to "extern" for C or "extern C" for C++, which works fine on
3624 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3625 SCOPE with something explicit in config.h. In addition:
3626
3627 (a) pcreposix.h still had just "extern" instead of either of these macros;
3628 I have replaced it with PCRE_DATA_SCOPE.
3629
3630 (b) Functions such as _pcre_xclass(), which are internal to the library,
3631 but external in the C sense, all had PCRE_EXPORT in their definitions.
3632 This is apparently wrong for the Windows case, so I have removed it.
3633 (It makes no difference on Unix-like systems.)
3634
3635 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3636 of recursive calls to match(). This is different to MATCH_LIMIT because
3637 that limits the total number of calls to match(), not all of which increase
3638 the depth of recursion. Limiting the recursion depth limits the amount of
3639 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3640 when PCRE is compiled, and changed at run time. A patch from Google adds
3641 this functionality to the C++ interface.
3642
3643 18. Changes to the handling of Unicode character properties:
3644
3645 (a) Updated the table to Unicode 4.1.0.
3646
3647 (b) Recognize characters that are not in the table as "Cn" (undefined).
3648
3649 (c) I revised the way the table is implemented to a much improved format
3650 which includes recognition of ranges. It now supports the ranges that
3651 are defined in UnicodeData.txt, and it also amalgamates other
3652 characters into ranges. This has reduced the number of entries in the
3653 table from around 16,000 to around 3,000, thus reducing its size
3654 considerably. I realized I did not need to use a tree structure after
3655 all - a binary chop search is just as efficient. Having reduced the
3656 number of entries, I extended their size from 6 bytes to 8 bytes to
3657 allow for more data.
3658
3659 (d) Added support for Unicode script names via properties such as \p{Han}.
3660
3661 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3662 matching that character.
3663
3664 20. When matching a repeated Unicode property with a minimum greater than zero,
3665 (for example \pL{2,}), PCRE could look past the end of the subject if it
3666 reached it while seeking the minimum number of characters. This could
3667 happen only if some of the characters were more than one byte long, because
3668 there is a check for at least the minimum number of bytes.
3669
3670 21. Refactored the implementation of \p and \P so as to be more general, to
3671 allow for more different types of property in future. This has changed the
3672 compiled form incompatibly. Anybody with saved compiled patterns that use
3673 \p or \P will have to recompile them.
3674
3675 22. Added "Any" and "L&" to the supported property types.
3676
3677 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3678 but give a compile time error if the value is greater than 0xff.
3679
3680 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3681 accidentally not being installed or uninstalled.
3682
3683 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3684 made were to insert the current release number. This seemed silly, because
3685 it made things harder for people building PCRE on systems that don't run
3686 "configure". I have turned pcre.h into a distributed file, no longer built
3687 by "configure", with the version identification directly included. There is
3688 no longer a pcre.h.in file.
3689
3690 However, this change necessitated a change to the pcre-config script as
3691 well. It is built from pcre-config.in, and one of the substitutions was the
3692 release number. I have updated configure.ac so that ./configure now finds
3693 the release number by grepping pcre.h.
3694
3695 26. Added the ability to run the tests under valgrind.
3696
3697
3698 Version 6.4 05-Sep-05
3699 ---------------------
3700
3701 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3702 "--" to be printed when multiple files were scanned, even when none of the
3703 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3704 consider it to be a bug, and have restored the previous behaviour.
3705
3706 2. A couple of code tidies to get rid of compiler warnings.
3707
3708 3. The pcretest program used to cheat by referring to symbols in the library
3709 whose names begin with _pcre_. These are internal symbols that are not
3710 really supposed to be visible externally, and in some environments it is
3711 possible to suppress them. The cheating is now confined to including
3712 certain files from the library's source, which is a bit cleaner.
3713
3714 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3715 file's purpose clearer.
3716
3717 5. Reorganized pcre_ucp_findchar().
3718
3719
3720 Version 6.3 15-Aug-05
3721 ---------------------
3722
3723 1. The file libpcre.pc.in did not have general read permission in the tarball.
3724
3725 2. There were some problems when building without C++ support:
3726
3727 (a) If C++ support was not built, "make install" and "make test" still
3728 tried to test it.
3729
3730 (b) There were problems when the value of CXX was explicitly set. Some
3731 changes have been made to try to fix these, and ...
3732
3733 (c) --disable-cpp can now be used to explicitly disable C++ support.
3734
3735 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3736 backslash in a target when C++ was disabled. This confuses some
3737 versions of "make", apparently. Using an intermediate variable solves
3738 this. (Same for CPP_LOBJ.)
3739
3740 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3741 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3742 necessary on certain architectures.
3743
3744 4. Added a setting of -export-symbols-regex to the link command to remove
3745 those symbols that are exported in the C sense, but actually are local
3746 within the library, and not documented. Their names all begin with
3747 "_pcre_". This is not a perfect job, because (a) we have to except some
3748 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3749 available (and never for static libraries). I have made a note to try to
3750 find a way round (a) in the future.
3751
3752
3753 Version 6.2 01-Aug-05
3754 ---------------------
3755
3756 1. There was no test for integer overflow of quantifier values. A construction
3757 such as {1111111111111111} would give undefined results. What is worse, if
3758 a minimum quantifier for a parenthesized subpattern overflowed and became
3759 negative, the calculation of the memory size went wrong. This could have
3760 led to memory overwriting.
3761
3762 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3763
3764 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3765 operating environments where this matters.
3766
3767 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3768 PCRE options from within the C++ wrapper.
3769
3770 5. Named capturing subpatterns were not being correctly counted when a pattern
3771 was compiled. This caused two problems: (a) If there were more than 100
3772 such subpatterns, the calculation of the memory needed for the whole
3773 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3774 back references of the form \12, where the number was greater than 9, were
3775 not recognized as back references, even though there were sufficient
3776 previous subpatterns.
3777
3778 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3779 versions of gcc, e.g. 2.95.4.
3780
3781
3782 Version 6.1 21-Jun-05
3783 ---------------------
3784
3785 1. There was one reference to the variable "posix" in pcretest.c that was not
3786 surrounded by "#if !defined NOPOSIX".
3787
3788 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3789 the cross-check on the old pcre_info() function, for the benefit of the
3790 cut-down version of PCRE that is currently imported into Exim.
3791
3792 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3793 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3794 patterns that start (?i) but I don't think that matters. The use of (?i) is
3795 just an example; this all applies to the other options as well.
3796
3797 4. Since libtool seems to echo the compile commands it is issuing, the output
3798 from "make" can be reduced a bit by putting "@" in front of each libtool
3799 compile command.
3800
3801 5. Patch from the folks at Google for configure.in to be a bit more thorough
3802 in checking for a suitable C++ installation before trying to compile the
3803 C++ stuff. This should fix a reported problem when a compiler was present,
3804 but no suitable headers.
3805
3806 6. The man pages all had just "PCRE" as their title. I have changed them to
3807 be the relevant file name. I have also arranged that these names are
3808 retained in the file doc/pcre.txt, which is a concatenation in text format
3809 of all the man pages except the little individual ones for each function.
3810
3811 7. The NON-UNIX-USE file had not been updated for the different set of source
3812 files that come with release 6. I also added a few comments about the C++
3813 wrapper.
3814
3815
3816 Version 6.0 07-Jun-05
3817 ---------------------
3818
3819 1. Some minor internal re-organization to help with my DFA experiments.
3820
3821 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3822 didn't matter for the library itself when fully configured, but did matter
3823 when compiling without UCP support, or within Exim, where the ucp files are
3824 not imported.
3825
3826 3. Refactoring of the library code to split up the various functions into
3827 different source modules. The addition of the new DFA matching code (see
3828 below) to a single monolithic source would have made it really too
3829 unwieldy, quite apart from causing all the code to be include in a
3830 statically linked application, when only some functions are used. This is
3831 relevant even without the DFA addition now that patterns can be compiled in
3832 one application and matched in another.
3833
3834 The downside of splitting up is that there have to be some external
3835 functions and data tables that are used internally in different modules of
3836 the library but which are not part of the API. These have all had their
3837 names changed to start with "_pcre_" so that they are unlikely to clash
3838 with other external names.
3839
3840 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3841 a different (DFA) algorithm. Although it is slower than the original
3842 function, it does have some advantages for certain types of matching
3843 problem.
3844
3845 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3846 including restarting after a partial match.
3847
3848 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3849 defined when compiling for Windows was sent to me. I have put it into the
3850 code, though I have no means of testing or verifying it.
3851
3852 7. Added the pcre_refcount() auxiliary function.
3853
3854 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3855 match before or at the first newline in the subject string. In pcretest,
3856 the /f option on a pattern can be used to set this.
3857
3858 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3859 would behave wrongly. This has been present in PCRE since release 4.0.
3860
3861 10. A number of changes to the pcregrep command:
3862
3863 (a) Refactored how -x works; insert ^(...)$ instead of setting
3864 PCRE_ANCHORED and checking the length, in preparation for adding
3865 something similar for -w.
3866
3867 (b) Added the -w (match as a word) option.
3868
3869 (c) Refactored the way lines are read and buffered so as to have more
3870 than one at a time available.
3871
3872 (d) Implemented a pcregrep test script.
3873
3874 (e) Added the -M (multiline match) option. This allows patterns to match
3875 over several lines of the subject. The buffering ensures that at least
3876 8K, or the rest of the document (whichever is the shorter) is available
3877 for matching (and similarly the previous 8K for lookbehind assertions).
3878
3879 (f) Changed the --help output so that it now says
3880
3881 -w, --word-regex(p)
3882
3883 instead of two lines, one with "regex" and the other with "regexp"
3884 because that confused at least one person since the short forms are the
3885 same. (This required a bit of code, as the output is generated
3886 automatically from a table. It wasn't just a text change.)
3887
3888 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3889 option but starts with a hyphen. Could be a pattern or a path name
3890 starting with a hyphen, for instance.
3891
3892 (h) "-" can be given as a file name to represent stdin.
3893
3894 (i) When file names are being printed, "(standard input)" is used for
3895 the standard input, for compatibility with GNU grep. Previously
3896 "<stdin>" was used.
3897
3898 (j) The option --label=xxx can be used to supply a name to be used for
3899 stdin when file names are being printed. There is no short form.
3900
3901 (k) Re-factored the options decoding logic because we are going to add
3902 two more options that take data. Such options can now be given in four
3903 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3904
3905 (l) Added the -A, -B, and -C options for requesting that lines of context
3906 around matches be printed.
3907
3908 (m) Added the -L option to print the names of files that do not contain
3909 any matching lines, that is, the complement of -l.
3910
3911 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3912 continue to scan other files.
3913
3914 (o) The -s option was incorrectly implemented. For compatibility with other
3915 greps, it now suppresses the error message for a non-existent or non-
3916 accessible file (but not the return code). There is a new option called
3917 -q that suppresses the output of matching lines, which was what -s was
3918 previously doing.
3919
3920 (p) Added --include and --exclude options to specify files for inclusion
3921 and exclusion when recursing.
3922
3923 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3924 Hopefully, it now does.
3925
3926 12. Missing cast in pcre_study().
3927
3928 13. Added an "uninstall" target to the makefile.
3929
3930 14. Replaced "extern" in the function prototypes in Makefile.in with
3931 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3932 world, but is set differently for Windows.
3933
3934 15. Added a second compiling function called pcre_compile2(). The only
3935 difference is that it has an extra argument, which is a pointer to an
3936 integer error code. When there is a compile-time failure, this is set
3937 non-zero, in addition to the error test pointer being set to point to an
3938 error message. The new argument may be NULL if no error number is required
3939 (but then you may as well call pcre_compile(), which is now just a
3940 wrapper). This facility is provided because some applications need a
3941 numeric error indication, but it has also enabled me to tidy up the way
3942 compile-time errors are handled in the POSIX wrapper.
3943
3944 16. Added VPATH=.libs to the makefile; this should help when building with one
3945 prefix path and installing with another. (Or so I'm told by someone who
3946 knows more about this stuff than I do.)
3947
3948 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3949 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3950 match everything, including newlines. This is not POSIX-compatible, but
3951 somebody wanted the feature. From pcretest it can be activated by using
3952 both the P and the s flags.
3953
3954 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3955
3956 19. libpcre.pc was being incorrectly installed as executable.
3957
3958 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3959 it now also looks for '\r' so that it will work unmodified on Windows.
3960
3961 21. Added Google's contributed C++ wrapper to the distribution.
3962
3963 22. Added some untidy missing memory free() calls in pcretest, to keep
3964 Electric Fence happy when testing.
3965
3966
3967
3968 Version 5.0 13-Sep-04
3969 ---------------------
3970
3971 1. Internal change: literal characters are no longer packed up into items
3972 containing multiple characters in a single byte-string. Each character
3973 is now matched using a separate opcode. However, there may be more than one
3974 byte in the character in UTF-8 mode.
3975
3976 2. The pcre_callout_block structure has two new fields: pattern_position and
3977 next_item_length. These contain the offset in the pattern to the next match
3978 item, and its length, respectively.
3979
3980 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3981 insertion of callouts before each pattern item. Added the /C option to
3982 pcretest to make use of this.
3983
3984 4. On the advice of a Windows user, the lines
3985
3986 #if defined(_WIN32) || defined(WIN32)
3987 _setmode( _fileno( stdout ), 0x8000 );
3988 #endif /* defined(_WIN32) || defined(WIN32) */
3989
3990 have been added to the source of pcretest. This apparently does useful
3991 magic in relation to line terminators.
3992
3993 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3994 for the benefit of those environments where the "b" makes a difference.
3995
3996 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3997 to know about it. I have put a hack into configure.in that adds in code
3998 to set GCC=yes if CC=icc. This seems to end up at a point in the
3999 generated configure script that is early enough to affect the setting of
4000 compiler options, which is what is needed, but I have no means of testing
4001 whether it really works. (The user who reported this had patched the
4002 generated configure script, which of course I cannot do.)
4003
4004 LATER: After change 22 below (new libtool files), the configure script
4005 seems to know about icc (and also ecc). Therefore, I have commented out
4006 this hack in configure.in.
4007
4008 7. Added support for pkg-config (2 patches were sent in).
4009
4010 8. Negated POSIX character classes that used a combination of internal tables
4011 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4012 [[:^ascii]]. Typically, they would match almost any characters. The other
4013 POSIX classes were not broken in this way.
4014
4015 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4016 to find the match, as PCRE was deluded into thinking that the match had to
4017 start at the start point or following a newline. The same bug applied to
4018 patterns with negative forward assertions or any backward assertions
4019 preceding ".*" at the start, unless the pattern required a fixed first
4020 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4021
4022 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4023 starting at the last subject character, bytes beyond the end of the subject
4024 string were read.
4025
4026 11. Renamed the variable "class" as "classbits" to make life easier for C++
4027 users. (Previously there was a macro definition, but it apparently wasn't
4028 enough.)
4029
4030 12. Added the new field "tables" to the extra data so that tables can be passed
4031 in at exec time, or the internal tables can be re-selected. This allows
4032 a compiled regex to be saved and re-used at a later time by a different
4033 program that might have everything at different addresses.
4034
4035 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4036 -R library as well as a -L library.
4037
4038 14. The debugging options of pcretest (-d on the command line or D on a
4039 pattern) showed incorrect output for anything following an extended class
4040 that contained multibyte characters and which was followed by a quantifier.
4041
4042 15. Added optional support for general category Unicode character properties
4043 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4044 support. It adds about 90K to the size of the library. The meanings of the
4045 inbuilt class escapes such as \d and \s have NOT been changed.
4046
4047 16. Updated pcredemo.c to include calls to free() to release the memory for the
4048 compiled pattern.
4049
4050 17. The generated file chartables.c was being created in the source directory
4051 instead of in the building directory. This caused the build to fail if the
4052 source directory was different from the building directory, and was
4053 read-only.
4054
4055 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4056 file. No doubt somebody will tell me if they don't make sense... Also added
4057 Dan Mooney's comments about building on OpenVMS.
4058
4059 19. Added support for partial matching via the PCRE_PARTIAL option for
4060 pcre_exec() and the \P data escape in pcretest.
4061
4062 20. Extended pcretest with 3 new pattern features:
4063
4064 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4065 write the compiled pattern to the file whose name is "rest-of-line".
4066 This is a straight binary dump of the data, with the saved pointer to
4067 the character tables forced to be NULL. The study data, if any, is
4068 written too. After writing, pcretest reads a new pattern.
4069
4070 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4071 compiled pattern from the given file. There must not be any
4072 occurrences of "<" in the file name (pretty unlikely); if there are,
4073 pcretest will instead treat the initial "<" as a pattern delimiter.
4074 After reading in the pattern, pcretest goes on to read data lines as
4075 usual.
4076
4077 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4078 and 16-bit fields in a compiled pattern, to simulate a pattern that
4079 was compiled on a host of opposite endianness.
4080
4081 21. The pcre-exec() function can now cope with patterns that were compiled on
4082 hosts of opposite endianness, with this restriction:
4083
4084 As for any compiled expression that is saved and used later, the tables
4085 pointer field cannot be preserved; the extra_data field in the arguments
4086 to pcre_exec() should be used to pass in a tables address if a value
4087 other than the default internal tables were used at compile time.
4088
4089 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4090 now diagnosed as an error. Previously, most of the time, a negative number
4091 would have been treated as zero, but if in addition "ovector" was passed as
4092 NULL, a crash could occur.
4093
4094 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4095 new versions from the libtool 1.5 distribution (the last one is a copy of
4096 a file called libtool.m4). This seems to have fixed the need to patch
4097 "configure" to support Darwin 1.3 (which I used to do). However, I still
4098 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4099 workstation).
4100
4101 24. Changed the PCRE licence to be the more standard "BSD" licence.
4102
4103
4104 Version 4.5 01-Dec-03
4105 ---------------------
4106
4107 1. There has been some re-arrangement of the code for the match() function so
4108 that it can be compiled in a version that does not call itself recursively.
4109 Instead, it keeps those local variables that need separate instances for
4110 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4111 needs to "recurse". Keeping track of where control must go is done by means
4112 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4113 hide most of the details from the main code, and operates only if
4114 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4115 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4116 operating.
4117
4118 To make it easier for callers to provide specially tailored get/free
4119 functions for this usage, two new functions, pcre_stack_malloc, and
4120 pcre_stack_free, are used. They are always called in strict stacking order,
4121 and the size of block requested is always the same.
4122
4123 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4124 PCRE has been compiled to use the stack or the heap for recursion. The
4125 -C option of pcretest uses this to show which version is compiled.
4126
4127 A new data escape \S, is added to pcretest; it causes the amounts of store
4128 obtained and freed by both kinds of malloc/free at match time to be added
4129 to the output.
4130
4131 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4132 what's available on my current Linux desktop machine.
4133
4134 3. When matching a UTF-8 string, the test for a valid string at the start has
4135 been extended. If start_offset is not zero, PCRE now checks that it points
4136 to a byte that is the start of a UTF-8 character. If not, it returns
4137 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4138 this is necessary because there may be backward assertions in the pattern.
4139 When matching the same subject several times, it may save resources to use
4140 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4141
4142 4. The code for checking the validity of UTF-8 strings has been tightened so
4143 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4144 containing "overlong sequences".
4145
4146 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4147 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4148 should have been "&", but it just so happened that all the cases this let
4149 through by mistake were picked up later in the function.
4150
4151 6. I had used a variable called "isblank" - this is a C99 function, causing
4152 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4153
4154 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4155 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4156 so the version is shown; (c) stop "make test" echoing "./RunTest".
4157
4158 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4159
4160 9. The prototype for memmove() for systems that don't have it was using
4161 size_t, but the inclusion of the header that defines size_t was later. I've
4162 moved the #includes for the C headers earlier to avoid this.
4163
4164 10. Added some adjustments to the code to make it easier to compiler on certain
4165 special systems:
4166
4167 (a) Some "const" qualifiers were missing.
4168 (b) Added the macro EXPORT before all exported functions; by default this
4169 is defined to be empty.
4170 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4171 that it reads its output file name as an argument instead of writing
4172 to the standard output and assuming this can be redirected.
4173
4174 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4175 class containing characters with values greater than 255, PCRE compilation
4176 went into a loop.
4177
4178 12. A recursive reference to a subpattern that was within another subpattern
4179 that had a minimum quantifier of zero caused PCRE to crash. For example,
4180 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4181 recursion. If the recursively-called subpattern itself had a zero repeat,
4182 that was OK.
4183
4184 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4185 buffer into which it was copied (for escape processing) was still set at
4186 1024, so long lines caused crashes.
4187
4188 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4189 "internal error: code overflow...". This applied to any character class
4190 that was followed by a possessive quantifier.
4191
4192 15. Modified the Makefile to add libpcre.la as a prerequisite for
4193 libpcreposix.la because I was told this is needed for a parallel build to
4194 work.
4195
4196 16. If a pattern that contained .* following optional items at the start was
4197 studied, the wrong optimizing data was generated, leading to matching
4198 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4199 matching string must start with a or b or c. The correct conclusion for
4200 this pattern is that a match can start with any character.
4201
4202
4203 Version 4.4 13-Aug-03
4204 ---------------------
4205
4206 1. In UTF-8 mode, a character class containing characters with values between
4207 127 and 255 was not handled correctly if the compiled pattern was studied.
4208 In fixing this, I have also improved the studying algorithm for such
4209 classes (slightly).
4210
4211 2. Three internal functions had redundant arguments passed to them. Removal
4212 might give a very teeny performance improvement.
4213
4214 3. Documentation bug: the value of the capture_top field in a callout is *one
4215 more than* the number of the hightest numbered captured substring.
4216
4217 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4218 in incorrectly linking with a previously installed version. They now link
4219 explicitly with libpcre.la.
4220
4221 5. configure.in no longer needs to recognize Cygwin specially.
4222
4223 6. A problem in pcre.in for Windows platforms is fixed.
4224
4225 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4226 pcretest, it used to include the size of the study block as part of its
4227 output. Unfortunately, the structure contains a field that has a different
4228 size on different hardware architectures. This meant that the tests that
4229 showed this size failed. As the block is currently always of a fixed size,
4230 this information isn't actually particularly useful in pcretest output, so
4231 I have just removed it.
4232
4233 8. Three pre-processor statements accidentally did not start in column 1.
4234 Sadly, there are *still* compilers around that complain, even though
4235 standard C has not required this for well over a decade. Sigh.
4236
4237 9. In pcretest, the code for checking callouts passed small integers in the
4238 callout_data field, which is a void * field. However, some picky compilers
4239 complained about the casts involved for this on 64-bit systems. Now
4240 pcretest passes the address of the small integer instead, which should get
4241 rid of the warnings.
4242
4243 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4244 both compile and run time, and gives an error if an invalid UTF-8 sequence
4245 is found. There is a option for disabling this check in cases where the
4246 string is known to be correct and/or the maximum performance is wanted.
4247
4248 11. In response to a bug report, I changed one line in Makefile.in from
4249
4250 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4251 to
4252 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4253
4254 to look similar to other lines, but I have no way of telling whether this
4255 is the right thing to do, as I do not use Windows. No doubt I'll get told
4256 if it's wrong...
4257
4258
4259 Version 4.3 21-May-03
4260 ---------------------
4261
4262 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4263 Makefile.
4264
4265 2. Some refactoring to improve the quality of the code:
4266
4267 (i) The utf8_table... variables are now declared "const".
4268
4269 (ii) The code for \cx, which used the "case flipping" table to upper case
4270 lower case letters, now just substracts 32. This is ASCII-specific,
4271 but the whole concept of \cx is ASCII-specific, so it seems
4272 reasonable.
4273
4274 (iii) PCRE was using its character types table to recognize decimal and
4275 hexadecimal digits in the pattern. This is silly, because it handles
4276 only 0-9, a-f, and A-F, but the character types table is locale-
4277 specific, which means strange things might happen. A private
4278 table is now used for this - though it costs 256 bytes, a table is
4279 much faster than multiple explicit tests. Of course, the standard
4280 character types table is still used for matching digits in subject
4281 strings against \d.
4282
4283 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4284 ending in _t are). So I've renamed it as ESC_tee.
4285
4286 3. The first argument for regexec() in the POSIX wrapper should have been
4287 defined as "const".
4288
4289 4. Changed pcretest to use malloc() for its buffers so that they can be
4290 Electric Fenced for debugging.
4291
4292 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4293 to read one or more bytes before the start of the subject string. Often this
4294 had no effect on PCRE's behaviour, but in some circumstances it could
4295 provoke a segmentation fault.
4296
4297 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4298 to try to read one or more bytes before the start of the subject string.
4299
4300 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4301 UTF-8 support could misbehave in various ways if the subject string
4302 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4303 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4304 back over UTF-8 characters.)
4305
4306
4307 Version 4.2 14-Apr-03
4308 ---------------------
4309
4310 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4311
4312 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4313 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4314 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4315 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4316 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4317 and BUILD_EXEEXT
4318 Note: automatic setting of the BUILD variables is not yet working
4319 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4320 compile-time but not at link-time
4321 [LINK]: use for linking executables only
4322 make different versions for Windows and non-Windows
4323 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4324 libraries
4325 [LINK_FOR_BUILD]: new variable
4326 [OBJEXT]: use throughout
4327 [EXEEXT]: use throughout
4328 <winshared>: new target
4329 <wininstall>: new target
4330 <dftables.o>: use native compiler
4331 <dftables>: use native linker
4332 <install>: handle Windows platform correctly
4333 <clean>: ditto
4334 <check>: ditto
4335 copy DLL to top builddir before testing
4336
4337 As part of these changes, -no-undefined was removed again. This was reported
4338 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4339 in any case.
4340
4341 3. Some tidies to get rid of compiler warnings:
4342
4343 . In the match_data structure, match_limit was an unsigned long int, whereas
4344 match_call_count was an int. I've made them both unsigned long ints.
4345
4346 . In pcretest the fact that a const uschar * doesn't automatically cast to
4347 a void * provoked a warning.
4348
4349 . Turning on some more compiler warnings threw up some "shadow" variables
4350 and a few more missing casts.
4351
4352 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4353 option, a class that contained a single character with a value between 128
4354 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4355
4356 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4357 option, a class that contained several characters, but with at least one
4358 whose value was between 128 and 255 caused PCRE to crash.
4359
4360
4361 Version 4.1 12-Mar-03
4362 ---------------------
4363
4364 1. Compiling with gcc -pedantic found a couple of places where casts were
4365 needed, and a string in dftables.c that was longer than standard compilers are
4366 required to support.
4367
4368 2. Compiling with Sun's compiler found a few more places where the code could
4369 be tidied up in order to avoid warnings.
4370
4371 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4372 first of these names is deprecated in the latest Autoconf in favour of the name
4373 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4374 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4375 analogy I have changed it to CFLAGS_FOR_BUILD.
4376
4377 4. Added -no-undefined to the linking command in the Makefile, because this is
4378 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4379 linking step for the pcreposix library.
4380
4381 5. PCRE was failing to diagnose the case of two named groups with the same
4382 name.
4383
4384 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4385 literal character that is needed in the subject for a match, and scans along to
4386 ensure that it is present before embarking on the full matching process. This
4387 saves time in cases of nested unlimited repeats that are never going to match.
4388 Problem: the scan can take a lot of time if the subject is very long (e.g.
4389 megabytes), thus penalizing straightforward matches. It is now done only if the
4390 amount of subject to be scanned is less than 1000 bytes.
4391
4392 7. A lesser problem with the same optimization is that it was recording the
4393 first character of an anchored pattern as "needed", thus provoking a search
4394 right along the subject, even when the first match of the pattern was going to
4395 fail. The "needed" character is now not set for anchored patterns, unless it
4396 follows something in the pattern that is of non-fixed length. Thus, it still
4397 fulfils its original purpose of finding quick non-matches in cases of nested
4398 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4399
4400
4401 Version 4.0 17-Feb-03
4402 ---------------------
4403
4404 1. If a comment in an extended regex that started immediately after a meta-item
4405 extended to the end of string, PCRE compiled incorrect data. This could lead to
4406 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4407
4408 2. Moved to autoconf 2.53 and libtool 1.4.2.
4409
4410 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4411 the special perltest8 script is no longer needed - all the tests can be run
4412 from a single perltest script.
4413
4414 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4415 by \s. It has now been removed in PCRE. This means it isn't recognized as
4416 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4417 class [:space:] *does* include VT, thereby creating a mess.
4418
4419 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4420 space and tab.
4421
4422 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4423 its new features into the main test script, reducing the number of scripts.
4424
4425 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4426 were backward compatible, and made the (?i) apply to the whole pattern, as if
4427 /i were given. Now it behaves more logically, and applies the option setting
4428 only to what follows. PCRE has been changed to follow suit. However, if it
4429 finds options settings right at the start of the pattern, it extracts them into
4430 the global options, as before. Thus, they show up in the info data.
4431
4432 8. Added support for the \Q...\E escape sequence. Characters in between are
4433 treated as literals. This is slightly different from Perl in that $ and @ are
4434 also handled as literals inside the quotes. In Perl, they will cause variable
4435 interpolation. Note the following examples:
4436
4437 Pattern PCRE matches Perl matches
4438
4439 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4440 \Qabc\$xyz\E abc\$xyz abc\$xyz
4441 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4442
4443 For compatibility with Perl, \Q...\E sequences are recognized inside character
4444 classes as well as outside them.
4445
4446 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4447 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4448 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4449 signed/unsigned warnings.
4450
4451 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4452 option for pcretest, so I've replaced it by a simple function that does just
4453 that job.
4454
4455 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4456 "pcregrep -".
4457
4458 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4459 Java package. This provides some syntactic sugar for simple cases of what my
4460 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4461 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4462 item, you can use this simplified notation. Note that only makes sense with
4463 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4464 greediness, whatever the setting of the PCRE_UNGREEDY option.
4465
4466 13. A change of greediness default within a pattern was not taking effect at
4467 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4468 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4469 was abstracted outside.
4470
4471 14. PCRE now supports the \G assertion. It is true when the current matching
4472 position is at the start point of the match. This differs from \A when the
4473 starting offset is non-zero. Used with the /g option of pcretest (or similar
4474 code), it works in the same way as it does for Perl's /g option. If all
4475 alternatives of a regex begin with \G, the expression is anchored to the start
4476 match position, and the "anchored" flag is set in the compiled expression.
4477
4478 15. Some bugs concerning the handling of certain option changes within patterns
4479 have been fixed. These applied to options other than (?ims). For example,
4480 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4481 been the other way round. Some of this was related to change 7 above.
4482
4483 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4484 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4485 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4486 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4487
4488 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4489 mode. Unlike ".", it always matches newline, whatever the setting of
4490 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4491 assertions. Perl allows it, but it doesn't (in general) work because it can't
4492 calculate the length of the lookbehind. At least, that's the case for Perl
4493 5.8.0 - I've been told they are going to document that it doesn't work in
4494 future.
4495
4496 18. Added an error diagnosis for escapes that PCRE does not support: these are
4497 \L, \l, \N, \P, \p, \U, \u, and \X.
4498
4499 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4500 reading past the end of the pattern in cases such as /[abcd/.
4501
4502 20. PCRE was getting more memory than necessary for patterns with classes that
4503 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4504
4505 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4506 compiling PCRE for use with Virtual Pascal.
4507
4508 22. Small fix to the Makefile to make it work properly if the build is done
4509 outside the source tree.
4510
4511 23. Added a new extension: a condition to go with recursion. If a conditional
4512 subpattern starts with (?(R) the "true" branch is used if recursion has
4513 happened, whereas the "false" branch is used only at the top level.
4514
4515 24. When there was a very long string of literal characters (over 255 bytes
4516 without UTF support, over 250 bytes with UTF support), the computation of how
4517 much memory was required could be incorrect, leading to segfaults or other
4518 strange effects.
4519
4520 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4521 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4522 there was a subsequent back reference to those brackets. This meant that, for
4523 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4524 possible to check for precisely this case. All we can do is abandon the
4525 optimization if .* occurs inside capturing brackets when there are any back
4526 references whatsoever. (See below for a better fix that came later.)
4527
4528 26. The handling of the optimization for finding the first character of a
4529 non-anchored pattern, and for finding a character that is required later in the
4530 match were failing in some cases. This didn't break the matching; it just
4531 failed to optimize when it could. The way this is done has been re-implemented.
4532
4533 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4534
4535 28. Added a new feature that provides some of the functionality that Perl
4536 provides with (?{...}). The facility is termed a "callout". The way it is done
4537 in PCRE is for the caller to provide an optional function, by setting
4538 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4539 global variable. By default it is unset, which disables all calling out. To get
4540 the function called, the regex must include (?C) at appropriate points. This
4541 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4542 This provides a means of identifying different callout points. When PCRE
4543 reaches such a point in the regex, if pcre_callout has been set, the external
4544 function is called. It is provided with data in a structure called
4545 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4546 matching continues; if it returns a non-zero value, the match at the current
4547 point fails. However, backtracking will occur if possible. [This was changed
4548 later and other features added - see item 49 below.]
4549
4550 29. pcretest is upgraded to test the callout functionality. It provides a
4551 callout function that displays information. By default, it shows the start of
4552 the match and the current position in the text. There are some new data escapes
4553 to vary what happens:
4554
4555 \C+ in addition, show current contents of captured substrings
4556 \C- do not supply a callout function
4557 \C!n return 1 when callout number n is reached
4558 \C!n!m return 1 when callout number n is reached for the mth time
4559
4560 30. If pcregrep was called with the -l option and just a single file name, it
4561 output "<stdin>" if a match was found, instead of the file name.
4562
4563 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4564 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4565 pcre_exec(). This saves a malloc/free per call. The default value of
4566 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4567 when configuring.
4568
4569 32. The default maximum size of a compiled pattern is 64K. There have been a
4570 few cases of people hitting this limit. The code now uses macros to handle the
4571 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4572 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4573 configuring. Tests 2 and 5 work only with 2-byte links because they output
4574 debugging information about compiled patterns.
4575
4576 33. Internal code re-arrangements:
4577
4578 (a) Moved the debugging function for printing out a compiled regex into
4579 its own source file (printint.c) and used #include to pull it into
4580 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4581 separate copies.
4582
4583 (b) Defined the list of op-code names for debugging as a macro in
4584 internal.h so that it is next to the definition of the opcodes.
4585
4586 (c) Defined a table of op-code lengths for simpler skipping along compiled
4587 code. This is again a macro in internal.h so that it is next to the
4588 definition of the opcodes.
4589
4590 34. Added support for recursive calls to individual subpatterns, along the
4591 lines of Robin Houston's patch (but implemented somewhat differently).
4592
4593 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4594 allow it to read and process whole directories in Win32. This code was
4595 contributed by Lionel Fourquaux; it has not been tested by me.
4596
4597 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4598 used to name a group. Names consist of alphanumerics and underscores, and must
4599 be unique. Back references use the syntax (?P=name) and recursive calls use
4600 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4601 numbers. The function pcre_fullinfo() can be used after compilation to extract
4602 a name/number map. There are three relevant calls:
4603
4604 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4605 PCRE_INFO_NAMECOUNT yields the number of entries
4606 PCRE_INFO_NAMETABLE yields a pointer to the map.
4607
4608 The map is a vector of fixed-size entries. The size of each entry depends on
4609 the length of the longest name used. The first two bytes of each entry are the
4610 group number, most significant byte first. There follows the corresponding
4611 name, zero terminated. The names are in alphabetical order.
4612
4613 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4614 case instead of 255. Making it the same both with and without UTF-8 support
4615 means that the same test output works with both.
4616
4617 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4618 calling malloc() with a zero argument.
4619
4620 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4621 optimization. I've improved things by keeping a bitmap of backreferences with
4622 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4623 fact referenced, the optimization can be applied. It is unlikely that a
4624 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4625 the match to follow \n) will appear inside brackets with a number greater than
4626 31, but if it does, any back reference > 31 suppresses the optimization.
4627
4628 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4629 of disabling numbered capturing parentheses. Any opening parenthesis that is
4630 not followed by ? behaves as if it were followed by ?: but named parentheses
4631 can still be used for capturing (and they will acquire numbers in the usual
4632 way).
4633
4634 41. Redesigned the return codes from the match() function into yes/no/error so
4635 that errors can be passed back from deep inside the nested calls. A malloc
4636 failure while inside a recursive subpattern call now causes the
4637 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4638
4639 42. It is now possible to set a limit on the number of times the match()
4640 function is called in a call to pcre_exec(). This facility makes it possible to
4641 limit the amount of recursion and backtracking, though not in a directly
4642 obvious way, because the match() function is used in a number of different
4643 circumstances. The count starts from zero for each position in the subject
4644 string (for non-anchored patterns). The default limit is, for compatibility, a
4645 large number, namely 10 000 000. You can change this in two ways:
4646
4647 (a) When configuring PCRE before making, you can use --with-match-limit=n
4648 to set a default value for the compiled library.
4649
4650 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4651 a different value is set. See 45 below.
4652
4653 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4654
4655 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4656 of things that can be changed at compile time. The first argument specifies
4657 what is wanted and the second points to where the information is to be placed.
4658 The current list of available information is:
4659
4660 PCRE_CONFIG_UTF8
4661
4662 The output is an integer that is set to one if UTF-8 support is available;
4663 otherwise it is set to zero.
4664
4665 PCRE_CONFIG_NEWLINE
4666
4667 The output is an integer that it set to the value of the code that is used for
4668 newline. It is either LF (10) or CR (13).
4669
4670 PCRE_CONFIG_LINK_SIZE
4671
4672 The output is an integer that contains the number of bytes used for internal
4673 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4674
4675 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4676
4677 The output is an integer that contains the threshold above which the POSIX
4678 interface uses malloc() for output vectors. See item 31 above.
4679
4680 PCRE_CONFIG_MATCH_LIMIT
4681
4682 The output is an unsigned integer that contains the default limit of the number
4683 of match() calls in a pcre_exec() execution. See 42 above.
4684
4685 44. pcretest has been upgraded by the addition of the -C option. This causes it
4686 to extract all the available output from the new pcre_config() function, and to
4687 output it. The program then exits immediately.
4688
4689 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4690 order to support additional features. One way would have been to define
4691 pcre_exec2() (for example) with extra arguments, but this would not have been
4692 extensible, and would also have required all calls to the original function to
4693 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4694 is used for passing in "extra" data from pcre_study().
4695
4696 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4697 contains the following fields:
4698
4699 flags a bitmap indicating which of the following fields are set
4700 study_data opaque data from pcre_study()
4701 match_limit a way of specifying a limit on match() calls for a specific
4702 call to pcre_exec()
4703 callout_data data for callouts (see 49 below)
4704
4705 The flag bits are also defined in pcre.h, and are
4706
4707 PCRE_EXTRA_STUDY_DATA
4708 PCRE_EXTRA_MATCH_LIMIT
4709 PCRE_EXTRA_CALLOUT_DATA
4710
4711 The pcre_study() function now returns one of these new pcre_extra blocks, with
4712 the actual study data pointed to by the study_data field, and the
4713 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4714 before. That is, this change is entirely upwards-compatible and requires no
4715 change to existing code.
4716
4717 If you want to pass in additional data to pcre_exec(), you can either place it
4718 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4719 block.
4720
4721 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4722 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4723 times with different match limits, until it finds the minimum value needed for
4724 pcre_exec() to complete. The value is then output. This can be instructive; for
4725 most simple matches the number is quite small, but for pathological cases it
4726 gets very large very quickly.
4727
4728 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4729 returns the size of the data block pointed to by the study_data field in a
4730 pcre_extra block, that is, the value that was passed as the argument to
4731 pcre_malloc() when PCRE was getting memory in which to place the information
4732 created by pcre_study(). The fourth argument should point to a size_t variable.
4733 pcretest has been extended so that this information is shown after a successful
4734 pcre_study() call when information about the compiled regex is being displayed.
4735
4736 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4737 because what follows is always an absolute path. (Later: it turns out that this
4738 is more than cosmetic for MinGW, because it doesn't like empty path
4739 components.)
4740
4741 49. Some changes have been made to the callout feature (see 28 above):
4742
4743 (i) A callout function now has three choices for what it returns:
4744
4745 0 => success, carry on matching
4746 > 0 => failure at this point, but backtrack if possible
4747 < 0 => serious error, return this value from pcre_exec()
4748
4749 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4750 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4751 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4752 use by callout functions. It will never be used by PCRE itself.
4753
4754 (ii) The pcre_extra structure (see 45 above) has a void * field called
4755 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4756 pcre_callout_block structure has a field of the same name. The contents of
4757 the field passed in the pcre_extra structure are passed to the callout
4758 function in the corresponding field in the callout block. This makes it
4759 easier to use the same callout-containing regex from multiple threads. For
4760 testing, the pcretest program has a new data escape
4761
4762 \C*n pass the number n (may be negative) as callout_data
4763
4764 If the callout function in pcretest receives a non-zero value as
4765 callout_data, it returns that value.
4766
4767 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4768 there were some redundant $(CFLAGS) in commands that are now specified as
4769 $(LINK), which already includes $(CFLAGS).
4770
4771 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4772 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4773 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4774 one-byte characters throughout. Note that case-insensitive matching applies
4775 only to characters whose values are less than 256. PCRE doesn't support the
4776 notion of cases for higher-valued characters.
4777
4778 (i) A character class whose characters are all within 0-255 is handled as
4779 a bit map, and the map is inverted for negative classes. Previously, a
4780 character > 255 always failed to match such a class; however it should
4781 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4782
4783 (ii) A negated character class with a single character < 255 is coded as
4784 "not this character" (OP_NOT). This wasn't working properly when the test
4785 character was multibyte, either singly or repeated.
4786
4787 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4788 mode, for example: \x{100}{2,3}.
4789
4790 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4791 singly or repeated) now correctly test multibyte characters. However,
4792 PCRE doesn't recognize any characters with values greater than 255 as
4793 digits, spaces, or word characters. Such characters always match \D, \S,
4794 and \W, and never match \d, \s, or \w.
4795
4796 (v) Classes may now contain characters and character ranges with values
4797 greater than 255. For example: [ab\x{100}-\x{400}].
4798
4799 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4800 PCRE in UTF-8 mode.
4801
4802 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4803 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4804 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4805 value.)
4806
4807 53. The single man page has become too large. I have therefore split it up into
4808 a number of separate man pages. These also give rise to individual HTML pages;
4809 these are now put in a separate directory, and there is an index.html page that
4810 lists them all. Some hyperlinking between the pages has been installed.
4811
4812 54. Added convenience functions for handling named capturing parentheses.
4813
4814 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4815 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4816 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4817 are faulted.
4818
4819 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4820 calling configure. These values are used when compiling the dftables.c program
4821 which is run to generate the source of the default character tables. They
4822 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4823 you will need to set these values.
4824
4825 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4826
4827
4828 Version 3.9 02-Jan-02
4829 ---------------------
4830
4831 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4832
4833 2. If --disable-static was given, the building process failed when trying to
4834 build pcretest and pcregrep. (For some reason it was using libtool to compile
4835 them, which is not right, as they aren't part of the library.)
4836
4837
4838 Version 3.8 18-Dec-01
4839 ---------------------
4840
4841 1. The experimental UTF-8 code was completely screwed up. It was packing the
4842 bytes in the wrong order. How dumb can you get?
4843
4844
4845 Version 3.7 29-Oct-01
4846 ---------------------
4847
4848 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4849 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4850 this didn't happen under Solaris 8, where I normally test things.
4851
4852 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4853 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4854 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4855 this fix an hour or so after the initial 3.7 release.)
4856
4857
4858 Version 3.6 23-Oct-01
4859 ---------------------
4860
4861 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4862 offsets passed as NULL with zero offset count.
4863
4864 2. The config.guess and config.sub files had not been updated when I moved to
4865 the latest autoconf.
4866
4867
4868 Version 3.5 15-Aug-01
4869 ---------------------
4870
4871 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4872 had been forgotten.
4873
4874 2. By using declared but undefined structures, we can avoid using "void"
4875 definitions in pcre.h while keeping the internal definitions of the structures
4876 private.
4877
4878 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4879 user point of view, this means that both static and shared libraries are built
4880 by default, but this can be individually controlled. More of the work of
4881 handling this static/shared cases is now inside libtool instead of PCRE's make
4882 file.
4883
4884 4. The pcretest utility is now installed along with pcregrep because it is
4885 useful for users (to test regexs) and by doing this, it automatically gets
4886 relinked by libtool. The documentation has been turned into a man page, so
4887 there are now .1, .txt, and .html versions in /doc.
4888
4889 5. Upgrades to pcregrep:
4890 (i) Added long-form option names like gnu grep.
4891 (ii) Added --help to list all options with an explanatory phrase.
4892 (iii) Added -r, --recursive to recurse into sub-directories.
4893 (iv) Added -f, --file to read patterns from a file.
4894
4895 6. pcre_exec() was referring to its "code" argument before testing that
4896 argument for NULL (and giving an error if it was NULL).
4897
4898 7. Upgraded Makefile.in to allow for compiling in a different directory from
4899 the source directory.
4900
4901 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4902 options bits, the pointer it was passed was to an int instead of to an unsigned
4903 long int. This mattered only on 64-bit systems.
4904
4905 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4906 generated) instead of pcre.in, which it its source. Also made the same change
4907 in several of the .c files.
4908
4909 10. A new release of gcc defines printf() as a macro, which broke pcretest
4910 because it had an ifdef in the middle of a string argument for printf(). Fixed
4911 by using separate calls to printf().
4912
4913 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4914 script, to force use of CR or LF instead of \n in the source. On non-Unix
4915 systems, the value can be set in config.h.
4916
4917 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4918 absolute limit. Changed the text of the error message to make this clear, and
4919 likewise updated the man page.
4920
4921 13. The limit of 99 on the number of capturing subpatterns has been removed.
4922 The new limit is 65535, which I hope will not be a "real" limit.
4923
4924
4925 Version 3.4 22-Aug-00
4926 ---------------------
4927
4928 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4929
4930 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4931
4932
4933 Version 3.3 01-Aug-00
4934 ---------------------
4935
4936 1. If an octal character was given, but the value was greater than \377, it
4937 was not getting masked to the least significant bits, as documented. This could
4938 lead to crashes in some systems.
4939
4940 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4941 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4942
4943 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4944 These just pass their arguments on to (pcre_free)(), but they are provided
4945 because some uses of PCRE bind it to non-C systems that can call its functions,
4946 but cannot call free() or pcre_free() directly.
4947
4948 4. Add "make test" as a synonym for "make check". Corrected some comments in
4949 the Makefile.
4950
4951 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4952 Makefile.
4953
4954 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4955 command called pgrep for grepping around the active processes.
4956
4957 7. Added the beginnings of support for UTF-8 character strings.
4958
4959 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4960 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4961 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4962 out for the ar command.)
4963
4964
4965 Version 3.2 12-May-00
4966 ---------------------
4967
4968 This is purely a bug fixing release.
4969
4970 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4971 of ZA. This was just one example of several cases that could provoke this bug,
4972 which was introduced by change 9 of version 2.00. The code for breaking
4973 infinite loops after an iteration that matches an empty string was't working
4974 correctly.
4975
4976 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4977 when matched against abbab (for example). After matching an empty string, it
4978 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4979 caused it to match further down the string than it should.
4980
4981 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4982 was there because it doesn't seem to be needed, and it causes trouble on some
4983 systems, as it is not a Standard C header. It has been removed.
4984
4985 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4986 were reported on the Macintosh. The changes were from
4987
4988 while ((c = *(++ptr)) != 0 && c != '\n');
4989 to
4990 while ((c = *(++ptr)) != 0 && c != '\n') ;
4991
4992 Totally extraordinary, but if that's what it takes...
4993
4994 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4995 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4996 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4997 assumes the way PCRE uses memmove() (always moving upwards).
4998
4999 6. PCRE is being used in one environment where strchr() is not available. There
5000 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5001 faster code anyway.
5002
5003
5004 Version 3.1 09-Feb-00
5005 ---------------------
5006
5007 The only change in this release is the fixing of some bugs in Makefile.in for
5008 the "install" target:
5009
5010 (1) It was failing to install pcreposix.h.
5011
5012 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5013
5014
5015 Version 3.0 01-Feb-00
5016 ---------------------
5017
5018 1. Add support for the /+ modifier to perltest (to output $` like it does in
5019 pcretest).
5020
5021 2. Add support for the /g modifier to perltest.
5022
5023 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5024 matches null strings.
5025
5026 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5027 pattern. Perl treats empty patterns specially - it reuses the most recent
5028 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5029 effect.
5030
5031 5. The POSIX interface was broken in that it was just handing over the POSIX
5032 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5033 required a bigger vector, with some working space on the end. This means that
5034 the POSIX wrapper now has to get and free some memory, and copy the results.
5035
5036 6. Added some simple autoconf support, placing the test data and the
5037 documentation in separate directories, re-organizing some of the
5038 information files, and making it build pcre-config (a GNU standard). Also added
5039 libtool support for building PCRE as a shared library, which is now the
5040 default.
5041
5042 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5043 09 are not valid octal constants. Single digits will be used for minor values
5044 less than 10.
5045
5046 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5047 existing programs that set these in the POSIX interface can use PCRE without
5048 modification.
5049
5050 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5051 return all that pcre_info() returns, plus additional data. The pcre_info()
5052 function is retained for compatibility, but is considered to be obsolete.
5053
5054 10. Added experimental recursion feature (?R) to handle one common case that
5055 Perl 5.6 will be able to do with (?p{...}).
5056
5057 11. Added support for POSIX character classes like [:alpha:], which Perl is
5058 adopting.
5059
5060
5061 Version 2.08 31-Aug-99
5062 ----------------------
5063
5064 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5065 trying to match at the startoffset position, but instead was moving forward to
5066 the next newline as if a previous match had failed.
5067
5068 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5069 and could get into a loop if a null string was matched other than at the start
5070 of the subject.
5071
5072 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5073 be distinguished at compile time, and for completeness also added PCRE_DATE.
5074
5075 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5076 in GnuWin32 environments.
5077
5078
5079 Version 2.07 29-Jul-99
5080 ----------------------
5081
5082 1. The documentation is now supplied in plain text form and HTML as well as in
5083 the form of man page sources.
5084
5085 2. C++ compilers don't like assigning (void *) values to other pointer types.
5086 In particular this affects malloc(). Although there is no problem in Standard
5087 C, I've put in casts to keep C++ compilers happy.
5088
5089 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5090 should be (const char *).
5091
5092 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5093 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5094 However, I haven't made this a standard facility. The documentation doesn't
5095 mention it, and the Makefile doesn't support it.
5096
5097 5. The Makefile now contains an "install" target, with editable destinations at
5098 the top of the file. The pcretest program is not installed.
5099
5100 6. pgrep -V now gives the PCRE version number and date.
5101
5102 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5103 causing the entire string to be ignored, instead of just the last character.
5104
5105 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5106 non-matching string, it can take a very, very long time, even for strings of
5107 quite modest length, because of the nested recursion. PCRE now does better in
5108 some of these cases. It does this by remembering the last required literal
5109 character in the pattern, and pre-searching the subject to ensure it is present
5110 before running the real match. In other words, it applies a heuristic to detect
5111 some types of certain failure quickly, and in the above example, if presented
5112 with a string that has no trailing " it gives "no match" very quickly.
5113
5114 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5115 other alternatives are tried instead.
5116
5117
5118 Version 2.06 09-Jun-99
5119 ----------------------
5120
5121 1. Change pcretest's output for amount of store used to show just the code
5122 space, because the remainder (the data block) varies in size between 32-bit and
5123 64-bit systems.
5124
5125 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5126 start matching at. This allows lookbehinds to work when searching for multiple
5127 occurrences in a string.
5128
5129 3. Added additional options to pcretest for testing multiple occurrences:
5130
5131 /+ outputs the rest of the string that follows a match
5132 /g loops for multiple occurrences, using the new startoffset argument
5133 /G loops for multiple occurrences by passing an incremented pointer
5134
5135 4. PCRE wasn't doing the "first character" optimization for patterns starting
5136 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5137 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5138 the letter 'x'. On long subject strings, this gives a significant speed-up.
5139
5140
5141 Version 2.05 21-Apr-99
5142 ----------------------
5143
5144 1. Changed the type of magic_number from int to long int so that it works
5145 properly on 16-bit systems.
5146
5147 2. Fixed a bug which caused patterns starting with .* not to work correctly
5148 when the subject string contained newline characters. PCRE was assuming
5149 anchoring for such patterns in all cases, which is not correct because .* will
5150 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5151 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5152 must be retried after every newline in the subject.
5153
5154
5155 Version 2.04 18-Feb-99
5156 ----------------------
5157
5158 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5159 computation of the store needed to hold the pattern was incorrect (too large).
5160 If such patterns were nested a few deep, this could multiply and become a real
5161 problem.
5162
5163 2. Added /M option to pcretest to show the memory requirement of a specific
5164 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5165
5166 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5167 compiled in such a way that the backtracking after subsequent failure was
5168 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5169 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5170
5171
5172 Version 2.03 02-Feb-99
5173 ----------------------
5174
5175 1. Fixed typo and small mistake in man page.
5176
5177 2. Added 4th condition (GPL supersedes if conflict) and created separate
5178 LICENCE file containing the conditions.
5179
5180 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5181 Perl, that is the internal \ allows the delimiter to be included in the
5182 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5183 the final delimiter, add \ to the end of the pattern (to test the error).
5184
5185 4. Added the convenience functions for extracting substrings after a successful
5186 match. Updated pcretest to make it able to test these functions.
5187
5188
5189 Version 2.02 14-Jan-99
5190 ----------------------
5191
5192 1. Initialized the working variables associated with each extraction so that
5193 their saving and restoring doesn't refer to uninitialized store.
5194
5195 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5196 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5197 fix the problem.
5198
5199 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5200 calls, and wasn't printing the correct value for compiling calls. Increased the
5201 default value of LOOPREPEAT, and the number of significant figures in the
5202 times.
5203
5204 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5205
5206 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5207 a building problem on Windows NT with a FAT file system.
5208
5209
5210 Version 2.01 21-Oct-98
5211 ----------------------
5212
5213 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5214 to character tables built by pcre_maketables() in the current locale. If NULL
5215 is passed, the default tables are used.
5216
5217
5218 Version 2.00 24-Sep-98
5219 ----------------------
5220
5221 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5222 it any more.
5223
5224 2. Allow quantification of (?>) groups, and make it work correctly.
5225
5226 3. The first character computation wasn't working for (?>) groups.
5227
5228 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5229 end of the subject) and add 5.005's \z, which really does match only at the
5230 very end of the subject.
5231
5232 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5233
5234 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5235 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5236 localized options. All options to pcre_study() were also removed.
5237
5238 7. Add other new features from 5.005:
5239
5240 $(?<= positive lookbehind
5241 $(?<! negative lookbehind
5242 (?imsx-imsx) added the unsetting capability
5243 such a setting is global if at outer level; local otherwise
5244 (?imsx-imsx:) non-capturing groups with option setting
5245 (?(cond)re|re) conditional pattern matching
5246
5247 A backreference to itself in a repeated group matches the previous
5248 captured string.
5249
5250 8. General tidying up of studying (both automatic and via "study")
5251 consequential on the addition of new assertions.
5252
5253 9. As in 5.005, unlimited repeated groups that could match an empty substring
5254 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5255 runtime if any iteration does actually match an empty substring.
5256
5257 10. Include the RunTest script in the distribution.
5258
5259 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5260 discrepancies, some of which were old and were also with respect to 5.004. They
5261 have now been fixed.
5262
5263
5264 Version 1.09 28-Apr-98
5265 ----------------------
5266
5267 1. A negated single character class followed by a quantifier with a minimum
5268 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5269 program crashes, or just wrong answers. This did not apply to negated classes
5270 containing more than one character, or to minima other than one.
5271
5272
5273 Version 1.08 27-Mar-98
5274 ----------------------
5275
5276 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5277
5278 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5279 latter must appear before anything that relies on it in the pattern.
5280
5281
5282 Version 1.07 16-Feb-98
5283 ----------------------
5284
5285 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5286 repeat of a potentially empty string).
5287
5288
5289 Version 1.06 23-Jan-98
5290 ----------------------
5291
5292 1. Added Markus Oberhumer's little patches for C++.
5293
5294 2. Literal strings longer than 255 characters were broken.
5295
5296
5297 Version 1.05 23-Dec-97
5298 ----------------------
5299
5300 1. Negated character classes containing more than one character were failing if
5301 PCRE_CASELESS was set at run time.
5302
5303
5304 Version 1.04 19-Dec-97
5305 ----------------------
5306
5307 1. Corrected the man page, where some "const" qualifiers had been omitted.
5308
5309 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5310 input syntax.
5311
5312 3. Fixed memory leak which occurred when a regex with back references was
5313 matched with an offsets vector that wasn't big enough. The temporary memory
5314 that is used in this case wasn't being freed if the match failed.
5315
5316 4. Tidied pcretest to ensure it frees memory that it gets.
5317
5318 5. Temporary memory was being obtained in the case where the passed offsets
5319 vector was exactly big enough.
5320
5321 6. Corrected definition of offsetof() from change 5 below.
5322
5323 7. I had screwed up change 6 below and broken the rules for the use of
5324 setjmp(). Now fixed.
5325
5326
5327 Version 1.03 18-Dec-97
5328 ----------------------
5329
5330 1. A erroneous regex with a missing opening parenthesis was correctly
5331 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5332 on some systems.
5333
5334 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5335 it was reported that one broken compiler failed on the former because "code" is
5336 also an independent variable.
5337
5338 3. The erroneous regex a[]b caused an array overrun reference.
5339
5340 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5341 fail on data ending with that character. (It was going on too far, and checking
5342 the next character, typically a binary zero.) This was specific to the
5343 optimized code for single-character negative classes.
5344
5345 5. Added a contributed patch from the TIN world which does the following:
5346
5347 + Add an undef for memmove, in case the the system defines a macro for it.
5348
5349 + Add a definition of offsetof(), in case there isn't one. (I don't know
5350 the reason behind this - offsetof() is part of the ANSI standard - but
5351 it does no harm).
5352
5353 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5354 most of the places where whitespace preceded '#'. I have given up and
5355 allowed the remaining 2 cases to be at the margin.
5356
5357 + Rename some variables in pcre to eliminate shadowing. This seems very
5358 pedantic, but does no harm, of course.
5359
5360 6. Moved the call to setjmp() into its own function, to get rid of warnings
5361 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5362
5363 7. Constructs such as \d{8,} were compiling into the equivalent of
5364 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5365 outcome, but in this particular case used more store than had been allocated,
5366 which caused the bug to be discovered because it threw up an internal error.
5367
5368 8. The debugging code in both pcre and pcretest for outputting the compiled
5369 form of a regex was going wrong in the case of back references followed by
5370 curly-bracketed repeats.
5371
5372
5373 Version 1.02 12-Dec-97
5374 ----------------------
5375
5376 1. Typos in pcre.3 and comments in the source fixed.
5377
5378 2. Applied a contributed patch to get rid of places where it used to remove
5379 'const' from variables, and fixed some signed/unsigned and uninitialized
5380 variable warnings.
5381
5382 3. Added the "runtest" target to Makefile.
5383
5384 4. Set default compiler flag to -O2 rather than just -O.
5385
5386
5387 Version 1.01 19-Nov-97
5388 ----------------------
5389
5390 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5391 like /([ab]*)*/, that is, for classes with more than one character in them.
5392
5393 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5394 as /((?>a*))*/ (a PCRE_EXTRA facility).
5395
5396
5397 Version 1.00 18-Nov-97
5398 ----------------------
5399
5400 1. Added compile-time macros to support systems such as SunOS4 which don't have
5401 memmove() or strerror() but have other things that can be used instead.
5402
5403 2. Arranged that "make clean" removes the executables.
5404
5405
5406 Version 0.99 27-Oct-97
5407 ----------------------
5408
5409 1. Fixed bug in code for optimizing classes with only one character. It was
5410 initializing a 32-byte map regardless, which could cause it to run off the end
5411 of the memory it had got.
5412
5413 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5414
5415
5416 Version 0.98 22-Oct-97
5417 ----------------------
5418
5419 1. Fixed bug in code for handling temporary memory usage when there are more
5420 back references than supplied space in the ovector. This could cause segfaults.
5421
5422
5423 Version 0.97 21-Oct-97
5424 ----------------------
5425
5426 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5427
5428 2. Optimized negated single characters not to use a bit map.
5429
5430 3. Brought error texts together as macro definitions; clarified some of them;
5431 fixed one that was wrong - it said "range out of order" when it meant "invalid
5432 escape sequence".
5433
5434 4. Changed some char * arguments to const char *.
5435
5436 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5437
5438 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5439 pcretest.
5440
5441
5442 Version 0.96 16-Oct-97
5443 ----------------------
5444
5445 1. Added a simple "pgrep" utility to the distribution.
5446
5447 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5448 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5449 where "ddd" means "one or more decimal digits".
5450
5451 3. Fixed serious bug. If a pattern had a back reference, but the call to
5452 pcre_exec() didn't supply a large enough ovector to record the related
5453 identifying subpattern, the match always failed. PCRE now remembers the number
5454 of the largest back reference, and gets some temporary memory in which to save
5455 the offsets during matching if necessary, in order to ensure that
5456 backreferences always work.
5457
5458 4. Increased the compatibility with Perl in a number of ways:
5459
5460 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5461 to request this handling. The option can be set at compile or exec time.
5462
5463 (b) $ matches before a terminating newline by default; an option
5464 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5465 mode). The option can be set at compile or exec time.
5466
5467 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5468 the same as Perl's. If the decimal number it represents is less than 10
5469 or there aren't that many previous left capturing parentheses, an octal
5470 escape is read. Inside a character class, it's always an octal escape,
5471 even if it is a single digit.
5472
5473 (d) An escaped but undefined alphabetic character is taken as a literal,
5474 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5475 escapes.
5476
5477 (e) {0} is now permitted. (The previous item is removed from the compiled
5478 pattern).
5479
5480 5. Changed all the names of code files so that the basic parts are no longer
5481 than 10 characters, and abolished the teeny "globals.c" file.
5482
5483 6. Changed the handling of character classes; they are now done with a 32-byte
5484 bit map always.
5485
5486 7. Added the -d and /D options to pcretest to make it possible to look at the
5487 internals of compilation without having to recompile pcre.
5488
5489
5490 Version 0.95 23-Sep-97
5491 ----------------------
5492
5493 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5494 \x20 at the start of a run of normal characters. These were being treated as
5495 real characters, instead of the source characters being re-checked.
5496
5497
5498 Version 0.94 18-Sep-97
5499 ----------------------
5500
5501 1. The functions are now thread-safe, with the caveat that the global variables
5502 containing pointers to malloc() and free() or alternative functions are the
5503 same for all threads.
5504
5505 2. Get pcre_study() to generate a bitmap of initial characters for non-
5506 anchored patterns when this is possible, and use it if passed to pcre_exec().
5507
5508
5509 Version 0.93 15-Sep-97
5510 ----------------------
5511
5512 1. /(b)|(:+)/ was computing an incorrect first character.
5513
5514 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5515 but not actually doing anything yet.
5516
5517 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5518 as Perl does (e.g. [-az] or [az-]).
5519
5520 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5521 all possible positions.
5522
5523 5. Split up into different modules to avoid including unneeded functions in a
5524 compiled binary. However, compile and exec are still in one module. The "study"
5525 function is split off.
5526
5527 6. The character tables are now in a separate module whose source is generated
5528 by an auxiliary program - but can then be edited by hand if required. There are
5529 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5530 toupper() in the code.
5531
5532 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5533 make them global. Abolish the function for setting them, as the caller can now
5534 set them directly.
5535
5536
5537 Version 0.92 11-Sep-97
5538 ----------------------
5539
5540 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5541 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5542
5543 2. Caseless matching was not working in character classes if the characters in
5544 the pattern were in upper case.
5545
5546 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5547
5548 4. Make PCRE_ANCHORED public and accept as a compile option.
5549
5550 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5551 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5552 pass them.
5553
5554 6. Give an error if bad option bits passed at compile or run time.
5555
5556 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5557 pcretest to cause it to pass that flag.
5558
5559 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5560 options, and the first character, if set.
5561
5562 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5563
5564
5565 Version 0.91 10-Sep-97
5566 ----------------------
5567
5568 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5569 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5570
5571 2. PCRE was looping on encountering an indefinitely repeated back reference to
5572 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5573 Perl does - treats the match as successful.
5574
5575 ****

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