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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1517 - (show annotations)
Fri Feb 6 12:10:13 2015 UTC (4 years, 9 months ago) by ph10
File size: 269575 byte(s)
Fix memory leak for subroutine call with more than 10 captured groups to be 
saved.
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
28 Version 8.36 26-September-2014
29 ------------------------------
30
31 1. Got rid of some compiler warnings in the C++ modules that were shown up by
32 -Wmissing-field-initializers and -Wunused-parameter.
33
34 2. The tests for quantifiers being too big (greater than 65535) were being
35 applied after reading the number, and stupidly assuming that integer
36 overflow would give a negative number. The tests are now applied as the
37 numbers are read.
38
39 3. Tidy code in pcre_exec.c where two branches that used to be different are
40 now the same.
41
42 4. The JIT compiler did not generate match limit checks for certain
43 bracketed expressions with quantifiers. This may lead to exponential
44 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
45 issue should be resolved now.
46
47 5. Fixed an issue, which occures when nested alternatives are optimized
48 with table jumps.
49
50 6. Inserted two casts and changed some ints to size_t in the light of some
51 reported 64-bit compiler warnings (Bugzilla 1477).
52
53 7. Fixed a bug concerned with zero-minimum possessive groups that could match
54 an empty string, which sometimes were behaving incorrectly in the
55 interpreter (though correctly in the JIT matcher). This pcretest input is
56 an example:
57
58 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
59 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
60
61 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
62 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
63 for an empty string was breaking the inner loop and carrying on at a lower
64 level, when possessive repeated groups should always return to a higher
65 level as they have no backtrack points in them. The empty string test now
66 occurs at the outer level.
67
68 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
69 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
70
71 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
72 doesn't).
73
74 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
75 to hold a pattern to be incorrectly computed (too small) when there were
76 named back references to duplicated names. This could cause "internal
77 error: code overflow" or "double free or corruption" or other memory
78 handling errors.
79
80 11. When named subpatterns had the same prefixes, back references could be
81 confused. For example, in this pattern:
82
83 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
84
85 the reference to 'Name' was incorrectly treated as a reference to a
86 duplicate name.
87
88 12. A pattern such as /^s?c/mi8 where the optional character has more than
89 one "other case" was incorrectly compiled such that it would only try to
90 match starting at "c".
91
92 13. When a pattern starting with \s was studied, VT was not included in the
93 list of possible starting characters; this should have been part of the
94 8.34/18 patch.
95
96 14. If a character class started [\Qx]... where x is any character, the class
97 was incorrectly terminated at the ].
98
99 15. If a pattern that started with a caseless match for a character with more
100 than one "other case" was studied, PCRE did not set up the starting code
101 unit bit map for the list of possible characters. Now it does. This is an
102 optimization improvement, not a bug fix.
103
104 16. The Unicode data tables have been updated to Unicode 7.0.0.
105
106 17. Fixed a number of memory leaks in pcregrep.
107
108 18. Avoid a compiler warning (from some compilers) for a function call with
109 a cast that removes "const" from an lvalue by using an intermediate
110 variable (to which the compiler does not object).
111
112 19. Incorrect code was compiled if a group that contained an internal recursive
113 back reference was optional (had quantifier with a minimum of zero). This
114 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
115 caused segmentation faults because of stack overflows at compile time.
116
117 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
118 group that is quantified with an indefinite repeat, caused a compile-time
119 loop which used up all the system stack and provoked a segmentation fault.
120 This was not the same bug as 19 above.
121
122 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
123 Patch by Mike Frysinger.
124
125
126 Version 8.35 04-April-2014
127 --------------------------
128
129 1. A new flag is set, when property checks are present in an XCLASS.
130 When this flag is not set, PCRE can perform certain optimizations
131 such as studying these XCLASS-es.
132
133 2. The auto-possessification of character sets were improved: a normal
134 and an extended character set can be compared now. Furthermore
135 the JIT compiler optimizes more character set checks.
136
137 3. Got rid of some compiler warnings for potentially uninitialized variables
138 that show up only when compiled with -O2.
139
140 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
141 of a match later then the end of the match. The pcretest program was not
142 handling the case sensibly - it was outputting from the start to the next
143 binary zero. It now reports this situation in a message, and outputs the
144 text from the end to the start.
145
146 5. Fast forward search is improved in JIT. Instead of the first three
147 characters, any three characters with fixed position can be searched.
148 Search order: first, last, middle.
149
150 6. Improve character range checks in JIT. Characters are read by an inprecise
151 function now, which returns with an unknown value if the character code is
152 above a certain threshold (e.g: 256). The only limitation is that the value
153 must be bigger than the threshold as well. This function is useful when
154 the characters above the threshold are handled in the same way.
155
156 7. The macros whose names start with RAWUCHAR are placeholders for a future
157 mode in which only the bottom 21 bits of 32-bit data items are used. To
158 make this more memorable for those maintaining the code, the names have
159 been changed to start with UCHAR21, and an extensive comment has been added
160 to their definition.
161
162 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
163 to the export list in Makefile.am (they were accidentally omitted from the
164 8.34 tarball).
165
166 9. The informational output from pcretest used the phrase "starting byte set"
167 which is inappropriate for the 16-bit and 32-bit libraries. As the output
168 for "first char" and "need char" really means "non-UTF-char", I've changed
169 "byte" to "char", and slightly reworded the output. The documentation about
170 these values has also been (I hope) clarified.
171
172 10. Another JIT related optimization: use table jumps for selecting the correct
173 backtracking path, when more than four alternatives are present inside a
174 bracket.
175
176 11. Empty match is not possible, when the minimum length is greater than zero,
177 and there is no \K in the pattern. JIT should avoid empty match checks in
178 such cases.
179
180 12. In a caseless character class with UCP support, when a character with more
181 than one alternative case was not the first character of a range, not all
182 the alternative cases were added to the class. For example, s and \x{17f}
183 are both alternative cases for S: the class [RST] was handled correctly,
184 but [R-T] was not.
185
186 13. The configure.ac file always checked for pthread support when JIT was
187 enabled. This is not used in Windows, so I have put this test inside a
188 check for the presence of windows.h (which was already tested for).
189
190 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
191 The algorithm provides a way to skip certain starting offsets, and usually
192 faster than linear prefix searches.
193
194 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
195 as for 'fr_FR' and 'french'. For some reason, however, it then used the
196 Windows-specific input and output files, which have 'french' screwed in.
197 So this could never have worked. One of the problems with locales is that
198 they aren't always the same. I have now updated RunTest so that it checks
199 the output of the locale test (test 3) against three different output
200 files, and it allows the test to pass if any one of them matches. With luck
201 this should make the test pass on some versions of Solaris where it was
202 failing. Because of the uncertainty, the script did not used to stop if
203 test 3 failed; it now does. If further versions of a French locale ever
204 come to light, they can now easily be added.
205
206 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
207 there was a message during ./configure, but it did not stop. This now
208 provokes an error. The invalid example in README has been corrected.
209 If a value less than the minimum is given, the minimum value has always
210 been used, but now a warning is given.
211
212 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
213 was a bug in the test system, which is now fixed. Also, the list of various
214 configurations that are tested for each release did not have one with both
215 16/32 bits and --enable-bar-anycrlf. It now does.
216
217 18. pcretest was missing "-C bsr" for displaying the \R default setting.
218
219 19. Little endian PowerPC systems are supported now by the JIT compiler.
220
221 20. The fast forward newline mechanism could enter to an infinite loop on
222 certain invalid UTF-8 input. Although we don't support these cases
223 this issue can be fixed by a performance optimization.
224
225 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
226 not take account if existing stack usage. There is now a new global
227 variable called pcre_stack_guard that can be set to point to an external
228 function to check stack availability. It is called at the start of
229 processing every parenthesized group.
230
231 22. A typo in the code meant that in ungreedy mode the max/min qualifier
232 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
233 not match "ab".
234
235 23. When UTF was disabled, the JIT program reported some incorrect compile
236 errors. These messages are silenced now.
237
238 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
239 compiler.
240
241 25. Change all the temporary files used in RunGrepTest to be different to those
242 used by RunTest so that the tests can be run simultaneously, for example by
243 "make -j check".
244
245
246 Version 8.34 15-December-2013
247 -----------------------------
248
249 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
250 executable memory. Patch inspired by Carsten Klein.
251
252 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
253 this macro is never tested and has no effect, because the work to support
254 coverage involves only compiling and linking options and special targets in
255 the Makefile. The comment in config.h implied that defining the macro would
256 enable coverage support, which is totally false. There was also support for
257 setting this macro in the CMake files (my fault, I just copied it from
258 configure). SUPPORT_GCOV has now been removed.
259
260 3. Make a small performance improvement in strlen16() and strlen32() in
261 pcretest.
262
263 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
264 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
265 these cases). There was also one in pcretest.c.
266
267 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
268
269 6. In UTF mode, the code for checking whether a group could match an empty
270 string (which is used for indefinitely repeated groups to allow for
271 breaking an infinite loop) was broken when the group contained a repeated
272 negated single-character class with a character that occupied more than one
273 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
274 UTF-8 mode). The effect was undefined: the group might or might not be
275 deemed as matching an empty string, or the program might have crashed.
276
277 7. The code for checking whether a group could match an empty string was not
278 recognizing that \h, \H, \v, \V, and \R must match a character.
279
280 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
281 an empty string. If it can, pcretest shows this in its information output.
282
283 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
284 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
285 matched by pcre_exec() without using JIT:
286
287 (a) If the rest of the pattern did not match after a maximal run of
288 grapheme clusters, the code for backing up to try with fewer of them
289 did not always back up over a full grapheme when characters that do not
290 have the modifier quality were involved, e.g. Hangul syllables.
291
292 (b) If the match point in a subject started with modifier character, and
293 there was no match, the code could incorrectly back up beyond the match
294 point, and potentially beyond the first character in the subject,
295 leading to a segfault or an incorrect match result.
296
297 10. A conditional group with an assertion condition could lead to PCRE
298 recording an incorrect first data item for a match if no other first data
299 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
300 first data item, and therefore matched "ca" after "c" instead of at the
301 start.
302
303 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
304 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
305
306 12. The source of pcregrep now includes z/OS-specific code so that it can be
307 compiled for z/OS as part of the special z/OS distribution.
308
309 13. Added the -T and -TM options to pcretest.
310
311 14. The code in pcre_compile.c for creating the table of named capturing groups
312 has been refactored. Instead of creating the table dynamically during the
313 actual compiling pass, the information is remembered during the pre-compile
314 pass (on the stack unless there are more than 20 named groups, in which
315 case malloc() is used) and the whole table is created before the actual
316 compile happens. This has simplified the code (it is now nearly 150 lines
317 shorter) and prepared the way for better handling of references to groups
318 with duplicate names.
319
320 15. A back reference to a named subpattern when there is more than one of the
321 same name now checks them in the order in which they appear in the pattern.
322 The first one that is set is used for the reference. Previously only the
323 first one was inspected. This change makes PCRE more compatible with Perl.
324
325 16. Unicode character properties were updated from Unicode 6.3.0.
326
327 17. The compile-time code for auto-possessification has been refactored, based
328 on a patch by Zoltan Herczeg. It now happens after instead of during
329 compilation. The code is cleaner, and more cases are handled. The option
330 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
331 options in pcretest are provided to set it. It can also be set by
332 (*NO_AUTO_POSSESS) at the start of a pattern.
333
334 18. The character VT has been added to the default ("C" locale) set of
335 characters that match \s and are generally treated as white space,
336 following this same change in Perl 5.18. There is now no difference between
337 "Perl space" and "POSIX space". Whether VT is treated as white space in
338 other locales depends on the locale.
339
340 19. The code for checking named groups as conditions, either for being set or
341 for being recursed, has been refactored (this is related to 14 and 15
342 above). Processing unduplicated named groups should now be as fast at
343 numerical groups, and processing duplicated groups should be faster than
344 before.
345
346 20. Two patches to the CMake build system, by Alexander Barkov:
347
348 (1) Replace the "source" command by "." in CMakeLists.txt because
349 "source" is a bash-ism.
350
351 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
352 without these the CMake build does not work on Solaris.
353
354 21. Perl has changed its handling of \8 and \9. If there is no previously
355 encountered capturing group of those numbers, they are treated as the
356 literal characters 8 and 9 instead of a binary zero followed by the
357 literals. PCRE now does the same.
358
359 22. Following Perl, added \o{} to specify codepoints in octal, making it
360 possible to specify values greater than 0777 and also making them
361 unambiguous.
362
363 23. Perl now gives an error for missing closing braces after \x{... instead of
364 treating the string as literal. PCRE now does the same.
365
366 24. RunTest used to grumble if an inappropriate test was selected explicitly,
367 but just skip it when running all tests. This make it awkward to run ranges
368 of tests when one of them was inappropriate. Now it just skips any
369 inappropriate tests, as it always did when running all tests.
370
371 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
372 character types such as \d or \w, too many callouts were inserted, and the
373 data that they returned was rubbish.
374
375 26. In UCP mode, \s was not matching two of the characters that Perl matches,
376 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
377 were matched by \h. The code has now been refactored so that the lists of
378 the horizontal and vertical whitespace characters used for \h and \v (which
379 are defined only in one place) are now also used for \s.
380
381 27. Add JIT support for the 64 bit TileGX architecture.
382 Patch by Jiong Wang (Tilera Corporation).
383
384 28. Possessive quantifiers for classes (both explicit and automatically
385 generated) now use special opcodes instead of wrapping in ONCE brackets.
386
387 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
388 (because it's meaningless), this was not happening when PCRE_CASELESS was
389 set. Not wrong, but inefficient.
390
391 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
392 properties for \w, \d, etc) is present in a test regex. Otherwise if the
393 test contains no characters greater than 255, Perl doesn't realise it
394 should be using Unicode semantics.
395
396 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
397 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
398 does in Unicode mode.
399
400 32. Added the "forbid" facility to pcretest so that putting tests into the
401 wrong test files can sometimes be quickly detected.
402
403 33. There is now a limit (default 250) on the depth of nesting of parentheses.
404 This limit is imposed to control the amount of system stack used at compile
405 time. It can be changed at build time by --with-parens-nest-limit=xxx or
406 the equivalent in CMake.
407
408 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
409 errors. Perl warns for these when in warning mode, but PCRE has no facility
410 for giving warnings.
411
412 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
413 However, this was not working for (?!) because it is optimized to (*FAIL),
414 for which PCRE does not allow quantifiers. The optimization is now disabled
415 when a quantifier follows (?!). I can't see any use for this, but it makes
416 things uniform.
417
418 36. Perl no longer allows group names to start with digits, so I have made this
419 change also in PCRE. It simplifies the code a bit.
420
421 37. In extended mode, Perl ignores spaces before a + that indicates a
422 possessive quantifier. PCRE allowed a space before the quantifier, but not
423 before the possessive +. It now does.
424
425 38. The use of \K (reset reported match start) within a repeated possessive
426 group such as (a\Kb)*+ was not working.
427
428 40. Document that the same character tables must be used at compile time and
429 run time, and that the facility to pass tables to pcre_exec() and
430 pcre_dfa_exec() is for use only with saved/restored patterns.
431
432 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
433 features for Builds with MSVC:
434
435 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
436 builds). The .rc files can be used to set FileDescription and many other
437 attributes.
438
439 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
440 This allows higher-level build scripts which want .pdb files to avoid
441 hard-coding the exact files needed."
442
443 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
444 mean "start of word" and "end of word", respectively, as a transition aid.
445
446 43. A minimizing repeat of a class containing codepoints greater than 255 in
447 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
448 compiled to use the heap for recursion.
449
450 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
451 is configured.
452
453
454 Version 8.33 28-May-2013
455 ------------------------
456
457 1. Added 'U' to some constants that are compared to unsigned integers, to
458 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
459 variables that are added to signed variables, to ensure the result is
460 signed and can be negated.
461
462 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
463 CMake config files.
464
465 3. Revise the creation of config.h.generic so that all boolean macros are
466 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
467 overriding via -D on the command line possible.
468
469 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
470 to unsigned int is reported to make a quite noticeable speed difference in
471 a specific Windows environment. Testing on Linux did also appear to show
472 some benefit (and it is clearly not harmful). Also fixed the definition of
473 Xop which should be unsigned.
474
475 5. Related to (4), changing the definition of the intermediate variable cc
476 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
477 improvements.
478
479 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
480 unnecessary spaces.
481
482 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
483 and later.
484
485 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
486
487 9. Optimizing fast_forward_start_bits in JIT.
488
489 10. Adding support for callouts in JIT, and fixing some issues revealed
490 during this work. Namely:
491
492 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
493
494 (b) Minimum length was not checked before the matching is started.
495
496 11. The value of capture_last that is passed to callouts was incorrect in some
497 cases when there was a capture on one path that was subsequently abandoned
498 after a backtrack. Also, the capture_last value is now reset after a
499 recursion, since all captures are also reset in this case.
500
501 12. The interpreter no longer returns the "too many substrings" error in the
502 case when an overflowing capture is in a branch that is subsequently
503 abandoned after a backtrack.
504
505 13. In the pathological case when an offset vector of size 2 is used, pcretest
506 now prints out the matched string after a yield of 0 or 1.
507
508 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
509 Only supported by the JIT compiler at the moment.
510
511 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
512
513 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
514 offset of the starting point of the matching process, provided the offsets
515 vector is large enough.
516
517 17. The \A escape now records a lookbehind value of 1, though its execution
518 does not actually inspect the previous character. This is to ensure that,
519 in partial multi-segment matching, at least one character from the old
520 segment is retained when a new segment is processed. Otherwise, if there
521 are no lookbehinds in the pattern, \A might match incorrectly at the start
522 of a new segment.
523
524 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
525
526 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
527 gives some modest performance improvement in 8-bit mode.
528
529 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
530 be expressed in certain programming languages using Universal Character
531 Names.
532
533 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
534 which points out that "non characters" are not "characters that may not
535 appear in Unicode strings" but rather "characters that are reserved for
536 internal use and have only local meaning".
537
538 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
539 there was a conditional group that depended on an assertion, if the
540 assertion was false, the callout that immediately followed the alternation
541 in the condition was skipped when pcre_exec() was used for matching.
542
543 23. Allow an explicit callout to be inserted before an assertion that is the
544 condition for a conditional group, for compatibility with automatic
545 callouts, which always insert a callout at this point.
546
547 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
548 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
549
550 25. (*PRUNE) is now supported by the JIT compiler.
551
552 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
553
554 27. Fix the case where there are two or more SKIPs with arguments that may be
555 ignored.
556
557 28. (*SKIP) is now supported by the JIT compiler.
558
559 29. (*THEN) is now supported by the JIT compiler.
560
561 30. Update RunTest with additional test selector options.
562
563 31. The way PCRE handles backtracking verbs has been changed in two ways.
564
565 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
566 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
567 backtracking. In some cases this makes it more Perl-compatible, but Perl's
568 rather obscure rules do not always do the same thing.
569
570 (2) Previously, backtracking verbs were confined within assertions. This is
571 no longer the case for positive assertions, except for (*ACCEPT). Again,
572 this sometimes improves Perl compatibility, and sometimes does not.
573
574 32. A number of tests that were in test 2 because Perl did things differently
575 have been moved to test 1, because either Perl or PCRE has changed, and
576 these tests are now compatible.
577
578 32. Backtracking control verbs are now handled in the same way in JIT and
579 interpreter.
580
581 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
582 contained a forward subroutine reference caused a compile error.
583
584 34. Auto-detect and optimize limited repetitions in JIT.
585
586 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
587 blocking (*UTF) etc.
588
589 36. In the interpreter, maximizing pattern repetitions for characters and
590 character types now use tail recursion, which reduces stack usage.
591
592 37. The value of the max lookbehind was not correctly preserved if a compiled
593 and saved regex was reloaded on a host of different endianness.
594
595 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
596 of the compiled pattern block, expand the flags field from 16 to 32 bits
597 because it was almost full.
598
599 39. Try madvise first before posix_madvise.
600
601 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
602 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
603 original change locked out all matches of empty strings. This has been
604 changed so that one match of an empty string per line is recognized.
605 Subsequent searches on the same line (for colouring or for --only-matching,
606 for example) do not recognize empty strings.
607
608 41. Applied a user patch to fix a number of spelling mistakes in comments.
609
610 42. Data lines longer than 65536 caused pcretest to crash.
611
612 43. Clarified the data type for length and startoffset arguments for pcre_exec
613 and pcre_dfa_exec in the function-specific man pages, where they were
614 explicitly stated to be in bytes, never having been updated. I also added
615 some clarification to the pcreapi man page.
616
617 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
618 a segmentation fault.
619
620
621 Version 8.32 30-November-2012
622 -----------------------------
623
624 1. Improved JIT compiler optimizations for first character search and single
625 character iterators.
626
627 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
628 Patch by Daniel Richard G.
629
630 3. Single character iterator optimizations in the JIT compiler.
631
632 4. Improved JIT compiler optimizations for character ranges.
633
634 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
635 Reported by Giuseppe D'Angelo.
636
637 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
638 of a line, was being set incorrectly in cases where .* appeared inside
639 atomic brackets at the start of a pattern, or where there was a subsequent
640 *PRUNE or *SKIP.
641
642 7. Improved instruction cache flush for POWER/PowerPC.
643 Patch by Daniel Richard G.
644
645 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
646 grep:
647
648 (a) There is now no limit to the number of patterns to be matched.
649
650 (b) An error is given if a pattern is too long.
651
652 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
653 are now supported.
654
655 (d) --exclude-from and --include-from (multiple use) have been added.
656
657 (e) Exclusions and inclusions now apply to all files and directories, not
658 just to those obtained from scanning a directory recursively.
659
660 (f) Multiple uses of -f and --file-list are now supported.
661
662 (g) In a Windows environment, the default for -d has been changed from
663 "read" (the GNU grep default) to "skip", because otherwise the presence
664 of a directory in the file list provokes an error.
665
666 (h) The documentation has been revised and clarified in places.
667
668 9. Improve the matching speed of capturing brackets.
669
670 10. Changed the meaning of \X so that it now matches a Unicode extended
671 grapheme cluster.
672
673 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
674 out POSIX threads when JIT support is configured.
675
676 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
677
678 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
679 structure could go wrong in environments where size_t is not the same size
680 as int.
681
682 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
683 set.
684
685 15. The EBCDIC support had decayed; later updates to the code had included
686 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
687 general tidy up of EBCDIC-related issues, and the documentation was also
688 not quite right. There is now a test that can be run on ASCII systems to
689 check some of the EBCDIC-related things (but is it not a full test).
690
691 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
692 in a small tidy to the code.
693
694 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
695
696 18. If the --only-matching (-o) option in pcregrep is specified multiple
697 times, each one causes appropriate output. For example, -o1 -o2 outputs the
698 substrings matched by the 1st and 2nd capturing parentheses. A separating
699 string can be specified by --om-separator (default empty).
700
701 19. Improving the first n character searches.
702
703 20. Turn case lists for horizontal and vertical white space into macros so that
704 they are defined only once.
705
706 21. This set of changes together give more compatible Unicode case-folding
707 behaviour for characters that have more than one other case when UCP
708 support is available.
709
710 (a) The Unicode property table now has offsets into a new table of sets of
711 three or more characters that are case-equivalent. The MultiStage2.py
712 script that generates these tables (the pcre_ucd.c file) now scans
713 CaseFolding.txt instead of UnicodeData.txt for character case
714 information.
715
716 (b) The code for adding characters or ranges of characters to a character
717 class has been abstracted into a generalized function that also handles
718 case-independence. In UTF-mode with UCP support, this uses the new data
719 to handle characters with more than one other case.
720
721 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
722 whose other case is greater than 256 are now correctly matched
723 caselessly. Previously, the high codepoint matched the low one, but not
724 vice versa.
725
726 (d) The processing of \h, \H, \v, and \ in character classes now makes use
727 of the new class addition function, using character lists defined as
728 macros alongside the case definitions of 20 above.
729
730 (e) Caseless back references now work with characters that have more than
731 one other case.
732
733 (f) General caseless matching of characters with more than one other case
734 is supported.
735
736 22. Unicode character properties were updated from Unicode 6.2.0
737
738 23. Improved CMake support under Windows. Patch by Daniel Richard G.
739
740 24. Add support for 32-bit character strings, and UTF-32
741
742 25. Major JIT compiler update (code refactoring and bugfixing).
743 Experimental Sparc 32 support is added.
744
745 26. Applied a modified version of Daniel Richard G's patch to create
746 pcre.h.generic and config.h.generic by "make" instead of in the
747 PrepareRelease script.
748
749 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
750 pcre_compile.c when checking for a zero character.
751
752 28. Introducing a native interface for JIT. Through this interface, the compiled
753 machine code can be directly executed. The purpose of this interface is to
754 provide fast pattern matching, so several sanity checks are not performed.
755 However, feature tests are still performed. The new interface provides
756 1.4x speedup compared to the old one.
757
758 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
759 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
760 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
761
762 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
763 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
764 the "old" RFC 2279). Instead, it ended up passing a negative length to
765 pcre_exec().
766
767 31. Add support for GCC's visibility feature to hide internal functions.
768
769 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
770 "unknown -C option" after outputting 0 or 1.
771
772 33. There is now support for generating a code coverage report for the test
773 suite in environments where gcc is the compiler and lcov is installed. This
774 is mainly for the benefit of the developers.
775
776 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
777 unaddressable using valgrind annotations, allowing valgrind to detect
778 invalid memory accesses. This is mainly for the benefit of the developers.
779
780 25. (*UTF) can now be used to start a pattern in any of the three libraries.
781
782 26. Give configure error if --enable-cpp but no C++ compiler found.
783
784
785 Version 8.31 06-July-2012
786 -------------------------
787
788 1. Fixing a wrong JIT test case and some compiler warnings.
789
790 2. Removed a bashism from the RunTest script.
791
792 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
793 to unsigned type, result still unsigned" that was given by an MS compiler
794 on encountering the code "-sizeof(xxx)".
795
796 4. Partial matching support is added to the JIT compiler.
797
798 5. Fixed several bugs concerned with partial matching of items that consist
799 of more than one character:
800
801 (a) /^(..)\1/ did not partially match "aba" because checking references was
802 done on an "all or nothing" basis. This also applied to repeated
803 references.
804
805 (b) \R did not give a hard partial match if \r was found at the end of the
806 subject.
807
808 (c) \X did not give a hard partial match after matching one or more
809 characters at the end of the subject.
810
811 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
812 a partial match for the string "\r".
813
814 (e) When newline was set to CRLF, the metacharacter "." did not recognize
815 a partial match for a CR character at the end of the subject string.
816
817 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
818 running pcretest, the text "(JIT)" added to the output whenever JIT is
819 actually used to run the match.
820
821 7. Individual JIT compile options can be set in pcretest by following -s+[+]
822 or /S+[+] with a digit between 1 and 7.
823
824 8. OP_NOT now supports any UTF character not just single-byte ones.
825
826 9. (*MARK) control verb is now supported by the JIT compiler.
827
828 10. The command "./RunTest list" lists the available tests without actually
829 running any of them. (Because I keep forgetting what they all are.)
830
831 11. Add PCRE_INFO_MAXLOOKBEHIND.
832
833 12. Applied a (slightly modified) user-supplied patch that improves performance
834 when the heap is used for recursion (compiled with --disable-stack-for-
835 recursion). Instead of malloc and free for each heap frame each time a
836 logical recursion happens, frames are retained on a chain and re-used where
837 possible. This sometimes gives as much as 30% improvement.
838
839 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
840 call.
841
842 14. As documented, (*COMMIT) is now confined to within a positive assertion.
843
844 15. It is now possible to link pcretest with libedit as an alternative to
845 libreadline.
846
847 16. (*COMMIT) control verb is now supported by the JIT compiler.
848
849 17. The Unicode data tables have been updated to Unicode 6.1.0.
850
851 18. Added --file-list option to pcregrep.
852
853 19. Added binary file support to pcregrep, including the -a, --binary-files,
854 -I, and --text options.
855
856 20. The madvise function is renamed for posix_madvise for QNX compatibility
857 reasons. Fixed by Giuseppe D'Angelo.
858
859 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
860
861 22. Changed the option for creating symbolic links for 16-bit man pages from
862 -s to -sf so that re-installing does not cause issues.
863
864 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
865
866 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
867 matches in certain environments (the workspace was not being correctly
868 retained). Also added to pcre_dfa_exec() a simple plausibility check on
869 some of the workspace data at the beginning of a restart.
870
871 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
872 was not doing so when it should - probably a typo introduced by SVN 528
873 (change 8.10/14).
874
875 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
876 \w+ when the character tables indicated that \x{c4} was a word character.
877 There were several related cases, all because the tests for doing a table
878 lookup were testing for characters less than 127 instead of 255.
879
880 27. If a pattern contains capturing parentheses that are not used in a match,
881 their slots in the ovector are set to -1. For those that are higher than
882 any matched groups, this happens at the end of processing. In the case when
883 there were back references that the ovector was too small to contain
884 (causing temporary malloc'd memory to be used during matching), and the
885 highest capturing number was not used, memory off the end of the ovector
886 was incorrectly being set to -1. (It was using the size of the temporary
887 memory instead of the true size.)
888
889 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
890 ovector size, it uses memory at the end of the block that it has got.
891
892 29. Check for an overlong MARK name and give an error at compile time. The
893 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
894
895 30. JIT compiler update.
896
897 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
898 Rill for the patch.
899
900 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
901
902 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
903
904 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
905 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
906 was enabled.
907
908 35. Improve JIT code generation for greedy plus quantifier.
909
910 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
911 1 to "aa" instead of to an empty string. The bug affected repeated groups
912 that could potentially match an empty string.
913
914 37. Optimizing single character iterators in JIT.
915
916 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
917 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
918 codepoints that are too big for the mode are faulted, and in a UTF mode,
919 disallowed codepoints are also faulted.
920
921 39. If PCRE was compiled with UTF support, in three places in the DFA
922 matcher there was code that should only have been obeyed in UTF mode, but
923 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
924 processing when bytes with values greater than 127 were present. In 16-bit
925 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
926 both cases the values are those that cannot be the first data item in a UTF
927 character. The three items that might have provoked this were recursions,
928 possessively repeated groups, and atomic groups.
929
930 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
931 and pcregrep, because some OS require shared objects to be explicitly
932 passed to ld, causing the link step to fail if they are not.
933
934 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
935 mode, patterns that started with \h* or \R* might be incorrectly matched.
936
937
938 Version 8.30 04-February-2012
939 -----------------------------
940
941 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
942 name is defined in ctype.h.
943
944 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
945 only in quite long subpatterns.
946
947 3. Removed the function pcre_info(), which has been obsolete and deprecated
948 since it was replaced by pcre_fullinfo() in February 2000.
949
950 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
951 match a (*MARK), and the match failed at the start of the subject, a
952 reference to memory before the start of the subject could occur. This bug
953 was introduced by fix 17 of release 8.21.
954
955 5. A reference to an unset group with zero minimum repetition was giving
956 totally wrong answers (in non-JavaScript-compatibility mode). For example,
957 /(another)?(\1?)test/ matched against "hello world test". This bug was
958 introduced in release 8.13.
959
960 6. Add support for 16-bit character strings (a large amount of work involving
961 many changes and refactorings).
962
963 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
964 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
965 from a file.
966
967 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
968 rounding is not applied in this particular case).
969
970 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
971 if they appear, or are escaped, in patterns.
972
973 10. Get rid of a number of -Wunused-but-set-variable warnings.
974
975 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
976 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
977 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
978 also returns the mark "x". This bug applied to capturing parentheses,
979 non-capturing parentheses, and atomic parentheses. It also applied to some
980 assertions.
981
982 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
983 information out of configure.ac instead of relying on pcre.h.generic, which
984 is not stored in the repository.
985
986 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
987 -lreadline.
988
989 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
990
991 15. Applied Graycode's patch to put the top-level frame on the stack rather
992 than the heap when not using the stack for recursion. This gives a
993 performance improvement in many cases when recursion is not deep.
994
995 16. Experimental code added to "pcretest -C" to output the stack frame size.
996
997
998 Version 8.21 12-Dec-2011
999 ------------------------
1000
1001 1. Updating the JIT compiler.
1002
1003 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1004 are added as well.
1005
1006 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1007 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1008 calling _pcre_jit_exec. Some extra comments are added.
1009
1010 4. (*MARK) settings inside atomic groups that do not contain any capturing
1011 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1012 was introduced by change 18 for 8.20.
1013
1014 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1015 ECMA-262 standard.
1016
1017 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1018 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1019 This bug was probably introduced by change 9 of 8.13.
1020
1021 7. While fixing 6 above, I noticed that a number of other items were being
1022 incorrectly rejected as "not fixed length". This arose partly because newer
1023 opcodes had not been added to the fixed-length checking code. I have (a)
1024 corrected the bug and added tests for these items, and (b) arranged for an
1025 error to occur if an unknown opcode is encountered while checking for fixed
1026 length instead of just assuming "not fixed length". The items that were
1027 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1028 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1029 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1030
1031 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1032 being incorrectly compiled and would have given unpredicatble results.
1033
1034 9. A possessively repeated subpattern with minimum repeat count greater than
1035 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1036 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1037 the first (A) could occur when it should not.
1038
1039 10. Add a cast and remove a redundant test from the code.
1040
1041 11. JIT should use pcre_malloc/pcre_free for allocation.
1042
1043 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1044 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1045 is anything other than /usr, -L is still shown).
1046
1047 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1048
1049 14. Perl does not support \N without a following name in a [] class; PCRE now
1050 also gives an error.
1051
1052 15. If a forward reference was repeated with an upper limit of around 2000,
1053 it caused the error "internal error: overran compiling workspace". The
1054 maximum number of forward references (including repeats) was limited by the
1055 internal workspace, and dependent on the LINK_SIZE. The code has been
1056 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1057 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1058 of around 200,000 forward references. While doing this, I also speeded up
1059 the filling in of repeated forward references.
1060
1061 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1062 incorrectly expecting the subject to contain another "a" after the start.
1063
1064 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1065 in the match, the SKIP should be ignored. This was not happening; instead
1066 the SKIP was being treated as NOMATCH. For patterns such as
1067 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1068 tested.
1069
1070 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1071 now much more compatible with Perl, in particular in cases where the result
1072 is a non-match for a non-anchored pattern. For example, if
1073 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1074 "m", where previously it did not return a name. A side effect of this
1075 change is that for partial matches, the last encountered mark name is
1076 returned, as for non matches. A number of tests that were previously not
1077 Perl-compatible have been moved into the Perl-compatible test files. The
1078 refactoring has had the pleasing side effect of removing one argument from
1079 the match() function, thus reducing its stack requirements.
1080
1081 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1082 subsequent uses of /S (without +) incorrectly behaved like /S+.
1083
1084 21. Retrieve executable code size support for the JIT compiler and fixing
1085 some warnings.
1086
1087 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1088 not work when the shorter character appeared right at the end of the
1089 subject string.
1090
1091 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1092 systems.
1093
1094 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1095 output it when the /M option is used in pcretest.
1096
1097 25. The CheckMan script was not being included in the distribution. Also, added
1098 an explicit "perl" to run Perl scripts from the PrepareRelease script
1099 because this is reportedly needed in Windows.
1100
1101 26. If study data was being save in a file and studying had not found a set of
1102 "starts with" bytes for the pattern, the data written to the file (though
1103 never used) was taken from uninitialized memory and so caused valgrind to
1104 complain.
1105
1106 27. Updated RunTest.bat as provided by Sheri Pierce.
1107
1108 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1109
1110 29. Computation of memory usage for the table of capturing group names was
1111 giving an unnecessarily large value.
1112
1113
1114 Version 8.20 21-Oct-2011
1115 ------------------------
1116
1117 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1118 a POSIX class. After further experiments with Perl, which convinced me that
1119 Perl has bugs and confusions, a closing square bracket is no longer allowed
1120 in a POSIX name. This bug also affected patterns with classes that started
1121 with full stops.
1122
1123 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1124 captured substring, but while checking the failing first alternative,
1125 substring 1 is temporarily captured. If the output vector supplied to
1126 pcre_exec() was not big enough for this capture, the yield of the function
1127 was still zero ("insufficient space for captured substrings"). This cannot
1128 be totally fixed without adding another stack variable, which seems a lot
1129 of expense for a edge case. However, I have improved the situation in cases
1130 such as /(a)(b)x|abc/ matched against "abc", where the return code
1131 indicates that fewer than the maximum number of slots in the ovector have
1132 been set.
1133
1134 3. Related to (2) above: when there are more back references in a pattern than
1135 slots in the output vector, pcre_exec() uses temporary memory during
1136 matching, and copies in the captures as far as possible afterwards. It was
1137 using the entire output vector, but this conflicts with the specification
1138 that only 2/3 is used for passing back captured substrings. Now it uses
1139 only the first 2/3, for compatibility. This is, of course, another edge
1140 case.
1141
1142 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1143 main code base, and can be used by building with --enable-jit. When this is
1144 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1145 runtime --no-jit option is given.
1146
1147 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1148 ovector, the return from the function was zero, implying that there were
1149 other matches that did not fit. The correct "exactly full" value is now
1150 returned.
1151
1152 6. If a subpattern that was called recursively or as a subroutine contained
1153 (*PRUNE) or any other control that caused it to give a non-standard return,
1154 invalid errors such as "Error -26 (nested recursion at the same subject
1155 position)" or even infinite loops could occur.
1156
1157 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1158 computing the minimum length on reaching *ACCEPT, and so ended up with the
1159 wrong value of 1 rather than 0. Further investigation indicates that
1160 computing a minimum subject length in the presence of *ACCEPT is difficult
1161 (think back references, subroutine calls), and so I have changed the code
1162 so that no minimum is registered for a pattern that contains *ACCEPT.
1163
1164 8. If (*THEN) was present in the first (true) branch of a conditional group,
1165 it was not handled as intended. [But see 16 below.]
1166
1167 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1168 Sheri Pierce.
1169
1170 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1171 the first byte in a match must be "a".
1172
1173 11. Change 17 for 8.13 increased the recursion depth for patterns like
1174 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1175 pattern contains any instances of (*THEN). If it does not, the old
1176 optimizations are restored. It would be nice to do this on a per-group
1177 basis, but at the moment that is not feasible.
1178
1179 12. In some environments, the output of pcretest -C is CRLF terminated. This
1180 broke RunTest's code that checks for the link size. A single white space
1181 character after the value is now allowed for.
1182
1183 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1184 For "fr", it uses the Windows-specific input and output files.
1185
1186 14. If (*THEN) appeared in a group that was called recursively or as a
1187 subroutine, it did not work as intended. [But see next item.]
1188
1189 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1190 pattern fragments (but not containing any | characters). If A and B are
1191 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1192 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1193 D. In other words, Perl considers parentheses that do not contain any |
1194 characters to be part of a surrounding alternative, whereas PCRE was
1195 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1196 differently. PCRE now behaves in the same way as Perl, except in the case
1197 of subroutine/recursion calls such as (?1) which have in any case always
1198 been different (but PCRE had them first :-).
1199
1200 16. Related to 15 above: Perl does not treat the | in a conditional group as
1201 creating alternatives. Such a group is treated in the same way as an
1202 ordinary group without any | characters when processing (*THEN). PCRE has
1203 been changed to match Perl's behaviour.
1204
1205 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1206 RunGrepTest script failed.
1207
1208 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1209 inevitable for groups that contain captures, but it can lead to a lot of
1210 stack use in large patterns. The old behaviour has been restored for atomic
1211 groups that do not contain any capturing parentheses.
1212
1213 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1214 suppress the check for a minimum subject length at run time. (If it was
1215 given to pcre_exec() or pcre_dfa_exec() it did work.)
1216
1217 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1218 fail to work when decoding hex characters in data strings in EBCDIC
1219 environments.
1220
1221 21. It appears that in at least one Mac OS environment, the isxdigit() function
1222 is implemented as a macro that evaluates to its argument more than once,
1223 contravening the C 90 Standard (I haven't checked a later standard). There
1224 was an instance in pcretest which caused it to go wrong when processing
1225 \x{...} escapes in subject strings. The has been rewritten to avoid using
1226 things like p++ in the argument of isxdigit().
1227
1228
1229 Version 8.13 16-Aug-2011
1230 ------------------------
1231
1232 1. The Unicode data tables have been updated to Unicode 6.0.0.
1233
1234 2. Two minor typos in pcre_internal.h have been fixed.
1235
1236 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1237 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1238 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1239
1240 4. There were a number of related bugs in the code for matching backrefences
1241 caselessly in UTF-8 mode when codes for the characters concerned were
1242 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1243 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1244 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1245 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1246 2-byte code at the end of the subject (it thought there wasn't enough data
1247 left).
1248
1249 5. Comprehensive information about what went wrong is now returned by
1250 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1251 as the output vector has at least 2 elements. The offset of the start of
1252 the failing character and a reason code are placed in the vector.
1253
1254 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1255 now returned is for the first byte of the failing character, instead of the
1256 last byte inspected. This is an incompatible change, but I hope it is small
1257 enough not to be a problem. It makes the returned offset consistent with
1258 pcre_exec() and pcre_dfa_exec().
1259
1260 7. pcretest now gives a text phrase as well as the error number when
1261 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1262 failure, the offset and reason code are output.
1263
1264 8. When \R was used with a maximizing quantifier it failed to skip backwards
1265 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1266 back over a single character (\n). This seems wrong (because it treated the
1267 two characters as a single entity when going forwards), conflicts with the
1268 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1269 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1270 has been changed.
1271
1272 9. Some internal refactoring has changed the processing so that the handling
1273 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1274 time (the PCRE_DOTALL option was changed this way some time ago: version
1275 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1276 which was always a bit of a fudge. It also means that there is one less
1277 argument for the match() function, which reduces its stack requirements
1278 slightly. This change also fixes an incompatibility with Perl: the pattern
1279 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1280
1281 10. More internal refactoring has drastically reduced the number of recursive
1282 calls to match() for possessively repeated groups such as (abc)++ when
1283 using pcre_exec().
1284
1285 11. While implementing 10, a number of bugs in the handling of groups were
1286 discovered and fixed:
1287
1288 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1289 (a|)*(?1) gave a compile-time internal error.
1290 ((a|)+)+ did not notice that the outer group could match an empty string.
1291 (^a|^)+ was not marked as anchored.
1292 (.*a|.*)+ was not marked as matching at start or after a newline.
1293
1294 12. Yet more internal refactoring has removed another argument from the match()
1295 function. Special calls to this function are now indicated by setting a
1296 value in a variable in the "match data" data block.
1297
1298 13. Be more explicit in pcre_study() instead of relying on "default" for
1299 opcodes that mean there is no starting character; this means that when new
1300 ones are added and accidentally left out of pcre_study(), testing should
1301 pick them up.
1302
1303 14. The -s option of pcretest has been documented for ages as being an old
1304 synonym of -m (show memory usage). I have changed it to mean "force study
1305 for every regex", that is, assume /S for every regex. This is similar to -i
1306 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1307 using it. It makes it easier to run collections of tests with and without
1308 study enabled, and thereby test pcre_study() more easily. All the standard
1309 tests are now run with and without -s (but some patterns can be marked as
1310 "never study" - see 20 below).
1311
1312 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1313 restoration of the capturing data to the outer values was not happening
1314 correctly.
1315
1316 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1317 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1318 pattern had matched an empty string, and so incorrectly returned a no
1319 match.
1320
1321 17. There was optimizing code for the last branch of non-capturing parentheses,
1322 and also for the obeyed branch of a conditional subexpression, which used
1323 tail recursion to cut down on stack usage. Unfortunately, now that there is
1324 the possibility of (*THEN) occurring in these branches, tail recursion is
1325 no longer possible because the return has to be checked for (*THEN). These
1326 two optimizations have therefore been removed. [But see 8.20/11 above.]
1327
1328 18. If a pattern containing \R was studied, it was assumed that \R always
1329 matched two bytes, thus causing the minimum subject length to be
1330 incorrectly computed because \R can also match just one byte.
1331
1332 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1333 was incorrectly computed.
1334
1335 20. If /S is present twice on a test pattern in pcretest input, it now
1336 *disables* studying, thereby overriding the use of -s on the command line
1337 (see 14 above). This is necessary for one or two tests to keep the output
1338 identical in both cases.
1339
1340 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1341 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1342
1343 22. When an atomic group that contained a capturing parenthesis was
1344 successfully matched, but the branch in which it appeared failed, the
1345 capturing was not being forgotten if a higher numbered group was later
1346 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1347 group 1 to "a", when in fact it should be unset. This applied to multi-
1348 branched capturing and non-capturing groups, repeated or not, and also to
1349 positive assertions (capturing in negative assertions does not happen
1350 in PCRE) and also to nested atomic groups.
1351
1352 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1353 subject after a captured substring, to make it easier to tell which of a
1354 number of identical substrings has been captured.
1355
1356 24. The way atomic groups are processed by pcre_exec() has been changed so that
1357 if they are repeated, backtracking one repetition now resets captured
1358 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1359 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1360 "aaa". Previously, it would have been "a". As part of this code
1361 refactoring, the way recursive calls are handled has also been changed.
1362
1363 25. If an assertion condition captured any substrings, they were not passed
1364 back unless some other capturing happened later. For example, if
1365 (?(?=(a))a) was matched against "a", no capturing was returned.
1366
1367 26. When studying a pattern that contained subroutine calls or assertions,
1368 the code for finding the minimum length of a possible match was handling
1369 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1370 group 1 called group 2 while simultaneously a separate group 2 called group
1371 1). A stack overflow occurred in this case. I have fixed this by limiting
1372 the recursion depth to 10.
1373
1374 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1375 Fortmann. This supports explicit test numbers on the command line, and has
1376 argument validation and error reporting.
1377
1378 28. An instance of \X with an unlimited repeat could fail if at any point the
1379 first character it looked at was a mark character.
1380
1381 29. Some minor code refactoring concerning Unicode properties and scripts
1382 should reduce the stack requirement of match() slightly.
1383
1384 30. Added the '=' option to pcretest to check the setting of unused capturing
1385 slots at the end of the pattern, which are documented as being -1, but are
1386 not included in the return count.
1387
1388 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1389 compiled something random. Now it gives a compile-time error (as does
1390 Perl).
1391
1392 32. A *MARK encountered during the processing of a positive assertion is now
1393 recorded and passed back (compatible with Perl).
1394
1395 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1396 had alternative anchored branches, the search for a second match in a line
1397 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1398 matched the line "0102" twice. The same bug affected patterns that started
1399 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1400 twice.
1401
1402 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1403 does, and because of capturing effects, quantifying parenthesized
1404 assertions may at times be useful. Quantifiers are now allowed for
1405 parenthesized assertions.
1406
1407 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1408
1409 36. \g was being checked for fancy things in a character class, when it should
1410 just be a literal "g".
1411
1412 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1413 appearance of a nested POSIX class supersedes an apparent external class.
1414 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1415 unescaped square brackets may also appear as part of class names. For
1416 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1417 more like Perl. (But see 8.20/1 above.)
1418
1419 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1420 was because it thought it was \N{name}, which is not supported).
1421
1422 39. Add minix to OS list not supporting the -S option in pcretest.
1423
1424 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1425 cannot analyze patterns in sufficient detail to catch mutual recursions
1426 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1427 subgroup is called recursively as a subpattern for a second time at the
1428 same position in the subject string. In previous releases this might have
1429 been caught by the recursion limit, or it might have run out of stack.
1430
1431 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1432 happen only once. PCRE was, however incorrectly giving a compile time error
1433 "recursive call could loop indefinitely" because it cannot analyze the
1434 pattern in sufficient detail. The compile time test no longer happens when
1435 PCRE is compiling a conditional subpattern, but actual runaway loops are
1436 now caught at runtime (see 40 above).
1437
1438 42. It seems that Perl allows any characters other than a closing parenthesis
1439 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1440 has been changed to be the same.
1441
1442 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1443 as not to get warnings when autogen.sh is called. Also changed
1444 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1445
1446 44. To help people who use pcregrep to scan files containing exceedingly long
1447 lines, the following changes have been made:
1448
1449 (a) The default value of the buffer size parameter has been increased from
1450 8K to 20K. (The actual buffer used is three times this size.)
1451
1452 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1453 PCRE is built.
1454
1455 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1456 to be set at run time.
1457
1458 (d) Numerical values in pcregrep options can be followed by K or M, for
1459 example --buffer-size=50K.
1460
1461 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1462 given and the return code is set to 2.
1463
1464 45. Add a pointer to the latest mark to the callout data block.
1465
1466 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1467 partial match of an empty string instead of no match. This was specific to
1468 the use of ".".
1469
1470 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1471 complete match instead of a partial match. This bug was dependent on both
1472 the PCRE_UTF8 and PCRE_DOTALL options being set.
1473
1474 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1475 starting byte set, because \b was not being ignored.
1476
1477
1478 Version 8.12 15-Jan-2011
1479 ------------------------
1480
1481 1. Fixed some typos in the markup of the man pages, and wrote a script that
1482 checks for such things as part of the documentation building process.
1483
1484 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1485 --match-limit and --recursion-limit options (added for 8.11). In
1486 particular, this made one of the standard tests fail. (The integer value
1487 went into the wrong half of a long int.)
1488
1489 3. If the --colour option was given to pcregrep with -v (invert match), it
1490 did strange things, either producing crazy output, or crashing. It should,
1491 of course, ignore a request for colour when reporting lines that do not
1492 match.
1493
1494 4. Another pcregrep bug caused similar problems if --colour was specified with
1495 -M (multiline) and the pattern match finished with a line ending.
1496
1497 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1498 matched in multiline mode, the following line was shown as part of the
1499 match. This seems wrong, so I have changed it.
1500
1501 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1502 the check for further matches in the same line (so they could be coloured)
1503 to overrun the end of the current line. If another match was found, it was
1504 incorrectly shown (and then shown again when found in the next line).
1505
1506 7. If pcregrep was compiled under Windows, there was a reference to the
1507 function pcregrep_exit() before it was defined. I am assuming this was
1508 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1509 reported by a user. I've moved the definition above the reference.
1510
1511
1512 Version 8.11 10-Dec-2010
1513 ------------------------
1514
1515 1. (*THEN) was not working properly if there were untried alternatives prior
1516 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1517 backtracked to try for "b" instead of moving to the next alternative branch
1518 at the same level (in this case, to look for "c"). The Perl documentation
1519 is clear that when (*THEN) is backtracked onto, it goes to the "next
1520 alternative in the innermost enclosing group".
1521
1522 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1523 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1524 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1525 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1526 (*THEN).
1527
1528 3. If \s appeared in a character class, it removed the VT character from
1529 the class, even if it had been included by some previous item, for example
1530 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1531 of \s, but is part of the POSIX "space" class.)
1532
1533 4. A partial match never returns an empty string (because you can always
1534 match an empty string at the end of the subject); however the checking for
1535 an empty string was starting at the "start of match" point. This has been
1536 changed to the "earliest inspected character" point, because the returned
1537 data for a partial match starts at this character. This means that, for
1538 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1539 (previously it gave "no match").
1540
1541 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1542 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1543 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1544 has an implication that the given string is incomplete (because a partial
1545 match is preferred over a full match). For this reason, these items now
1546 give a partial match in this situation. [Aside: previously, the one case
1547 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1548 match rather than a full match, which was wrong by the old rules, but is
1549 now correct.]
1550
1551 6. There was a bug in the handling of #-introduced comments, recognized when
1552 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1553 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1554 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1555 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1556 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1557 places in pcre_compile().
1558
1559 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1560 comments when looking ahead for named forward references to subpatterns,
1561 the only newline sequence it recognized was NL. It now handles newlines
1562 according to the set newline convention.
1563
1564 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1565 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1566 cater for a lack of strerror(). These oversights have been fixed.
1567
1568 9. Added --match-limit and --recursion-limit to pcregrep.
1569
1570 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1571
1572 11. When the -o option was used, pcregrep was setting a return code of 1, even
1573 when matches were found, and --line-buffered was not being honoured.
1574
1575 12. Added an optional parentheses number to the -o and --only-matching options
1576 of pcregrep.
1577
1578 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1579 can match an empty string. The code to do it in pcretest and pcredemo
1580 needed fixing:
1581
1582 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1583 only one byte after an empty string match just before CRLF (this case
1584 just got forgotten; "any" and "anycrlf" were OK).
1585
1586 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1587 mode when an empty string match preceded an ASCII character followed by
1588 a non-ASCII character. (The code for advancing by one character rather
1589 than one byte was nonsense.)
1590
1591 (c) The pcredemo.c sample program did not have any code at all to handle
1592 the cases when CRLF is a valid newline sequence.
1593
1594 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1595 as a starting offset was within the subject string. There is now a new
1596 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1597 negative or greater than the length of the string. In order to test this,
1598 pcretest is extended to allow the setting of negative starting offsets.
1599
1600 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1601 starting offset points to the beginning of a UTF-8 character was
1602 unnecessarily clumsy. I tidied it up.
1603
1604 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1605 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1606
1607 17. Nobody had reported that the --include_dir option, which was added in
1608 release 7.7 should have been called --include-dir (hyphen, not underscore)
1609 for compatibility with GNU grep. I have changed it to --include-dir, but
1610 left --include_dir as an undocumented synonym, and the same for
1611 --exclude-dir, though that is not available in GNU grep, at least as of
1612 release 2.5.4.
1613
1614 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1615 characters from a string of bytes) have been redefined so as not to use
1616 loops, in order to improve performance in some environments. At the same
1617 time, I abstracted some of the common code into auxiliary macros to save
1618 repetition (this should not affect the compiled code).
1619
1620 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1621 compile-time error is now given if \c is not followed by an ASCII
1622 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1623 different, and any byte value is allowed.)
1624
1625 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1626 START_OPTIMIZE option, which is now allowed at compile time - but just
1627 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1628 to pcregrep and other applications that have no direct access to PCRE
1629 options. The new /Y option in pcretest sets this option when calling
1630 pcre_compile().
1631
1632 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1633 back references. Groups containing recursive back references were forced to
1634 be atomic by that change, but in the case of named groups, the amount of
1635 memory required was incorrectly computed, leading to "Failed: internal
1636 error: code overflow". This has been fixed.
1637
1638 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1639 pcretest.c, to avoid build problems in some Borland environments.
1640
1641
1642 Version 8.10 25-Jun-2010
1643 ------------------------
1644
1645 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1646 THEN.
1647
1648 2. (*ACCEPT) was not working when inside an atomic group.
1649
1650 3. Inside a character class, \B is treated as a literal by default, but
1651 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1652 causes the error). The code is unchanged, but I tidied the documentation.
1653
1654 4. Inside a character class, PCRE always treated \R and \X as literals,
1655 whereas Perl faults them if its -w option is set. I have changed PCRE so
1656 that it faults them when PCRE_EXTRA is set.
1657
1658 5. Added support for \N, which always matches any character other than
1659 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1660
1661 6. When compiling pcregrep with newer versions of gcc which may have
1662 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1663 declared with attribute warn_unused_result" were given. Just casting the
1664 result to (void) does not stop the warnings; a more elaborate fudge is
1665 needed. I've used a macro to implement this.
1666
1667 7. Minor change to pcretest.c to avoid a compiler warning.
1668
1669 8. Added four artifical Unicode properties to help with an option to make
1670 \s etc use properties (see next item). The new properties are: Xan
1671 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1672
1673 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1674 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1675 this option. Modified pcretest to add /W to test this facility. Added
1676 REG_UCP to make it available via the POSIX interface.
1677
1678 10. Added --line-buffered to pcregrep.
1679
1680 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1681 studied, and the match started with a letter with a code point greater than
1682 127 whose first byte was different to the first byte of the other case of
1683 the letter, the other case of this starting letter was not recognized
1684 (#976).
1685
1686 12. If a pattern that was studied started with a repeated Unicode property
1687 test, for example, \p{Nd}+, there was the theoretical possibility of
1688 setting up an incorrect bitmap of starting bytes, but fortunately it could
1689 not have actually happened in practice until change 8 above was made (it
1690 added property types that matched character-matching opcodes).
1691
1692 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1693 possible starting bytes for non-anchored patterns.
1694
1695 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1696 \R, and also a number of cases that involve Unicode properties, both
1697 explicit and implicit when PCRE_UCP is set.
1698
1699 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1700 input, it could crash or give wrong results if characters with values
1701 greater than 0xc0 were present in the subject string. (Detail: it assumed
1702 UTF-8 input when processing these items.)
1703
1704 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1705 size_t is 64-bit (#991).
1706
1707 17. Added a check for running out of memory when PCRE is compiled with
1708 --disable-stack-for-recursion (#990).
1709
1710 18. If the last data line in a file for pcretest does not have a newline on
1711 the end, a newline was missing in the output.
1712
1713 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1714 less than 128) in its various bitmaps. However, there is a facility for
1715 generating tables according to the current locale when PCRE is compiled. It
1716 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1717 characters, are recognized by isspace() and therefore were getting set in
1718 these tables, and indeed these tables seem to approximate to ISO 8859. This
1719 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1720 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1721 which of course cannot start UTF-8 characters. I have changed the code so
1722 that only real ASCII characters (less than 128) and the correct starting
1723 bytes for UTF-8 encodings are set for characters greater than 127 when in
1724 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1725 altogether.)
1726
1727 20. Added the /T option to pcretest so as to be able to run tests with non-
1728 standard character tables, thus making it possible to include the tests
1729 used for 19 above in the standard set of tests.
1730
1731 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1732 reference to a subpattern the other side of a comment that contains an
1733 opening parenthesis caused either an internal compiling error, or a
1734 reference to the wrong subpattern.
1735
1736
1737 Version 8.02 19-Mar-2010
1738 ------------------------
1739
1740 1. The Unicode data tables have been updated to Unicode 5.2.0.
1741
1742 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1743 configured.
1744
1745 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1746 original author of that file, following a query about its status.
1747
1748 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1749 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1750
1751 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1752 quantifier applied to a forward-referencing subroutine call, could compile
1753 incorrect code or give the error "internal error: previously-checked
1754 referenced subpattern not found".
1755
1756 6. Both MS Visual Studio and Symbian OS have problems with initializing
1757 variables to point to external functions. For these systems, therefore,
1758 pcre_malloc etc. are now initialized to local functions that call the
1759 relevant global functions.
1760
1761 7. There were two entries missing in the vectors called coptable and poptable
1762 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1763 I've fixed the data, and added a kludgy way of testing at compile time that
1764 the lengths are correct (equal to the number of opcodes).
1765
1766 8. Following on from 7, I added a similar kludge to check the length of the
1767 eint vector in pcreposix.c.
1768
1769 9. Error texts for pcre_compile() are held as one long string to avoid too
1770 much relocation at load time. To find a text, the string is searched,
1771 counting zeros. There was no check for running off the end of the string,
1772 which could happen if a new error number was added without updating the
1773 string.
1774
1775 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1776
1777 11. \K was not working if it appeared in an atomic group or in a group that
1778 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1779 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1780 the assertion is positive, but not if it is negative.
1781
1782 12. Change 11 fortuitously reduced the size of the stack frame used in the
1783 "match()" function of pcre_exec.c by one pointer. Forthcoming
1784 implementation of support for (*MARK) will need an extra pointer on the
1785 stack; I have reserved it now, so that the stack frame size does not
1786 decrease.
1787
1788 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1789 item in branch that calls a recursion is a subroutine call - as in the
1790 second branch in the above example - was incorrectly given the compile-
1791 time error "recursive call could loop indefinitely" because pcre_compile()
1792 was not correctly checking the subroutine for matching a non-empty string.
1793
1794 14. The checks for overrunning compiling workspace could trigger after an
1795 overrun had occurred. This is a "should never occur" error, but it can be
1796 triggered by pathological patterns such as hundreds of nested parentheses.
1797 The checks now trigger 100 bytes before the end of the workspace.
1798
1799 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1800
1801
1802 Version 8.01 19-Jan-2010
1803 ------------------------
1804
1805 1. If a pattern contained a conditional subpattern with only one branch (in
1806 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1807 computed the wrong minimum data length (which is of course zero for such
1808 subpatterns). This could cause incorrect "no match" results.
1809
1810 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1811 the pattern is reset in the first branch, pcre_compile() failed with
1812 "internal error: code overflow at offset...". This happened only when
1813 the reset was to the original external option setting. (An optimization
1814 abstracts leading options settings into an external setting, which was the
1815 cause of this.)
1816
1817 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1818 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1819 assertion pattern did not match (meaning that the assertion was true), it
1820 was incorrectly treated as false if the SKIP had been reached during the
1821 matching. This also applied to assertions used as conditions.
1822
1823 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1824 assertion subpattern, including such a pattern used as a condition,
1825 unpredictable results occurred, instead of the error return
1826 PCRE_ERROR_DFA_UITEM.
1827
1828 5. The C++ GlobalReplace function was not working like Perl for the special
1829 situation when an empty string is matched. It now does the fancy magic
1830 stuff that is necessary.
1831
1832 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1833 removed. (These were left over from very, very early versions of PCRE.)
1834
1835 7. Some cosmetic changes to the code to make life easier when compiling it
1836 as part of something else:
1837
1838 (a) Change DEBUG to PCRE_DEBUG.
1839
1840 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1841 called "current" as "current_branch", to prevent a collision with the
1842 Linux macro when compiled as a kernel module.
1843
1844 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1845 prevent a collision with the Linux macro when compiled as a kernel
1846 module.
1847
1848 8. In pcre_compile() there are some checks for integer overflows that used to
1849 cast potentially large values to (double). This has been changed to that
1850 when building, a check for int64_t is made, and if it is found, it is used
1851 instead, thus avoiding the use of floating point arithmetic. (There is no
1852 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1853 double.
1854
1855 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1856 2005 (difference between two addresses compared to an unsigned value).
1857
1858 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1859 custom one, because of the following reported problem in Windows:
1860
1861 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1862 under Win32.
1863 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1864 therefore missing the function definition.
1865 - The compiler thus generates a "C" signature for the test function.
1866 - The linker fails to find the "C" function.
1867 - PCRE fails to configure if asked to do so against libbz2.
1868
1869 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1870 messages were output:
1871
1872 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1873 rerunning libtoolize, to keep the correct libtool macros in-tree.
1874 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1875
1876 I have done both of these things.
1877
1878 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1879 most of the time, it *can* run out if it is given a pattern that contains a
1880 runaway infinite recursion. I updated the discussion in the pcrestack man
1881 page.
1882
1883 13. Now that we have gone to the x.xx style of version numbers, the minor
1884 version may start with zero. Using 08 or 09 is a bad idea because users
1885 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1886 interpreted as invalid octal numbers. I've updated the previous comment in
1887 configure.ac, and also added a check that gives an error if 08 or 09 are
1888 used.
1889
1890 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1891 causing partial matching to fail when the end of the subject matched \W
1892 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1893
1894 15. There were some discrepancies between the declarations in pcre_internal.h
1895 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1896 their definitions. The declarations used "const uschar *" and the
1897 definitions used USPTR. Even though USPTR is normally defined as "const
1898 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1899 reported that: "This difference in casting confuses some C++ compilers, for
1900 example, SunCC recognizes above declarations as different functions and
1901 generates broken code for hbpcre." I have changed the declarations to use
1902 USPTR.
1903
1904 16. GNU libtool is named differently on some systems. The autogen.sh script now
1905 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1906 (FreeBSD).
1907
1908 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1909 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1910 comment: "Figure out how to create a longlong from a string: strtoll and
1911 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1912 instance, but it only takes 2 args instead of 3!"
1913
1914 18. A subtle bug concerned with back references has been fixed by a change of
1915 specification, with a corresponding code fix. A pattern such as
1916 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1917 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1918 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1919 same bug. Such groups have to be quantified to be useful, or contained
1920 inside another quantified group. (If there's no repetition, the reference
1921 can never match.) The problem arises because, having left the group and
1922 moved on to the rest of the pattern, a later failure that backtracks into
1923 the group uses the captured value from the final iteration of the group
1924 rather than the correct earlier one. I have fixed this in PCRE by forcing
1925 any group that contains a reference to itself to be an atomic group; that
1926 is, there cannot be any backtracking into it once it has completed. This is
1927 similar to recursive and subroutine calls.
1928
1929
1930 Version 8.00 19-Oct-09
1931 ----------------------
1932
1933 1. The table for translating pcre_compile() error codes into POSIX error codes
1934 was out-of-date, and there was no check on the pcre_compile() error code
1935 being within the table. This could lead to an OK return being given in
1936 error.
1937
1938 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1939 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1940 in a Windows environment.
1941
1942 3. The pcregrep --count option prints the count for each file even when it is
1943 zero, as does GNU grep. However, pcregrep was also printing all files when
1944 --files-with-matches was added. Now, when both options are given, it prints
1945 counts only for those files that have at least one match. (GNU grep just
1946 prints the file name in this circumstance, but including the count seems
1947 more useful - otherwise, why use --count?) Also ensured that the
1948 combination -clh just lists non-zero counts, with no names.
1949
1950 4. The long form of the pcregrep -F option was incorrectly implemented as
1951 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1952 but it seems right to fix it, and I didn't think it was worth preserving
1953 the old behaviour.
1954
1955 5. The command line items --regex=pattern and --regexp=pattern were not
1956 recognized by pcregrep, which required --regex pattern or --regexp pattern
1957 (with a space rather than an '='). The man page documented the '=' forms,
1958 which are compatible with GNU grep; these now work.
1959
1960 6. No libpcreposix.pc file was created for pkg-config; there was just
1961 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1962
1963 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1964 when UCP support is not needed, by modifying the Python script that
1965 generates it from Unicode data files. This should not matter if the module
1966 is correctly used as a library, but I received one complaint about 50K of
1967 unwanted data. My guess is that the person linked everything into his
1968 program rather than using a library. Anyway, it does no harm.
1969
1970 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1971 was a minimum greater than 1 for a wide character in a possessive
1972 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1973 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1974 character. Chaos in the form of incorrect output or a compiling loop could
1975 result.
1976
1977 9. The restrictions on what a pattern can contain when partial matching is
1978 requested for pcre_exec() have been removed. All patterns can now be
1979 partially matched by this function. In addition, if there are at least two
1980 slots in the offset vector, the offset of the earliest inspected character
1981 for the match and the offset of the end of the subject are set in them when
1982 PCRE_ERROR_PARTIAL is returned.
1983
1984 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1985 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1986 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1987 and may be more useful for multi-segment matching.
1988
1989 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1990 used to be given if ever the end of the subject was reached; now it is
1991 given only if matching could not proceed because another character was
1992 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1993 string "Z", which now yields "no match" instead of "partial match". In the
1994 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1995 final character ended with (*FAIL).
1996
1997 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1998 if the pattern had a "must contain" character that was already found in the
1999 earlier partial match, unless partial matching was again requested. For
2000 example, with the pattern /dog.(body)?/, the "must contain" character is
2001 "g". If the first part-match was for the string "dog", restarting with
2002 "sbody" failed. This bug has been fixed.
2003
2004 13. The string returned by pcre_dfa_exec() after a partial match has been
2005 changed so that it starts at the first inspected character rather than the
2006 first character of the match. This makes a difference only if the pattern
2007 starts with a lookbehind assertion or \b or \B (\K is not supported by
2008 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2009 matching functions compatible, and I think it's the right thing to do.
2010
2011 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2012 so that the demonstration program is easily available in environments where
2013 PCRE has not been installed from source.
2014
2015 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2016 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2017 library.
2018
2019 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2020 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2021 is not the first non-POSIX option to be added. Clearly some people find
2022 these options useful.
2023
2024 17. If a caller to the POSIX matching function regexec() passes a non-zero
2025 value for nmatch with a NULL value for pmatch, the value of
2026 nmatch is forced to zero.
2027
2028 18. RunGrepTest did not have a test for the availability of the -u option of
2029 the diff command, as RunTest does. It now checks in the same way as
2030 RunTest, and also checks for the -b option.
2031
2032 19. If an odd number of negated classes containing just a single character
2033 interposed, within parentheses, between a forward reference to a named
2034 subpattern and the definition of the subpattern, compilation crashed with
2035 an internal error, complaining that it could not find the referenced
2036 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2037 [The bug was that it was starting one character too far in when skipping
2038 over the character class, thus treating the ] as data rather than
2039 terminating the class. This meant it could skip too much.]
2040
2041 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2042 /g option in pcretest when the pattern contains \K, which makes it possible
2043 to have an empty string match not at the start, even when the pattern is
2044 anchored. Updated pcretest and pcredemo to use this option.
2045
2046 21. If the maximum number of capturing subpatterns in a recursion was greater
2047 than the maximum at the outer level, the higher number was returned, but
2048 with unset values at the outer level. The correct (outer level) value is
2049 now given.
2050
2051 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2052 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2053 make it do so. The string so far is captured, making this feature
2054 compatible with Perl.
2055
2056 23. The tests have been re-organized, adding tests 11 and 12, to make it
2057 possible to check the Perl 5.10 features against Perl 5.10.
2058
2059 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2060 pattern matches a fixed length string. PCRE did not allow this; now it
2061 does. Neither allows recursion.
2062
2063 25. I finally figured out how to implement a request to provide the minimum
2064 length of subject string that was needed in order to match a given pattern.
2065 (It was back references and recursion that I had previously got hung up
2066 on.) This code has now been added to pcre_study(); it finds a lower bound
2067 to the length of subject needed. It is not necessarily the greatest lower
2068 bound, but using it to avoid searching strings that are too short does give
2069 some useful speed-ups. The value is available to calling programs via
2070 pcre_fullinfo().
2071
2072 26. While implementing 25, I discovered to my embarrassment that pcretest had
2073 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2074 study optimizations had never been tested with that matching function.
2075 Oops. What is worse, even when it was passed study data, there was a bug in
2076 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2077 were also very few tests of studied patterns with pcre_dfa_exec().
2078
2079 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2080 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2081 on the other side of the coin, they are no longer allowed to have different
2082 names, because these cannot be distinguished in PCRE, and this has caused
2083 confusion. (This is a difference from Perl.)
2084
2085 28. When duplicate subpattern names are present (necessarily with different
2086 numbers, as required by 27 above), and a test is made by name in a
2087 conditional pattern, either for a subpattern having been matched, or for
2088 recursion in such a pattern, all the associated numbered subpatterns are
2089 tested, and the overall condition is true if the condition is true for any
2090 one of them. This is the way Perl works, and is also more like the way
2091 testing by number works.
2092
2093
2094 Version 7.9 11-Apr-09
2095 ---------------------
2096
2097 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2098 (pcretest), all targets were linked against these libraries. This included
2099 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2100 libraries. This caused unwanted dependencies to be created. This problem
2101 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2102 pcretest is linked with readline.
2103
2104 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2105 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2106 moved outside it again, because FALSE and TRUE are already defined in AIX,
2107 but BOOL is not.
2108
2109 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2110 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2111
2112 4. The pcregrep documentation said spaces were inserted as well as colons (or
2113 hyphens) following file names and line numbers when outputting matching
2114 lines. This is not true; no spaces are inserted. I have also clarified the
2115 wording for the --colour (or --color) option.
2116
2117 5. In pcregrep, when --colour was used with -o, the list of matching strings
2118 was not coloured; this is different to GNU grep, so I have changed it to be
2119 the same.
2120
2121 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2122 each matching line was coloured. Now it goes on to look for further matches
2123 of any of the test patterns, which is the same behaviour as GNU grep.
2124
2125 7. A pattern that could match an empty string could cause pcregrep to loop; it
2126 doesn't make sense to accept an empty string match in pcregrep, so I have
2127 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2128 seems to be how GNU grep behaves. [But see later change 40 for release
2129 8.33.]
2130
2131 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2132 start or after a newline", because the conditional assertion was not being
2133 correctly handled. The rule now is that both the assertion and what follows
2134 in the first alternative must satisfy the test.
2135
2136 9. If auto-callout was enabled in a pattern with a conditional group whose
2137 condition was an assertion, PCRE could crash during matching, both with
2138 pcre_exec() and pcre_dfa_exec().
2139
2140 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2141 used for matching.
2142
2143 11. Unicode property support in character classes was not working for
2144 characters (bytes) greater than 127 when not in UTF-8 mode.
2145
2146 12. Added the -M command line option to pcretest.
2147
2148 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2149
2150 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2151
2152 16. Added comments and documentation about mis-use of no_arg in the C++
2153 wrapper.
2154
2155 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2156 from Martin Jerabek that uses macro names for all relevant character and
2157 string constants.
2158
2159 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2160 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2161 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2162 these, but not everybody uses configure.
2163
2164 19. A conditional group that had only one branch was not being correctly
2165 recognized as an item that could match an empty string. This meant that an
2166 enclosing group might also not be so recognized, causing infinite looping
2167 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2168 with the subject "ab", where knowledge that the repeated group can match
2169 nothing is needed in order to break the loop.
2170
2171 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2172 exec(), but without supplying a callout function, matching went wrong.
2173
2174 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2175 leak if the size of the offset vector was greater than 30. When the vector
2176 is smaller, the saved offsets during recursion go onto a local stack
2177 vector, but for larger vectors malloc() is used. It was failing to free
2178 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2179 error, in fact).
2180
2181 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2182 heapframe that is used only when UTF-8 support is enabled. This caused no
2183 problem, but was untidy.
2184
2185 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2186 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2187 included within another project.
2188
2189 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2190 slightly modified by me:
2191
2192 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2193 not building pcregrep.
2194
2195 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2196 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2197
2198 25. Forward references, both numeric and by name, in patterns that made use of
2199 duplicate group numbers, could behave incorrectly or give incorrect errors,
2200 because when scanning forward to find the reference group, PCRE was not
2201 taking into account the duplicate group numbers. A pattern such as
2202 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2203
2204 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2205 the feature of a custom pointer more persuasive (as requested by a user).
2206
2207 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2208 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2209 pre-defined.
2210
2211 28. Added support for (*UTF8) at the start of a pattern.
2212
2213 29. Arrange for flags added by the "release type" setting in CMake to be shown
2214 in the configuration summary.
2215
2216
2217 Version 7.8 05-Sep-08
2218 ---------------------
2219
2220 1. Replaced UCP searching code with optimized version as implemented for Ad
2221 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2222 stage table and inline lookup instead of a function, giving speed ups of 2
2223 to 5 times on some simple patterns that I tested. Permission was given to
2224 distribute the MultiStage2.py script that generates the tables (it's not in
2225 the tarball, but is in the Subversion repository).
2226
2227 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2228 scripts.
2229
2230 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2231 a group with a zero qualifier. The result of the study could be incorrect,
2232 or the function might crash, depending on the pattern.
2233
2234 4. Caseless matching was not working for non-ASCII characters in back
2235 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2236 It now works when Unicode Property Support is available.
2237
2238 5. In pcretest, an escape such as \x{de} in the data was always generating
2239 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2240 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2241 truncation.
2242
2243 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2244
2245 7. Added two (int) casts to pcregrep when printing the difference of two
2246 pointers, in case they are 64-bit values.
2247
2248 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2249 test 2 if it fails.
2250
2251 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2252 and a #define of that name to empty if it is not externally set. This is to
2253 allow users of MSVC to set it if necessary.
2254
2255 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2256 the convenience functions in the pcre_get.c source file.
2257
2258 11. An option change at the start of a pattern that had top-level alternatives
2259 could cause overwriting and/or a crash. This command provoked a crash in
2260 some environments:
2261
2262 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2263
2264 This potential security problem was recorded as CVE-2008-2371.
2265
2266 12. For a pattern where the match had to start at the beginning or immediately
2267 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2268 pcre_dfa_exec() could read past the end of the passed subject if there was
2269 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2270 pcretest so that it places the subject at the end of its malloc-ed buffer.
2271
2272 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2273 exec() might read past the end of the data buffer in UTF-8 mode.
2274
2275 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2276 the data contained the byte 0x85 as part of a UTF-8 character within its
2277 first line. This applied both to normal and DFA matching.
2278
2279 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2280 /^[^d]*?$/8 failed to match "abc".
2281
2282 16. Added a missing copyright notice to pcrecpp_internal.h.
2283
2284 17. Make it more clear in the documentation that values returned from
2285 pcre_exec() in ovector are byte offsets, not character counts.
2286
2287 18. Tidied a few places to stop certain compilers from issuing warnings.
2288
2289 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2290 supplied by Stefan Weber. I made a further small update for 7.8 because
2291 there is a change of source arrangements: the pcre_searchfuncs.c module is
2292 replaced by pcre_ucd.c.
2293
2294
2295 Version 7.7 07-May-08
2296 ---------------------
2297
2298 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2299 a string to a long long, pretend we don't even have a long long." This is
2300 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2301
2302 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2303 pre-7.6 versions, which defined a global no_arg variable instead of putting
2304 it in the RE class. (See also #8 below.)
2305
2306 3. Remove a line of dead code, identified by coverity and reported by Nuno
2307 Lopes.
2308
2309 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2310
2311 (1) The include/exclude patterns were being applied to the whole pathnames
2312 of files, instead of just to the final components.
2313
2314 (2) If there was more than one level of directory, the subdirectories were
2315 skipped unless they satisfied the include/exclude conditions. This is
2316 inconsistent with GNU grep (and could even be seen as contrary to the
2317 pcregrep specification - which I improved to make it absolutely clear).
2318 The action now is always to scan all levels of directory, and just
2319 apply the include/exclude patterns to regular files.
2320
2321 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2322 --exclude_dir in the tests to avoid scanning .svn directories.
2323
2324 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2325 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2326 doesn't support NULs in patterns.
2327
2328 7. Added some missing "const"s to declarations of static tables in
2329 pcre_compile.c and pcre_dfa_exec.c.
2330
2331 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2332 caused by fix #2 above. (Subsequently also a second patch to fix the
2333 first patch. And a third patch - this was a messy problem.)
2334
2335 9. Applied Craig's patch to remove the use of push_back().
2336
2337 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2338 matching function regexec().
2339
2340 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2341 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2342 references. PCRE supports relative numbers with this syntax (I don't think
2343 Oniguruma does).
2344
2345 12. Previously, a group with a zero repeat such as (...){0} was completely
2346 omitted from the compiled regex. However, this means that if the group
2347 was called as a subroutine from elsewhere in the pattern, things went wrong
2348 (an internal error was given). Such groups are now left in the compiled
2349 pattern, with a new opcode that causes them to be skipped at execution
2350 time.
2351
2352 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2353 to the way PCRE behaves:
2354
2355 (a) A lone ] character is dis-allowed (Perl treats it as data).
2356
2357 (b) A back reference to an unmatched subpattern matches an empty string
2358 (Perl fails the current match path).
2359
2360 (c) A data ] in a character class must be notated as \] because if the
2361 first data character in a class is ], it defines an empty class. (In
2362 Perl it is not possible to have an empty class.) The empty class []
2363 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2364 The negative empty class [^] matches any one character, independently
2365 of the DOTALL setting.
2366
2367 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2368 non-existent subpattern following a character class starting with ']' and
2369 containing () gave an internal compiling error instead of "reference to
2370 non-existent subpattern". Fortunately, when the pattern did exist, the
2371 compiled code was correct. (When scanning forwards to check for the
2372 existence of the subpattern, it was treating the data ']' as terminating
2373 the class, so got the count wrong. When actually compiling, the reference
2374 was subsequently set up correctly.)
2375
2376 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2377 it was being rejected as not supported by pcre_dfa_exec(), even though
2378 other assertions are supported. I have made pcre_dfa_exec() support
2379 (*FAIL).
2380
2381 16. The implementation of 13c above involved the invention of a new opcode,
2382 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2383 cannot be changed at match time, I realized I could make a small
2384 improvement to matching performance by compiling OP_ALLANY instead of
2385 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2386 on the OP_ANY path.
2387
2388 17. Compiling pcretest on Windows with readline support failed without the
2389 following two fixes: (1) Make the unistd.h include conditional on
2390 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2391
2392 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2393 ncurses library to be included for pcretest when ReadLine support is
2394 requested, but also to allow for it to be overridden. This patch came from
2395 Daniel Bergström.
2396
2397 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2398 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2399 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2400 spotting this.
2401
2402
2403 Version 7.6 28-Jan-08
2404 ---------------------
2405
2406 1. A character class containing a very large number of characters with
2407 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2408 overflow.
2409
2410 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2411 HAVE_LONG_LONG is not defined.
2412
2413 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2414 bring them up to date and include new features. This patch includes:
2415
2416 - Fixed PH's badly added libz and libbz2 support.
2417 - Fixed a problem with static linking.
2418 - Added pcredemo. [But later removed - see 7 below.]
2419 - Fixed dftables problem and added an option.
2420 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2421 HAVE_LONG_LONG.
2422 - Added readline support for pcretest.
2423 - Added an listing of the option settings after cmake has run.
2424
2425 4. A user submitted a patch to Makefile that makes it easy to create
2426 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2427 Makefile.am that cause it to include this special target, without
2428 affecting anything else. Note that the same mingw target plus all
2429 the other distribution libraries and programs are now supported
2430 when configuring with CMake (see 6 below) instead of with
2431 Configure/Make.
2432
2433 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2434 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2435 exported in the Windows port". It has not yet been confirmed that the patch
2436 solves the problem, but it does no harm.
2437
2438 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2439 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2440 with CMake, and also correct the comment about stack recursion.
2441
2442 7. Remove the automatic building of pcredemo from the ./configure system and
2443 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2444 of a program that users should build themselves after PCRE is installed, so
2445 building it automatically is not really right. What is more, it gave
2446 trouble in some build environments.
2447
2448 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2449
2450
2451 Version 7.5 10-Jan-08
2452 ---------------------
2453
2454 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2455 values in parens when parsing an RE using the C++ wrapper."
2456
2457 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2458 Characters greater than 255 were excluded from the class instead of being
2459 included.
2460
2461 3. The same bug as (2) above applied to negated POSIX classes such as
2462 [:^space:].
2463
2464 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2465 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2466 I have changed it.
2467
2468 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2469 first named subpattern) and a construct such as (?&a) would reference the
2470 first named subpattern whose name started with "a" (in other words, the
2471 length check was missing). Both these problems are fixed. "Subpattern name
2472 expected" is now given for (?&) (a zero-length name), and this patch also
2473 makes it give the same error for \k'' (previously it complained that that
2474 was a reference to a non-existent subpattern).
2475
2476 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2477 this is right because (?- can be followed by option settings as well as by
2478 digits. I have, however, made the messages clearer.
2479
2480 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2481 than the number used in the conditional) now cause a compile-time error.
2482 This is actually not compatible with Perl, which accepts such patterns, but
2483 treats the conditional as always being FALSE (as PCRE used to), but it
2484 seems to me that giving a diagnostic is better.
2485
2486 8. Change "alphameric" to the more common word "alphanumeric" in comments
2487 and messages.
2488
2489 9. Fix two occurrences of "backslash" in comments that should have been
2490 "backspace".
2491
2492 10. Remove two redundant lines of code that can never be obeyed (their function
2493 was moved elsewhere).
2494
2495 11. The program that makes PCRE's Unicode character property table had a bug
2496 which caused it to generate incorrect table entries for sequences of
2497 characters that have the same character type, but are in different scripts.
2498 It amalgamated them into a single range, with the script of the first of
2499 them. In other words, some characters were in the wrong script. There were
2500 thirteen such cases, affecting characters in the following ranges:
2501
2502 U+002b0 - U+002c1
2503 U+0060c - U+0060d
2504 U+0061e - U+00612
2505 U+0064b - U+0065e
2506 U+0074d - U+0076d
2507 U+01800 - U+01805
2508 U+01d00 - U+01d77
2509 U+01d9b - U+01dbf
2510 U+0200b - U+0200f
2511 U+030fc - U+030fe
2512 U+03260 - U+0327f
2513 U+0fb46 - U+0fbb1
2514 U+10450 - U+1049d
2515
2516 12. The -o option (show only the matching part of a line) for pcregrep was not
2517 compatible with GNU grep in that, if there was more than one match in a
2518 line, it showed only the first of them. It now behaves in the same way as
2519 GNU grep.
2520
2521 13. If the -o and -v options were combined for pcregrep, it printed a blank
2522 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2523 does the same. The return code can be used to tell if there were any
2524 non-matching lines.
2525
2526 14. Added --file-offsets and --line-offsets to pcregrep.
2527
2528 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2529 infinitely looping recursion. The bug was that positive lookaheads were not
2530 being skipped when checking for a possible empty match (negative lookaheads
2531 and both kinds of lookbehind were skipped).
2532
2533 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2534 inclusion of <windows.h> to before rather than after the definition of
2535 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2536
2537 17. Specifying a possessive quantifier with a specific limit for a Unicode
2538 character property caused pcre_compile() to compile bad code, which led at
2539 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2540 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2541 caused the error; without that there was no problem.
2542
2543 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2544
2545 19. Added --enable-pcretest-libreadline.
2546
2547 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2548 RE::GlobalReplace(). As a result, the number of replacements returned was
2549 double what it should be. I removed one of the increments, but Craig sent a
2550 later patch that removed the other one (the right fix) and added unit tests
2551 that check the return values (which was not done before).
2552
2553 21. Several CMake things:
2554
2555 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2556 the names libpcre and libpcreposix, not just pcre and pcreposix.
2557
2558 (2) The above change means that pcretest and pcregrep are now correctly
2559 linked with the newly-built libraries, not previously installed ones.
2560
2561 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2562
2563 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2564 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2565 UTF-8 newline character). The key issue is that the pattern starts .*;
2566 this means that the match must be either at the beginning, or after a
2567 newline. The bug was in the code for advancing after a failed match and
2568 checking that the new position followed a newline. It was not taking
2569 account of UTF-8 characters correctly.
2570
2571 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2572 character classes. PCRE was not treating the sequence [:...:] as a
2573 character class unless the ... were all letters. Perl, however, seems to
2574 allow any characters between [: and :], though of course it rejects as
2575 unknown any "names" that contain non-letters, because all the known class
2576 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2577 for example, whereas PCRE did not - it did not recognize a POSIX character
2578 class. This seemed a bit dangerous, so the code has been changed to be
2579 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2580 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2581 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2582 Perl does, and where it didn't before.
2583
2584 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2585 Windows environments %n is disabled by default.
2586
2587
2588 Version 7.4 21-Sep-07
2589 ---------------------
2590
2591 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2592 means that a class such as [\s] counted as "explicit reference to CR or
2593 LF". That isn't really right - the whole point of the change was to try to
2594 help when there was an actual mention of one of the two characters. So now
2595 the change happens only if \r or \n (or a literal CR or LF) character is
2596 encountered.
2597
2598 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2599 of both had grown to the point where there were only 3 bits left.
2600 Fortunately, there was spare space in the data structure, and so I have
2601 moved the internal flags into a new 16-bit field to free up more option
2602 bits.
2603
2604 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2605 but did not set the internal JCHANGED flag - either of these is enough to
2606 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2607 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2608 start sets both bits.
2609
2610 4. Added options (at build time, compile time, exec time) to change \R from
2611 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2612
2613 5. doc/pcresyntax.html was missing from the distribution.
2614
2615 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2616 compatibility, even though it is no longer used.
2617
2618 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2619 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2620 windows.h file is present (where different names are used). [This was
2621 reversed later after testing - see 16 below.]
2622
2623 8. Changed all #include <config.h> to #include "config.h". There were also
2624 some further <pcre.h> cases that I changed to "pcre.h".
2625
2626 9. When pcregrep was used with the --colour option, it missed the line ending
2627 sequence off the lines that it output.
2628
2629 10. It was pointed out to me that arrays of string pointers cause lots of
2630 relocations when a shared library is dynamically loaded. A technique of
2631 using a single long string with a table of offsets can drastically reduce
2632 these. I have refactored PCRE in four places to do this. The result is
2633 dramatic:
2634
2635 Originally: 290
2636 After changing UCP table: 187
2637 After changing error message table: 43
2638 After changing table of "verbs" 36
2639 After changing table of Posix names 22
2640
2641 Thanks to the folks working on Gregex for glib for this insight.
2642
2643 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2644 unicode-properties was also set.
2645
2646 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2647
2648 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2649 checked only for CRLF.
2650
2651 14. Added casts to pcretest.c to avoid compiler warnings.
2652
2653 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2654
2655 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2656 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2657 entirely. This removes changes made in 7 above.
2658
2659 17. The CMake files have been updated, and there is now more information about
2660 building with CMake in the NON-UNIX-USE document.
2661
2662
2663 Version 7.3 28-Aug-07
2664 ---------------------
2665
2666 1. In the rejigging of the build system that eventually resulted in 7.1, the
2667 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2668 brackets there is not right, since it causes compilers to look for an
2669 installed pcre.h, not the version that is in the source that is being
2670 compiled (which of course may be different). I have changed it back to:
2671
2672 #include "pcre.h"
2673
2674 I have a vague recollection that the change was concerned with compiling in
2675 different directories, but in the new build system, that is taken care of
2676 by the VPATH setting the Makefile.
2677
2678 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2679 when the subject happened to end in the byte 0x85 (e.g. if the last
2680 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2681 characters but of course it shouldn't be taken as a newline when it is part
2682 of another character. The bug was that, for an unlimited repeat of . in
2683 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2684 characters when looking for a newline.
2685
2686 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2687
2688 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2689 in debug output.
2690
2691 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2692 long printing in the pcrecpp unittest when running under MinGW.
2693
2694 6. ESC_K was left out of the EBCDIC table.
2695
2696 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2697 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2698 limit also applies to "virtual nesting" when a pattern is recursive, and in
2699 this case 1000 isn't so big. I have been able to remove this limit at the
2700 expense of backing off one optimization in certain circumstances. Normally,
2701 when pcre_exec() would call its internal match() function recursively and
2702 immediately return the result unconditionally, it uses a "tail recursion"
2703 feature to save stack. However, when a subpattern that can match an empty
2704 string has an unlimited repetition quantifier, it no longer makes this
2705 optimization. That gives it a stack frame in which to save the data for
2706 checking that an empty string has been matched. Previously this was taken
2707 from the 1000-entry workspace that had been reserved. So now there is no
2708 explicit limit, but more stack is used.
2709
2710 8. Applied Daniel's patches to solve problems with the import/export magic
2711 syntax that is required for Windows, and which was going wrong for the
2712 pcreposix and pcrecpp parts of the library. These were overlooked when this
2713 problem was solved for the main library.
2714
2715 9. There were some crude static tests to avoid integer overflow when computing
2716 the size of patterns that contain repeated groups with explicit upper
2717 limits. As the maximum quantifier is 65535, the maximum group length was
2718 set at 30,000 so that the product of these two numbers did not overflow a
2719 32-bit integer. However, it turns out that people want to use groups that
2720 are longer than 30,000 bytes (though not repeat them that many times).
2721 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2722 made it possible to implement the integer overflow checks in a much more
2723 dynamic way, which I have now done. The artificial limitation on group
2724 length has been removed - we now have only the limit on the total length of
2725 the compiled pattern, which depends on the LINK_SIZE setting.
2726
2727 10. Fixed a bug in the documentation for get/copy named substring when
2728 duplicate names are permitted. If none of the named substrings are set, the
2729 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2730 empty string.
2731
2732 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2733 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2734 because the ] is interpreted as the first data character and the
2735 terminating ] is not found. PCRE has been made compatible with Perl in this
2736 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2737 cause memory overwriting.
2738
2739 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2740 string has been matched (to stop an infinite loop). It was not recognizing
2741 a conditional subpattern that could match an empty string if that
2742 subpattern was within another subpattern. For example, it looped when
2743 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2744 condition was not nested. This bug has been fixed.
2745
2746 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2747 past the start of the subject in the presence of bytes with the top bit
2748 set, for example "\x8aBCD".
2749
2750 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2751 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2752
2753 14. Optimized (?!) to (*FAIL).
2754
2755 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2756 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2757 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2758 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2759 does: it's just the validity check that is more restrictive.
2760
2761 16. Inserted checks for integer overflows during escape sequence (backslash)
2762 processing, and also fixed erroneous offset values for syntax errors during
2763 backslash processing.
2764
2765 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2766 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2767
2768 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2769 caused an overrun.
2770
2771 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2772 something other than just ASCII characters) inside a group that had an
2773 unlimited repeat caused a loop at compile time (while checking to see
2774 whether the group could match an empty string).
2775
2776 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2777 [\P{Any}] did so. (Error in the code for printing property names.)
2778
2779 21. An orphan \E inside a character class could cause a crash.
2780
2781 22. A repeated capturing bracket such as (A)? could cause a wild memory
2782 reference during compilation.
2783
2784 23. There are several functions in pcre_compile() that scan along a compiled
2785 expression for various reasons (e.g. to see if it's fixed length for look
2786 behind). There were bugs in these functions when a repeated \p or \P was
2787 present in the pattern. These operators have additional parameters compared
2788 with \d, etc, and these were not being taken into account when moving along
2789 the compiled data. Specifically:
2790
2791 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2792 length.
2793
2794 (b) An item such as \pL+ within a repeated group could cause crashes or
2795 loops.
2796
2797 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2798 "reference to non-existent subpattern" error.
2799
2800 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2801
2802 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2803 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2804
2805 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2806
2807 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2808 character were causing crashes (broken optimization).
2809
2810 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2811 \p or \P) caused a compile-time loop.
2812
2813 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2814 break. For example, the unstudied pattern [\r\n]A does not match the string
2815 "\r\nA" because change 7.0/46 below moves the current point on by two
2816 characters after failing to match at the start. However, the pattern \nA
2817 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2818 the same is true. There doesn't seem any very clean way out of this, but
2819 what I have chosen to do makes the common cases work: PCRE now takes note
2820 of whether there can be an explicit match for \r or \n anywhere in the
2821 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2822 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2823 pattern has explicit CR or LF references.
2824
2825 29. Added (*CR) etc for changing newline setting at start of pattern.
2826
2827
2828 Version 7.2 19-Jun-07
2829 ---------------------
2830
2831 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2832 which is apparently normally available under Windows.
2833
2834 2. Re-jig the pcregrep tests with different newline settings in an attempt
2835 to make them independent of the local environment's newline setting.
2836
2837 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2838
2839 4. Some of the "internals" tests were previously cut out when the link size
2840 was not 2, because the output contained actual offsets. The recent new
2841 "Z" feature of pcretest means that these can be cut out, making the tests
2842 usable with all link sizes.
2843
2844 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2845 stack recursion. This gives a massive performance boost under BSD, but just
2846 a small improvement under Linux. However, it saves one field in the frame
2847 in all cases.
2848
2849 6. Added more features from the forthcoming Perl 5.10:
2850
2851 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2852 recursion call. It refers to the nth most recently opened parentheses.
2853
2854 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2855 to be opened parentheses.
2856
2857 (c) Conditions that refer to capturing parentheses can be specified
2858 relatively, for example, (?(-2)... or (?(+3)...
2859
2860 (d) \K resets the start of the current match so that everything before
2861 is not part of it.
2862
2863 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2864
2865 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2866 reference syntax.
2867
2868 (g) (?| introduces a group in which the numbering of parentheses in each
2869 alternative starts with the same number.
2870
2871 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2872
2873 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2874 PCRE_INFO_JCHANGED.
2875
2876 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2877 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2878 for detecting groups that can match an empty string.
2879
2880 9. A pattern with a very large number of alternatives (more than several
2881 hundred) was running out of internal workspace during the pre-compile
2882 phase, where pcre_compile() figures out how much memory will be needed. A
2883 bit of new cunning has reduced the workspace needed for groups with
2884 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2885 workspace instead of running out of the 4096 that are available.
2886
2887 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2888
2889 11. Applied patch from Google to remove an optimization that didn't quite work.
2890 The report of the bug said:
2891
2892 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2893 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2894 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2895
2896 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2897 it matched the wrong number of bytes.
2898
2899
2900 Version 7.1 24-Apr-07
2901 ---------------------
2902
2903 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2904 that is more "standard", making use of automake and other Autotools. There
2905 is some re-arrangement of the files and adjustment of comments consequent
2906 on this.
2907
2908 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2909 for recursive directory scanning broke on some systems because the files
2910 are not scanned in any specific order and on different systems the order
2911 was different. A call to "sort" has been inserted into RunGrepTest for the
2912 approprate test as a short-term fix. In the longer term there may be an
2913 alternative.
2914
2915 3. I had an email from Eric Raymond about problems translating some of PCRE's
2916 man pages to HTML (despite the fact that I distribute HTML pages, some
2917 people do their own conversions for various reasons). The problems
2918 concerned the use of low-level troff macros .br and .in. I have therefore
2919 removed all such uses from the man pages (some were redundant, some could
2920 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2921 HTML has been updated to handle .nf/.fi and to complain if it encounters
2922 .br or .in.
2923
2924 4. Updated comments in configure.ac that get placed in config.h.in and also
2925 arranged for config.h to be included in the distribution, with the name
2926 config.h.generic, for the benefit of those who have to compile without
2927 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2928
2929 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2930 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2931 makevp.bat for the current PCRE, using the additional files
2932 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2933
2934 6. A Windows user reported a minor discrepancy with test 2, which turned out
2935 to be caused by a trailing space on an input line that had got lost in his
2936 copy. The trailing space was an accident, so I've just removed it.
2937
2938 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2939 that is needed.
2940
2941 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2942 as "const" (a) because they are and (b) because it helps the PHP
2943 maintainers who have recently made a script to detect big data structures
2944 in the php code that should be moved to the .rodata section. I remembered
2945 to update Builducptable as well, so it won't revert if ucptable.h is ever
2946 re-created.
2947
2948 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2949 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2950 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2951 support is not required. This saves 1.5-2K of code, which is important in
2952 some applications.
2953
2954 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2955 so as not to refer to the tables, even though these functions will never be
2956 called when UTF-8 support is disabled. Otherwise there are problems with a
2957 shared library.
2958
2959 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2960
2961 (a) It was defining its arguments as char * instead of void *.
2962
2963 (b) It was assuming that all moves were upwards in memory; this was true
2964 a long time ago when I wrote it, but is no longer the case.
2965
2966 The emulated memove() is provided for those environments that have neither
2967 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2968 is clearly not the case, as these two bugs were recently reported.
2969
2970 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2971 and Detrail to create the HTML documentation, the .txt form of the man
2972 pages, and it removes trailing spaces from listed files. It also creates
2973 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2974 case, it wraps all the #defines with #ifndefs. This script should be run
2975 before "make dist".
2976
2977 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2978 with Unicode property support.
2979
2980 (a) For a maximizing quantifier, if the two different cases of the
2981 character were of different lengths in their UTF-8 codings (there are
2982 some cases like this - I found 11), and the matching function had to
2983 back up over a mixture of the two cases, it incorrectly assumed they
2984 were both the same length.
2985
2986 (b) When PCRE was configured to use the heap rather than the stack for
2987 recursion during matching, it was not correctly preserving the data for
2988 the other case of a UTF-8 character when checking ahead for a match
2989 while processing a minimizing repeat. If the check also involved
2990 matching a wide character, but failed, corruption could cause an
2991 erroneous result when trying to check for a repeat of the original
2992 character.
2993
2994 13. Some tidying changes to the testing mechanism:
2995
2996 (a) The RunTest script now detects the internal link size and whether there
2997 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2998 values substituted by "configure". (The RunGrepTest script already did
2999 this for UTF-8.) The configure.ac script no longer substitutes the
3000 relevant variables.
3001
3002 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3003 with length and offset values. This means that the output is different
3004 for different internal link sizes. Test 2 is skipped for link sizes
3005 other than 2 because of this, bypassing the problem. Unfortunately,
3006 there was also a test in test 3 (the locale tests) that used /B and
3007 failed for link sizes other than 2. Rather than cut the whole test out,
3008 I have added a new /Z option to pcretest that replaces the length and
3009 offset values with spaces. This is now used to make test 3 independent
3010 of link size. (Test 2 will be tidied up later.)
3011
3012 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3013 segmentation fault instead of returning the appropriate error message.
3014
3015 15. In multiline mode when the newline sequence was set to "any", the pattern
3016 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3017 This doesn't seem right; it now treats the CRLF combination as the line
3018 ending, and so does not match in that case. It's only a pattern such as ^$
3019 that would hit this one: something like ^ABC$ would have failed after \r
3020 and then tried again after \r\n.
3021
3022 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3023 in an attempt to make files that differ only in their line terminators
3024 compare equal. This works on Linux.
3025
3026 17. Under certain error circumstances pcregrep might try to free random memory
3027 as it exited. This is now fixed, thanks to valgrind.
3028
3029 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3030 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3031 was because its rules for how to advance for /g after matching an empty
3032 string at the end of a line did not allow for this case. They now check for
3033 it specially.
3034
3035 20. pcretest is supposed to handle patterns and data of any length, by
3036 extending its buffers when necessary. It was getting this wrong when the
3037 buffer for a data line had to be extended.
3038
3039 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3040 CRLF as a newline sequence.
3041
3042 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3043 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3044 I have nevertheless tidied it up.
3045
3046 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3047
3048 24. Added a man page for pcre-config.
3049
3050
3051 Version 7.0 19-Dec-06
3052 ---------------------
3053
3054 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3055 moving to gcc 4.1.1.
3056
3057 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3058 sys/time.h, which is documented as needed for this function. It doesn't
3059 seem to matter on Linux, but it showed up on some releases of OS X.
3060
3061 3. It seems that there are systems where bytes whose values are greater than
3062 127 match isprint() in the "C" locale. The "C" locale should be the
3063 default when a C program starts up. In most systems, only ASCII printing
3064 characters match isprint(). This difference caused the output from pcretest
3065 to vary, making some of the tests fail. I have changed pcretest so that:
3066
3067 (a) When it is outputting text in the compiled version of a pattern, bytes
3068 other than 32-126 are always shown as hex escapes.
3069
3070 (b) When it is outputting text that is a matched part of a subject string,
3071 it does the same, unless a different locale has been set for the match
3072 (using the /L modifier). In this case, it uses isprint() to decide.
3073
3074 4. Fixed a major bug that caused incorrect computation of the amount of memory
3075 required for a compiled pattern when options that changed within the
3076 pattern affected the logic of the preliminary scan that determines the
3077 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3078 that the computed length was too small. The symptoms of this bug were
3079 either the PCRE error "internal error: code overflow" from pcre_compile(),
3080 or a glibc crash with a message such as "pcretest: free(): invalid next
3081 size (fast)". Examples of patterns that provoked this bug (shown in
3082 pcretest format) are:
3083
3084 /(?-x: )/x
3085 /(?x)(?-x: \s*#\s*)/
3086 /((?i)[\x{c0}])/8
3087 /(?i:[\x{c0}])/8
3088
3089 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3090 is now done differently.
3091
3092 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3093 wrapper classes; (b) implement a new function in the C++ scanner that is
3094 more efficient than the old way of doing things because it avoids levels of
3095 recursion in the regex matching; (c) add a paragraph to the documentation
3096 for the FullMatch() function.
3097
3098 6. The escape sequence \n was being treated as whatever was defined as
3099 "newline". Not only was this contrary to the documentation, which states
3100 that \n is character 10 (hex 0A), but it also went horribly wrong when
3101 "newline" was defined as CRLF. This has been fixed.
3102
3103 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3104 was being set to -1 for the "end of line" case (supposedly a value that no
3105 character can have). Though this value is never used (the check for end of
3106 line is "zero bytes in current character"), it caused compiler complaints.
3107 I've changed it to 0xffffffff.
3108
3109 8. In pcre_version.c, the version string was being built by a sequence of
3110 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3111 string (as it is for production releases) called a macro with an empty
3112 argument. The C standard says the result of this is undefined. The gcc
3113 compiler treats it as an empty string (which was what was wanted) but it is
3114 reported that Visual C gives an error. The source has been hacked around to
3115 avoid this problem.
3116
3117 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3118 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3119 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3120 of them did).
3121
3122 10. Originally, pcretest opened its input and output without "b"; then I was
3123 told that "b" was needed in some environments, so it was added for release
3124 5.0 to both the input and output. (It makes no difference on Unix-like
3125 systems.) Later I was told that it is wrong for the input on Windows. I've
3126 now abstracted the modes into two macros, to make it easier to fiddle with
3127 them, and removed "b" from the input mode under Windows.
3128
3129 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3130
3131 12. Added -help and --help to pcretest as an official way of being reminded
3132 of the options.
3133
3134 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3135 and pcrecpp.cc because they annoy compilers at high warning levels.
3136
3137 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3138
3139 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3140 scripts are not C programs :-) and which was not noticed because it works
3141 on Linux.
3142
3143 16. pcretest is supposed to handle any length of pattern and data line (as one
3144 line or as a continued sequence of lines) by extending its input buffer if
3145 necessary. This feature was broken for very long pattern lines, leading to
3146 a string of junk being passed to pcre_compile() if the pattern was longer
3147 than about 50K.
3148
3149 17. I have done a major re-factoring of the way pcre_compile() computes the
3150 amount of memory needed for a compiled pattern. Previously, there was code
3151 that made a preliminary scan of the pattern in order to do this. That was
3152 OK when PCRE was new, but as the facilities have expanded, it has become
3153 harder and harder to keep it in step with the real compile phase, and there
3154 have been a number of bugs (see for example, 4 above). I have now found a
3155 cunning way of running the real compile function in a "fake" mode that
3156 enables it to compute how much memory it would need, while actually only
3157 ever using a few hundred bytes of working memory and without too many
3158 tests of the mode. This should make future maintenance and development
3159 easier. A side effect of this work is that the limit of 200 on the nesting
3160 depth of parentheses has been removed (though this was never a serious
3161 limitation, I suspect). However, there is a downside: pcre_compile() now
3162 runs more slowly than before (30% or more, depending on the pattern). I
3163 hope this isn't a big issue. There is no effect on runtime performance.
3164
3165 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3166 newline (only possible for the last line of a file) and it was a
3167 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3168
3169 19. Added additional timing features to pcretest. (1) The -tm option now times
3170 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3171 separate command line item, by a number that specifies the number of
3172 repeats to use when timing. The default is 50000; this gives better
3173 precision, but takes uncomfortably long for very large patterns.
3174
3175 20. Extended pcre_study() to be more clever in cases where a branch of a
3176 subpattern has no definite first character. For example, (a*|b*)[cd] would
3177 previously give no result from pcre_study(). Now it recognizes that the
3178 first character must be a, b, c, or d.
3179
3180 21. There was an incorrect error "recursive call could loop indefinitely" if
3181 a subpattern (or the entire pattern) that was being tested for matching an
3182 empty string contained only one non-empty item after a nested subpattern.
3183 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3184 incorrectly, because the \d was being skipped in the check.
3185
3186 22. The pcretest program now has a new pattern option /B and a command line
3187 option -b, which is equivalent to adding /B to every pattern. This causes
3188 it to show the compiled bytecode, without the additional information that
3189 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3190 is the same as /B/I).
3191
3192 23. A new optimization is now able automatically to treat some sequences such
3193 as a*b as a*+b. More specifically, if something simple (such as a character
3194 or a simple class like \d) has an unlimited quantifier, and is followed by
3195 something that cannot possibly match the quantified thing, the quantifier
3196 is automatically "possessified".
3197
3198 24. A recursive reference to a subpattern whose number was greater than 39
3199 went wrong under certain circumstances in UTF-8 mode. This bug could also
3200 have affected the operation of pcre_study().
3201
3202 25. Realized that a little bit of performance could be had by replacing
3203 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3204
3205 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3206
3207 27. Possessive quantifiers such as a++ were previously implemented by turning
3208 them into atomic groups such as ($>a+). Now they have their own opcodes,
3209 which improves performance. This includes the automatically created ones
3210 from 23 above.
3211
3212 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3213 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3214 the first matched character to be a colon. This applied both to named and
3215 numbered groups.
3216
3217 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3218
3219 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3220 building PCRE on Windows easier, so I have included it in the distribution.
3221
3222 31. There is now a check in pcretest against a ridiculously large number being
3223 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3224 loop, the loop is abandoned.
3225
3226 32. Forward references to subpatterns in conditions such as (?(2)...) where
3227 subpattern 2 is defined later cause pcre_compile() to search forwards in
3228 the pattern for the relevant set of parentheses. This search went wrong
3229 when there were unescaped parentheses in a character class, parentheses
3230 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3231
3232 33. "Subroutine" calls and backreferences were previously restricted to
3233 referencing subpatterns earlier in the regex. This restriction has now
3234 been removed.
3235
3236 34. Added a number of extra features that are going to be in Perl 5.10. On the
3237 whole, these are just syntactic alternatives for features that PCRE had
3238 previously implemented using the Python syntax or my own invention. The
3239 other formats are all retained for compatibility.
3240
3241 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3242 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3243 also .NET compatible.
3244
3245 (b) A recursion or subroutine call to a named group can now be defined as
3246 (?&name) as well as (?P>name).
3247
3248 (c) A backreference to a named group can now be defined as \k<name> or
3249 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3250 5.10, are also .NET compatible.
3251
3252 (d) A conditional reference to a named group can now use the syntax
3253 (?(<name>) or (?('name') as well as (?(name).
3254
3255 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3256 groups (named and numbered) that are never evaluated inline, but can be
3257 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3258 is always false. There may be only one alternative in such a group.
3259
3260 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3261 as the simple (?(R). The condition is true only if the most recent
3262 recursion is that of the given number or name. It does not search out
3263 through the entire recursion stack.
3264
3265 (g) The escape \gN or \g{N} has been added, where N is a positive or
3266 negative number, specifying an absolute or relative reference.
3267
3268 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3269 some "unreachable code" warnings.
3270
3271 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3272 things, this adds five new scripts.
3273
3274 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3275 There were also incompatibilities regarding the handling of \Q..\E inside
3276 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3277 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3278
3279 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3280 matches an empty string, and forcibly breaks the loop. There were bugs in
3281 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3282 against aaaa the result was just "a" rather than "aaaa", for example. Two
3283 separate and independent bugs (that affected different cases) have been
3284 fixed.
3285
3286 39. Refactored the code to abolish the use of different opcodes for small
3287 capturing bracket numbers. This is a tidy that I avoided doing when I
3288 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3289 The new approach is not only tidier, it makes it possible to reduce the
3290 memory needed to fix the previous bug (38).
3291
3292 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3293 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3294 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3295 mode.
3296
3297 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3298 report.
3299
3300 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3301 copy construction and assignment in the C++ wrapper.
3302
3303 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3304 couple of bugs that could have given wrong results in the "--newline=crlf"
3305 case.
3306
3307 44. Added a number of casts and did some reorganization of signed/unsigned int
3308 variables following suggestions from Dair Grant. Also renamed the variable
3309 "this" as "item" because it is a C++ keyword.
3310
3311 45. Arranged for dftables to add
3312
3313 #include "pcre_internal.h"
3314
3315 to pcre_chartables.c because without it, gcc 4.x may remove the array
3316 definition from the final binary if PCRE is built into a static library and
3317 dead code stripping is activated.
3318
3319 46. For an unanchored pattern, if a match attempt fails at the start of a
3320 newline sequence, and the newline setting is CRLF or ANY, and the next two
3321 characters are CRLF, advance by two characters instead of one.
3322
3323
3324 Version 6.7 04-Jul-06
3325 ---------------------
3326
3327 1. In order to handle tests when input lines are enormously long, pcretest has
3328 been re-factored so that it automatically extends its buffers when
3329 necessary. The code is crude, but this _is_ just a test program. The
3330 default size has been increased from 32K to 50K.
3331
3332 2. The code in pcre_study() was using the value of the re argument before
3333 testing it for NULL. (Of course, in any sensible call of the function, it
3334 won't be NULL.)
3335
3336 3. The memmove() emulation function in pcre_internal.h, which is used on
3337 systems that lack both memmove() and bcopy() - that is, hardly ever -
3338 was missing a "static" storage class specifier.
3339
3340 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3341 containing an extended class (one that cannot be represented by a bitmap
3342 because it contains high-valued characters or Unicode property items, e.g.
3343 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3344 pattern, but PCRE should not loop if you do not (it no longer does).
3345 [Detail: two cases were found: (a) a repeated subpattern containing an
3346 extended class; (b) a recursive reference to a subpattern that followed a
3347 previous extended class. It wasn't skipping over the extended class
3348 correctly when UTF-8 mode was not set.]
3349
3350 5. A negated single-character class was not being recognized as fixed-length
3351 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3352 compile error "lookbehind assertion is not fixed length".
3353
3354 6. The RunPerlTest auxiliary script was showing an unexpected difference
3355 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3356 write a Perl script that can interpret lines of an input file either as
3357 byte characters or as UTF-8, which is what "perltest" was being required to
3358 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3359 can't do is switch easily at run time between having the "use utf8;" pragma
3360 or not. In the end, I fudged it by using the RunPerlTest script to insert
3361 "use utf8;" explicitly for the UTF-8 tests.
3362
3363 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3364 the end of the subject string, contrary to the documentation and to what
3365 Perl does. This was true of both matching functions. Now it matches only at
3366 the start of the subject and immediately after *internal* newlines.
3367
3368 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3369 a pointer to an int instead of a pointer to an unsigned long int. This
3370 caused problems on 64-bit systems.
3371
3372 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3373 instance of the 'standard' template library not being so standard".
3374
3375 10. There was no check on the number of named subpatterns nor the maximum
3376 length of a subpattern name. The product of these values is used to compute
3377 the size of the memory block for a compiled pattern. By supplying a very
3378 long subpattern name and a large number of named subpatterns, the size
3379 computation could be caused to overflow. This is now prevented by limiting
3380 the length of names to 32 characters, and the number of named subpatterns
3381 to 10,000.
3382
3383 11. Subpatterns that are repeated with specific counts have to be replicated in
3384 the compiled pattern. The size of memory for this was computed from the
3385 length of the subpattern and the repeat count. The latter is limited to
3386 65535, but there was no limit on the former, meaning that integer overflow
3387 could in principle occur. The compiled length of a repeated subpattern is
3388 now limited to 30,000 bytes in order to prevent this.
3389
3390 12. Added the optional facility to have named substrings with the same name.
3391
3392 13. Added the ability to use a named substring as a condition, using the
3393 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3394 are numbers (not recommended). Forward references are permitted.
3395
3396 14. Added forward references in named backreferences (if you see what I mean).
3397
3398 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3399 pattern could run off the end of the subject. For example, the pattern
3400 "(?s)(.{1,5})"8 did this with the subject "ab".
3401
3402 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3403 PCRE_CASELESS was set when matching characters that were quantified with ?
3404 or *.
3405
3406 17. A character class other than a single negated character that had a minimum
3407 but no maximum quantifier - for example [ab]{6,} - was not handled
3408 correctly by pce_dfa_exec(). It would match only one character.
3409
3410 18. A valid (though odd) pattern that looked like a POSIX character
3411 class but used an invalid character after [ (for example [[,abc,]]) caused
3412 pcre_compile() to give the error "Failed: internal error: code overflow" or
3413 in some cases to crash with a glibc free() error. This could even happen if
3414 the pattern terminated after [[ but there just happened to be a sequence of
3415 letters, a binary zero, and a closing ] in the memory that followed.
3416
3417 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3418 over the years. Originally (before any Unicode support), just the bottom 8
3419 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3420 output from "man perlunicode" includes this:
3421
3422 The regular expression compiler produces polymorphic opcodes. That
3423 is, the pattern adapts to the data and automatically switches to
3424 the Unicode character scheme when presented with Unicode data--or
3425 instead uses a traditional byte scheme when presented with byte
3426 data.
3427
3428 Sadly, a wide octal escape does not cause a switch, and in a string with
3429 no other multibyte characters, these octal escapes are treated as before.
3430 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3431 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3432 Unicode string.
3433
3434 I have not perpetrated such confusion in PCRE. Up till now, it took just
3435 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3436 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3437 translate to the appropriate multibyte character.
3438
3439 29. Applied some refactoring to reduce the number of warnings from Microsoft
3440 and Borland compilers. This has included removing the fudge introduced
3441 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3442 a warning about an unused variable.
3443
3444 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3445 characters since release 4.0, because Perl (from release 5.004) does not.
3446 [Or at least, is documented not to: some releases seem to be in conflict
3447 with the documentation.] However, when a pattern was studied with
3448 pcre_study() and all its branches started with \s, PCRE still included VT
3449 as a possible starting character. Of course, this did no harm; it just
3450 caused an unnecessary match attempt.
3451
3452 22. Removed a now-redundant internal flag bit that recorded the fact that case
3453 dependency changed within the pattern. This was once needed for "required
3454 byte" processing, but is no longer used. This recovers a now-scarce options
3455 bit. Also moved the least significant internal flag bit to the most-
3456 significant bit of the word, which was not previously used (hangover from
3457 the days when it was an int rather than a uint) to free up another bit for
3458 the future.
3459
3460 23. Added support for CRLF line endings as well as CR and LF. As well as the
3461 default being selectable at build time, it can now be changed at runtime
3462 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3463 specify that it is scanning data with non-default line endings.
3464
3465 24. Changed the definition of CXXLINK to make it agree with the definition of
3466 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3467
3468 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3469 recursions. This makes a big different to stack usage for some patterns.
3470
3471 26. If a subpattern containing a named recursion or subroutine reference such
3472 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3473 the space required for the compiled pattern went wrong and gave too small a
3474 value. Depending on the environment, this could lead to "Failed: internal
3475 error: code overflow at offset 49" or "glibc detected double free or
3476 corruption" errors.
3477
3478 27. Applied patches from Google (a) to support the new newline modes and (b) to
3479 advance over multibyte UTF-8 characters in GlobalReplace.
3480
3481 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3482 difference for some implementation of PCRE in some Windows version.
3483
3484 29. Added some extra testing facilities to pcretest:
3485
3486 \q<number> in a data line sets the "match limit" value
3487 \Q<number> in a data line sets the "match recursion limt" value
3488 -S <number> sets the stack size, where <number> is in megabytes
3489
3490 The -S option isn't available for Windows.
3491
3492
3493 Version 6.6 06-Feb-06
3494 ---------------------
3495
3496 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3497 in pcreposix.h. I have copied the definition from pcre.h.
3498
3499 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3500 because pcre.h is no longer a built file.
3501
3502 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3503 not normally included in the compiled code.
3504
3505
3506 Version 6.5 01-Feb-06
3507 ---------------------
3508
3509 1. When using the partial match feature with pcre_dfa_exec(), it was not
3510 anchoring the second and subsequent partial matches at the new starting
3511 point. This could lead to incorrect results. For example, with the pattern
3512 /1234/, partially matching against "123" and then "a4" gave a match.
3513
3514 2. Changes to pcregrep:
3515
3516 (a) All non-match returns from pcre_exec() were being treated as failures
3517 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3518 error message is output. Some extra information is given for the
3519 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3520 probably the only errors that are likely to be caused by users (by
3521 specifying a regex that has nested indefinite repeats, for instance).
3522 If there are more than 20 of these errors, pcregrep is abandoned.
3523
3524 (b) A binary zero was treated as data while matching, but terminated the
3525 output line if it was written out. This has been fixed: binary zeroes
3526 are now no different to any other data bytes.
3527
3528 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3529 used to set a locale for matching. The --locale=xxxx long option has
3530 been added (no short equivalent) to specify a locale explicitly on the
3531 pcregrep command, overriding the environment variables.
3532
3533 (d) When -B was used with -n, some line numbers in the output were one less
3534 than they should have been.
3535
3536 (e) Added the -o (--only-matching) option.
3537
3538 (f) If -A or -C was used with -c (count only), some lines of context were
3539 accidentally printed for the final match.
3540
3541 (g) Added the -H (--with-filename) option.
3542
3543 (h) The combination of options -rh failed to suppress file names for files
3544 that were found from directory arguments.
3545
3546 (i) Added the -D (--devices) and -d (--directories) options.
3547
3548 (j) Added the -F (--fixed-strings) option.
3549
3550 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3551
3552 (l) Added the --colo(u)r option.
3553
3554 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3555 is not present by default.
3556
3557 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3558 items such as (?R) or (?1), when the recursion could match a number of
3559 alternatives. If it matched one of the alternatives, but subsequently,
3560 outside the recursion, there was a failure, the code tried to back up into
3561 the recursion. However, because of the way PCRE is implemented, this is not
3562 possible, and the result was an incorrect result from the match.
3563
3564 In order to prevent this happening, the specification of recursion has
3565 been changed so that all such subpatterns are automatically treated as
3566 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3567
3568 4. I had overlooked the fact that, in some locales, there are characters for
3569 which isalpha() is true but neither isupper() nor islower() are true. In
3570 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3571 and ordfeminine) are like this. This affected the treatment of \w and \W
3572 when they appeared in character classes, but not when they appeared outside
3573 a character class. The bit map for "word" characters is now created
3574 separately from the results of isalnum() instead of just taking it from the
3575 upper, lower, and digit maps. (Plus the underscore character, of course.)
3576
3577 5. The above bug also affected the handling of POSIX character classes such as
3578 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3579 permanent tables. Instead, the bit maps for such a class were previously
3580 created as the appropriate unions of the upper, lower, and digit bitmaps.
3581 Now they are created by subtraction from the [[:word:]] class, which has
3582 its own bitmap.
3583
3584 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3585 It is created by subtracting the vertical space characters (\x09, \x0a,
3586 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3587 subtraction was done in the overall bitmap for a character class, meaning
3588 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3589 be recognized. This bug has been fixed.
3590
3591 7. Patches from the folks at Google:
3592
3593 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3594 real life, but is still worth protecting against".
3595
3596 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3597 regular expressions".
3598
3599 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3600 have it.
3601
3602 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3603 "configure" and the latter not, in order to fix a problem somebody had
3604 with compiling the Arg class on HP-UX.
3605
3606 (e) Improve the error-handling of the C++ wrapper a little bit.
3607
3608 (f) New tests for checking recursion limiting.
3609
3610 8. The pcre_memmove() function, which is used only if the environment does not
3611 have a standard memmove() function (and is therefore rarely compiled),
3612 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3613 returning a result (though PCRE never actually uses the result).
3614
3615 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3616 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3617 returned instead of calling malloc() with an overflowing number that would
3618 most likely cause subsequent chaos.
3619
3620 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3621
3622 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3623 with this option is matched, the nmatch and pmatch options of regexec() are
3624 ignored.
3625
3626 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3627 provided in case anyone wants to the the POSIX interface with UTF-8
3628 strings.
3629
3630 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3631 C++ linking (needed for some HP-UX environments).
3632
3633 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3634 (unused parameter) and in the pcre_printint() function (omitted "default"
3635 switch label when the default is to do nothing).
3636
3637 15. Added some code to make it possible, when PCRE is compiled as a C++
3638 library, to replace subject pointers for pcre_exec() with a smart pointer
3639 class, thus making it possible to process discontinuous strings.
3640
3641 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3642 much the same function. They were added by different people who were trying
3643 to make PCRE easy to compile on non-Unix systems. It has been suggested
3644 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3645 for compiling on Windows systems. I have therefore replaced it with
3646 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3647 defaults to "extern" for C or "extern C" for C++, which works fine on
3648 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3649 SCOPE with something explicit in config.h. In addition:
3650
3651 (a) pcreposix.h still had just "extern" instead of either of these macros;
3652 I have replaced it with PCRE_DATA_SCOPE.
3653
3654 (b) Functions such as _pcre_xclass(), which are internal to the library,
3655 but external in the C sense, all had PCRE_EXPORT in their definitions.
3656 This is apparently wrong for the Windows case, so I have removed it.
3657 (It makes no difference on Unix-like systems.)
3658
3659 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3660 of recursive calls to match(). This is different to MATCH_LIMIT because
3661 that limits the total number of calls to match(), not all of which increase
3662 the depth of recursion. Limiting the recursion depth limits the amount of
3663 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3664 when PCRE is compiled, and changed at run time. A patch from Google adds
3665 this functionality to the C++ interface.
3666
3667 18. Changes to the handling of Unicode character properties:
3668
3669 (a) Updated the table to Unicode 4.1.0.
3670
3671 (b) Recognize characters that are not in the table as "Cn" (undefined).
3672
3673 (c) I revised the way the table is implemented to a much improved format
3674 which includes recognition of ranges. It now supports the ranges that
3675 are defined in UnicodeData.txt, and it also amalgamates other
3676 characters into ranges. This has reduced the number of entries in the
3677 table from around 16,000 to around 3,000, thus reducing its size
3678 considerably. I realized I did not need to use a tree structure after
3679 all - a binary chop search is just as efficient. Having reduced the
3680 number of entries, I extended their size from 6 bytes to 8 bytes to
3681 allow for more data.
3682
3683 (d) Added support for Unicode script names via properties such as \p{Han}.
3684
3685 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3686 matching that character.
3687
3688 20. When matching a repeated Unicode property with a minimum greater than zero,
3689 (for example \pL{2,}), PCRE could look past the end of the subject if it
3690 reached it while seeking the minimum number of characters. This could
3691 happen only if some of the characters were more than one byte long, because
3692 there is a check for at least the minimum number of bytes.
3693
3694 21. Refactored the implementation of \p and \P so as to be more general, to
3695 allow for more different types of property in future. This has changed the
3696 compiled form incompatibly. Anybody with saved compiled patterns that use
3697 \p or \P will have to recompile them.
3698
3699 22. Added "Any" and "L&" to the supported property types.
3700
3701 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3702 but give a compile time error if the value is greater than 0xff.
3703
3704 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3705 accidentally not being installed or uninstalled.
3706
3707 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3708 made were to insert the current release number. This seemed silly, because
3709 it made things harder for people building PCRE on systems that don't run
3710 "configure". I have turned pcre.h into a distributed file, no longer built
3711 by "configure", with the version identification directly included. There is
3712 no longer a pcre.h.in file.
3713
3714 However, this change necessitated a change to the pcre-config script as
3715 well. It is built from pcre-config.in, and one of the substitutions was the
3716 release number. I have updated configure.ac so that ./configure now finds
3717 the release number by grepping pcre.h.
3718
3719 26. Added the ability to run the tests under valgrind.
3720
3721
3722 Version 6.4 05-Sep-05
3723 ---------------------
3724
3725 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3726 "--" to be printed when multiple files were scanned, even when none of the
3727 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3728 consider it to be a bug, and have restored the previous behaviour.
3729
3730 2. A couple of code tidies to get rid of compiler warnings.
3731
3732 3. The pcretest program used to cheat by referring to symbols in the library
3733 whose names begin with _pcre_. These are internal symbols that are not
3734 really supposed to be visible externally, and in some environments it is
3735 possible to suppress them. The cheating is now confined to including
3736 certain files from the library's source, which is a bit cleaner.
3737
3738 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3739 file's purpose clearer.
3740
3741 5. Reorganized pcre_ucp_findchar().
3742
3743
3744 Version 6.3 15-Aug-05
3745 ---------------------
3746
3747 1. The file libpcre.pc.in did not have general read permission in the tarball.
3748
3749 2. There were some problems when building without C++ support:
3750
3751 (a) If C++ support was not built, "make install" and "make test" still
3752 tried to test it.
3753
3754 (b) There were problems when the value of CXX was explicitly set. Some
3755 changes have been made to try to fix these, and ...
3756
3757 (c) --disable-cpp can now be used to explicitly disable C++ support.
3758
3759 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3760 backslash in a target when C++ was disabled. This confuses some
3761 versions of "make", apparently. Using an intermediate variable solves
3762 this. (Same for CPP_LOBJ.)
3763
3764 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3765 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3766 necessary on certain architectures.
3767
3768 4. Added a setting of -export-symbols-regex to the link command to remove
3769 those symbols that are exported in the C sense, but actually are local
3770 within the library, and not documented. Their names all begin with
3771 "_pcre_". This is not a perfect job, because (a) we have to except some
3772 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3773 available (and never for static libraries). I have made a note to try to
3774 find a way round (a) in the future.
3775
3776
3777 Version 6.2 01-Aug-05
3778 ---------------------
3779
3780 1. There was no test for integer overflow of quantifier values. A construction
3781 such as {1111111111111111} would give undefined results. What is worse, if
3782 a minimum quantifier for a parenthesized subpattern overflowed and became
3783 negative, the calculation of the memory size went wrong. This could have
3784 led to memory overwriting.
3785
3786 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3787
3788 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3789 operating environments where this matters.
3790
3791 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3792 PCRE options from within the C++ wrapper.
3793
3794 5. Named capturing subpatterns were not being correctly counted when a pattern
3795 was compiled. This caused two problems: (a) If there were more than 100
3796 such subpatterns, the calculation of the memory needed for the whole
3797 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3798 back references of the form \12, where the number was greater than 9, were
3799 not recognized as back references, even though there were sufficient
3800 previous subpatterns.
3801
3802 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3803 versions of gcc, e.g. 2.95.4.
3804
3805
3806 Version 6.1 21-Jun-05
3807 ---------------------
3808
3809 1. There was one reference to the variable "posix" in pcretest.c that was not
3810 surrounded by "#if !defined NOPOSIX".
3811
3812 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3813 the cross-check on the old pcre_info() function, for the benefit of the
3814 cut-down version of PCRE that is currently imported into Exim.
3815
3816 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3817 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3818 patterns that start (?i) but I don't think that matters. The use of (?i) is
3819 just an example; this all applies to the other options as well.
3820
3821 4. Since libtool seems to echo the compile commands it is issuing, the output
3822 from "make" can be reduced a bit by putting "@" in front of each libtool
3823 compile command.
3824
3825 5. Patch from the folks at Google for configure.in to be a bit more thorough
3826 in checking for a suitable C++ installation before trying to compile the
3827 C++ stuff. This should fix a reported problem when a compiler was present,
3828 but no suitable headers.
3829
3830 6. The man pages all had just "PCRE" as their title. I have changed them to
3831 be the relevant file name. I have also arranged that these names are
3832 retained in the file doc/pcre.txt, which is a concatenation in text format
3833 of all the man pages except the little individual ones for each function.
3834
3835 7. The NON-UNIX-USE file had not been updated for the different set of source
3836 files that come with release 6. I also added a few comments about the C++
3837 wrapper.
3838
3839
3840 Version 6.0 07-Jun-05
3841 ---------------------
3842
3843 1. Some minor internal re-organization to help with my DFA experiments.
3844
3845 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3846 didn't matter for the library itself when fully configured, but did matter
3847 when compiling without UCP support, or within Exim, where the ucp files are
3848 not imported.
3849
3850 3. Refactoring of the library code to split up the various functions into
3851 different source modules. The addition of the new DFA matching code (see
3852 below) to a single monolithic source would have made it really too
3853 unwieldy, quite apart from causing all the code to be include in a
3854 statically linked application, when only some functions are used. This is
3855 relevant even without the DFA addition now that patterns can be compiled in
3856 one application and matched in another.
3857
3858 The downside of splitting up is that there have to be some external
3859 functions and data tables that are used internally in different modules of
3860 the library but which are not part of the API. These have all had their
3861 names changed to start with "_pcre_" so that they are unlikely to clash
3862 with other external names.
3863
3864 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3865 a different (DFA) algorithm. Although it is slower than the original
3866 function, it does have some advantages for certain types of matching
3867 problem.
3868
3869 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3870 including restarting after a partial match.
3871
3872 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3873 defined when compiling for Windows was sent to me. I have put it into the
3874 code, though I have no means of testing or verifying it.
3875
3876 7. Added the pcre_refcount() auxiliary function.
3877
3878 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3879 match before or at the first newline in the subject string. In pcretest,
3880 the /f option on a pattern can be used to set this.
3881
3882 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3883 would behave wrongly. This has been present in PCRE since release 4.0.
3884
3885 10. A number of changes to the pcregrep command:
3886
3887 (a) Refactored how -x works; insert ^(...)$ instead of setting
3888 PCRE_ANCHORED and checking the length, in preparation for adding
3889 something similar for -w.
3890
3891 (b) Added the -w (match as a word) option.
3892
3893 (c) Refactored the way lines are read and buffered so as to have more
3894 than one at a time available.
3895
3896 (d) Implemented a pcregrep test script.
3897
3898 (e) Added the -M (multiline match) option. This allows patterns to match
3899 over several lines of the subject. The buffering ensures that at least
3900 8K, or the rest of the document (whichever is the shorter) is available
3901 for matching (and similarly the previous 8K for lookbehind assertions).
3902
3903 (f) Changed the --help output so that it now says
3904
3905 -w, --word-regex(p)
3906
3907 instead of two lines, one with "regex" and the other with "regexp"
3908 because that confused at least one person since the short forms are the
3909 same. (This required a bit of code, as the output is generated
3910 automatically from a table. It wasn't just a text change.)
3911
3912 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3913 option but starts with a hyphen. Could be a pattern or a path name
3914 starting with a hyphen, for instance.
3915
3916 (h) "-" can be given as a file name to represent stdin.
3917
3918 (i) When file names are being printed, "(standard input)" is used for
3919 the standard input, for compatibility with GNU grep. Previously
3920 "<stdin>" was used.
3921
3922 (j) The option --label=xxx can be used to supply a name to be used for
3923 stdin when file names are being printed. There is no short form.
3924
3925 (k) Re-factored the options decoding logic because we are going to add
3926 two more options that take data. Such options can now be given in four
3927 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3928
3929 (l) Added the -A, -B, and -C options for requesting that lines of context
3930 around matches be printed.
3931
3932 (m) Added the -L option to print the names of files that do not contain
3933 any matching lines, that is, the complement of -l.
3934
3935 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3936 continue to scan other files.
3937
3938 (o) The -s option was incorrectly implemented. For compatibility with other
3939 greps, it now suppresses the error message for a non-existent or non-
3940 accessible file (but not the return code). There is a new option called
3941 -q that suppresses the output of matching lines, which was what -s was
3942 previously doing.
3943
3944 (p) Added --include and --exclude options to specify files for inclusion
3945 and exclusion when recursing.
3946
3947 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3948 Hopefully, it now does.
3949
3950 12. Missing cast in pcre_study().
3951
3952 13. Added an "uninstall" target to the makefile.
3953
3954 14. Replaced "extern" in the function prototypes in Makefile.in with
3955 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3956 world, but is set differently for Windows.
3957
3958 15. Added a second compiling function called pcre_compile2(). The only
3959 difference is that it has an extra argument, which is a pointer to an
3960 integer error code. When there is a compile-time failure, this is set
3961 non-zero, in addition to the error test pointer being set to point to an
3962 error message. The new argument may be NULL if no error number is required
3963 (but then you may as well call pcre_compile(), which is now just a
3964 wrapper). This facility is provided because some applications need a
3965 numeric error indication, but it has also enabled me to tidy up the way
3966 compile-time errors are handled in the POSIX wrapper.
3967
3968 16. Added VPATH=.libs to the makefile; this should help when building with one
3969 prefix path and installing with another. (Or so I'm told by someone who
3970 knows more about this stuff than I do.)
3971
3972 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3973 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3974 match everything, including newlines. This is not POSIX-compatible, but
3975 somebody wanted the feature. From pcretest it can be activated by using
3976 both the P and the s flags.
3977
3978 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3979
3980 19. libpcre.pc was being incorrectly installed as executable.
3981
3982 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3983 it now also looks for '\r' so that it will work unmodified on Windows.
3984
3985 21. Added Google's contributed C++ wrapper to the distribution.
3986
3987 22. Added some untidy missing memory free() calls in pcretest, to keep
3988 Electric Fence happy when testing.
3989
3990
3991
3992 Version 5.0 13-Sep-04
3993 ---------------------
3994
3995 1. Internal change: literal characters are no longer packed up into items
3996 containing multiple characters in a single byte-string. Each character
3997 is now matched using a separate opcode. However, there may be more than one
3998 byte in the character in UTF-8 mode.
3999
4000 2. The pcre_callout_block structure has two new fields: pattern_position and
4001 next_item_length. These contain the offset in the pattern to the next match
4002 item, and its length, respectively.
4003
4004 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4005 insertion of callouts before each pattern item. Added the /C option to
4006 pcretest to make use of this.
4007
4008 4. On the advice of a Windows user, the lines
4009
4010 #if defined(_WIN32) || defined(WIN32)
4011 _setmode( _fileno( stdout ), 0x8000 );
4012 #endif /* defined(_WIN32) || defined(WIN32) */
4013
4014 have been added to the source of pcretest. This apparently does useful
4015 magic in relation to line terminators.
4016
4017 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4018 for the benefit of those environments where the "b" makes a difference.
4019
4020 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4021 to know about it. I have put a hack into configure.in that adds in code
4022 to set GCC=yes if CC=icc. This seems to end up at a point in the
4023 generated configure script that is early enough to affect the setting of
4024 compiler options, which is what is needed, but I have no means of testing
4025 whether it really works. (The user who reported this had patched the
4026 generated configure script, which of course I cannot do.)
4027
4028 LATER: After change 22 below (new libtool files), the configure script
4029 seems to know about icc (and also ecc). Therefore, I have commented out
4030 this hack in configure.in.
4031
4032 7. Added support for pkg-config (2 patches were sent in).
4033
4034 8. Negated POSIX character classes that used a combination of internal tables
4035 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4036 [[:^ascii]]. Typically, they would match almost any characters. The other
4037 POSIX classes were not broken in this way.
4038
4039 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4040 to find the match, as PCRE was deluded into thinking that the match had to
4041 start at the start point or following a newline. The same bug applied to
4042 patterns with negative forward assertions or any backward assertions
4043 preceding ".*" at the start, unless the pattern required a fixed first
4044 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4045
4046 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4047 starting at the last subject character, bytes beyond the end of the subject
4048 string were read.
4049
4050 11. Renamed the variable "class" as "classbits" to make life easier for C++
4051 users. (Previously there was a macro definition, but it apparently wasn't
4052 enough.)
4053
4054 12. Added the new field "tables" to the extra data so that tables can be passed
4055 in at exec time, or the internal tables can be re-selected. This allows
4056 a compiled regex to be saved and re-used at a later time by a different
4057 program that might have everything at different addresses.
4058
4059 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4060 -R library as well as a -L library.
4061
4062 14. The debugging options of pcretest (-d on the command line or D on a
4063 pattern) showed incorrect output for anything following an extended class
4064 that contained multibyte characters and which was followed by a quantifier.
4065
4066 15. Added optional support for general category Unicode character properties
4067 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4068 support. It adds about 90K to the size of the library. The meanings of the
4069 inbuilt class escapes such as \d and \s have NOT been changed.
4070
4071 16. Updated pcredemo.c to include calls to free() to release the memory for the
4072 compiled pattern.
4073
4074 17. The generated file chartables.c was being created in the source directory
4075 instead of in the building directory. This caused the build to fail if the
4076 source directory was different from the building directory, and was
4077 read-only.
4078
4079 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4080 file. No doubt somebody will tell me if they don't make sense... Also added
4081 Dan Mooney's comments about building on OpenVMS.
4082
4083 19. Added support for partial matching via the PCRE_PARTIAL option for
4084 pcre_exec() and the \P data escape in pcretest.
4085
4086 20. Extended pcretest with 3 new pattern features:
4087
4088 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4089 write the compiled pattern to the file whose name is "rest-of-line".
4090 This is a straight binary dump of the data, with the saved pointer to
4091 the character tables forced to be NULL. The study data, if any, is
4092 written too. After writing, pcretest reads a new pattern.
4093
4094 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4095 compiled pattern from the given file. There must not be any
4096 occurrences of "<" in the file name (pretty unlikely); if there are,
4097 pcretest will instead treat the initial "<" as a pattern delimiter.
4098 After reading in the pattern, pcretest goes on to read data lines as
4099 usual.
4100
4101 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4102 and 16-bit fields in a compiled pattern, to simulate a pattern that
4103 was compiled on a host of opposite endianness.
4104
4105 21. The pcre-exec() function can now cope with patterns that were compiled on
4106 hosts of opposite endianness, with this restriction:
4107
4108 As for any compiled expression that is saved and used later, the tables
4109 pointer field cannot be preserved; the extra_data field in the arguments
4110 to pcre_exec() should be used to pass in a tables address if a value
4111 other than the default internal tables were used at compile time.
4112
4113 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4114 now diagnosed as an error. Previously, most of the time, a negative number
4115 would have been treated as zero, but if in addition "ovector" was passed as
4116 NULL, a crash could occur.
4117
4118 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4119 new versions from the libtool 1.5 distribution (the last one is a copy of
4120 a file called libtool.m4). This seems to have fixed the need to patch
4121 "configure" to support Darwin 1.3 (which I used to do). However, I still
4122 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4123 workstation).
4124
4125 24. Changed the PCRE licence to be the more standard "BSD" licence.
4126
4127
4128 Version 4.5 01-Dec-03
4129 ---------------------
4130
4131 1. There has been some re-arrangement of the code for the match() function so
4132 that it can be compiled in a version that does not call itself recursively.
4133 Instead, it keeps those local variables that need separate instances for
4134 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4135 needs to "recurse". Keeping track of where control must go is done by means
4136 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4137 hide most of the details from the main code, and operates only if
4138 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4139 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4140 operating.
4141
4142 To make it easier for callers to provide specially tailored get/free
4143 functions for this usage, two new functions, pcre_stack_malloc, and
4144 pcre_stack_free, are used. They are always called in strict stacking order,
4145 and the size of block requested is always the same.
4146
4147 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4148 PCRE has been compiled to use the stack or the heap for recursion. The
4149 -C option of pcretest uses this to show which version is compiled.
4150
4151 A new data escape \S, is added to pcretest; it causes the amounts of store
4152 obtained and freed by both kinds of malloc/free at match time to be added
4153 to the output.
4154
4155 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4156 what's available on my current Linux desktop machine.
4157
4158 3. When matching a UTF-8 string, the test for a valid string at the start has
4159 been extended. If start_offset is not zero, PCRE now checks that it points
4160 to a byte that is the start of a UTF-8 character. If not, it returns
4161 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4162 this is necessary because there may be backward assertions in the pattern.
4163 When matching the same subject several times, it may save resources to use
4164 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4165
4166 4. The code for checking the validity of UTF-8 strings has been tightened so
4167 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4168 containing "overlong sequences".
4169
4170 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4171 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4172 should have been "&", but it just so happened that all the cases this let
4173 through by mistake were picked up later in the function.
4174
4175 6. I had used a variable called "isblank" - this is a C99 function, causing
4176 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4177
4178 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4179 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4180 so the version is shown; (c) stop "make test" echoing "./RunTest".
4181
4182 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4183
4184 9. The prototype for memmove() for systems that don't have it was using
4185 size_t, but the inclusion of the header that defines size_t was later. I've
4186 moved the #includes for the C headers earlier to avoid this.
4187
4188 10. Added some adjustments to the code to make it easier to compiler on certain
4189 special systems:
4190
4191 (a) Some "const" qualifiers were missing.
4192 (b) Added the macro EXPORT before all exported functions; by default this
4193 is defined to be empty.
4194 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4195 that it reads its output file name as an argument instead of writing
4196 to the standard output and assuming this can be redirected.
4197
4198 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4199 class containing characters with values greater than 255, PCRE compilation
4200 went into a loop.
4201
4202 12. A recursive reference to a subpattern that was within another subpattern
4203 that had a minimum quantifier of zero caused PCRE to crash. For example,
4204 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4205 recursion. If the recursively-called subpattern itself had a zero repeat,
4206 that was OK.
4207
4208 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4209 buffer into which it was copied (for escape processing) was still set at
4210 1024, so long lines caused crashes.
4211
4212 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4213 "internal error: code overflow...". This applied to any character class
4214 that was followed by a possessive quantifier.
4215
4216 15. Modified the Makefile to add libpcre.la as a prerequisite for
4217 libpcreposix.la because I was told this is needed for a parallel build to
4218 work.
4219
4220 16. If a pattern that contained .* following optional items at the start was
4221 studied, the wrong optimizing data was generated, leading to matching
4222 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4223 matching string must start with a or b or c. The correct conclusion for
4224 this pattern is that a match can start with any character.
4225
4226
4227 Version 4.4 13-Aug-03
4228 ---------------------
4229
4230 1. In UTF-8 mode, a character class containing characters with values between
4231 127 and 255 was not handled correctly if the compiled pattern was studied.
4232 In fixing this, I have also improved the studying algorithm for such
4233 classes (slightly).
4234
4235 2. Three internal functions had redundant arguments passed to them. Removal
4236 might give a very teeny performance improvement.
4237
4238 3. Documentation bug: the value of the capture_top field in a callout is *one
4239 more than* the number of the hightest numbered captured substring.
4240
4241 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4242 in incorrectly linking with a previously installed version. They now link
4243 explicitly with libpcre.la.
4244
4245 5. configure.in no longer needs to recognize Cygwin specially.
4246
4247 6. A problem in pcre.in for Windows platforms is fixed.
4248
4249 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4250 pcretest, it used to include the size of the study block as part of its
4251 output. Unfortunately, the structure contains a field that has a different
4252 size on different hardware architectures. This meant that the tests that
4253 showed this size failed. As the block is currently always of a fixed size,
4254 this information isn't actually particularly useful in pcretest output, so
4255 I have just removed it.
4256
4257 8. Three pre-processor statements accidentally did not start in column 1.
4258 Sadly, there are *still* compilers around that complain, even though
4259 standard C has not required this for well over a decade. Sigh.
4260
4261 9. In pcretest, the code for checking callouts passed small integers in the
4262 callout_data field, which is a void * field. However, some picky compilers
4263 complained about the casts involved for this on 64-bit systems. Now
4264 pcretest passes the address of the small integer instead, which should get
4265 rid of the warnings.
4266
4267 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4268 both compile and run time, and gives an error if an invalid UTF-8 sequence
4269 is found. There is a option for disabling this check in cases where the
4270 string is known to be correct and/or the maximum performance is wanted.
4271
4272 11. In response to a bug report, I changed one line in Makefile.in from
4273
4274 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4275 to
4276 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4277
4278 to look similar to other lines, but I have no way of telling whether this
4279 is the right thing to do, as I do not use Windows. No doubt I'll get told
4280 if it's wrong...
4281
4282
4283 Version 4.3 21-May-03
4284 ---------------------
4285
4286 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4287 Makefile.
4288
4289 2. Some refactoring to improve the quality of the code:
4290
4291 (i) The utf8_table... variables are now declared "const".
4292
4293 (ii) The code for \cx, which used the "case flipping" table to upper case
4294 lower case letters, now just substracts 32. This is ASCII-specific,
4295 but the whole concept of \cx is ASCII-specific, so it seems
4296 reasonable.
4297
4298 (iii) PCRE was using its character types table to recognize decimal and
4299 hexadecimal digits in the pattern. This is silly, because it handles
4300 only 0-9, a-f, and A-F, but the character types table is locale-
4301 specific, which means strange things might happen. A private
4302 table is now used for this - though it costs 256 bytes, a table is
4303 much faster than multiple explicit tests. Of course, the standard
4304 character types table is still used for matching digits in subject
4305 strings against \d.
4306
4307 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4308 ending in _t are). So I've renamed it as ESC_tee.
4309
4310 3. The first argument for regexec() in the POSIX wrapper should have been
4311 defined as "const".
4312
4313 4. Changed pcretest to use malloc() for its buffers so that they can be
4314 Electric Fenced for debugging.
4315
4316 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4317 to read one or more bytes before the start of the subject string. Often this
4318 had no effect on PCRE's behaviour, but in some circumstances it could
4319 provoke a segmentation fault.
4320
4321 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4322 to try to read one or more bytes before the start of the subject string.
4323
4324 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4325 UTF-8 support could misbehave in various ways if the subject string
4326 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4327 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4328 back over UTF-8 characters.)
4329
4330
4331 Version 4.2 14-Apr-03
4332 ---------------------
4333
4334 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4335
4336 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4337 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4338 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4339 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4340 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4341 and BUILD_EXEEXT
4342 Note: automatic setting of the BUILD variables is not yet working
4343 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4344 compile-time but not at link-time
4345 [LINK]: use for linking executables only
4346 make different versions for Windows and non-Windows
4347 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4348 libraries
4349 [LINK_FOR_BUILD]: new variable
4350 [OBJEXT]: use throughout
4351 [EXEEXT]: use throughout
4352 <winshared>: new target
4353 <wininstall>: new target
4354 <dftables.o>: use native compiler
4355 <dftables>: use native linker
4356 <install>: handle Windows platform correctly
4357 <clean>: ditto
4358 <check>: ditto
4359 copy DLL to top builddir before testing
4360
4361 As part of these changes, -no-undefined was removed again. This was reported
4362 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4363 in any case.
4364
4365 3. Some tidies to get rid of compiler warnings:
4366
4367 . In the match_data structure, match_limit was an unsigned long int, whereas
4368 match_call_count was an int. I've made them both unsigned long ints.
4369
4370 . In pcretest the fact that a const uschar * doesn't automatically cast to
4371 a void * provoked a warning.
4372
4373 . Turning on some more compiler warnings threw up some "shadow" variables
4374 and a few more missing casts.
4375
4376 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4377 option, a class that contained a single character with a value between 128
4378 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4379
4380 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4381 option, a class that contained several characters, but with at least one
4382 whose value was between 128 and 255 caused PCRE to crash.
4383
4384
4385 Version 4.1 12-Mar-03
4386 ---------------------
4387
4388 1. Compiling with gcc -pedantic found a couple of places where casts were
4389 needed, and a string in dftables.c that was longer than standard compilers are
4390 required to support.
4391
4392 2. Compiling with Sun's compiler found a few more places where the code could
4393 be tidied up in order to avoid warnings.
4394
4395 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4396 first of these names is deprecated in the latest Autoconf in favour of the name
4397 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4398 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4399 analogy I have changed it to CFLAGS_FOR_BUILD.
4400
4401 4. Added -no-undefined to the linking command in the Makefile, because this is
4402 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4403 linking step for the pcreposix library.
4404
4405 5. PCRE was failing to diagnose the case of two named groups with the same
4406 name.
4407
4408 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4409 literal character that is needed in the subject for a match, and scans along to
4410 ensure that it is present before embarking on the full matching process. This
4411 saves time in cases of nested unlimited repeats that are never going to match.
4412 Problem: the scan can take a lot of time if the subject is very long (e.g.
4413 megabytes), thus penalizing straightforward matches. It is now done only if the
4414 amount of subject to be scanned is less than 1000 bytes.
4415
4416 7. A lesser problem with the same optimization is that it was recording the
4417 first character of an anchored pattern as "needed", thus provoking a search
4418 right along the subject, even when the first match of the pattern was going to
4419 fail. The "needed" character is now not set for anchored patterns, unless it
4420 follows something in the pattern that is of non-fixed length. Thus, it still
4421 fulfils its original purpose of finding quick non-matches in cases of nested
4422 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4423
4424
4425 Version 4.0 17-Feb-03
4426 ---------------------
4427
4428 1. If a comment in an extended regex that started immediately after a meta-item
4429 extended to the end of string, PCRE compiled incorrect data. This could lead to
4430 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4431
4432 2. Moved to autoconf 2.53 and libtool 1.4.2.
4433
4434 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4435 the special perltest8 script is no longer needed - all the tests can be run
4436 from a single perltest script.
4437
4438 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4439 by \s. It has now been removed in PCRE. This means it isn't recognized as
4440 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4441 class [:space:] *does* include VT, thereby creating a mess.
4442
4443 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4444 space and tab.
4445
4446 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4447 its new features into the main test script, reducing the number of scripts.
4448
4449 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4450 were backward compatible, and made the (?i) apply to the whole pattern, as if
4451 /i were given. Now it behaves more logically, and applies the option setting
4452 only to what follows. PCRE has been changed to follow suit. However, if it
4453 finds options settings right at the start of the pattern, it extracts them into
4454 the global options, as before. Thus, they show up in the info data.
4455
4456 8. Added support for the \Q...\E escape sequence. Characters in between are
4457 treated as literals. This is slightly different from Perl in that $ and @ are
4458 also handled as literals inside the quotes. In Perl, they will cause variable
4459 interpolation. Note the following examples:
4460
4461 Pattern PCRE matches Perl matches
4462
4463 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4464 \Qabc\$xyz\E abc\$xyz abc\$xyz
4465 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4466
4467 For compatibility with Perl, \Q...\E sequences are recognized inside character
4468 classes as well as outside them.
4469
4470 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4471 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4472 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4473 signed/unsigned warnings.
4474
4475 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4476 option for pcretest, so I've replaced it by a simple function that does just
4477 that job.
4478
4479 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4480 "pcregrep -".
4481
4482 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4483 Java package. This provides some syntactic sugar for simple cases of what my
4484 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4485 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4486 item, you can use this simplified notation. Note that only makes sense with
4487 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4488 greediness, whatever the setting of the PCRE_UNGREEDY option.
4489
4490 13. A change of greediness default within a pattern was not taking effect at
4491 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4492 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4493 was abstracted outside.
4494
4495 14. PCRE now supports the \G assertion. It is true when the current matching
4496 position is at the start point of the match. This differs from \A when the
4497 starting offset is non-zero. Used with the /g option of pcretest (or similar
4498 code), it works in the same way as it does for Perl's /g option. If all
4499 alternatives of a regex begin with \G, the expression is anchored to the start
4500 match position, and the "anchored" flag is set in the compiled expression.
4501
4502 15. Some bugs concerning the handling of certain option changes within patterns
4503 have been fixed. These applied to options other than (?ims). For example,
4504 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4505 been the other way round. Some of this was related to change 7 above.
4506
4507 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4508 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4509 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4510 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4511
4512 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4513 mode. Unlike ".", it always matches newline, whatever the setting of
4514 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4515 assertions. Perl allows it, but it doesn't (in general) work because it can't
4516 calculate the length of the lookbehind. At least, that's the case for Perl
4517 5.8.0 - I've been told they are going to document that it doesn't work in
4518 future.
4519
4520 18. Added an error diagnosis for escapes that PCRE does not support: these are
4521 \L, \l, \N, \P, \p, \U, \u, and \X.
4522
4523 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4524 reading past the end of the pattern in cases such as /[abcd/.
4525
4526 20. PCRE was getting more memory than necessary for patterns with classes that
4527 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4528
4529 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4530 compiling PCRE for use with Virtual Pascal.
4531
4532 22. Small fix to the Makefile to make it work properly if the build is done
4533 outside the source tree.
4534
4535 23. Added a new extension: a condition to go with recursion. If a conditional
4536 subpattern starts with (?(R) the "true" branch is used if recursion has
4537 happened, whereas the "false" branch is used only at the top level.
4538
4539 24. When there was a very long string of literal characters (over 255 bytes
4540 without UTF support, over 250 bytes with UTF support), the computation of how
4541 much memory was required could be incorrect, leading to segfaults or other
4542 strange effects.
4543
4544 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4545 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4546 there was a subsequent back reference to those brackets. This meant that, for
4547 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4548 possible to check for precisely this case. All we can do is abandon the
4549 optimization if .* occurs inside capturing brackets when there are any back
4550 references whatsoever. (See below for a better fix that came later.)
4551
4552 26. The handling of the optimization for finding the first character of a
4553 non-anchored pattern, and for finding a character that is required later in the
4554 match were failing in some cases. This didn't break the matching; it just
4555 failed to optimize when it could. The way this is done has been re-implemented.
4556
4557 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4558
4559 28. Added a new feature that provides some of the functionality that Perl
4560 provides with (?{...}). The facility is termed a "callout". The way it is done
4561 in PCRE is for the caller to provide an optional function, by setting
4562 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4563 global variable. By default it is unset, which disables all calling out. To get
4564 the function called, the regex must include (?C) at appropriate points. This
4565 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4566 This provides a means of identifying different callout points. When PCRE
4567 reaches such a point in the regex, if pcre_callout has been set, the external
4568 function is called. It is provided with data in a structure called
4569 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4570 matching continues; if it returns a non-zero value, the match at the current
4571 point fails. However, backtracking will occur if possible. [This was changed
4572 later and other features added - see item 49 below.]
4573
4574 29. pcretest is upgraded to test the callout functionality. It provides a
4575 callout function that displays information. By default, it shows the start of
4576 the match and the current position in the text. There are some new data escapes
4577 to vary what happens:
4578
4579 \C+ in addition, show current contents of captured substrings
4580 \C- do not supply a callout function
4581 \C!n return 1 when callout number n is reached
4582 \C!n!m return 1 when callout number n is reached for the mth time
4583
4584 30. If pcregrep was called with the -l option and just a single file name, it
4585 output "<stdin>" if a match was found, instead of the file name.
4586
4587 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4588 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4589 pcre_exec(). This saves a malloc/free per call. The default value of
4590 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4591 when configuring.
4592
4593 32. The default maximum size of a compiled pattern is 64K. There have been a
4594 few cases of people hitting this limit. The code now uses macros to handle the
4595 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4596 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4597 configuring. Tests 2 and 5 work only with 2-byte links because they output
4598 debugging information about compiled patterns.
4599
4600 33. Internal code re-arrangements:
4601
4602 (a) Moved the debugging function for printing out a compiled regex into
4603 its own source file (printint.c) and used #include to pull it into
4604 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4605 separate copies.
4606
4607 (b) Defined the list of op-code names for debugging as a macro in
4608 internal.h so that it is next to the definition of the opcodes.
4609
4610 (c) Defined a table of op-code lengths for simpler skipping along compiled
4611 code. This is again a macro in internal.h so that it is next to the
4612 definition of the opcodes.
4613
4614 34. Added support for recursive calls to individual subpatterns, along the
4615 lines of Robin Houston's patch (but implemented somewhat differently).
4616
4617 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4618 allow it to read and process whole directories in Win32. This code was
4619 contributed by Lionel Fourquaux; it has not been tested by me.
4620
4621 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4622 used to name a group. Names consist of alphanumerics and underscores, and must
4623 be unique. Back references use the syntax (?P=name) and recursive calls use
4624 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4625 numbers. The function pcre_fullinfo() can be used after compilation to extract
4626 a name/number map. There are three relevant calls:
4627
4628 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4629 PCRE_INFO_NAMECOUNT yields the number of entries
4630 PCRE_INFO_NAMETABLE yields a pointer to the map.
4631
4632 The map is a vector of fixed-size entries. The size of each entry depends on
4633 the length of the longest name used. The first two bytes of each entry are the
4634 group number, most significant byte first. There follows the corresponding
4635 name, zero terminated. The names are in alphabetical order.
4636
4637 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4638 case instead of 255. Making it the same both with and without UTF-8 support
4639 means that the same test output works with both.
4640
4641 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4642 calling malloc() with a zero argument.
4643
4644 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4645 optimization. I've improved things by keeping a bitmap of backreferences with
4646 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4647 fact referenced, the optimization can be applied. It is unlikely that a
4648 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4649 the match to follow \n) will appear inside brackets with a number greater than
4650 31, but if it does, any back reference > 31 suppresses the optimization.
4651
4652 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4653 of disabling numbered capturing parentheses. Any opening parenthesis that is
4654 not followed by ? behaves as if it were followed by ?: but named parentheses
4655 can still be used for capturing (and they will acquire numbers in the usual
4656 way).
4657
4658 41. Redesigned the return codes from the match() function into yes/no/error so
4659 that errors can be passed back from deep inside the nested calls. A malloc
4660 failure while inside a recursive subpattern call now causes the
4661 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4662
4663 42. It is now possible to set a limit on the number of times the match()
4664 function is called in a call to pcre_exec(). This facility makes it possible to
4665 limit the amount of recursion and backtracking, though not in a directly
4666 obvious way, because the match() function is used in a number of different
4667 circumstances. The count starts from zero for each position in the subject
4668 string (for non-anchored patterns). The default limit is, for compatibility, a
4669 large number, namely 10 000 000. You can change this in two ways:
4670
4671 (a) When configuring PCRE before making, you can use --with-match-limit=n
4672 to set a default value for the compiled library.
4673
4674 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4675 a different value is set. See 45 below.
4676
4677 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4678
4679 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4680 of things that can be changed at compile time. The first argument specifies
4681 what is wanted and the second points to where the information is to be placed.
4682 The current list of available information is:
4683
4684 PCRE_CONFIG_UTF8
4685
4686 The output is an integer that is set to one if UTF-8 support is available;
4687 otherwise it is set to zero.
4688
4689 PCRE_CONFIG_NEWLINE
4690
4691 The output is an integer that it set to the value of the code that is used for
4692 newline. It is either LF (10) or CR (13).
4693
4694 PCRE_CONFIG_LINK_SIZE
4695
4696 The output is an integer that contains the number of bytes used for internal
4697 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4698
4699 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4700
4701 The output is an integer that contains the threshold above which the POSIX
4702 interface uses malloc() for output vectors. See item 31 above.
4703
4704 PCRE_CONFIG_MATCH_LIMIT
4705
4706 The output is an unsigned integer that contains the default limit of the number
4707 of match() calls in a pcre_exec() execution. See 42 above.
4708
4709 44. pcretest has been upgraded by the addition of the -C option. This causes it
4710 to extract all the available output from the new pcre_config() function, and to
4711 output it. The program then exits immediately.
4712
4713 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4714 order to support additional features. One way would have been to define
4715 pcre_exec2() (for example) with extra arguments, but this would not have been
4716 extensible, and would also have required all calls to the original function to
4717 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4718 is used for passing in "extra" data from pcre_study().
4719
4720 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4721 contains the following fields:
4722
4723 flags a bitmap indicating which of the following fields are set
4724 study_data opaque data from pcre_study()
4725 match_limit a way of specifying a limit on match() calls for a specific
4726 call to pcre_exec()
4727 callout_data data for callouts (see 49 below)
4728
4729 The flag bits are also defined in pcre.h, and are
4730
4731 PCRE_EXTRA_STUDY_DATA
4732 PCRE_EXTRA_MATCH_LIMIT
4733 PCRE_EXTRA_CALLOUT_DATA
4734
4735 The pcre_study() function now returns one of these new pcre_extra blocks, with
4736 the actual study data pointed to by the study_data field, and the
4737 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4738 before. That is, this change is entirely upwards-compatible and requires no
4739 change to existing code.
4740
4741 If you want to pass in additional data to pcre_exec(), you can either place it
4742 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4743 block.
4744
4745 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4746 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4747 times with different match limits, until it finds the minimum value needed for
4748 pcre_exec() to complete. The value is then output. This can be instructive; for
4749 most simple matches the number is quite small, but for pathological cases it
4750 gets very large very quickly.
4751
4752 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4753 returns the size of the data block pointed to by the study_data field in a
4754 pcre_extra block, that is, the value that was passed as the argument to
4755 pcre_malloc() when PCRE was getting memory in which to place the information
4756 created by pcre_study(). The fourth argument should point to a size_t variable.
4757 pcretest has been extended so that this information is shown after a successful
4758 pcre_study() call when information about the compiled regex is being displayed.
4759
4760 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4761 because what follows is always an absolute path. (Later: it turns out that this
4762 is more than cosmetic for MinGW, because it doesn't like empty path
4763 components.)
4764
4765 49. Some changes have been made to the callout feature (see 28 above):
4766
4767 (i) A callout function now has three choices for what it returns:
4768
4769 0 => success, carry on matching
4770 > 0 => failure at this point, but backtrack if possible
4771 < 0 => serious error, return this value from pcre_exec()
4772
4773 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4774 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4775 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4776 use by callout functions. It will never be used by PCRE itself.
4777
4778 (ii) The pcre_extra structure (see 45 above) has a void * field called
4779 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4780 pcre_callout_block structure has a field of the same name. The contents of
4781 the field passed in the pcre_extra structure are passed to the callout
4782 function in the corresponding field in the callout block. This makes it
4783 easier to use the same callout-containing regex from multiple threads. For
4784 testing, the pcretest program has a new data escape
4785
4786 \C*n pass the number n (may be negative) as callout_data
4787
4788 If the callout function in pcretest receives a non-zero value as
4789 callout_data, it returns that value.
4790
4791 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4792 there were some redundant $(CFLAGS) in commands that are now specified as
4793 $(LINK), which already includes $(CFLAGS).
4794
4795 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4796 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4797 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4798 one-byte characters throughout. Note that case-insensitive matching applies
4799 only to characters whose values are less than 256. PCRE doesn't support the
4800 notion of cases for higher-valued characters.
4801
4802 (i) A character class whose characters are all within 0-255 is handled as
4803 a bit map, and the map is inverted for negative classes. Previously, a
4804 character > 255 always failed to match such a class; however it should
4805 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4806
4807 (ii) A negated character class with a single character < 255 is coded as
4808 "not this character" (OP_NOT). This wasn't working properly when the test
4809 character was multibyte, either singly or repeated.
4810
4811 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4812 mode, for example: \x{100}{2,3}.
4813
4814 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4815 singly or repeated) now correctly test multibyte characters. However,
4816 PCRE doesn't recognize any characters with values greater than 255 as
4817 digits, spaces, or word characters. Such characters always match \D, \S,
4818 and \W, and never match \d, \s, or \w.
4819
4820 (v) Classes may now contain characters and character ranges with values
4821 greater than 255. For example: [ab\x{100}-\x{400}].
4822
4823 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4824 PCRE in UTF-8 mode.
4825
4826 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4827 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4828 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4829 value.)
4830
4831 53. The single man page has become too large. I have therefore split it up into
4832 a number of separate man pages. These also give rise to individual HTML pages;
4833 these are now put in a separate directory, and there is an index.html page that
4834 lists them all. Some hyperlinking between the pages has been installed.
4835
4836 54. Added convenience functions for handling named capturing parentheses.
4837
4838 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4839 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4840 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4841 are faulted.
4842
4843 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4844 calling configure. These values are used when compiling the dftables.c program
4845 which is run to generate the source of the default character tables. They
4846 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4847 you will need to set these values.
4848
4849 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4850
4851
4852 Version 3.9 02-Jan-02
4853 ---------------------
4854
4855 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4856
4857 2. If --disable-static was given, the building process failed when trying to
4858 build pcretest and pcregrep. (For some reason it was using libtool to compile
4859 them, which is not right, as they aren't part of the library.)
4860
4861
4862 Version 3.8 18-Dec-01
4863 ---------------------
4864
4865 1. The experimental UTF-8 code was completely screwed up. It was packing the
4866 bytes in the wrong order. How dumb can you get?
4867
4868
4869 Version 3.7 29-Oct-01
4870 ---------------------
4871
4872 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4873 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4874 this didn't happen under Solaris 8, where I normally test things.
4875
4876 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4877 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4878 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4879 this fix an hour or so after the initial 3.7 release.)
4880
4881
4882 Version 3.6 23-Oct-01
4883 ---------------------
4884
4885 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4886 offsets passed as NULL with zero offset count.
4887
4888 2. The config.guess and config.sub files had not been updated when I moved to
4889 the latest autoconf.
4890
4891
4892 Version 3.5 15-Aug-01
4893 ---------------------
4894
4895 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4896 had been forgotten.
4897
4898 2. By using declared but undefined structures, we can avoid using "void"
4899 definitions in pcre.h while keeping the internal definitions of the structures
4900 private.
4901
4902 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4903 user point of view, this means that both static and shared libraries are built
4904 by default, but this can be individually controlled. More of the work of
4905 handling this static/shared cases is now inside libtool instead of PCRE's make
4906 file.
4907
4908 4. The pcretest utility is now installed along with pcregrep because it is
4909 useful for users (to test regexs) and by doing this, it automatically gets
4910 relinked by libtool. The documentation has been turned into a man page, so
4911 there are now .1, .txt, and .html versions in /doc.
4912
4913 5. Upgrades to pcregrep:
4914 (i) Added long-form option names like gnu grep.
4915 (ii) Added --help to list all options with an explanatory phrase.
4916 (iii) Added -r, --recursive to recurse into sub-directories.
4917 (iv) Added -f, --file to read patterns from a file.
4918
4919 6. pcre_exec() was referring to its "code" argument before testing that
4920 argument for NULL (and giving an error if it was NULL).
4921
4922 7. Upgraded Makefile.in to allow for compiling in a different directory from
4923 the source directory.
4924
4925 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4926 options bits, the pointer it was passed was to an int instead of to an unsigned
4927 long int. This mattered only on 64-bit systems.
4928
4929 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4930 generated) instead of pcre.in, which it its source. Also made the same change
4931 in several of the .c files.
4932
4933 10. A new release of gcc defines printf() as a macro, which broke pcretest
4934 because it had an ifdef in the middle of a string argument for printf(). Fixed
4935 by using separate calls to printf().
4936
4937 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4938 script, to force use of CR or LF instead of \n in the source. On non-Unix
4939 systems, the value can be set in config.h.
4940
4941 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4942 absolute limit. Changed the text of the error message to make this clear, and
4943 likewise updated the man page.
4944
4945 13. The limit of 99 on the number of capturing subpatterns has been removed.
4946 The new limit is 65535, which I hope will not be a "real" limit.
4947
4948
4949 Version 3.4 22-Aug-00
4950 ---------------------
4951
4952 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4953
4954 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4955
4956
4957 Version 3.3 01-Aug-00
4958 ---------------------
4959
4960 1. If an octal character was given, but the value was greater than \377, it
4961 was not getting masked to the least significant bits, as documented. This could
4962 lead to crashes in some systems.
4963
4964 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4965 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4966
4967 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4968 These just pass their arguments on to (pcre_free)(), but they are provided
4969 because some uses of PCRE bind it to non-C systems that can call its functions,
4970 but cannot call free() or pcre_free() directly.
4971
4972 4. Add "make test" as a synonym for "make check". Corrected some comments in
4973 the Makefile.
4974
4975 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4976 Makefile.
4977
4978 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4979 command called pgrep for grepping around the active processes.
4980
4981 7. Added the beginnings of support for UTF-8 character strings.
4982
4983 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4984 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4985 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4986 out for the ar command.)
4987
4988
4989 Version 3.2 12-May-00
4990 ---------------------
4991
4992 This is purely a bug fixing release.
4993
4994 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4995 of ZA. This was just one example of several cases that could provoke this bug,
4996 which was introduced by change 9 of version 2.00. The code for breaking
4997 infinite loops after an iteration that matches an empty string was't working
4998 correctly.
4999
5000 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5001 when matched against abbab (for example). After matching an empty string, it
5002 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5003 caused it to match further down the string than it should.
5004
5005 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5006 was there because it doesn't seem to be needed, and it causes trouble on some
5007 systems, as it is not a Standard C header. It has been removed.
5008
5009 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5010 were reported on the Macintosh. The changes were from
5011
5012 while ((c = *(++ptr)) != 0 && c != '\n');
5013 to
5014 while ((c = *(++ptr)) != 0 && c != '\n') ;
5015
5016 Totally extraordinary, but if that's what it takes...
5017
5018 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5019 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5020 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5021 assumes the way PCRE uses memmove() (always moving upwards).
5022
5023 6. PCRE is being used in one environment where strchr() is not available. There
5024 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5025 faster code anyway.
5026
5027
5028 Version 3.1 09-Feb-00
5029 ---------------------
5030
5031 The only change in this release is the fixing of some bugs in Makefile.in for
5032 the "install" target:
5033
5034 (1) It was failing to install pcreposix.h.
5035
5036 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5037
5038
5039 Version 3.0 01-Feb-00
5040 ---------------------
5041
5042 1. Add support for the /+ modifier to perltest (to output $` like it does in
5043 pcretest).
5044
5045 2. Add support for the /g modifier to perltest.
5046
5047 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5048 matches null strings.
5049
5050 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5051 pattern. Perl treats empty patterns specially - it reuses the most recent
5052 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5053 effect.
5054
5055 5. The POSIX interface was broken in that it was just handing over the POSIX
5056 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5057 required a bigger vector, with some working space on the end. This means that
5058 the POSIX wrapper now has to get and free some memory, and copy the results.
5059
5060 6. Added some simple autoconf support, placing the test data and the
5061 documentation in separate directories, re-organizing some of the
5062 information files, and making it build pcre-config (a GNU standard). Also added
5063 libtool support for building PCRE as a shared library, which is now the
5064 default.
5065
5066 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5067 09 are not valid octal constants. Single digits will be used for minor values
5068 less than 10.
5069
5070 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5071 existing programs that set these in the POSIX interface can use PCRE without
5072 modification.
5073
5074 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5075 return all that pcre_info() returns, plus additional data. The pcre_info()
5076 function is retained for compatibility, but is considered to be obsolete.
5077
5078 10. Added experimental recursion feature (?R) to handle one common case that
5079 Perl 5.6 will be able to do with (?p{...}).
5080
5081 11. Added support for POSIX character classes like [:alpha:], which Perl is
5082 adopting.
5083
5084
5085 Version 2.08 31-Aug-99
5086 ----------------------
5087
5088 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5089 trying to match at the startoffset position, but instead was moving forward to
5090 the next newline as if a previous match had failed.
5091
5092 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5093 and could get into a loop if a null string was matched other than at the start
5094 of the subject.
5095
5096 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5097 be distinguished at compile time, and for completeness also added PCRE_DATE.
5098
5099 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5100 in GnuWin32 environments.
5101
5102
5103 Version 2.07 29-Jul-99
5104 ----------------------
5105
5106 1. The documentation is now supplied in plain text form and HTML as well as in
5107 the form of man page sources.
5108
5109 2. C++ compilers don't like assigning (void *) values to other pointer types.
5110 In particular this affects malloc(). Although there is no problem in Standard
5111 C, I've put in casts to keep C++ compilers happy.
5112
5113 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5114 should be (const char *).
5115
5116 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5117 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5118 However, I haven't made this a standard facility. The documentation doesn't
5119 mention it, and the Makefile doesn't support it.
5120
5121 5. The Makefile now contains an "install" target, with editable destinations at
5122 the top of the file. The pcretest program is not installed.
5123
5124 6. pgrep -V now gives the PCRE version number and date.
5125
5126 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5127 causing the entire string to be ignored, instead of just the last character.
5128
5129 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5130 non-matching string, it can take a very, very long time, even for strings of
5131 quite modest length, because of the nested recursion. PCRE now does better in
5132 some of these cases. It does this by remembering the last required literal
5133 character in the pattern, and pre-searching the subject to ensure it is present
5134 before running the real match. In other words, it applies a heuristic to detect
5135 some types of certain failure quickly, and in the above example, if presented
5136 with a string that has no trailing " it gives "no match" very quickly.
5137
5138 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5139 other alternatives are tried instead.
5140
5141
5142 Version 2.06 09-Jun-99
5143 ----------------------
5144
5145 1. Change pcretest's output for amount of store used to show just the code
5146 space, because the remainder (the data block) varies in size between 32-bit and
5147 64-bit systems.
5148
5149 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5150 start matching at. This allows lookbehinds to work when searching for multiple
5151 occurrences in a string.
5152
5153 3. Added additional options to pcretest for testing multiple occurrences:
5154
5155 /+ outputs the rest of the string that follows a match
5156 /g loops for multiple occurrences, using the new startoffset argument
5157 /G loops for multiple occurrences by passing an incremented pointer
5158
5159 4. PCRE wasn't doing the "first character" optimization for patterns starting
5160 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5161 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5162 the letter 'x'. On long subject strings, this gives a significant speed-up.
5163
5164
5165 Version 2.05 21-Apr-99
5166 ----------------------
5167
5168 1. Changed the type of magic_number from int to long int so that it works
5169 properly on 16-bit systems.
5170
5171 2. Fixed a bug which caused patterns starting with .* not to work correctly
5172 when the subject string contained newline characters. PCRE was assuming
5173 anchoring for such patterns in all cases, which is not correct because .* will
5174 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5175 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5176 must be retried after every newline in the subject.
5177
5178
5179 Version 2.04 18-Feb-99
5180 ----------------------
5181
5182 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5183 computation of the store needed to hold the pattern was incorrect (too large).
5184 If such patterns were nested a few deep, this could multiply and become a real
5185 problem.
5186
5187 2. Added /M option to pcretest to show the memory requirement of a specific
5188 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5189
5190 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5191 compiled in such a way that the backtracking after subsequent failure was
5192 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5193 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5194
5195
5196 Version 2.03 02-Feb-99
5197 ----------------------
5198
5199 1. Fixed typo and small mistake in man page.
5200
5201 2. Added 4th condition (GPL supersedes if conflict) and created separate
5202 LICENCE file containing the conditions.
5203
5204 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5205 Perl, that is the internal \ allows the delimiter to be included in the
5206 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5207 the final delimiter, add \ to the end of the pattern (to test the error).
5208
5209 4. Added the convenience functions for extracting substrings after a successful
5210 match. Updated pcretest to make it able to test these functions.
5211
5212
5213 Version 2.02 14-Jan-99
5214 ----------------------
5215
5216 1. Initialized the working variables associated with each extraction so that
5217 their saving and restoring doesn't refer to uninitialized store.
5218
5219 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5220 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5221 fix the problem.
5222
5223 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5224 calls, and wasn't printing the correct value for compiling calls. Increased the
5225 default value of LOOPREPEAT, and the number of significant figures in the
5226 times.
5227
5228 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5229
5230 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5231 a building problem on Windows NT with a FAT file system.
5232
5233
5234 Version 2.01 21-Oct-98
5235 ----------------------
5236
5237 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5238 to character tables built by pcre_maketables() in the current locale. If NULL
5239 is passed, the default tables are used.
5240
5241
5242 Version 2.00 24-Sep-98
5243 ----------------------
5244
5245 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5246 it any more.
5247
5248 2. Allow quantification of (?>) groups, and make it work correctly.
5249
5250 3. The first character computation wasn't working for (?>) groups.
5251
5252 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5253 end of the subject) and add 5.005's \z, which really does match only at the
5254 very end of the subject.
5255
5256 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5257
5258 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5259 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5260 localized options. All options to pcre_study() were also removed.
5261
5262 7. Add other new features from 5.005:
5263
5264 $(?<= positive lookbehind
5265 $(?<! negative lookbehind
5266 (?imsx-imsx) added the unsetting capability
5267 such a setting is global if at outer level; local otherwise
5268 (?imsx-imsx:) non-capturing groups with option setting
5269 (?(cond)re|re) conditional pattern matching
5270
5271 A backreference to itself in a repeated group matches the previous
5272 captured string.
5273
5274 8. General tidying up of studying (both automatic and via "study")
5275 consequential on the addition of new assertions.
5276
5277 9. As in 5.005, unlimited repeated groups that could match an empty substring
5278 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5279 runtime if any iteration does actually match an empty substring.
5280
5281 10. Include the RunTest script in the distribution.
5282
5283 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5284 discrepancies, some of which were old and were also with respect to 5.004. They
5285 have now been fixed.
5286
5287
5288 Version 1.09 28-Apr-98
5289 ----------------------
5290
5291 1. A negated single character class followed by a quantifier with a minimum
5292 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5293 program crashes, or just wrong answers. This did not apply to negated classes
5294 containing more than one character, or to minima other than one.
5295
5296
5297 Version 1.08 27-Mar-98
5298 ----------------------
5299
5300 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5301
5302 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5303 latter must appear before anything that relies on it in the pattern.
5304
5305
5306 Version 1.07 16-Feb-98
5307 ----------------------
5308
5309 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5310 repeat of a potentially empty string).
5311
5312
5313 Version 1.06 23-Jan-98
5314 ----------------------
5315
5316 1. Added Markus Oberhumer's little patches for C++.
5317
5318 2. Literal strings longer than 255 characters were broken.
5319
5320
5321 Version 1.05 23-Dec-97
5322 ----------------------
5323
5324 1. Negated character classes containing more than one character were failing if
5325 PCRE_CASELESS was set at run time.
5326
5327
5328 Version 1.04 19-Dec-97
5329 ----------------------
5330
5331 1. Corrected the man page, where some "const" qualifiers had been omitted.
5332
5333 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5334 input syntax.
5335
5336 3. Fixed memory leak which occurred when a regex with back references was
5337 matched with an offsets vector that wasn't big enough. The temporary memory
5338 that is used in this case wasn't being freed if the match failed.
5339
5340 4. Tidied pcretest to ensure it frees memory that it gets.
5341
5342 5. Temporary memory was being obtained in the case where the passed offsets
5343 vector was exactly big enough.
5344
5345 6. Corrected definition of offsetof() from change 5 below.
5346
5347 7. I had screwed up change 6 below and broken the rules for the use of
5348 setjmp(). Now fixed.
5349
5350
5351 Version 1.03 18-Dec-97
5352 ----------------------
5353
5354 1. A erroneous regex with a missing opening parenthesis was correctly
5355 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5356 on some systems.
5357
5358 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5359 it was reported that one broken compiler failed on the former because "code" is
5360 also an independent variable.
5361
5362 3. The erroneous regex a[]b caused an array overrun reference.
5363
5364 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5365 fail on data ending with that character. (It was going on too far, and checking
5366 the next character, typically a binary zero.) This was specific to the
5367 optimized code for single-character negative classes.
5368
5369 5. Added a contributed patch from the TIN world which does the following:
5370
5371 + Add an undef for memmove, in case the the system defines a macro for it.
5372
5373 + Add a definition of offsetof(), in case there isn't one. (I don't know
5374 the reason behind this - offsetof() is part of the ANSI standard - but
5375 it does no harm).
5376
5377 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5378 most of the places where whitespace preceded '#'. I have given up and
5379 allowed the remaining 2 cases to be at the margin.
5380
5381 + Rename some variables in pcre to eliminate shadowing. This seems very
5382 pedantic, but does no harm, of course.
5383
5384 6. Moved the call to setjmp() into its own function, to get rid of warnings
5385 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5386
5387 7. Constructs such as \d{8,} were compiling into the equivalent of
5388 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5389 outcome, but in this particular case used more store than had been allocated,
5390 which caused the bug to be discovered because it threw up an internal error.
5391
5392 8. The debugging code in both pcre and pcretest for outputting the compiled
5393 form of a regex was going wrong in the case of back references followed by
5394 curly-bracketed repeats.
5395
5396
5397 Version 1.02 12-Dec-97
5398 ----------------------
5399
5400 1. Typos in pcre.3 and comments in the source fixed.
5401
5402 2. Applied a contributed patch to get rid of places where it used to remove
5403 'const' from variables, and fixed some signed/unsigned and uninitialized
5404 variable warnings.
5405
5406 3. Added the "runtest" target to Makefile.
5407
5408 4. Set default compiler flag to -O2 rather than just -O.
5409
5410
5411 Version 1.01 19-Nov-97
5412 ----------------------
5413
5414 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5415 like /([ab]*)*/, that is, for classes with more than one character in them.
5416
5417 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5418 as /((?>a*))*/ (a PCRE_EXTRA facility).
5419
5420
5421 Version 1.00 18-Nov-97
5422 ----------------------
5423
5424 1. Added compile-time macros to support systems such as SunOS4 which don't have
5425 memmove() or strerror() but have other things that can be used instead.
5426
5427 2. Arranged that "make clean" removes the executables.
5428
5429
5430 Version 0.99 27-Oct-97
5431 ----------------------
5432
5433 1. Fixed bug in code for optimizing classes with only one character. It was
5434 initializing a 32-byte map regardless, which could cause it to run off the end
5435 of the memory it had got.
5436
5437 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5438
5439
5440 Version 0.98 22-Oct-97
5441 ----------------------
5442
5443 1. Fixed bug in code for handling temporary memory usage when there are more
5444 back references than supplied space in the ovector. This could cause segfaults.
5445
5446
5447 Version 0.97 21-Oct-97
5448 ----------------------
5449
5450 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5451
5452 2. Optimized negated single characters not to use a bit map.
5453
5454 3. Brought error texts together as macro definitions; clarified some of them;
5455 fixed one that was wrong - it said "range out of order" when it meant "invalid
5456 escape sequence".
5457
5458 4. Changed some char * arguments to const char *.
5459
5460 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5461
5462 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5463 pcretest.
5464
5465
5466 Version 0.96 16-Oct-97
5467 ----------------------
5468
5469 1. Added a simple "pgrep" utility to the distribution.
5470
5471 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5472 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5473 where "ddd" means "one or more decimal digits".
5474
5475 3. Fixed serious bug. If a pattern had a back reference, but the call to
5476 pcre_exec() didn't supply a large enough ovector to record the related
5477 identifying subpattern, the match always failed. PCRE now remembers the number
5478 of the largest back reference, and gets some temporary memory in which to save
5479 the offsets during matching if necessary, in order to ensure that
5480 backreferences always work.
5481
5482 4. Increased the compatibility with Perl in a number of ways:
5483
5484 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5485 to request this handling. The option can be set at compile or exec time.
5486
5487 (b) $ matches before a terminating newline by default; an option
5488 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5489 mode). The option can be set at compile or exec time.
5490
5491 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5492 the same as Perl's. If the decimal number it represents is less than 10
5493 or there aren't that many previous left capturing parentheses, an octal
5494 escape is read. Inside a character class, it's always an octal escape,
5495 even if it is a single digit.
5496
5497 (d) An escaped but undefined alphabetic character is taken as a literal,
5498 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5499 escapes.
5500
5501 (e) {0} is now permitted. (The previous item is removed from the compiled
5502 pattern).
5503
5504 5. Changed all the names of code files so that the basic parts are no longer
5505 than 10 characters, and abolished the teeny "globals.c" file.
5506
5507 6. Changed the handling of character classes; they are now done with a 32-byte
5508 bit map always.
5509
5510 7. Added the -d and /D options to pcretest to make it possible to look at the
5511 internals of compilation without having to recompile pcre.
5512
5513
5514 Version 0.95 23-Sep-97
5515 ----------------------
5516
5517 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5518 \x20 at the start of a run of normal characters. These were being treated as
5519 real characters, instead of the source characters being re-checked.
5520
5521
5522 Version 0.94 18-Sep-97
5523 ----------------------
5524
5525 1. The functions are now thread-safe, with the caveat that the global variables
5526 containing pointers to malloc() and free() or alternative functions are the
5527 same for all threads.
5528
5529 2. Get pcre_study() to generate a bitmap of initial characters for non-
5530 anchored patterns when this is possible, and use it if passed to pcre_exec().
5531
5532
5533 Version 0.93 15-Sep-97
5534 ----------------------
5535
5536 1. /(b)|(:+)/ was computing an incorrect first character.
5537
5538 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5539 but not actually doing anything yet.
5540
5541 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5542 as Perl does (e.g. [-az] or [az-]).
5543
5544 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5545 all possible positions.
5546
5547 5. Split up into different modules to avoid including unneeded functions in a
5548 compiled binary. However, compile and exec are still in one module. The "study"
5549 function is split off.
5550
5551 6. The character tables are now in a separate module whose source is generated
5552 by an auxiliary program - but can then be edited by hand if required. There are
5553 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5554 toupper() in the code.
5555
5556 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5557 make them global. Abolish the function for setting them, as the caller can now
5558 set them directly.
5559
5560
5561 Version 0.92 11-Sep-97
5562 ----------------------
5563
5564 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5565 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5566
5567 2. Caseless matching was not working in character classes if the characters in
5568 the pattern were in upper case.
5569
5570 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5571
5572 4. Make PCRE_ANCHORED public and accept as a compile option.
5573
5574 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5575 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5576 pass them.
5577
5578 6. Give an error if bad option bits passed at compile or run time.
5579
5580 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5581 pcretest to cause it to pass that flag.
5582
5583 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5584 options, and the first character, if set.
5585
5586 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5587
5588
5589 Version 0.91 10-Sep-97
5590 ----------------------
5591
5592 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5593 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5594
5595 2. PCRE was looping on encountering an indefinitely repeated back reference to
5596 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5597 Perl does - treats the match as successful.
5598
5599 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5