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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1515 - (show annotations)
Tue Feb 3 17:08:54 2015 UTC (4 years, 7 months ago) by ph10
File size: 269372 byte(s)
Error occurred while calculating annotation data.
Fix pcretest crash with /g and zero-length ovector.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.37 xx-xxx-201x
5 ------------------------
6
7 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8 for those parentheses to be closed with whatever has been captured so far.
9 However, it was failing to mark any other groups between the hightest
10 capture so far and the currrent group as "unset". Thus, the ovector for
11 those groups contained whatever was previously there. An example is the
12 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
13
14 2. If an assertion condition was quantified with a minimum of zero (an odd
15 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
16
17 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
18 unrecognized modifier, a crash could occur.
19
20 4. An attempt to do global matching in pcretest with a zero-length ovector
21 caused a crash.
22
23
24
25 Version 8.36 26-September-2014
26 ------------------------------
27
28 1. Got rid of some compiler warnings in the C++ modules that were shown up by
29 -Wmissing-field-initializers and -Wunused-parameter.
30
31 2. The tests for quantifiers being too big (greater than 65535) were being
32 applied after reading the number, and stupidly assuming that integer
33 overflow would give a negative number. The tests are now applied as the
34 numbers are read.
35
36 3. Tidy code in pcre_exec.c where two branches that used to be different are
37 now the same.
38
39 4. The JIT compiler did not generate match limit checks for certain
40 bracketed expressions with quantifiers. This may lead to exponential
41 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
42 issue should be resolved now.
43
44 5. Fixed an issue, which occures when nested alternatives are optimized
45 with table jumps.
46
47 6. Inserted two casts and changed some ints to size_t in the light of some
48 reported 64-bit compiler warnings (Bugzilla 1477).
49
50 7. Fixed a bug concerned with zero-minimum possessive groups that could match
51 an empty string, which sometimes were behaving incorrectly in the
52 interpreter (though correctly in the JIT matcher). This pcretest input is
53 an example:
54
55 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
56 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
57
58 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
59 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
60 for an empty string was breaking the inner loop and carrying on at a lower
61 level, when possessive repeated groups should always return to a higher
62 level as they have no backtrack points in them. The empty string test now
63 occurs at the outer level.
64
65 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
66 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
67
68 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
69 doesn't).
70
71 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
72 to hold a pattern to be incorrectly computed (too small) when there were
73 named back references to duplicated names. This could cause "internal
74 error: code overflow" or "double free or corruption" or other memory
75 handling errors.
76
77 11. When named subpatterns had the same prefixes, back references could be
78 confused. For example, in this pattern:
79
80 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
81
82 the reference to 'Name' was incorrectly treated as a reference to a
83 duplicate name.
84
85 12. A pattern such as /^s?c/mi8 where the optional character has more than
86 one "other case" was incorrectly compiled such that it would only try to
87 match starting at "c".
88
89 13. When a pattern starting with \s was studied, VT was not included in the
90 list of possible starting characters; this should have been part of the
91 8.34/18 patch.
92
93 14. If a character class started [\Qx]... where x is any character, the class
94 was incorrectly terminated at the ].
95
96 15. If a pattern that started with a caseless match for a character with more
97 than one "other case" was studied, PCRE did not set up the starting code
98 unit bit map for the list of possible characters. Now it does. This is an
99 optimization improvement, not a bug fix.
100
101 16. The Unicode data tables have been updated to Unicode 7.0.0.
102
103 17. Fixed a number of memory leaks in pcregrep.
104
105 18. Avoid a compiler warning (from some compilers) for a function call with
106 a cast that removes "const" from an lvalue by using an intermediate
107 variable (to which the compiler does not object).
108
109 19. Incorrect code was compiled if a group that contained an internal recursive
110 back reference was optional (had quantifier with a minimum of zero). This
111 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
112 caused segmentation faults because of stack overflows at compile time.
113
114 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
115 group that is quantified with an indefinite repeat, caused a compile-time
116 loop which used up all the system stack and provoked a segmentation fault.
117 This was not the same bug as 19 above.
118
119 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
120 Patch by Mike Frysinger.
121
122
123 Version 8.35 04-April-2014
124 --------------------------
125
126 1. A new flag is set, when property checks are present in an XCLASS.
127 When this flag is not set, PCRE can perform certain optimizations
128 such as studying these XCLASS-es.
129
130 2. The auto-possessification of character sets were improved: a normal
131 and an extended character set can be compared now. Furthermore
132 the JIT compiler optimizes more character set checks.
133
134 3. Got rid of some compiler warnings for potentially uninitialized variables
135 that show up only when compiled with -O2.
136
137 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
138 of a match later then the end of the match. The pcretest program was not
139 handling the case sensibly - it was outputting from the start to the next
140 binary zero. It now reports this situation in a message, and outputs the
141 text from the end to the start.
142
143 5. Fast forward search is improved in JIT. Instead of the first three
144 characters, any three characters with fixed position can be searched.
145 Search order: first, last, middle.
146
147 6. Improve character range checks in JIT. Characters are read by an inprecise
148 function now, which returns with an unknown value if the character code is
149 above a certain threshold (e.g: 256). The only limitation is that the value
150 must be bigger than the threshold as well. This function is useful when
151 the characters above the threshold are handled in the same way.
152
153 7. The macros whose names start with RAWUCHAR are placeholders for a future
154 mode in which only the bottom 21 bits of 32-bit data items are used. To
155 make this more memorable for those maintaining the code, the names have
156 been changed to start with UCHAR21, and an extensive comment has been added
157 to their definition.
158
159 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
160 to the export list in Makefile.am (they were accidentally omitted from the
161 8.34 tarball).
162
163 9. The informational output from pcretest used the phrase "starting byte set"
164 which is inappropriate for the 16-bit and 32-bit libraries. As the output
165 for "first char" and "need char" really means "non-UTF-char", I've changed
166 "byte" to "char", and slightly reworded the output. The documentation about
167 these values has also been (I hope) clarified.
168
169 10. Another JIT related optimization: use table jumps for selecting the correct
170 backtracking path, when more than four alternatives are present inside a
171 bracket.
172
173 11. Empty match is not possible, when the minimum length is greater than zero,
174 and there is no \K in the pattern. JIT should avoid empty match checks in
175 such cases.
176
177 12. In a caseless character class with UCP support, when a character with more
178 than one alternative case was not the first character of a range, not all
179 the alternative cases were added to the class. For example, s and \x{17f}
180 are both alternative cases for S: the class [RST] was handled correctly,
181 but [R-T] was not.
182
183 13. The configure.ac file always checked for pthread support when JIT was
184 enabled. This is not used in Windows, so I have put this test inside a
185 check for the presence of windows.h (which was already tested for).
186
187 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
188 The algorithm provides a way to skip certain starting offsets, and usually
189 faster than linear prefix searches.
190
191 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
192 as for 'fr_FR' and 'french'. For some reason, however, it then used the
193 Windows-specific input and output files, which have 'french' screwed in.
194 So this could never have worked. One of the problems with locales is that
195 they aren't always the same. I have now updated RunTest so that it checks
196 the output of the locale test (test 3) against three different output
197 files, and it allows the test to pass if any one of them matches. With luck
198 this should make the test pass on some versions of Solaris where it was
199 failing. Because of the uncertainty, the script did not used to stop if
200 test 3 failed; it now does. If further versions of a French locale ever
201 come to light, they can now easily be added.
202
203 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
204 there was a message during ./configure, but it did not stop. This now
205 provokes an error. The invalid example in README has been corrected.
206 If a value less than the minimum is given, the minimum value has always
207 been used, but now a warning is given.
208
209 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
210 was a bug in the test system, which is now fixed. Also, the list of various
211 configurations that are tested for each release did not have one with both
212 16/32 bits and --enable-bar-anycrlf. It now does.
213
214 18. pcretest was missing "-C bsr" for displaying the \R default setting.
215
216 19. Little endian PowerPC systems are supported now by the JIT compiler.
217
218 20. The fast forward newline mechanism could enter to an infinite loop on
219 certain invalid UTF-8 input. Although we don't support these cases
220 this issue can be fixed by a performance optimization.
221
222 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
223 not take account if existing stack usage. There is now a new global
224 variable called pcre_stack_guard that can be set to point to an external
225 function to check stack availability. It is called at the start of
226 processing every parenthesized group.
227
228 22. A typo in the code meant that in ungreedy mode the max/min qualifier
229 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
230 not match "ab".
231
232 23. When UTF was disabled, the JIT program reported some incorrect compile
233 errors. These messages are silenced now.
234
235 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
236 compiler.
237
238 25. Change all the temporary files used in RunGrepTest to be different to those
239 used by RunTest so that the tests can be run simultaneously, for example by
240 "make -j check".
241
242
243 Version 8.34 15-December-2013
244 -----------------------------
245
246 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
247 executable memory. Patch inspired by Carsten Klein.
248
249 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
250 this macro is never tested and has no effect, because the work to support
251 coverage involves only compiling and linking options and special targets in
252 the Makefile. The comment in config.h implied that defining the macro would
253 enable coverage support, which is totally false. There was also support for
254 setting this macro in the CMake files (my fault, I just copied it from
255 configure). SUPPORT_GCOV has now been removed.
256
257 3. Make a small performance improvement in strlen16() and strlen32() in
258 pcretest.
259
260 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
261 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
262 these cases). There was also one in pcretest.c.
263
264 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
265
266 6. In UTF mode, the code for checking whether a group could match an empty
267 string (which is used for indefinitely repeated groups to allow for
268 breaking an infinite loop) was broken when the group contained a repeated
269 negated single-character class with a character that occupied more than one
270 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
271 UTF-8 mode). The effect was undefined: the group might or might not be
272 deemed as matching an empty string, or the program might have crashed.
273
274 7. The code for checking whether a group could match an empty string was not
275 recognizing that \h, \H, \v, \V, and \R must match a character.
276
277 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
278 an empty string. If it can, pcretest shows this in its information output.
279
280 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
281 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
282 matched by pcre_exec() without using JIT:
283
284 (a) If the rest of the pattern did not match after a maximal run of
285 grapheme clusters, the code for backing up to try with fewer of them
286 did not always back up over a full grapheme when characters that do not
287 have the modifier quality were involved, e.g. Hangul syllables.
288
289 (b) If the match point in a subject started with modifier character, and
290 there was no match, the code could incorrectly back up beyond the match
291 point, and potentially beyond the first character in the subject,
292 leading to a segfault or an incorrect match result.
293
294 10. A conditional group with an assertion condition could lead to PCRE
295 recording an incorrect first data item for a match if no other first data
296 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
297 first data item, and therefore matched "ca" after "c" instead of at the
298 start.
299
300 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
301 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
302
303 12. The source of pcregrep now includes z/OS-specific code so that it can be
304 compiled for z/OS as part of the special z/OS distribution.
305
306 13. Added the -T and -TM options to pcretest.
307
308 14. The code in pcre_compile.c for creating the table of named capturing groups
309 has been refactored. Instead of creating the table dynamically during the
310 actual compiling pass, the information is remembered during the pre-compile
311 pass (on the stack unless there are more than 20 named groups, in which
312 case malloc() is used) and the whole table is created before the actual
313 compile happens. This has simplified the code (it is now nearly 150 lines
314 shorter) and prepared the way for better handling of references to groups
315 with duplicate names.
316
317 15. A back reference to a named subpattern when there is more than one of the
318 same name now checks them in the order in which they appear in the pattern.
319 The first one that is set is used for the reference. Previously only the
320 first one was inspected. This change makes PCRE more compatible with Perl.
321
322 16. Unicode character properties were updated from Unicode 6.3.0.
323
324 17. The compile-time code for auto-possessification has been refactored, based
325 on a patch by Zoltan Herczeg. It now happens after instead of during
326 compilation. The code is cleaner, and more cases are handled. The option
327 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
328 options in pcretest are provided to set it. It can also be set by
329 (*NO_AUTO_POSSESS) at the start of a pattern.
330
331 18. The character VT has been added to the default ("C" locale) set of
332 characters that match \s and are generally treated as white space,
333 following this same change in Perl 5.18. There is now no difference between
334 "Perl space" and "POSIX space". Whether VT is treated as white space in
335 other locales depends on the locale.
336
337 19. The code for checking named groups as conditions, either for being set or
338 for being recursed, has been refactored (this is related to 14 and 15
339 above). Processing unduplicated named groups should now be as fast at
340 numerical groups, and processing duplicated groups should be faster than
341 before.
342
343 20. Two patches to the CMake build system, by Alexander Barkov:
344
345 (1) Replace the "source" command by "." in CMakeLists.txt because
346 "source" is a bash-ism.
347
348 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
349 without these the CMake build does not work on Solaris.
350
351 21. Perl has changed its handling of \8 and \9. If there is no previously
352 encountered capturing group of those numbers, they are treated as the
353 literal characters 8 and 9 instead of a binary zero followed by the
354 literals. PCRE now does the same.
355
356 22. Following Perl, added \o{} to specify codepoints in octal, making it
357 possible to specify values greater than 0777 and also making them
358 unambiguous.
359
360 23. Perl now gives an error for missing closing braces after \x{... instead of
361 treating the string as literal. PCRE now does the same.
362
363 24. RunTest used to grumble if an inappropriate test was selected explicitly,
364 but just skip it when running all tests. This make it awkward to run ranges
365 of tests when one of them was inappropriate. Now it just skips any
366 inappropriate tests, as it always did when running all tests.
367
368 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
369 character types such as \d or \w, too many callouts were inserted, and the
370 data that they returned was rubbish.
371
372 26. In UCP mode, \s was not matching two of the characters that Perl matches,
373 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
374 were matched by \h. The code has now been refactored so that the lists of
375 the horizontal and vertical whitespace characters used for \h and \v (which
376 are defined only in one place) are now also used for \s.
377
378 27. Add JIT support for the 64 bit TileGX architecture.
379 Patch by Jiong Wang (Tilera Corporation).
380
381 28. Possessive quantifiers for classes (both explicit and automatically
382 generated) now use special opcodes instead of wrapping in ONCE brackets.
383
384 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
385 (because it's meaningless), this was not happening when PCRE_CASELESS was
386 set. Not wrong, but inefficient.
387
388 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
389 properties for \w, \d, etc) is present in a test regex. Otherwise if the
390 test contains no characters greater than 255, Perl doesn't realise it
391 should be using Unicode semantics.
392
393 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
394 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
395 does in Unicode mode.
396
397 32. Added the "forbid" facility to pcretest so that putting tests into the
398 wrong test files can sometimes be quickly detected.
399
400 33. There is now a limit (default 250) on the depth of nesting of parentheses.
401 This limit is imposed to control the amount of system stack used at compile
402 time. It can be changed at build time by --with-parens-nest-limit=xxx or
403 the equivalent in CMake.
404
405 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
406 errors. Perl warns for these when in warning mode, but PCRE has no facility
407 for giving warnings.
408
409 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
410 However, this was not working for (?!) because it is optimized to (*FAIL),
411 for which PCRE does not allow quantifiers. The optimization is now disabled
412 when a quantifier follows (?!). I can't see any use for this, but it makes
413 things uniform.
414
415 36. Perl no longer allows group names to start with digits, so I have made this
416 change also in PCRE. It simplifies the code a bit.
417
418 37. In extended mode, Perl ignores spaces before a + that indicates a
419 possessive quantifier. PCRE allowed a space before the quantifier, but not
420 before the possessive +. It now does.
421
422 38. The use of \K (reset reported match start) within a repeated possessive
423 group such as (a\Kb)*+ was not working.
424
425 40. Document that the same character tables must be used at compile time and
426 run time, and that the facility to pass tables to pcre_exec() and
427 pcre_dfa_exec() is for use only with saved/restored patterns.
428
429 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
430 features for Builds with MSVC:
431
432 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
433 builds). The .rc files can be used to set FileDescription and many other
434 attributes.
435
436 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
437 This allows higher-level build scripts which want .pdb files to avoid
438 hard-coding the exact files needed."
439
440 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
441 mean "start of word" and "end of word", respectively, as a transition aid.
442
443 43. A minimizing repeat of a class containing codepoints greater than 255 in
444 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
445 compiled to use the heap for recursion.
446
447 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
448 is configured.
449
450
451 Version 8.33 28-May-2013
452 ------------------------
453
454 1. Added 'U' to some constants that are compared to unsigned integers, to
455 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
456 variables that are added to signed variables, to ensure the result is
457 signed and can be negated.
458
459 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
460 CMake config files.
461
462 3. Revise the creation of config.h.generic so that all boolean macros are
463 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
464 overriding via -D on the command line possible.
465
466 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
467 to unsigned int is reported to make a quite noticeable speed difference in
468 a specific Windows environment. Testing on Linux did also appear to show
469 some benefit (and it is clearly not harmful). Also fixed the definition of
470 Xop which should be unsigned.
471
472 5. Related to (4), changing the definition of the intermediate variable cc
473 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
474 improvements.
475
476 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
477 unnecessary spaces.
478
479 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
480 and later.
481
482 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
483
484 9. Optimizing fast_forward_start_bits in JIT.
485
486 10. Adding support for callouts in JIT, and fixing some issues revealed
487 during this work. Namely:
488
489 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
490
491 (b) Minimum length was not checked before the matching is started.
492
493 11. The value of capture_last that is passed to callouts was incorrect in some
494 cases when there was a capture on one path that was subsequently abandoned
495 after a backtrack. Also, the capture_last value is now reset after a
496 recursion, since all captures are also reset in this case.
497
498 12. The interpreter no longer returns the "too many substrings" error in the
499 case when an overflowing capture is in a branch that is subsequently
500 abandoned after a backtrack.
501
502 13. In the pathological case when an offset vector of size 2 is used, pcretest
503 now prints out the matched string after a yield of 0 or 1.
504
505 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
506 Only supported by the JIT compiler at the moment.
507
508 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
509
510 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
511 offset of the starting point of the matching process, provided the offsets
512 vector is large enough.
513
514 17. The \A escape now records a lookbehind value of 1, though its execution
515 does not actually inspect the previous character. This is to ensure that,
516 in partial multi-segment matching, at least one character from the old
517 segment is retained when a new segment is processed. Otherwise, if there
518 are no lookbehinds in the pattern, \A might match incorrectly at the start
519 of a new segment.
520
521 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
522
523 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
524 gives some modest performance improvement in 8-bit mode.
525
526 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
527 be expressed in certain programming languages using Universal Character
528 Names.
529
530 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
531 which points out that "non characters" are not "characters that may not
532 appear in Unicode strings" but rather "characters that are reserved for
533 internal use and have only local meaning".
534
535 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
536 there was a conditional group that depended on an assertion, if the
537 assertion was false, the callout that immediately followed the alternation
538 in the condition was skipped when pcre_exec() was used for matching.
539
540 23. Allow an explicit callout to be inserted before an assertion that is the
541 condition for a conditional group, for compatibility with automatic
542 callouts, which always insert a callout at this point.
543
544 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
545 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
546
547 25. (*PRUNE) is now supported by the JIT compiler.
548
549 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
550
551 27. Fix the case where there are two or more SKIPs with arguments that may be
552 ignored.
553
554 28. (*SKIP) is now supported by the JIT compiler.
555
556 29. (*THEN) is now supported by the JIT compiler.
557
558 30. Update RunTest with additional test selector options.
559
560 31. The way PCRE handles backtracking verbs has been changed in two ways.
561
562 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
563 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
564 backtracking. In some cases this makes it more Perl-compatible, but Perl's
565 rather obscure rules do not always do the same thing.
566
567 (2) Previously, backtracking verbs were confined within assertions. This is
568 no longer the case for positive assertions, except for (*ACCEPT). Again,
569 this sometimes improves Perl compatibility, and sometimes does not.
570
571 32. A number of tests that were in test 2 because Perl did things differently
572 have been moved to test 1, because either Perl or PCRE has changed, and
573 these tests are now compatible.
574
575 32. Backtracking control verbs are now handled in the same way in JIT and
576 interpreter.
577
578 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
579 contained a forward subroutine reference caused a compile error.
580
581 34. Auto-detect and optimize limited repetitions in JIT.
582
583 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
584 blocking (*UTF) etc.
585
586 36. In the interpreter, maximizing pattern repetitions for characters and
587 character types now use tail recursion, which reduces stack usage.
588
589 37. The value of the max lookbehind was not correctly preserved if a compiled
590 and saved regex was reloaded on a host of different endianness.
591
592 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
593 of the compiled pattern block, expand the flags field from 16 to 32 bits
594 because it was almost full.
595
596 39. Try madvise first before posix_madvise.
597
598 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
599 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
600 original change locked out all matches of empty strings. This has been
601 changed so that one match of an empty string per line is recognized.
602 Subsequent searches on the same line (for colouring or for --only-matching,
603 for example) do not recognize empty strings.
604
605 41. Applied a user patch to fix a number of spelling mistakes in comments.
606
607 42. Data lines longer than 65536 caused pcretest to crash.
608
609 43. Clarified the data type for length and startoffset arguments for pcre_exec
610 and pcre_dfa_exec in the function-specific man pages, where they were
611 explicitly stated to be in bytes, never having been updated. I also added
612 some clarification to the pcreapi man page.
613
614 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
615 a segmentation fault.
616
617
618 Version 8.32 30-November-2012
619 -----------------------------
620
621 1. Improved JIT compiler optimizations for first character search and single
622 character iterators.
623
624 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
625 Patch by Daniel Richard G.
626
627 3. Single character iterator optimizations in the JIT compiler.
628
629 4. Improved JIT compiler optimizations for character ranges.
630
631 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
632 Reported by Giuseppe D'Angelo.
633
634 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
635 of a line, was being set incorrectly in cases where .* appeared inside
636 atomic brackets at the start of a pattern, or where there was a subsequent
637 *PRUNE or *SKIP.
638
639 7. Improved instruction cache flush for POWER/PowerPC.
640 Patch by Daniel Richard G.
641
642 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
643 grep:
644
645 (a) There is now no limit to the number of patterns to be matched.
646
647 (b) An error is given if a pattern is too long.
648
649 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
650 are now supported.
651
652 (d) --exclude-from and --include-from (multiple use) have been added.
653
654 (e) Exclusions and inclusions now apply to all files and directories, not
655 just to those obtained from scanning a directory recursively.
656
657 (f) Multiple uses of -f and --file-list are now supported.
658
659 (g) In a Windows environment, the default for -d has been changed from
660 "read" (the GNU grep default) to "skip", because otherwise the presence
661 of a directory in the file list provokes an error.
662
663 (h) The documentation has been revised and clarified in places.
664
665 9. Improve the matching speed of capturing brackets.
666
667 10. Changed the meaning of \X so that it now matches a Unicode extended
668 grapheme cluster.
669
670 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
671 out POSIX threads when JIT support is configured.
672
673 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
674
675 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
676 structure could go wrong in environments where size_t is not the same size
677 as int.
678
679 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
680 set.
681
682 15. The EBCDIC support had decayed; later updates to the code had included
683 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
684 general tidy up of EBCDIC-related issues, and the documentation was also
685 not quite right. There is now a test that can be run on ASCII systems to
686 check some of the EBCDIC-related things (but is it not a full test).
687
688 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
689 in a small tidy to the code.
690
691 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
692
693 18. If the --only-matching (-o) option in pcregrep is specified multiple
694 times, each one causes appropriate output. For example, -o1 -o2 outputs the
695 substrings matched by the 1st and 2nd capturing parentheses. A separating
696 string can be specified by --om-separator (default empty).
697
698 19. Improving the first n character searches.
699
700 20. Turn case lists for horizontal and vertical white space into macros so that
701 they are defined only once.
702
703 21. This set of changes together give more compatible Unicode case-folding
704 behaviour for characters that have more than one other case when UCP
705 support is available.
706
707 (a) The Unicode property table now has offsets into a new table of sets of
708 three or more characters that are case-equivalent. The MultiStage2.py
709 script that generates these tables (the pcre_ucd.c file) now scans
710 CaseFolding.txt instead of UnicodeData.txt for character case
711 information.
712
713 (b) The code for adding characters or ranges of characters to a character
714 class has been abstracted into a generalized function that also handles
715 case-independence. In UTF-mode with UCP support, this uses the new data
716 to handle characters with more than one other case.
717
718 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
719 whose other case is greater than 256 are now correctly matched
720 caselessly. Previously, the high codepoint matched the low one, but not
721 vice versa.
722
723 (d) The processing of \h, \H, \v, and \ in character classes now makes use
724 of the new class addition function, using character lists defined as
725 macros alongside the case definitions of 20 above.
726
727 (e) Caseless back references now work with characters that have more than
728 one other case.
729
730 (f) General caseless matching of characters with more than one other case
731 is supported.
732
733 22. Unicode character properties were updated from Unicode 6.2.0
734
735 23. Improved CMake support under Windows. Patch by Daniel Richard G.
736
737 24. Add support for 32-bit character strings, and UTF-32
738
739 25. Major JIT compiler update (code refactoring and bugfixing).
740 Experimental Sparc 32 support is added.
741
742 26. Applied a modified version of Daniel Richard G's patch to create
743 pcre.h.generic and config.h.generic by "make" instead of in the
744 PrepareRelease script.
745
746 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
747 pcre_compile.c when checking for a zero character.
748
749 28. Introducing a native interface for JIT. Through this interface, the compiled
750 machine code can be directly executed. The purpose of this interface is to
751 provide fast pattern matching, so several sanity checks are not performed.
752 However, feature tests are still performed. The new interface provides
753 1.4x speedup compared to the old one.
754
755 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
756 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
757 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
758
759 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
760 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
761 the "old" RFC 2279). Instead, it ended up passing a negative length to
762 pcre_exec().
763
764 31. Add support for GCC's visibility feature to hide internal functions.
765
766 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
767 "unknown -C option" after outputting 0 or 1.
768
769 33. There is now support for generating a code coverage report for the test
770 suite in environments where gcc is the compiler and lcov is installed. This
771 is mainly for the benefit of the developers.
772
773 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
774 unaddressable using valgrind annotations, allowing valgrind to detect
775 invalid memory accesses. This is mainly for the benefit of the developers.
776
777 25. (*UTF) can now be used to start a pattern in any of the three libraries.
778
779 26. Give configure error if --enable-cpp but no C++ compiler found.
780
781
782 Version 8.31 06-July-2012
783 -------------------------
784
785 1. Fixing a wrong JIT test case and some compiler warnings.
786
787 2. Removed a bashism from the RunTest script.
788
789 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
790 to unsigned type, result still unsigned" that was given by an MS compiler
791 on encountering the code "-sizeof(xxx)".
792
793 4. Partial matching support is added to the JIT compiler.
794
795 5. Fixed several bugs concerned with partial matching of items that consist
796 of more than one character:
797
798 (a) /^(..)\1/ did not partially match "aba" because checking references was
799 done on an "all or nothing" basis. This also applied to repeated
800 references.
801
802 (b) \R did not give a hard partial match if \r was found at the end of the
803 subject.
804
805 (c) \X did not give a hard partial match after matching one or more
806 characters at the end of the subject.
807
808 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
809 a partial match for the string "\r".
810
811 (e) When newline was set to CRLF, the metacharacter "." did not recognize
812 a partial match for a CR character at the end of the subject string.
813
814 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
815 running pcretest, the text "(JIT)" added to the output whenever JIT is
816 actually used to run the match.
817
818 7. Individual JIT compile options can be set in pcretest by following -s+[+]
819 or /S+[+] with a digit between 1 and 7.
820
821 8. OP_NOT now supports any UTF character not just single-byte ones.
822
823 9. (*MARK) control verb is now supported by the JIT compiler.
824
825 10. The command "./RunTest list" lists the available tests without actually
826 running any of them. (Because I keep forgetting what they all are.)
827
828 11. Add PCRE_INFO_MAXLOOKBEHIND.
829
830 12. Applied a (slightly modified) user-supplied patch that improves performance
831 when the heap is used for recursion (compiled with --disable-stack-for-
832 recursion). Instead of malloc and free for each heap frame each time a
833 logical recursion happens, frames are retained on a chain and re-used where
834 possible. This sometimes gives as much as 30% improvement.
835
836 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
837 call.
838
839 14. As documented, (*COMMIT) is now confined to within a positive assertion.
840
841 15. It is now possible to link pcretest with libedit as an alternative to
842 libreadline.
843
844 16. (*COMMIT) control verb is now supported by the JIT compiler.
845
846 17. The Unicode data tables have been updated to Unicode 6.1.0.
847
848 18. Added --file-list option to pcregrep.
849
850 19. Added binary file support to pcregrep, including the -a, --binary-files,
851 -I, and --text options.
852
853 20. The madvise function is renamed for posix_madvise for QNX compatibility
854 reasons. Fixed by Giuseppe D'Angelo.
855
856 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
857
858 22. Changed the option for creating symbolic links for 16-bit man pages from
859 -s to -sf so that re-installing does not cause issues.
860
861 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
862
863 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
864 matches in certain environments (the workspace was not being correctly
865 retained). Also added to pcre_dfa_exec() a simple plausibility check on
866 some of the workspace data at the beginning of a restart.
867
868 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
869 was not doing so when it should - probably a typo introduced by SVN 528
870 (change 8.10/14).
871
872 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
873 \w+ when the character tables indicated that \x{c4} was a word character.
874 There were several related cases, all because the tests for doing a table
875 lookup were testing for characters less than 127 instead of 255.
876
877 27. If a pattern contains capturing parentheses that are not used in a match,
878 their slots in the ovector are set to -1. For those that are higher than
879 any matched groups, this happens at the end of processing. In the case when
880 there were back references that the ovector was too small to contain
881 (causing temporary malloc'd memory to be used during matching), and the
882 highest capturing number was not used, memory off the end of the ovector
883 was incorrectly being set to -1. (It was using the size of the temporary
884 memory instead of the true size.)
885
886 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
887 ovector size, it uses memory at the end of the block that it has got.
888
889 29. Check for an overlong MARK name and give an error at compile time. The
890 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
891
892 30. JIT compiler update.
893
894 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
895 Rill for the patch.
896
897 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
898
899 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
900
901 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
902 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
903 was enabled.
904
905 35. Improve JIT code generation for greedy plus quantifier.
906
907 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
908 1 to "aa" instead of to an empty string. The bug affected repeated groups
909 that could potentially match an empty string.
910
911 37. Optimizing single character iterators in JIT.
912
913 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
914 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
915 codepoints that are too big for the mode are faulted, and in a UTF mode,
916 disallowed codepoints are also faulted.
917
918 39. If PCRE was compiled with UTF support, in three places in the DFA
919 matcher there was code that should only have been obeyed in UTF mode, but
920 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
921 processing when bytes with values greater than 127 were present. In 16-bit
922 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
923 both cases the values are those that cannot be the first data item in a UTF
924 character. The three items that might have provoked this were recursions,
925 possessively repeated groups, and atomic groups.
926
927 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
928 and pcregrep, because some OS require shared objects to be explicitly
929 passed to ld, causing the link step to fail if they are not.
930
931 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
932 mode, patterns that started with \h* or \R* might be incorrectly matched.
933
934
935 Version 8.30 04-February-2012
936 -----------------------------
937
938 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
939 name is defined in ctype.h.
940
941 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
942 only in quite long subpatterns.
943
944 3. Removed the function pcre_info(), which has been obsolete and deprecated
945 since it was replaced by pcre_fullinfo() in February 2000.
946
947 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
948 match a (*MARK), and the match failed at the start of the subject, a
949 reference to memory before the start of the subject could occur. This bug
950 was introduced by fix 17 of release 8.21.
951
952 5. A reference to an unset group with zero minimum repetition was giving
953 totally wrong answers (in non-JavaScript-compatibility mode). For example,
954 /(another)?(\1?)test/ matched against "hello world test". This bug was
955 introduced in release 8.13.
956
957 6. Add support for 16-bit character strings (a large amount of work involving
958 many changes and refactorings).
959
960 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
961 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
962 from a file.
963
964 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
965 rounding is not applied in this particular case).
966
967 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
968 if they appear, or are escaped, in patterns.
969
970 10. Get rid of a number of -Wunused-but-set-variable warnings.
971
972 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
973 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
974 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
975 also returns the mark "x". This bug applied to capturing parentheses,
976 non-capturing parentheses, and atomic parentheses. It also applied to some
977 assertions.
978
979 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
980 information out of configure.ac instead of relying on pcre.h.generic, which
981 is not stored in the repository.
982
983 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
984 -lreadline.
985
986 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
987
988 15. Applied Graycode's patch to put the top-level frame on the stack rather
989 than the heap when not using the stack for recursion. This gives a
990 performance improvement in many cases when recursion is not deep.
991
992 16. Experimental code added to "pcretest -C" to output the stack frame size.
993
994
995 Version 8.21 12-Dec-2011
996 ------------------------
997
998 1. Updating the JIT compiler.
999
1000 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1001 are added as well.
1002
1003 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1004 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1005 calling _pcre_jit_exec. Some extra comments are added.
1006
1007 4. (*MARK) settings inside atomic groups that do not contain any capturing
1008 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1009 was introduced by change 18 for 8.20.
1010
1011 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1012 ECMA-262 standard.
1013
1014 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1015 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1016 This bug was probably introduced by change 9 of 8.13.
1017
1018 7. While fixing 6 above, I noticed that a number of other items were being
1019 incorrectly rejected as "not fixed length". This arose partly because newer
1020 opcodes had not been added to the fixed-length checking code. I have (a)
1021 corrected the bug and added tests for these items, and (b) arranged for an
1022 error to occur if an unknown opcode is encountered while checking for fixed
1023 length instead of just assuming "not fixed length". The items that were
1024 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1025 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1026 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1027
1028 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1029 being incorrectly compiled and would have given unpredicatble results.
1030
1031 9. A possessively repeated subpattern with minimum repeat count greater than
1032 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1033 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1034 the first (A) could occur when it should not.
1035
1036 10. Add a cast and remove a redundant test from the code.
1037
1038 11. JIT should use pcre_malloc/pcre_free for allocation.
1039
1040 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1041 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1042 is anything other than /usr, -L is still shown).
1043
1044 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1045
1046 14. Perl does not support \N without a following name in a [] class; PCRE now
1047 also gives an error.
1048
1049 15. If a forward reference was repeated with an upper limit of around 2000,
1050 it caused the error "internal error: overran compiling workspace". The
1051 maximum number of forward references (including repeats) was limited by the
1052 internal workspace, and dependent on the LINK_SIZE. The code has been
1053 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1054 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1055 of around 200,000 forward references. While doing this, I also speeded up
1056 the filling in of repeated forward references.
1057
1058 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1059 incorrectly expecting the subject to contain another "a" after the start.
1060
1061 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1062 in the match, the SKIP should be ignored. This was not happening; instead
1063 the SKIP was being treated as NOMATCH. For patterns such as
1064 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1065 tested.
1066
1067 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1068 now much more compatible with Perl, in particular in cases where the result
1069 is a non-match for a non-anchored pattern. For example, if
1070 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1071 "m", where previously it did not return a name. A side effect of this
1072 change is that for partial matches, the last encountered mark name is
1073 returned, as for non matches. A number of tests that were previously not
1074 Perl-compatible have been moved into the Perl-compatible test files. The
1075 refactoring has had the pleasing side effect of removing one argument from
1076 the match() function, thus reducing its stack requirements.
1077
1078 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1079 subsequent uses of /S (without +) incorrectly behaved like /S+.
1080
1081 21. Retrieve executable code size support for the JIT compiler and fixing
1082 some warnings.
1083
1084 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1085 not work when the shorter character appeared right at the end of the
1086 subject string.
1087
1088 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1089 systems.
1090
1091 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1092 output it when the /M option is used in pcretest.
1093
1094 25. The CheckMan script was not being included in the distribution. Also, added
1095 an explicit "perl" to run Perl scripts from the PrepareRelease script
1096 because this is reportedly needed in Windows.
1097
1098 26. If study data was being save in a file and studying had not found a set of
1099 "starts with" bytes for the pattern, the data written to the file (though
1100 never used) was taken from uninitialized memory and so caused valgrind to
1101 complain.
1102
1103 27. Updated RunTest.bat as provided by Sheri Pierce.
1104
1105 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1106
1107 29. Computation of memory usage for the table of capturing group names was
1108 giving an unnecessarily large value.
1109
1110
1111 Version 8.20 21-Oct-2011
1112 ------------------------
1113
1114 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1115 a POSIX class. After further experiments with Perl, which convinced me that
1116 Perl has bugs and confusions, a closing square bracket is no longer allowed
1117 in a POSIX name. This bug also affected patterns with classes that started
1118 with full stops.
1119
1120 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1121 captured substring, but while checking the failing first alternative,
1122 substring 1 is temporarily captured. If the output vector supplied to
1123 pcre_exec() was not big enough for this capture, the yield of the function
1124 was still zero ("insufficient space for captured substrings"). This cannot
1125 be totally fixed without adding another stack variable, which seems a lot
1126 of expense for a edge case. However, I have improved the situation in cases
1127 such as /(a)(b)x|abc/ matched against "abc", where the return code
1128 indicates that fewer than the maximum number of slots in the ovector have
1129 been set.
1130
1131 3. Related to (2) above: when there are more back references in a pattern than
1132 slots in the output vector, pcre_exec() uses temporary memory during
1133 matching, and copies in the captures as far as possible afterwards. It was
1134 using the entire output vector, but this conflicts with the specification
1135 that only 2/3 is used for passing back captured substrings. Now it uses
1136 only the first 2/3, for compatibility. This is, of course, another edge
1137 case.
1138
1139 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1140 main code base, and can be used by building with --enable-jit. When this is
1141 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1142 runtime --no-jit option is given.
1143
1144 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1145 ovector, the return from the function was zero, implying that there were
1146 other matches that did not fit. The correct "exactly full" value is now
1147 returned.
1148
1149 6. If a subpattern that was called recursively or as a subroutine contained
1150 (*PRUNE) or any other control that caused it to give a non-standard return,
1151 invalid errors such as "Error -26 (nested recursion at the same subject
1152 position)" or even infinite loops could occur.
1153
1154 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1155 computing the minimum length on reaching *ACCEPT, and so ended up with the
1156 wrong value of 1 rather than 0. Further investigation indicates that
1157 computing a minimum subject length in the presence of *ACCEPT is difficult
1158 (think back references, subroutine calls), and so I have changed the code
1159 so that no minimum is registered for a pattern that contains *ACCEPT.
1160
1161 8. If (*THEN) was present in the first (true) branch of a conditional group,
1162 it was not handled as intended. [But see 16 below.]
1163
1164 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1165 Sheri Pierce.
1166
1167 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1168 the first byte in a match must be "a".
1169
1170 11. Change 17 for 8.13 increased the recursion depth for patterns like
1171 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1172 pattern contains any instances of (*THEN). If it does not, the old
1173 optimizations are restored. It would be nice to do this on a per-group
1174 basis, but at the moment that is not feasible.
1175
1176 12. In some environments, the output of pcretest -C is CRLF terminated. This
1177 broke RunTest's code that checks for the link size. A single white space
1178 character after the value is now allowed for.
1179
1180 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1181 For "fr", it uses the Windows-specific input and output files.
1182
1183 14. If (*THEN) appeared in a group that was called recursively or as a
1184 subroutine, it did not work as intended. [But see next item.]
1185
1186 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1187 pattern fragments (but not containing any | characters). If A and B are
1188 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1189 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1190 D. In other words, Perl considers parentheses that do not contain any |
1191 characters to be part of a surrounding alternative, whereas PCRE was
1192 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1193 differently. PCRE now behaves in the same way as Perl, except in the case
1194 of subroutine/recursion calls such as (?1) which have in any case always
1195 been different (but PCRE had them first :-).
1196
1197 16. Related to 15 above: Perl does not treat the | in a conditional group as
1198 creating alternatives. Such a group is treated in the same way as an
1199 ordinary group without any | characters when processing (*THEN). PCRE has
1200 been changed to match Perl's behaviour.
1201
1202 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1203 RunGrepTest script failed.
1204
1205 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1206 inevitable for groups that contain captures, but it can lead to a lot of
1207 stack use in large patterns. The old behaviour has been restored for atomic
1208 groups that do not contain any capturing parentheses.
1209
1210 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1211 suppress the check for a minimum subject length at run time. (If it was
1212 given to pcre_exec() or pcre_dfa_exec() it did work.)
1213
1214 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1215 fail to work when decoding hex characters in data strings in EBCDIC
1216 environments.
1217
1218 21. It appears that in at least one Mac OS environment, the isxdigit() function
1219 is implemented as a macro that evaluates to its argument more than once,
1220 contravening the C 90 Standard (I haven't checked a later standard). There
1221 was an instance in pcretest which caused it to go wrong when processing
1222 \x{...} escapes in subject strings. The has been rewritten to avoid using
1223 things like p++ in the argument of isxdigit().
1224
1225
1226 Version 8.13 16-Aug-2011
1227 ------------------------
1228
1229 1. The Unicode data tables have been updated to Unicode 6.0.0.
1230
1231 2. Two minor typos in pcre_internal.h have been fixed.
1232
1233 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1234 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1235 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1236
1237 4. There were a number of related bugs in the code for matching backrefences
1238 caselessly in UTF-8 mode when codes for the characters concerned were
1239 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1240 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1241 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1242 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1243 2-byte code at the end of the subject (it thought there wasn't enough data
1244 left).
1245
1246 5. Comprehensive information about what went wrong is now returned by
1247 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1248 as the output vector has at least 2 elements. The offset of the start of
1249 the failing character and a reason code are placed in the vector.
1250
1251 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1252 now returned is for the first byte of the failing character, instead of the
1253 last byte inspected. This is an incompatible change, but I hope it is small
1254 enough not to be a problem. It makes the returned offset consistent with
1255 pcre_exec() and pcre_dfa_exec().
1256
1257 7. pcretest now gives a text phrase as well as the error number when
1258 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1259 failure, the offset and reason code are output.
1260
1261 8. When \R was used with a maximizing quantifier it failed to skip backwards
1262 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1263 back over a single character (\n). This seems wrong (because it treated the
1264 two characters as a single entity when going forwards), conflicts with the
1265 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1266 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1267 has been changed.
1268
1269 9. Some internal refactoring has changed the processing so that the handling
1270 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1271 time (the PCRE_DOTALL option was changed this way some time ago: version
1272 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1273 which was always a bit of a fudge. It also means that there is one less
1274 argument for the match() function, which reduces its stack requirements
1275 slightly. This change also fixes an incompatibility with Perl: the pattern
1276 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1277
1278 10. More internal refactoring has drastically reduced the number of recursive
1279 calls to match() for possessively repeated groups such as (abc)++ when
1280 using pcre_exec().
1281
1282 11. While implementing 10, a number of bugs in the handling of groups were
1283 discovered and fixed:
1284
1285 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1286 (a|)*(?1) gave a compile-time internal error.
1287 ((a|)+)+ did not notice that the outer group could match an empty string.
1288 (^a|^)+ was not marked as anchored.
1289 (.*a|.*)+ was not marked as matching at start or after a newline.
1290
1291 12. Yet more internal refactoring has removed another argument from the match()
1292 function. Special calls to this function are now indicated by setting a
1293 value in a variable in the "match data" data block.
1294
1295 13. Be more explicit in pcre_study() instead of relying on "default" for
1296 opcodes that mean there is no starting character; this means that when new
1297 ones are added and accidentally left out of pcre_study(), testing should
1298 pick them up.
1299
1300 14. The -s option of pcretest has been documented for ages as being an old
1301 synonym of -m (show memory usage). I have changed it to mean "force study
1302 for every regex", that is, assume /S for every regex. This is similar to -i
1303 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1304 using it. It makes it easier to run collections of tests with and without
1305 study enabled, and thereby test pcre_study() more easily. All the standard
1306 tests are now run with and without -s (but some patterns can be marked as
1307 "never study" - see 20 below).
1308
1309 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1310 restoration of the capturing data to the outer values was not happening
1311 correctly.
1312
1313 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1314 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1315 pattern had matched an empty string, and so incorrectly returned a no
1316 match.
1317
1318 17. There was optimizing code for the last branch of non-capturing parentheses,
1319 and also for the obeyed branch of a conditional subexpression, which used
1320 tail recursion to cut down on stack usage. Unfortunately, now that there is
1321 the possibility of (*THEN) occurring in these branches, tail recursion is
1322 no longer possible because the return has to be checked for (*THEN). These
1323 two optimizations have therefore been removed. [But see 8.20/11 above.]
1324
1325 18. If a pattern containing \R was studied, it was assumed that \R always
1326 matched two bytes, thus causing the minimum subject length to be
1327 incorrectly computed because \R can also match just one byte.
1328
1329 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1330 was incorrectly computed.
1331
1332 20. If /S is present twice on a test pattern in pcretest input, it now
1333 *disables* studying, thereby overriding the use of -s on the command line
1334 (see 14 above). This is necessary for one or two tests to keep the output
1335 identical in both cases.
1336
1337 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1338 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1339
1340 22. When an atomic group that contained a capturing parenthesis was
1341 successfully matched, but the branch in which it appeared failed, the
1342 capturing was not being forgotten if a higher numbered group was later
1343 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1344 group 1 to "a", when in fact it should be unset. This applied to multi-
1345 branched capturing and non-capturing groups, repeated or not, and also to
1346 positive assertions (capturing in negative assertions does not happen
1347 in PCRE) and also to nested atomic groups.
1348
1349 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1350 subject after a captured substring, to make it easier to tell which of a
1351 number of identical substrings has been captured.
1352
1353 24. The way atomic groups are processed by pcre_exec() has been changed so that
1354 if they are repeated, backtracking one repetition now resets captured
1355 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1356 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1357 "aaa". Previously, it would have been "a". As part of this code
1358 refactoring, the way recursive calls are handled has also been changed.
1359
1360 25. If an assertion condition captured any substrings, they were not passed
1361 back unless some other capturing happened later. For example, if
1362 (?(?=(a))a) was matched against "a", no capturing was returned.
1363
1364 26. When studying a pattern that contained subroutine calls or assertions,
1365 the code for finding the minimum length of a possible match was handling
1366 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1367 group 1 called group 2 while simultaneously a separate group 2 called group
1368 1). A stack overflow occurred in this case. I have fixed this by limiting
1369 the recursion depth to 10.
1370
1371 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1372 Fortmann. This supports explicit test numbers on the command line, and has
1373 argument validation and error reporting.
1374
1375 28. An instance of \X with an unlimited repeat could fail if at any point the
1376 first character it looked at was a mark character.
1377
1378 29. Some minor code refactoring concerning Unicode properties and scripts
1379 should reduce the stack requirement of match() slightly.
1380
1381 30. Added the '=' option to pcretest to check the setting of unused capturing
1382 slots at the end of the pattern, which are documented as being -1, but are
1383 not included in the return count.
1384
1385 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1386 compiled something random. Now it gives a compile-time error (as does
1387 Perl).
1388
1389 32. A *MARK encountered during the processing of a positive assertion is now
1390 recorded and passed back (compatible with Perl).
1391
1392 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1393 had alternative anchored branches, the search for a second match in a line
1394 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1395 matched the line "0102" twice. The same bug affected patterns that started
1396 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1397 twice.
1398
1399 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1400 does, and because of capturing effects, quantifying parenthesized
1401 assertions may at times be useful. Quantifiers are now allowed for
1402 parenthesized assertions.
1403
1404 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1405
1406 36. \g was being checked for fancy things in a character class, when it should
1407 just be a literal "g".
1408
1409 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1410 appearance of a nested POSIX class supersedes an apparent external class.
1411 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1412 unescaped square brackets may also appear as part of class names. For
1413 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1414 more like Perl. (But see 8.20/1 above.)
1415
1416 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1417 was because it thought it was \N{name}, which is not supported).
1418
1419 39. Add minix to OS list not supporting the -S option in pcretest.
1420
1421 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1422 cannot analyze patterns in sufficient detail to catch mutual recursions
1423 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1424 subgroup is called recursively as a subpattern for a second time at the
1425 same position in the subject string. In previous releases this might have
1426 been caught by the recursion limit, or it might have run out of stack.
1427
1428 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1429 happen only once. PCRE was, however incorrectly giving a compile time error
1430 "recursive call could loop indefinitely" because it cannot analyze the
1431 pattern in sufficient detail. The compile time test no longer happens when
1432 PCRE is compiling a conditional subpattern, but actual runaway loops are
1433 now caught at runtime (see 40 above).
1434
1435 42. It seems that Perl allows any characters other than a closing parenthesis
1436 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1437 has been changed to be the same.
1438
1439 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1440 as not to get warnings when autogen.sh is called. Also changed
1441 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1442
1443 44. To help people who use pcregrep to scan files containing exceedingly long
1444 lines, the following changes have been made:
1445
1446 (a) The default value of the buffer size parameter has been increased from
1447 8K to 20K. (The actual buffer used is three times this size.)
1448
1449 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1450 PCRE is built.
1451
1452 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1453 to be set at run time.
1454
1455 (d) Numerical values in pcregrep options can be followed by K or M, for
1456 example --buffer-size=50K.
1457
1458 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1459 given and the return code is set to 2.
1460
1461 45. Add a pointer to the latest mark to the callout data block.
1462
1463 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1464 partial match of an empty string instead of no match. This was specific to
1465 the use of ".".
1466
1467 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1468 complete match instead of a partial match. This bug was dependent on both
1469 the PCRE_UTF8 and PCRE_DOTALL options being set.
1470
1471 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1472 starting byte set, because \b was not being ignored.
1473
1474
1475 Version 8.12 15-Jan-2011
1476 ------------------------
1477
1478 1. Fixed some typos in the markup of the man pages, and wrote a script that
1479 checks for such things as part of the documentation building process.
1480
1481 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1482 --match-limit and --recursion-limit options (added for 8.11). In
1483 particular, this made one of the standard tests fail. (The integer value
1484 went into the wrong half of a long int.)
1485
1486 3. If the --colour option was given to pcregrep with -v (invert match), it
1487 did strange things, either producing crazy output, or crashing. It should,
1488 of course, ignore a request for colour when reporting lines that do not
1489 match.
1490
1491 4. Another pcregrep bug caused similar problems if --colour was specified with
1492 -M (multiline) and the pattern match finished with a line ending.
1493
1494 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1495 matched in multiline mode, the following line was shown as part of the
1496 match. This seems wrong, so I have changed it.
1497
1498 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1499 the check for further matches in the same line (so they could be coloured)
1500 to overrun the end of the current line. If another match was found, it was
1501 incorrectly shown (and then shown again when found in the next line).
1502
1503 7. If pcregrep was compiled under Windows, there was a reference to the
1504 function pcregrep_exit() before it was defined. I am assuming this was
1505 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1506 reported by a user. I've moved the definition above the reference.
1507
1508
1509 Version 8.11 10-Dec-2010
1510 ------------------------
1511
1512 1. (*THEN) was not working properly if there were untried alternatives prior
1513 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1514 backtracked to try for "b" instead of moving to the next alternative branch
1515 at the same level (in this case, to look for "c"). The Perl documentation
1516 is clear that when (*THEN) is backtracked onto, it goes to the "next
1517 alternative in the innermost enclosing group".
1518
1519 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1520 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1521 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1522 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1523 (*THEN).
1524
1525 3. If \s appeared in a character class, it removed the VT character from
1526 the class, even if it had been included by some previous item, for example
1527 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1528 of \s, but is part of the POSIX "space" class.)
1529
1530 4. A partial match never returns an empty string (because you can always
1531 match an empty string at the end of the subject); however the checking for
1532 an empty string was starting at the "start of match" point. This has been
1533 changed to the "earliest inspected character" point, because the returned
1534 data for a partial match starts at this character. This means that, for
1535 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1536 (previously it gave "no match").
1537
1538 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1539 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1540 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1541 has an implication that the given string is incomplete (because a partial
1542 match is preferred over a full match). For this reason, these items now
1543 give a partial match in this situation. [Aside: previously, the one case
1544 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1545 match rather than a full match, which was wrong by the old rules, but is
1546 now correct.]
1547
1548 6. There was a bug in the handling of #-introduced comments, recognized when
1549 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1550 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1551 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1552 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1553 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1554 places in pcre_compile().
1555
1556 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1557 comments when looking ahead for named forward references to subpatterns,
1558 the only newline sequence it recognized was NL. It now handles newlines
1559 according to the set newline convention.
1560
1561 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1562 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1563 cater for a lack of strerror(). These oversights have been fixed.
1564
1565 9. Added --match-limit and --recursion-limit to pcregrep.
1566
1567 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1568
1569 11. When the -o option was used, pcregrep was setting a return code of 1, even
1570 when matches were found, and --line-buffered was not being honoured.
1571
1572 12. Added an optional parentheses number to the -o and --only-matching options
1573 of pcregrep.
1574
1575 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1576 can match an empty string. The code to do it in pcretest and pcredemo
1577 needed fixing:
1578
1579 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1580 only one byte after an empty string match just before CRLF (this case
1581 just got forgotten; "any" and "anycrlf" were OK).
1582
1583 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1584 mode when an empty string match preceded an ASCII character followed by
1585 a non-ASCII character. (The code for advancing by one character rather
1586 than one byte was nonsense.)
1587
1588 (c) The pcredemo.c sample program did not have any code at all to handle
1589 the cases when CRLF is a valid newline sequence.
1590
1591 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1592 as a starting offset was within the subject string. There is now a new
1593 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1594 negative or greater than the length of the string. In order to test this,
1595 pcretest is extended to allow the setting of negative starting offsets.
1596
1597 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1598 starting offset points to the beginning of a UTF-8 character was
1599 unnecessarily clumsy. I tidied it up.
1600
1601 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1602 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1603
1604 17. Nobody had reported that the --include_dir option, which was added in
1605 release 7.7 should have been called --include-dir (hyphen, not underscore)
1606 for compatibility with GNU grep. I have changed it to --include-dir, but
1607 left --include_dir as an undocumented synonym, and the same for
1608 --exclude-dir, though that is not available in GNU grep, at least as of
1609 release 2.5.4.
1610
1611 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1612 characters from a string of bytes) have been redefined so as not to use
1613 loops, in order to improve performance in some environments. At the same
1614 time, I abstracted some of the common code into auxiliary macros to save
1615 repetition (this should not affect the compiled code).
1616
1617 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1618 compile-time error is now given if \c is not followed by an ASCII
1619 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1620 different, and any byte value is allowed.)
1621
1622 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1623 START_OPTIMIZE option, which is now allowed at compile time - but just
1624 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1625 to pcregrep and other applications that have no direct access to PCRE
1626 options. The new /Y option in pcretest sets this option when calling
1627 pcre_compile().
1628
1629 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1630 back references. Groups containing recursive back references were forced to
1631 be atomic by that change, but in the case of named groups, the amount of
1632 memory required was incorrectly computed, leading to "Failed: internal
1633 error: code overflow". This has been fixed.
1634
1635 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1636 pcretest.c, to avoid build problems in some Borland environments.
1637
1638
1639 Version 8.10 25-Jun-2010
1640 ------------------------
1641
1642 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1643 THEN.
1644
1645 2. (*ACCEPT) was not working when inside an atomic group.
1646
1647 3. Inside a character class, \B is treated as a literal by default, but
1648 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1649 causes the error). The code is unchanged, but I tidied the documentation.
1650
1651 4. Inside a character class, PCRE always treated \R and \X as literals,
1652 whereas Perl faults them if its -w option is set. I have changed PCRE so
1653 that it faults them when PCRE_EXTRA is set.
1654
1655 5. Added support for \N, which always matches any character other than
1656 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1657
1658 6. When compiling pcregrep with newer versions of gcc which may have
1659 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1660 declared with attribute warn_unused_result" were given. Just casting the
1661 result to (void) does not stop the warnings; a more elaborate fudge is
1662 needed. I've used a macro to implement this.
1663
1664 7. Minor change to pcretest.c to avoid a compiler warning.
1665
1666 8. Added four artifical Unicode properties to help with an option to make
1667 \s etc use properties (see next item). The new properties are: Xan
1668 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1669
1670 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1671 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1672 this option. Modified pcretest to add /W to test this facility. Added
1673 REG_UCP to make it available via the POSIX interface.
1674
1675 10. Added --line-buffered to pcregrep.
1676
1677 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1678 studied, and the match started with a letter with a code point greater than
1679 127 whose first byte was different to the first byte of the other case of
1680 the letter, the other case of this starting letter was not recognized
1681 (#976).
1682
1683 12. If a pattern that was studied started with a repeated Unicode property
1684 test, for example, \p{Nd}+, there was the theoretical possibility of
1685 setting up an incorrect bitmap of starting bytes, but fortunately it could
1686 not have actually happened in practice until change 8 above was made (it
1687 added property types that matched character-matching opcodes).
1688
1689 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1690 possible starting bytes for non-anchored patterns.
1691
1692 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1693 \R, and also a number of cases that involve Unicode properties, both
1694 explicit and implicit when PCRE_UCP is set.
1695
1696 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1697 input, it could crash or give wrong results if characters with values
1698 greater than 0xc0 were present in the subject string. (Detail: it assumed
1699 UTF-8 input when processing these items.)
1700
1701 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1702 size_t is 64-bit (#991).
1703
1704 17. Added a check for running out of memory when PCRE is compiled with
1705 --disable-stack-for-recursion (#990).
1706
1707 18. If the last data line in a file for pcretest does not have a newline on
1708 the end, a newline was missing in the output.
1709
1710 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1711 less than 128) in its various bitmaps. However, there is a facility for
1712 generating tables according to the current locale when PCRE is compiled. It
1713 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1714 characters, are recognized by isspace() and therefore were getting set in
1715 these tables, and indeed these tables seem to approximate to ISO 8859. This
1716 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1717 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1718 which of course cannot start UTF-8 characters. I have changed the code so
1719 that only real ASCII characters (less than 128) and the correct starting
1720 bytes for UTF-8 encodings are set for characters greater than 127 when in
1721 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1722 altogether.)
1723
1724 20. Added the /T option to pcretest so as to be able to run tests with non-
1725 standard character tables, thus making it possible to include the tests
1726 used for 19 above in the standard set of tests.
1727
1728 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1729 reference to a subpattern the other side of a comment that contains an
1730 opening parenthesis caused either an internal compiling error, or a
1731 reference to the wrong subpattern.
1732
1733
1734 Version 8.02 19-Mar-2010
1735 ------------------------
1736
1737 1. The Unicode data tables have been updated to Unicode 5.2.0.
1738
1739 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1740 configured.
1741
1742 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1743 original author of that file, following a query about its status.
1744
1745 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1746 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1747
1748 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1749 quantifier applied to a forward-referencing subroutine call, could compile
1750 incorrect code or give the error "internal error: previously-checked
1751 referenced subpattern not found".
1752
1753 6. Both MS Visual Studio and Symbian OS have problems with initializing
1754 variables to point to external functions. For these systems, therefore,
1755 pcre_malloc etc. are now initialized to local functions that call the
1756 relevant global functions.
1757
1758 7. There were two entries missing in the vectors called coptable and poptable
1759 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1760 I've fixed the data, and added a kludgy way of testing at compile time that
1761 the lengths are correct (equal to the number of opcodes).
1762
1763 8. Following on from 7, I added a similar kludge to check the length of the
1764 eint vector in pcreposix.c.
1765
1766 9. Error texts for pcre_compile() are held as one long string to avoid too
1767 much relocation at load time. To find a text, the string is searched,
1768 counting zeros. There was no check for running off the end of the string,
1769 which could happen if a new error number was added without updating the
1770 string.
1771
1772 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1773
1774 11. \K was not working if it appeared in an atomic group or in a group that
1775 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1776 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1777 the assertion is positive, but not if it is negative.
1778
1779 12. Change 11 fortuitously reduced the size of the stack frame used in the
1780 "match()" function of pcre_exec.c by one pointer. Forthcoming
1781 implementation of support for (*MARK) will need an extra pointer on the
1782 stack; I have reserved it now, so that the stack frame size does not
1783 decrease.
1784
1785 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1786 item in branch that calls a recursion is a subroutine call - as in the
1787 second branch in the above example - was incorrectly given the compile-
1788 time error "recursive call could loop indefinitely" because pcre_compile()
1789 was not correctly checking the subroutine for matching a non-empty string.
1790
1791 14. The checks for overrunning compiling workspace could trigger after an
1792 overrun had occurred. This is a "should never occur" error, but it can be
1793 triggered by pathological patterns such as hundreds of nested parentheses.
1794 The checks now trigger 100 bytes before the end of the workspace.
1795
1796 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1797
1798
1799 Version 8.01 19-Jan-2010
1800 ------------------------
1801
1802 1. If a pattern contained a conditional subpattern with only one branch (in
1803 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1804 computed the wrong minimum data length (which is of course zero for such
1805 subpatterns). This could cause incorrect "no match" results.
1806
1807 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1808 the pattern is reset in the first branch, pcre_compile() failed with
1809 "internal error: code overflow at offset...". This happened only when
1810 the reset was to the original external option setting. (An optimization
1811 abstracts leading options settings into an external setting, which was the
1812 cause of this.)
1813
1814 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1815 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1816 assertion pattern did not match (meaning that the assertion was true), it
1817 was incorrectly treated as false if the SKIP had been reached during the
1818 matching. This also applied to assertions used as conditions.
1819
1820 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1821 assertion subpattern, including such a pattern used as a condition,
1822 unpredictable results occurred, instead of the error return
1823 PCRE_ERROR_DFA_UITEM.
1824
1825 5. The C++ GlobalReplace function was not working like Perl for the special
1826 situation when an empty string is matched. It now does the fancy magic
1827 stuff that is necessary.
1828
1829 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1830 removed. (These were left over from very, very early versions of PCRE.)
1831
1832 7. Some cosmetic changes to the code to make life easier when compiling it
1833 as part of something else:
1834
1835 (a) Change DEBUG to PCRE_DEBUG.
1836
1837 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1838 called "current" as "current_branch", to prevent a collision with the
1839 Linux macro when compiled as a kernel module.
1840
1841 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1842 prevent a collision with the Linux macro when compiled as a kernel
1843 module.
1844
1845 8. In pcre_compile() there are some checks for integer overflows that used to
1846 cast potentially large values to (double). This has been changed to that
1847 when building, a check for int64_t is made, and if it is found, it is used
1848 instead, thus avoiding the use of floating point arithmetic. (There is no
1849 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1850 double.
1851
1852 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1853 2005 (difference between two addresses compared to an unsigned value).
1854
1855 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1856 custom one, because of the following reported problem in Windows:
1857
1858 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1859 under Win32.
1860 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1861 therefore missing the function definition.
1862 - The compiler thus generates a "C" signature for the test function.
1863 - The linker fails to find the "C" function.
1864 - PCRE fails to configure if asked to do so against libbz2.
1865
1866 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1867 messages were output:
1868
1869 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1870 rerunning libtoolize, to keep the correct libtool macros in-tree.
1871 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1872
1873 I have done both of these things.
1874
1875 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1876 most of the time, it *can* run out if it is given a pattern that contains a
1877 runaway infinite recursion. I updated the discussion in the pcrestack man
1878 page.
1879
1880 13. Now that we have gone to the x.xx style of version numbers, the minor
1881 version may start with zero. Using 08 or 09 is a bad idea because users
1882 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1883 interpreted as invalid octal numbers. I've updated the previous comment in
1884 configure.ac, and also added a check that gives an error if 08 or 09 are
1885 used.
1886
1887 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1888 causing partial matching to fail when the end of the subject matched \W
1889 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1890
1891 15. There were some discrepancies between the declarations in pcre_internal.h
1892 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1893 their definitions. The declarations used "const uschar *" and the
1894 definitions used USPTR. Even though USPTR is normally defined as "const
1895 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1896 reported that: "This difference in casting confuses some C++ compilers, for
1897 example, SunCC recognizes above declarations as different functions and
1898 generates broken code for hbpcre." I have changed the declarations to use
1899 USPTR.
1900
1901 16. GNU libtool is named differently on some systems. The autogen.sh script now
1902 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1903 (FreeBSD).
1904
1905 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1906 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1907 comment: "Figure out how to create a longlong from a string: strtoll and
1908 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1909 instance, but it only takes 2 args instead of 3!"
1910
1911 18. A subtle bug concerned with back references has been fixed by a change of
1912 specification, with a corresponding code fix. A pattern such as
1913 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1914 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1915 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1916 same bug. Such groups have to be quantified to be useful, or contained
1917 inside another quantified group. (If there's no repetition, the reference
1918 can never match.) The problem arises because, having left the group and
1919 moved on to the rest of the pattern, a later failure that backtracks into
1920 the group uses the captured value from the final iteration of the group
1921 rather than the correct earlier one. I have fixed this in PCRE by forcing
1922 any group that contains a reference to itself to be an atomic group; that
1923 is, there cannot be any backtracking into it once it has completed. This is
1924 similar to recursive and subroutine calls.
1925
1926
1927 Version 8.00 19-Oct-09
1928 ----------------------
1929
1930 1. The table for translating pcre_compile() error codes into POSIX error codes
1931 was out-of-date, and there was no check on the pcre_compile() error code
1932 being within the table. This could lead to an OK return being given in
1933 error.
1934
1935 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1936 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1937 in a Windows environment.
1938
1939 3. The pcregrep --count option prints the count for each file even when it is
1940 zero, as does GNU grep. However, pcregrep was also printing all files when
1941 --files-with-matches was added. Now, when both options are given, it prints
1942 counts only for those files that have at least one match. (GNU grep just
1943 prints the file name in this circumstance, but including the count seems
1944 more useful - otherwise, why use --count?) Also ensured that the
1945 combination -clh just lists non-zero counts, with no names.
1946
1947 4. The long form of the pcregrep -F option was incorrectly implemented as
1948 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1949 but it seems right to fix it, and I didn't think it was worth preserving
1950 the old behaviour.
1951
1952 5. The command line items --regex=pattern and --regexp=pattern were not
1953 recognized by pcregrep, which required --regex pattern or --regexp pattern
1954 (with a space rather than an '='). The man page documented the '=' forms,
1955 which are compatible with GNU grep; these now work.
1956
1957 6. No libpcreposix.pc file was created for pkg-config; there was just
1958 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1959
1960 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1961 when UCP support is not needed, by modifying the Python script that
1962 generates it from Unicode data files. This should not matter if the module
1963 is correctly used as a library, but I received one complaint about 50K of
1964 unwanted data. My guess is that the person linked everything into his
1965 program rather than using a library. Anyway, it does no harm.
1966
1967 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1968 was a minimum greater than 1 for a wide character in a possessive
1969 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1970 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1971 character. Chaos in the form of incorrect output or a compiling loop could
1972 result.
1973
1974 9. The restrictions on what a pattern can contain when partial matching is
1975 requested for pcre_exec() have been removed. All patterns can now be
1976 partially matched by this function. In addition, if there are at least two
1977 slots in the offset vector, the offset of the earliest inspected character
1978 for the match and the offset of the end of the subject are set in them when
1979 PCRE_ERROR_PARTIAL is returned.
1980
1981 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1982 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1983 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1984 and may be more useful for multi-segment matching.
1985
1986 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1987 used to be given if ever the end of the subject was reached; now it is
1988 given only if matching could not proceed because another character was
1989 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1990 string "Z", which now yields "no match" instead of "partial match". In the
1991 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1992 final character ended with (*FAIL).
1993
1994 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1995 if the pattern had a "must contain" character that was already found in the
1996 earlier partial match, unless partial matching was again requested. For
1997 example, with the pattern /dog.(body)?/, the "must contain" character is
1998 "g". If the first part-match was for the string "dog", restarting with
1999 "sbody" failed. This bug has been fixed.
2000
2001 13. The string returned by pcre_dfa_exec() after a partial match has been
2002 changed so that it starts at the first inspected character rather than the
2003 first character of the match. This makes a difference only if the pattern
2004 starts with a lookbehind assertion or \b or \B (\K is not supported by
2005 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2006 matching functions compatible, and I think it's the right thing to do.
2007
2008 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2009 so that the demonstration program is easily available in environments where
2010 PCRE has not been installed from source.
2011
2012 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2013 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2014 library.
2015
2016 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2017 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2018 is not the first non-POSIX option to be added. Clearly some people find
2019 these options useful.
2020
2021 17. If a caller to the POSIX matching function regexec() passes a non-zero
2022 value for nmatch with a NULL value for pmatch, the value of
2023 nmatch is forced to zero.
2024
2025 18. RunGrepTest did not have a test for the availability of the -u option of
2026 the diff command, as RunTest does. It now checks in the same way as
2027 RunTest, and also checks for the -b option.
2028
2029 19. If an odd number of negated classes containing just a single character
2030 interposed, within parentheses, between a forward reference to a named
2031 subpattern and the definition of the subpattern, compilation crashed with
2032 an internal error, complaining that it could not find the referenced
2033 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2034 [The bug was that it was starting one character too far in when skipping
2035 over the character class, thus treating the ] as data rather than
2036 terminating the class. This meant it could skip too much.]
2037
2038 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2039 /g option in pcretest when the pattern contains \K, which makes it possible
2040 to have an empty string match not at the start, even when the pattern is
2041 anchored. Updated pcretest and pcredemo to use this option.
2042
2043 21. If the maximum number of capturing subpatterns in a recursion was greater
2044 than the maximum at the outer level, the higher number was returned, but
2045 with unset values at the outer level. The correct (outer level) value is
2046 now given.
2047
2048 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2049 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2050 make it do so. The string so far is captured, making this feature
2051 compatible with Perl.
2052
2053 23. The tests have been re-organized, adding tests 11 and 12, to make it
2054 possible to check the Perl 5.10 features against Perl 5.10.
2055
2056 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2057 pattern matches a fixed length string. PCRE did not allow this; now it
2058 does. Neither allows recursion.
2059
2060 25. I finally figured out how to implement a request to provide the minimum
2061 length of subject string that was needed in order to match a given pattern.
2062 (It was back references and recursion that I had previously got hung up
2063 on.) This code has now been added to pcre_study(); it finds a lower bound
2064 to the length of subject needed. It is not necessarily the greatest lower
2065 bound, but using it to avoid searching strings that are too short does give
2066 some useful speed-ups. The value is available to calling programs via
2067 pcre_fullinfo().
2068
2069 26. While implementing 25, I discovered to my embarrassment that pcretest had
2070 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2071 study optimizations had never been tested with that matching function.
2072 Oops. What is worse, even when it was passed study data, there was a bug in
2073 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2074 were also very few tests of studied patterns with pcre_dfa_exec().
2075
2076 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2077 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2078 on the other side of the coin, they are no longer allowed to have different
2079 names, because these cannot be distinguished in PCRE, and this has caused
2080 confusion. (This is a difference from Perl.)
2081
2082 28. When duplicate subpattern names are present (necessarily with different
2083 numbers, as required by 27 above), and a test is made by name in a
2084 conditional pattern, either for a subpattern having been matched, or for
2085 recursion in such a pattern, all the associated numbered subpatterns are
2086 tested, and the overall condition is true if the condition is true for any
2087 one of them. This is the way Perl works, and is also more like the way
2088 testing by number works.
2089
2090
2091 Version 7.9 11-Apr-09
2092 ---------------------
2093
2094 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2095 (pcretest), all targets were linked against these libraries. This included
2096 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2097 libraries. This caused unwanted dependencies to be created. This problem
2098 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2099 pcretest is linked with readline.
2100
2101 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2102 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2103 moved outside it again, because FALSE and TRUE are already defined in AIX,
2104 but BOOL is not.
2105
2106 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2107 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2108
2109 4. The pcregrep documentation said spaces were inserted as well as colons (or
2110 hyphens) following file names and line numbers when outputting matching
2111 lines. This is not true; no spaces are inserted. I have also clarified the
2112 wording for the --colour (or --color) option.
2113
2114 5. In pcregrep, when --colour was used with -o, the list of matching strings
2115 was not coloured; this is different to GNU grep, so I have changed it to be
2116 the same.
2117
2118 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2119 each matching line was coloured. Now it goes on to look for further matches
2120 of any of the test patterns, which is the same behaviour as GNU grep.
2121
2122 7. A pattern that could match an empty string could cause pcregrep to loop; it
2123 doesn't make sense to accept an empty string match in pcregrep, so I have
2124 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2125 seems to be how GNU grep behaves. [But see later change 40 for release
2126 8.33.]
2127
2128 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2129 start or after a newline", because the conditional assertion was not being
2130 correctly handled. The rule now is that both the assertion and what follows
2131 in the first alternative must satisfy the test.
2132
2133 9. If auto-callout was enabled in a pattern with a conditional group whose
2134 condition was an assertion, PCRE could crash during matching, both with
2135 pcre_exec() and pcre_dfa_exec().
2136
2137 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2138 used for matching.
2139
2140 11. Unicode property support in character classes was not working for
2141 characters (bytes) greater than 127 when not in UTF-8 mode.
2142
2143 12. Added the -M command line option to pcretest.
2144
2145 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2146
2147 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2148
2149 16. Added comments and documentation about mis-use of no_arg in the C++
2150 wrapper.
2151
2152 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2153 from Martin Jerabek that uses macro names for all relevant character and
2154 string constants.
2155
2156 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2157 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2158 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2159 these, but not everybody uses configure.
2160
2161 19. A conditional group that had only one branch was not being correctly
2162 recognized as an item that could match an empty string. This meant that an
2163 enclosing group might also not be so recognized, causing infinite looping
2164 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2165 with the subject "ab", where knowledge that the repeated group can match
2166 nothing is needed in order to break the loop.
2167
2168 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2169 exec(), but without supplying a callout function, matching went wrong.
2170
2171 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2172 leak if the size of the offset vector was greater than 30. When the vector
2173 is smaller, the saved offsets during recursion go onto a local stack
2174 vector, but for larger vectors malloc() is used. It was failing to free
2175 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2176 error, in fact).
2177
2178 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2179 heapframe that is used only when UTF-8 support is enabled. This caused no
2180 problem, but was untidy.
2181
2182 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2183 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2184 included within another project.
2185
2186 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2187 slightly modified by me:
2188
2189 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2190 not building pcregrep.
2191
2192 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2193 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2194
2195 25. Forward references, both numeric and by name, in patterns that made use of
2196 duplicate group numbers, could behave incorrectly or give incorrect errors,
2197 because when scanning forward to find the reference group, PCRE was not
2198 taking into account the duplicate group numbers. A pattern such as
2199 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2200
2201 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2202 the feature of a custom pointer more persuasive (as requested by a user).
2203
2204 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2205 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2206 pre-defined.
2207
2208 28. Added support for (*UTF8) at the start of a pattern.
2209
2210 29. Arrange for flags added by the "release type" setting in CMake to be shown
2211 in the configuration summary.
2212
2213
2214 Version 7.8 05-Sep-08
2215 ---------------------
2216
2217 1. Replaced UCP searching code with optimized version as implemented for Ad
2218 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2219 stage table and inline lookup instead of a function, giving speed ups of 2
2220 to 5 times on some simple patterns that I tested. Permission was given to
2221 distribute the MultiStage2.py script that generates the tables (it's not in
2222 the tarball, but is in the Subversion repository).
2223
2224 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2225 scripts.
2226
2227 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2228 a group with a zero qualifier. The result of the study could be incorrect,
2229 or the function might crash, depending on the pattern.
2230
2231 4. Caseless matching was not working for non-ASCII characters in back
2232 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2233 It now works when Unicode Property Support is available.
2234
2235 5. In pcretest, an escape such as \x{de} in the data was always generating
2236 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2237 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2238 truncation.
2239
2240 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2241
2242 7. Added two (int) casts to pcregrep when printing the difference of two
2243 pointers, in case they are 64-bit values.
2244
2245 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2246 test 2 if it fails.
2247
2248 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2249 and a #define of that name to empty if it is not externally set. This is to
2250 allow users of MSVC to set it if necessary.
2251
2252 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2253 the convenience functions in the pcre_get.c source file.
2254
2255 11. An option change at the start of a pattern that had top-level alternatives
2256 could cause overwriting and/or a crash. This command provoked a crash in
2257 some environments:
2258
2259 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2260
2261 This potential security problem was recorded as CVE-2008-2371.
2262
2263 12. For a pattern where the match had to start at the beginning or immediately
2264 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2265 pcre_dfa_exec() could read past the end of the passed subject if there was
2266 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2267 pcretest so that it places the subject at the end of its malloc-ed buffer.
2268
2269 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2270 exec() might read past the end of the data buffer in UTF-8 mode.
2271
2272 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2273 the data contained the byte 0x85 as part of a UTF-8 character within its
2274 first line. This applied both to normal and DFA matching.
2275
2276 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2277 /^[^d]*?$/8 failed to match "abc".
2278
2279 16. Added a missing copyright notice to pcrecpp_internal.h.
2280
2281 17. Make it more clear in the documentation that values returned from
2282 pcre_exec() in ovector are byte offsets, not character counts.
2283
2284 18. Tidied a few places to stop certain compilers from issuing warnings.
2285
2286 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2287 supplied by Stefan Weber. I made a further small update for 7.8 because
2288 there is a change of source arrangements: the pcre_searchfuncs.c module is
2289 replaced by pcre_ucd.c.
2290
2291
2292 Version 7.7 07-May-08
2293 ---------------------
2294
2295 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2296 a string to a long long, pretend we don't even have a long long." This is
2297 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2298
2299 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2300 pre-7.6 versions, which defined a global no_arg variable instead of putting
2301 it in the RE class. (See also #8 below.)
2302
2303 3. Remove a line of dead code, identified by coverity and reported by Nuno
2304 Lopes.
2305
2306 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2307
2308 (1) The include/exclude patterns were being applied to the whole pathnames
2309 of files, instead of just to the final components.
2310
2311 (2) If there was more than one level of directory, the subdirectories were
2312 skipped unless they satisfied the include/exclude conditions. This is
2313 inconsistent with GNU grep (and could even be seen as contrary to the
2314 pcregrep specification - which I improved to make it absolutely clear).
2315 The action now is always to scan all levels of directory, and just
2316 apply the include/exclude patterns to regular files.
2317
2318 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2319 --exclude_dir in the tests to avoid scanning .svn directories.
2320
2321 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2322 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2323 doesn't support NULs in patterns.
2324
2325 7. Added some missing "const"s to declarations of static tables in
2326 pcre_compile.c and pcre_dfa_exec.c.
2327
2328 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2329 caused by fix #2 above. (Subsequently also a second patch to fix the
2330 first patch. And a third patch - this was a messy problem.)
2331
2332 9. Applied Craig's patch to remove the use of push_back().
2333
2334 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2335 matching function regexec().
2336
2337 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2338 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2339 references. PCRE supports relative numbers with this syntax (I don't think
2340 Oniguruma does).
2341
2342 12. Previously, a group with a zero repeat such as (...){0} was completely
2343 omitted from the compiled regex. However, this means that if the group
2344 was called as a subroutine from elsewhere in the pattern, things went wrong
2345 (an internal error was given). Such groups are now left in the compiled
2346 pattern, with a new opcode that causes them to be skipped at execution
2347 time.
2348
2349 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2350 to the way PCRE behaves:
2351
2352 (a) A lone ] character is dis-allowed (Perl treats it as data).
2353
2354 (b) A back reference to an unmatched subpattern matches an empty string
2355 (Perl fails the current match path).
2356
2357 (c) A data ] in a character class must be notated as \] because if the
2358 first data character in a class is ], it defines an empty class. (In
2359 Perl it is not possible to have an empty class.) The empty class []
2360 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2361 The negative empty class [^] matches any one character, independently
2362 of the DOTALL setting.
2363
2364 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2365 non-existent subpattern following a character class starting with ']' and
2366 containing () gave an internal compiling error instead of "reference to
2367 non-existent subpattern". Fortunately, when the pattern did exist, the
2368 compiled code was correct. (When scanning forwards to check for the
2369 existence of the subpattern, it was treating the data ']' as terminating
2370 the class, so got the count wrong. When actually compiling, the reference
2371 was subsequently set up correctly.)
2372
2373 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2374 it was being rejected as not supported by pcre_dfa_exec(), even though
2375 other assertions are supported. I have made pcre_dfa_exec() support
2376 (*FAIL).
2377
2378 16. The implementation of 13c above involved the invention of a new opcode,
2379 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2380 cannot be changed at match time, I realized I could make a small
2381 improvement to matching performance by compiling OP_ALLANY instead of
2382 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2383 on the OP_ANY path.
2384
2385 17. Compiling pcretest on Windows with readline support failed without the
2386 following two fixes: (1) Make the unistd.h include conditional on
2387 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2388
2389 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2390 ncurses library to be included for pcretest when ReadLine support is
2391 requested, but also to allow for it to be overridden. This patch came from
2392 Daniel Bergström.
2393
2394 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2395 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2396 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2397 spotting this.
2398
2399
2400 Version 7.6 28-Jan-08
2401 ---------------------
2402
2403 1. A character class containing a very large number of characters with
2404 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2405 overflow.
2406
2407 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2408 HAVE_LONG_LONG is not defined.
2409
2410 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2411 bring them up to date and include new features. This patch includes:
2412
2413 - Fixed PH's badly added libz and libbz2 support.
2414 - Fixed a problem with static linking.
2415 - Added pcredemo. [But later removed - see 7 below.]
2416 - Fixed dftables problem and added an option.
2417 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2418 HAVE_LONG_LONG.
2419 - Added readline support for pcretest.
2420 - Added an listing of the option settings after cmake has run.
2421
2422 4. A user submitted a patch to Makefile that makes it easy to create
2423 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2424 Makefile.am that cause it to include this special target, without
2425 affecting anything else. Note that the same mingw target plus all
2426 the other distribution libraries and programs are now supported
2427 when configuring with CMake (see 6 below) instead of with
2428 Configure/Make.
2429
2430 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2431 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2432 exported in the Windows port". It has not yet been confirmed that the patch
2433 solves the problem, but it does no harm.
2434
2435 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2436 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2437 with CMake, and also correct the comment about stack recursion.
2438
2439 7. Remove the automatic building of pcredemo from the ./configure system and
2440 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2441 of a program that users should build themselves after PCRE is installed, so
2442 building it automatically is not really right. What is more, it gave
2443 trouble in some build environments.
2444
2445 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2446
2447
2448 Version 7.5 10-Jan-08
2449 ---------------------
2450
2451 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2452 values in parens when parsing an RE using the C++ wrapper."
2453
2454 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2455 Characters greater than 255 were excluded from the class instead of being
2456 included.
2457
2458 3. The same bug as (2) above applied to negated POSIX classes such as
2459 [:^space:].
2460
2461 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2462 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2463 I have changed it.
2464
2465 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2466 first named subpattern) and a construct such as (?&a) would reference the
2467 first named subpattern whose name started with "a" (in other words, the
2468 length check was missing). Both these problems are fixed. "Subpattern name
2469 expected" is now given for (?&) (a zero-length name), and this patch also
2470 makes it give the same error for \k'' (previously it complained that that
2471 was a reference to a non-existent subpattern).
2472
2473 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2474 this is right because (?- can be followed by option settings as well as by
2475 digits. I have, however, made the messages clearer.
2476
2477 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2478 than the number used in the conditional) now cause a compile-time error.
2479 This is actually not compatible with Perl, which accepts such patterns, but
2480 treats the conditional as always being FALSE (as PCRE used to), but it
2481 seems to me that giving a diagnostic is better.
2482
2483 8. Change "alphameric" to the more common word "alphanumeric" in comments
2484 and messages.
2485
2486 9. Fix two occurrences of "backslash" in comments that should have been
2487 "backspace".
2488
2489 10. Remove two redundant lines of code that can never be obeyed (their function
2490 was moved elsewhere).
2491
2492 11. The program that makes PCRE's Unicode character property table had a bug
2493 which caused it to generate incorrect table entries for sequences of
2494 characters that have the same character type, but are in different scripts.
2495 It amalgamated them into a single range, with the script of the first of
2496 them. In other words, some characters were in the wrong script. There were
2497 thirteen such cases, affecting characters in the following ranges:
2498
2499 U+002b0 - U+002c1
2500 U+0060c - U+0060d
2501 U+0061e - U+00612
2502 U+0064b - U+0065e
2503 U+0074d - U+0076d
2504 U+01800 - U+01805
2505 U+01d00 - U+01d77
2506 U+01d9b - U+01dbf
2507 U+0200b - U+0200f
2508 U+030fc - U+030fe
2509 U+03260 - U+0327f
2510 U+0fb46 - U+0fbb1
2511 U+10450 - U+1049d
2512
2513 12. The -o option (show only the matching part of a line) for pcregrep was not
2514 compatible with GNU grep in that, if there was more than one match in a
2515 line, it showed only the first of them. It now behaves in the same way as
2516 GNU grep.
2517
2518 13. If the -o and -v options were combined for pcregrep, it printed a blank
2519 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2520 does the same. The return code can be used to tell if there were any
2521 non-matching lines.
2522
2523 14. Added --file-offsets and --line-offsets to pcregrep.
2524
2525 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2526 infinitely looping recursion. The bug was that positive lookaheads were not
2527 being skipped when checking for a possible empty match (negative lookaheads
2528 and both kinds of lookbehind were skipped).
2529
2530 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2531 inclusion of <windows.h> to before rather than after the definition of
2532 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2533
2534 17. Specifying a possessive quantifier with a specific limit for a Unicode
2535 character property caused pcre_compile() to compile bad code, which led at
2536 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2537 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2538 caused the error; without that there was no problem.
2539
2540 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2541
2542 19. Added --enable-pcretest-libreadline.
2543
2544 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2545 RE::GlobalReplace(). As a result, the number of replacements returned was
2546 double what it should be. I removed one of the increments, but Craig sent a
2547 later patch that removed the other one (the right fix) and added unit tests
2548 that check the return values (which was not done before).
2549
2550 21. Several CMake things:
2551
2552 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2553 the names libpcre and libpcreposix, not just pcre and pcreposix.
2554
2555 (2) The above change means that pcretest and pcregrep are now correctly
2556 linked with the newly-built libraries, not previously installed ones.
2557
2558 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2559
2560 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2561 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2562 UTF-8 newline character). The key issue is that the pattern starts .*;
2563 this means that the match must be either at the beginning, or after a
2564 newline. The bug was in the code for advancing after a failed match and
2565 checking that the new position followed a newline. It was not taking
2566 account of UTF-8 characters correctly.
2567
2568 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2569 character classes. PCRE was not treating the sequence [:...:] as a
2570 character class unless the ... were all letters. Perl, however, seems to
2571 allow any characters between [: and :], though of course it rejects as
2572 unknown any "names" that contain non-letters, because all the known class
2573 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2574 for example, whereas PCRE did not - it did not recognize a POSIX character
2575 class. This seemed a bit dangerous, so the code has been changed to be
2576 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2577 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2578 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2579 Perl does, and where it didn't before.
2580
2581 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2582 Windows environments %n is disabled by default.
2583
2584
2585 Version 7.4 21-Sep-07
2586 ---------------------
2587
2588 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2589 means that a class such as [\s] counted as "explicit reference to CR or
2590 LF". That isn't really right - the whole point of the change was to try to
2591 help when there was an actual mention of one of the two characters. So now
2592 the change happens only if \r or \n (or a literal CR or LF) character is
2593 encountered.
2594
2595 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2596 of both had grown to the point where there were only 3 bits left.
2597 Fortunately, there was spare space in the data structure, and so I have
2598 moved the internal flags into a new 16-bit field to free up more option
2599 bits.
2600
2601 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2602 but did not set the internal JCHANGED flag - either of these is enough to
2603 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2604 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2605 start sets both bits.
2606
2607 4. Added options (at build time, compile time, exec time) to change \R from
2608 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2609
2610 5. doc/pcresyntax.html was missing from the distribution.
2611
2612 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2613 compatibility, even though it is no longer used.
2614
2615 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2616 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2617 windows.h file is present (where different names are used). [This was
2618 reversed later after testing - see 16 below.]
2619
2620 8. Changed all #include <config.h> to #include "config.h". There were also
2621 some further <pcre.h> cases that I changed to "pcre.h".
2622
2623 9. When pcregrep was used with the --colour option, it missed the line ending
2624 sequence off the lines that it output.
2625
2626 10. It was pointed out to me that arrays of string pointers cause lots of
2627 relocations when a shared library is dynamically loaded. A technique of
2628 using a single long string with a table of offsets can drastically reduce
2629 these. I have refactored PCRE in four places to do this. The result is
2630 dramatic:
2631
2632 Originally: 290
2633 After changing UCP table: 187
2634 After changing error message table: 43
2635 After changing table of "verbs" 36
2636 After changing table of Posix names 22
2637
2638 Thanks to the folks working on Gregex for glib for this insight.
2639
2640 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2641 unicode-properties was also set.
2642
2643 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2644
2645 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2646 checked only for CRLF.
2647
2648 14. Added casts to pcretest.c to avoid compiler warnings.
2649
2650 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2651
2652 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2653 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2654 entirely. This removes changes made in 7 above.
2655
2656 17. The CMake files have been updated, and there is now more information about
2657 building with CMake in the NON-UNIX-USE document.
2658
2659
2660 Version 7.3 28-Aug-07
2661 ---------------------
2662
2663 1. In the rejigging of the build system that eventually resulted in 7.1, the
2664 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2665 brackets there is not right, since it causes compilers to look for an
2666 installed pcre.h, not the version that is in the source that is being
2667 compiled (which of course may be different). I have changed it back to:
2668
2669 #include "pcre.h"
2670
2671 I have a vague recollection that the change was concerned with compiling in
2672 different directories, but in the new build system, that is taken care of
2673 by the VPATH setting the Makefile.
2674
2675 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2676 when the subject happened to end in the byte 0x85 (e.g. if the last
2677 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2678 characters but of course it shouldn't be taken as a newline when it is part
2679 of another character. The bug was that, for an unlimited repeat of . in
2680 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2681 characters when looking for a newline.
2682
2683 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2684
2685 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2686 in debug output.
2687
2688 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2689 long printing in the pcrecpp unittest when running under MinGW.
2690
2691 6. ESC_K was left out of the EBCDIC table.
2692
2693 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2694 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2695 limit also applies to "virtual nesting" when a pattern is recursive, and in
2696 this case 1000 isn't so big. I have been able to remove this limit at the
2697 expense of backing off one optimization in certain circumstances. Normally,
2698 when pcre_exec() would call its internal match() function recursively and
2699 immediately return the result unconditionally, it uses a "tail recursion"
2700 feature to save stack. However, when a subpattern that can match an empty
2701 string has an unlimited repetition quantifier, it no longer makes this
2702 optimization. That gives it a stack frame in which to save the data for
2703 checking that an empty string has been matched. Previously this was taken
2704 from the 1000-entry workspace that had been reserved. So now there is no
2705 explicit limit, but more stack is used.
2706
2707 8. Applied Daniel's patches to solve problems with the import/export magic
2708 syntax that is required for Windows, and which was going wrong for the
2709 pcreposix and pcrecpp parts of the library. These were overlooked when this
2710 problem was solved for the main library.
2711
2712 9. There were some crude static tests to avoid integer overflow when computing
2713 the size of patterns that contain repeated groups with explicit upper
2714 limits. As the maximum quantifier is 65535, the maximum group length was
2715 set at 30,000 so that the product of these two numbers did not overflow a
2716 32-bit integer. However, it turns out that people want to use groups that
2717 are longer than 30,000 bytes (though not repeat them that many times).
2718 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2719 made it possible to implement the integer overflow checks in a much more
2720 dynamic way, which I have now done. The artificial limitation on group
2721 length has been removed - we now have only the limit on the total length of
2722 the compiled pattern, which depends on the LINK_SIZE setting.
2723
2724 10. Fixed a bug in the documentation for get/copy named substring when
2725 duplicate names are permitted. If none of the named substrings are set, the
2726 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2727 empty string.
2728
2729 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2730 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2731 because the ] is interpreted as the first data character and the
2732 terminating ] is not found. PCRE has been made compatible with Perl in this
2733 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2734 cause memory overwriting.
2735
2736 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2737 string has been matched (to stop an infinite loop). It was not recognizing
2738 a conditional subpattern that could match an empty string if that
2739 subpattern was within another subpattern. For example, it looped when
2740 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2741 condition was not nested. This bug has been fixed.
2742
2743 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2744 past the start of the subject in the presence of bytes with the top bit
2745 set, for example "\x8aBCD".
2746
2747 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2748 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2749
2750 14. Optimized (?!) to (*FAIL).
2751
2752 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2753 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2754 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2755 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2756 does: it's just the validity check that is more restrictive.
2757
2758 16. Inserted checks for integer overflows during escape sequence (backslash)
2759 processing, and also fixed erroneous offset values for syntax errors during
2760 backslash processing.
2761
2762 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2763 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2764
2765 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2766 caused an overrun.
2767
2768 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2769 something other than just ASCII characters) inside a group that had an
2770 unlimited repeat caused a loop at compile time (while checking to see
2771 whether the group could match an empty string).
2772
2773 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2774 [\P{Any}] did so. (Error in the code for printing property names.)
2775
2776 21. An orphan \E inside a character class could cause a crash.
2777
2778 22. A repeated capturing bracket such as (A)? could cause a wild memory
2779 reference during compilation.
2780
2781 23. There are several functions in pcre_compile() that scan along a compiled
2782 expression for various reasons (e.g. to see if it's fixed length for look
2783 behind). There were bugs in these functions when a repeated \p or \P was
2784 present in the pattern. These operators have additional parameters compared
2785 with \d, etc, and these were not being taken into account when moving along
2786 the compiled data. Specifically:
2787
2788 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2789 length.
2790
2791 (b) An item such as \pL+ within a repeated group could cause crashes or
2792 loops.
2793
2794 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2795 "reference to non-existent subpattern" error.
2796
2797 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2798
2799 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2800 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2801
2802 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2803
2804 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2805 character were causing crashes (broken optimization).
2806
2807 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2808 \p or \P) caused a compile-time loop.
2809
2810 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2811 break. For example, the unstudied pattern [\r\n]A does not match the string
2812 "\r\nA" because change 7.0/46 below moves the current point on by two
2813 characters after failing to match at the start. However, the pattern \nA
2814 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2815 the same is true. There doesn't seem any very clean way out of this, but
2816 what I have chosen to do makes the common cases work: PCRE now takes note
2817 of whether there can be an explicit match for \r or \n anywhere in the
2818 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2819 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2820 pattern has explicit CR or LF references.
2821
2822 29. Added (*CR) etc for changing newline setting at start of pattern.
2823
2824
2825 Version 7.2 19-Jun-07
2826 ---------------------
2827
2828 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2829 which is apparently normally available under Windows.
2830
2831 2. Re-jig the pcregrep tests with different newline settings in an attempt
2832 to make them independent of the local environment's newline setting.
2833
2834 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2835
2836 4. Some of the "internals" tests were previously cut out when the link size
2837 was not 2, because the output contained actual offsets. The recent new
2838 "Z" feature of pcretest means that these can be cut out, making the tests
2839 usable with all link sizes.
2840
2841 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2842 stack recursion. This gives a massive performance boost under BSD, but just
2843 a small improvement under Linux. However, it saves one field in the frame
2844 in all cases.
2845
2846 6. Added more features from the forthcoming Perl 5.10:
2847
2848 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2849 recursion call. It refers to the nth most recently opened parentheses.
2850
2851 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2852 to be opened parentheses.
2853
2854 (c) Conditions that refer to capturing parentheses can be specified
2855 relatively, for example, (?(-2)... or (?(+3)...
2856
2857 (d) \K resets the start of the current match so that everything before
2858 is not part of it.
2859
2860 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2861
2862 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2863 reference syntax.
2864
2865 (g) (?| introduces a group in which the numbering of parentheses in each
2866 alternative starts with the same number.
2867
2868 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2869
2870 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2871 PCRE_INFO_JCHANGED.
2872
2873 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2874 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2875 for detecting groups that can match an empty string.
2876
2877 9. A pattern with a very large number of alternatives (more than several
2878 hundred) was running out of internal workspace during the pre-compile
2879 phase, where pcre_compile() figures out how much memory will be needed. A
2880 bit of new cunning has reduced the workspace needed for groups with
2881 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2882 workspace instead of running out of the 4096 that are available.
2883
2884 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2885
2886 11. Applied patch from Google to remove an optimization that didn't quite work.
2887 The report of the bug said:
2888
2889 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2890 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2891 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2892
2893 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2894 it matched the wrong number of bytes.
2895
2896
2897 Version 7.1 24-Apr-07
2898 ---------------------
2899
2900 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2901 that is more "standard", making use of automake and other Autotools. There
2902 is some re-arrangement of the files and adjustment of comments consequent
2903 on this.
2904
2905 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2906 for recursive directory scanning broke on some systems because the files
2907 are not scanned in any specific order and on different systems the order
2908 was different. A call to "sort" has been inserted into RunGrepTest for the
2909 approprate test as a short-term fix. In the longer term there may be an
2910 alternative.
2911
2912 3. I had an email from Eric Raymond about problems translating some of PCRE's
2913 man pages to HTML (despite the fact that I distribute HTML pages, some
2914 people do their own conversions for various reasons). The problems
2915 concerned the use of low-level troff macros .br and .in. I have therefore
2916 removed all such uses from the man pages (some were redundant, some could
2917 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2918 HTML has been updated to handle .nf/.fi and to complain if it encounters
2919 .br or .in.
2920
2921 4. Updated comments in configure.ac that get placed in config.h.in and also
2922 arranged for config.h to be included in the distribution, with the name
2923 config.h.generic, for the benefit of those who have to compile without
2924 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2925
2926 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2927 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2928 makevp.bat for the current PCRE, using the additional files
2929 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2930
2931 6. A Windows user reported a minor discrepancy with test 2, which turned out
2932 to be caused by a trailing space on an input line that had got lost in his
2933 copy. The trailing space was an accident, so I've just removed it.
2934
2935 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2936 that is needed.
2937
2938 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2939 as "const" (a) because they are and (b) because it helps the PHP
2940 maintainers who have recently made a script to detect big data structures
2941 in the php code that should be moved to the .rodata section. I remembered
2942 to update Builducptable as well, so it won't revert if ucptable.h is ever
2943 re-created.
2944
2945 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2946 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2947 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2948 support is not required. This saves 1.5-2K of code, which is important in
2949 some applications.
2950
2951 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2952 so as not to refer to the tables, even though these functions will never be
2953 called when UTF-8 support is disabled. Otherwise there are problems with a
2954 shared library.
2955
2956 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2957
2958 (a) It was defining its arguments as char * instead of void *.
2959
2960 (b) It was assuming that all moves were upwards in memory; this was true
2961 a long time ago when I wrote it, but is no longer the case.
2962
2963 The emulated memove() is provided for those environments that have neither
2964 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2965 is clearly not the case, as these two bugs were recently reported.
2966
2967 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2968 and Detrail to create the HTML documentation, the .txt form of the man
2969 pages, and it removes trailing spaces from listed files. It also creates
2970 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2971 case, it wraps all the #defines with #ifndefs. This script should be run
2972 before "make dist".
2973
2974 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2975 with Unicode property support.
2976
2977 (a) For a maximizing quantifier, if the two different cases of the
2978 character were of different lengths in their UTF-8 codings (there are
2979 some cases like this - I found 11), and the matching function had to
2980 back up over a mixture of the two cases, it incorrectly assumed they
2981 were both the same length.
2982
2983 (b) When PCRE was configured to use the heap rather than the stack for
2984 recursion during matching, it was not correctly preserving the data for
2985 the other case of a UTF-8 character when checking ahead for a match
2986 while processing a minimizing repeat. If the check also involved
2987 matching a wide character, but failed, corruption could cause an
2988 erroneous result when trying to check for a repeat of the original
2989 character.
2990
2991 13. Some tidying changes to the testing mechanism:
2992
2993 (a) The RunTest script now detects the internal link size and whether there
2994 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2995 values substituted by "configure". (The RunGrepTest script already did
2996 this for UTF-8.) The configure.ac script no longer substitutes the
2997 relevant variables.
2998
2999 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3000 with length and offset values. This means that the output is different
3001 for different internal link sizes. Test 2 is skipped for link sizes
3002 other than 2 because of this, bypassing the problem. Unfortunately,
3003 there was also a test in test 3 (the locale tests) that used /B and
3004 failed for link sizes other than 2. Rather than cut the whole test out,
3005 I have added a new /Z option to pcretest that replaces the length and
3006 offset values with spaces. This is now used to make test 3 independent
3007 of link size. (Test 2 will be tidied up later.)
3008
3009 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3010 segmentation fault instead of returning the appropriate error message.
3011
3012 15. In multiline mode when the newline sequence was set to "any", the pattern
3013 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3014 This doesn't seem right; it now treats the CRLF combination as the line
3015 ending, and so does not match in that case. It's only a pattern such as ^$
3016 that would hit this one: something like ^ABC$ would have failed after \r
3017 and then tried again after \r\n.
3018
3019 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3020 in an attempt to make files that differ only in their line terminators
3021 compare equal. This works on Linux.
3022
3023 17. Under certain error circumstances pcregrep might try to free random memory
3024 as it exited. This is now fixed, thanks to valgrind.
3025
3026 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3027 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3028 was because its rules for how to advance for /g after matching an empty
3029 string at the end of a line did not allow for this case. They now check for
3030 it specially.
3031
3032 20. pcretest is supposed to handle patterns and data of any length, by
3033 extending its buffers when necessary. It was getting this wrong when the
3034 buffer for a data line had to be extended.
3035
3036 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3037 CRLF as a newline sequence.
3038
3039 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3040 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3041 I have nevertheless tidied it up.
3042
3043 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3044
3045 24. Added a man page for pcre-config.
3046
3047
3048 Version 7.0 19-Dec-06
3049 ---------------------
3050
3051 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3052 moving to gcc 4.1.1.
3053
3054 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3055 sys/time.h, which is documented as needed for this function. It doesn't
3056 seem to matter on Linux, but it showed up on some releases of OS X.
3057
3058 3. It seems that there are systems where bytes whose values are greater than
3059 127 match isprint() in the "C" locale. The "C" locale should be the
3060 default when a C program starts up. In most systems, only ASCII printing
3061 characters match isprint(). This difference caused the output from pcretest
3062 to vary, making some of the tests fail. I have changed pcretest so that:
3063
3064 (a) When it is outputting text in the compiled version of a pattern, bytes
3065 other than 32-126 are always shown as hex escapes.
3066
3067 (b) When it is outputting text that is a matched part of a subject string,
3068 it does the same, unless a different locale has been set for the match
3069 (using the /L modifier). In this case, it uses isprint() to decide.
3070
3071 4. Fixed a major bug that caused incorrect computation of the amount of memory
3072 required for a compiled pattern when options that changed within the
3073 pattern affected the logic of the preliminary scan that determines the
3074 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3075 that the computed length was too small. The symptoms of this bug were
3076 either the PCRE error "internal error: code overflow" from pcre_compile(),
3077 or a glibc crash with a message such as "pcretest: free(): invalid next
3078 size (fast)". Examples of patterns that provoked this bug (shown in
3079 pcretest format) are:
3080
3081 /(?-x: )/x
3082 /(?x)(?-x: \s*#\s*)/
3083 /((?i)[\x{c0}])/8
3084 /(?i:[\x{c0}])/8
3085
3086 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3087 is now done differently.
3088
3089 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3090 wrapper classes; (b) implement a new function in the C++ scanner that is
3091 more efficient than the old way of doing things because it avoids levels of
3092 recursion in the regex matching; (c) add a paragraph to the documentation
3093 for the FullMatch() function.
3094
3095 6. The escape sequence \n was being treated as whatever was defined as
3096 "newline". Not only was this contrary to the documentation, which states
3097 that \n is character 10 (hex 0A), but it also went horribly wrong when
3098 "newline" was defined as CRLF. This has been fixed.
3099
3100 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3101 was being set to -1 for the "end of line" case (supposedly a value that no
3102 character can have). Though this value is never used (the check for end of
3103 line is "zero bytes in current character"), it caused compiler complaints.
3104 I've changed it to 0xffffffff.
3105
3106 8. In pcre_version.c, the version string was being built by a sequence of
3107 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3108 string (as it is for production releases) called a macro with an empty
3109 argument. The C standard says the result of this is undefined. The gcc
3110 compiler treats it as an empty string (which was what was wanted) but it is
3111 reported that Visual C gives an error. The source has been hacked around to
3112 avoid this problem.
3113
3114 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3115 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3116 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3117 of them did).
3118
3119 10. Originally, pcretest opened its input and output without "b"; then I was
3120 told that "b" was needed in some environments, so it was added for release
3121 5.0 to both the input and output. (It makes no difference on Unix-like
3122 systems.) Later I was told that it is wrong for the input on Windows. I've
3123 now abstracted the modes into two macros, to make it easier to fiddle with
3124 them, and removed "b" from the input mode under Windows.
3125
3126 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3127
3128 12. Added -help and --help to pcretest as an official way of being reminded
3129 of the options.
3130
3131 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3132 and pcrecpp.cc because they annoy compilers at high warning levels.
3133
3134 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3135
3136 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3137 scripts are not C programs :-) and which was not noticed because it works
3138 on Linux.
3139
3140 16. pcretest is supposed to handle any length of pattern and data line (as one
3141 line or as a continued sequence of lines) by extending its input buffer if
3142 necessary. This feature was broken for very long pattern lines, leading to
3143 a string of junk being passed to pcre_compile() if the pattern was longer
3144 than about 50K.
3145
3146 17. I have done a major re-factoring of the way pcre_compile() computes the
3147 amount of memory needed for a compiled pattern. Previously, there was code
3148 that made a preliminary scan of the pattern in order to do this. That was
3149 OK when PCRE was new, but as the facilities have expanded, it has become
3150 harder and harder to keep it in step with the real compile phase, and there
3151 have been a number of bugs (see for example, 4 above). I have now found a
3152 cunning way of running the real compile function in a "fake" mode that
3153 enables it to compute how much memory it would need, while actually only
3154 ever using a few hundred bytes of working memory and without too many
3155 tests of the mode. This should make future maintenance and development
3156 easier. A side effect of this work is that the limit of 200 on the nesting
3157 depth of parentheses has been removed (though this was never a serious
3158 limitation, I suspect). However, there is a downside: pcre_compile() now
3159 runs more slowly than before (30% or more, depending on the pattern). I
3160 hope this isn't a big issue. There is no effect on runtime performance.
3161
3162 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3163 newline (only possible for the last line of a file) and it was a
3164 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3165
3166 19. Added additional timing features to pcretest. (1) The -tm option now times
3167 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3168 separate command line item, by a number that specifies the number of
3169 repeats to use when timing. The default is 50000; this gives better
3170 precision, but takes uncomfortably long for very large patterns.
3171
3172 20. Extended pcre_study() to be more clever in cases where a branch of a
3173 subpattern has no definite first character. For example, (a*|b*)[cd] would
3174 previously give no result from pcre_study(). Now it recognizes that the
3175 first character must be a, b, c, or d.
3176
3177 21. There was an incorrect error "recursive call could loop indefinitely" if
3178 a subpattern (or the entire pattern) that was being tested for matching an
3179 empty string contained only one non-empty item after a nested subpattern.
3180 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3181 incorrectly, because the \d was being skipped in the check.
3182
3183 22. The pcretest program now has a new pattern option /B and a command line
3184 option -b, which is equivalent to adding /B to every pattern. This causes
3185 it to show the compiled bytecode, without the additional information that
3186 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3187 is the same as /B/I).
3188
3189 23. A new optimization is now able automatically to treat some sequences such
3190 as a*b as a*+b. More specifically, if something simple (such as a character
3191 or a simple class like \d) has an unlimited quantifier, and is followed by
3192 something that cannot possibly match the quantified thing, the quantifier
3193 is automatically "possessified".
3194
3195 24. A recursive reference to a subpattern whose number was greater than 39
3196 went wrong under certain circumstances in UTF-8 mode. This bug could also
3197 have affected the operation of pcre_study().
3198
3199 25. Realized that a little bit of performance could be had by replacing
3200 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3201
3202 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3203
3204 27. Possessive quantifiers such as a++ were previously implemented by turning
3205 them into atomic groups such as ($>a+). Now they have their own opcodes,
3206 which improves performance. This includes the automatically created ones
3207 from 23 above.
3208
3209 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3210 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3211 the first matched character to be a colon. This applied both to named and
3212 numbered groups.
3213
3214 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3215
3216 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3217 building PCRE on Windows easier, so I have included it in the distribution.
3218
3219 31. There is now a check in pcretest against a ridiculously large number being
3220 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3221 loop, the loop is abandoned.
3222
3223 32. Forward references to subpatterns in conditions such as (?(2)...) where
3224 subpattern 2 is defined later cause pcre_compile() to search forwards in
3225 the pattern for the relevant set of parentheses. This search went wrong
3226 when there were unescaped parentheses in a character class, parentheses
3227 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3228
3229 33. "Subroutine" calls and backreferences were previously restricted to
3230 referencing subpatterns earlier in the regex. This restriction has now
3231 been removed.
3232
3233 34. Added a number of extra features that are going to be in Perl 5.10. On the
3234 whole, these are just syntactic alternatives for features that PCRE had
3235 previously implemented using the Python syntax or my own invention. The
3236 other formats are all retained for compatibility.
3237
3238 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3239 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3240 also .NET compatible.
3241
3242 (b) A recursion or subroutine call to a named group can now be defined as
3243 (?&name) as well as (?P>name).
3244
3245 (c) A backreference to a named group can now be defined as \k<name> or
3246 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3247 5.10, are also .NET compatible.
3248
3249 (d) A conditional reference to a named group can now use the syntax
3250 (?(<name>) or (?('name') as well as (?(name).
3251
3252 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3253 groups (named and numbered) that are never evaluated inline, but can be
3254 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3255 is always false. There may be only one alternative in such a group.
3256
3257 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3258 as the simple (?(R). The condition is true only if the most recent
3259 recursion is that of the given number or name. It does not search out
3260 through the entire recursion stack.
3261
3262 (g) The escape \gN or \g{N} has been added, where N is a positive or
3263 negative number, specifying an absolute or relative reference.
3264
3265 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3266 some "unreachable code" warnings.
3267
3268 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3269 things, this adds five new scripts.
3270
3271 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3272 There were also incompatibilities regarding the handling of \Q..\E inside
3273 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3274 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3275
3276 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3277 matches an empty string, and forcibly breaks the loop. There were bugs in
3278 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3279 against aaaa the result was just "a" rather than "aaaa", for example. Two
3280 separate and independent bugs (that affected different cases) have been
3281 fixed.
3282
3283 39. Refactored the code to abolish the use of different opcodes for small
3284 capturing bracket numbers. This is a tidy that I avoided doing when I
3285 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3286 The new approach is not only tidier, it makes it possible to reduce the
3287 memory needed to fix the previous bug (38).
3288
3289 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3290 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3291 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3292 mode.
3293
3294 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3295 report.
3296
3297 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3298 copy construction and assignment in the C++ wrapper.
3299
3300 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3301 couple of bugs that could have given wrong results in the "--newline=crlf"
3302 case.
3303
3304 44. Added a number of casts and did some reorganization of signed/unsigned int
3305 variables following suggestions from Dair Grant. Also renamed the variable
3306 "this" as "item" because it is a C++ keyword.
3307
3308 45. Arranged for dftables to add
3309
3310 #include "pcre_internal.h"
3311
3312 to pcre_chartables.c because without it, gcc 4.x may remove the array
3313 definition from the final binary if PCRE is built into a static library and
3314 dead code stripping is activated.
3315
3316 46. For an unanchored pattern, if a match attempt fails at the start of a
3317 newline sequence, and the newline setting is CRLF or ANY, and the next two
3318 characters are CRLF, advance by two characters instead of one.
3319
3320
3321 Version 6.7 04-Jul-06
3322 ---------------------
3323
3324 1. In order to handle tests when input lines are enormously long, pcretest has
3325 been re-factored so that it automatically extends its buffers when
3326 necessary. The code is crude, but this _is_ just a test program. The
3327 default size has been increased from 32K to 50K.
3328
3329 2. The code in pcre_study() was using the value of the re argument before
3330 testing it for NULL. (Of course, in any sensible call of the function, it
3331 won't be NULL.)
3332
3333 3. The memmove() emulation function in pcre_internal.h, which is used on
3334 systems that lack both memmove() and bcopy() - that is, hardly ever -
3335 was missing a "static" storage class specifier.
3336
3337 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3338 containing an extended class (one that cannot be represented by a bitmap
3339 because it contains high-valued characters or Unicode property items, e.g.
3340 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3341 pattern, but PCRE should not loop if you do not (it no longer does).
3342 [Detail: two cases were found: (a) a repeated subpattern containing an
3343 extended class; (b) a recursive reference to a subpattern that followed a
3344 previous extended class. It wasn't skipping over the extended class
3345 correctly when UTF-8 mode was not set.]
3346
3347 5. A negated single-character class was not being recognized as fixed-length
3348 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3349 compile error "lookbehind assertion is not fixed length".
3350
3351 6. The RunPerlTest auxiliary script was showing an unexpected difference
3352 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3353 write a Perl script that can interpret lines of an input file either as
3354 byte characters or as UTF-8, which is what "perltest" was being required to
3355 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3356 can't do is switch easily at run time between having the "use utf8;" pragma
3357 or not. In the end, I fudged it by using the RunPerlTest script to insert
3358 "use utf8;" explicitly for the UTF-8 tests.
3359
3360 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3361 the end of the subject string, contrary to the documentation and to what
3362 Perl does. This was true of both matching functions. Now it matches only at
3363 the start of the subject and immediately after *internal* newlines.
3364
3365 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3366 a pointer to an int instead of a pointer to an unsigned long int. This
3367 caused problems on 64-bit systems.
3368
3369 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3370 instance of the 'standard' template library not being so standard".
3371
3372 10. There was no check on the number of named subpatterns nor the maximum
3373 length of a subpattern name. The product of these values is used to compute
3374 the size of the memory block for a compiled pattern. By supplying a very
3375 long subpattern name and a large number of named subpatterns, the size
3376 computation could be caused to overflow. This is now prevented by limiting
3377 the length of names to 32 characters, and the number of named subpatterns
3378 to 10,000.
3379
3380 11. Subpatterns that are repeated with specific counts have to be replicated in
3381 the compiled pattern. The size of memory for this was computed from the
3382 length of the subpattern and the repeat count. The latter is limited to
3383 65535, but there was no limit on the former, meaning that integer overflow
3384 could in principle occur. The compiled length of a repeated subpattern is
3385 now limited to 30,000 bytes in order to prevent this.
3386
3387 12. Added the optional facility to have named substrings with the same name.
3388
3389 13. Added the ability to use a named substring as a condition, using the
3390 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3391 are numbers (not recommended). Forward references are permitted.
3392
3393 14. Added forward references in named backreferences (if you see what I mean).
3394
3395 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3396 pattern could run off the end of the subject. For example, the pattern
3397 "(?s)(.{1,5})"8 did this with the subject "ab".
3398
3399 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3400 PCRE_CASELESS was set when matching characters that were quantified with ?
3401 or *.
3402
3403 17. A character class other than a single negated character that had a minimum
3404 but no maximum quantifier - for example [ab]{6,} - was not handled
3405 correctly by pce_dfa_exec(). It would match only one character.
3406
3407 18. A valid (though odd) pattern that looked like a POSIX character
3408 class but used an invalid character after [ (for example [[,abc,]]) caused
3409 pcre_compile() to give the error "Failed: internal error: code overflow" or
3410 in some cases to crash with a glibc free() error. This could even happen if
3411 the pattern terminated after [[ but there just happened to be a sequence of
3412 letters, a binary zero, and a closing ] in the memory that followed.
3413
3414 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3415 over the years. Originally (before any Unicode support), just the bottom 8
3416 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3417 output from "man perlunicode" includes this:
3418
3419 The regular expression compiler produces polymorphic opcodes. That
3420 is, the pattern adapts to the data and automatically switches to
3421 the Unicode character scheme when presented with Unicode data--or
3422 instead uses a traditional byte scheme when presented with byte
3423 data.
3424
3425 Sadly, a wide octal escape does not cause a switch, and in a string with
3426 no other multibyte characters, these octal escapes are treated as before.
3427 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3428 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3429 Unicode string.
3430
3431 I have not perpetrated such confusion in PCRE. Up till now, it took just
3432 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3433 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3434 translate to the appropriate multibyte character.
3435
3436 29. Applied some refactoring to reduce the number of warnings from Microsoft
3437 and Borland compilers. This has included removing the fudge introduced
3438 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3439 a warning about an unused variable.
3440
3441 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3442 characters since release 4.0, because Perl (from release 5.004) does not.
3443 [Or at least, is documented not to: some releases seem to be in conflict
3444 with the documentation.] However, when a pattern was studied with
3445 pcre_study() and all its branches started with \s, PCRE still included VT
3446 as a possible starting character. Of course, this did no harm; it just
3447 caused an unnecessary match attempt.
3448
3449 22. Removed a now-redundant internal flag bit that recorded the fact that case
3450 dependency changed within the pattern. This was once needed for "required
3451 byte" processing, but is no longer used. This recovers a now-scarce options
3452 bit. Also moved the least significant internal flag bit to the most-
3453 significant bit of the word, which was not previously used (hangover from
3454 the days when it was an int rather than a uint) to free up another bit for
3455 the future.
3456
3457 23. Added support for CRLF line endings as well as CR and LF. As well as the
3458 default being selectable at build time, it can now be changed at runtime
3459 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3460 specify that it is scanning data with non-default line endings.
3461
3462 24. Changed the definition of CXXLINK to make it agree with the definition of
3463 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3464
3465 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3466 recursions. This makes a big different to stack usage for some patterns.
3467
3468 26. If a subpattern containing a named recursion or subroutine reference such
3469 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3470 the space required for the compiled pattern went wrong and gave too small a
3471 value. Depending on the environment, this could lead to "Failed: internal
3472 error: code overflow at offset 49" or "glibc detected double free or
3473 corruption" errors.
3474
3475 27. Applied patches from Google (a) to support the new newline modes and (b) to
3476 advance over multibyte UTF-8 characters in GlobalReplace.
3477
3478 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3479 difference for some implementation of PCRE in some Windows version.
3480
3481 29. Added some extra testing facilities to pcretest:
3482
3483 \q<number> in a data line sets the "match limit" value
3484 \Q<number> in a data line sets the "match recursion limt" value
3485 -S <number> sets the stack size, where <number> is in megabytes
3486
3487 The -S option isn't available for Windows.
3488
3489
3490 Version 6.6 06-Feb-06
3491 ---------------------
3492
3493 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3494 in pcreposix.h. I have copied the definition from pcre.h.
3495
3496 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3497 because pcre.h is no longer a built file.
3498
3499 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3500 not normally included in the compiled code.
3501
3502
3503 Version 6.5 01-Feb-06
3504 ---------------------
3505
3506 1. When using the partial match feature with pcre_dfa_exec(), it was not
3507 anchoring the second and subsequent partial matches at the new starting
3508 point. This could lead to incorrect results. For example, with the pattern
3509 /1234/, partially matching against "123" and then "a4" gave a match.
3510
3511 2. Changes to pcregrep:
3512
3513 (a) All non-match returns from pcre_exec() were being treated as failures
3514 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3515 error message is output. Some extra information is given for the
3516 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3517 probably the only errors that are likely to be caused by users (by
3518 specifying a regex that has nested indefinite repeats, for instance).
3519 If there are more than 20 of these errors, pcregrep is abandoned.
3520
3521 (b) A binary zero was treated as data while matching, but terminated the
3522 output line if it was written out. This has been fixed: binary zeroes
3523 are now no different to any other data bytes.
3524
3525 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3526 used to set a locale for matching. The --locale=xxxx long option has
3527 been added (no short equivalent) to specify a locale explicitly on the
3528 pcregrep command, overriding the environment variables.
3529
3530 (d) When -B was used with -n, some line numbers in the output were one less
3531 than they should have been.
3532
3533 (e) Added the -o (--only-matching) option.
3534
3535 (f) If -A or -C was used with -c (count only), some lines of context were
3536 accidentally printed for the final match.
3537
3538 (g) Added the -H (--with-filename) option.
3539
3540 (h) The combination of options -rh failed to suppress file names for files
3541 that were found from directory arguments.
3542
3543 (i) Added the -D (--devices) and -d (--directories) options.
3544
3545 (j) Added the -F (--fixed-strings) option.
3546
3547 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3548
3549 (l) Added the --colo(u)r option.
3550
3551 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3552 is not present by default.
3553
3554 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3555 items such as (?R) or (?1), when the recursion could match a number of
3556 alternatives. If it matched one of the alternatives, but subsequently,
3557 outside the recursion, there was a failure, the code tried to back up into
3558 the recursion. However, because of the way PCRE is implemented, this is not
3559 possible, and the result was an incorrect result from the match.
3560
3561 In order to prevent this happening, the specification of recursion has
3562 been changed so that all such subpatterns are automatically treated as
3563 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3564
3565 4. I had overlooked the fact that, in some locales, there are characters for
3566 which isalpha() is true but neither isupper() nor islower() are true. In
3567 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3568 and ordfeminine) are like this. This affected the treatment of \w and \W
3569 when they appeared in character classes, but not when they appeared outside
3570 a character class. The bit map for "word" characters is now created
3571 separately from the results of isalnum() instead of just taking it from the
3572 upper, lower, and digit maps. (Plus the underscore character, of course.)
3573
3574 5. The above bug also affected the handling of POSIX character classes such as
3575 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3576 permanent tables. Instead, the bit maps for such a class were previously
3577 created as the appropriate unions of the upper, lower, and digit bitmaps.
3578 Now they are created by subtraction from the [[:word:]] class, which has
3579 its own bitmap.
3580
3581 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3582 It is created by subtracting the vertical space characters (\x09, \x0a,
3583 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3584 subtraction was done in the overall bitmap for a character class, meaning
3585 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3586 be recognized. This bug has been fixed.
3587
3588 7. Patches from the folks at Google:
3589
3590 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3591 real life, but is still worth protecting against".
3592
3593 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3594 regular expressions".
3595
3596 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3597 have it.
3598
3599 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3600 "configure" and the latter not, in order to fix a problem somebody had
3601 with compiling the Arg class on HP-UX.
3602
3603 (e) Improve the error-handling of the C++ wrapper a little bit.
3604
3605 (f) New tests for checking recursion limiting.
3606
3607 8. The pcre_memmove() function, which is used only if the environment does not
3608 have a standard memmove() function (and is therefore rarely compiled),
3609 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3610 returning a result (though PCRE never actually uses the result).
3611
3612 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3613 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3614 returned instead of calling malloc() with an overflowing number that would
3615 most likely cause subsequent chaos.
3616
3617 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3618
3619 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3620 with this option is matched, the nmatch and pmatch options of regexec() are
3621 ignored.
3622
3623 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3624 provided in case anyone wants to the the POSIX interface with UTF-8
3625 strings.
3626
3627 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3628 C++ linking (needed for some HP-UX environments).
3629
3630 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3631 (unused parameter) and in the pcre_printint() function (omitted "default"
3632 switch label when the default is to do nothing).
3633
3634 15. Added some code to make it possible, when PCRE is compiled as a C++
3635 library, to replace subject pointers for pcre_exec() with a smart pointer
3636 class, thus making it possible to process discontinuous strings.
3637
3638 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3639 much the same function. They were added by different people who were trying
3640 to make PCRE easy to compile on non-Unix systems. It has been suggested
3641 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3642 for compiling on Windows systems. I have therefore replaced it with
3643 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3644 defaults to "extern" for C or "extern C" for C++, which works fine on
3645 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3646 SCOPE with something explicit in config.h. In addition:
3647
3648 (a) pcreposix.h still had just "extern" instead of either of these macros;
3649 I have replaced it with PCRE_DATA_SCOPE.
3650
3651 (b) Functions such as _pcre_xclass(), which are internal to the library,
3652 but external in the C sense, all had PCRE_EXPORT in their definitions.
3653 This is apparently wrong for the Windows case, so I have removed it.
3654 (It makes no difference on Unix-like systems.)
3655
3656 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3657 of recursive calls to match(). This is different to MATCH_LIMIT because
3658 that limits the total number of calls to match(), not all of which increase
3659 the depth of recursion. Limiting the recursion depth limits the amount of
3660 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3661 when PCRE is compiled, and changed at run time. A patch from Google adds
3662 this functionality to the C++ interface.
3663
3664 18. Changes to the handling of Unicode character properties:
3665
3666 (a) Updated the table to Unicode 4.1.0.
3667
3668 (b) Recognize characters that are not in the table as "Cn" (undefined).
3669
3670 (c) I revised the way the table is implemented to a much improved format
3671 which includes recognition of ranges. It now supports the ranges that
3672 are defined in UnicodeData.txt, and it also amalgamates other
3673 characters into ranges. This has reduced the number of entries in the
3674 table from around 16,000 to around 3,000, thus reducing its size
3675 considerably. I realized I did not need to use a tree structure after
3676 all - a binary chop search is just as efficient. Having reduced the
3677 number of entries, I extended their size from 6 bytes to 8 bytes to
3678 allow for more data.
3679
3680 (d) Added support for Unicode script names via properties such as \p{Han}.
3681
3682 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3683 matching that character.
3684
3685 20. When matching a repeated Unicode property with a minimum greater than zero,
3686 (for example \pL{2,}), PCRE could look past the end of the subject if it
3687 reached it while seeking the minimum number of characters. This could
3688 happen only if some of the characters were more than one byte long, because
3689 there is a check for at least the minimum number of bytes.
3690
3691 21. Refactored the implementation of \p and \P so as to be more general, to
3692 allow for more different types of property in future. This has changed the
3693 compiled form incompatibly. Anybody with saved compiled patterns that use
3694 \p or \P will have to recompile them.
3695
3696 22. Added "Any" and "L&" to the supported property types.
3697
3698 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3699 but give a compile time error if the value is greater than 0xff.
3700
3701 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3702 accidentally not being installed or uninstalled.
3703
3704 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3705 made were to insert the current release number. This seemed silly, because
3706 it made things harder for people building PCRE on systems that don't run
3707 "configure". I have turned pcre.h into a distributed file, no longer built
3708 by "configure", with the version identification directly included. There is
3709 no longer a pcre.h.in file.
3710
3711 However, this change necessitated a change to the pcre-config script as
3712 well. It is built from pcre-config.in, and one of the substitutions was the
3713 release number. I have updated configure.ac so that ./configure now finds
3714 the release number by grepping pcre.h.
3715
3716 26. Added the ability to run the tests under valgrind.
3717
3718
3719 Version 6.4 05-Sep-05
3720 ---------------------
3721
3722 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3723 "--" to be printed when multiple files were scanned, even when none of the
3724 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3725 consider it to be a bug, and have restored the previous behaviour.
3726
3727 2. A couple of code tidies to get rid of compiler warnings.
3728
3729 3. The pcretest program used to cheat by referring to symbols in the library
3730 whose names begin with _pcre_. These are internal symbols that are not
3731 really supposed to be visible externally, and in some environments it is
3732 possible to suppress them. The cheating is now confined to including
3733 certain files from the library's source, which is a bit cleaner.
3734
3735 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3736 file's purpose clearer.
3737
3738 5. Reorganized pcre_ucp_findchar().
3739
3740
3741 Version 6.3 15-Aug-05
3742 ---------------------
3743
3744 1. The file libpcre.pc.in did not have general read permission in the tarball.
3745
3746 2. There were some problems when building without C++ support:
3747
3748 (a) If C++ support was not built, "make install" and "make test" still
3749 tried to test it.
3750
3751 (b) There were problems when the value of CXX was explicitly set. Some
3752 changes have been made to try to fix these, and ...
3753
3754 (c) --disable-cpp can now be used to explicitly disable C++ support.
3755
3756 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3757 backslash in a target when C++ was disabled. This confuses some
3758 versions of "make", apparently. Using an intermediate variable solves
3759 this. (Same for CPP_LOBJ.)
3760
3761 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3762 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3763 necessary on certain architectures.
3764
3765 4. Added a setting of -export-symbols-regex to the link command to remove
3766 those symbols that are exported in the C sense, but actually are local
3767 within the library, and not documented. Their names all begin with
3768 "_pcre_". This is not a perfect job, because (a) we have to except some
3769 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3770 available (and never for static libraries). I have made a note to try to
3771 find a way round (a) in the future.
3772
3773
3774 Version 6.2 01-Aug-05
3775 ---------------------
3776
3777 1. There was no test for integer overflow of quantifier values. A construction
3778 such as {1111111111111111} would give undefined results. What is worse, if
3779 a minimum quantifier for a parenthesized subpattern overflowed and became
3780 negative, the calculation of the memory size went wrong. This could have
3781 led to memory overwriting.
3782
3783 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3784
3785 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3786 operating environments where this matters.
3787
3788 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3789 PCRE options from within the C++ wrapper.
3790
3791 5. Named capturing subpatterns were not being correctly counted when a pattern
3792 was compiled. This caused two problems: (a) If there were more than 100
3793 such subpatterns, the calculation of the memory needed for the whole
3794 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3795 back references of the form \12, where the number was greater than 9, were
3796 not recognized as back references, even though there were sufficient
3797 previous subpatterns.
3798
3799 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3800 versions of gcc, e.g. 2.95.4.
3801
3802
3803 Version 6.1 21-Jun-05
3804 ---------------------
3805
3806 1. There was one reference to the variable "posix" in pcretest.c that was not
3807 surrounded by "#if !defined NOPOSIX".
3808
3809 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3810 the cross-check on the old pcre_info() function, for the benefit of the
3811 cut-down version of PCRE that is currently imported into Exim.
3812
3813 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3814 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3815 patterns that start (?i) but I don't think that matters. The use of (?i) is
3816 just an example; this all applies to the other options as well.
3817
3818 4. Since libtool seems to echo the compile commands it is issuing, the output
3819 from "make" can be reduced a bit by putting "@" in front of each libtool
3820 compile command.
3821
3822 5. Patch from the folks at Google for configure.in to be a bit more thorough
3823 in checking for a suitable C++ installation before trying to compile the
3824 C++ stuff. This should fix a reported problem when a compiler was present,
3825 but no suitable headers.
3826
3827 6. The man pages all had just "PCRE" as their title. I have changed them to
3828 be the relevant file name. I have also arranged that these names are
3829 retained in the file doc/pcre.txt, which is a concatenation in text format
3830 of all the man pages except the little individual ones for each function.
3831
3832 7. The NON-UNIX-USE file had not been updated for the different set of source
3833 files that come with release 6. I also added a few comments about the C++
3834 wrapper.
3835
3836
3837 Version 6.0 07-Jun-05
3838 ---------------------
3839
3840 1. Some minor internal re-organization to help with my DFA experiments.
3841
3842 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3843 didn't matter for the library itself when fully configured, but did matter
3844 when compiling without UCP support, or within Exim, where the ucp files are
3845 not imported.
3846
3847 3. Refactoring of the library code to split up the various functions into
3848 different source modules. The addition of the new DFA matching code (see
3849 below) to a single monolithic source would have made it really too
3850 unwieldy, quite apart from causing all the code to be include in a
3851 statically linked application, when only some functions are used. This is
3852 relevant even without the DFA addition now that patterns can be compiled in
3853 one application and matched in another.
3854
3855 The downside of splitting up is that there have to be some external
3856 functions and data tables that are used internally in different modules of
3857 the library but which are not part of the API. These have all had their
3858 names changed to start with "_pcre_" so that they are unlikely to clash
3859 with other external names.
3860
3861 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3862 a different (DFA) algorithm. Although it is slower than the original
3863 function, it does have some advantages for certain types of matching
3864 problem.
3865
3866 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3867 including restarting after a partial match.
3868
3869 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3870 defined when compiling for Windows was sent to me. I have put it into the
3871 code, though I have no means of testing or verifying it.
3872
3873 7. Added the pcre_refcount() auxiliary function.
3874
3875 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3876 match before or at the first newline in the subject string. In pcretest,
3877 the /f option on a pattern can be used to set this.
3878
3879 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3880 would behave wrongly. This has been present in PCRE since release 4.0.
3881
3882 10. A number of changes to the pcregrep command:
3883
3884 (a) Refactored how -x works; insert ^(...)$ instead of setting
3885 PCRE_ANCHORED and checking the length, in preparation for adding
3886 something similar for -w.
3887
3888 (b) Added the -w (match as a word) option.
3889
3890 (c) Refactored the way lines are read and buffered so as to have more
3891 than one at a time available.
3892
3893 (d) Implemented a pcregrep test script.
3894
3895 (e) Added the -M (multiline match) option. This allows patterns to match
3896 over several lines of the subject. The buffering ensures that at least
3897 8K, or the rest of the document (whichever is the shorter) is available
3898 for matching (and similarly the previous 8K for lookbehind assertions).
3899
3900 (f) Changed the --help output so that it now says
3901
3902 -w, --word-regex(p)
3903
3904 instead of two lines, one with "regex" and the other with "regexp"
3905 because that confused at least one person since the short forms are the
3906 same. (This required a bit of code, as the output is generated
3907 automatically from a table. It wasn't just a text change.)
3908
3909 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3910 option but starts with a hyphen. Could be a pattern or a path name
3911 starting with a hyphen, for instance.
3912
3913 (h) "-" can be given as a file name to represent stdin.
3914
3915 (i) When file names are being printed, "(standard input)" is used for
3916 the standard input, for compatibility with GNU grep. Previously
3917 "<stdin>" was used.
3918
3919 (j) The option --label=xxx can be used to supply a name to be used for
3920 stdin when file names are being printed. There is no short form.
3921
3922 (k) Re-factored the options decoding logic because we are going to add
3923 two more options that take data. Such options can now be given in four
3924 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3925
3926 (l) Added the -A, -B, and -C options for requesting that lines of context
3927 around matches be printed.
3928
3929 (m) Added the -L option to print the names of files that do not contain
3930 any matching lines, that is, the complement of -l.
3931
3932 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3933 continue to scan other files.
3934
3935 (o) The -s option was incorrectly implemented. For compatibility with other
3936 greps, it now suppresses the error message for a non-existent or non-
3937 accessible file (but not the return code). There is a new option called
3938 -q that suppresses the output of matching lines, which was what -s was
3939 previously doing.
3940
3941 (p) Added --include and --exclude options to specify files for inclusion
3942 and exclusion when recursing.
3943
3944 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3945 Hopefully, it now does.
3946
3947 12. Missing cast in pcre_study().
3948
3949 13. Added an "uninstall" target to the makefile.
3950
3951 14. Replaced "extern" in the function prototypes in Makefile.in with
3952 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3953 world, but is set differently for Windows.
3954
3955 15. Added a second compiling function called pcre_compile2(). The only
3956 difference is that it has an extra argument, which is a pointer to an
3957 integer error code. When there is a compile-time failure, this is set
3958 non-zero, in addition to the error test pointer being set to point to an
3959 error message. The new argument may be NULL if no error number is required
3960 (but then you may as well call pcre_compile(), which is now just a
3961 wrapper). This facility is provided because some applications need a
3962 numeric error indication, but it has also enabled me to tidy up the way
3963 compile-time errors are handled in the POSIX wrapper.
3964
3965 16. Added VPATH=.libs to the makefile; this should help when building with one
3966 prefix path and installing with another. (Or so I'm told by someone who
3967 knows more about this stuff than I do.)
3968
3969 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3970 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3971 match everything, including newlines. This is not POSIX-compatible, but
3972 somebody wanted the feature. From pcretest it can be activated by using
3973 both the P and the s flags.
3974
3975 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3976
3977 19. libpcre.pc was being incorrectly installed as executable.
3978
3979 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3980 it now also looks for '\r' so that it will work unmodified on Windows.
3981
3982 21. Added Google's contributed C++ wrapper to the distribution.
3983
3984 22. Added some untidy missing memory free() calls in pcretest, to keep
3985 Electric Fence happy when testing.
3986
3987
3988
3989 Version 5.0 13-Sep-04
3990 ---------------------
3991
3992 1. Internal change: literal characters are no longer packed up into items
3993 containing multiple characters in a single byte-string. Each character
3994 is now matched using a separate opcode. However, there may be more than one
3995 byte in the character in UTF-8 mode.
3996
3997 2. The pcre_callout_block structure has two new fields: pattern_position and
3998 next_item_length. These contain the offset in the pattern to the next match
3999 item, and its length, respectively.
4000
4001 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4002 insertion of callouts before each pattern item. Added the /C option to
4003 pcretest to make use of this.
4004
4005 4. On the advice of a Windows user, the lines
4006
4007 #if defined(_WIN32) || defined(WIN32)
4008 _setmode( _fileno( stdout ), 0x8000 );
4009 #endif /* defined(_WIN32) || defined(WIN32) */
4010
4011 have been added to the source of pcretest. This apparently does useful
4012 magic in relation to line terminators.
4013
4014 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4015 for the benefit of those environments where the "b" makes a difference.
4016
4017 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4018 to know about it. I have put a hack into configure.in that adds in code
4019 to set GCC=yes if CC=icc. This seems to end up at a point in the
4020 generated configure script that is early enough to affect the setting of
4021 compiler options, which is what is needed, but I have no means of testing
4022 whether it really works. (The user who reported this had patched the
4023 generated configure script, which of course I cannot do.)
4024
4025 LATER: After change 22 below (new libtool files), the configure script
4026 seems to know about icc (and also ecc). Therefore, I have commented out
4027 this hack in configure.in.
4028
4029 7. Added support for pkg-config (2 patches were sent in).
4030
4031 8. Negated POSIX character classes that used a combination of internal tables
4032 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4033 [[:^ascii]]. Typically, they would match almost any characters. The other
4034 POSIX classes were not broken in this way.
4035
4036 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4037 to find the match, as PCRE was deluded into thinking that the match had to
4038 start at the start point or following a newline. The same bug applied to
4039 patterns with negative forward assertions or any backward assertions
4040 preceding ".*" at the start, unless the pattern required a fixed first
4041 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4042
4043 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4044 starting at the last subject character, bytes beyond the end of the subject
4045 string were read.
4046
4047 11. Renamed the variable "class" as "classbits" to make life easier for C++
4048 users. (Previously there was a macro definition, but it apparently wasn't
4049 enough.)
4050
4051 12. Added the new field "tables" to the extra data so that tables can be passed
4052 in at exec time, or the internal tables can be re-selected. This allows
4053 a compiled regex to be saved and re-used at a later time by a different
4054 program that might have everything at different addresses.
4055
4056 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4057 -R library as well as a -L library.
4058
4059 14. The debugging options of pcretest (-d on the command line or D on a
4060 pattern) showed incorrect output for anything following an extended class
4061 that contained multibyte characters and which was followed by a quantifier.
4062
4063 15. Added optional support for general category Unicode character properties
4064 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4065 support. It adds about 90K to the size of the library. The meanings of the
4066 inbuilt class escapes such as \d and \s have NOT been changed.
4067
4068 16. Updated pcredemo.c to include calls to free() to release the memory for the
4069 compiled pattern.
4070
4071 17. The generated file chartables.c was being created in the source directory
4072 instead of in the building directory. This caused the build to fail if the
4073 source directory was different from the building directory, and was
4074 read-only.
4075
4076 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4077 file. No doubt somebody will tell me if they don't make sense... Also added
4078 Dan Mooney's comments about building on OpenVMS.
4079
4080 19. Added support for partial matching via the PCRE_PARTIAL option for
4081 pcre_exec() and the \P data escape in pcretest.
4082
4083 20. Extended pcretest with 3 new pattern features:
4084
4085 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4086 write the compiled pattern to the file whose name is "rest-of-line".
4087 This is a straight binary dump of the data, with the saved pointer to
4088 the character tables forced to be NULL. The study data, if any, is
4089 written too. After writing, pcretest reads a new pattern.
4090
4091 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4092 compiled pattern from the given file. There must not be any
4093 occurrences of "<" in the file name (pretty unlikely); if there are,
4094 pcretest will instead treat the initial "<" as a pattern delimiter.
4095 After reading in the pattern, pcretest goes on to read data lines as
4096 usual.
4097
4098 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4099 and 16-bit fields in a compiled pattern, to simulate a pattern that
4100 was compiled on a host of opposite endianness.
4101
4102 21. The pcre-exec() function can now cope with patterns that were compiled on
4103 hosts of opposite endianness, with this restriction:
4104
4105 As for any compiled expression that is saved and used later, the tables
4106 pointer field cannot be preserved; the extra_data field in the arguments
4107 to pcre_exec() should be used to pass in a tables address if a value
4108 other than the default internal tables were used at compile time.
4109
4110 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4111 now diagnosed as an error. Previously, most of the time, a negative number
4112 would have been treated as zero, but if in addition "ovector" was passed as
4113 NULL, a crash could occur.
4114
4115 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4116 new versions from the libtool 1.5 distribution (the last one is a copy of
4117 a file called libtool.m4). This seems to have fixed the need to patch
4118 "configure" to support Darwin 1.3 (which I used to do). However, I still
4119 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4120 workstation).
4121
4122 24. Changed the PCRE licence to be the more standard "BSD" licence.
4123
4124
4125 Version 4.5 01-Dec-03
4126 ---------------------
4127
4128 1. There has been some re-arrangement of the code for the match() function so
4129 that it can be compiled in a version that does not call itself recursively.
4130 Instead, it keeps those local variables that need separate instances for
4131 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4132 needs to "recurse". Keeping track of where control must go is done by means
4133 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4134 hide most of the details from the main code, and operates only if
4135 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4136 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4137 operating.
4138
4139 To make it easier for callers to provide specially tailored get/free
4140 functions for this usage, two new functions, pcre_stack_malloc, and
4141 pcre_stack_free, are used. They are always called in strict stacking order,
4142 and the size of block requested is always the same.
4143
4144 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4145 PCRE has been compiled to use the stack or the heap for recursion. The
4146 -C option of pcretest uses this to show which version is compiled.
4147
4148 A new data escape \S, is added to pcretest; it causes the amounts of store
4149 obtained and freed by both kinds of malloc/free at match time to be added
4150 to the output.
4151
4152 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4153 what's available on my current Linux desktop machine.
4154
4155 3. When matching a UTF-8 string, the test for a valid string at the start has
4156 been extended. If start_offset is not zero, PCRE now checks that it points
4157 to a byte that is the start of a UTF-8 character. If not, it returns
4158 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4159 this is necessary because there may be backward assertions in the pattern.
4160 When matching the same subject several times, it may save resources to use
4161 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4162
4163 4. The code for checking the validity of UTF-8 strings has been tightened so
4164 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4165 containing "overlong sequences".
4166
4167 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4168 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4169 should have been "&", but it just so happened that all the cases this let
4170 through by mistake were picked up later in the function.
4171
4172 6. I had used a variable called "isblank" - this is a C99 function, causing
4173 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4174
4175 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4176 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4177 so the version is shown; (c) stop "make test" echoing "./RunTest".
4178
4179 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4180
4181 9. The prototype for memmove() for systems that don't have it was using
4182 size_t, but the inclusion of the header that defines size_t was later. I've
4183 moved the #includes for the C headers earlier to avoid this.
4184
4185 10. Added some adjustments to the code to make it easier to compiler on certain
4186 special systems:
4187
4188 (a) Some "const" qualifiers were missing.
4189 (b) Added the macro EXPORT before all exported functions; by default this
4190 is defined to be empty.
4191 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4192 that it reads its output file name as an argument instead of writing
4193 to the standard output and assuming this can be redirected.
4194
4195 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4196 class containing characters with values greater than 255, PCRE compilation
4197 went into a loop.
4198
4199 12. A recursive reference to a subpattern that was within another subpattern
4200 that had a minimum quantifier of zero caused PCRE to crash. For example,
4201 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4202 recursion. If the recursively-called subpattern itself had a zero repeat,
4203 that was OK.
4204
4205 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4206 buffer into which it was copied (for escape processing) was still set at
4207 1024, so long lines caused crashes.
4208
4209 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4210 "internal error: code overflow...". This applied to any character class
4211 that was followed by a possessive quantifier.
4212
4213 15. Modified the Makefile to add libpcre.la as a prerequisite for
4214 libpcreposix.la because I was told this is needed for a parallel build to
4215 work.
4216
4217 16. If a pattern that contained .* following optional items at the start was
4218 studied, the wrong optimizing data was generated, leading to matching
4219 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4220 matching string must start with a or b or c. The correct conclusion for
4221 this pattern is that a match can start with any character.
4222
4223
4224 Version 4.4 13-Aug-03
4225 ---------------------
4226
4227 1. In UTF-8 mode, a character class containing characters with values between
4228 127 and 255 was not handled correctly if the compiled pattern was studied.
4229 In fixing this, I have also improved the studying algorithm for such
4230 classes (slightly).
4231
4232 2. Three internal functions had redundant arguments passed to them. Removal
4233 might give a very teeny performance improvement.
4234
4235 3. Documentation bug: the value of the capture_top field in a callout is *one
4236 more than* the number of the hightest numbered captured substring.
4237
4238 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4239 in incorrectly linking with a previously installed version. They now link
4240 explicitly with libpcre.la.
4241
4242 5. configure.in no longer needs to recognize Cygwin specially.
4243
4244 6. A problem in pcre.in for Windows platforms is fixed.
4245
4246 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4247 pcretest, it used to include the size of the study block as part of its
4248 output. Unfortunately, the structure contains a field that has a different
4249 size on different hardware architectures. This meant that the tests that
4250 showed this size failed. As the block is currently always of a fixed size,
4251 this information isn't actually particularly useful in pcretest output, so
4252 I have just removed it.
4253
4254 8. Three pre-processor statements accidentally did not start in column 1.
4255 Sadly, there are *still* compilers around that complain, even though
4256 standard C has not required this for well over a decade. Sigh.
4257
4258 9. In pcretest, the code for checking callouts passed small integers in the
4259 callout_data field, which is a void * field. However, some picky compilers
4260 complained about the casts involved for this on 64-bit systems. Now
4261 pcretest passes the address of the small integer instead, which should get
4262 rid of the warnings.
4263
4264 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4265 both compile and run time, and gives an error if an invalid UTF-8 sequence
4266 is found. There is a option for disabling this check in cases where the
4267 string is known to be correct and/or the maximum performance is wanted.
4268
4269 11. In response to a bug report, I changed one line in Makefile.in from
4270
4271 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4272 to
4273 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4274
4275 to look similar to other lines, but I have no way of telling whether this
4276 is the right thing to do, as I do not use Windows. No doubt I'll get told
4277 if it's wrong...
4278
4279
4280 Version 4.3 21-May-03
4281 ---------------------
4282
4283 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4284 Makefile.
4285
4286 2. Some refactoring to improve the quality of the code:
4287
4288 (i) The utf8_table... variables are now declared "const".
4289
4290 (ii) The code for \cx, which used the "case flipping" table to upper case
4291 lower case letters, now just substracts 32. This is ASCII-specific,
4292 but the whole concept of \cx is ASCII-specific, so it seems
4293 reasonable.
4294
4295 (iii) PCRE was using its character types table to recognize decimal and
4296 hexadecimal digits in the pattern. This is silly, because it handles
4297 only 0-9, a-f, and A-F, but the character types table is locale-
4298 specific, which means strange things might happen. A private
4299 table is now used for this - though it costs 256 bytes, a table is
4300 much faster than multiple explicit tests. Of course, the standard
4301 character types table is still used for matching digits in subject
4302 strings against \d.
4303
4304 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4305 ending in _t are). So I've renamed it as ESC_tee.
4306
4307 3. The first argument for regexec() in the POSIX wrapper should have been
4308 defined as "const".
4309
4310 4. Changed pcretest to use malloc() for its buffers so that they can be
4311 Electric Fenced for debugging.
4312
4313 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4314 to read one or more bytes before the start of the subject string. Often this
4315 had no effect on PCRE's behaviour, but in some circumstances it could
4316 provoke a segmentation fault.
4317
4318 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4319 to try to read one or more bytes before the start of the subject string.
4320
4321 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4322 UTF-8 support could misbehave in various ways if the subject string
4323 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4324 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4325 back over UTF-8 characters.)
4326
4327
4328 Version 4.2 14-Apr-03
4329 ---------------------
4330
4331 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4332
4333 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4334 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4335 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4336 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4337 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4338 and BUILD_EXEEXT
4339 Note: automatic setting of the BUILD variables is not yet working
4340 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4341 compile-time but not at link-time
4342 [LINK]: use for linking executables only
4343 make different versions for Windows and non-Windows
4344 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4345 libraries
4346 [LINK_FOR_BUILD]: new variable
4347 [OBJEXT]: use throughout
4348 [EXEEXT]: use throughout
4349 <winshared>: new target
4350 <wininstall>: new target
4351 <dftables.o>: use native compiler
4352 <dftables>: use native linker
4353 <install>: handle Windows platform correctly
4354 <clean>: ditto
4355 <check>: ditto
4356 copy DLL to top builddir before testing
4357
4358 As part of these changes, -no-undefined was removed again. This was reported
4359 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4360 in any case.
4361
4362 3. Some tidies to get rid of compiler warnings:
4363
4364 . In the match_data structure, match_limit was an unsigned long int, whereas
4365 match_call_count was an int. I've made them both unsigned long ints.
4366
4367 . In pcretest the fact that a const uschar * doesn't automatically cast to
4368 a void * provoked a warning.
4369
4370 . Turning on some more compiler warnings threw up some "shadow" variables
4371 and a few more missing casts.
4372
4373 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4374 option, a class that contained a single character with a value between 128
4375 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4376
4377 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4378 option, a class that contained several characters, but with at least one
4379 whose value was between 128 and 255 caused PCRE to crash.
4380
4381
4382 Version 4.1 12-Mar-03
4383 ---------------------
4384
4385 1. Compiling with gcc -pedantic found a couple of places where casts were
4386 needed, and a string in dftables.c that was longer than standard compilers are
4387 required to support.
4388
4389 2. Compiling with Sun's compiler found a few more places where the code could
4390 be tidied up in order to avoid warnings.
4391
4392 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4393 first of these names is deprecated in the latest Autoconf in favour of the name
4394 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4395 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4396 analogy I have changed it to CFLAGS_FOR_BUILD.
4397
4398 4. Added -no-undefined to the linking command in the Makefile, because this is
4399 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4400 linking step for the pcreposix library.
4401
4402 5. PCRE was failing to diagnose the case of two named groups with the same
4403 name.
4404
4405 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4406 literal character that is needed in the subject for a match, and scans along to
4407 ensure that it is present before embarking on the full matching process. This
4408 saves time in cases of nested unlimited repeats that are never going to match.
4409 Problem: the scan can take a lot of time if the subject is very long (e.g.
4410 megabytes), thus penalizing straightforward matches. It is now done only if the
4411 amount of subject to be scanned is less than 1000 bytes.
4412
4413 7. A lesser problem with the same optimization is that it was recording the
4414 first character of an anchored pattern as "needed", thus provoking a search
4415 right along the subject, even when the first match of the pattern was going to
4416 fail. The "needed" character is now not set for anchored patterns, unless it
4417 follows something in the pattern that is of non-fixed length. Thus, it still
4418 fulfils its original purpose of finding quick non-matches in cases of nested
4419 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4420
4421
4422 Version 4.0 17-Feb-03
4423 ---------------------
4424
4425 1. If a comment in an extended regex that started immediately after a meta-item
4426 extended to the end of string, PCRE compiled incorrect data. This could lead to
4427 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4428
4429 2. Moved to autoconf 2.53 and libtool 1.4.2.
4430
4431 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4432 the special perltest8 script is no longer needed - all the tests can be run
4433 from a single perltest script.
4434
4435 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4436 by \s. It has now been removed in PCRE. This means it isn't recognized as
4437 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4438 class [:space:] *does* include VT, thereby creating a mess.
4439
4440 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4441 space and tab.
4442
4443 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4444 its new features into the main test script, reducing the number of scripts.
4445
4446 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4447 were backward compatible, and made the (?i) apply to the whole pattern, as if
4448 /i were given. Now it behaves more logically, and applies the option setting
4449 only to what follows. PCRE has been changed to follow suit. However, if it
4450 finds options settings right at the start of the pattern, it extracts them into
4451 the global options, as before. Thus, they show up in the info data.
4452
4453 8. Added support for the \Q...\E escape sequence. Characters in between are
4454 treated as literals. This is slightly different from Perl in that $ and @ are
4455 also handled as literals inside the quotes. In Perl, they will cause variable
4456 interpolation. Note the following examples:
4457
4458 Pattern PCRE matches Perl matches
4459
4460 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4461 \Qabc\$xyz\E abc\$xyz abc\$xyz
4462 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4463
4464 For compatibility with Perl, \Q...\E sequences are recognized inside character
4465 classes as well as outside them.
4466
4467 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4468 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4469 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4470 signed/unsigned warnings.
4471
4472 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4473 option for pcretest, so I've replaced it by a simple function that does just
4474 that job.
4475
4476 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4477 "pcregrep -".
4478
4479 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4480 Java package. This provides some syntactic sugar for simple cases of what my
4481 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4482 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4483 item, you can use this simplified notation. Note that only makes sense with
4484 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4485 greediness, whatever the setting of the PCRE_UNGREEDY option.
4486
4487 13. A change of greediness default within a pattern was not taking effect at
4488 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4489 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4490 was abstracted outside.
4491
4492 14. PCRE now supports the \G assertion. It is true when the current matching
4493 position is at the start point of the match. This differs from \A when the
4494 starting offset is non-zero. Used with the /g option of pcretest (or similar
4495 code), it works in the same way as it does for Perl's /g option. If all
4496 alternatives of a regex begin with \G, the expression is anchored to the start
4497 match position, and the "anchored" flag is set in the compiled expression.
4498
4499 15. Some bugs concerning the handling of certain option changes within patterns
4500 have been fixed. These applied to options other than (?ims). For example,
4501 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4502 been the other way round. Some of this was related to change 7 above.
4503
4504 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4505 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4506 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4507 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4508
4509 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4510 mode. Unlike ".", it always matches newline, whatever the setting of
4511 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4512 assertions. Perl allows it, but it doesn't (in general) work because it can't
4513 calculate the length of the lookbehind. At least, that's the case for Perl
4514 5.8.0 - I've been told they are going to document that it doesn't work in
4515 future.
4516
4517 18. Added an error diagnosis for escapes that PCRE does not support: these are
4518 \L, \l, \N, \P, \p, \U, \u, and \X.
4519
4520 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4521 reading past the end of the pattern in cases such as /[abcd/.
4522
4523 20. PCRE was getting more memory than necessary for patterns with classes that
4524 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4525
4526 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4527 compiling PCRE for use with Virtual Pascal.
4528
4529 22. Small fix to the Makefile to make it work properly if the build is done
4530 outside the source tree.
4531
4532 23. Added a new extension: a condition to go with recursion. If a conditional
4533 subpattern starts with (?(R) the "true" branch is used if recursion has
4534 happened, whereas the "false" branch is used only at the top level.
4535
4536 24. When there was a very long string of literal characters (over 255 bytes
4537 without UTF support, over 250 bytes with UTF support), the computation of how
4538 much memory was required could be incorrect, leading to segfaults or other
4539 strange effects.
4540
4541 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4542 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4543 there was a subsequent back reference to those brackets. This meant that, for
4544 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4545 possible to check for precisely this case. All we can do is abandon the
4546 optimization if .* occurs inside capturing brackets when there are any back
4547 references whatsoever. (See below for a better fix that came later.)
4548
4549 26. The handling of the optimization for finding the first character of a
4550 non-anchored pattern, and for finding a character that is required later in the
4551 match were failing in some cases. This didn't break the matching; it just
4552 failed to optimize when it could. The way this is done has been re-implemented.
4553
4554 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4555
4556 28. Added a new feature that provides some of the functionality that Perl
4557 provides with (?{...}). The facility is termed a "callout". The way it is done
4558 in PCRE is for the caller to provide an optional function, by setting
4559 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4560 global variable. By default it is unset, which disables all calling out. To get
4561 the function called, the regex must include (?C) at appropriate points. This
4562 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4563 This provides a means of identifying different callout points. When PCRE
4564 reaches such a point in the regex, if pcre_callout has been set, the external
4565 function is called. It is provided with data in a structure called
4566 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4567 matching continues; if it returns a non-zero value, the match at the current
4568 point fails. However, backtracking will occur if possible. [This was changed
4569 later and other features added - see item 49 below.]
4570
4571 29. pcretest is upgraded to test the callout functionality. It provides a
4572 callout function that displays information. By default, it shows the start of
4573 the match and the current position in the text. There are some new data escapes
4574 to vary what happens:
4575
4576 \C+ in addition, show current contents of captured substrings
4577 \C- do not supply a callout function
4578 \C!n return 1 when callout number n is reached
4579 \C!n!m return 1 when callout number n is reached for the mth time
4580
4581 30. If pcregrep was called with the -l option and just a single file name, it
4582 output "<stdin>" if a match was found, instead of the file name.
4583
4584 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4585 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4586 pcre_exec(). This saves a malloc/free per call. The default value of
4587 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4588 when configuring.
4589
4590 32. The default maximum size of a compiled pattern is 64K. There have been a
4591 few cases of people hitting this limit. The code now uses macros to handle the
4592 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4593 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4594 configuring. Tests 2 and 5 work only with 2-byte links because they output
4595 debugging information about compiled patterns.
4596
4597 33. Internal code re-arrangements:
4598
4599 (a) Moved the debugging function for printing out a compiled regex into
4600 its own source file (printint.c) and used #include to pull it into
4601 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4602 separate copies.
4603
4604 (b) Defined the list of op-code names for debugging as a macro in
4605 internal.h so that it is next to the definition of the opcodes.
4606
4607 (c) Defined a table of op-code lengths for simpler skipping along compiled
4608 code. This is again a macro in internal.h so that it is next to the
4609 definition of the opcodes.
4610
4611 34. Added support for recursive calls to individual subpatterns, along the
4612 lines of Robin Houston's patch (but implemented somewhat differently).
4613
4614 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4615 allow it to read and process whole directories in Win32. This code was
4616 contributed by Lionel Fourquaux; it has not been tested by me.
4617
4618 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4619 used to name a group. Names consist of alphanumerics and underscores, and must
4620 be unique. Back references use the syntax (?P=name) and recursive calls use
4621 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4622 numbers. The function pcre_fullinfo() can be used after compilation to extract
4623 a name/number map. There are three relevant calls:
4624
4625 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4626 PCRE_INFO_NAMECOUNT yields the number of entries
4627 PCRE_INFO_NAMETABLE yields a pointer to the map.
4628
4629 The map is a vector of fixed-size entries. The size of each entry depends on
4630 the length of the longest name used. The first two bytes of each entry are the
4631 group number, most significant byte first. There follows the corresponding
4632 name, zero terminated. The names are in alphabetical order.
4633
4634 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4635 case instead of 255. Making it the same both with and without UTF-8 support
4636 means that the same test output works with both.
4637
4638 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4639 calling malloc() with a zero argument.
4640
4641 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4642 optimization. I've improved things by keeping a bitmap of backreferences with
4643 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4644 fact referenced, the optimization can be applied. It is unlikely that a
4645 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4646 the match to follow \n) will appear inside brackets with a number greater than
4647 31, but if it does, any back reference > 31 suppresses the optimization.
4648
4649 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4650 of disabling numbered capturing parentheses. Any opening parenthesis that is
4651 not followed by ? behaves as if it were followed by ?: but named parentheses
4652 can still be used for capturing (and they will acquire numbers in the usual
4653 way).
4654
4655 41. Redesigned the return codes from the match() function into yes/no/error so
4656 that errors can be passed back from deep inside the nested calls. A malloc
4657 failure while inside a recursive subpattern call now causes the
4658 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4659
4660 42. It is now possible to set a limit on the number of times the match()
4661 function is called in a call to pcre_exec(). This facility makes it possible to
4662 limit the amount of recursion and backtracking, though not in a directly
4663 obvious way, because the match() function is used in a number of different
4664 circumstances. The count starts from zero for each position in the subject
4665 string (for non-anchored patterns). The default limit is, for compatibility, a
4666 large number, namely 10 000 000. You can change this in two ways:
4667
4668 (a) When configuring PCRE before making, you can use --with-match-limit=n
4669 to set a default value for the compiled library.
4670
4671 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4672 a different value is set. See 45 below.
4673
4674 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4675
4676 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4677 of things that can be changed at compile time. The first argument specifies
4678 what is wanted and the second points to where the information is to be placed.
4679 The current list of available information is:
4680
4681 PCRE_CONFIG_UTF8
4682
4683 The output is an integer that is set to one if UTF-8 support is available;
4684 otherwise it is set to zero.
4685
4686 PCRE_CONFIG_NEWLINE
4687
4688 The output is an integer that it set to the value of the code that is used for
4689 newline. It is either LF (10) or CR (13).
4690
4691 PCRE_CONFIG_LINK_SIZE
4692
4693 The output is an integer that contains the number of bytes used for internal
4694 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4695
4696 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4697
4698 The output is an integer that contains the threshold above which the POSIX
4699 interface uses malloc() for output vectors. See item 31 above.
4700
4701 PCRE_CONFIG_MATCH_LIMIT
4702
4703 The output is an unsigned integer that contains the default limit of the number
4704 of match() calls in a pcre_exec() execution. See 42 above.
4705
4706 44. pcretest has been upgraded by the addition of the -C option. This causes it
4707 to extract all the available output from the new pcre_config() function, and to
4708 output it. The program then exits immediately.
4709
4710 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4711 order to support additional features. One way would have been to define
4712 pcre_exec2() (for example) with extra arguments, but this would not have been
4713 extensible, and would also have required all calls to the original function to
4714 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4715 is used for passing in "extra" data from pcre_study().
4716
4717 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4718 contains the following fields:
4719
4720 flags a bitmap indicating which of the following fields are set
4721 study_data opaque data from pcre_study()
4722 match_limit a way of specifying a limit on match() calls for a specific
4723 call to pcre_exec()
4724 callout_data data for callouts (see 49 below)
4725
4726 The flag bits are also defined in pcre.h, and are
4727
4728 PCRE_EXTRA_STUDY_DATA
4729 PCRE_EXTRA_MATCH_LIMIT
4730 PCRE_EXTRA_CALLOUT_DATA
4731
4732 The pcre_study() function now returns one of these new pcre_extra blocks, with
4733 the actual study data pointed to by the study_data field, and the
4734 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4735 before. That is, this change is entirely upwards-compatible and requires no
4736 change to existing code.
4737
4738 If you want to pass in additional data to pcre_exec(), you can either place it
4739 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4740 block.
4741
4742 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4743 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4744 times with different match limits, until it finds the minimum value needed for
4745 pcre_exec() to complete. The value is then output. This can be instructive; for
4746 most simple matches the number is quite small, but for pathological cases it
4747 gets very large very quickly.
4748
4749 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4750 returns the size of the data block pointed to by the study_data field in a
4751 pcre_extra block, that is, the value that was passed as the argument to
4752 pcre_malloc() when PCRE was getting memory in which to place the information
4753 created by pcre_study(). The fourth argument should point to a size_t variable.
4754 pcretest has been extended so that this information is shown after a successful
4755 pcre_study() call when information about the compiled regex is being displayed.
4756
4757 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4758 because what follows is always an absolute path. (Later: it turns out that this
4759 is more than cosmetic for MinGW, because it doesn't like empty path
4760 components.)
4761
4762 49. Some changes have been made to the callout feature (see 28 above):
4763
4764 (i) A callout function now has three choices for what it returns:
4765
4766 0 => success, carry on matching
4767 > 0 => failure at this point, but backtrack if possible
4768 < 0 => serious error, return this value from pcre_exec()
4769
4770 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4771 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4772 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4773 use by callout functions. It will never be used by PCRE itself.
4774
4775 (ii) The pcre_extra structure (see 45 above) has a void * field called
4776 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4777 pcre_callout_block structure has a field of the same name. The contents of
4778 the field passed in the pcre_extra structure are passed to the callout
4779 function in the corresponding field in the callout block. This makes it
4780 easier to use the same callout-containing regex from multiple threads. For
4781 testing, the pcretest program has a new data escape
4782
4783 \C*n pass the number n (may be negative) as callout_data
4784
4785 If the callout function in pcretest receives a non-zero value as
4786 callout_data, it returns that value.
4787
4788 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4789 there were some redundant $(CFLAGS) in commands that are now specified as
4790 $(LINK), which already includes $(CFLAGS).
4791
4792 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4793 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4794 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4795 one-byte characters throughout. Note that case-insensitive matching applies
4796 only to characters whose values are less than 256. PCRE doesn't support the
4797 notion of cases for higher-valued characters.
4798
4799 (i) A character class whose characters are all within 0-255 is handled as
4800 a bit map, and the map is inverted for negative classes. Previously, a
4801 character > 255 always failed to match such a class; however it should
4802 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4803
4804 (ii) A negated character class with a single character < 255 is coded as
4805 "not this character" (OP_NOT). This wasn't working properly when the test
4806 character was multibyte, either singly or repeated.
4807
4808 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4809 mode, for example: \x{100}{2,3}.
4810
4811 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4812 singly or repeated) now correctly test multibyte characters. However,
4813 PCRE doesn't recognize any characters with values greater than 255 as
4814 digits, spaces, or word characters. Such characters always match \D, \S,
4815 and \W, and never match \d, \s, or \w.
4816
4817 (v) Classes may now contain characters and character ranges with values
4818 greater than 255. For example: [ab\x{100}-\x{400}].
4819
4820 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4821 PCRE in UTF-8 mode.
4822
4823 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4824 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4825 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4826 value.)
4827
4828 53. The single man page has become too large. I have therefore split it up into
4829 a number of separate man pages. These also give rise to individual HTML pages;
4830 these are now put in a separate directory, and there is an index.html page that
4831 lists them all. Some hyperlinking between the pages has been installed.
4832
4833 54. Added convenience functions for handling named capturing parentheses.
4834
4835 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4836 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4837 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4838 are faulted.
4839
4840 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4841 calling configure. These values are used when compiling the dftables.c program
4842 which is run to generate the source of the default character tables. They
4843 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4844 you will need to set these values.
4845
4846 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4847
4848
4849 Version 3.9 02-Jan-02
4850 ---------------------
4851
4852 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4853
4854 2. If --disable-static was given, the building process failed when trying to
4855 build pcretest and pcregrep. (For some reason it was using libtool to compile
4856 them, which is not right, as they aren't part of the library.)
4857
4858
4859 Version 3.8 18-Dec-01
4860 ---------------------
4861
4862 1. The experimental UTF-8 code was completely screwed up. It was packing the
4863 bytes in the wrong order. How dumb can you get?
4864
4865
4866 Version 3.7 29-Oct-01
4867 ---------------------
4868
4869 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4870 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4871 this didn't happen under Solaris 8, where I normally test things.
4872
4873 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4874 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4875 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4876 this fix an hour or so after the initial 3.7 release.)
4877
4878
4879 Version 3.6 23-Oct-01
4880 ---------------------
4881
4882 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4883 offsets passed as NULL with zero offset count.
4884
4885 2. The config.guess and config.sub files had not been updated when I moved to
4886 the latest autoconf.
4887
4888
4889 Version 3.5 15-Aug-01
4890 ---------------------
4891
4892 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4893 had been forgotten.
4894
4895 2. By using declared but undefined structures, we can avoid using "void"
4896 definitions in pcre.h while keeping the internal definitions of the structures
4897 private.
4898
4899 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4900 user point of view, this means that both static and shared libraries are built
4901 by default, but this can be individually controlled. More of the work of
4902 handling this static/shared cases is now inside libtool instead of PCRE's make
4903 file.
4904
4905 4. The pcretest utility is now installed along with pcregrep because it is
4906 useful for users (to test regexs) and by doing this, it automatically gets
4907 relinked by libtool. The documentation has been turned into a man page, so
4908 there are now .1, .txt, and .html versions in /doc.
4909
4910 5. Upgrades to pcregrep:
4911 (i) Added long-form option names like gnu grep.
4912 (ii) Added --help to list all options with an explanatory phrase.
4913 (iii) Added -r, --recursive to recurse into sub-directories.
4914 (iv) Added -f, --file to read patterns from a file.
4915
4916 6. pcre_exec() was referring to its "code" argument before testing that
4917 argument for NULL (and giving an error if it was NULL).
4918
4919 7. Upgraded Makefile.in to allow for compiling in a different directory from
4920 the source directory.
4921
4922 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4923 options bits, the pointer it was passed was to an int instead of to an unsigned
4924 long int. This mattered only on 64-bit systems.
4925
4926 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4927 generated) instead of pcre.in, which it its source. Also made the same change
4928 in several of the .c files.
4929
4930 10. A new release of gcc defines printf() as a macro, which broke pcretest
4931 because it had an ifdef in the middle of a string argument for printf(). Fixed
4932 by using separate calls to printf().
4933
4934 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4935 script, to force use of CR or LF instead of \n in the source. On non-Unix
4936 systems, the value can be set in config.h.
4937
4938 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4939 absolute limit. Changed the text of the error message to make this clear, and
4940 likewise updated the man page.
4941
4942 13. The limit of 99 on the number of capturing subpatterns has been removed.
4943 The new limit is 65535, which I hope will not be a "real" limit.
4944
4945
4946 Version 3.4 22-Aug-00
4947 ---------------------
4948
4949 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4950
4951 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4952
4953
4954 Version 3.3 01-Aug-00
4955 ---------------------
4956
4957 1. If an octal character was given, but the value was greater than \377, it
4958 was not getting masked to the least significant bits, as documented. This could
4959 lead to crashes in some systems.
4960
4961 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4962 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4963
4964 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4965 These just pass their arguments on to (pcre_free)(), but they are provided
4966 because some uses of PCRE bind it to non-C systems that can call its functions,
4967 but cannot call free() or pcre_free() directly.
4968
4969 4. Add "make test" as a synonym for "make check". Corrected some comments in
4970 the Makefile.
4971
4972 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4973 Makefile.
4974
4975 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4976 command called pgrep for grepping around the active processes.
4977
4978 7. Added the beginnings of support for UTF-8 character strings.
4979
4980 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4981 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4982 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4983 out for the ar command.)
4984
4985
4986 Version 3.2 12-May-00
4987 ---------------------
4988
4989 This is purely a bug fixing release.
4990
4991 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4992 of ZA. This was just one example of several cases that could provoke this bug,
4993 which was introduced by change 9 of version 2.00. The code for breaking
4994 infinite loops after an iteration that matches an empty string was't working
4995 correctly.
4996
4997 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4998 when matched against abbab (for example). After matching an empty string, it
4999 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5000 caused it to match further down the string than it should.
5001
5002 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5003 was there because it doesn't seem to be needed, and it causes trouble on some
5004 systems, as it is not a Standard C header. It has been removed.
5005
5006 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5007 were reported on the Macintosh. The changes were from
5008
5009 while ((c = *(++ptr)) != 0 && c != '\n');
5010 to
5011 while ((c = *(++ptr)) != 0 && c != '\n') ;
5012
5013 Totally extraordinary, but if that's what it takes...
5014
5015 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5016 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5017 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5018 assumes the way PCRE uses memmove() (always moving upwards).
5019
5020 6. PCRE is being used in one environment where strchr() is not available. There
5021 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5022 faster code anyway.
5023
5024
5025 Version 3.1 09-Feb-00
5026 ---------------------
5027
5028 The only change in this release is the fixing of some bugs in Makefile.in for
5029 the "install" target:
5030
5031 (1) It was failing to install pcreposix.h.
5032
5033 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5034
5035
5036 Version 3.0 01-Feb-00
5037 ---------------------
5038
5039 1. Add support for the /+ modifier to perltest (to output $` like it does in
5040 pcretest).
5041
5042 2. Add support for the /g modifier to perltest.
5043
5044 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5045 matches null strings.
5046
5047 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5048 pattern. Perl treats empty patterns specially - it reuses the most recent
5049 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5050 effect.
5051
5052 5. The POSIX interface was broken in that it was just handing over the POSIX
5053 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5054 required a bigger vector, with some working space on the end. This means that
5055 the POSIX wrapper now has to get and free some memory, and copy the results.
5056
5057 6. Added some simple autoconf support, placing the test data and the
5058 documentation in separate directories, re-organizing some of the
5059 information files, and making it build pcre-config (a GNU standard). Also added
5060 libtool support for building PCRE as a shared library, which is now the
5061 default.
5062
5063 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5064 09 are not valid octal constants. Single digits will be used for minor values
5065 less than 10.
5066
5067 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5068 existing programs that set these in the POSIX interface can use PCRE without
5069 modification.
5070
5071 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5072 return all that pcre_info() returns, plus additional data. The pcre_info()
5073 function is retained for compatibility, but is considered to be obsolete.
5074
5075 10. Added experimental recursion feature (?R) to handle one common case that
5076 Perl 5.6 will be able to do with (?p{...}).
5077
5078 11. Added support for POSIX character classes like [:alpha:], which Perl is
5079 adopting.
5080
5081
5082 Version 2.08 31-Aug-99
5083 ----------------------
5084
5085 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5086 trying to match at the startoffset position, but instead was moving forward to
5087 the next newline as if a previous match had failed.
5088
5089 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5090 and could get into a loop if a null string was matched other than at the start
5091 of the subject.
5092
5093 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5094 be distinguished at compile time, and for completeness also added PCRE_DATE.
5095
5096 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5097 in GnuWin32 environments.
5098
5099
5100 Version 2.07 29-Jul-99
5101 ----------------------
5102
5103 1. The documentation is now supplied in plain text form and HTML as well as in
5104 the form of man page sources.
5105
5106 2. C++ compilers don't like assigning (void *) values to other pointer types.
5107 In particular this affects malloc(). Although there is no problem in Standard
5108 C, I've put in casts to keep C++ compilers happy.
5109
5110 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5111 should be (const char *).
5112
5113 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5114 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5115 However, I haven't made this a standard facility. The documentation doesn't
5116 mention it, and the Makefile doesn't support it.
5117
5118 5. The Makefile now contains an "install" target, with editable destinations at
5119 the top of the file. The pcretest program is not installed.
5120
5121 6. pgrep -V now gives the PCRE version number and date.
5122
5123 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5124 causing the entire string to be ignored, instead of just the last character.
5125
5126 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5127 non-matching string, it can take a very, very long time, even for strings of
5128 quite modest length, because of the nested recursion. PCRE now does better in
5129 some of these cases. It does this by remembering the last required literal
5130 character in the pattern, and pre-searching the subject to ensure it is present
5131 before running the real match. In other words, it applies a heuristic to detect
5132 some types of certain failure quickly, and in the above example, if presented
5133 with a string that has no trailing " it gives "no match" very quickly.
5134
5135 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5136 other alternatives are tried instead.
5137
5138
5139 Version 2.06 09-Jun-99
5140 ----------------------
5141
5142 1. Change pcretest's output for amount of store used to show just the code
5143 space, because the remainder (the data block) varies in size between 32-bit and
5144 64-bit systems.
5145
5146 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5147 start matching at. This allows lookbehinds to work when searching for multiple
5148 occurrences in a string.
5149
5150 3. Added additional options to pcretest for testing multiple occurrences:
5151
5152 /+ outputs the rest of the string that follows a match
5153 /g loops for multiple occurrences, using the new startoffset argument
5154 /G loops for multiple occurrences by passing an incremented pointer
5155
5156 4. PCRE wasn't doing the "first character" optimization for patterns starting
5157 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5158 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5159 the letter 'x'. On long subject strings, this gives a significant speed-up.
5160
5161
5162 Version 2.05 21-Apr-99
5163 ----------------------
5164
5165 1. Changed the type of magic_number from int to long int so that it works
5166 properly on 16-bit systems.
5167
5168 2. Fixed a bug which caused patterns starting with .* not to work correctly
5169 when the subject string contained newline characters. PCRE was assuming
5170 anchoring for such patterns in all cases, which is not correct because .* will
5171 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5172 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5173 must be retried after every newline in the subject.
5174
5175
5176 Version 2.04 18-Feb-99
5177 ----------------------
5178
5179 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5180 computation of the store needed to hold the pattern was incorrect (too large).
5181 If such patterns were nested a few deep, this could multiply and become a real
5182 problem.
5183
5184 2. Added /M option to pcretest to show the memory requirement of a specific
5185 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5186
5187 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5188 compiled in such a way that the backtracking after subsequent failure was
5189 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5190 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5191
5192
5193 Version 2.03 02-Feb-99
5194 ----------------------
5195
5196 1. Fixed typo and small mistake in man page.
5197
5198 2. Added 4th condition (GPL supersedes if conflict) and created separate
5199 LICENCE file containing the conditions.
5200
5201 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5202 Perl, that is the internal \ allows the delimiter to be included in the
5203 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5204 the final delimiter, add \ to the end of the pattern (to test the error).
5205
5206 4. Added the convenience functions for extracting substrings after a successful
5207 match. Updated pcretest to make it able to test these functions.
5208
5209
5210 Version 2.02 14-Jan-99
5211 ----------------------
5212
5213 1. Initialized the working variables associated with each extraction so that
5214 their saving and restoring doesn't refer to uninitialized store.
5215
5216 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5217 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5218 fix the problem.
5219
5220 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5221 calls, and wasn't printing the correct value for compiling calls. Increased the
5222 default value of LOOPREPEAT, and the number of significant figures in the
5223 times.
5224
5225 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5226
5227 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5228 a building problem on Windows NT with a FAT file system.
5229
5230
5231 Version 2.01 21-Oct-98
5232 ----------------------
5233
5234 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5235 to character tables built by pcre_maketables() in the current locale. If NULL
5236 is passed, the default tables are used.
5237
5238
5239 Version 2.00 24-Sep-98
5240 ----------------------
5241
5242 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5243 it any more.
5244
5245 2. Allow quantification of (?>) groups, and make it work correctly.
5246
5247 3. The first character computation wasn't working for (?>) groups.
5248
5249 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5250 end of the subject) and add 5.005's \z, which really does match only at the
5251 very end of the subject.
5252
5253 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5254
5255 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5256 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5257 localized options. All options to pcre_study() were also removed.
5258
5259 7. Add other new features from 5.005:
5260
5261 $(?<= positive lookbehind
5262 $(?<! negative lookbehind
5263 (?imsx-imsx) added the unsetting capability
5264 such a setting is global if at outer level; local otherwise
5265 (?imsx-imsx:) non-capturing groups with option setting
5266 (?(cond)re|re) conditional pattern matching
5267
5268 A backreference to itself in a repeated group matches the previous
5269 captured string.
5270
5271 8. General tidying up of studying (both automatic and via "study")
5272 consequential on the addition of new assertions.
5273
5274 9. As in 5.005, unlimited repeated groups that could match an empty substring
5275 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5276 runtime if any iteration does actually match an empty substring.
5277
5278 10. Include the RunTest script in the distribution.
5279
5280 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5281 discrepancies, some of which were old and were also with respect to 5.004. They
5282 have now been fixed.
5283
5284
5285 Version 1.09 28-Apr-98
5286 ----------------------
5287
5288 1. A negated single character class followed by a quantifier with a minimum
5289 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5290 program crashes, or just wrong answers. This did not apply to negated classes
5291 containing more than one character, or to minima other than one.
5292
5293
5294 Version 1.08 27-Mar-98
5295 ----------------------
5296
5297 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5298
5299 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5300 latter must appear before anything that relies on it in the pattern.
5301
5302
5303 Version 1.07 16-Feb-98
5304 ----------------------
5305
5306 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5307 repeat of a potentially empty string).
5308
5309
5310 Version 1.06 23-Jan-98
5311 ----------------------
5312
5313 1. Added Markus Oberhumer's little patches for C++.
5314
5315 2. Literal strings longer than 255 characters were broken.
5316
5317
5318 Version 1.05 23-Dec-97
5319 ----------------------
5320
5321 1. Negated character classes containing more than one character were failing if
5322 PCRE_CASELESS was set at run time.
5323
5324
5325 Version 1.04 19-Dec-97
5326 ----------------------
5327
5328 1. Corrected the man page, where some "const" qualifiers had been omitted.
5329
5330 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5331 input syntax.
5332
5333 3. Fixed memory leak which occurred when a regex with back references was
5334 matched with an offsets vector that wasn't big enough. The temporary memory
5335 that is used in this case wasn't being freed if the match failed.
5336
5337 4. Tidied pcretest to ensure it frees memory that it gets.
5338
5339 5. Temporary memory was being obtained in the case where the passed offsets
5340 vector was exactly big enough.
5341
5342 6. Corrected definition of offsetof() from change 5 below.
5343
5344 7. I had screwed up change 6 below and broken the rules for the use of
5345 setjmp(). Now fixed.
5346
5347
5348 Version 1.03 18-Dec-97
5349 ----------------------
5350
5351 1. A erroneous regex with a missing opening parenthesis was correctly
5352 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5353 on some systems.
5354
5355 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5356 it was reported that one broken compiler failed on the former because "code" is
5357 also an independent variable.
5358
5359 3. The erroneous regex a[]b caused an array overrun reference.
5360
5361 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5362 fail on data ending with that character. (It was going on too far, and checking
5363 the next character, typically a binary zero.) This was specific to the
5364 optimized code for single-character negative classes.
5365
5366 5. Added a contributed patch from the TIN world which does the following:
5367
5368 + Add an undef for memmove, in case the the system defines a macro for it.
5369
5370 + Add a definition of offsetof(), in case there isn't one. (I don't know
5371 the reason behind this - offsetof() is part of the ANSI standard - but
5372 it does no harm).
5373
5374 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5375 most of the places where whitespace preceded '#'. I have given up and
5376 allowed the remaining 2 cases to be at the margin.
5377
5378 + Rename some variables in pcre to eliminate shadowing. This seems very
5379 pedantic, but does no harm, of course.
5380
5381 6. Moved the call to setjmp() into its own function, to get rid of warnings
5382 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5383
5384 7. Constructs such as \d{8,} were compiling into the equivalent of
5385 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5386 outcome, but in this particular case used more store than had been allocated,
5387 which caused the bug to be discovered because it threw up an internal error.
5388
5389 8. The debugging code in both pcre and pcretest for outputting the compiled
5390 form of a regex was going wrong in the case of back references followed by
5391 curly-bracketed repeats.
5392
5393
5394 Version 1.02 12-Dec-97
5395 ----------------------
5396
5397 1. Typos in pcre.3 and comments in the source fixed.
5398
5399 2. Applied a contributed patch to get rid of places where it used to remove
5400 'const' from variables, and fixed some signed/unsigned and uninitialized
5401 variable warnings.
5402
5403 3. Added the "runtest" target to Makefile.
5404
5405 4. Set default compiler flag to -O2 rather than just -O.
5406
5407
5408 Version 1.01 19-Nov-97
5409 ----------------------
5410
5411 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5412 like /([ab]*)*/, that is, for classes with more than one character in them.
5413
5414 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5415 as /((?>a*))*/ (a PCRE_EXTRA facility).
5416
5417
5418 Version 1.00 18-Nov-97
5419 ----------------------
5420
5421 1. Added compile-time macros to support systems such as SunOS4 which don't have
5422 memmove() or strerror() but have other things that can be used instead.
5423
5424 2. Arranged that "make clean" removes the executables.
5425
5426
5427 Version 0.99 27-Oct-97
5428 ----------------------
5429
5430 1. Fixed bug in code for optimizing classes with only one character. It was
5431 initializing a 32-byte map regardless, which could cause it to run off the end
5432 of the memory it had got.
5433
5434 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5435
5436
5437 Version 0.98 22-Oct-97
5438 ----------------------
5439
5440 1. Fixed bug in code for handling temporary memory usage when there are more
5441 back references than supplied space in the ovector. This could cause segfaults.
5442
5443
5444 Version 0.97 21-Oct-97
5445 ----------------------
5446
5447 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5448
5449 2. Optimized negated single characters not to use a bit map.
5450
5451 3. Brought error texts together as macro definitions; clarified some of them;
5452 fixed one that was wrong - it said "range out of order" when it meant "invalid
5453 escape sequence".
5454
5455 4. Changed some char * arguments to const char *.
5456
5457 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5458
5459 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5460 pcretest.
5461
5462
5463 Version 0.96 16-Oct-97
5464 ----------------------
5465
5466 1. Added a simple "pgrep" utility to the distribution.
5467
5468 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5469 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5470 where "ddd" means "one or more decimal digits".
5471
5472 3. Fixed serious bug. If a pattern had a back reference, but the call to
5473 pcre_exec() didn't supply a large enough ovector to record the related
5474 identifying subpattern, the match always failed. PCRE now remembers the number
5475 of the largest back reference, and gets some temporary memory in which to save
5476 the offsets during matching if necessary, in order to ensure that
5477 backreferences always work.
5478
5479 4. Increased the compatibility with Perl in a number of ways:
5480
5481 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5482 to request this handling. The option can be set at compile or exec time.
5483
5484 (b) $ matches before a terminating newline by default; an option
5485 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5486 mode). The option can be set at compile or exec time.
5487
5488 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5489 the same as Perl's. If the decimal number it represents is less than 10
5490 or there aren't that many previous left capturing parentheses, an octal
5491 escape is read. Inside a character class, it's always an octal escape,
5492 even if it is a single digit.
5493
5494 (d) An escaped but undefined alphabetic character is taken as a literal,
5495 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5496 escapes.
5497
5498 (e) {0} is now permitted. (The previous item is removed from the compiled
5499 pattern).
5500
5501 5. Changed all the names of code files so that the basic parts are no longer
5502 than 10 characters, and abolished the teeny "globals.c" file.
5503
5504 6. Changed the handling of character classes; they are now done with a 32-byte
5505 bit map always.
5506
5507 7. Added the -d and /D options to pcretest to make it possible to look at the
5508 internals of compilation without having to recompile pcre.
5509
5510
5511 Version 0.95 23-Sep-97
5512 ----------------------
5513
5514 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5515 \x20 at the start of a run of normal characters. These were being treated as
5516 real characters, instead of the source characters being re-checked.
5517
5518
5519 Version 0.94 18-Sep-97
5520 ----------------------
5521
5522 1. The functions are now thread-safe, with the caveat that the global variables
5523 containing pointers to malloc() and free() or alternative functions are the
5524 same for all threads.
5525
5526 2. Get pcre_study() to generate a bitmap of initial characters for non-
5527 anchored patterns when this is possible, and use it if passed to pcre_exec().
5528
5529
5530 Version 0.93 15-Sep-97
5531 ----------------------
5532
5533 1. /(b)|(:+)/ was computing an incorrect first character.
5534
5535 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5536 but not actually doing anything yet.
5537
5538 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5539 as Perl does (e.g. [-az] or [az-]).
5540
5541 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5542 all possible positions.
5543
5544 5. Split up into different modules to avoid including unneeded functions in a
5545 compiled binary. However, compile and exec are still in one module. The "study"
5546 function is split off.
5547
5548 6. The character tables are now in a separate module whose source is generated
5549 by an auxiliary program - but can then be edited by hand if required. There are
5550 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5551 toupper() in the code.
5552
5553 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5554 make them global. Abolish the function for setting them, as the caller can now
5555 set them directly.
5556
5557
5558 Version 0.92 11-Sep-97
5559 ----------------------
5560
5561 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5562 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5563
5564 2. Caseless matching was not working in character classes if the characters in
5565 the pattern were in upper case.
5566
5567 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5568
5569 4. Make PCRE_ANCHORED public and accept as a compile option.
5570
5571 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5572 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5573 pass them.
5574
5575 6. Give an error if bad option bits passed at compile or run time.
5576
5577 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5578 pcretest to cause it to pass that flag.
5579
5580 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5581 options, and the first character, if set.
5582
5583 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5584
5585
5586 Version 0.91 10-Sep-97
5587 ----------------------
5588
5589 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5590 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5591
5592 2. PCRE was looping on encountering an indefinitely repeated back reference to
5593 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5594 Perl does - treats the match as successful.
5595
5596 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5