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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1518 - (show annotations)
Fri Feb 6 16:10:27 2015 UTC (4 years, 8 months ago) by ph10
File size: 269810 byte(s)
Catch auto-possessification potential loop for bad UTF pattern with 
NO_UTF_CHECK.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.37 xx-xxx-2015
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 5. Fixed a memory leak during matching that could occur for a subpattern
24 subroutine call (recursive or otherwise) if the number of captured groups
25 that had to be saved was greater than ten.
26
27 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
28 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
29 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
30
31
32 Version 8.36 26-September-2014
33 ------------------------------
34
35 1. Got rid of some compiler warnings in the C++ modules that were shown up by
36 -Wmissing-field-initializers and -Wunused-parameter.
37
38 2. The tests for quantifiers being too big (greater than 65535) were being
39 applied after reading the number, and stupidly assuming that integer
40 overflow would give a negative number. The tests are now applied as the
41 numbers are read.
42
43 3. Tidy code in pcre_exec.c where two branches that used to be different are
44 now the same.
45
46 4. The JIT compiler did not generate match limit checks for certain
47 bracketed expressions with quantifiers. This may lead to exponential
48 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
49 issue should be resolved now.
50
51 5. Fixed an issue, which occures when nested alternatives are optimized
52 with table jumps.
53
54 6. Inserted two casts and changed some ints to size_t in the light of some
55 reported 64-bit compiler warnings (Bugzilla 1477).
56
57 7. Fixed a bug concerned with zero-minimum possessive groups that could match
58 an empty string, which sometimes were behaving incorrectly in the
59 interpreter (though correctly in the JIT matcher). This pcretest input is
60 an example:
61
62 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
63 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
64
65 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
66 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
67 for an empty string was breaking the inner loop and carrying on at a lower
68 level, when possessive repeated groups should always return to a higher
69 level as they have no backtrack points in them. The empty string test now
70 occurs at the outer level.
71
72 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
73 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
74
75 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
76 doesn't).
77
78 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
79 to hold a pattern to be incorrectly computed (too small) when there were
80 named back references to duplicated names. This could cause "internal
81 error: code overflow" or "double free or corruption" or other memory
82 handling errors.
83
84 11. When named subpatterns had the same prefixes, back references could be
85 confused. For example, in this pattern:
86
87 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
88
89 the reference to 'Name' was incorrectly treated as a reference to a
90 duplicate name.
91
92 12. A pattern such as /^s?c/mi8 where the optional character has more than
93 one "other case" was incorrectly compiled such that it would only try to
94 match starting at "c".
95
96 13. When a pattern starting with \s was studied, VT was not included in the
97 list of possible starting characters; this should have been part of the
98 8.34/18 patch.
99
100 14. If a character class started [\Qx]... where x is any character, the class
101 was incorrectly terminated at the ].
102
103 15. If a pattern that started with a caseless match for a character with more
104 than one "other case" was studied, PCRE did not set up the starting code
105 unit bit map for the list of possible characters. Now it does. This is an
106 optimization improvement, not a bug fix.
107
108 16. The Unicode data tables have been updated to Unicode 7.0.0.
109
110 17. Fixed a number of memory leaks in pcregrep.
111
112 18. Avoid a compiler warning (from some compilers) for a function call with
113 a cast that removes "const" from an lvalue by using an intermediate
114 variable (to which the compiler does not object).
115
116 19. Incorrect code was compiled if a group that contained an internal recursive
117 back reference was optional (had quantifier with a minimum of zero). This
118 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
119 caused segmentation faults because of stack overflows at compile time.
120
121 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
122 group that is quantified with an indefinite repeat, caused a compile-time
123 loop which used up all the system stack and provoked a segmentation fault.
124 This was not the same bug as 19 above.
125
126 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
127 Patch by Mike Frysinger.
128
129
130 Version 8.35 04-April-2014
131 --------------------------
132
133 1. A new flag is set, when property checks are present in an XCLASS.
134 When this flag is not set, PCRE can perform certain optimizations
135 such as studying these XCLASS-es.
136
137 2. The auto-possessification of character sets were improved: a normal
138 and an extended character set can be compared now. Furthermore
139 the JIT compiler optimizes more character set checks.
140
141 3. Got rid of some compiler warnings for potentially uninitialized variables
142 that show up only when compiled with -O2.
143
144 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
145 of a match later then the end of the match. The pcretest program was not
146 handling the case sensibly - it was outputting from the start to the next
147 binary zero. It now reports this situation in a message, and outputs the
148 text from the end to the start.
149
150 5. Fast forward search is improved in JIT. Instead of the first three
151 characters, any three characters with fixed position can be searched.
152 Search order: first, last, middle.
153
154 6. Improve character range checks in JIT. Characters are read by an inprecise
155 function now, which returns with an unknown value if the character code is
156 above a certain threshold (e.g: 256). The only limitation is that the value
157 must be bigger than the threshold as well. This function is useful when
158 the characters above the threshold are handled in the same way.
159
160 7. The macros whose names start with RAWUCHAR are placeholders for a future
161 mode in which only the bottom 21 bits of 32-bit data items are used. To
162 make this more memorable for those maintaining the code, the names have
163 been changed to start with UCHAR21, and an extensive comment has been added
164 to their definition.
165
166 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
167 to the export list in Makefile.am (they were accidentally omitted from the
168 8.34 tarball).
169
170 9. The informational output from pcretest used the phrase "starting byte set"
171 which is inappropriate for the 16-bit and 32-bit libraries. As the output
172 for "first char" and "need char" really means "non-UTF-char", I've changed
173 "byte" to "char", and slightly reworded the output. The documentation about
174 these values has also been (I hope) clarified.
175
176 10. Another JIT related optimization: use table jumps for selecting the correct
177 backtracking path, when more than four alternatives are present inside a
178 bracket.
179
180 11. Empty match is not possible, when the minimum length is greater than zero,
181 and there is no \K in the pattern. JIT should avoid empty match checks in
182 such cases.
183
184 12. In a caseless character class with UCP support, when a character with more
185 than one alternative case was not the first character of a range, not all
186 the alternative cases were added to the class. For example, s and \x{17f}
187 are both alternative cases for S: the class [RST] was handled correctly,
188 but [R-T] was not.
189
190 13. The configure.ac file always checked for pthread support when JIT was
191 enabled. This is not used in Windows, so I have put this test inside a
192 check for the presence of windows.h (which was already tested for).
193
194 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
195 The algorithm provides a way to skip certain starting offsets, and usually
196 faster than linear prefix searches.
197
198 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
199 as for 'fr_FR' and 'french'. For some reason, however, it then used the
200 Windows-specific input and output files, which have 'french' screwed in.
201 So this could never have worked. One of the problems with locales is that
202 they aren't always the same. I have now updated RunTest so that it checks
203 the output of the locale test (test 3) against three different output
204 files, and it allows the test to pass if any one of them matches. With luck
205 this should make the test pass on some versions of Solaris where it was
206 failing. Because of the uncertainty, the script did not used to stop if
207 test 3 failed; it now does. If further versions of a French locale ever
208 come to light, they can now easily be added.
209
210 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
211 there was a message during ./configure, but it did not stop. This now
212 provokes an error. The invalid example in README has been corrected.
213 If a value less than the minimum is given, the minimum value has always
214 been used, but now a warning is given.
215
216 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
217 was a bug in the test system, which is now fixed. Also, the list of various
218 configurations that are tested for each release did not have one with both
219 16/32 bits and --enable-bar-anycrlf. It now does.
220
221 18. pcretest was missing "-C bsr" for displaying the \R default setting.
222
223 19. Little endian PowerPC systems are supported now by the JIT compiler.
224
225 20. The fast forward newline mechanism could enter to an infinite loop on
226 certain invalid UTF-8 input. Although we don't support these cases
227 this issue can be fixed by a performance optimization.
228
229 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
230 not take account if existing stack usage. There is now a new global
231 variable called pcre_stack_guard that can be set to point to an external
232 function to check stack availability. It is called at the start of
233 processing every parenthesized group.
234
235 22. A typo in the code meant that in ungreedy mode the max/min qualifier
236 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
237 not match "ab".
238
239 23. When UTF was disabled, the JIT program reported some incorrect compile
240 errors. These messages are silenced now.
241
242 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
243 compiler.
244
245 25. Change all the temporary files used in RunGrepTest to be different to those
246 used by RunTest so that the tests can be run simultaneously, for example by
247 "make -j check".
248
249
250 Version 8.34 15-December-2013
251 -----------------------------
252
253 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
254 executable memory. Patch inspired by Carsten Klein.
255
256 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
257 this macro is never tested and has no effect, because the work to support
258 coverage involves only compiling and linking options and special targets in
259 the Makefile. The comment in config.h implied that defining the macro would
260 enable coverage support, which is totally false. There was also support for
261 setting this macro in the CMake files (my fault, I just copied it from
262 configure). SUPPORT_GCOV has now been removed.
263
264 3. Make a small performance improvement in strlen16() and strlen32() in
265 pcretest.
266
267 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
268 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
269 these cases). There was also one in pcretest.c.
270
271 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
272
273 6. In UTF mode, the code for checking whether a group could match an empty
274 string (which is used for indefinitely repeated groups to allow for
275 breaking an infinite loop) was broken when the group contained a repeated
276 negated single-character class with a character that occupied more than one
277 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
278 UTF-8 mode). The effect was undefined: the group might or might not be
279 deemed as matching an empty string, or the program might have crashed.
280
281 7. The code for checking whether a group could match an empty string was not
282 recognizing that \h, \H, \v, \V, and \R must match a character.
283
284 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
285 an empty string. If it can, pcretest shows this in its information output.
286
287 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
288 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
289 matched by pcre_exec() without using JIT:
290
291 (a) If the rest of the pattern did not match after a maximal run of
292 grapheme clusters, the code for backing up to try with fewer of them
293 did not always back up over a full grapheme when characters that do not
294 have the modifier quality were involved, e.g. Hangul syllables.
295
296 (b) If the match point in a subject started with modifier character, and
297 there was no match, the code could incorrectly back up beyond the match
298 point, and potentially beyond the first character in the subject,
299 leading to a segfault or an incorrect match result.
300
301 10. A conditional group with an assertion condition could lead to PCRE
302 recording an incorrect first data item for a match if no other first data
303 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
304 first data item, and therefore matched "ca" after "c" instead of at the
305 start.
306
307 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
308 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
309
310 12. The source of pcregrep now includes z/OS-specific code so that it can be
311 compiled for z/OS as part of the special z/OS distribution.
312
313 13. Added the -T and -TM options to pcretest.
314
315 14. The code in pcre_compile.c for creating the table of named capturing groups
316 has been refactored. Instead of creating the table dynamically during the
317 actual compiling pass, the information is remembered during the pre-compile
318 pass (on the stack unless there are more than 20 named groups, in which
319 case malloc() is used) and the whole table is created before the actual
320 compile happens. This has simplified the code (it is now nearly 150 lines
321 shorter) and prepared the way for better handling of references to groups
322 with duplicate names.
323
324 15. A back reference to a named subpattern when there is more than one of the
325 same name now checks them in the order in which they appear in the pattern.
326 The first one that is set is used for the reference. Previously only the
327 first one was inspected. This change makes PCRE more compatible with Perl.
328
329 16. Unicode character properties were updated from Unicode 6.3.0.
330
331 17. The compile-time code for auto-possessification has been refactored, based
332 on a patch by Zoltan Herczeg. It now happens after instead of during
333 compilation. The code is cleaner, and more cases are handled. The option
334 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
335 options in pcretest are provided to set it. It can also be set by
336 (*NO_AUTO_POSSESS) at the start of a pattern.
337
338 18. The character VT has been added to the default ("C" locale) set of
339 characters that match \s and are generally treated as white space,
340 following this same change in Perl 5.18. There is now no difference between
341 "Perl space" and "POSIX space". Whether VT is treated as white space in
342 other locales depends on the locale.
343
344 19. The code for checking named groups as conditions, either for being set or
345 for being recursed, has been refactored (this is related to 14 and 15
346 above). Processing unduplicated named groups should now be as fast at
347 numerical groups, and processing duplicated groups should be faster than
348 before.
349
350 20. Two patches to the CMake build system, by Alexander Barkov:
351
352 (1) Replace the "source" command by "." in CMakeLists.txt because
353 "source" is a bash-ism.
354
355 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
356 without these the CMake build does not work on Solaris.
357
358 21. Perl has changed its handling of \8 and \9. If there is no previously
359 encountered capturing group of those numbers, they are treated as the
360 literal characters 8 and 9 instead of a binary zero followed by the
361 literals. PCRE now does the same.
362
363 22. Following Perl, added \o{} to specify codepoints in octal, making it
364 possible to specify values greater than 0777 and also making them
365 unambiguous.
366
367 23. Perl now gives an error for missing closing braces after \x{... instead of
368 treating the string as literal. PCRE now does the same.
369
370 24. RunTest used to grumble if an inappropriate test was selected explicitly,
371 but just skip it when running all tests. This make it awkward to run ranges
372 of tests when one of them was inappropriate. Now it just skips any
373 inappropriate tests, as it always did when running all tests.
374
375 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
376 character types such as \d or \w, too many callouts were inserted, and the
377 data that they returned was rubbish.
378
379 26. In UCP mode, \s was not matching two of the characters that Perl matches,
380 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
381 were matched by \h. The code has now been refactored so that the lists of
382 the horizontal and vertical whitespace characters used for \h and \v (which
383 are defined only in one place) are now also used for \s.
384
385 27. Add JIT support for the 64 bit TileGX architecture.
386 Patch by Jiong Wang (Tilera Corporation).
387
388 28. Possessive quantifiers for classes (both explicit and automatically
389 generated) now use special opcodes instead of wrapping in ONCE brackets.
390
391 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
392 (because it's meaningless), this was not happening when PCRE_CASELESS was
393 set. Not wrong, but inefficient.
394
395 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
396 properties for \w, \d, etc) is present in a test regex. Otherwise if the
397 test contains no characters greater than 255, Perl doesn't realise it
398 should be using Unicode semantics.
399
400 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
401 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
402 does in Unicode mode.
403
404 32. Added the "forbid" facility to pcretest so that putting tests into the
405 wrong test files can sometimes be quickly detected.
406
407 33. There is now a limit (default 250) on the depth of nesting of parentheses.
408 This limit is imposed to control the amount of system stack used at compile
409 time. It can be changed at build time by --with-parens-nest-limit=xxx or
410 the equivalent in CMake.
411
412 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
413 errors. Perl warns for these when in warning mode, but PCRE has no facility
414 for giving warnings.
415
416 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
417 However, this was not working for (?!) because it is optimized to (*FAIL),
418 for which PCRE does not allow quantifiers. The optimization is now disabled
419 when a quantifier follows (?!). I can't see any use for this, but it makes
420 things uniform.
421
422 36. Perl no longer allows group names to start with digits, so I have made this
423 change also in PCRE. It simplifies the code a bit.
424
425 37. In extended mode, Perl ignores spaces before a + that indicates a
426 possessive quantifier. PCRE allowed a space before the quantifier, but not
427 before the possessive +. It now does.
428
429 38. The use of \K (reset reported match start) within a repeated possessive
430 group such as (a\Kb)*+ was not working.
431
432 40. Document that the same character tables must be used at compile time and
433 run time, and that the facility to pass tables to pcre_exec() and
434 pcre_dfa_exec() is for use only with saved/restored patterns.
435
436 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
437 features for Builds with MSVC:
438
439 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
440 builds). The .rc files can be used to set FileDescription and many other
441 attributes.
442
443 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
444 This allows higher-level build scripts which want .pdb files to avoid
445 hard-coding the exact files needed."
446
447 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
448 mean "start of word" and "end of word", respectively, as a transition aid.
449
450 43. A minimizing repeat of a class containing codepoints greater than 255 in
451 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
452 compiled to use the heap for recursion.
453
454 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
455 is configured.
456
457
458 Version 8.33 28-May-2013
459 ------------------------
460
461 1. Added 'U' to some constants that are compared to unsigned integers, to
462 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
463 variables that are added to signed variables, to ensure the result is
464 signed and can be negated.
465
466 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
467 CMake config files.
468
469 3. Revise the creation of config.h.generic so that all boolean macros are
470 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
471 overriding via -D on the command line possible.
472
473 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
474 to unsigned int is reported to make a quite noticeable speed difference in
475 a specific Windows environment. Testing on Linux did also appear to show
476 some benefit (and it is clearly not harmful). Also fixed the definition of
477 Xop which should be unsigned.
478
479 5. Related to (4), changing the definition of the intermediate variable cc
480 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
481 improvements.
482
483 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
484 unnecessary spaces.
485
486 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
487 and later.
488
489 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
490
491 9. Optimizing fast_forward_start_bits in JIT.
492
493 10. Adding support for callouts in JIT, and fixing some issues revealed
494 during this work. Namely:
495
496 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
497
498 (b) Minimum length was not checked before the matching is started.
499
500 11. The value of capture_last that is passed to callouts was incorrect in some
501 cases when there was a capture on one path that was subsequently abandoned
502 after a backtrack. Also, the capture_last value is now reset after a
503 recursion, since all captures are also reset in this case.
504
505 12. The interpreter no longer returns the "too many substrings" error in the
506 case when an overflowing capture is in a branch that is subsequently
507 abandoned after a backtrack.
508
509 13. In the pathological case when an offset vector of size 2 is used, pcretest
510 now prints out the matched string after a yield of 0 or 1.
511
512 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
513 Only supported by the JIT compiler at the moment.
514
515 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
516
517 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
518 offset of the starting point of the matching process, provided the offsets
519 vector is large enough.
520
521 17. The \A escape now records a lookbehind value of 1, though its execution
522 does not actually inspect the previous character. This is to ensure that,
523 in partial multi-segment matching, at least one character from the old
524 segment is retained when a new segment is processed. Otherwise, if there
525 are no lookbehinds in the pattern, \A might match incorrectly at the start
526 of a new segment.
527
528 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
529
530 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
531 gives some modest performance improvement in 8-bit mode.
532
533 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
534 be expressed in certain programming languages using Universal Character
535 Names.
536
537 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
538 which points out that "non characters" are not "characters that may not
539 appear in Unicode strings" but rather "characters that are reserved for
540 internal use and have only local meaning".
541
542 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
543 there was a conditional group that depended on an assertion, if the
544 assertion was false, the callout that immediately followed the alternation
545 in the condition was skipped when pcre_exec() was used for matching.
546
547 23. Allow an explicit callout to be inserted before an assertion that is the
548 condition for a conditional group, for compatibility with automatic
549 callouts, which always insert a callout at this point.
550
551 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
552 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
553
554 25. (*PRUNE) is now supported by the JIT compiler.
555
556 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
557
558 27. Fix the case where there are two or more SKIPs with arguments that may be
559 ignored.
560
561 28. (*SKIP) is now supported by the JIT compiler.
562
563 29. (*THEN) is now supported by the JIT compiler.
564
565 30. Update RunTest with additional test selector options.
566
567 31. The way PCRE handles backtracking verbs has been changed in two ways.
568
569 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
570 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
571 backtracking. In some cases this makes it more Perl-compatible, but Perl's
572 rather obscure rules do not always do the same thing.
573
574 (2) Previously, backtracking verbs were confined within assertions. This is
575 no longer the case for positive assertions, except for (*ACCEPT). Again,
576 this sometimes improves Perl compatibility, and sometimes does not.
577
578 32. A number of tests that were in test 2 because Perl did things differently
579 have been moved to test 1, because either Perl or PCRE has changed, and
580 these tests are now compatible.
581
582 32. Backtracking control verbs are now handled in the same way in JIT and
583 interpreter.
584
585 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
586 contained a forward subroutine reference caused a compile error.
587
588 34. Auto-detect and optimize limited repetitions in JIT.
589
590 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
591 blocking (*UTF) etc.
592
593 36. In the interpreter, maximizing pattern repetitions for characters and
594 character types now use tail recursion, which reduces stack usage.
595
596 37. The value of the max lookbehind was not correctly preserved if a compiled
597 and saved regex was reloaded on a host of different endianness.
598
599 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
600 of the compiled pattern block, expand the flags field from 16 to 32 bits
601 because it was almost full.
602
603 39. Try madvise first before posix_madvise.
604
605 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
606 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
607 original change locked out all matches of empty strings. This has been
608 changed so that one match of an empty string per line is recognized.
609 Subsequent searches on the same line (for colouring or for --only-matching,
610 for example) do not recognize empty strings.
611
612 41. Applied a user patch to fix a number of spelling mistakes in comments.
613
614 42. Data lines longer than 65536 caused pcretest to crash.
615
616 43. Clarified the data type for length and startoffset arguments for pcre_exec
617 and pcre_dfa_exec in the function-specific man pages, where they were
618 explicitly stated to be in bytes, never having been updated. I also added
619 some clarification to the pcreapi man page.
620
621 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
622 a segmentation fault.
623
624
625 Version 8.32 30-November-2012
626 -----------------------------
627
628 1. Improved JIT compiler optimizations for first character search and single
629 character iterators.
630
631 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
632 Patch by Daniel Richard G.
633
634 3. Single character iterator optimizations in the JIT compiler.
635
636 4. Improved JIT compiler optimizations for character ranges.
637
638 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
639 Reported by Giuseppe D'Angelo.
640
641 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
642 of a line, was being set incorrectly in cases where .* appeared inside
643 atomic brackets at the start of a pattern, or where there was a subsequent
644 *PRUNE or *SKIP.
645
646 7. Improved instruction cache flush for POWER/PowerPC.
647 Patch by Daniel Richard G.
648
649 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
650 grep:
651
652 (a) There is now no limit to the number of patterns to be matched.
653
654 (b) An error is given if a pattern is too long.
655
656 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
657 are now supported.
658
659 (d) --exclude-from and --include-from (multiple use) have been added.
660
661 (e) Exclusions and inclusions now apply to all files and directories, not
662 just to those obtained from scanning a directory recursively.
663
664 (f) Multiple uses of -f and --file-list are now supported.
665
666 (g) In a Windows environment, the default for -d has been changed from
667 "read" (the GNU grep default) to "skip", because otherwise the presence
668 of a directory in the file list provokes an error.
669
670 (h) The documentation has been revised and clarified in places.
671
672 9. Improve the matching speed of capturing brackets.
673
674 10. Changed the meaning of \X so that it now matches a Unicode extended
675 grapheme cluster.
676
677 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
678 out POSIX threads when JIT support is configured.
679
680 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
681
682 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
683 structure could go wrong in environments where size_t is not the same size
684 as int.
685
686 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
687 set.
688
689 15. The EBCDIC support had decayed; later updates to the code had included
690 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
691 general tidy up of EBCDIC-related issues, and the documentation was also
692 not quite right. There is now a test that can be run on ASCII systems to
693 check some of the EBCDIC-related things (but is it not a full test).
694
695 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
696 in a small tidy to the code.
697
698 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
699
700 18. If the --only-matching (-o) option in pcregrep is specified multiple
701 times, each one causes appropriate output. For example, -o1 -o2 outputs the
702 substrings matched by the 1st and 2nd capturing parentheses. A separating
703 string can be specified by --om-separator (default empty).
704
705 19. Improving the first n character searches.
706
707 20. Turn case lists for horizontal and vertical white space into macros so that
708 they are defined only once.
709
710 21. This set of changes together give more compatible Unicode case-folding
711 behaviour for characters that have more than one other case when UCP
712 support is available.
713
714 (a) The Unicode property table now has offsets into a new table of sets of
715 three or more characters that are case-equivalent. The MultiStage2.py
716 script that generates these tables (the pcre_ucd.c file) now scans
717 CaseFolding.txt instead of UnicodeData.txt for character case
718 information.
719
720 (b) The code for adding characters or ranges of characters to a character
721 class has been abstracted into a generalized function that also handles
722 case-independence. In UTF-mode with UCP support, this uses the new data
723 to handle characters with more than one other case.
724
725 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
726 whose other case is greater than 256 are now correctly matched
727 caselessly. Previously, the high codepoint matched the low one, but not
728 vice versa.
729
730 (d) The processing of \h, \H, \v, and \ in character classes now makes use
731 of the new class addition function, using character lists defined as
732 macros alongside the case definitions of 20 above.
733
734 (e) Caseless back references now work with characters that have more than
735 one other case.
736
737 (f) General caseless matching of characters with more than one other case
738 is supported.
739
740 22. Unicode character properties were updated from Unicode 6.2.0
741
742 23. Improved CMake support under Windows. Patch by Daniel Richard G.
743
744 24. Add support for 32-bit character strings, and UTF-32
745
746 25. Major JIT compiler update (code refactoring and bugfixing).
747 Experimental Sparc 32 support is added.
748
749 26. Applied a modified version of Daniel Richard G's patch to create
750 pcre.h.generic and config.h.generic by "make" instead of in the
751 PrepareRelease script.
752
753 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
754 pcre_compile.c when checking for a zero character.
755
756 28. Introducing a native interface for JIT. Through this interface, the compiled
757 machine code can be directly executed. The purpose of this interface is to
758 provide fast pattern matching, so several sanity checks are not performed.
759 However, feature tests are still performed. The new interface provides
760 1.4x speedup compared to the old one.
761
762 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
763 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
764 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
765
766 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
767 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
768 the "old" RFC 2279). Instead, it ended up passing a negative length to
769 pcre_exec().
770
771 31. Add support for GCC's visibility feature to hide internal functions.
772
773 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
774 "unknown -C option" after outputting 0 or 1.
775
776 33. There is now support for generating a code coverage report for the test
777 suite in environments where gcc is the compiler and lcov is installed. This
778 is mainly for the benefit of the developers.
779
780 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
781 unaddressable using valgrind annotations, allowing valgrind to detect
782 invalid memory accesses. This is mainly for the benefit of the developers.
783
784 25. (*UTF) can now be used to start a pattern in any of the three libraries.
785
786 26. Give configure error if --enable-cpp but no C++ compiler found.
787
788
789 Version 8.31 06-July-2012
790 -------------------------
791
792 1. Fixing a wrong JIT test case and some compiler warnings.
793
794 2. Removed a bashism from the RunTest script.
795
796 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
797 to unsigned type, result still unsigned" that was given by an MS compiler
798 on encountering the code "-sizeof(xxx)".
799
800 4. Partial matching support is added to the JIT compiler.
801
802 5. Fixed several bugs concerned with partial matching of items that consist
803 of more than one character:
804
805 (a) /^(..)\1/ did not partially match "aba" because checking references was
806 done on an "all or nothing" basis. This also applied to repeated
807 references.
808
809 (b) \R did not give a hard partial match if \r was found at the end of the
810 subject.
811
812 (c) \X did not give a hard partial match after matching one or more
813 characters at the end of the subject.
814
815 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
816 a partial match for the string "\r".
817
818 (e) When newline was set to CRLF, the metacharacter "." did not recognize
819 a partial match for a CR character at the end of the subject string.
820
821 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
822 running pcretest, the text "(JIT)" added to the output whenever JIT is
823 actually used to run the match.
824
825 7. Individual JIT compile options can be set in pcretest by following -s+[+]
826 or /S+[+] with a digit between 1 and 7.
827
828 8. OP_NOT now supports any UTF character not just single-byte ones.
829
830 9. (*MARK) control verb is now supported by the JIT compiler.
831
832 10. The command "./RunTest list" lists the available tests without actually
833 running any of them. (Because I keep forgetting what they all are.)
834
835 11. Add PCRE_INFO_MAXLOOKBEHIND.
836
837 12. Applied a (slightly modified) user-supplied patch that improves performance
838 when the heap is used for recursion (compiled with --disable-stack-for-
839 recursion). Instead of malloc and free for each heap frame each time a
840 logical recursion happens, frames are retained on a chain and re-used where
841 possible. This sometimes gives as much as 30% improvement.
842
843 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
844 call.
845
846 14. As documented, (*COMMIT) is now confined to within a positive assertion.
847
848 15. It is now possible to link pcretest with libedit as an alternative to
849 libreadline.
850
851 16. (*COMMIT) control verb is now supported by the JIT compiler.
852
853 17. The Unicode data tables have been updated to Unicode 6.1.0.
854
855 18. Added --file-list option to pcregrep.
856
857 19. Added binary file support to pcregrep, including the -a, --binary-files,
858 -I, and --text options.
859
860 20. The madvise function is renamed for posix_madvise for QNX compatibility
861 reasons. Fixed by Giuseppe D'Angelo.
862
863 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
864
865 22. Changed the option for creating symbolic links for 16-bit man pages from
866 -s to -sf so that re-installing does not cause issues.
867
868 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
869
870 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
871 matches in certain environments (the workspace was not being correctly
872 retained). Also added to pcre_dfa_exec() a simple plausibility check on
873 some of the workspace data at the beginning of a restart.
874
875 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
876 was not doing so when it should - probably a typo introduced by SVN 528
877 (change 8.10/14).
878
879 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
880 \w+ when the character tables indicated that \x{c4} was a word character.
881 There were several related cases, all because the tests for doing a table
882 lookup were testing for characters less than 127 instead of 255.
883
884 27. If a pattern contains capturing parentheses that are not used in a match,
885 their slots in the ovector are set to -1. For those that are higher than
886 any matched groups, this happens at the end of processing. In the case when
887 there were back references that the ovector was too small to contain
888 (causing temporary malloc'd memory to be used during matching), and the
889 highest capturing number was not used, memory off the end of the ovector
890 was incorrectly being set to -1. (It was using the size of the temporary
891 memory instead of the true size.)
892
893 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
894 ovector size, it uses memory at the end of the block that it has got.
895
896 29. Check for an overlong MARK name and give an error at compile time. The
897 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
898
899 30. JIT compiler update.
900
901 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
902 Rill for the patch.
903
904 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
905
906 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
907
908 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
909 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
910 was enabled.
911
912 35. Improve JIT code generation for greedy plus quantifier.
913
914 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
915 1 to "aa" instead of to an empty string. The bug affected repeated groups
916 that could potentially match an empty string.
917
918 37. Optimizing single character iterators in JIT.
919
920 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
921 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
922 codepoints that are too big for the mode are faulted, and in a UTF mode,
923 disallowed codepoints are also faulted.
924
925 39. If PCRE was compiled with UTF support, in three places in the DFA
926 matcher there was code that should only have been obeyed in UTF mode, but
927 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
928 processing when bytes with values greater than 127 were present. In 16-bit
929 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
930 both cases the values are those that cannot be the first data item in a UTF
931 character. The three items that might have provoked this were recursions,
932 possessively repeated groups, and atomic groups.
933
934 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
935 and pcregrep, because some OS require shared objects to be explicitly
936 passed to ld, causing the link step to fail if they are not.
937
938 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
939 mode, patterns that started with \h* or \R* might be incorrectly matched.
940
941
942 Version 8.30 04-February-2012
943 -----------------------------
944
945 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
946 name is defined in ctype.h.
947
948 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
949 only in quite long subpatterns.
950
951 3. Removed the function pcre_info(), which has been obsolete and deprecated
952 since it was replaced by pcre_fullinfo() in February 2000.
953
954 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
955 match a (*MARK), and the match failed at the start of the subject, a
956 reference to memory before the start of the subject could occur. This bug
957 was introduced by fix 17 of release 8.21.
958
959 5. A reference to an unset group with zero minimum repetition was giving
960 totally wrong answers (in non-JavaScript-compatibility mode). For example,
961 /(another)?(\1?)test/ matched against "hello world test". This bug was
962 introduced in release 8.13.
963
964 6. Add support for 16-bit character strings (a large amount of work involving
965 many changes and refactorings).
966
967 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
968 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
969 from a file.
970
971 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
972 rounding is not applied in this particular case).
973
974 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
975 if they appear, or are escaped, in patterns.
976
977 10. Get rid of a number of -Wunused-but-set-variable warnings.
978
979 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
980 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
981 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
982 also returns the mark "x". This bug applied to capturing parentheses,
983 non-capturing parentheses, and atomic parentheses. It also applied to some
984 assertions.
985
986 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
987 information out of configure.ac instead of relying on pcre.h.generic, which
988 is not stored in the repository.
989
990 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
991 -lreadline.
992
993 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
994
995 15. Applied Graycode's patch to put the top-level frame on the stack rather
996 than the heap when not using the stack for recursion. This gives a
997 performance improvement in many cases when recursion is not deep.
998
999 16. Experimental code added to "pcretest -C" to output the stack frame size.
1000
1001
1002 Version 8.21 12-Dec-2011
1003 ------------------------
1004
1005 1. Updating the JIT compiler.
1006
1007 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1008 are added as well.
1009
1010 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1011 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1012 calling _pcre_jit_exec. Some extra comments are added.
1013
1014 4. (*MARK) settings inside atomic groups that do not contain any capturing
1015 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1016 was introduced by change 18 for 8.20.
1017
1018 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1019 ECMA-262 standard.
1020
1021 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1022 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1023 This bug was probably introduced by change 9 of 8.13.
1024
1025 7. While fixing 6 above, I noticed that a number of other items were being
1026 incorrectly rejected as "not fixed length". This arose partly because newer
1027 opcodes had not been added to the fixed-length checking code. I have (a)
1028 corrected the bug and added tests for these items, and (b) arranged for an
1029 error to occur if an unknown opcode is encountered while checking for fixed
1030 length instead of just assuming "not fixed length". The items that were
1031 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1032 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1033 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1034
1035 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1036 being incorrectly compiled and would have given unpredicatble results.
1037
1038 9. A possessively repeated subpattern with minimum repeat count greater than
1039 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1040 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1041 the first (A) could occur when it should not.
1042
1043 10. Add a cast and remove a redundant test from the code.
1044
1045 11. JIT should use pcre_malloc/pcre_free for allocation.
1046
1047 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1048 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1049 is anything other than /usr, -L is still shown).
1050
1051 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1052
1053 14. Perl does not support \N without a following name in a [] class; PCRE now
1054 also gives an error.
1055
1056 15. If a forward reference was repeated with an upper limit of around 2000,
1057 it caused the error "internal error: overran compiling workspace". The
1058 maximum number of forward references (including repeats) was limited by the
1059 internal workspace, and dependent on the LINK_SIZE. The code has been
1060 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1061 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1062 of around 200,000 forward references. While doing this, I also speeded up
1063 the filling in of repeated forward references.
1064
1065 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1066 incorrectly expecting the subject to contain another "a" after the start.
1067
1068 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1069 in the match, the SKIP should be ignored. This was not happening; instead
1070 the SKIP was being treated as NOMATCH. For patterns such as
1071 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1072 tested.
1073
1074 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1075 now much more compatible with Perl, in particular in cases where the result
1076 is a non-match for a non-anchored pattern. For example, if
1077 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1078 "m", where previously it did not return a name. A side effect of this
1079 change is that for partial matches, the last encountered mark name is
1080 returned, as for non matches. A number of tests that were previously not
1081 Perl-compatible have been moved into the Perl-compatible test files. The
1082 refactoring has had the pleasing side effect of removing one argument from
1083 the match() function, thus reducing its stack requirements.
1084
1085 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1086 subsequent uses of /S (without +) incorrectly behaved like /S+.
1087
1088 21. Retrieve executable code size support for the JIT compiler and fixing
1089 some warnings.
1090
1091 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1092 not work when the shorter character appeared right at the end of the
1093 subject string.
1094
1095 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1096 systems.
1097
1098 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1099 output it when the /M option is used in pcretest.
1100
1101 25. The CheckMan script was not being included in the distribution. Also, added
1102 an explicit "perl" to run Perl scripts from the PrepareRelease script
1103 because this is reportedly needed in Windows.
1104
1105 26. If study data was being save in a file and studying had not found a set of
1106 "starts with" bytes for the pattern, the data written to the file (though
1107 never used) was taken from uninitialized memory and so caused valgrind to
1108 complain.
1109
1110 27. Updated RunTest.bat as provided by Sheri Pierce.
1111
1112 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1113
1114 29. Computation of memory usage for the table of capturing group names was
1115 giving an unnecessarily large value.
1116
1117
1118 Version 8.20 21-Oct-2011
1119 ------------------------
1120
1121 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1122 a POSIX class. After further experiments with Perl, which convinced me that
1123 Perl has bugs and confusions, a closing square bracket is no longer allowed
1124 in a POSIX name. This bug also affected patterns with classes that started
1125 with full stops.
1126
1127 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1128 captured substring, but while checking the failing first alternative,
1129 substring 1 is temporarily captured. If the output vector supplied to
1130 pcre_exec() was not big enough for this capture, the yield of the function
1131 was still zero ("insufficient space for captured substrings"). This cannot
1132 be totally fixed without adding another stack variable, which seems a lot
1133 of expense for a edge case. However, I have improved the situation in cases
1134 such as /(a)(b)x|abc/ matched against "abc", where the return code
1135 indicates that fewer than the maximum number of slots in the ovector have
1136 been set.
1137
1138 3. Related to (2) above: when there are more back references in a pattern than
1139 slots in the output vector, pcre_exec() uses temporary memory during
1140 matching, and copies in the captures as far as possible afterwards. It was
1141 using the entire output vector, but this conflicts with the specification
1142 that only 2/3 is used for passing back captured substrings. Now it uses
1143 only the first 2/3, for compatibility. This is, of course, another edge
1144 case.
1145
1146 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1147 main code base, and can be used by building with --enable-jit. When this is
1148 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1149 runtime --no-jit option is given.
1150
1151 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1152 ovector, the return from the function was zero, implying that there were
1153 other matches that did not fit. The correct "exactly full" value is now
1154 returned.
1155
1156 6. If a subpattern that was called recursively or as a subroutine contained
1157 (*PRUNE) or any other control that caused it to give a non-standard return,
1158 invalid errors such as "Error -26 (nested recursion at the same subject
1159 position)" or even infinite loops could occur.
1160
1161 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1162 computing the minimum length on reaching *ACCEPT, and so ended up with the
1163 wrong value of 1 rather than 0. Further investigation indicates that
1164 computing a minimum subject length in the presence of *ACCEPT is difficult
1165 (think back references, subroutine calls), and so I have changed the code
1166 so that no minimum is registered for a pattern that contains *ACCEPT.
1167
1168 8. If (*THEN) was present in the first (true) branch of a conditional group,
1169 it was not handled as intended. [But see 16 below.]
1170
1171 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1172 Sheri Pierce.
1173
1174 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1175 the first byte in a match must be "a".
1176
1177 11. Change 17 for 8.13 increased the recursion depth for patterns like
1178 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1179 pattern contains any instances of (*THEN). If it does not, the old
1180 optimizations are restored. It would be nice to do this on a per-group
1181 basis, but at the moment that is not feasible.
1182
1183 12. In some environments, the output of pcretest -C is CRLF terminated. This
1184 broke RunTest's code that checks for the link size. A single white space
1185 character after the value is now allowed for.
1186
1187 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1188 For "fr", it uses the Windows-specific input and output files.
1189
1190 14. If (*THEN) appeared in a group that was called recursively or as a
1191 subroutine, it did not work as intended. [But see next item.]
1192
1193 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1194 pattern fragments (but not containing any | characters). If A and B are
1195 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1196 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1197 D. In other words, Perl considers parentheses that do not contain any |
1198 characters to be part of a surrounding alternative, whereas PCRE was
1199 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1200 differently. PCRE now behaves in the same way as Perl, except in the case
1201 of subroutine/recursion calls such as (?1) which have in any case always
1202 been different (but PCRE had them first :-).
1203
1204 16. Related to 15 above: Perl does not treat the | in a conditional group as
1205 creating alternatives. Such a group is treated in the same way as an
1206 ordinary group without any | characters when processing (*THEN). PCRE has
1207 been changed to match Perl's behaviour.
1208
1209 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1210 RunGrepTest script failed.
1211
1212 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1213 inevitable for groups that contain captures, but it can lead to a lot of
1214 stack use in large patterns. The old behaviour has been restored for atomic
1215 groups that do not contain any capturing parentheses.
1216
1217 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1218 suppress the check for a minimum subject length at run time. (If it was
1219 given to pcre_exec() or pcre_dfa_exec() it did work.)
1220
1221 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1222 fail to work when decoding hex characters in data strings in EBCDIC
1223 environments.
1224
1225 21. It appears that in at least one Mac OS environment, the isxdigit() function
1226 is implemented as a macro that evaluates to its argument more than once,
1227 contravening the C 90 Standard (I haven't checked a later standard). There
1228 was an instance in pcretest which caused it to go wrong when processing
1229 \x{...} escapes in subject strings. The has been rewritten to avoid using
1230 things like p++ in the argument of isxdigit().
1231
1232
1233 Version 8.13 16-Aug-2011
1234 ------------------------
1235
1236 1. The Unicode data tables have been updated to Unicode 6.0.0.
1237
1238 2. Two minor typos in pcre_internal.h have been fixed.
1239
1240 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1241 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1242 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1243
1244 4. There were a number of related bugs in the code for matching backrefences
1245 caselessly in UTF-8 mode when codes for the characters concerned were
1246 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1247 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1248 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1249 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1250 2-byte code at the end of the subject (it thought there wasn't enough data
1251 left).
1252
1253 5. Comprehensive information about what went wrong is now returned by
1254 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1255 as the output vector has at least 2 elements. The offset of the start of
1256 the failing character and a reason code are placed in the vector.
1257
1258 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1259 now returned is for the first byte of the failing character, instead of the
1260 last byte inspected. This is an incompatible change, but I hope it is small
1261 enough not to be a problem. It makes the returned offset consistent with
1262 pcre_exec() and pcre_dfa_exec().
1263
1264 7. pcretest now gives a text phrase as well as the error number when
1265 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1266 failure, the offset and reason code are output.
1267
1268 8. When \R was used with a maximizing quantifier it failed to skip backwards
1269 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1270 back over a single character (\n). This seems wrong (because it treated the
1271 two characters as a single entity when going forwards), conflicts with the
1272 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1273 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1274 has been changed.
1275
1276 9. Some internal refactoring has changed the processing so that the handling
1277 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1278 time (the PCRE_DOTALL option was changed this way some time ago: version
1279 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1280 which was always a bit of a fudge. It also means that there is one less
1281 argument for the match() function, which reduces its stack requirements
1282 slightly. This change also fixes an incompatibility with Perl: the pattern
1283 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1284
1285 10. More internal refactoring has drastically reduced the number of recursive
1286 calls to match() for possessively repeated groups such as (abc)++ when
1287 using pcre_exec().
1288
1289 11. While implementing 10, a number of bugs in the handling of groups were
1290 discovered and fixed:
1291
1292 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1293 (a|)*(?1) gave a compile-time internal error.
1294 ((a|)+)+ did not notice that the outer group could match an empty string.
1295 (^a|^)+ was not marked as anchored.
1296 (.*a|.*)+ was not marked as matching at start or after a newline.
1297
1298 12. Yet more internal refactoring has removed another argument from the match()
1299 function. Special calls to this function are now indicated by setting a
1300 value in a variable in the "match data" data block.
1301
1302 13. Be more explicit in pcre_study() instead of relying on "default" for
1303 opcodes that mean there is no starting character; this means that when new
1304 ones are added and accidentally left out of pcre_study(), testing should
1305 pick them up.
1306
1307 14. The -s option of pcretest has been documented for ages as being an old
1308 synonym of -m (show memory usage). I have changed it to mean "force study
1309 for every regex", that is, assume /S for every regex. This is similar to -i
1310 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1311 using it. It makes it easier to run collections of tests with and without
1312 study enabled, and thereby test pcre_study() more easily. All the standard
1313 tests are now run with and without -s (but some patterns can be marked as
1314 "never study" - see 20 below).
1315
1316 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1317 restoration of the capturing data to the outer values was not happening
1318 correctly.
1319
1320 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1321 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1322 pattern had matched an empty string, and so incorrectly returned a no
1323 match.
1324
1325 17. There was optimizing code for the last branch of non-capturing parentheses,
1326 and also for the obeyed branch of a conditional subexpression, which used
1327 tail recursion to cut down on stack usage. Unfortunately, now that there is
1328 the possibility of (*THEN) occurring in these branches, tail recursion is
1329 no longer possible because the return has to be checked for (*THEN). These
1330 two optimizations have therefore been removed. [But see 8.20/11 above.]
1331
1332 18. If a pattern containing \R was studied, it was assumed that \R always
1333 matched two bytes, thus causing the minimum subject length to be
1334 incorrectly computed because \R can also match just one byte.
1335
1336 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1337 was incorrectly computed.
1338
1339 20. If /S is present twice on a test pattern in pcretest input, it now
1340 *disables* studying, thereby overriding the use of -s on the command line
1341 (see 14 above). This is necessary for one or two tests to keep the output
1342 identical in both cases.
1343
1344 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1345 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1346
1347 22. When an atomic group that contained a capturing parenthesis was
1348 successfully matched, but the branch in which it appeared failed, the
1349 capturing was not being forgotten if a higher numbered group was later
1350 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1351 group 1 to "a", when in fact it should be unset. This applied to multi-
1352 branched capturing and non-capturing groups, repeated or not, and also to
1353 positive assertions (capturing in negative assertions does not happen
1354 in PCRE) and also to nested atomic groups.
1355
1356 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1357 subject after a captured substring, to make it easier to tell which of a
1358 number of identical substrings has been captured.
1359
1360 24. The way atomic groups are processed by pcre_exec() has been changed so that
1361 if they are repeated, backtracking one repetition now resets captured
1362 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1363 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1364 "aaa". Previously, it would have been "a". As part of this code
1365 refactoring, the way recursive calls are handled has also been changed.
1366
1367 25. If an assertion condition captured any substrings, they were not passed
1368 back unless some other capturing happened later. For example, if
1369 (?(?=(a))a) was matched against "a", no capturing was returned.
1370
1371 26. When studying a pattern that contained subroutine calls or assertions,
1372 the code for finding the minimum length of a possible match was handling
1373 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1374 group 1 called group 2 while simultaneously a separate group 2 called group
1375 1). A stack overflow occurred in this case. I have fixed this by limiting
1376 the recursion depth to 10.
1377
1378 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1379 Fortmann. This supports explicit test numbers on the command line, and has
1380 argument validation and error reporting.
1381
1382 28. An instance of \X with an unlimited repeat could fail if at any point the
1383 first character it looked at was a mark character.
1384
1385 29. Some minor code refactoring concerning Unicode properties and scripts
1386 should reduce the stack requirement of match() slightly.
1387
1388 30. Added the '=' option to pcretest to check the setting of unused capturing
1389 slots at the end of the pattern, which are documented as being -1, but are
1390 not included in the return count.
1391
1392 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1393 compiled something random. Now it gives a compile-time error (as does
1394 Perl).
1395
1396 32. A *MARK encountered during the processing of a positive assertion is now
1397 recorded and passed back (compatible with Perl).
1398
1399 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1400 had alternative anchored branches, the search for a second match in a line
1401 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1402 matched the line "0102" twice. The same bug affected patterns that started
1403 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1404 twice.
1405
1406 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1407 does, and because of capturing effects, quantifying parenthesized
1408 assertions may at times be useful. Quantifiers are now allowed for
1409 parenthesized assertions.
1410
1411 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1412
1413 36. \g was being checked for fancy things in a character class, when it should
1414 just be a literal "g".
1415
1416 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1417 appearance of a nested POSIX class supersedes an apparent external class.
1418 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1419 unescaped square brackets may also appear as part of class names. For
1420 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1421 more like Perl. (But see 8.20/1 above.)
1422
1423 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1424 was because it thought it was \N{name}, which is not supported).
1425
1426 39. Add minix to OS list not supporting the -S option in pcretest.
1427
1428 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1429 cannot analyze patterns in sufficient detail to catch mutual recursions
1430 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1431 subgroup is called recursively as a subpattern for a second time at the
1432 same position in the subject string. In previous releases this might have
1433 been caught by the recursion limit, or it might have run out of stack.
1434
1435 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1436 happen only once. PCRE was, however incorrectly giving a compile time error
1437 "recursive call could loop indefinitely" because it cannot analyze the
1438 pattern in sufficient detail. The compile time test no longer happens when
1439 PCRE is compiling a conditional subpattern, but actual runaway loops are
1440 now caught at runtime (see 40 above).
1441
1442 42. It seems that Perl allows any characters other than a closing parenthesis
1443 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1444 has been changed to be the same.
1445
1446 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1447 as not to get warnings when autogen.sh is called. Also changed
1448 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1449
1450 44. To help people who use pcregrep to scan files containing exceedingly long
1451 lines, the following changes have been made:
1452
1453 (a) The default value of the buffer size parameter has been increased from
1454 8K to 20K. (The actual buffer used is three times this size.)
1455
1456 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1457 PCRE is built.
1458
1459 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1460 to be set at run time.
1461
1462 (d) Numerical values in pcregrep options can be followed by K or M, for
1463 example --buffer-size=50K.
1464
1465 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1466 given and the return code is set to 2.
1467
1468 45. Add a pointer to the latest mark to the callout data block.
1469
1470 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1471 partial match of an empty string instead of no match. This was specific to
1472 the use of ".".
1473
1474 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1475 complete match instead of a partial match. This bug was dependent on both
1476 the PCRE_UTF8 and PCRE_DOTALL options being set.
1477
1478 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1479 starting byte set, because \b was not being ignored.
1480
1481
1482 Version 8.12 15-Jan-2011
1483 ------------------------
1484
1485 1. Fixed some typos in the markup of the man pages, and wrote a script that
1486 checks for such things as part of the documentation building process.
1487
1488 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1489 --match-limit and --recursion-limit options (added for 8.11). In
1490 particular, this made one of the standard tests fail. (The integer value
1491 went into the wrong half of a long int.)
1492
1493 3. If the --colour option was given to pcregrep with -v (invert match), it
1494 did strange things, either producing crazy output, or crashing. It should,
1495 of course, ignore a request for colour when reporting lines that do not
1496 match.
1497
1498 4. Another pcregrep bug caused similar problems if --colour was specified with
1499 -M (multiline) and the pattern match finished with a line ending.
1500
1501 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1502 matched in multiline mode, the following line was shown as part of the
1503 match. This seems wrong, so I have changed it.
1504
1505 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1506 the check for further matches in the same line (so they could be coloured)
1507 to overrun the end of the current line. If another match was found, it was
1508 incorrectly shown (and then shown again when found in the next line).
1509
1510 7. If pcregrep was compiled under Windows, there was a reference to the
1511 function pcregrep_exit() before it was defined. I am assuming this was
1512 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1513 reported by a user. I've moved the definition above the reference.
1514
1515
1516 Version 8.11 10-Dec-2010
1517 ------------------------
1518
1519 1. (*THEN) was not working properly if there were untried alternatives prior
1520 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1521 backtracked to try for "b" instead of moving to the next alternative branch
1522 at the same level (in this case, to look for "c"). The Perl documentation
1523 is clear that when (*THEN) is backtracked onto, it goes to the "next
1524 alternative in the innermost enclosing group".
1525
1526 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1527 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1528 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1529 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1530 (*THEN).
1531
1532 3. If \s appeared in a character class, it removed the VT character from
1533 the class, even if it had been included by some previous item, for example
1534 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1535 of \s, but is part of the POSIX "space" class.)
1536
1537 4. A partial match never returns an empty string (because you can always
1538 match an empty string at the end of the subject); however the checking for
1539 an empty string was starting at the "start of match" point. This has been
1540 changed to the "earliest inspected character" point, because the returned
1541 data for a partial match starts at this character. This means that, for
1542 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1543 (previously it gave "no match").
1544
1545 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1546 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1547 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1548 has an implication that the given string is incomplete (because a partial
1549 match is preferred over a full match). For this reason, these items now
1550 give a partial match in this situation. [Aside: previously, the one case
1551 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1552 match rather than a full match, which was wrong by the old rules, but is
1553 now correct.]
1554
1555 6. There was a bug in the handling of #-introduced comments, recognized when
1556 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1557 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1558 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1559 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1560 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1561 places in pcre_compile().
1562
1563 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1564 comments when looking ahead for named forward references to subpatterns,
1565 the only newline sequence it recognized was NL. It now handles newlines
1566 according to the set newline convention.
1567
1568 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1569 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1570 cater for a lack of strerror(). These oversights have been fixed.
1571
1572 9. Added --match-limit and --recursion-limit to pcregrep.
1573
1574 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1575
1576 11. When the -o option was used, pcregrep was setting a return code of 1, even
1577 when matches were found, and --line-buffered was not being honoured.
1578
1579 12. Added an optional parentheses number to the -o and --only-matching options
1580 of pcregrep.
1581
1582 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1583 can match an empty string. The code to do it in pcretest and pcredemo
1584 needed fixing:
1585
1586 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1587 only one byte after an empty string match just before CRLF (this case
1588 just got forgotten; "any" and "anycrlf" were OK).
1589
1590 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1591 mode when an empty string match preceded an ASCII character followed by
1592 a non-ASCII character. (The code for advancing by one character rather
1593 than one byte was nonsense.)
1594
1595 (c) The pcredemo.c sample program did not have any code at all to handle
1596 the cases when CRLF is a valid newline sequence.
1597
1598 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1599 as a starting offset was within the subject string. There is now a new
1600 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1601 negative or greater than the length of the string. In order to test this,
1602 pcretest is extended to allow the setting of negative starting offsets.
1603
1604 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1605 starting offset points to the beginning of a UTF-8 character was
1606 unnecessarily clumsy. I tidied it up.
1607
1608 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1609 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1610
1611 17. Nobody had reported that the --include_dir option, which was added in
1612 release 7.7 should have been called --include-dir (hyphen, not underscore)
1613 for compatibility with GNU grep. I have changed it to --include-dir, but
1614 left --include_dir as an undocumented synonym, and the same for
1615 --exclude-dir, though that is not available in GNU grep, at least as of
1616 release 2.5.4.
1617
1618 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1619 characters from a string of bytes) have been redefined so as not to use
1620 loops, in order to improve performance in some environments. At the same
1621 time, I abstracted some of the common code into auxiliary macros to save
1622 repetition (this should not affect the compiled code).
1623
1624 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1625 compile-time error is now given if \c is not followed by an ASCII
1626 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1627 different, and any byte value is allowed.)
1628
1629 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1630 START_OPTIMIZE option, which is now allowed at compile time - but just
1631 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1632 to pcregrep and other applications that have no direct access to PCRE
1633 options. The new /Y option in pcretest sets this option when calling
1634 pcre_compile().
1635
1636 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1637 back references. Groups containing recursive back references were forced to
1638 be atomic by that change, but in the case of named groups, the amount of
1639 memory required was incorrectly computed, leading to "Failed: internal
1640 error: code overflow". This has been fixed.
1641
1642 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1643 pcretest.c, to avoid build problems in some Borland environments.
1644
1645
1646 Version 8.10 25-Jun-2010
1647 ------------------------
1648
1649 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1650 THEN.
1651
1652 2. (*ACCEPT) was not working when inside an atomic group.
1653
1654 3. Inside a character class, \B is treated as a literal by default, but
1655 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1656 causes the error). The code is unchanged, but I tidied the documentation.
1657
1658 4. Inside a character class, PCRE always treated \R and \X as literals,
1659 whereas Perl faults them if its -w option is set. I have changed PCRE so
1660 that it faults them when PCRE_EXTRA is set.
1661
1662 5. Added support for \N, which always matches any character other than
1663 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1664
1665 6. When compiling pcregrep with newer versions of gcc which may have
1666 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1667 declared with attribute warn_unused_result" were given. Just casting the
1668 result to (void) does not stop the warnings; a more elaborate fudge is
1669 needed. I've used a macro to implement this.
1670
1671 7. Minor change to pcretest.c to avoid a compiler warning.
1672
1673 8. Added four artifical Unicode properties to help with an option to make
1674 \s etc use properties (see next item). The new properties are: Xan
1675 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1676
1677 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1678 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1679 this option. Modified pcretest to add /W to test this facility. Added
1680 REG_UCP to make it available via the POSIX interface.
1681
1682 10. Added --line-buffered to pcregrep.
1683
1684 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1685 studied, and the match started with a letter with a code point greater than
1686 127 whose first byte was different to the first byte of the other case of
1687 the letter, the other case of this starting letter was not recognized
1688 (#976).
1689
1690 12. If a pattern that was studied started with a repeated Unicode property
1691 test, for example, \p{Nd}+, there was the theoretical possibility of
1692 setting up an incorrect bitmap of starting bytes, but fortunately it could
1693 not have actually happened in practice until change 8 above was made (it
1694 added property types that matched character-matching opcodes).
1695
1696 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1697 possible starting bytes for non-anchored patterns.
1698
1699 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1700 \R, and also a number of cases that involve Unicode properties, both
1701 explicit and implicit when PCRE_UCP is set.
1702
1703 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1704 input, it could crash or give wrong results if characters with values
1705 greater than 0xc0 were present in the subject string. (Detail: it assumed
1706 UTF-8 input when processing these items.)
1707
1708 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1709 size_t is 64-bit (#991).
1710
1711 17. Added a check for running out of memory when PCRE is compiled with
1712 --disable-stack-for-recursion (#990).
1713
1714 18. If the last data line in a file for pcretest does not have a newline on
1715 the end, a newline was missing in the output.
1716
1717 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1718 less than 128) in its various bitmaps. However, there is a facility for
1719 generating tables according to the current locale when PCRE is compiled. It
1720 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1721 characters, are recognized by isspace() and therefore were getting set in
1722 these tables, and indeed these tables seem to approximate to ISO 8859. This
1723 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1724 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1725 which of course cannot start UTF-8 characters. I have changed the code so
1726 that only real ASCII characters (less than 128) and the correct starting
1727 bytes for UTF-8 encodings are set for characters greater than 127 when in
1728 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1729 altogether.)
1730
1731 20. Added the /T option to pcretest so as to be able to run tests with non-
1732 standard character tables, thus making it possible to include the tests
1733 used for 19 above in the standard set of tests.
1734
1735 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1736 reference to a subpattern the other side of a comment that contains an
1737 opening parenthesis caused either an internal compiling error, or a
1738 reference to the wrong subpattern.
1739
1740
1741 Version 8.02 19-Mar-2010
1742 ------------------------
1743
1744 1. The Unicode data tables have been updated to Unicode 5.2.0.
1745
1746 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1747 configured.
1748
1749 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1750 original author of that file, following a query about its status.
1751
1752 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1753 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1754
1755 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1756 quantifier applied to a forward-referencing subroutine call, could compile
1757 incorrect code or give the error "internal error: previously-checked
1758 referenced subpattern not found".
1759
1760 6. Both MS Visual Studio and Symbian OS have problems with initializing
1761 variables to point to external functions. For these systems, therefore,
1762 pcre_malloc etc. are now initialized to local functions that call the
1763 relevant global functions.
1764
1765 7. There were two entries missing in the vectors called coptable and poptable
1766 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1767 I've fixed the data, and added a kludgy way of testing at compile time that
1768 the lengths are correct (equal to the number of opcodes).
1769
1770 8. Following on from 7, I added a similar kludge to check the length of the
1771 eint vector in pcreposix.c.
1772
1773 9. Error texts for pcre_compile() are held as one long string to avoid too
1774 much relocation at load time. To find a text, the string is searched,
1775 counting zeros. There was no check for running off the end of the string,
1776 which could happen if a new error number was added without updating the
1777 string.
1778
1779 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1780
1781 11. \K was not working if it appeared in an atomic group or in a group that
1782 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1783 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1784 the assertion is positive, but not if it is negative.
1785
1786 12. Change 11 fortuitously reduced the size of the stack frame used in the
1787 "match()" function of pcre_exec.c by one pointer. Forthcoming
1788 implementation of support for (*MARK) will need an extra pointer on the
1789 stack; I have reserved it now, so that the stack frame size does not
1790 decrease.
1791
1792 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1793 item in branch that calls a recursion is a subroutine call - as in the
1794 second branch in the above example - was incorrectly given the compile-
1795 time error "recursive call could loop indefinitely" because pcre_compile()
1796 was not correctly checking the subroutine for matching a non-empty string.
1797
1798 14. The checks for overrunning compiling workspace could trigger after an
1799 overrun had occurred. This is a "should never occur" error, but it can be
1800 triggered by pathological patterns such as hundreds of nested parentheses.
1801 The checks now trigger 100 bytes before the end of the workspace.
1802
1803 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1804
1805
1806 Version 8.01 19-Jan-2010
1807 ------------------------
1808
1809 1. If a pattern contained a conditional subpattern with only one branch (in
1810 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1811 computed the wrong minimum data length (which is of course zero for such
1812 subpatterns). This could cause incorrect "no match" results.
1813
1814 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1815 the pattern is reset in the first branch, pcre_compile() failed with
1816 "internal error: code overflow at offset...". This happened only when
1817 the reset was to the original external option setting. (An optimization
1818 abstracts leading options settings into an external setting, which was the
1819 cause of this.)
1820
1821 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1822 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1823 assertion pattern did not match (meaning that the assertion was true), it
1824 was incorrectly treated as false if the SKIP had been reached during the
1825 matching. This also applied to assertions used as conditions.
1826
1827 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1828 assertion subpattern, including such a pattern used as a condition,
1829 unpredictable results occurred, instead of the error return
1830 PCRE_ERROR_DFA_UITEM.
1831
1832 5. The C++ GlobalReplace function was not working like Perl for the special
1833 situation when an empty string is matched. It now does the fancy magic
1834 stuff that is necessary.
1835
1836 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1837 removed. (These were left over from very, very early versions of PCRE.)
1838
1839 7. Some cosmetic changes to the code to make life easier when compiling it
1840 as part of something else:
1841
1842 (a) Change DEBUG to PCRE_DEBUG.
1843
1844 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1845 called "current" as "current_branch", to prevent a collision with the
1846 Linux macro when compiled as a kernel module.
1847
1848 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1849 prevent a collision with the Linux macro when compiled as a kernel
1850 module.
1851
1852 8. In pcre_compile() there are some checks for integer overflows that used to
1853 cast potentially large values to (double). This has been changed to that
1854 when building, a check for int64_t is made, and if it is found, it is used
1855 instead, thus avoiding the use of floating point arithmetic. (There is no
1856 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1857 double.
1858
1859 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1860 2005 (difference between two addresses compared to an unsigned value).
1861
1862 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1863 custom one, because of the following reported problem in Windows:
1864
1865 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1866 under Win32.
1867 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1868 therefore missing the function definition.
1869 - The compiler thus generates a "C" signature for the test function.
1870 - The linker fails to find the "C" function.
1871 - PCRE fails to configure if asked to do so against libbz2.
1872
1873 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1874 messages were output:
1875
1876 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1877 rerunning libtoolize, to keep the correct libtool macros in-tree.
1878 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1879
1880 I have done both of these things.
1881
1882 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1883 most of the time, it *can* run out if it is given a pattern that contains a
1884 runaway infinite recursion. I updated the discussion in the pcrestack man
1885 page.
1886
1887 13. Now that we have gone to the x.xx style of version numbers, the minor
1888 version may start with zero. Using 08 or 09 is a bad idea because users
1889 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1890 interpreted as invalid octal numbers. I've updated the previous comment in
1891 configure.ac, and also added a check that gives an error if 08 or 09 are
1892 used.
1893
1894 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1895 causing partial matching to fail when the end of the subject matched \W
1896 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1897
1898 15. There were some discrepancies between the declarations in pcre_internal.h
1899 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1900 their definitions. The declarations used "const uschar *" and the
1901 definitions used USPTR. Even though USPTR is normally defined as "const
1902 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1903 reported that: "This difference in casting confuses some C++ compilers, for
1904 example, SunCC recognizes above declarations as different functions and
1905 generates broken code for hbpcre." I have changed the declarations to use
1906 USPTR.
1907
1908 16. GNU libtool is named differently on some systems. The autogen.sh script now
1909 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1910 (FreeBSD).
1911
1912 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1913 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1914 comment: "Figure out how to create a longlong from a string: strtoll and
1915 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1916 instance, but it only takes 2 args instead of 3!"
1917
1918 18. A subtle bug concerned with back references has been fixed by a change of
1919 specification, with a corresponding code fix. A pattern such as
1920 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1921 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1922 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1923 same bug. Such groups have to be quantified to be useful, or contained
1924 inside another quantified group. (If there's no repetition, the reference
1925 can never match.) The problem arises because, having left the group and
1926 moved on to the rest of the pattern, a later failure that backtracks into
1927 the group uses the captured value from the final iteration of the group
1928 rather than the correct earlier one. I have fixed this in PCRE by forcing
1929 any group that contains a reference to itself to be an atomic group; that
1930 is, there cannot be any backtracking into it once it has completed. This is
1931 similar to recursive and subroutine calls.
1932
1933
1934 Version 8.00 19-Oct-09
1935 ----------------------
1936
1937 1. The table for translating pcre_compile() error codes into POSIX error codes
1938 was out-of-date, and there was no check on the pcre_compile() error code
1939 being within the table. This could lead to an OK return being given in
1940 error.
1941
1942 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1943 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1944 in a Windows environment.
1945
1946 3. The pcregrep --count option prints the count for each file even when it is
1947 zero, as does GNU grep. However, pcregrep was also printing all files when
1948 --files-with-matches was added. Now, when both options are given, it prints
1949 counts only for those files that have at least one match. (GNU grep just
1950 prints the file name in this circumstance, but including the count seems
1951 more useful - otherwise, why use --count?) Also ensured that the
1952 combination -clh just lists non-zero counts, with no names.
1953
1954 4. The long form of the pcregrep -F option was incorrectly implemented as
1955 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1956 but it seems right to fix it, and I didn't think it was worth preserving
1957 the old behaviour.
1958
1959 5. The command line items --regex=pattern and --regexp=pattern were not
1960 recognized by pcregrep, which required --regex pattern or --regexp pattern
1961 (with a space rather than an '='). The man page documented the '=' forms,
1962 which are compatible with GNU grep; these now work.
1963
1964 6. No libpcreposix.pc file was created for pkg-config; there was just
1965 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1966
1967 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1968 when UCP support is not needed, by modifying the Python script that
1969 generates it from Unicode data files. This should not matter if the module
1970 is correctly used as a library, but I received one complaint about 50K of
1971 unwanted data. My guess is that the person linked everything into his
1972 program rather than using a library. Anyway, it does no harm.
1973
1974 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1975 was a minimum greater than 1 for a wide character in a possessive
1976 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1977 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1978 character. Chaos in the form of incorrect output or a compiling loop could
1979 result.
1980
1981 9. The restrictions on what a pattern can contain when partial matching is
1982 requested for pcre_exec() have been removed. All patterns can now be
1983 partially matched by this function. In addition, if there are at least two
1984 slots in the offset vector, the offset of the earliest inspected character
1985 for the match and the offset of the end of the subject are set in them when
1986 PCRE_ERROR_PARTIAL is returned.
1987
1988 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1989 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1990 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1991 and may be more useful for multi-segment matching.
1992
1993 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1994 used to be given if ever the end of the subject was reached; now it is
1995 given only if matching could not proceed because another character was
1996 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1997 string "Z", which now yields "no match" instead of "partial match". In the
1998 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1999 final character ended with (*FAIL).
2000
2001 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2002 if the pattern had a "must contain" character that was already found in the
2003 earlier partial match, unless partial matching was again requested. For
2004 example, with the pattern /dog.(body)?/, the "must contain" character is
2005 "g". If the first part-match was for the string "dog", restarting with
2006 "sbody" failed. This bug has been fixed.
2007
2008 13. The string returned by pcre_dfa_exec() after a partial match has been
2009 changed so that it starts at the first inspected character rather than the
2010 first character of the match. This makes a difference only if the pattern
2011 starts with a lookbehind assertion or \b or \B (\K is not supported by
2012 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2013 matching functions compatible, and I think it's the right thing to do.
2014
2015 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2016 so that the demonstration program is easily available in environments where
2017 PCRE has not been installed from source.
2018
2019 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2020 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2021 library.
2022
2023 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2024 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2025 is not the first non-POSIX option to be added. Clearly some people find
2026 these options useful.
2027
2028 17. If a caller to the POSIX matching function regexec() passes a non-zero
2029 value for nmatch with a NULL value for pmatch, the value of
2030 nmatch is forced to zero.
2031
2032 18. RunGrepTest did not have a test for the availability of the -u option of
2033 the diff command, as RunTest does. It now checks in the same way as
2034 RunTest, and also checks for the -b option.
2035
2036 19. If an odd number of negated classes containing just a single character
2037 interposed, within parentheses, between a forward reference to a named
2038 subpattern and the definition of the subpattern, compilation crashed with
2039 an internal error, complaining that it could not find the referenced
2040 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2041 [The bug was that it was starting one character too far in when skipping
2042 over the character class, thus treating the ] as data rather than
2043 terminating the class. This meant it could skip too much.]
2044
2045 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2046 /g option in pcretest when the pattern contains \K, which makes it possible
2047 to have an empty string match not at the start, even when the pattern is
2048 anchored. Updated pcretest and pcredemo to use this option.
2049
2050 21. If the maximum number of capturing subpatterns in a recursion was greater
2051 than the maximum at the outer level, the higher number was returned, but
2052 with unset values at the outer level. The correct (outer level) value is
2053 now given.
2054
2055 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2056 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2057 make it do so. The string so far is captured, making this feature
2058 compatible with Perl.
2059
2060 23. The tests have been re-organized, adding tests 11 and 12, to make it
2061 possible to check the Perl 5.10 features against Perl 5.10.
2062
2063 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2064 pattern matches a fixed length string. PCRE did not allow this; now it
2065 does. Neither allows recursion.
2066
2067 25. I finally figured out how to implement a request to provide the minimum
2068 length of subject string that was needed in order to match a given pattern.
2069 (It was back references and recursion that I had previously got hung up
2070 on.) This code has now been added to pcre_study(); it finds a lower bound
2071 to the length of subject needed. It is not necessarily the greatest lower
2072 bound, but using it to avoid searching strings that are too short does give
2073 some useful speed-ups. The value is available to calling programs via
2074 pcre_fullinfo().
2075
2076 26. While implementing 25, I discovered to my embarrassment that pcretest had
2077 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2078 study optimizations had never been tested with that matching function.
2079 Oops. What is worse, even when it was passed study data, there was a bug in
2080 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2081 were also very few tests of studied patterns with pcre_dfa_exec().
2082
2083 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2084 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2085 on the other side of the coin, they are no longer allowed to have different
2086 names, because these cannot be distinguished in PCRE, and this has caused
2087 confusion. (This is a difference from Perl.)
2088
2089 28. When duplicate subpattern names are present (necessarily with different
2090 numbers, as required by 27 above), and a test is made by name in a
2091 conditional pattern, either for a subpattern having been matched, or for
2092 recursion in such a pattern, all the associated numbered subpatterns are
2093 tested, and the overall condition is true if the condition is true for any
2094 one of them. This is the way Perl works, and is also more like the way
2095 testing by number works.
2096
2097
2098 Version 7.9 11-Apr-09
2099 ---------------------
2100
2101 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2102 (pcretest), all targets were linked against these libraries. This included
2103 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2104 libraries. This caused unwanted dependencies to be created. This problem
2105 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2106 pcretest is linked with readline.
2107
2108 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2109 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2110 moved outside it again, because FALSE and TRUE are already defined in AIX,
2111 but BOOL is not.
2112
2113 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2114 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2115
2116 4. The pcregrep documentation said spaces were inserted as well as colons (or
2117 hyphens) following file names and line numbers when outputting matching
2118 lines. This is not true; no spaces are inserted. I have also clarified the
2119 wording for the --colour (or --color) option.
2120
2121 5. In pcregrep, when --colour was used with -o, the list of matching strings
2122 was not coloured; this is different to GNU grep, so I have changed it to be
2123 the same.
2124
2125 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2126 each matching line was coloured. Now it goes on to look for further matches
2127 of any of the test patterns, which is the same behaviour as GNU grep.
2128
2129 7. A pattern that could match an empty string could cause pcregrep to loop; it
2130 doesn't make sense to accept an empty string match in pcregrep, so I have
2131 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2132 seems to be how GNU grep behaves. [But see later change 40 for release
2133 8.33.]
2134
2135 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2136 start or after a newline", because the conditional assertion was not being
2137 correctly handled. The rule now is that both the assertion and what follows
2138 in the first alternative must satisfy the test.
2139
2140 9. If auto-callout was enabled in a pattern with a conditional group whose
2141 condition was an assertion, PCRE could crash during matching, both with
2142 pcre_exec() and pcre_dfa_exec().
2143
2144 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2145 used for matching.
2146
2147 11. Unicode property support in character classes was not working for
2148 characters (bytes) greater than 127 when not in UTF-8 mode.
2149
2150 12. Added the -M command line option to pcretest.
2151
2152 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2153
2154 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2155
2156 16. Added comments and documentation about mis-use of no_arg in the C++
2157 wrapper.
2158
2159 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2160 from Martin Jerabek that uses macro names for all relevant character and
2161 string constants.
2162
2163 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2164 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2165 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2166 these, but not everybody uses configure.
2167
2168 19. A conditional group that had only one branch was not being correctly
2169 recognized as an item that could match an empty string. This meant that an
2170 enclosing group might also not be so recognized, causing infinite looping
2171 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2172 with the subject "ab", where knowledge that the repeated group can match
2173 nothing is needed in order to break the loop.
2174
2175 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2176 exec(), but without supplying a callout function, matching went wrong.
2177
2178 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2179 leak if the size of the offset vector was greater than 30. When the vector
2180 is smaller, the saved offsets during recursion go onto a local stack
2181 vector, but for larger vectors malloc() is used. It was failing to free
2182 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2183 error, in fact).
2184
2185 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2186 heapframe that is used only when UTF-8 support is enabled. This caused no
2187 problem, but was untidy.
2188
2189 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2190 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2191 included within another project.
2192
2193 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2194 slightly modified by me:
2195
2196 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2197 not building pcregrep.
2198
2199 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2200 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2201
2202 25. Forward references, both numeric and by name, in patterns that made use of
2203 duplicate group numbers, could behave incorrectly or give incorrect errors,
2204 because when scanning forward to find the reference group, PCRE was not
2205 taking into account the duplicate group numbers. A pattern such as
2206 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2207
2208 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2209 the feature of a custom pointer more persuasive (as requested by a user).
2210
2211 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2212 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2213 pre-defined.
2214
2215 28. Added support for (*UTF8) at the start of a pattern.
2216
2217 29. Arrange for flags added by the "release type" setting in CMake to be shown
2218 in the configuration summary.
2219
2220
2221 Version 7.8 05-Sep-08
2222 ---------------------
2223
2224 1. Replaced UCP searching code with optimized version as implemented for Ad
2225 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2226 stage table and inline lookup instead of a function, giving speed ups of 2
2227 to 5 times on some simple patterns that I tested. Permission was given to
2228 distribute the MultiStage2.py script that generates the tables (it's not in
2229 the tarball, but is in the Subversion repository).
2230
2231 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2232 scripts.
2233
2234 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2235 a group with a zero qualifier. The result of the study could be incorrect,
2236 or the function might crash, depending on the pattern.
2237
2238 4. Caseless matching was not working for non-ASCII characters in back
2239 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2240 It now works when Unicode Property Support is available.
2241
2242 5. In pcretest, an escape such as \x{de} in the data was always generating
2243 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2244 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2245 truncation.
2246
2247 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2248
2249 7. Added two (int) casts to pcregrep when printing the difference of two
2250 pointers, in case they are 64-bit values.
2251
2252 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2253 test 2 if it fails.
2254
2255 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2256 and a #define of that name to empty if it is not externally set. This is to
2257 allow users of MSVC to set it if necessary.
2258
2259 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2260 the convenience functions in the pcre_get.c source file.
2261
2262 11. An option change at the start of a pattern that had top-level alternatives
2263 could cause overwriting and/or a crash. This command provoked a crash in
2264 some environments:
2265
2266 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2267
2268 This potential security problem was recorded as CVE-2008-2371.
2269
2270 12. For a pattern where the match had to start at the beginning or immediately
2271 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2272 pcre_dfa_exec() could read past the end of the passed subject if there was
2273 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2274 pcretest so that it places the subject at the end of its malloc-ed buffer.
2275
2276 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2277 exec() might read past the end of the data buffer in UTF-8 mode.
2278
2279 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2280 the data contained the byte 0x85 as part of a UTF-8 character within its
2281 first line. This applied both to normal and DFA matching.
2282
2283 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2284 /^[^d]*?$/8 failed to match "abc".
2285
2286 16. Added a missing copyright notice to pcrecpp_internal.h.
2287
2288 17. Make it more clear in the documentation that values returned from
2289 pcre_exec() in ovector are byte offsets, not character counts.
2290
2291 18. Tidied a few places to stop certain compilers from issuing warnings.
2292
2293 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2294 supplied by Stefan Weber. I made a further small update for 7.8 because
2295 there is a change of source arrangements: the pcre_searchfuncs.c module is
2296 replaced by pcre_ucd.c.
2297
2298
2299 Version 7.7 07-May-08
2300 ---------------------
2301
2302 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2303 a string to a long long, pretend we don't even have a long long." This is
2304 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2305
2306 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2307 pre-7.6 versions, which defined a global no_arg variable instead of putting
2308 it in the RE class. (See also #8 below.)
2309
2310 3. Remove a line of dead code, identified by coverity and reported by Nuno
2311 Lopes.
2312
2313 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2314
2315 (1) The include/exclude patterns were being applied to the whole pathnames
2316 of files, instead of just to the final components.
2317
2318 (2) If there was more than one level of directory, the subdirectories were
2319 skipped unless they satisfied the include/exclude conditions. This is
2320 inconsistent with GNU grep (and could even be seen as contrary to the
2321 pcregrep specification - which I improved to make it absolutely clear).
2322 The action now is always to scan all levels of directory, and just
2323 apply the include/exclude patterns to regular files.
2324
2325 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2326 --exclude_dir in the tests to avoid scanning .svn directories.
2327
2328 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2329 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2330 doesn't support NULs in patterns.
2331
2332 7. Added some missing "const"s to declarations of static tables in
2333 pcre_compile.c and pcre_dfa_exec.c.
2334
2335 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2336 caused by fix #2 above. (Subsequently also a second patch to fix the
2337 first patch. And a third patch - this was a messy problem.)
2338
2339 9. Applied Craig's patch to remove the use of push_back().
2340
2341 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2342 matching function regexec().
2343
2344 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2345 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2346 references. PCRE supports relative numbers with this syntax (I don't think
2347 Oniguruma does).
2348
2349 12. Previously, a group with a zero repeat such as (...){0} was completely
2350 omitted from the compiled regex. However, this means that if the group
2351 was called as a subroutine from elsewhere in the pattern, things went wrong
2352 (an internal error was given). Such groups are now left in the compiled
2353 pattern, with a new opcode that causes them to be skipped at execution
2354 time.
2355
2356 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2357 to the way PCRE behaves:
2358
2359 (a) A lone ] character is dis-allowed (Perl treats it as data).
2360
2361 (b) A back reference to an unmatched subpattern matches an empty string
2362 (Perl fails the current match path).
2363
2364 (c) A data ] in a character class must be notated as \] because if the
2365 first data character in a class is ], it defines an empty class. (In
2366 Perl it is not possible to have an empty class.) The empty class []
2367 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2368 The negative empty class [^] matches any one character, independently
2369 of the DOTALL setting.
2370
2371 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2372 non-existent subpattern following a character class starting with ']' and
2373 containing () gave an internal compiling error instead of "reference to
2374 non-existent subpattern". Fortunately, when the pattern did exist, the
2375 compiled code was correct. (When scanning forwards to check for the
2376 existence of the subpattern, it was treating the data ']' as terminating
2377 the class, so got the count wrong. When actually compiling, the reference
2378 was subsequently set up correctly.)
2379
2380 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2381 it was being rejected as not supported by pcre_dfa_exec(), even though
2382 other assertions are supported. I have made pcre_dfa_exec() support
2383 (*FAIL).
2384
2385 16. The implementation of 13c above involved the invention of a new opcode,
2386 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2387 cannot be changed at match time, I realized I could make a small
2388 improvement to matching performance by compiling OP_ALLANY instead of
2389 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2390 on the OP_ANY path.
2391
2392 17. Compiling pcretest on Windows with readline support failed without the
2393 following two fixes: (1) Make the unistd.h include conditional on
2394 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2395
2396 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2397 ncurses library to be included for pcretest when ReadLine support is
2398 requested, but also to allow for it to be overridden. This patch came from
2399 Daniel Bergström.
2400
2401 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2402 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2403 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2404 spotting this.
2405
2406
2407 Version 7.6 28-Jan-08
2408 ---------------------
2409
2410 1. A character class containing a very large number of characters with
2411 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2412 overflow.
2413
2414 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2415 HAVE_LONG_LONG is not defined.
2416
2417 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2418 bring them up to date and include new features. This patch includes:
2419
2420 - Fixed PH's badly added libz and libbz2 support.
2421 - Fixed a problem with static linking.
2422 - Added pcredemo. [But later removed - see 7 below.]
2423 - Fixed dftables problem and added an option.
2424 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2425 HAVE_LONG_LONG.
2426 - Added readline support for pcretest.
2427 - Added an listing of the option settings after cmake has run.
2428
2429 4. A user submitted a patch to Makefile that makes it easy to create
2430 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2431 Makefile.am that cause it to include this special target, without
2432 affecting anything else. Note that the same mingw target plus all
2433 the other distribution libraries and programs are now supported
2434 when configuring with CMake (see 6 below) instead of with
2435 Configure/Make.
2436
2437 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2438 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2439 exported in the Windows port". It has not yet been confirmed that the patch
2440 solves the problem, but it does no harm.
2441
2442 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2443 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2444 with CMake, and also correct the comment about stack recursion.
2445
2446 7. Remove the automatic building of pcredemo from the ./configure system and
2447 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2448 of a program that users should build themselves after PCRE is installed, so
2449 building it automatically is not really right. What is more, it gave
2450 trouble in some build environments.
2451
2452 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2453
2454
2455 Version 7.5 10-Jan-08
2456 ---------------------
2457
2458 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2459 values in parens when parsing an RE using the C++ wrapper."
2460
2461 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2462 Characters greater than 255 were excluded from the class instead of being
2463 included.
2464
2465 3. The same bug as (2) above applied to negated POSIX classes such as
2466 [:^space:].
2467
2468 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2469 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2470 I have changed it.
2471
2472 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2473 first named subpattern) and a construct such as (?&a) would reference the
2474 first named subpattern whose name started with "a" (in other words, the
2475 length check was missing). Both these problems are fixed. "Subpattern name
2476 expected" is now given for (?&) (a zero-length name), and this patch also
2477 makes it give the same error for \k'' (previously it complained that that
2478 was a reference to a non-existent subpattern).
2479
2480 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2481 this is right because (?- can be followed by option settings as well as by
2482 digits. I have, however, made the messages clearer.
2483
2484 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2485 than the number used in the conditional) now cause a compile-time error.
2486 This is actually not compatible with Perl, which accepts such patterns, but
2487 treats the conditional as always being FALSE (as PCRE used to), but it
2488 seems to me that giving a diagnostic is better.
2489
2490 8. Change "alphameric" to the more common word "alphanumeric" in comments
2491 and messages.
2492
2493 9. Fix two occurrences of "backslash" in comments that should have been
2494 "backspace".
2495
2496 10. Remove two redundant lines of code that can never be obeyed (their function
2497 was moved elsewhere).
2498
2499 11. The program that makes PCRE's Unicode character property table had a bug
2500 which caused it to generate incorrect table entries for sequences of
2501 characters that have the same character type, but are in different scripts.
2502 It amalgamated them into a single range, with the script of the first of
2503 them. In other words, some characters were in the wrong script. There were
2504 thirteen such cases, affecting characters in the following ranges:
2505
2506 U+002b0 - U+002c1
2507 U+0060c - U+0060d
2508 U+0061e - U+00612
2509 U+0064b - U+0065e
2510 U+0074d - U+0076d
2511 U+01800 - U+01805
2512 U+01d00 - U+01d77
2513 U+01d9b - U+01dbf
2514 U+0200b - U+0200f
2515 U+030fc - U+030fe
2516 U+03260 - U+0327f
2517 U+0fb46 - U+0fbb1
2518 U+10450 - U+1049d
2519
2520 12. The -o option (show only the matching part of a line) for pcregrep was not
2521 compatible with GNU grep in that, if there was more than one match in a
2522 line, it showed only the first of them. It now behaves in the same way as
2523 GNU grep.
2524
2525 13. If the -o and -v options were combined for pcregrep, it printed a blank
2526 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2527 does the same. The return code can be used to tell if there were any
2528 non-matching lines.
2529
2530 14. Added --file-offsets and --line-offsets to pcregrep.
2531
2532 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2533 infinitely looping recursion. The bug was that positive lookaheads were not
2534 being skipped when checking for a possible empty match (negative lookaheads
2535 and both kinds of lookbehind were skipped).
2536
2537 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2538 inclusion of <windows.h> to before rather than after the definition of
2539 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2540
2541 17. Specifying a possessive quantifier with a specific limit for a Unicode
2542 character property caused pcre_compile() to compile bad code, which led at
2543 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2544 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2545 caused the error; without that there was no problem.
2546
2547 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2548
2549 19. Added --enable-pcretest-libreadline.
2550
2551 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2552 RE::GlobalReplace(). As a result, the number of replacements returned was
2553 double what it should be. I removed one of the increments, but Craig sent a
2554 later patch that removed the other one (the right fix) and added unit tests
2555 that check the return values (which was not done before).
2556
2557 21. Several CMake things:
2558
2559 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2560 the names libpcre and libpcreposix, not just pcre and pcreposix.
2561
2562 (2) The above change means that pcretest and pcregrep are now correctly
2563 linked with the newly-built libraries, not previously installed ones.
2564
2565 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2566
2567 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2568 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2569 UTF-8 newline character). The key issue is that the pattern starts .*;
2570 this means that the match must be either at the beginning, or after a
2571 newline. The bug was in the code for advancing after a failed match and
2572 checking that the new position followed a newline. It was not taking
2573 account of UTF-8 characters correctly.
2574
2575 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2576 character classes. PCRE was not treating the sequence [:...:] as a
2577 character class unless the ... were all letters. Perl, however, seems to
2578 allow any characters between [: and :], though of course it rejects as
2579 unknown any "names" that contain non-letters, because all the known class
2580 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2581 for example, whereas PCRE did not - it did not recognize a POSIX character
2582 class. This seemed a bit dangerous, so the code has been changed to be
2583 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2584 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2585 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2586 Perl does, and where it didn't before.
2587
2588 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2589 Windows environments %n is disabled by default.
2590
2591
2592 Version 7.4 21-Sep-07
2593 ---------------------
2594
2595 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2596 means that a class such as [\s] counted as "explicit reference to CR or
2597 LF". That isn't really right - the whole point of the change was to try to
2598 help when there was an actual mention of one of the two characters. So now
2599 the change happens only if \r or \n (or a literal CR or LF) character is
2600 encountered.
2601
2602 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2603 of both had grown to the point where there were only 3 bits left.
2604 Fortunately, there was spare space in the data structure, and so I have
2605 moved the internal flags into a new 16-bit field to free up more option
2606 bits.
2607
2608 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2609 but did not set the internal JCHANGED flag - either of these is enough to
2610 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2611 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2612 start sets both bits.
2613
2614 4. Added options (at build time, compile time, exec time) to change \R from
2615 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2616
2617 5. doc/pcresyntax.html was missing from the distribution.
2618
2619 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2620 compatibility, even though it is no longer used.
2621
2622 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2623 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2624 windows.h file is present (where different names are used). [This was
2625 reversed later after testing - see 16 below.]
2626
2627 8. Changed all #include <config.h> to #include "config.h". There were also
2628 some further <pcre.h> cases that I changed to "pcre.h".
2629
2630 9. When pcregrep was used with the --colour option, it missed the line ending
2631 sequence off the lines that it output.
2632
2633 10. It was pointed out to me that arrays of string pointers cause lots of
2634 relocations when a shared library is dynamically loaded. A technique of
2635 using a single long string with a table of offsets can drastically reduce
2636 these. I have refactored PCRE in four places to do this. The result is
2637 dramatic:
2638
2639 Originally: 290
2640 After changing UCP table: 187
2641 After changing error message table: 43
2642 After changing table of "verbs" 36
2643 After changing table of Posix names 22
2644
2645 Thanks to the folks working on Gregex for glib for this insight.
2646
2647 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2648 unicode-properties was also set.
2649
2650 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2651
2652 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2653 checked only for CRLF.
2654
2655 14. Added casts to pcretest.c to avoid compiler warnings.
2656
2657 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2658
2659 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2660 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2661 entirely. This removes changes made in 7 above.
2662
2663 17. The CMake files have been updated, and there is now more information about
2664 building with CMake in the NON-UNIX-USE document.
2665
2666
2667 Version 7.3 28-Aug-07
2668 ---------------------
2669
2670 1. In the rejigging of the build system that eventually resulted in 7.1, the
2671 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2672 brackets there is not right, since it causes compilers to look for an
2673 installed pcre.h, not the version that is in the source that is being
2674 compiled (which of course may be different). I have changed it back to:
2675
2676 #include "pcre.h"
2677
2678 I have a vague recollection that the change was concerned with compiling in
2679 different directories, but in the new build system, that is taken care of
2680 by the VPATH setting the Makefile.
2681
2682 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2683 when the subject happened to end in the byte 0x85 (e.g. if the last
2684 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2685 characters but of course it shouldn't be taken as a newline when it is part
2686 of another character. The bug was that, for an unlimited repeat of . in
2687 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2688 characters when looking for a newline.
2689
2690 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2691
2692 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2693 in debug output.
2694
2695 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2696 long printing in the pcrecpp unittest when running under MinGW.
2697
2698 6. ESC_K was left out of the EBCDIC table.
2699
2700 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2701 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2702 limit also applies to "virtual nesting" when a pattern is recursive, and in
2703 this case 1000 isn't so big. I have been able to remove this limit at the
2704 expense of backing off one optimization in certain circumstances. Normally,
2705 when pcre_exec() would call its internal match() function recursively and
2706 immediately return the result unconditionally, it uses a "tail recursion"
2707 feature to save stack. However, when a subpattern that can match an empty
2708 string has an unlimited repetition quantifier, it no longer makes this
2709 optimization. That gives it a stack frame in which to save the data for
2710 checking that an empty string has been matched. Previously this was taken
2711 from the 1000-entry workspace that had been reserved. So now there is no
2712 explicit limit, but more stack is used.
2713
2714 8. Applied Daniel's patches to solve problems with the import/export magic
2715 syntax that is required for Windows, and which was going wrong for the
2716 pcreposix and pcrecpp parts of the library. These were overlooked when this
2717 problem was solved for the main library.
2718
2719 9. There were some crude static tests to avoid integer overflow when computing
2720 the size of patterns that contain repeated groups with explicit upper
2721 limits. As the maximum quantifier is 65535, the maximum group length was
2722 set at 30,000 so that the product of these two numbers did not overflow a
2723 32-bit integer. However, it turns out that people want to use groups that
2724 are longer than 30,000 bytes (though not repeat them that many times).
2725 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2726 made it possible to implement the integer overflow checks in a much more
2727 dynamic way, which I have now done. The artificial limitation on group
2728 length has been removed - we now have only the limit on the total length of
2729 the compiled pattern, which depends on the LINK_SIZE setting.
2730
2731 10. Fixed a bug in the documentation for get/copy named substring when
2732 duplicate names are permitted. If none of the named substrings are set, the
2733 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2734 empty string.
2735
2736 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2737 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2738 because the ] is interpreted as the first data character and the
2739 terminating ] is not found. PCRE has been made compatible with Perl in this
2740 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2741 cause memory overwriting.
2742
2743 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2744 string has been matched (to stop an infinite loop). It was not recognizing
2745 a conditional subpattern that could match an empty string if that
2746 subpattern was within another subpattern. For example, it looped when
2747 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2748 condition was not nested. This bug has been fixed.
2749
2750 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2751 past the start of the subject in the presence of bytes with the top bit
2752 set, for example "\x8aBCD".
2753
2754 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2755 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2756
2757 14. Optimized (?!) to (*FAIL).
2758
2759 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2760 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2761 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2762 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2763 does: it's just the validity check that is more restrictive.
2764
2765 16. Inserted checks for integer overflows during escape sequence (backslash)
2766 processing, and also fixed erroneous offset values for syntax errors during
2767 backslash processing.
2768
2769 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2770 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2771
2772 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2773 caused an overrun.
2774
2775 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2776 something other than just ASCII characters) inside a group that had an
2777 unlimited repeat caused a loop at compile time (while checking to see
2778 whether the group could match an empty string).
2779
2780 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2781 [\P{Any}] did so. (Error in the code for printing property names.)
2782
2783 21. An orphan \E inside a character class could cause a crash.
2784
2785 22. A repeated capturing bracket such as (A)? could cause a wild memory
2786 reference during compilation.
2787
2788 23. There are several functions in pcre_compile() that scan along a compiled
2789 expression for various reasons (e.g. to see if it's fixed length for look
2790 behind). There were bugs in these functions when a repeated \p or \P was
2791 present in the pattern. These operators have additional parameters compared
2792 with \d, etc, and these were not being taken into account when moving along
2793 the compiled data. Specifically:
2794
2795 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2796 length.
2797
2798 (b) An item such as \pL+ within a repeated group could cause crashes or
2799 loops.
2800
2801 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2802 "reference to non-existent subpattern" error.
2803
2804 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2805
2806 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2807 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2808
2809 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2810
2811 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2812 character were causing crashes (broken optimization).
2813
2814 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2815 \p or \P) caused a compile-time loop.
2816
2817 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2818 break. For example, the unstudied pattern [\r\n]A does not match the string
2819 "\r\nA" because change 7.0/46 below moves the current point on by two
2820 characters after failing to match at the start. However, the pattern \nA
2821 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2822 the same is true. There doesn't seem any very clean way out of this, but
2823 what I have chosen to do makes the common cases work: PCRE now takes note
2824 of whether there can be an explicit match for \r or \n anywhere in the
2825 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2826 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2827 pattern has explicit CR or LF references.
2828
2829 29. Added (*CR) etc for changing newline setting at start of pattern.
2830
2831
2832 Version 7.2 19-Jun-07
2833 ---------------------
2834
2835 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2836 which is apparently normally available under Windows.
2837
2838 2. Re-jig the pcregrep tests with different newline settings in an attempt
2839 to make them independent of the local environment's newline setting.
2840
2841 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2842
2843 4. Some of the "internals" tests were previously cut out when the link size
2844 was not 2, because the output contained actual offsets. The recent new
2845 "Z" feature of pcretest means that these can be cut out, making the tests
2846 usable with all link sizes.
2847
2848 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2849 stack recursion. This gives a massive performance boost under BSD, but just
2850 a small improvement under Linux. However, it saves one field in the frame
2851 in all cases.
2852
2853 6. Added more features from the forthcoming Perl 5.10:
2854
2855 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2856 recursion call. It refers to the nth most recently opened parentheses.
2857
2858 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2859 to be opened parentheses.
2860
2861 (c) Conditions that refer to capturing parentheses can be specified
2862 relatively, for example, (?(-2)... or (?(+3)...
2863
2864 (d) \K resets the start of the current match so that everything before
2865 is not part of it.
2866
2867 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2868
2869 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2870 reference syntax.
2871
2872 (g) (?| introduces a group in which the numbering of parentheses in each
2873 alternative starts with the same number.
2874
2875 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2876
2877 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2878 PCRE_INFO_JCHANGED.
2879
2880 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2881 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2882 for detecting groups that can match an empty string.
2883
2884 9. A pattern with a very large number of alternatives (more than several
2885 hundred) was running out of internal workspace during the pre-compile
2886 phase, where pcre_compile() figures out how much memory will be needed. A
2887 bit of new cunning has reduced the workspace needed for groups with
2888 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2889 workspace instead of running out of the 4096 that are available.
2890
2891 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2892
2893 11. Applied patch from Google to remove an optimization that didn't quite work.
2894 The report of the bug said:
2895
2896 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2897 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2898 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2899
2900 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2901 it matched the wrong number of bytes.
2902
2903
2904 Version 7.1 24-Apr-07
2905 ---------------------
2906
2907 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2908 that is more "standard", making use of automake and other Autotools. There
2909 is some re-arrangement of the files and adjustment of comments consequent
2910 on this.
2911
2912 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2913 for recursive directory scanning broke on some systems because the files
2914 are not scanned in any specific order and on different systems the order
2915 was different. A call to "sort" has been inserted into RunGrepTest for the
2916 approprate test as a short-term fix. In the longer term there may be an
2917 alternative.
2918
2919 3. I had an email from Eric Raymond about problems translating some of PCRE's
2920 man pages to HTML (despite the fact that I distribute HTML pages, some
2921 people do their own conversions for various reasons). The problems
2922 concerned the use of low-level troff macros .br and .in. I have therefore
2923 removed all such uses from the man pages (some were redundant, some could
2924 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2925 HTML has been updated to handle .nf/.fi and to complain if it encounters
2926 .br or .in.
2927
2928 4. Updated comments in configure.ac that get placed in config.h.in and also
2929 arranged for config.h to be included in the distribution, with the name
2930 config.h.generic, for the benefit of those who have to compile without
2931 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2932
2933 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2934 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2935 makevp.bat for the current PCRE, using the additional files
2936 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2937
2938 6. A Windows user reported a minor discrepancy with test 2, which turned out
2939 to be caused by a trailing space on an input line that had got lost in his
2940 copy. The trailing space was an accident, so I've just removed it.
2941
2942 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2943 that is needed.
2944
2945 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2946 as "const" (a) because they are and (b) because it helps the PHP
2947 maintainers who have recently made a script to detect big data structures
2948 in the php code that should be moved to the .rodata section. I remembered
2949 to update Builducptable as well, so it won't revert if ucptable.h is ever
2950 re-created.
2951
2952 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2953 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2954 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2955 support is not required. This saves 1.5-2K of code, which is important in
2956 some applications.
2957
2958 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2959 so as not to refer to the tables, even though these functions will never be
2960 called when UTF-8 support is disabled. Otherwise there are problems with a
2961 shared library.
2962
2963 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2964
2965 (a) It was defining its arguments as char * instead of void *.
2966
2967 (b) It was assuming that all moves were upwards in memory; this was true
2968 a long time ago when I wrote it, but is no longer the case.
2969
2970 The emulated memove() is provided for those environments that have neither
2971 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2972 is clearly not the case, as these two bugs were recently reported.
2973
2974 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2975 and Detrail to create the HTML documentation, the .txt form of the man
2976 pages, and it removes trailing spaces from listed files. It also creates
2977 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2978 case, it wraps all the #defines with #ifndefs. This script should be run
2979 before "make dist".
2980
2981 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2982 with Unicode property support.
2983
2984 (a) For a maximizing quantifier, if the two different cases of the
2985 character were of different lengths in their UTF-8 codings (there are
2986 some cases like this - I found 11), and the matching function had to
2987 back up over a mixture of the two cases, it incorrectly assumed they
2988 were both the same length.
2989
2990 (b) When PCRE was configured to use the heap rather than the stack for
2991 recursion during matching, it was not correctly preserving the data for
2992 the other case of a UTF-8 character when checking ahead for a match
2993 while processing a minimizing repeat. If the check also involved
2994 matching a wide character, but failed, corruption could cause an
2995 erroneous result when trying to check for a repeat of the original
2996 character.
2997
2998 13. Some tidying changes to the testing mechanism:
2999
3000 (a) The RunTest script now detects the internal link size and whether there
3001 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3002 values substituted by "configure". (The RunGrepTest script already did
3003 this for UTF-8.) The configure.ac script no longer substitutes the
3004 relevant variables.
3005
3006 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3007 with length and offset values. This means that the output is different
3008 for different internal link sizes. Test 2 is skipped for link sizes
3009 other than 2 because of this, bypassing the problem. Unfortunately,
3010 there was also a test in test 3 (the locale tests) that used /B and
3011 failed for link sizes other than 2. Rather than cut the whole test out,
3012 I have added a new /Z option to pcretest that replaces the length and
3013 offset values with spaces. This is now used to make test 3 independent
3014 of link size. (Test 2 will be tidied up later.)
3015
3016 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3017 segmentation fault instead of returning the appropriate error message.
3018
3019 15. In multiline mode when the newline sequence was set to "any", the pattern
3020 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3021 This doesn't seem right; it now treats the CRLF combination as the line
3022 ending, and so does not match in that case. It's only a pattern such as ^$
3023 that would hit this one: something like ^ABC$ would have failed after \r
3024 and then tried again after \r\n.
3025
3026 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3027 in an attempt to make files that differ only in their line terminators
3028 compare equal. This works on Linux.
3029
3030 17. Under certain error circumstances pcregrep might try to free random memory
3031 as it exited. This is now fixed, thanks to valgrind.
3032
3033 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3034 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3035 was because its rules for how to advance for /g after matching an empty
3036 string at the end of a line did not allow for this case. They now check for
3037 it specially.
3038
3039 20. pcretest is supposed to handle patterns and data of any length, by
3040 extending its buffers when necessary. It was getting this wrong when the
3041 buffer for a data line had to be extended.
3042
3043 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3044 CRLF as a newline sequence.
3045
3046 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3047 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3048 I have nevertheless tidied it up.
3049
3050 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3051
3052 24. Added a man page for pcre-config.
3053
3054
3055 Version 7.0 19-Dec-06
3056 ---------------------
3057
3058 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3059 moving to gcc 4.1.1.
3060
3061 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3062 sys/time.h, which is documented as needed for this function. It doesn't
3063 seem to matter on Linux, but it showed up on some releases of OS X.
3064
3065 3. It seems that there are systems where bytes whose values are greater than
3066 127 match isprint() in the "C" locale. The "C" locale should be the
3067 default when a C program starts up. In most systems, only ASCII printing
3068 characters match isprint(). This difference caused the output from pcretest
3069 to vary, making some of the tests fail. I have changed pcretest so that:
3070
3071 (a) When it is outputting text in the compiled version of a pattern, bytes
3072 other than 32-126 are always shown as hex escapes.
3073
3074 (b) When it is outputting text that is a matched part of a subject string,
3075 it does the same, unless a different locale has been set for the match
3076 (using the /L modifier). In this case, it uses isprint() to decide.
3077
3078 4. Fixed a major bug that caused incorrect computation of the amount of memory
3079 required for a compiled pattern when options that changed within the
3080 pattern affected the logic of the preliminary scan that determines the
3081 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3082 that the computed length was too small. The symptoms of this bug were
3083 either the PCRE error "internal error: code overflow" from pcre_compile(),
3084 or a glibc crash with a message such as "pcretest: free(): invalid next
3085 size (fast)". Examples of patterns that provoked this bug (shown in
3086 pcretest format) are:
3087
3088 /(?-x: )/x
3089 /(?x)(?-x: \s*#\s*)/
3090 /((?i)[\x{c0}])/8
3091 /(?i:[\x{c0}])/8
3092
3093 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3094 is now done differently.
3095
3096 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3097 wrapper classes; (b) implement a new function in the C++ scanner that is
3098 more efficient than the old way of doing things because it avoids levels of
3099 recursion in the regex matching; (c) add a paragraph to the documentation
3100 for the FullMatch() function.
3101
3102 6. The escape sequence \n was being treated as whatever was defined as
3103 "newline". Not only was this contrary to the documentation, which states
3104 that \n is character 10 (hex 0A), but it also went horribly wrong when
3105 "newline" was defined as CRLF. This has been fixed.
3106
3107 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3108 was being set to -1 for the "end of line" case (supposedly a value that no
3109 character can have). Though this value is never used (the check for end of
3110 line is "zero bytes in current character"), it caused compiler complaints.
3111 I've changed it to 0xffffffff.
3112
3113 8. In pcre_version.c, the version string was being built by a sequence of
3114 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3115 string (as it is for production releases) called a macro with an empty
3116 argument. The C standard says the result of this is undefined. The gcc
3117 compiler treats it as an empty string (which was what was wanted) but it is
3118 reported that Visual C gives an error. The source has been hacked around to
3119 avoid this problem.
3120
3121 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3122 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3123 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3124 of them did).
3125
3126 10. Originally, pcretest opened its input and output without "b"; then I was
3127 told that "b" was needed in some environments, so it was added for release
3128 5.0 to both the input and output. (It makes no difference on Unix-like
3129 systems.) Later I was told that it is wrong for the input on Windows. I've
3130 now abstracted the modes into two macros, to make it easier to fiddle with
3131 them, and removed "b" from the input mode under Windows.
3132
3133 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3134
3135 12. Added -help and --help to pcretest as an official way of being reminded
3136 of the options.
3137
3138 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3139 and pcrecpp.cc because they annoy compilers at high warning levels.
3140
3141 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3142
3143 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3144 scripts are not C programs :-) and which was not noticed because it works
3145 on Linux.
3146
3147 16. pcretest is supposed to handle any length of pattern and data line (as one
3148 line or as a continued sequence of lines) by extending its input buffer if
3149 necessary. This feature was broken for very long pattern lines, leading to
3150 a string of junk being passed to pcre_compile() if the pattern was longer
3151 than about 50K.
3152
3153 17. I have done a major re-factoring of the way pcre_compile() computes the
3154 amount of memory needed for a compiled pattern. Previously, there was code
3155 that made a preliminary scan of the pattern in order to do this. That was
3156 OK when PCRE was new, but as the facilities have expanded, it has become
3157 harder and harder to keep it in step with the real compile phase, and there
3158 have been a number of bugs (see for example, 4 above). I have now found a
3159 cunning way of running the real compile function in a "fake" mode that
3160 enables it to compute how much memory it would need, while actually only
3161 ever using a few hundred bytes of working memory and without too many
3162 tests of the mode. This should make future maintenance and development
3163 easier. A side effect of this work is that the limit of 200 on the nesting
3164 depth of parentheses has been removed (though this was never a serious
3165 limitation, I suspect). However, there is a downside: pcre_compile() now
3166 runs more slowly than before (30% or more, depending on the pattern). I
3167 hope this isn't a big issue. There is no effect on runtime performance.
3168
3169 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3170 newline (only possible for the last line of a file) and it was a
3171 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3172
3173 19. Added additional timing features to pcretest. (1) The -tm option now times
3174 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3175 separate command line item, by a number that specifies the number of
3176 repeats to use when timing. The default is 50000; this gives better
3177 precision, but takes uncomfortably long for very large patterns.
3178
3179 20. Extended pcre_study() to be more clever in cases where a branch of a
3180 subpattern has no definite first character. For example, (a*|b*)[cd] would
3181 previously give no result from pcre_study(). Now it recognizes that the
3182 first character must be a, b, c, or d.
3183
3184 21. There was an incorrect error "recursive call could loop indefinitely" if
3185 a subpattern (or the entire pattern) that was being tested for matching an
3186 empty string contained only one non-empty item after a nested subpattern.
3187 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3188 incorrectly, because the \d was being skipped in the check.
3189
3190 22. The pcretest program now has a new pattern option /B and a command line
3191 option -b, which is equivalent to adding /B to every pattern. This causes
3192 it to show the compiled bytecode, without the additional information that
3193 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3194 is the same as /B/I).
3195
3196 23. A new optimization is now able automatically to treat some sequences such
3197 as a*b as a*+b. More specifically, if something simple (such as a character
3198 or a simple class like \d) has an unlimited quantifier, and is followed by
3199 something that cannot possibly match the quantified thing, the quantifier
3200 is automatically "possessified".
3201
3202 24. A recursive reference to a subpattern whose number was greater than 39
3203 went wrong under certain circumstances in UTF-8 mode. This bug could also
3204 have affected the operation of pcre_study().
3205
3206 25. Realized that a little bit of performance could be had by replacing
3207 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3208
3209 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3210
3211 27. Possessive quantifiers such as a++ were previously implemented by turning
3212 them into atomic groups such as ($>a+). Now they have their own opcodes,
3213 which improves performance. This includes the automatically created ones
3214 from 23 above.
3215
3216 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3217 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3218 the first matched character to be a colon. This applied both to named and
3219 numbered groups.
3220
3221 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3222
3223 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3224 building PCRE on Windows easier, so I have included it in the distribution.
3225
3226 31. There is now a check in pcretest against a ridiculously large number being
3227 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3228 loop, the loop is abandoned.
3229
3230 32. Forward references to subpatterns in conditions such as (?(2)...) where
3231 subpattern 2 is defined later cause pcre_compile() to search forwards in
3232 the pattern for the relevant set of parentheses. This search went wrong
3233 when there were unescaped parentheses in a character class, parentheses
3234 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3235
3236 33. "Subroutine" calls and backreferences were previously restricted to
3237 referencing subpatterns earlier in the regex. This restriction has now
3238 been removed.
3239
3240 34. Added a number of extra features that are going to be in Perl 5.10. On the
3241 whole, these are just syntactic alternatives for features that PCRE had
3242 previously implemented using the Python syntax or my own invention. The
3243 other formats are all retained for compatibility.
3244
3245 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3246 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3247 also .NET compatible.
3248
3249 (b) A recursion or subroutine call to a named group can now be defined as
3250 (?&name) as well as (?P>name).
3251
3252 (c) A backreference to a named group can now be defined as \k<name> or
3253 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3254 5.10, are also .NET compatible.
3255
3256 (d) A conditional reference to a named group can now use the syntax
3257 (?(<name>) or (?('name') as well as (?(name).
3258
3259 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3260 groups (named and numbered) that are never evaluated inline, but can be
3261 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3262 is always false. There may be only one alternative in such a group.
3263
3264 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3265 as the simple (?(R). The condition is true only if the most recent
3266 recursion is that of the given number or name. It does not search out
3267 through the entire recursion stack.
3268
3269 (g) The escape \gN or \g{N} has been added, where N is a positive or
3270 negative number, specifying an absolute or relative reference.
3271
3272 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3273 some "unreachable code" warnings.
3274
3275 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3276 things, this adds five new scripts.
3277
3278 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3279 There were also incompatibilities regarding the handling of \Q..\E inside
3280 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3281 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3282
3283 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3284 matches an empty string, and forcibly breaks the loop. There were bugs in
3285 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3286 against aaaa the result was just "a" rather than "aaaa", for example. Two
3287 separate and independent bugs (that affected different cases) have been
3288 fixed.
3289
3290 39. Refactored the code to abolish the use of different opcodes for small
3291 capturing bracket numbers. This is a tidy that I avoided doing when I
3292 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3293 The new approach is not only tidier, it makes it possible to reduce the
3294 memory needed to fix the previous bug (38).
3295
3296 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3297 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3298 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3299 mode.
3300
3301 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3302 report.
3303
3304 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3305 copy construction and assignment in the C++ wrapper.
3306
3307 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3308 couple of bugs that could have given wrong results in the "--newline=crlf"
3309 case.
3310
3311 44. Added a number of casts and did some reorganization of signed/unsigned int
3312 variables following suggestions from Dair Grant. Also renamed the variable
3313 "this" as "item" because it is a C++ keyword.
3314
3315 45. Arranged for dftables to add
3316
3317 #include "pcre_internal.h"
3318
3319 to pcre_chartables.c because without it, gcc 4.x may remove the array
3320 definition from the final binary if PCRE is built into a static library and
3321 dead code stripping is activated.
3322
3323 46. For an unanchored pattern, if a match attempt fails at the start of a
3324 newline sequence, and the newline setting is CRLF or ANY, and the next two
3325 characters are CRLF, advance by two characters instead of one.
3326
3327
3328 Version 6.7 04-Jul-06
3329 ---------------------
3330
3331 1. In order to handle tests when input lines are enormously long, pcretest has
3332 been re-factored so that it automatically extends its buffers when
3333 necessary. The code is crude, but this _is_ just a test program. The
3334 default size has been increased from 32K to 50K.
3335
3336 2. The code in pcre_study() was using the value of the re argument before
3337 testing it for NULL. (Of course, in any sensible call of the function, it
3338 won't be NULL.)
3339
3340 3. The memmove() emulation function in pcre_internal.h, which is used on
3341 systems that lack both memmove() and bcopy() - that is, hardly ever -
3342 was missing a "static" storage class specifier.
3343
3344 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3345 containing an extended class (one that cannot be represented by a bitmap
3346 because it contains high-valued characters or Unicode property items, e.g.
3347 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3348 pattern, but PCRE should not loop if you do not (it no longer does).
3349 [Detail: two cases were found: (a) a repeated subpattern containing an
3350 extended class; (b) a recursive reference to a subpattern that followed a
3351 previous extended class. It wasn't skipping over the extended class
3352 correctly when UTF-8 mode was not set.]
3353
3354 5. A negated single-character class was not being recognized as fixed-length
3355 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3356 compile error "lookbehind assertion is not fixed length".
3357
3358 6. The RunPerlTest auxiliary script was showing an unexpected difference
3359 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3360 write a Perl script that can interpret lines of an input file either as
3361 byte characters or as UTF-8, which is what "perltest" was being required to
3362 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3363 can't do is switch easily at run time between having the "use utf8;" pragma
3364 or not. In the end, I fudged it by using the RunPerlTest script to insert
3365 "use utf8;" explicitly for the UTF-8 tests.
3366
3367 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3368 the end of the subject string, contrary to the documentation and to what
3369 Perl does. This was true of both matching functions. Now it matches only at
3370 the start of the subject and immediately after *internal* newlines.
3371
3372 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3373 a pointer to an int instead of a pointer to an unsigned long int. This
3374 caused problems on 64-bit systems.
3375
3376 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3377 instance of the 'standard' template library not being so standard".
3378
3379 10. There was no check on the number of named subpatterns nor the maximum
3380 length of a subpattern name. The product of these values is used to compute
3381 the size of the memory block for a compiled pattern. By supplying a very
3382 long subpattern name and a large number of named subpatterns, the size
3383 computation could be caused to overflow. This is now prevented by limiting
3384 the length of names to 32 characters, and the number of named subpatterns
3385 to 10,000.
3386
3387 11. Subpatterns that are repeated with specific counts have to be replicated in
3388 the compiled pattern. The size of memory for this was computed from the
3389 length of the subpattern and the repeat count. The latter is limited to
3390 65535, but there was no limit on the former, meaning that integer overflow
3391 could in principle occur. The compiled length of a repeated subpattern is
3392 now limited to 30,000 bytes in order to prevent this.
3393
3394 12. Added the optional facility to have named substrings with the same name.
3395
3396 13. Added the ability to use a named substring as a condition, using the
3397 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3398 are numbers (not recommended). Forward references are permitted.
3399
3400 14. Added forward references in named backreferences (if you see what I mean).
3401
3402 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3403 pattern could run off the end of the subject. For example, the pattern
3404 "(?s)(.{1,5})"8 did this with the subject "ab".
3405
3406 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3407 PCRE_CASELESS was set when matching characters that were quantified with ?
3408 or *.
3409
3410 17. A character class other than a single negated character that had a minimum
3411 but no maximum quantifier - for example [ab]{6,} - was not handled
3412 correctly by pce_dfa_exec(). It would match only one character.
3413
3414 18. A valid (though odd) pattern that looked like a POSIX character
3415 class but used an invalid character after [ (for example [[,abc,]]) caused
3416 pcre_compile() to give the error "Failed: internal error: code overflow" or
3417 in some cases to crash with a glibc free() error. This could even happen if
3418 the pattern terminated after [[ but there just happened to be a sequence of
3419 letters, a binary zero, and a closing ] in the memory that followed.
3420
3421 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3422 over the years. Originally (before any Unicode support), just the bottom 8
3423 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3424 output from "man perlunicode" includes this:
3425
3426 The regular expression compiler produces polymorphic opcodes. That
3427 is, the pattern adapts to the data and automatically switches to
3428 the Unicode character scheme when presented with Unicode data--or
3429 instead uses a traditional byte scheme when presented with byte
3430 data.
3431
3432 Sadly, a wide octal escape does not cause a switch, and in a string with
3433 no other multibyte characters, these octal escapes are treated as before.
3434 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3435 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3436 Unicode string.
3437
3438 I have not perpetrated such confusion in PCRE. Up till now, it took just
3439 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3440 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3441 translate to the appropriate multibyte character.
3442
3443 29. Applied some refactoring to reduce the number of warnings from Microsoft
3444 and Borland compilers. This has included removing the fudge introduced
3445 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3446 a warning about an unused variable.
3447
3448 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3449 characters since release 4.0, because Perl (from release 5.004) does not.
3450 [Or at least, is documented not to: some releases seem to be in conflict
3451 with the documentation.] However, when a pattern was studied with
3452 pcre_study() and all its branches started with \s, PCRE still included VT
3453 as a possible starting character. Of course, this did no harm; it just
3454 caused an unnecessary match attempt.
3455
3456 22. Removed a now-redundant internal flag bit that recorded the fact that case
3457 dependency changed within the pattern. This was once needed for "required
3458 byte" processing, but is no longer used. This recovers a now-scarce options
3459 bit. Also moved the least significant internal flag bit to the most-
3460 significant bit of the word, which was not previously used (hangover from
3461 the days when it was an int rather than a uint) to free up another bit for
3462 the future.
3463
3464 23. Added support for CRLF line endings as well as CR and LF. As well as the
3465 default being selectable at build time, it can now be changed at runtime
3466 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3467 specify that it is scanning data with non-default line endings.
3468
3469 24. Changed the definition of CXXLINK to make it agree with the definition of
3470 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3471
3472 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3473 recursions. This makes a big different to stack usage for some patterns.
3474
3475 26. If a subpattern containing a named recursion or subroutine reference such
3476 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3477 the space required for the compiled pattern went wrong and gave too small a
3478 value. Depending on the environment, this could lead to "Failed: internal
3479 error: code overflow at offset 49" or "glibc detected double free or
3480 corruption" errors.
3481
3482 27. Applied patches from Google (a) to support the new newline modes and (b) to
3483 advance over multibyte UTF-8 characters in GlobalReplace.
3484
3485 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3486 difference for some implementation of PCRE in some Windows version.
3487
3488 29. Added some extra testing facilities to pcretest:
3489
3490 \q<number> in a data line sets the "match limit" value
3491 \Q<number> in a data line sets the "match recursion limt" value
3492 -S <number> sets the stack size, where <number> is in megabytes
3493
3494 The -S option isn't available for Windows.
3495
3496
3497 Version 6.6 06-Feb-06
3498 ---------------------
3499
3500 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3501 in pcreposix.h. I have copied the definition from pcre.h.
3502
3503 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3504 because pcre.h is no longer a built file.
3505
3506 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3507 not normally included in the compiled code.
3508
3509
3510 Version 6.5 01-Feb-06
3511 ---------------------
3512
3513 1. When using the partial match feature with pcre_dfa_exec(), it was not
3514 anchoring the second and subsequent partial matches at the new starting
3515 point. This could lead to incorrect results. For example, with the pattern
3516 /1234/, partially matching against "123" and then "a4" gave a match.
3517
3518 2. Changes to pcregrep:
3519
3520 (a) All non-match returns from pcre_exec() were being treated as failures
3521 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3522 error message is output. Some extra information is given for the
3523 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3524 probably the only errors that are likely to be caused by users (by
3525 specifying a regex that has nested indefinite repeats, for instance).
3526 If there are more than 20 of these errors, pcregrep is abandoned.
3527
3528 (b) A binary zero was treated as data while matching, but terminated the
3529 output line if it was written out. This has been fixed: binary zeroes
3530 are now no different to any other data bytes.
3531
3532 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3533 used to set a locale for matching. The --locale=xxxx long option has
3534 been added (no short equivalent) to specify a locale explicitly on the
3535 pcregrep command, overriding the environment variables.
3536
3537 (d) When -B was used with -n, some line numbers in the output were one less
3538 than they should have been.
3539
3540 (e) Added the -o (--only-matching) option.
3541
3542 (f) If -A or -C was used with -c (count only), some lines of context were
3543 accidentally printed for the final match.
3544
3545 (g) Added the -H (--with-filename) option.
3546
3547 (h) The combination of options -rh failed to suppress file names for files
3548 that were found from directory arguments.
3549
3550 (i) Added the -D (--devices) and -d (--directories) options.
3551
3552 (j) Added the -F (--fixed-strings) option.
3553
3554 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3555
3556 (l) Added the --colo(u)r option.
3557
3558 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3559 is not present by default.
3560
3561 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3562 items such as (?R) or (?1), when the recursion could match a number of
3563 alternatives. If it matched one of the alternatives, but subsequently,
3564 outside the recursion, there was a failure, the code tried to back up into
3565 the recursion. However, because of the way PCRE is implemented, this is not
3566 possible, and the result was an incorrect result from the match.
3567
3568 In order to prevent this happening, the specification of recursion has
3569 been changed so that all such subpatterns are automatically treated as
3570 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3571
3572 4. I had overlooked the fact that, in some locales, there are characters for
3573 which isalpha() is true but neither isupper() nor islower() are true. In
3574 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3575 and ordfeminine) are like this. This affected the treatment of \w and \W
3576 when they appeared in character classes, but not when they appeared outside
3577 a character class. The bit map for "word" characters is now created
3578 separately from the results of isalnum() instead of just taking it from the
3579 upper, lower, and digit maps. (Plus the underscore character, of course.)
3580
3581 5. The above bug also affected the handling of POSIX character classes such as
3582 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3583 permanent tables. Instead, the bit maps for such a class were previously
3584 created as the appropriate unions of the upper, lower, and digit bitmaps.
3585 Now they are created by subtraction from the [[:word:]] class, which has
3586 its own bitmap.
3587
3588 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3589 It is created by subtracting the vertical space characters (\x09, \x0a,
3590 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3591 subtraction was done in the overall bitmap for a character class, meaning
3592 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3593 be recognized. This bug has been fixed.
3594
3595 7. Patches from the folks at Google:
3596
3597 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3598 real life, but is still worth protecting against".
3599
3600 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3601 regular expressions".
3602
3603 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3604 have it.
3605
3606 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3607 "configure" and the latter not, in order to fix a problem somebody had
3608 with compiling the Arg class on HP-UX.
3609
3610 (e) Improve the error-handling of the C++ wrapper a little bit.
3611
3612 (f) New tests for checking recursion limiting.
3613
3614 8. The pcre_memmove() function, which is used only if the environment does not
3615 have a standard memmove() function (and is therefore rarely compiled),
3616 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3617 returning a result (though PCRE never actually uses the result).
3618
3619 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3620 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3621 returned instead of calling malloc() with an overflowing number that would
3622 most likely cause subsequent chaos.
3623
3624 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3625
3626 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3627 with this option is matched, the nmatch and pmatch options of regexec() are
3628 ignored.
3629
3630 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3631 provided in case anyone wants to the the POSIX interface with UTF-8
3632 strings.
3633
3634 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3635 C++ linking (needed for some HP-UX environments).
3636
3637 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3638 (unused parameter) and in the pcre_printint() function (omitted "default"
3639 switch label when the default is to do nothing).
3640
3641 15. Added some code to make it possible, when PCRE is compiled as a C++
3642 library, to replace subject pointers for pcre_exec() with a smart pointer
3643 class, thus making it possible to process discontinuous strings.
3644
3645 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3646 much the same function. They were added by different people who were trying
3647 to make PCRE easy to compile on non-Unix systems. It has been suggested
3648 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3649 for compiling on Windows systems. I have therefore replaced it with
3650 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3651 defaults to "extern" for C or "extern C" for C++, which works fine on
3652 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3653 SCOPE with something explicit in config.h. In addition:
3654
3655 (a) pcreposix.h still had just "extern" instead of either of these macros;
3656 I have replaced it with PCRE_DATA_SCOPE.
3657
3658 (b) Functions such as _pcre_xclass(), which are internal to the library,
3659 but external in the C sense, all had PCRE_EXPORT in their definitions.
3660 This is apparently wrong for the Windows case, so I have removed it.
3661 (It makes no difference on Unix-like systems.)
3662
3663 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3664 of recursive calls to match(). This is different to MATCH_LIMIT because
3665 that limits the total number of calls to match(), not all of which increase
3666 the depth of recursion. Limiting the recursion depth limits the amount of
3667 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3668 when PCRE is compiled, and changed at run time. A patch from Google adds
3669 this functionality to the C++ interface.
3670
3671 18. Changes to the handling of Unicode character properties:
3672
3673 (a) Updated the table to Unicode 4.1.0.
3674
3675 (b) Recognize characters that are not in the table as "Cn" (undefined).
3676
3677 (c) I revised the way the table is implemented to a much improved format
3678 which includes recognition of ranges. It now supports the ranges that
3679 are defined in UnicodeData.txt, and it also amalgamates other
3680 characters into ranges. This has reduced the number of entries in the
3681 table from around 16,000 to around 3,000, thus reducing its size
3682 considerably. I realized I did not need to use a tree structure after
3683 all - a binary chop search is just as efficient. Having reduced the
3684 number of entries, I extended their size from 6 bytes to 8 bytes to
3685 allow for more data.
3686
3687 (d) Added support for Unicode script names via properties such as \p{Han}.
3688
3689 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3690 matching that character.
3691
3692 20. When matching a repeated Unicode property with a minimum greater than zero,
3693 (for example \pL{2,}), PCRE could look past the end of the subject if it
3694 reached it while seeking the minimum number of characters. This could
3695 happen only if some of the characters were more than one byte long, because
3696 there is a check for at least the minimum number of bytes.
3697
3698 21. Refactored the implementation of \p and \P so as to be more general, to
3699 allow for more different types of property in future. This has changed the
3700 compiled form incompatibly. Anybody with saved compiled patterns that use
3701 \p or \P will have to recompile them.
3702
3703 22. Added "Any" and "L&" to the supported property types.
3704
3705 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3706 but give a compile time error if the value is greater than 0xff.
3707
3708 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3709 accidentally not being installed or uninstalled.
3710
3711 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3712 made were to insert the current release number. This seemed silly, because
3713 it made things harder for people building PCRE on systems that don't run
3714 "configure". I have turned pcre.h into a distributed file, no longer built
3715 by "configure", with the version identification directly included. There is
3716 no longer a pcre.h.in file.
3717
3718 However, this change necessitated a change to the pcre-config script as
3719 well. It is built from pcre-config.in, and one of the substitutions was the
3720 release number. I have updated configure.ac so that ./configure now finds
3721 the release number by grepping pcre.h.
3722
3723 26. Added the ability to run the tests under valgrind.
3724
3725
3726 Version 6.4 05-Sep-05
3727 ---------------------
3728
3729 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3730 "--" to be printed when multiple files were scanned, even when none of the
3731 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3732 consider it to be a bug, and have restored the previous behaviour.
3733
3734 2. A couple of code tidies to get rid of compiler warnings.
3735
3736 3. The pcretest program used to cheat by referring to symbols in the library
3737 whose names begin with _pcre_. These are internal symbols that are not
3738 really supposed to be visible externally, and in some environments it is
3739 possible to suppress them. The cheating is now confined to including
3740 certain files from the library's source, which is a bit cleaner.
3741
3742 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3743 file's purpose clearer.
3744
3745 5. Reorganized pcre_ucp_findchar().
3746
3747
3748 Version 6.3 15-Aug-05
3749 ---------------------
3750
3751 1. The file libpcre.pc.in did not have general read permission in the tarball.
3752
3753 2. There were some problems when building without C++ support:
3754
3755 (a) If C++ support was not built, "make install" and "make test" still
3756 tried to test it.
3757
3758 (b) There were problems when the value of CXX was explicitly set. Some
3759 changes have been made to try to fix these, and ...
3760
3761 (c) --disable-cpp can now be used to explicitly disable C++ support.
3762
3763 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3764 backslash in a target when C++ was disabled. This confuses some
3765 versions of "make", apparently. Using an intermediate variable solves
3766 this. (Same for CPP_LOBJ.)
3767
3768 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3769 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3770 necessary on certain architectures.
3771
3772 4. Added a setting of -export-symbols-regex to the link command to remove
3773 those symbols that are exported in the C sense, but actually are local
3774 within the library, and not documented. Their names all begin with
3775 "_pcre_". This is not a perfect job, because (a) we have to except some
3776 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3777 available (and never for static libraries). I have made a note to try to
3778 find a way round (a) in the future.
3779
3780
3781 Version 6.2 01-Aug-05
3782 ---------------------
3783
3784 1. There was no test for integer overflow of quantifier values. A construction
3785 such as {1111111111111111} would give undefined results. What is worse, if
3786 a minimum quantifier for a parenthesized subpattern overflowed and became
3787 negative, the calculation of the memory size went wrong. This could have
3788 led to memory overwriting.
3789
3790 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3791
3792 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3793 operating environments where this matters.
3794
3795 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3796 PCRE options from within the C++ wrapper.
3797
3798 5. Named capturing subpatterns were not being correctly counted when a pattern
3799 was compiled. This caused two problems: (a) If there were more than 100
3800 such subpatterns, the calculation of the memory needed for the whole
3801 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3802 back references of the form \12, where the number was greater than 9, were
3803 not recognized as back references, even though there were sufficient
3804 previous subpatterns.
3805
3806 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3807 versions of gcc, e.g. 2.95.4.
3808
3809
3810 Version 6.1 21-Jun-05
3811 ---------------------
3812
3813 1. There was one reference to the variable "posix" in pcretest.c that was not
3814 surrounded by "#if !defined NOPOSIX".
3815
3816 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3817 the cross-check on the old pcre_info() function, for the benefit of the
3818 cut-down version of PCRE that is currently imported into Exim.
3819
3820 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3821 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3822 patterns that start (?i) but I don't think that matters. The use of (?i) is
3823 just an example; this all applies to the other options as well.
3824
3825 4. Since libtool seems to echo the compile commands it is issuing, the output
3826 from "make" can be reduced a bit by putting "@" in front of each libtool
3827 compile command.
3828
3829 5. Patch from the folks at Google for configure.in to be a bit more thorough
3830 in checking for a suitable C++ installation before trying to compile the
3831 C++ stuff. This should fix a reported problem when a compiler was present,
3832 but no suitable headers.
3833
3834 6. The man pages all had just "PCRE" as their title. I have changed them to
3835 be the relevant file name. I have also arranged that these names are
3836 retained in the file doc/pcre.txt, which is a concatenation in text format
3837 of all the man pages except the little individual ones for each function.
3838
3839 7. The NON-UNIX-USE file had not been updated for the different set of source
3840 files that come with release 6. I also added a few comments about the C++
3841 wrapper.
3842
3843
3844 Version 6.0 07-Jun-05
3845 ---------------------
3846
3847 1. Some minor internal re-organization to help with my DFA experiments.
3848
3849 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3850 didn't matter for the library itself when fully configured, but did matter
3851 when compiling without UCP support, or within Exim, where the ucp files are
3852 not imported.
3853
3854 3. Refactoring of the library code to split up the various functions into
3855 different source modules. The addition of the new DFA matching code (see
3856 below) to a single monolithic source would have made it really too
3857 unwieldy, quite apart from causing all the code to be include in a
3858 statically linked application, when only some functions are used. This is
3859 relevant even without the DFA addition now that patterns can be compiled in
3860 one application and matched in another.
3861
3862 The downside of splitting up is that there have to be some external
3863 functions and data tables that are used internally in different modules of
3864 the library but which are not part of the API. These have all had their
3865 names changed to start with "_pcre_" so that they are unlikely to clash
3866 with other external names.
3867
3868 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3869 a different (DFA) algorithm. Although it is slower than the original
3870 function, it does have some advantages for certain types of matching
3871 problem.
3872
3873 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3874 including restarting after a partial match.
3875
3876 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3877 defined when compiling for Windows was sent to me. I have put it into the
3878 code, though I have no means of testing or verifying it.
3879
3880 7. Added the pcre_refcount() auxiliary function.
3881
3882 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3883 match before or at the first newline in the subject string. In pcretest,
3884 the /f option on a pattern can be used to set this.
3885
3886 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3887 would behave wrongly. This has been present in PCRE since release 4.0.
3888
3889 10. A number of changes to the pcregrep command:
3890
3891 (a) Refactored how -x works; insert ^(...)$ instead of setting
3892 PCRE_ANCHORED and checking the length, in preparation for adding
3893 something similar for -w.
3894
3895 (b) Added the -w (match as a word) option.
3896
3897 (c) Refactored the way lines are read and buffered so as to have more
3898 than one at a time available.
3899
3900 (d) Implemented a pcregrep test script.
3901
3902 (e) Added the -M (multiline match) option. This allows patterns to match
3903 over several lines of the subject. The buffering ensures that at least
3904 8K, or the rest of the document (whichever is the shorter) is available
3905 for matching (and similarly the previous 8K for lookbehind assertions).
3906
3907 (f) Changed the --help output so that it now says
3908
3909 -w, --word-regex(p)
3910
3911 instead of two lines, one with "regex" and the other with "regexp"
3912 because that confused at least one person since the short forms are the
3913 same. (This required a bit of code, as the output is generated
3914 automatically from a table. It wasn't just a text change.)
3915
3916 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3917 option but starts with a hyphen. Could be a pattern or a path name
3918 starting with a hyphen, for instance.
3919
3920 (h) "-" can be given as a file name to represent stdin.
3921
3922 (i) When file names are being printed, "(standard input)" is used for
3923 the standard input, for compatibility with GNU grep. Previously
3924 "<stdin>" was used.
3925
3926 (j) The option --label=xxx can be used to supply a name to be used for
3927 stdin when file names are being printed. There is no short form.
3928
3929 (k) Re-factored the options decoding logic because we are going to add
3930 two more options that take data. Such options can now be given in four
3931 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3932
3933 (l) Added the -A, -B, and -C options for requesting that lines of context
3934 around matches be printed.
3935
3936 (m) Added the -L option to print the names of files that do not contain
3937 any matching lines, that is, the complement of -l.
3938
3939 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3940 continue to scan other files.
3941
3942 (o) The -s option was incorrectly implemented. For compatibility with other
3943 greps, it now suppresses the error message for a non-existent or non-
3944 accessible file (but not the return code). There is a new option called
3945 -q that suppresses the output of matching lines, which was what -s was
3946 previously doing.
3947
3948 (p) Added --include and --exclude options to specify files for inclusion
3949 and exclusion when recursing.
3950
3951 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3952 Hopefully, it now does.
3953
3954 12. Missing cast in pcre_study().
3955
3956 13. Added an "uninstall" target to the makefile.
3957
3958 14. Replaced "extern" in the function prototypes in Makefile.in with
3959 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3960 world, but is set differently for Windows.
3961
3962 15. Added a second compiling function called pcre_compile2(). The only
3963 difference is that it has an extra argument, which is a pointer to an
3964 integer error code. When there is a compile-time failure, this is set
3965 non-zero, in addition to the error test pointer being set to point to an
3966 error message. The new argument may be NULL if no error number is required
3967 (but then you may as well call pcre_compile(), which is now just a
3968 wrapper). This facility is provided because some applications need a
3969 numeric error indication, but it has also enabled me to tidy up the way
3970 compile-time errors are handled in the POSIX wrapper.
3971
3972 16. Added VPATH=.libs to the makefile; this should help when building with one
3973 prefix path and installing with another. (Or so I'm told by someone who
3974 knows more about this stuff than I do.)
3975
3976 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3977 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3978 match everything, including newlines. This is not POSIX-compatible, but
3979 somebody wanted the feature. From pcretest it can be activated by using
3980 both the P and the s flags.
3981
3982 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3983
3984 19. libpcre.pc was being incorrectly installed as executable.
3985
3986 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3987 it now also looks for '\r' so that it will work unmodified on Windows.
3988
3989 21. Added Google's contributed C++ wrapper to the distribution.
3990
3991 22. Added some untidy missing memory free() calls in pcretest, to keep
3992 Electric Fence happy when testing.
3993
3994
3995
3996 Version 5.0 13-Sep-04
3997 ---------------------
3998
3999 1. Internal change: literal characters are no longer packed up into items
4000 containing multiple characters in a single byte-string. Each character
4001 is now matched using a separate opcode. However, there may be more than one
4002 byte in the character in UTF-8 mode.
4003
4004 2. The pcre_callout_block structure has two new fields: pattern_position and
4005 next_item_length. These contain the offset in the pattern to the next match
4006 item, and its length, respectively.
4007
4008 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4009 insertion of callouts before each pattern item. Added the /C option to
4010 pcretest to make use of this.
4011
4012 4. On the advice of a Windows user, the lines
4013
4014 #if defined(_WIN32) || defined(WIN32)
4015 _setmode( _fileno( stdout ), 0x8000 );
4016 #endif /* defined(_WIN32) || defined(WIN32) */
4017
4018 have been added to the source of pcretest. This apparently does useful
4019 magic in relation to line terminators.
4020
4021 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4022 for the benefit of those environments where the "b" makes a difference.
4023
4024 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4025 to know about it. I have put a hack into configure.in that adds in code
4026 to set GCC=yes if CC=icc. This seems to end up at a point in the
4027 generated configure script that is early enough to affect the setting of
4028 compiler options, which is what is needed, but I have no means of testing
4029 whether it really works. (The user who reported this had patched the
4030 generated configure script, which of course I cannot do.)
4031
4032 LATER: After change 22 below (new libtool files), the configure script
4033 seems to know about icc (and also ecc). Therefore, I have commented out
4034 this hack in configure.in.
4035
4036 7. Added support for pkg-config (2 patches were sent in).
4037
4038 8. Negated POSIX character classes that used a combination of internal tables
4039 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4040 [[:^ascii]]. Typically, they would match almost any characters. The other
4041 POSIX classes were not broken in this way.
4042
4043 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4044 to find the match, as PCRE was deluded into thinking that the match had to
4045 start at the start point or following a newline. The same bug applied to
4046 patterns with negative forward assertions or any backward assertions
4047 preceding ".*" at the start, unless the pattern required a fixed first
4048 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4049
4050 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4051 starting at the last subject character, bytes beyond the end of the subject
4052 string were read.
4053
4054 11. Renamed the variable "class" as "classbits" to make life easier for C++
4055 users. (Previously there was a macro definition, but it apparently wasn't
4056 enough.)
4057
4058 12. Added the new field "tables" to the extra data so that tables can be passed
4059 in at exec time, or the internal tables can be re-selected. This allows
4060 a compiled regex to be saved and re-used at a later time by a different
4061 program that might have everything at different addresses.
4062
4063 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4064 -R library as well as a -L library.
4065
4066 14. The debugging options of pcretest (-d on the command line or D on a
4067 pattern) showed incorrect output for anything following an extended class
4068 that contained multibyte characters and which was followed by a quantifier.
4069
4070 15. Added optional support for general category Unicode character properties
4071 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4072 support. It adds about 90K to the size of the library. The meanings of the
4073 inbuilt class escapes such as \d and \s have NOT been changed.
4074
4075 16. Updated pcredemo.c to include calls to free() to release the memory for the
4076 compiled pattern.
4077
4078 17. The generated file chartables.c was being created in the source directory
4079 instead of in the building directory. This caused the build to fail if the
4080 source directory was different from the building directory, and was
4081 read-only.
4082
4083 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4084 file. No doubt somebody will tell me if they don't make sense... Also added
4085 Dan Mooney's comments about building on OpenVMS.
4086
4087 19. Added support for partial matching via the PCRE_PARTIAL option for
4088 pcre_exec() and the \P data escape in pcretest.
4089
4090 20. Extended pcretest with 3 new pattern features:
4091
4092 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4093 write the compiled pattern to the file whose name is "rest-of-line".
4094 This is a straight binary dump of the data, with the saved pointer to
4095 the character tables forced to be NULL. The study data, if any, is
4096 written too. After writing, pcretest reads a new pattern.
4097
4098 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4099 compiled pattern from the given file. There must not be any
4100 occurrences of "<" in the file name (pretty unlikely); if there are,
4101 pcretest will instead treat the initial "<" as a pattern delimiter.
4102 After reading in the pattern, pcretest goes on to read data lines as
4103 usual.
4104
4105 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4106 and 16-bit fields in a compiled pattern, to simulate a pattern that
4107 was compiled on a host of opposite endianness.
4108
4109 21. The pcre-exec() function can now cope with patterns that were compiled on
4110 hosts of opposite endianness, with this restriction:
4111
4112 As for any compiled expression that is saved and used later, the tables
4113 pointer field cannot be preserved; the extra_data field in the arguments
4114 to pcre_exec() should be used to pass in a tables address if a value
4115 other than the default internal tables were used at compile time.
4116
4117 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4118 now diagnosed as an error. Previously, most of the time, a negative number
4119 would have been treated as zero, but if in addition "ovector" was passed as
4120 NULL, a crash could occur.
4121
4122 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4123 new versions from the libtool 1.5 distribution (the last one is a copy of
4124 a file called libtool.m4). This seems to have fixed the need to patch
4125 "configure" to support Darwin 1.3 (which I used to do). However, I still
4126 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4127 workstation).
4128
4129 24. Changed the PCRE licence to be the more standard "BSD" licence.
4130
4131
4132 Version 4.5 01-Dec-03
4133 ---------------------
4134
4135 1. There has been some re-arrangement of the code for the match() function so
4136 that it can be compiled in a version that does not call itself recursively.
4137 Instead, it keeps those local variables that need separate instances for
4138 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4139 needs to "recurse". Keeping track of where control must go is done by means
4140 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4141 hide most of the details from the main code, and operates only if
4142 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4143 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4144 operating.
4145
4146 To make it easier for callers to provide specially tailored get/free
4147 functions for this usage, two new functions, pcre_stack_malloc, and
4148 pcre_stack_free, are used. They are always called in strict stacking order,
4149 and the size of block requested is always the same.
4150
4151 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4152 PCRE has been compiled to use the stack or the heap for recursion. The
4153 -C option of pcretest uses this to show which version is compiled.
4154
4155 A new data escape \S, is added to pcretest; it causes the amounts of store
4156 obtained and freed by both kinds of malloc/free at match time to be added
4157 to the output.
4158
4159 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4160 what's available on my current Linux desktop machine.
4161
4162 3. When matching a UTF-8 string, the test for a valid string at the start has
4163 been extended. If start_offset is not zero, PCRE now checks that it points
4164 to a byte that is the start of a UTF-8 character. If not, it returns
4165 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4166 this is necessary because there may be backward assertions in the pattern.
4167 When matching the same subject several times, it may save resources to use
4168 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4169
4170 4. The code for checking the validity of UTF-8 strings has been tightened so
4171 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4172 containing "overlong sequences".
4173
4174 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4175 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4176 should have been "&", but it just so happened that all the cases this let
4177 through by mistake were picked up later in the function.
4178
4179 6. I had used a variable called "isblank" - this is a C99 function, causing
4180 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4181
4182 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4183 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4184 so the version is shown; (c) stop "make test" echoing "./RunTest".
4185
4186 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4187
4188 9. The prototype for memmove() for systems that don't have it was using
4189 size_t, but the inclusion of the header that defines size_t was later. I've
4190 moved the #includes for the C headers earlier to avoid this.
4191
4192 10. Added some adjustments to the code to make it easier to compiler on certain
4193 special systems:
4194
4195 (a) Some "const" qualifiers were missing.
4196 (b) Added the macro EXPORT before all exported functions; by default this
4197 is defined to be empty.
4198 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4199 that it reads its output file name as an argument instead of writing
4200 to the standard output and assuming this can be redirected.
4201
4202 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4203 class containing characters with values greater than 255, PCRE compilation
4204 went into a loop.
4205
4206 12. A recursive reference to a subpattern that was within another subpattern
4207 that had a minimum quantifier of zero caused PCRE to crash. For example,
4208 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4209 recursion. If the recursively-called subpattern itself had a zero repeat,
4210 that was OK.
4211
4212 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4213 buffer into which it was copied (for escape processing) was still set at
4214 1024, so long lines caused crashes.
4215
4216 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4217 "internal error: code overflow...". This applied to any character class
4218 that was followed by a possessive quantifier.
4219
4220 15. Modified the Makefile to add libpcre.la as a prerequisite for
4221 libpcreposix.la because I was told this is needed for a parallel build to
4222 work.
4223
4224 16. If a pattern that contained .* following optional items at the start was
4225 studied, the wrong optimizing data was generated, leading to matching
4226 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4227 matching string must start with a or b or c. The correct conclusion for
4228 this pattern is that a match can start with any character.
4229
4230
4231 Version 4.4 13-Aug-03
4232 ---------------------
4233
4234 1. In UTF-8 mode, a character class containing characters with values between
4235 127 and 255 was not handled correctly if the compiled pattern was studied.
4236 In fixing this, I have also improved the studying algorithm for such
4237 classes (slightly).
4238
4239 2. Three internal functions had redundant arguments passed to them. Removal
4240 might give a very teeny performance improvement.
4241
4242 3. Documentation bug: the value of the capture_top field in a callout is *one
4243 more than* the number of the hightest numbered captured substring.
4244
4245 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4246 in incorrectly linking with a previously installed version. They now link
4247 explicitly with libpcre.la.
4248
4249 5. configure.in no longer needs to recognize Cygwin specially.
4250
4251 6. A problem in pcre.in for Windows platforms is fixed.
4252
4253 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4254 pcretest, it used to include the size of the study block as part of its
4255 output. Unfortunately, the structure contains a field that has a different
4256 size on different hardware architectures. This meant that the tests that
4257 showed this size failed. As the block is currently always of a fixed size,
4258 this information isn't actually particularly useful in pcretest output, so
4259 I have just removed it.
4260
4261 8. Three pre-processor statements accidentally did not start in column 1.
4262 Sadly, there are *still* compilers around that complain, even though
4263 standard C has not required this for well over a decade. Sigh.
4264
4265 9. In pcretest, the code for checking callouts passed small integers in the
4266 callout_data field, which is a void * field. However, some picky compilers
4267 complained about the casts involved for this on 64-bit systems. Now
4268 pcretest passes the address of the small integer instead, which should get
4269 rid of the warnings.
4270
4271 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4272 both compile and run time, and gives an error if an invalid UTF-8 sequence
4273 is found. There is a option for disabling this check in cases where the
4274 string is known to be correct and/or the maximum performance is wanted.
4275
4276 11. In response to a bug report, I changed one line in Makefile.in from
4277
4278 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4279 to
4280 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4281
4282 to look similar to other lines, but I have no way of telling whether this
4283 is the right thing to do, as I do not use Windows. No doubt I'll get told
4284 if it's wrong...
4285
4286
4287 Version 4.3 21-May-03
4288 ---------------------
4289
4290 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4291 Makefile.
4292
4293 2. Some refactoring to improve the quality of the code:
4294
4295 (i) The utf8_table... variables are now declared "const".
4296
4297 (ii) The code for \cx, which used the "case flipping" table to upper case
4298 lower case letters, now just substracts 32. This is ASCII-specific,
4299 but the whole concept of \cx is ASCII-specific, so it seems
4300 reasonable.
4301
4302 (iii) PCRE was using its character types table to recognize decimal and
4303 hexadecimal digits in the pattern. This is silly, because it handles
4304 only 0-9, a-f, and A-F, but the character types table is locale-
4305 specific, which means strange things might happen. A private
4306 table is now used for this - though it costs 256 bytes, a table is
4307 much faster than multiple explicit tests. Of course, the standard
4308 character types table is still used for matching digits in subject
4309 strings against \d.
4310
4311 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4312 ending in _t are). So I've renamed it as ESC_tee.
4313
4314 3. The first argument for regexec() in the POSIX wrapper should have been
4315 defined as "const".
4316
4317 4. Changed pcretest to use malloc() for its buffers so that they can be
4318 Electric Fenced for debugging.
4319
4320 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4321 to read one or more bytes before the start of the subject string. Often this
4322 had no effect on PCRE's behaviour, but in some circumstances it could
4323 provoke a segmentation fault.
4324
4325 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4326 to try to read one or more bytes before the start of the subject string.
4327
4328 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4329 UTF-8 support could misbehave in various ways if the subject string
4330 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4331 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4332 back over UTF-8 characters.)
4333
4334
4335 Version 4.2 14-Apr-03
4336 ---------------------
4337
4338 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4339
4340 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4341 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4342 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4343 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4344 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4345 and BUILD_EXEEXT
4346 Note: automatic setting of the BUILD variables is not yet working
4347 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4348 compile-time but not at link-time
4349 [LINK]: use for linking executables only
4350 make different versions for Windows and non-Windows
4351 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4352 libraries
4353 [LINK_FOR_BUILD]: new variable
4354 [OBJEXT]: use throughout
4355 [EXEEXT]: use throughout
4356 <winshared>: new target
4357 <wininstall>: new target
4358 <dftables.o>: use native compiler
4359 <dftables>: use native linker
4360 <install>: handle Windows platform correctly
4361 <clean>: ditto
4362 <check>: ditto
4363 copy DLL to top builddir before testing
4364
4365 As part of these changes, -no-undefined was removed again. This was reported
4366 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4367 in any case.
4368
4369 3. Some tidies to get rid of compiler warnings:
4370
4371 . In the match_data structure, match_limit was an unsigned long int, whereas
4372 match_call_count was an int. I've made them both unsigned long ints.
4373
4374 . In pcretest the fact that a const uschar * doesn't automatically cast to
4375 a void * provoked a warning.
4376
4377 . Turning on some more compiler warnings threw up some "shadow" variables
4378 and a few more missing casts.
4379
4380 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4381 option, a class that contained a single character with a value between 128
4382 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4383
4384 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4385 option, a class that contained several characters, but with at least one
4386 whose value was between 128 and 255 caused PCRE to crash.
4387
4388
4389 Version 4.1 12-Mar-03
4390 ---------------------
4391
4392 1. Compiling with gcc -pedantic found a couple of places where casts were
4393 needed, and a string in dftables.c that was longer than standard compilers are
4394 required to support.
4395
4396 2. Compiling with Sun's compiler found a few more places where the code could
4397 be tidied up in order to avoid warnings.
4398
4399 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4400 first of these names is deprecated in the latest Autoconf in favour of the name
4401 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4402 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4403 analogy I have changed it to CFLAGS_FOR_BUILD.
4404
4405 4. Added -no-undefined to the linking command in the Makefile, because this is
4406 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4407 linking step for the pcreposix library.
4408
4409 5. PCRE was failing to diagnose the case of two named groups with the same
4410 name.
4411
4412 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4413 literal character that is needed in the subject for a match, and scans along to
4414 ensure that it is present before embarking on the full matching process. This
4415 saves time in cases of nested unlimited repeats that are never going to match.
4416 Problem: the scan can take a lot of time if the subject is very long (e.g.
4417 megabytes), thus penalizing straightforward matches. It is now done only if the
4418 amount of subject to be scanned is less than 1000 bytes.
4419
4420 7. A lesser problem with the same optimization is that it was recording the
4421 first character of an anchored pattern as "needed", thus provoking a search
4422 right along the subject, even when the first match of the pattern was going to
4423 fail. The "needed" character is now not set for anchored patterns, unless it
4424 follows something in the pattern that is of non-fixed length. Thus, it still
4425 fulfils its original purpose of finding quick non-matches in cases of nested
4426 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4427
4428
4429 Version 4.0 17-Feb-03
4430 ---------------------
4431
4432 1. If a comment in an extended regex that started immediately after a meta-item
4433 extended to the end of string, PCRE compiled incorrect data. This could lead to
4434 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4435
4436 2. Moved to autoconf 2.53 and libtool 1.4.2.
4437
4438 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4439 the special perltest8 script is no longer needed - all the tests can be run
4440 from a single perltest script.
4441
4442 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4443 by \s. It has now been removed in PCRE. This means it isn't recognized as
4444 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4445 class [:space:] *does* include VT, thereby creating a mess.
4446
4447 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4448 space and tab.
4449
4450 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4451 its new features into the main test script, reducing the number of scripts.
4452
4453 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4454 were backward compatible, and made the (?i) apply to the whole pattern, as if
4455 /i were given. Now it behaves more logically, and applies the option setting
4456 only to what follows. PCRE has been changed to follow suit. However, if it
4457 finds options settings right at the start of the pattern, it extracts them into
4458 the global options, as before. Thus, they show up in the info data.
4459
4460 8. Added support for the \Q...\E escape sequence. Characters in between are
4461 treated as literals. This is slightly different from Perl in that $ and @ are
4462 also handled as literals inside the quotes. In Perl, they will cause variable
4463 interpolation. Note the following examples:
4464
4465 Pattern PCRE matches Perl matches
4466
4467 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4468 \Qabc\$xyz\E abc\$xyz abc\$xyz
4469 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4470
4471 For compatibility with Perl, \Q...\E sequences are recognized inside character
4472 classes as well as outside them.
4473
4474 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4475 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4476 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4477 signed/unsigned warnings.
4478
4479 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4480 option for pcretest, so I've replaced it by a simple function that does just
4481 that job.
4482
4483 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4484 "pcregrep -".
4485
4486 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4487 Java package. This provides some syntactic sugar for simple cases of what my
4488 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4489 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4490 item, you can use this simplified notation. Note that only makes sense with
4491 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4492 greediness, whatever the setting of the PCRE_UNGREEDY option.
4493
4494 13. A change of greediness default within a pattern was not taking effect at
4495 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4496 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4497 was abstracted outside.
4498
4499 14. PCRE now supports the \G assertion. It is true when the current matching
4500 position is at the start point of the match. This differs from \A when the
4501 starting offset is non-zero. Used with the /g option of pcretest (or similar
4502 code), it works in the same way as it does for Perl's /g option. If all
4503 alternatives of a regex begin with \G, the expression is anchored to the start
4504 match position, and the "anchored" flag is set in the compiled expression.
4505
4506 15. Some bugs concerning the handling of certain option changes within patterns
4507 have been fixed. These applied to options other than (?ims). For example,
4508 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4509 been the other way round. Some of this was related to change 7 above.
4510
4511 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4512 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4513 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4514 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4515
4516 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4517 mode. Unlike ".", it always matches newline, whatever the setting of
4518 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4519 assertions. Perl allows it, but it doesn't (in general) work because it can't
4520 calculate the length of the lookbehind. At least, that's the case for Perl
4521 5.8.0 - I've been told they are going to document that it doesn't work in
4522 future.
4523
4524 18. Added an error diagnosis for escapes that PCRE does not support: these are
4525 \L, \l, \N, \P, \p, \U, \u, and \X.
4526
4527 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4528 reading past the end of the pattern in cases such as /[abcd/.
4529
4530 20. PCRE was getting more memory than necessary for patterns with classes that
4531 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4532
4533 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4534 compiling PCRE for use with Virtual Pascal.
4535
4536 22. Small fix to the Makefile to make it work properly if the build is done
4537 outside the source tree.
4538
4539 23. Added a new extension: a condition to go with recursion. If a conditional
4540 subpattern starts with (?(R) the "true" branch is used if recursion has
4541 happened, whereas the "false" branch is used only at the top level.
4542
4543 24. When there was a very long string of literal characters (over 255 bytes
4544 without UTF support, over 250 bytes with UTF support), the computation of how
4545 much memory was required could be incorrect, leading to segfaults or other
4546 strange effects.
4547
4548 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4549 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4550 there was a subsequent back reference to those brackets. This meant that, for
4551 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4552 possible to check for precisely this case. All we can do is abandon the
4553 optimization if .* occurs inside capturing brackets when there are any back
4554 references whatsoever. (See below for a better fix that came later.)
4555
4556 26. The handling of the optimization for finding the first character of a
4557 non-anchored pattern, and for finding a character that is required later in the
4558 match were failing in some cases. This didn't break the matching; it just
4559 failed to optimize when it could. The way this is done has been re-implemented.
4560
4561 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4562
4563 28. Added a new feature that provides some of the functionality that Perl
4564 provides with (?{...}). The facility is termed a "callout". The way it is done
4565 in PCRE is for the caller to provide an optional function, by setting
4566 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4567 global variable. By default it is unset, which disables all calling out. To get
4568 the function called, the regex must include (?C) at appropriate points. This
4569 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4570 This provides a means of identifying different callout points. When PCRE
4571 reaches such a point in the regex, if pcre_callout has been set, the external
4572 function is called. It is provided with data in a structure called
4573 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4574 matching continues; if it returns a non-zero value, the match at the current
4575 point fails. However, backtracking will occur if possible. [This was changed
4576 later and other features added - see item 49 below.]
4577
4578 29. pcretest is upgraded to test the callout functionality. It provides a
4579 callout function that displays information. By default, it shows the start of
4580 the match and the current position in the text. There are some new data escapes
4581 to vary what happens:
4582
4583 \C+ in addition, show current contents of captured substrings
4584 \C- do not supply a callout function
4585 \C!n return 1 when callout number n is reached
4586 \C!n!m return 1 when callout number n is reached for the mth time
4587
4588 30. If pcregrep was called with the -l option and just a single file name, it
4589 output "<stdin>" if a match was found, instead of the file name.
4590
4591 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4592 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4593 pcre_exec(). This saves a malloc/free per call. The default value of
4594 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4595 when configuring.
4596
4597 32. The default maximum size of a compiled pattern is 64K. There have been a
4598 few cases of people hitting this limit. The code now uses macros to handle the
4599 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4600 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4601 configuring. Tests 2 and 5 work only with 2-byte links because they output
4602 debugging information about compiled patterns.
4603
4604 33. Internal code re-arrangements:
4605
4606 (a) Moved the debugging function for printing out a compiled regex into
4607 its own source file (printint.c) and used #include to pull it into
4608 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4609 separate copies.
4610
4611 (b) Defined the list of op-code names for debugging as a macro in
4612 internal.h so that it is next to the definition of the opcodes.
4613
4614 (c) Defined a table of op-code lengths for simpler skipping along compiled
4615 code. This is again a macro in internal.h so that it is next to the
4616 definition of the opcodes.
4617
4618 34. Added support for recursive calls to individual subpatterns, along the
4619 lines of Robin Houston's patch (but implemented somewhat differently).
4620
4621 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4622 allow it to read and process whole directories in Win32. This code was
4623 contributed by Lionel Fourquaux; it has not been tested by me.
4624
4625 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4626 used to name a group. Names consist of alphanumerics and underscores, and must
4627 be unique. Back references use the syntax (?P=name) and recursive calls use
4628 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4629 numbers. The function pcre_fullinfo() can be used after compilation to extract
4630 a name/number map. There are three relevant calls:
4631
4632 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4633 PCRE_INFO_NAMECOUNT yields the number of entries
4634 PCRE_INFO_NAMETABLE yields a pointer to the map.
4635
4636 The map is a vector of fixed-size entries. The size of each entry depends on
4637 the length of the longest name used. The first two bytes of each entry are the
4638 group number, most significant byte first. There follows the corresponding
4639 name, zero terminated. The names are in alphabetical order.
4640
4641 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4642 case instead of 255. Making it the same both with and without UTF-8 support
4643 means that the same test output works with both.
4644
4645 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4646 calling malloc() with a zero argument.
4647
4648 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4649 optimization. I've improved things by keeping a bitmap of backreferences with
4650 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4651 fact referenced, the optimization can be applied. It is unlikely that a
4652 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4653 the match to follow \n) will appear inside brackets with a number greater than
4654 31, but if it does, any back reference > 31 suppresses the optimization.
4655
4656 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4657 of disabling numbered capturing parentheses. Any opening parenthesis that is
4658 not followed by ? behaves as if it were followed by ?: but named parentheses
4659 can still be used for capturing (and they will acquire numbers in the usual
4660 way).
4661
4662 41. Redesigned the return codes from the match() function into yes/no/error so
4663 that errors can be passed back from deep inside the nested calls. A malloc
4664 failure while inside a recursive subpattern call now causes the
4665 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4666
4667 42. It is now possible to set a limit on the number of times the match()
4668 function is called in a call to pcre_exec(). This facility makes it possible to
4669 limit the amount of recursion and backtracking, though not in a directly
4670 obvious way, because the match() function is used in a number of different
4671 circumstances. The count starts from zero for each position in the subject
4672 string (for non-anchored patterns). The default limit is, for compatibility, a
4673 large number, namely 10 000 000. You can change this in two ways:
4674
4675 (a) When configuring PCRE before making, you can use --with-match-limit=n
4676 to set a default value for the compiled library.
4677
4678 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4679 a different value is set. See 45 below.
4680
4681 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4682
4683 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4684 of things that can be changed at compile time. The first argument specifies
4685 what is wanted and the second points to where the information is to be placed.
4686 The current list of available information is:
4687
4688 PCRE_CONFIG_UTF8
4689
4690 The output is an integer that is set to one if UTF-8 support is available;
4691 otherwise it is set to zero.
4692
4693 PCRE_CONFIG_NEWLINE
4694
4695 The output is an integer that it set to the value of the code that is used for
4696 newline. It is either LF (10) or CR (13).
4697
4698 PCRE_CONFIG_LINK_SIZE
4699
4700 The output is an integer that contains the number of bytes used for internal
4701 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4702
4703 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4704
4705 The output is an integer that contains the threshold above which the POSIX
4706 interface uses malloc() for output vectors. See item 31 above.
4707
4708 PCRE_CONFIG_MATCH_LIMIT
4709
4710 The output is an unsigned integer that contains the default limit of the number
4711 of match() calls in a pcre_exec() execution. See 42 above.
4712
4713 44. pcretest has been upgraded by the addition of the -C option. This causes it
4714 to extract all the available output from the new pcre_config() function, and to
4715 output it. The program then exits immediately.
4716
4717 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4718 order to support additional features. One way would have been to define
4719 pcre_exec2() (for example) with extra arguments, but this would not have been
4720 extensible, and would also have required all calls to the original function to
4721 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4722 is used for passing in "extra" data from pcre_study().
4723
4724 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4725 contains the following fields:
4726
4727 flags a bitmap indicating which of the following fields are set
4728 study_data opaque data from pcre_study()
4729 match_limit a way of specifying a limit on match() calls for a specific
4730 call to pcre_exec()
4731 callout_data data for callouts (see 49 below)
4732
4733 The flag bits are also defined in pcre.h, and are
4734
4735 PCRE_EXTRA_STUDY_DATA
4736 PCRE_EXTRA_MATCH_LIMIT
4737 PCRE_EXTRA_CALLOUT_DATA
4738
4739 The pcre_study() function now returns one of these new pcre_extra blocks, with
4740 the actual study data pointed to by the study_data field, and the
4741 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4742 before. That is, this change is entirely upwards-compatible and requires no
4743 change to existing code.
4744
4745 If you want to pass in additional data to pcre_exec(), you can either place it
4746 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4747 block.
4748
4749 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4750 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4751 times with different match limits, until it finds the minimum value needed for
4752 pcre_exec() to complete. The value is then output. This can be instructive; for
4753 most simple matches the number is quite small, but for pathological cases it
4754 gets very large very quickly.
4755
4756 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4757 returns the size of the data block pointed to by the study_data field in a
4758 pcre_extra block, that is, the value that was passed as the argument to
4759 pcre_malloc() when PCRE was getting memory in which to place the information
4760 created by pcre_study(). The fourth argument should point to a size_t variable.
4761 pcretest has been extended so that this information is shown after a successful
4762 pcre_study() call when information about the compiled regex is being displayed.
4763
4764 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4765 because what follows is always an absolute path. (Later: it turns out that this
4766 is more than cosmetic for MinGW, because it doesn't like empty path
4767 components.)
4768
4769 49. Some changes have been made to the callout feature (see 28 above):
4770
4771 (i) A callout function now has three choices for what it returns:
4772
4773 0 => success, carry on matching
4774 > 0 => failure at this point, but backtrack if possible
4775 < 0 => serious error, return this value from pcre_exec()
4776
4777 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4778 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4779 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4780 use by callout functions. It will never be used by PCRE itself.
4781
4782 (ii) The pcre_extra structure (see 45 above) has a void * field called
4783 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4784 pcre_callout_block structure has a field of the same name. The contents of
4785 the field passed in the pcre_extra structure are passed to the callout
4786 function in the corresponding field in the callout block. This makes it
4787 easier to use the same callout-containing regex from multiple threads. For
4788 testing, the pcretest program has a new data escape
4789
4790 \C*n pass the number n (may be negative) as callout_data
4791
4792 If the callout function in pcretest receives a non-zero value as
4793 callout_data, it returns that value.
4794
4795 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4796 there were some redundant $(CFLAGS) in commands that are now specified as
4797 $(LINK), which already includes $(CFLAGS).
4798
4799 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4800 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4801 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4802 one-byte characters throughout. Note that case-insensitive matching applies
4803 only to characters whose values are less than 256. PCRE doesn't support the
4804 notion of cases for higher-valued characters.
4805
4806 (i) A character class whose characters are all within 0-255 is handled as
4807 a bit map, and the map is inverted for negative classes. Previously, a
4808 character > 255 always failed to match such a class; however it should
4809 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4810
4811 (ii) A negated character class with a single character < 255 is coded as
4812 "not this character" (OP_NOT). This wasn't working properly when the test
4813 character was multibyte, either singly or repeated.
4814
4815 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4816 mode, for example: \x{100}{2,3}.
4817
4818 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4819 singly or repeated) now correctly test multibyte characters. However,
4820 PCRE doesn't recognize any characters with values greater than 255 as
4821 digits, spaces, or word characters. Such characters always match \D, \S,
4822 and \W, and never match \d, \s, or \w.
4823
4824 (v) Classes may now contain characters and character ranges with values
4825 greater than 255. For example: [ab\x{100}-\x{400}].
4826
4827 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4828 PCRE in UTF-8 mode.
4829
4830 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4831 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4832 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4833 value.)
4834
4835 53. The single man page has become too large. I have therefore split it up into
4836 a number of separate man pages. These also give rise to individual HTML pages;
4837 these are now put in a separate directory, and there is an index.html page that
4838 lists them all. Some hyperlinking between the pages has been installed.
4839
4840 54. Added convenience functions for handling named capturing parentheses.
4841
4842 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4843 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4844 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4845 are faulted.
4846
4847 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4848 calling configure. These values are used when compiling the dftables.c program
4849 which is run to generate the source of the default character tables. They
4850 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4851 you will need to set these values.
4852
4853 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4854
4855
4856 Version 3.9 02-Jan-02
4857 ---------------------
4858
4859 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4860
4861 2. If --disable-static was given, the building process failed when trying to
4862 build pcretest and pcregrep. (For some reason it was using libtool to compile
4863 them, which is not right, as they aren't part of the library.)
4864
4865
4866 Version 3.8 18-Dec-01
4867 ---------------------
4868
4869 1. The experimental UTF-8 code was completely screwed up. It was packing the
4870 bytes in the wrong order. How dumb can you get?
4871
4872
4873 Version 3.7 29-Oct-01
4874 ---------------------
4875
4876 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4877 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4878 this didn't happen under Solaris 8, where I normally test things.
4879
4880 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4881 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4882 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4883 this fix an hour or so after the initial 3.7 release.)
4884
4885
4886 Version 3.6 23-Oct-01
4887 ---------------------
4888
4889 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4890 offsets passed as NULL with zero offset count.
4891
4892 2. The config.guess and config.sub files had not been updated when I moved to
4893 the latest autoconf.
4894
4895
4896 Version 3.5 15-Aug-01
4897 ---------------------
4898
4899 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4900 had been forgotten.
4901
4902 2. By using declared but undefined structures, we can avoid using "void"
4903 definitions in pcre.h while keeping the internal definitions of the structures
4904 private.
4905
4906 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4907 user point of view, this means that both static and shared libraries are built
4908 by default, but this can be individually controlled. More of the work of
4909 handling this static/shared cases is now inside libtool instead of PCRE's make
4910 file.
4911
4912 4. The pcretest utility is now installed along with pcregrep because it is
4913 useful for users (to test regexs) and by doing this, it automatically gets
4914 relinked by libtool. The documentation has been turned into a man page, so
4915 there are now .1, .txt, and .html versions in /doc.
4916
4917 5. Upgrades to pcregrep:
4918 (i) Added long-form option names like gnu grep.
4919 (ii) Added --help to list all options with an explanatory phrase.
4920 (iii) Added -r, --recursive to recurse into sub-directories.
4921 (iv) Added -f, --file to read patterns from a file.
4922
4923 6. pcre_exec() was referring to its "code" argument before testing that
4924 argument for NULL (and giving an error if it was NULL).
4925
4926 7. Upgraded Makefile.in to allow for compiling in a different directory from
4927 the source directory.
4928
4929 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4930 options bits, the pointer it was passed was to an int instead of to an unsigned
4931 long int. This mattered only on 64-bit systems.
4932
4933 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4934 generated) instead of pcre.in, which it its source. Also made the same change
4935 in several of the .c files.
4936
4937 10. A new release of gcc defines printf() as a macro, which broke pcretest
4938 because it had an ifdef in the middle of a string argument for printf(). Fixed
4939 by using separate calls to printf().
4940
4941 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4942 script, to force use of CR or LF instead of \n in the source. On non-Unix
4943 systems, the value can be set in config.h.
4944
4945 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4946 absolute limit. Changed the text of the error message to make this clear, and
4947 likewise updated the man page.
4948
4949 13. The limit of 99 on the number of capturing subpatterns has been removed.
4950 The new limit is 65535, which I hope will not be a "real" limit.
4951
4952
4953 Version 3.4 22-Aug-00
4954 ---------------------
4955
4956 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4957
4958 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4959
4960
4961 Version 3.3 01-Aug-00
4962 ---------------------
4963
4964 1. If an octal character was given, but the value was greater than \377, it
4965 was not getting masked to the least significant bits, as documented. This could
4966 lead to crashes in some systems.
4967
4968 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4969 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4970
4971 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4972 These just pass their arguments on to (pcre_free)(), but they are provided
4973 because some uses of PCRE bind it to non-C systems that can call its functions,
4974 but cannot call free() or pcre_free() directly.
4975
4976 4. Add "make test" as a synonym for "make check". Corrected some comments in
4977 the Makefile.
4978
4979 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4980 Makefile.
4981
4982 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4983 command called pgrep for grepping around the active processes.
4984
4985 7. Added the beginnings of support for UTF-8 character strings.
4986
4987 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4988 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4989 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4990 out for the ar command.)
4991
4992
4993 Version 3.2 12-May-00
4994 ---------------------
4995
4996 This is purely a bug fixing release.
4997
4998 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4999 of ZA. This was just one example of several cases that could provoke this bug,
5000 which was introduced by change 9 of version 2.00. The code for breaking
5001 infinite loops after an iteration that matches an empty string was't working
5002 correctly.
5003
5004 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5005 when matched against abbab (for example). After matching an empty string, it
5006 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5007 caused it to match further down the string than it should.
5008
5009 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5010 was there because it doesn't seem to be needed, and it causes trouble on some
5011 systems, as it is not a Standard C header. It has been removed.
5012
5013 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5014 were reported on the Macintosh. The changes were from
5015
5016 while ((c = *(++ptr)) != 0 && c != '\n');
5017 to
5018 while ((c = *(++ptr)) != 0 && c != '\n') ;
5019
5020 Totally extraordinary, but if that's what it takes...
5021
5022 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5023 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5024 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5025 assumes the way PCRE uses memmove() (always moving upwards).
5026
5027 6. PCRE is being used in one environment where strchr() is not available. There
5028 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5029 faster code anyway.
5030
5031
5032 Version 3.1 09-Feb-00
5033 ---------------------
5034
5035 The only change in this release is the fixing of some bugs in Makefile.in for
5036 the "install" target:
5037
5038 (1) It was failing to install pcreposix.h.
5039
5040 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5041
5042
5043 Version 3.0 01-Feb-00
5044 ---------------------
5045
5046 1. Add support for the /+ modifier to perltest (to output $` like it does in
5047 pcretest).
5048
5049 2. Add support for the /g modifier to perltest.
5050
5051 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5052 matches null strings.
5053
5054 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5055 pattern. Perl treats empty patterns specially - it reuses the most recent
5056 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5057 effect.
5058
5059 5. The POSIX interface was broken in that it was just handing over the POSIX
5060 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5061 required a bigger vector, with some working space on the end. This means that
5062 the POSIX wrapper now has to get and free some memory, and copy the results.
5063
5064 6. Added some simple autoconf support, placing the test data and the
5065 documentation in separate directories, re-organizing some of the
5066 information files, and making it build pcre-config (a GNU standard). Also added
5067 libtool support for building PCRE as a shared library, which is now the
5068 default.
5069
5070 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5071 09 are not valid octal constants. Single digits will be used for minor values
5072 less than 10.
5073
5074 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5075 existing programs that set these in the POSIX interface can use PCRE without
5076 modification.
5077
5078 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5079 return all that pcre_info() returns, plus additional data. The pcre_info()
5080 function is retained for compatibility, but is considered to be obsolete.
5081
5082 10. Added experimental recursion feature (?R) to handle one common case that
5083 Perl 5.6 will be able to do with (?p{...}).
5084
5085 11. Added support for POSIX character classes like [:alpha:], which Perl is
5086 adopting.
5087
5088
5089 Version 2.08 31-Aug-99
5090 ----------------------
5091
5092 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5093 trying to match at the startoffset position, but instead was moving forward to
5094 the next newline as if a previous match had failed.
5095
5096 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5097 and could get into a loop if a null string was matched other than at the start
5098 of the subject.
5099
5100 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5101 be distinguished at compile time, and for completeness also added PCRE_DATE.
5102
5103 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5104 in GnuWin32 environments.
5105
5106
5107 Version 2.07 29-Jul-99
5108 ----------------------
5109
5110 1. The documentation is now supplied in plain text form and HTML as well as in
5111 the form of man page sources.
5112
5113 2. C++ compilers don't like assigning (void *) values to other pointer types.
5114 In particular this affects malloc(). Although there is no problem in Standard
5115 C, I've put in casts to keep C++ compilers happy.
5116
5117 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5118 should be (const char *).
5119
5120 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5121 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5122 However, I haven't made this a standard facility. The documentation doesn't
5123 mention it, and the Makefile doesn't support it.
5124
5125 5. The Makefile now contains an "install" target, with editable destinations at
5126 the top of the file. The pcretest program is not installed.
5127
5128 6. pgrep -V now gives the PCRE version number and date.
5129
5130 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5131 causing the entire string to be ignored, instead of just the last character.
5132
5133 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5134 non-matching string, it can take a very, very long time, even for strings of
5135 quite modest length, because of the nested recursion. PCRE now does better in
5136 some of these cases. It does this by remembering the last required literal
5137 character in the pattern, and pre-searching the subject to ensure it is present
5138 before running the real match. In other words, it applies a heuristic to detect
5139 some types of certain failure quickly, and in the above example, if presented
5140 with a string that has no trailing " it gives "no match" very quickly.
5141
5142 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5143 other alternatives are tried instead.
5144
5145
5146 Version 2.06 09-Jun-99
5147 ----------------------
5148
5149 1. Change pcretest's output for amount of store used to show just the code
5150 space, because the remainder (the data block) varies in size between 32-bit and
5151 64-bit systems.
5152
5153 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5154 start matching at. This allows lookbehinds to work when searching for multiple
5155 occurrences in a string.
5156
5157 3. Added additional options to pcretest for testing multiple occurrences:
5158
5159 /+ outputs the rest of the string that follows a match
5160 /g loops for multiple occurrences, using the new startoffset argument
5161 /G loops for multiple occurrences by passing an incremented pointer
5162
5163 4. PCRE wasn't doing the "first character" optimization for patterns starting
5164 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5165 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5166 the letter 'x'. On long subject strings, this gives a significant speed-up.
5167
5168
5169 Version 2.05 21-Apr-99
5170 ----------------------
5171
5172 1. Changed the type of magic_number from int to long int so that it works
5173 properly on 16-bit systems.
5174
5175 2. Fixed a bug which caused patterns starting with .* not to work correctly
5176 when the subject string contained newline characters. PCRE was assuming
5177 anchoring for such patterns in all cases, which is not correct because .* will
5178 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5179 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5180 must be retried after every newline in the subject.
5181
5182
5183 Version 2.04 18-Feb-99
5184 ----------------------
5185
5186 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5187 computation of the store needed to hold the pattern was incorrect (too large).
5188 If such patterns were nested a few deep, this could multiply and become a real
5189 problem.
5190
5191 2. Added /M option to pcretest to show the memory requirement of a specific
5192 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5193
5194 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5195 compiled in such a way that the backtracking after subsequent failure was
5196 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5197 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5198
5199
5200 Version 2.03 02-Feb-99
5201 ----------------------
5202
5203 1. Fixed typo and small mistake in man page.
5204
5205 2. Added 4th condition (GPL supersedes if conflict) and created separate
5206 LICENCE file containing the conditions.
5207
5208 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5209 Perl, that is the internal \ allows the delimiter to be included in the
5210 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5211 the final delimiter, add \ to the end of the pattern (to test the error).
5212
5213 4. Added the convenience functions for extracting substrings after a successful
5214 match. Updated pcretest to make it able to test these functions.
5215
5216
5217 Version 2.02 14-Jan-99
5218 ----------------------
5219
5220 1. Initialized the working variables associated with each extraction so that
5221 their saving and restoring doesn't refer to uninitialized store.
5222
5223 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5224 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5225 fix the problem.
5226
5227 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5228 calls, and wasn't printing the correct value for compiling calls. Increased the
5229 default value of LOOPREPEAT, and the number of significant figures in the
5230 times.
5231
5232 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5233
5234 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5235 a building problem on Windows NT with a FAT file system.
5236
5237
5238 Version 2.01 21-Oct-98
5239 ----------------------
5240
5241 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5242 to character tables built by pcre_maketables() in the current locale. If NULL
5243 is passed, the default tables are used.
5244
5245
5246 Version 2.00 24-Sep-98
5247 ----------------------
5248
5249 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5250 it any more.
5251
5252 2. Allow quantification of (?>) groups, and make it work correctly.
5253
5254 3. The first character computation wasn't working for (?>) groups.
5255
5256 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5257 end of the subject) and add 5.005's \z, which really does match only at the
5258 very end of the subject.
5259
5260 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5261
5262 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5263 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5264 localized options. All options to pcre_study() were also removed.
5265
5266 7. Add other new features from 5.005:
5267
5268 $(?<= positive lookbehind
5269 $(?<! negative lookbehind
5270 (?imsx-imsx) added the unsetting capability
5271 such a setting is global if at outer level; local otherwise
5272 (?imsx-imsx:) non-capturing groups with option setting
5273 (?(cond)re|re) conditional pattern matching
5274
5275 A backreference to itself in a repeated group matches the previous
5276 captured string.
5277
5278 8. General tidying up of studying (both automatic and via "study")
5279 consequential on the addition of new assertions.
5280
5281 9. As in 5.005, unlimited repeated groups that could match an empty substring
5282 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5283 runtime if any iteration does actually match an empty substring.
5284
5285 10. Include the RunTest script in the distribution.
5286
5287 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5288 discrepancies, some of which were old and were also with respect to 5.004. They
5289 have now been fixed.
5290
5291
5292 Version 1.09 28-Apr-98
5293 ----------------------
5294
5295 1. A negated single character class followed by a quantifier with a minimum
5296 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5297 program crashes, or just wrong answers. This did not apply to negated classes
5298 containing more than one character, or to minima other than one.
5299
5300
5301 Version 1.08 27-Mar-98
5302 ----------------------
5303
5304 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5305
5306 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5307 latter must appear before anything that relies on it in the pattern.
5308
5309
5310 Version 1.07 16-Feb-98
5311 ----------------------
5312
5313 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5314 repeat of a potentially empty string).
5315
5316
5317 Version 1.06 23-Jan-98
5318 ----------------------
5319
5320 1. Added Markus Oberhumer's little patches for C++.
5321
5322 2. Literal strings longer than 255 characters were broken.
5323
5324
5325 Version 1.05 23-Dec-97
5326 ----------------------
5327
5328 1. Negated character classes containing more than one character were failing if
5329 PCRE_CASELESS was set at run time.
5330
5331
5332 Version 1.04 19-Dec-97
5333 ----------------------
5334
5335 1. Corrected the man page, where some "const" qualifiers had been omitted.
5336
5337 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5338 input syntax.
5339
5340 3. Fixed memory leak which occurred when a regex with back references was
5341 matched with an offsets vector that wasn't big enough. The temporary memory
5342 that is used in this case wasn't being freed if the match failed.
5343
5344 4. Tidied pcretest to ensure it frees memory that it gets.
5345
5346 5. Temporary memory was being obtained in the case where the passed offsets
5347 vector was exactly big enough.
5348
5349 6. Corrected definition of offsetof() from change 5 below.
5350
5351 7. I had screwed up change 6 below and broken the rules for the use of
5352 setjmp(). Now fixed.
5353
5354
5355 Version 1.03 18-Dec-97
5356 ----------------------
5357
5358 1. A erroneous regex with a missing opening parenthesis was correctly
5359 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5360 on some systems.
5361
5362 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5363 it was reported that one broken compiler failed on the former because "code" is
5364 also an independent variable.
5365
5366 3. The erroneous regex a[]b caused an array overrun reference.
5367
5368 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5369 fail on data ending with that character. (It was going on too far, and checking
5370 the next character, typically a binary zero.) This was specific to the
5371 optimized code for single-character negative classes.
5372
5373 5. Added a contributed patch from the TIN world which does the following:
5374
5375 + Add an undef for memmove, in case the the system defines a macro for it.
5376
5377 + Add a definition of offsetof(), in case there isn't one. (I don't know
5378 the reason behind this - offsetof() is part of the ANSI standard - but
5379 it does no harm).
5380
5381 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5382 most of the places where whitespace preceded '#'. I have given up and
5383 allowed the remaining 2 cases to be at the margin.
5384
5385 + Rename some variables in pcre to eliminate shadowing. This seems very
5386 pedantic, but does no harm, of course.
5387
5388 6. Moved the call to setjmp() into its own function, to get rid of warnings
5389 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5390
5391 7. Constructs such as \d{8,} were compiling into the equivalent of
5392 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5393 outcome, but in this particular case used more store than had been allocated,
5394 which caused the bug to be discovered because it threw up an internal error.
5395
5396 8. The debugging code in both pcre and pcretest for outputting the compiled
5397 form of a regex was going wrong in the case of back references followed by
5398 curly-bracketed repeats.
5399
5400
5401 Version 1.02 12-Dec-97
5402 ----------------------
5403
5404 1. Typos in pcre.3 and comments in the source fixed.
5405
5406 2. Applied a contributed patch to get rid of places where it used to remove
5407 'const' from variables, and fixed some signed/unsigned and uninitialized
5408 variable warnings.
5409
5410 3. Added the "runtest" target to Makefile.
5411
5412 4. Set default compiler flag to -O2 rather than just -O.
5413
5414
5415 Version 1.01 19-Nov-97
5416 ----------------------
5417
5418 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5419 like /([ab]*)*/, that is, for classes with more than one character in them.
5420
5421 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5422 as /((?>a*))*/ (a PCRE_EXTRA facility).
5423
5424
5425 Version 1.00 18-Nov-97
5426 ----------------------
5427
5428 1. Added compile-time macros to support systems such as SunOS4 which don't have
5429 memmove() or strerror() but have other things that can be used instead.
5430
5431 2. Arranged that "make clean" removes the executables.
5432
5433
5434 Version 0.99 27-Oct-97
5435 ----------------------
5436
5437 1. Fixed bug in code for optimizing classes with only one character. It was
5438 initializing a 32-byte map regardless, which could cause it to run off the end
5439 of the memory it had got.
5440
5441 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5442
5443
5444 Version 0.98 22-Oct-97
5445 ----------------------
5446
5447 1. Fixed bug in code for handling temporary memory usage when there are more
5448 back references than supplied space in the ovector. This could cause segfaults.
5449
5450
5451 Version 0.97 21-Oct-97
5452 ----------------------
5453
5454 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5455
5456 2. Optimized negated single characters not to use a bit map.
5457
5458 3. Brought error texts together as macro definitions; clarified some of them;
5459 fixed one that was wrong - it said "range out of order" when it meant "invalid
5460 escape sequence".
5461
5462 4. Changed some char * arguments to const char *.
5463
5464 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5465
5466 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5467 pcretest.
5468
5469
5470 Version 0.96 16-Oct-97
5471 ----------------------
5472
5473 1. Added a simple "pgrep" utility to the distribution.
5474
5475 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5476 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5477 where "ddd" means "one or more decimal digits".
5478
5479 3. Fixed serious bug. If a pattern had a back reference, but the call to
5480 pcre_exec() didn't supply a large enough ovector to record the related
5481 identifying subpattern, the match always failed. PCRE now remembers the number
5482 of the largest back reference, and gets some temporary memory in which to save
5483 the offsets during matching if necessary, in order to ensure that
5484 backreferences always work.
5485
5486 4. Increased the compatibility with Perl in a number of ways:
5487
5488 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5489 to request this handling. The option can be set at compile or exec time.
5490
5491 (b) $ matches before a terminating newline by default; an option
5492 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5493 mode). The option can be set at compile or exec time.
5494
5495 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5496 the same as Perl's. If the decimal number it represents is less than 10
5497 or there aren't that many previous left capturing parentheses, an octal
5498 escape is read. Inside a character class, it's always an octal escape,
5499 even if it is a single digit.
5500
5501 (d) An escaped but undefined alphabetic character is taken as a literal,
5502 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5503 escapes.
5504
5505 (e) {0} is now permitted. (The previous item is removed from the compiled
5506 pattern).
5507
5508 5. Changed all the names of code files so that the basic parts are no longer
5509 than 10 characters, and abolished the teeny "globals.c" file.
5510
5511 6. Changed the handling of character classes; they are now done with a 32-byte
5512 bit map always.
5513
5514 7. Added the -d and /D options to pcretest to make it possible to look at the
5515 internals of compilation without having to recompile pcre.
5516
5517
5518 Version 0.95 23-Sep-97
5519 ----------------------
5520
5521 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5522 \x20 at the start of a run of normal characters. These were being treated as
5523 real characters, instead of the source characters being re-checked.
5524
5525
5526 Version 0.94 18-Sep-97
5527 ----------------------
5528
5529 1. The functions are now thread-safe, with the caveat that the global variables
5530 containing pointers to malloc() and free() or alternative functions are the
5531 same for all threads.
5532
5533 2. Get pcre_study() to generate a bitmap of initial characters for non-
5534 anchored patterns when this is possible, and use it if passed to pcre_exec().
5535
5536
5537 Version 0.93 15-Sep-97
5538 ----------------------
5539
5540 1. /(b)|(:+)/ was computing an incorrect first character.
5541
5542 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5543 but not actually doing anything yet.
5544
5545 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5546 as Perl does (e.g. [-az] or [az-]).
5547
5548 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5549 all possible positions.
5550
5551 5. Split up into different modules to avoid including unneeded functions in a
5552 compiled binary. However, compile and exec are still in one module. The "study"
5553 function is split off.
5554
5555 6. The character tables are now in a separate module whose source is generated
5556 by an auxiliary program - but can then be edited by hand if required. There are
5557 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5558 toupper() in the code.
5559
5560 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5561 make them global. Abolish the function for setting them, as the caller can now
5562 set them directly.
5563
5564
5565 Version 0.92 11-Sep-97
5566 ----------------------
5567
5568 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5569 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5570
5571 2. Caseless matching was not working in character classes if the characters in
5572 the pattern were in upper case.
5573
5574 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5575
5576 4. Make PCRE_ANCHORED public and accept as a compile option.
5577
5578 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5579 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5580 pass them.
5581
5582 6. Give an error if bad option bits passed at compile or run time.
5583
5584 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5585 pcretest to cause it to pass that flag.
5586
5587 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5588 options, and the first character, if set.
5589
5590 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5591
5592
5593 Version 0.91 10-Sep-97
5594 ----------------------
5595
5596 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5597 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5598
5599 2. PCRE was looping on encountering an indefinitely repeated back reference to
5600 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5601 Perl does - treats the match as successful.
5602
5603 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5