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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1514 - (show annotations)
Tue Feb 3 14:34:44 2015 UTC (4 years, 11 months ago) by ph10
File size: 269270 byte(s)
Error occurred while calculating annotation data.
Crash in pcretest after bad input.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.37 xx-xxx-201x
5 ------------------------
6
7 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8 for those parentheses to be closed with whatever has been captured so far.
9 However, it was failing to mark any other groups between the hightest
10 capture so far and the currrent group as "unset". Thus, the ovector for
11 those groups contained whatever was previously there. An example is the
12 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
13
14 2. If an assertion condition was quantified with a minimum of zero (an odd
15 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
16
17 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
18 unrecognized modifier, a crash could occur.
19
20
21
22 Version 8.36 26-September-2014
23 ------------------------------
24
25 1. Got rid of some compiler warnings in the C++ modules that were shown up by
26 -Wmissing-field-initializers and -Wunused-parameter.
27
28 2. The tests for quantifiers being too big (greater than 65535) were being
29 applied after reading the number, and stupidly assuming that integer
30 overflow would give a negative number. The tests are now applied as the
31 numbers are read.
32
33 3. Tidy code in pcre_exec.c where two branches that used to be different are
34 now the same.
35
36 4. The JIT compiler did not generate match limit checks for certain
37 bracketed expressions with quantifiers. This may lead to exponential
38 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
39 issue should be resolved now.
40
41 5. Fixed an issue, which occures when nested alternatives are optimized
42 with table jumps.
43
44 6. Inserted two casts and changed some ints to size_t in the light of some
45 reported 64-bit compiler warnings (Bugzilla 1477).
46
47 7. Fixed a bug concerned with zero-minimum possessive groups that could match
48 an empty string, which sometimes were behaving incorrectly in the
49 interpreter (though correctly in the JIT matcher). This pcretest input is
50 an example:
51
52 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
53 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
54
55 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
56 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
57 for an empty string was breaking the inner loop and carrying on at a lower
58 level, when possessive repeated groups should always return to a higher
59 level as they have no backtrack points in them. The empty string test now
60 occurs at the outer level.
61
62 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
63 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
64
65 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
66 doesn't).
67
68 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
69 to hold a pattern to be incorrectly computed (too small) when there were
70 named back references to duplicated names. This could cause "internal
71 error: code overflow" or "double free or corruption" or other memory
72 handling errors.
73
74 11. When named subpatterns had the same prefixes, back references could be
75 confused. For example, in this pattern:
76
77 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
78
79 the reference to 'Name' was incorrectly treated as a reference to a
80 duplicate name.
81
82 12. A pattern such as /^s?c/mi8 where the optional character has more than
83 one "other case" was incorrectly compiled such that it would only try to
84 match starting at "c".
85
86 13. When a pattern starting with \s was studied, VT was not included in the
87 list of possible starting characters; this should have been part of the
88 8.34/18 patch.
89
90 14. If a character class started [\Qx]... where x is any character, the class
91 was incorrectly terminated at the ].
92
93 15. If a pattern that started with a caseless match for a character with more
94 than one "other case" was studied, PCRE did not set up the starting code
95 unit bit map for the list of possible characters. Now it does. This is an
96 optimization improvement, not a bug fix.
97
98 16. The Unicode data tables have been updated to Unicode 7.0.0.
99
100 17. Fixed a number of memory leaks in pcregrep.
101
102 18. Avoid a compiler warning (from some compilers) for a function call with
103 a cast that removes "const" from an lvalue by using an intermediate
104 variable (to which the compiler does not object).
105
106 19. Incorrect code was compiled if a group that contained an internal recursive
107 back reference was optional (had quantifier with a minimum of zero). This
108 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
109 caused segmentation faults because of stack overflows at compile time.
110
111 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
112 group that is quantified with an indefinite repeat, caused a compile-time
113 loop which used up all the system stack and provoked a segmentation fault.
114 This was not the same bug as 19 above.
115
116 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
117 Patch by Mike Frysinger.
118
119
120 Version 8.35 04-April-2014
121 --------------------------
122
123 1. A new flag is set, when property checks are present in an XCLASS.
124 When this flag is not set, PCRE can perform certain optimizations
125 such as studying these XCLASS-es.
126
127 2. The auto-possessification of character sets were improved: a normal
128 and an extended character set can be compared now. Furthermore
129 the JIT compiler optimizes more character set checks.
130
131 3. Got rid of some compiler warnings for potentially uninitialized variables
132 that show up only when compiled with -O2.
133
134 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
135 of a match later then the end of the match. The pcretest program was not
136 handling the case sensibly - it was outputting from the start to the next
137 binary zero. It now reports this situation in a message, and outputs the
138 text from the end to the start.
139
140 5. Fast forward search is improved in JIT. Instead of the first three
141 characters, any three characters with fixed position can be searched.
142 Search order: first, last, middle.
143
144 6. Improve character range checks in JIT. Characters are read by an inprecise
145 function now, which returns with an unknown value if the character code is
146 above a certain threshold (e.g: 256). The only limitation is that the value
147 must be bigger than the threshold as well. This function is useful when
148 the characters above the threshold are handled in the same way.
149
150 7. The macros whose names start with RAWUCHAR are placeholders for a future
151 mode in which only the bottom 21 bits of 32-bit data items are used. To
152 make this more memorable for those maintaining the code, the names have
153 been changed to start with UCHAR21, and an extensive comment has been added
154 to their definition.
155
156 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
157 to the export list in Makefile.am (they were accidentally omitted from the
158 8.34 tarball).
159
160 9. The informational output from pcretest used the phrase "starting byte set"
161 which is inappropriate for the 16-bit and 32-bit libraries. As the output
162 for "first char" and "need char" really means "non-UTF-char", I've changed
163 "byte" to "char", and slightly reworded the output. The documentation about
164 these values has also been (I hope) clarified.
165
166 10. Another JIT related optimization: use table jumps for selecting the correct
167 backtracking path, when more than four alternatives are present inside a
168 bracket.
169
170 11. Empty match is not possible, when the minimum length is greater than zero,
171 and there is no \K in the pattern. JIT should avoid empty match checks in
172 such cases.
173
174 12. In a caseless character class with UCP support, when a character with more
175 than one alternative case was not the first character of a range, not all
176 the alternative cases were added to the class. For example, s and \x{17f}
177 are both alternative cases for S: the class [RST] was handled correctly,
178 but [R-T] was not.
179
180 13. The configure.ac file always checked for pthread support when JIT was
181 enabled. This is not used in Windows, so I have put this test inside a
182 check for the presence of windows.h (which was already tested for).
183
184 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
185 The algorithm provides a way to skip certain starting offsets, and usually
186 faster than linear prefix searches.
187
188 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
189 as for 'fr_FR' and 'french'. For some reason, however, it then used the
190 Windows-specific input and output files, which have 'french' screwed in.
191 So this could never have worked. One of the problems with locales is that
192 they aren't always the same. I have now updated RunTest so that it checks
193 the output of the locale test (test 3) against three different output
194 files, and it allows the test to pass if any one of them matches. With luck
195 this should make the test pass on some versions of Solaris where it was
196 failing. Because of the uncertainty, the script did not used to stop if
197 test 3 failed; it now does. If further versions of a French locale ever
198 come to light, they can now easily be added.
199
200 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
201 there was a message during ./configure, but it did not stop. This now
202 provokes an error. The invalid example in README has been corrected.
203 If a value less than the minimum is given, the minimum value has always
204 been used, but now a warning is given.
205
206 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
207 was a bug in the test system, which is now fixed. Also, the list of various
208 configurations that are tested for each release did not have one with both
209 16/32 bits and --enable-bar-anycrlf. It now does.
210
211 18. pcretest was missing "-C bsr" for displaying the \R default setting.
212
213 19. Little endian PowerPC systems are supported now by the JIT compiler.
214
215 20. The fast forward newline mechanism could enter to an infinite loop on
216 certain invalid UTF-8 input. Although we don't support these cases
217 this issue can be fixed by a performance optimization.
218
219 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
220 not take account if existing stack usage. There is now a new global
221 variable called pcre_stack_guard that can be set to point to an external
222 function to check stack availability. It is called at the start of
223 processing every parenthesized group.
224
225 22. A typo in the code meant that in ungreedy mode the max/min qualifier
226 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
227 not match "ab".
228
229 23. When UTF was disabled, the JIT program reported some incorrect compile
230 errors. These messages are silenced now.
231
232 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
233 compiler.
234
235 25. Change all the temporary files used in RunGrepTest to be different to those
236 used by RunTest so that the tests can be run simultaneously, for example by
237 "make -j check".
238
239
240 Version 8.34 15-December-2013
241 -----------------------------
242
243 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
244 executable memory. Patch inspired by Carsten Klein.
245
246 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
247 this macro is never tested and has no effect, because the work to support
248 coverage involves only compiling and linking options and special targets in
249 the Makefile. The comment in config.h implied that defining the macro would
250 enable coverage support, which is totally false. There was also support for
251 setting this macro in the CMake files (my fault, I just copied it from
252 configure). SUPPORT_GCOV has now been removed.
253
254 3. Make a small performance improvement in strlen16() and strlen32() in
255 pcretest.
256
257 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
258 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
259 these cases). There was also one in pcretest.c.
260
261 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
262
263 6. In UTF mode, the code for checking whether a group could match an empty
264 string (which is used for indefinitely repeated groups to allow for
265 breaking an infinite loop) was broken when the group contained a repeated
266 negated single-character class with a character that occupied more than one
267 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
268 UTF-8 mode). The effect was undefined: the group might or might not be
269 deemed as matching an empty string, or the program might have crashed.
270
271 7. The code for checking whether a group could match an empty string was not
272 recognizing that \h, \H, \v, \V, and \R must match a character.
273
274 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
275 an empty string. If it can, pcretest shows this in its information output.
276
277 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
278 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
279 matched by pcre_exec() without using JIT:
280
281 (a) If the rest of the pattern did not match after a maximal run of
282 grapheme clusters, the code for backing up to try with fewer of them
283 did not always back up over a full grapheme when characters that do not
284 have the modifier quality were involved, e.g. Hangul syllables.
285
286 (b) If the match point in a subject started with modifier character, and
287 there was no match, the code could incorrectly back up beyond the match
288 point, and potentially beyond the first character in the subject,
289 leading to a segfault or an incorrect match result.
290
291 10. A conditional group with an assertion condition could lead to PCRE
292 recording an incorrect first data item for a match if no other first data
293 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
294 first data item, and therefore matched "ca" after "c" instead of at the
295 start.
296
297 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
298 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
299
300 12. The source of pcregrep now includes z/OS-specific code so that it can be
301 compiled for z/OS as part of the special z/OS distribution.
302
303 13. Added the -T and -TM options to pcretest.
304
305 14. The code in pcre_compile.c for creating the table of named capturing groups
306 has been refactored. Instead of creating the table dynamically during the
307 actual compiling pass, the information is remembered during the pre-compile
308 pass (on the stack unless there are more than 20 named groups, in which
309 case malloc() is used) and the whole table is created before the actual
310 compile happens. This has simplified the code (it is now nearly 150 lines
311 shorter) and prepared the way for better handling of references to groups
312 with duplicate names.
313
314 15. A back reference to a named subpattern when there is more than one of the
315 same name now checks them in the order in which they appear in the pattern.
316 The first one that is set is used for the reference. Previously only the
317 first one was inspected. This change makes PCRE more compatible with Perl.
318
319 16. Unicode character properties were updated from Unicode 6.3.0.
320
321 17. The compile-time code for auto-possessification has been refactored, based
322 on a patch by Zoltan Herczeg. It now happens after instead of during
323 compilation. The code is cleaner, and more cases are handled. The option
324 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
325 options in pcretest are provided to set it. It can also be set by
326 (*NO_AUTO_POSSESS) at the start of a pattern.
327
328 18. The character VT has been added to the default ("C" locale) set of
329 characters that match \s and are generally treated as white space,
330 following this same change in Perl 5.18. There is now no difference between
331 "Perl space" and "POSIX space". Whether VT is treated as white space in
332 other locales depends on the locale.
333
334 19. The code for checking named groups as conditions, either for being set or
335 for being recursed, has been refactored (this is related to 14 and 15
336 above). Processing unduplicated named groups should now be as fast at
337 numerical groups, and processing duplicated groups should be faster than
338 before.
339
340 20. Two patches to the CMake build system, by Alexander Barkov:
341
342 (1) Replace the "source" command by "." in CMakeLists.txt because
343 "source" is a bash-ism.
344
345 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
346 without these the CMake build does not work on Solaris.
347
348 21. Perl has changed its handling of \8 and \9. If there is no previously
349 encountered capturing group of those numbers, they are treated as the
350 literal characters 8 and 9 instead of a binary zero followed by the
351 literals. PCRE now does the same.
352
353 22. Following Perl, added \o{} to specify codepoints in octal, making it
354 possible to specify values greater than 0777 and also making them
355 unambiguous.
356
357 23. Perl now gives an error for missing closing braces after \x{... instead of
358 treating the string as literal. PCRE now does the same.
359
360 24. RunTest used to grumble if an inappropriate test was selected explicitly,
361 but just skip it when running all tests. This make it awkward to run ranges
362 of tests when one of them was inappropriate. Now it just skips any
363 inappropriate tests, as it always did when running all tests.
364
365 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
366 character types such as \d or \w, too many callouts were inserted, and the
367 data that they returned was rubbish.
368
369 26. In UCP mode, \s was not matching two of the characters that Perl matches,
370 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
371 were matched by \h. The code has now been refactored so that the lists of
372 the horizontal and vertical whitespace characters used for \h and \v (which
373 are defined only in one place) are now also used for \s.
374
375 27. Add JIT support for the 64 bit TileGX architecture.
376 Patch by Jiong Wang (Tilera Corporation).
377
378 28. Possessive quantifiers for classes (both explicit and automatically
379 generated) now use special opcodes instead of wrapping in ONCE brackets.
380
381 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
382 (because it's meaningless), this was not happening when PCRE_CASELESS was
383 set. Not wrong, but inefficient.
384
385 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
386 properties for \w, \d, etc) is present in a test regex. Otherwise if the
387 test contains no characters greater than 255, Perl doesn't realise it
388 should be using Unicode semantics.
389
390 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
391 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
392 does in Unicode mode.
393
394 32. Added the "forbid" facility to pcretest so that putting tests into the
395 wrong test files can sometimes be quickly detected.
396
397 33. There is now a limit (default 250) on the depth of nesting of parentheses.
398 This limit is imposed to control the amount of system stack used at compile
399 time. It can be changed at build time by --with-parens-nest-limit=xxx or
400 the equivalent in CMake.
401
402 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
403 errors. Perl warns for these when in warning mode, but PCRE has no facility
404 for giving warnings.
405
406 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
407 However, this was not working for (?!) because it is optimized to (*FAIL),
408 for which PCRE does not allow quantifiers. The optimization is now disabled
409 when a quantifier follows (?!). I can't see any use for this, but it makes
410 things uniform.
411
412 36. Perl no longer allows group names to start with digits, so I have made this
413 change also in PCRE. It simplifies the code a bit.
414
415 37. In extended mode, Perl ignores spaces before a + that indicates a
416 possessive quantifier. PCRE allowed a space before the quantifier, but not
417 before the possessive +. It now does.
418
419 38. The use of \K (reset reported match start) within a repeated possessive
420 group such as (a\Kb)*+ was not working.
421
422 40. Document that the same character tables must be used at compile time and
423 run time, and that the facility to pass tables to pcre_exec() and
424 pcre_dfa_exec() is for use only with saved/restored patterns.
425
426 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
427 features for Builds with MSVC:
428
429 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
430 builds). The .rc files can be used to set FileDescription and many other
431 attributes.
432
433 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
434 This allows higher-level build scripts which want .pdb files to avoid
435 hard-coding the exact files needed."
436
437 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
438 mean "start of word" and "end of word", respectively, as a transition aid.
439
440 43. A minimizing repeat of a class containing codepoints greater than 255 in
441 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
442 compiled to use the heap for recursion.
443
444 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
445 is configured.
446
447
448 Version 8.33 28-May-2013
449 ------------------------
450
451 1. Added 'U' to some constants that are compared to unsigned integers, to
452 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
453 variables that are added to signed variables, to ensure the result is
454 signed and can be negated.
455
456 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
457 CMake config files.
458
459 3. Revise the creation of config.h.generic so that all boolean macros are
460 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
461 overriding via -D on the command line possible.
462
463 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
464 to unsigned int is reported to make a quite noticeable speed difference in
465 a specific Windows environment. Testing on Linux did also appear to show
466 some benefit (and it is clearly not harmful). Also fixed the definition of
467 Xop which should be unsigned.
468
469 5. Related to (4), changing the definition of the intermediate variable cc
470 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
471 improvements.
472
473 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
474 unnecessary spaces.
475
476 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
477 and later.
478
479 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
480
481 9. Optimizing fast_forward_start_bits in JIT.
482
483 10. Adding support for callouts in JIT, and fixing some issues revealed
484 during this work. Namely:
485
486 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
487
488 (b) Minimum length was not checked before the matching is started.
489
490 11. The value of capture_last that is passed to callouts was incorrect in some
491 cases when there was a capture on one path that was subsequently abandoned
492 after a backtrack. Also, the capture_last value is now reset after a
493 recursion, since all captures are also reset in this case.
494
495 12. The interpreter no longer returns the "too many substrings" error in the
496 case when an overflowing capture is in a branch that is subsequently
497 abandoned after a backtrack.
498
499 13. In the pathological case when an offset vector of size 2 is used, pcretest
500 now prints out the matched string after a yield of 0 or 1.
501
502 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
503 Only supported by the JIT compiler at the moment.
504
505 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
506
507 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
508 offset of the starting point of the matching process, provided the offsets
509 vector is large enough.
510
511 17. The \A escape now records a lookbehind value of 1, though its execution
512 does not actually inspect the previous character. This is to ensure that,
513 in partial multi-segment matching, at least one character from the old
514 segment is retained when a new segment is processed. Otherwise, if there
515 are no lookbehinds in the pattern, \A might match incorrectly at the start
516 of a new segment.
517
518 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
519
520 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
521 gives some modest performance improvement in 8-bit mode.
522
523 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
524 be expressed in certain programming languages using Universal Character
525 Names.
526
527 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
528 which points out that "non characters" are not "characters that may not
529 appear in Unicode strings" but rather "characters that are reserved for
530 internal use and have only local meaning".
531
532 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
533 there was a conditional group that depended on an assertion, if the
534 assertion was false, the callout that immediately followed the alternation
535 in the condition was skipped when pcre_exec() was used for matching.
536
537 23. Allow an explicit callout to be inserted before an assertion that is the
538 condition for a conditional group, for compatibility with automatic
539 callouts, which always insert a callout at this point.
540
541 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
542 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
543
544 25. (*PRUNE) is now supported by the JIT compiler.
545
546 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
547
548 27. Fix the case where there are two or more SKIPs with arguments that may be
549 ignored.
550
551 28. (*SKIP) is now supported by the JIT compiler.
552
553 29. (*THEN) is now supported by the JIT compiler.
554
555 30. Update RunTest with additional test selector options.
556
557 31. The way PCRE handles backtracking verbs has been changed in two ways.
558
559 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
560 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
561 backtracking. In some cases this makes it more Perl-compatible, but Perl's
562 rather obscure rules do not always do the same thing.
563
564 (2) Previously, backtracking verbs were confined within assertions. This is
565 no longer the case for positive assertions, except for (*ACCEPT). Again,
566 this sometimes improves Perl compatibility, and sometimes does not.
567
568 32. A number of tests that were in test 2 because Perl did things differently
569 have been moved to test 1, because either Perl or PCRE has changed, and
570 these tests are now compatible.
571
572 32. Backtracking control verbs are now handled in the same way in JIT and
573 interpreter.
574
575 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
576 contained a forward subroutine reference caused a compile error.
577
578 34. Auto-detect and optimize limited repetitions in JIT.
579
580 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
581 blocking (*UTF) etc.
582
583 36. In the interpreter, maximizing pattern repetitions for characters and
584 character types now use tail recursion, which reduces stack usage.
585
586 37. The value of the max lookbehind was not correctly preserved if a compiled
587 and saved regex was reloaded on a host of different endianness.
588
589 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
590 of the compiled pattern block, expand the flags field from 16 to 32 bits
591 because it was almost full.
592
593 39. Try madvise first before posix_madvise.
594
595 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
596 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
597 original change locked out all matches of empty strings. This has been
598 changed so that one match of an empty string per line is recognized.
599 Subsequent searches on the same line (for colouring or for --only-matching,
600 for example) do not recognize empty strings.
601
602 41. Applied a user patch to fix a number of spelling mistakes in comments.
603
604 42. Data lines longer than 65536 caused pcretest to crash.
605
606 43. Clarified the data type for length and startoffset arguments for pcre_exec
607 and pcre_dfa_exec in the function-specific man pages, where they were
608 explicitly stated to be in bytes, never having been updated. I also added
609 some clarification to the pcreapi man page.
610
611 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
612 a segmentation fault.
613
614
615 Version 8.32 30-November-2012
616 -----------------------------
617
618 1. Improved JIT compiler optimizations for first character search and single
619 character iterators.
620
621 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
622 Patch by Daniel Richard G.
623
624 3. Single character iterator optimizations in the JIT compiler.
625
626 4. Improved JIT compiler optimizations for character ranges.
627
628 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
629 Reported by Giuseppe D'Angelo.
630
631 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
632 of a line, was being set incorrectly in cases where .* appeared inside
633 atomic brackets at the start of a pattern, or where there was a subsequent
634 *PRUNE or *SKIP.
635
636 7. Improved instruction cache flush for POWER/PowerPC.
637 Patch by Daniel Richard G.
638
639 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
640 grep:
641
642 (a) There is now no limit to the number of patterns to be matched.
643
644 (b) An error is given if a pattern is too long.
645
646 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
647 are now supported.
648
649 (d) --exclude-from and --include-from (multiple use) have been added.
650
651 (e) Exclusions and inclusions now apply to all files and directories, not
652 just to those obtained from scanning a directory recursively.
653
654 (f) Multiple uses of -f and --file-list are now supported.
655
656 (g) In a Windows environment, the default for -d has been changed from
657 "read" (the GNU grep default) to "skip", because otherwise the presence
658 of a directory in the file list provokes an error.
659
660 (h) The documentation has been revised and clarified in places.
661
662 9. Improve the matching speed of capturing brackets.
663
664 10. Changed the meaning of \X so that it now matches a Unicode extended
665 grapheme cluster.
666
667 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
668 out POSIX threads when JIT support is configured.
669
670 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
671
672 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
673 structure could go wrong in environments where size_t is not the same size
674 as int.
675
676 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
677 set.
678
679 15. The EBCDIC support had decayed; later updates to the code had included
680 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
681 general tidy up of EBCDIC-related issues, and the documentation was also
682 not quite right. There is now a test that can be run on ASCII systems to
683 check some of the EBCDIC-related things (but is it not a full test).
684
685 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
686 in a small tidy to the code.
687
688 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
689
690 18. If the --only-matching (-o) option in pcregrep is specified multiple
691 times, each one causes appropriate output. For example, -o1 -o2 outputs the
692 substrings matched by the 1st and 2nd capturing parentheses. A separating
693 string can be specified by --om-separator (default empty).
694
695 19. Improving the first n character searches.
696
697 20. Turn case lists for horizontal and vertical white space into macros so that
698 they are defined only once.
699
700 21. This set of changes together give more compatible Unicode case-folding
701 behaviour for characters that have more than one other case when UCP
702 support is available.
703
704 (a) The Unicode property table now has offsets into a new table of sets of
705 three or more characters that are case-equivalent. The MultiStage2.py
706 script that generates these tables (the pcre_ucd.c file) now scans
707 CaseFolding.txt instead of UnicodeData.txt for character case
708 information.
709
710 (b) The code for adding characters or ranges of characters to a character
711 class has been abstracted into a generalized function that also handles
712 case-independence. In UTF-mode with UCP support, this uses the new data
713 to handle characters with more than one other case.
714
715 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
716 whose other case is greater than 256 are now correctly matched
717 caselessly. Previously, the high codepoint matched the low one, but not
718 vice versa.
719
720 (d) The processing of \h, \H, \v, and \ in character classes now makes use
721 of the new class addition function, using character lists defined as
722 macros alongside the case definitions of 20 above.
723
724 (e) Caseless back references now work with characters that have more than
725 one other case.
726
727 (f) General caseless matching of characters with more than one other case
728 is supported.
729
730 22. Unicode character properties were updated from Unicode 6.2.0
731
732 23. Improved CMake support under Windows. Patch by Daniel Richard G.
733
734 24. Add support for 32-bit character strings, and UTF-32
735
736 25. Major JIT compiler update (code refactoring and bugfixing).
737 Experimental Sparc 32 support is added.
738
739 26. Applied a modified version of Daniel Richard G's patch to create
740 pcre.h.generic and config.h.generic by "make" instead of in the
741 PrepareRelease script.
742
743 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
744 pcre_compile.c when checking for a zero character.
745
746 28. Introducing a native interface for JIT. Through this interface, the compiled
747 machine code can be directly executed. The purpose of this interface is to
748 provide fast pattern matching, so several sanity checks are not performed.
749 However, feature tests are still performed. The new interface provides
750 1.4x speedup compared to the old one.
751
752 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
753 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
754 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
755
756 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
757 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
758 the "old" RFC 2279). Instead, it ended up passing a negative length to
759 pcre_exec().
760
761 31. Add support for GCC's visibility feature to hide internal functions.
762
763 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
764 "unknown -C option" after outputting 0 or 1.
765
766 33. There is now support for generating a code coverage report for the test
767 suite in environments where gcc is the compiler and lcov is installed. This
768 is mainly for the benefit of the developers.
769
770 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
771 unaddressable using valgrind annotations, allowing valgrind to detect
772 invalid memory accesses. This is mainly for the benefit of the developers.
773
774 25. (*UTF) can now be used to start a pattern in any of the three libraries.
775
776 26. Give configure error if --enable-cpp but no C++ compiler found.
777
778
779 Version 8.31 06-July-2012
780 -------------------------
781
782 1. Fixing a wrong JIT test case and some compiler warnings.
783
784 2. Removed a bashism from the RunTest script.
785
786 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
787 to unsigned type, result still unsigned" that was given by an MS compiler
788 on encountering the code "-sizeof(xxx)".
789
790 4. Partial matching support is added to the JIT compiler.
791
792 5. Fixed several bugs concerned with partial matching of items that consist
793 of more than one character:
794
795 (a) /^(..)\1/ did not partially match "aba" because checking references was
796 done on an "all or nothing" basis. This also applied to repeated
797 references.
798
799 (b) \R did not give a hard partial match if \r was found at the end of the
800 subject.
801
802 (c) \X did not give a hard partial match after matching one or more
803 characters at the end of the subject.
804
805 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
806 a partial match for the string "\r".
807
808 (e) When newline was set to CRLF, the metacharacter "." did not recognize
809 a partial match for a CR character at the end of the subject string.
810
811 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
812 running pcretest, the text "(JIT)" added to the output whenever JIT is
813 actually used to run the match.
814
815 7. Individual JIT compile options can be set in pcretest by following -s+[+]
816 or /S+[+] with a digit between 1 and 7.
817
818 8. OP_NOT now supports any UTF character not just single-byte ones.
819
820 9. (*MARK) control verb is now supported by the JIT compiler.
821
822 10. The command "./RunTest list" lists the available tests without actually
823 running any of them. (Because I keep forgetting what they all are.)
824
825 11. Add PCRE_INFO_MAXLOOKBEHIND.
826
827 12. Applied a (slightly modified) user-supplied patch that improves performance
828 when the heap is used for recursion (compiled with --disable-stack-for-
829 recursion). Instead of malloc and free for each heap frame each time a
830 logical recursion happens, frames are retained on a chain and re-used where
831 possible. This sometimes gives as much as 30% improvement.
832
833 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
834 call.
835
836 14. As documented, (*COMMIT) is now confined to within a positive assertion.
837
838 15. It is now possible to link pcretest with libedit as an alternative to
839 libreadline.
840
841 16. (*COMMIT) control verb is now supported by the JIT compiler.
842
843 17. The Unicode data tables have been updated to Unicode 6.1.0.
844
845 18. Added --file-list option to pcregrep.
846
847 19. Added binary file support to pcregrep, including the -a, --binary-files,
848 -I, and --text options.
849
850 20. The madvise function is renamed for posix_madvise for QNX compatibility
851 reasons. Fixed by Giuseppe D'Angelo.
852
853 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
854
855 22. Changed the option for creating symbolic links for 16-bit man pages from
856 -s to -sf so that re-installing does not cause issues.
857
858 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
859
860 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
861 matches in certain environments (the workspace was not being correctly
862 retained). Also added to pcre_dfa_exec() a simple plausibility check on
863 some of the workspace data at the beginning of a restart.
864
865 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
866 was not doing so when it should - probably a typo introduced by SVN 528
867 (change 8.10/14).
868
869 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
870 \w+ when the character tables indicated that \x{c4} was a word character.
871 There were several related cases, all because the tests for doing a table
872 lookup were testing for characters less than 127 instead of 255.
873
874 27. If a pattern contains capturing parentheses that are not used in a match,
875 their slots in the ovector are set to -1. For those that are higher than
876 any matched groups, this happens at the end of processing. In the case when
877 there were back references that the ovector was too small to contain
878 (causing temporary malloc'd memory to be used during matching), and the
879 highest capturing number was not used, memory off the end of the ovector
880 was incorrectly being set to -1. (It was using the size of the temporary
881 memory instead of the true size.)
882
883 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
884 ovector size, it uses memory at the end of the block that it has got.
885
886 29. Check for an overlong MARK name and give an error at compile time. The
887 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
888
889 30. JIT compiler update.
890
891 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
892 Rill for the patch.
893
894 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
895
896 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
897
898 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
899 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
900 was enabled.
901
902 35. Improve JIT code generation for greedy plus quantifier.
903
904 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
905 1 to "aa" instead of to an empty string. The bug affected repeated groups
906 that could potentially match an empty string.
907
908 37. Optimizing single character iterators in JIT.
909
910 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
911 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
912 codepoints that are too big for the mode are faulted, and in a UTF mode,
913 disallowed codepoints are also faulted.
914
915 39. If PCRE was compiled with UTF support, in three places in the DFA
916 matcher there was code that should only have been obeyed in UTF mode, but
917 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
918 processing when bytes with values greater than 127 were present. In 16-bit
919 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
920 both cases the values are those that cannot be the first data item in a UTF
921 character. The three items that might have provoked this were recursions,
922 possessively repeated groups, and atomic groups.
923
924 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
925 and pcregrep, because some OS require shared objects to be explicitly
926 passed to ld, causing the link step to fail if they are not.
927
928 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
929 mode, patterns that started with \h* or \R* might be incorrectly matched.
930
931
932 Version 8.30 04-February-2012
933 -----------------------------
934
935 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
936 name is defined in ctype.h.
937
938 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
939 only in quite long subpatterns.
940
941 3. Removed the function pcre_info(), which has been obsolete and deprecated
942 since it was replaced by pcre_fullinfo() in February 2000.
943
944 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
945 match a (*MARK), and the match failed at the start of the subject, a
946 reference to memory before the start of the subject could occur. This bug
947 was introduced by fix 17 of release 8.21.
948
949 5. A reference to an unset group with zero minimum repetition was giving
950 totally wrong answers (in non-JavaScript-compatibility mode). For example,
951 /(another)?(\1?)test/ matched against "hello world test". This bug was
952 introduced in release 8.13.
953
954 6. Add support for 16-bit character strings (a large amount of work involving
955 many changes and refactorings).
956
957 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
958 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
959 from a file.
960
961 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
962 rounding is not applied in this particular case).
963
964 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
965 if they appear, or are escaped, in patterns.
966
967 10. Get rid of a number of -Wunused-but-set-variable warnings.
968
969 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
970 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
971 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
972 also returns the mark "x". This bug applied to capturing parentheses,
973 non-capturing parentheses, and atomic parentheses. It also applied to some
974 assertions.
975
976 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
977 information out of configure.ac instead of relying on pcre.h.generic, which
978 is not stored in the repository.
979
980 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
981 -lreadline.
982
983 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
984
985 15. Applied Graycode's patch to put the top-level frame on the stack rather
986 than the heap when not using the stack for recursion. This gives a
987 performance improvement in many cases when recursion is not deep.
988
989 16. Experimental code added to "pcretest -C" to output the stack frame size.
990
991
992 Version 8.21 12-Dec-2011
993 ------------------------
994
995 1. Updating the JIT compiler.
996
997 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
998 are added as well.
999
1000 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1001 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1002 calling _pcre_jit_exec. Some extra comments are added.
1003
1004 4. (*MARK) settings inside atomic groups that do not contain any capturing
1005 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1006 was introduced by change 18 for 8.20.
1007
1008 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1009 ECMA-262 standard.
1010
1011 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1012 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1013 This bug was probably introduced by change 9 of 8.13.
1014
1015 7. While fixing 6 above, I noticed that a number of other items were being
1016 incorrectly rejected as "not fixed length". This arose partly because newer
1017 opcodes had not been added to the fixed-length checking code. I have (a)
1018 corrected the bug and added tests for these items, and (b) arranged for an
1019 error to occur if an unknown opcode is encountered while checking for fixed
1020 length instead of just assuming "not fixed length". The items that were
1021 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1022 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1023 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1024
1025 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1026 being incorrectly compiled and would have given unpredicatble results.
1027
1028 9. A possessively repeated subpattern with minimum repeat count greater than
1029 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1030 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1031 the first (A) could occur when it should not.
1032
1033 10. Add a cast and remove a redundant test from the code.
1034
1035 11. JIT should use pcre_malloc/pcre_free for allocation.
1036
1037 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1038 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1039 is anything other than /usr, -L is still shown).
1040
1041 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1042
1043 14. Perl does not support \N without a following name in a [] class; PCRE now
1044 also gives an error.
1045
1046 15. If a forward reference was repeated with an upper limit of around 2000,
1047 it caused the error "internal error: overran compiling workspace". The
1048 maximum number of forward references (including repeats) was limited by the
1049 internal workspace, and dependent on the LINK_SIZE. The code has been
1050 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1051 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1052 of around 200,000 forward references. While doing this, I also speeded up
1053 the filling in of repeated forward references.
1054
1055 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1056 incorrectly expecting the subject to contain another "a" after the start.
1057
1058 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1059 in the match, the SKIP should be ignored. This was not happening; instead
1060 the SKIP was being treated as NOMATCH. For patterns such as
1061 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1062 tested.
1063
1064 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1065 now much more compatible with Perl, in particular in cases where the result
1066 is a non-match for a non-anchored pattern. For example, if
1067 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1068 "m", where previously it did not return a name. A side effect of this
1069 change is that for partial matches, the last encountered mark name is
1070 returned, as for non matches. A number of tests that were previously not
1071 Perl-compatible have been moved into the Perl-compatible test files. The
1072 refactoring has had the pleasing side effect of removing one argument from
1073 the match() function, thus reducing its stack requirements.
1074
1075 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1076 subsequent uses of /S (without +) incorrectly behaved like /S+.
1077
1078 21. Retrieve executable code size support for the JIT compiler and fixing
1079 some warnings.
1080
1081 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1082 not work when the shorter character appeared right at the end of the
1083 subject string.
1084
1085 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1086 systems.
1087
1088 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1089 output it when the /M option is used in pcretest.
1090
1091 25. The CheckMan script was not being included in the distribution. Also, added
1092 an explicit "perl" to run Perl scripts from the PrepareRelease script
1093 because this is reportedly needed in Windows.
1094
1095 26. If study data was being save in a file and studying had not found a set of
1096 "starts with" bytes for the pattern, the data written to the file (though
1097 never used) was taken from uninitialized memory and so caused valgrind to
1098 complain.
1099
1100 27. Updated RunTest.bat as provided by Sheri Pierce.
1101
1102 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1103
1104 29. Computation of memory usage for the table of capturing group names was
1105 giving an unnecessarily large value.
1106
1107
1108 Version 8.20 21-Oct-2011
1109 ------------------------
1110
1111 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1112 a POSIX class. After further experiments with Perl, which convinced me that
1113 Perl has bugs and confusions, a closing square bracket is no longer allowed
1114 in a POSIX name. This bug also affected patterns with classes that started
1115 with full stops.
1116
1117 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1118 captured substring, but while checking the failing first alternative,
1119 substring 1 is temporarily captured. If the output vector supplied to
1120 pcre_exec() was not big enough for this capture, the yield of the function
1121 was still zero ("insufficient space for captured substrings"). This cannot
1122 be totally fixed without adding another stack variable, which seems a lot
1123 of expense for a edge case. However, I have improved the situation in cases
1124 such as /(a)(b)x|abc/ matched against "abc", where the return code
1125 indicates that fewer than the maximum number of slots in the ovector have
1126 been set.
1127
1128 3. Related to (2) above: when there are more back references in a pattern than
1129 slots in the output vector, pcre_exec() uses temporary memory during
1130 matching, and copies in the captures as far as possible afterwards. It was
1131 using the entire output vector, but this conflicts with the specification
1132 that only 2/3 is used for passing back captured substrings. Now it uses
1133 only the first 2/3, for compatibility. This is, of course, another edge
1134 case.
1135
1136 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1137 main code base, and can be used by building with --enable-jit. When this is
1138 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1139 runtime --no-jit option is given.
1140
1141 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1142 ovector, the return from the function was zero, implying that there were
1143 other matches that did not fit. The correct "exactly full" value is now
1144 returned.
1145
1146 6. If a subpattern that was called recursively or as a subroutine contained
1147 (*PRUNE) or any other control that caused it to give a non-standard return,
1148 invalid errors such as "Error -26 (nested recursion at the same subject
1149 position)" or even infinite loops could occur.
1150
1151 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1152 computing the minimum length on reaching *ACCEPT, and so ended up with the
1153 wrong value of 1 rather than 0. Further investigation indicates that
1154 computing a minimum subject length in the presence of *ACCEPT is difficult
1155 (think back references, subroutine calls), and so I have changed the code
1156 so that no minimum is registered for a pattern that contains *ACCEPT.
1157
1158 8. If (*THEN) was present in the first (true) branch of a conditional group,
1159 it was not handled as intended. [But see 16 below.]
1160
1161 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1162 Sheri Pierce.
1163
1164 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1165 the first byte in a match must be "a".
1166
1167 11. Change 17 for 8.13 increased the recursion depth for patterns like
1168 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1169 pattern contains any instances of (*THEN). If it does not, the old
1170 optimizations are restored. It would be nice to do this on a per-group
1171 basis, but at the moment that is not feasible.
1172
1173 12. In some environments, the output of pcretest -C is CRLF terminated. This
1174 broke RunTest's code that checks for the link size. A single white space
1175 character after the value is now allowed for.
1176
1177 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1178 For "fr", it uses the Windows-specific input and output files.
1179
1180 14. If (*THEN) appeared in a group that was called recursively or as a
1181 subroutine, it did not work as intended. [But see next item.]
1182
1183 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1184 pattern fragments (but not containing any | characters). If A and B are
1185 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1186 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1187 D. In other words, Perl considers parentheses that do not contain any |
1188 characters to be part of a surrounding alternative, whereas PCRE was
1189 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1190 differently. PCRE now behaves in the same way as Perl, except in the case
1191 of subroutine/recursion calls such as (?1) which have in any case always
1192 been different (but PCRE had them first :-).
1193
1194 16. Related to 15 above: Perl does not treat the | in a conditional group as
1195 creating alternatives. Such a group is treated in the same way as an
1196 ordinary group without any | characters when processing (*THEN). PCRE has
1197 been changed to match Perl's behaviour.
1198
1199 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1200 RunGrepTest script failed.
1201
1202 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1203 inevitable for groups that contain captures, but it can lead to a lot of
1204 stack use in large patterns. The old behaviour has been restored for atomic
1205 groups that do not contain any capturing parentheses.
1206
1207 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1208 suppress the check for a minimum subject length at run time. (If it was
1209 given to pcre_exec() or pcre_dfa_exec() it did work.)
1210
1211 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1212 fail to work when decoding hex characters in data strings in EBCDIC
1213 environments.
1214
1215 21. It appears that in at least one Mac OS environment, the isxdigit() function
1216 is implemented as a macro that evaluates to its argument more than once,
1217 contravening the C 90 Standard (I haven't checked a later standard). There
1218 was an instance in pcretest which caused it to go wrong when processing
1219 \x{...} escapes in subject strings. The has been rewritten to avoid using
1220 things like p++ in the argument of isxdigit().
1221
1222
1223 Version 8.13 16-Aug-2011
1224 ------------------------
1225
1226 1. The Unicode data tables have been updated to Unicode 6.0.0.
1227
1228 2. Two minor typos in pcre_internal.h have been fixed.
1229
1230 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1231 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1232 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1233
1234 4. There were a number of related bugs in the code for matching backrefences
1235 caselessly in UTF-8 mode when codes for the characters concerned were
1236 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1237 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1238 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1239 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1240 2-byte code at the end of the subject (it thought there wasn't enough data
1241 left).
1242
1243 5. Comprehensive information about what went wrong is now returned by
1244 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1245 as the output vector has at least 2 elements. The offset of the start of
1246 the failing character and a reason code are placed in the vector.
1247
1248 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1249 now returned is for the first byte of the failing character, instead of the
1250 last byte inspected. This is an incompatible change, but I hope it is small
1251 enough not to be a problem. It makes the returned offset consistent with
1252 pcre_exec() and pcre_dfa_exec().
1253
1254 7. pcretest now gives a text phrase as well as the error number when
1255 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1256 failure, the offset and reason code are output.
1257
1258 8. When \R was used with a maximizing quantifier it failed to skip backwards
1259 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1260 back over a single character (\n). This seems wrong (because it treated the
1261 two characters as a single entity when going forwards), conflicts with the
1262 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1263 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1264 has been changed.
1265
1266 9. Some internal refactoring has changed the processing so that the handling
1267 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1268 time (the PCRE_DOTALL option was changed this way some time ago: version
1269 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1270 which was always a bit of a fudge. It also means that there is one less
1271 argument for the match() function, which reduces its stack requirements
1272 slightly. This change also fixes an incompatibility with Perl: the pattern
1273 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1274
1275 10. More internal refactoring has drastically reduced the number of recursive
1276 calls to match() for possessively repeated groups such as (abc)++ when
1277 using pcre_exec().
1278
1279 11. While implementing 10, a number of bugs in the handling of groups were
1280 discovered and fixed:
1281
1282 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1283 (a|)*(?1) gave a compile-time internal error.
1284 ((a|)+)+ did not notice that the outer group could match an empty string.
1285 (^a|^)+ was not marked as anchored.
1286 (.*a|.*)+ was not marked as matching at start or after a newline.
1287
1288 12. Yet more internal refactoring has removed another argument from the match()
1289 function. Special calls to this function are now indicated by setting a
1290 value in a variable in the "match data" data block.
1291
1292 13. Be more explicit in pcre_study() instead of relying on "default" for
1293 opcodes that mean there is no starting character; this means that when new
1294 ones are added and accidentally left out of pcre_study(), testing should
1295 pick them up.
1296
1297 14. The -s option of pcretest has been documented for ages as being an old
1298 synonym of -m (show memory usage). I have changed it to mean "force study
1299 for every regex", that is, assume /S for every regex. This is similar to -i
1300 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1301 using it. It makes it easier to run collections of tests with and without
1302 study enabled, and thereby test pcre_study() more easily. All the standard
1303 tests are now run with and without -s (but some patterns can be marked as
1304 "never study" - see 20 below).
1305
1306 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1307 restoration of the capturing data to the outer values was not happening
1308 correctly.
1309
1310 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1311 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1312 pattern had matched an empty string, and so incorrectly returned a no
1313 match.
1314
1315 17. There was optimizing code for the last branch of non-capturing parentheses,
1316 and also for the obeyed branch of a conditional subexpression, which used
1317 tail recursion to cut down on stack usage. Unfortunately, now that there is
1318 the possibility of (*THEN) occurring in these branches, tail recursion is
1319 no longer possible because the return has to be checked for (*THEN). These
1320 two optimizations have therefore been removed. [But see 8.20/11 above.]
1321
1322 18. If a pattern containing \R was studied, it was assumed that \R always
1323 matched two bytes, thus causing the minimum subject length to be
1324 incorrectly computed because \R can also match just one byte.
1325
1326 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1327 was incorrectly computed.
1328
1329 20. If /S is present twice on a test pattern in pcretest input, it now
1330 *disables* studying, thereby overriding the use of -s on the command line
1331 (see 14 above). This is necessary for one or two tests to keep the output
1332 identical in both cases.
1333
1334 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1335 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1336
1337 22. When an atomic group that contained a capturing parenthesis was
1338 successfully matched, but the branch in which it appeared failed, the
1339 capturing was not being forgotten if a higher numbered group was later
1340 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1341 group 1 to "a", when in fact it should be unset. This applied to multi-
1342 branched capturing and non-capturing groups, repeated or not, and also to
1343 positive assertions (capturing in negative assertions does not happen
1344 in PCRE) and also to nested atomic groups.
1345
1346 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1347 subject after a captured substring, to make it easier to tell which of a
1348 number of identical substrings has been captured.
1349
1350 24. The way atomic groups are processed by pcre_exec() has been changed so that
1351 if they are repeated, backtracking one repetition now resets captured
1352 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1353 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1354 "aaa". Previously, it would have been "a". As part of this code
1355 refactoring, the way recursive calls are handled has also been changed.
1356
1357 25. If an assertion condition captured any substrings, they were not passed
1358 back unless some other capturing happened later. For example, if
1359 (?(?=(a))a) was matched against "a", no capturing was returned.
1360
1361 26. When studying a pattern that contained subroutine calls or assertions,
1362 the code for finding the minimum length of a possible match was handling
1363 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1364 group 1 called group 2 while simultaneously a separate group 2 called group
1365 1). A stack overflow occurred in this case. I have fixed this by limiting
1366 the recursion depth to 10.
1367
1368 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1369 Fortmann. This supports explicit test numbers on the command line, and has
1370 argument validation and error reporting.
1371
1372 28. An instance of \X with an unlimited repeat could fail if at any point the
1373 first character it looked at was a mark character.
1374
1375 29. Some minor code refactoring concerning Unicode properties and scripts
1376 should reduce the stack requirement of match() slightly.
1377
1378 30. Added the '=' option to pcretest to check the setting of unused capturing
1379 slots at the end of the pattern, which are documented as being -1, but are
1380 not included in the return count.
1381
1382 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1383 compiled something random. Now it gives a compile-time error (as does
1384 Perl).
1385
1386 32. A *MARK encountered during the processing of a positive assertion is now
1387 recorded and passed back (compatible with Perl).
1388
1389 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1390 had alternative anchored branches, the search for a second match in a line
1391 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1392 matched the line "0102" twice. The same bug affected patterns that started
1393 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1394 twice.
1395
1396 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1397 does, and because of capturing effects, quantifying parenthesized
1398 assertions may at times be useful. Quantifiers are now allowed for
1399 parenthesized assertions.
1400
1401 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1402
1403 36. \g was being checked for fancy things in a character class, when it should
1404 just be a literal "g".
1405
1406 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1407 appearance of a nested POSIX class supersedes an apparent external class.
1408 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1409 unescaped square brackets may also appear as part of class names. For
1410 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1411 more like Perl. (But see 8.20/1 above.)
1412
1413 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1414 was because it thought it was \N{name}, which is not supported).
1415
1416 39. Add minix to OS list not supporting the -S option in pcretest.
1417
1418 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1419 cannot analyze patterns in sufficient detail to catch mutual recursions
1420 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1421 subgroup is called recursively as a subpattern for a second time at the
1422 same position in the subject string. In previous releases this might have
1423 been caught by the recursion limit, or it might have run out of stack.
1424
1425 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1426 happen only once. PCRE was, however incorrectly giving a compile time error
1427 "recursive call could loop indefinitely" because it cannot analyze the
1428 pattern in sufficient detail. The compile time test no longer happens when
1429 PCRE is compiling a conditional subpattern, but actual runaway loops are
1430 now caught at runtime (see 40 above).
1431
1432 42. It seems that Perl allows any characters other than a closing parenthesis
1433 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1434 has been changed to be the same.
1435
1436 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1437 as not to get warnings when autogen.sh is called. Also changed
1438 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1439
1440 44. To help people who use pcregrep to scan files containing exceedingly long
1441 lines, the following changes have been made:
1442
1443 (a) The default value of the buffer size parameter has been increased from
1444 8K to 20K. (The actual buffer used is three times this size.)
1445
1446 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1447 PCRE is built.
1448
1449 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1450 to be set at run time.
1451
1452 (d) Numerical values in pcregrep options can be followed by K or M, for
1453 example --buffer-size=50K.
1454
1455 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1456 given and the return code is set to 2.
1457
1458 45. Add a pointer to the latest mark to the callout data block.
1459
1460 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1461 partial match of an empty string instead of no match. This was specific to
1462 the use of ".".
1463
1464 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1465 complete match instead of a partial match. This bug was dependent on both
1466 the PCRE_UTF8 and PCRE_DOTALL options being set.
1467
1468 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1469 starting byte set, because \b was not being ignored.
1470
1471
1472 Version 8.12 15-Jan-2011
1473 ------------------------
1474
1475 1. Fixed some typos in the markup of the man pages, and wrote a script that
1476 checks for such things as part of the documentation building process.
1477
1478 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1479 --match-limit and --recursion-limit options (added for 8.11). In
1480 particular, this made one of the standard tests fail. (The integer value
1481 went into the wrong half of a long int.)
1482
1483 3. If the --colour option was given to pcregrep with -v (invert match), it
1484 did strange things, either producing crazy output, or crashing. It should,
1485 of course, ignore a request for colour when reporting lines that do not
1486 match.
1487
1488 4. Another pcregrep bug caused similar problems if --colour was specified with
1489 -M (multiline) and the pattern match finished with a line ending.
1490
1491 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1492 matched in multiline mode, the following line was shown as part of the
1493 match. This seems wrong, so I have changed it.
1494
1495 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1496 the check for further matches in the same line (so they could be coloured)
1497 to overrun the end of the current line. If another match was found, it was
1498 incorrectly shown (and then shown again when found in the next line).
1499
1500 7. If pcregrep was compiled under Windows, there was a reference to the
1501 function pcregrep_exit() before it was defined. I am assuming this was
1502 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1503 reported by a user. I've moved the definition above the reference.
1504
1505
1506 Version 8.11 10-Dec-2010
1507 ------------------------
1508
1509 1. (*THEN) was not working properly if there were untried alternatives prior
1510 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1511 backtracked to try for "b" instead of moving to the next alternative branch
1512 at the same level (in this case, to look for "c"). The Perl documentation
1513 is clear that when (*THEN) is backtracked onto, it goes to the "next
1514 alternative in the innermost enclosing group".
1515
1516 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1517 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1518 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1519 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1520 (*THEN).
1521
1522 3. If \s appeared in a character class, it removed the VT character from
1523 the class, even if it had been included by some previous item, for example
1524 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1525 of \s, but is part of the POSIX "space" class.)
1526
1527 4. A partial match never returns an empty string (because you can always
1528 match an empty string at the end of the subject); however the checking for
1529 an empty string was starting at the "start of match" point. This has been
1530 changed to the "earliest inspected character" point, because the returned
1531 data for a partial match starts at this character. This means that, for
1532 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1533 (previously it gave "no match").
1534
1535 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1536 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1537 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1538 has an implication that the given string is incomplete (because a partial
1539 match is preferred over a full match). For this reason, these items now
1540 give a partial match in this situation. [Aside: previously, the one case
1541 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1542 match rather than a full match, which was wrong by the old rules, but is
1543 now correct.]
1544
1545 6. There was a bug in the handling of #-introduced comments, recognized when
1546 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1547 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1548 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1549 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1550 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1551 places in pcre_compile().
1552
1553 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1554 comments when looking ahead for named forward references to subpatterns,
1555 the only newline sequence it recognized was NL. It now handles newlines
1556 according to the set newline convention.
1557
1558 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1559 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1560 cater for a lack of strerror(). These oversights have been fixed.
1561
1562 9. Added --match-limit and --recursion-limit to pcregrep.
1563
1564 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1565
1566 11. When the -o option was used, pcregrep was setting a return code of 1, even
1567 when matches were found, and --line-buffered was not being honoured.
1568
1569 12. Added an optional parentheses number to the -o and --only-matching options
1570 of pcregrep.
1571
1572 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1573 can match an empty string. The code to do it in pcretest and pcredemo
1574 needed fixing:
1575
1576 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1577 only one byte after an empty string match just before CRLF (this case
1578 just got forgotten; "any" and "anycrlf" were OK).
1579
1580 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1581 mode when an empty string match preceded an ASCII character followed by
1582 a non-ASCII character. (The code for advancing by one character rather
1583 than one byte was nonsense.)
1584
1585 (c) The pcredemo.c sample program did not have any code at all to handle
1586 the cases when CRLF is a valid newline sequence.
1587
1588 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1589 as a starting offset was within the subject string. There is now a new
1590 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1591 negative or greater than the length of the string. In order to test this,
1592 pcretest is extended to allow the setting of negative starting offsets.
1593
1594 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1595 starting offset points to the beginning of a UTF-8 character was
1596 unnecessarily clumsy. I tidied it up.
1597
1598 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1599 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1600
1601 17. Nobody had reported that the --include_dir option, which was added in
1602 release 7.7 should have been called --include-dir (hyphen, not underscore)
1603 for compatibility with GNU grep. I have changed it to --include-dir, but
1604 left --include_dir as an undocumented synonym, and the same for
1605 --exclude-dir, though that is not available in GNU grep, at least as of
1606 release 2.5.4.
1607
1608 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1609 characters from a string of bytes) have been redefined so as not to use
1610 loops, in order to improve performance in some environments. At the same
1611 time, I abstracted some of the common code into auxiliary macros to save
1612 repetition (this should not affect the compiled code).
1613
1614 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1615 compile-time error is now given if \c is not followed by an ASCII
1616 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1617 different, and any byte value is allowed.)
1618
1619 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1620 START_OPTIMIZE option, which is now allowed at compile time - but just
1621 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1622 to pcregrep and other applications that have no direct access to PCRE
1623 options. The new /Y option in pcretest sets this option when calling
1624 pcre_compile().
1625
1626 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1627 back references. Groups containing recursive back references were forced to
1628 be atomic by that change, but in the case of named groups, the amount of
1629 memory required was incorrectly computed, leading to "Failed: internal
1630 error: code overflow". This has been fixed.
1631
1632 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1633 pcretest.c, to avoid build problems in some Borland environments.
1634
1635
1636 Version 8.10 25-Jun-2010
1637 ------------------------
1638
1639 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1640 THEN.
1641
1642 2. (*ACCEPT) was not working when inside an atomic group.
1643
1644 3. Inside a character class, \B is treated as a literal by default, but
1645 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1646 causes the error). The code is unchanged, but I tidied the documentation.
1647
1648 4. Inside a character class, PCRE always treated \R and \X as literals,
1649 whereas Perl faults them if its -w option is set. I have changed PCRE so
1650 that it faults them when PCRE_EXTRA is set.
1651
1652 5. Added support for \N, which always matches any character other than
1653 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1654
1655 6. When compiling pcregrep with newer versions of gcc which may have
1656 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1657 declared with attribute warn_unused_result" were given. Just casting the
1658 result to (void) does not stop the warnings; a more elaborate fudge is
1659 needed. I've used a macro to implement this.
1660
1661 7. Minor change to pcretest.c to avoid a compiler warning.
1662
1663 8. Added four artifical Unicode properties to help with an option to make
1664 \s etc use properties (see next item). The new properties are: Xan
1665 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1666
1667 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1668 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1669 this option. Modified pcretest to add /W to test this facility. Added
1670 REG_UCP to make it available via the POSIX interface.
1671
1672 10. Added --line-buffered to pcregrep.
1673
1674 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1675 studied, and the match started with a letter with a code point greater than
1676 127 whose first byte was different to the first byte of the other case of
1677 the letter, the other case of this starting letter was not recognized
1678 (#976).
1679
1680 12. If a pattern that was studied started with a repeated Unicode property
1681 test, for example, \p{Nd}+, there was the theoretical possibility of
1682 setting up an incorrect bitmap of starting bytes, but fortunately it could
1683 not have actually happened in practice until change 8 above was made (it
1684 added property types that matched character-matching opcodes).
1685
1686 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1687 possible starting bytes for non-anchored patterns.
1688
1689 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1690 \R, and also a number of cases that involve Unicode properties, both
1691 explicit and implicit when PCRE_UCP is set.
1692
1693 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1694 input, it could crash or give wrong results if characters with values
1695 greater than 0xc0 were present in the subject string. (Detail: it assumed
1696 UTF-8 input when processing these items.)
1697
1698 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1699 size_t is 64-bit (#991).
1700
1701 17. Added a check for running out of memory when PCRE is compiled with
1702 --disable-stack-for-recursion (#990).
1703
1704 18. If the last data line in a file for pcretest does not have a newline on
1705 the end, a newline was missing in the output.
1706
1707 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1708 less than 128) in its various bitmaps. However, there is a facility for
1709 generating tables according to the current locale when PCRE is compiled. It
1710 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1711 characters, are recognized by isspace() and therefore were getting set in
1712 these tables, and indeed these tables seem to approximate to ISO 8859. This
1713 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1714 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1715 which of course cannot start UTF-8 characters. I have changed the code so
1716 that only real ASCII characters (less than 128) and the correct starting
1717 bytes for UTF-8 encodings are set for characters greater than 127 when in
1718 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1719 altogether.)
1720
1721 20. Added the /T option to pcretest so as to be able to run tests with non-
1722 standard character tables, thus making it possible to include the tests
1723 used for 19 above in the standard set of tests.
1724
1725 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1726 reference to a subpattern the other side of a comment that contains an
1727 opening parenthesis caused either an internal compiling error, or a
1728 reference to the wrong subpattern.
1729
1730
1731 Version 8.02 19-Mar-2010
1732 ------------------------
1733
1734 1. The Unicode data tables have been updated to Unicode 5.2.0.
1735
1736 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1737 configured.
1738
1739 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1740 original author of that file, following a query about its status.
1741
1742 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1743 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1744
1745 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1746 quantifier applied to a forward-referencing subroutine call, could compile
1747 incorrect code or give the error "internal error: previously-checked
1748 referenced subpattern not found".
1749
1750 6. Both MS Visual Studio and Symbian OS have problems with initializing
1751 variables to point to external functions. For these systems, therefore,
1752 pcre_malloc etc. are now initialized to local functions that call the
1753 relevant global functions.
1754
1755 7. There were two entries missing in the vectors called coptable and poptable
1756 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1757 I've fixed the data, and added a kludgy way of testing at compile time that
1758 the lengths are correct (equal to the number of opcodes).
1759
1760 8. Following on from 7, I added a similar kludge to check the length of the
1761 eint vector in pcreposix.c.
1762
1763 9. Error texts for pcre_compile() are held as one long string to avoid too
1764 much relocation at load time. To find a text, the string is searched,
1765 counting zeros. There was no check for running off the end of the string,
1766 which could happen if a new error number was added without updating the
1767 string.
1768
1769 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1770
1771 11. \K was not working if it appeared in an atomic group or in a group that
1772 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1773 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1774 the assertion is positive, but not if it is negative.
1775
1776 12. Change 11 fortuitously reduced the size of the stack frame used in the
1777 "match()" function of pcre_exec.c by one pointer. Forthcoming
1778 implementation of support for (*MARK) will need an extra pointer on the
1779 stack; I have reserved it now, so that the stack frame size does not
1780 decrease.
1781
1782 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1783 item in branch that calls a recursion is a subroutine call - as in the
1784 second branch in the above example - was incorrectly given the compile-
1785 time error "recursive call could loop indefinitely" because pcre_compile()
1786 was not correctly checking the subroutine for matching a non-empty string.
1787
1788 14. The checks for overrunning compiling workspace could trigger after an
1789 overrun had occurred. This is a "should never occur" error, but it can be
1790 triggered by pathological patterns such as hundreds of nested parentheses.
1791 The checks now trigger 100 bytes before the end of the workspace.
1792
1793 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1794
1795
1796 Version 8.01 19-Jan-2010
1797 ------------------------
1798
1799 1. If a pattern contained a conditional subpattern with only one branch (in
1800 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1801 computed the wrong minimum data length (which is of course zero for such
1802 subpatterns). This could cause incorrect "no match" results.
1803
1804 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1805 the pattern is reset in the first branch, pcre_compile() failed with
1806 "internal error: code overflow at offset...". This happened only when
1807 the reset was to the original external option setting. (An optimization
1808 abstracts leading options settings into an external setting, which was the
1809 cause of this.)
1810
1811 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1812 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1813 assertion pattern did not match (meaning that the assertion was true), it
1814 was incorrectly treated as false if the SKIP had been reached during the
1815 matching. This also applied to assertions used as conditions.
1816
1817 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1818 assertion subpattern, including such a pattern used as a condition,
1819 unpredictable results occurred, instead of the error return
1820 PCRE_ERROR_DFA_UITEM.
1821
1822 5. The C++ GlobalReplace function was not working like Perl for the special
1823 situation when an empty string is matched. It now does the fancy magic
1824 stuff that is necessary.
1825
1826 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1827 removed. (These were left over from very, very early versions of PCRE.)
1828
1829 7. Some cosmetic changes to the code to make life easier when compiling it
1830 as part of something else:
1831
1832 (a) Change DEBUG to PCRE_DEBUG.
1833
1834 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1835 called "current" as "current_branch", to prevent a collision with the
1836 Linux macro when compiled as a kernel module.
1837
1838 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1839 prevent a collision with the Linux macro when compiled as a kernel
1840 module.
1841
1842 8. In pcre_compile() there are some checks for integer overflows that used to
1843 cast potentially large values to (double). This has been changed to that
1844 when building, a check for int64_t is made, and if it is found, it is used
1845 instead, thus avoiding the use of floating point arithmetic. (There is no
1846 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1847 double.
1848
1849 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1850 2005 (difference between two addresses compared to an unsigned value).
1851
1852 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1853 custom one, because of the following reported problem in Windows:
1854
1855 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1856 under Win32.
1857 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1858 therefore missing the function definition.
1859 - The compiler thus generates a "C" signature for the test function.
1860 - The linker fails to find the "C" function.
1861 - PCRE fails to configure if asked to do so against libbz2.
1862
1863 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1864 messages were output:
1865
1866 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1867 rerunning libtoolize, to keep the correct libtool macros in-tree.
1868 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1869
1870 I have done both of these things.
1871
1872 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1873 most of the time, it *can* run out if it is given a pattern that contains a
1874 runaway infinite recursion. I updated the discussion in the pcrestack man
1875 page.
1876
1877 13. Now that we have gone to the x.xx style of version numbers, the minor
1878 version may start with zero. Using 08 or 09 is a bad idea because users
1879 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1880 interpreted as invalid octal numbers. I've updated the previous comment in
1881 configure.ac, and also added a check that gives an error if 08 or 09 are
1882 used.
1883
1884 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1885 causing partial matching to fail when the end of the subject matched \W
1886 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1887
1888 15. There were some discrepancies between the declarations in pcre_internal.h
1889 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1890 their definitions. The declarations used "const uschar *" and the
1891 definitions used USPTR. Even though USPTR is normally defined as "const
1892 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1893 reported that: "This difference in casting confuses some C++ compilers, for
1894 example, SunCC recognizes above declarations as different functions and
1895 generates broken code for hbpcre." I have changed the declarations to use
1896 USPTR.
1897
1898 16. GNU libtool is named differently on some systems. The autogen.sh script now
1899 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1900 (FreeBSD).
1901
1902 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1903 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1904 comment: "Figure out how to create a longlong from a string: strtoll and
1905 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1906 instance, but it only takes 2 args instead of 3!"
1907
1908 18. A subtle bug concerned with back references has been fixed by a change of
1909 specification, with a corresponding code fix. A pattern such as
1910 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1911 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1912 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1913 same bug. Such groups have to be quantified to be useful, or contained
1914 inside another quantified group. (If there's no repetition, the reference
1915 can never match.) The problem arises because, having left the group and
1916 moved on to the rest of the pattern, a later failure that backtracks into
1917 the group uses the captured value from the final iteration of the group
1918 rather than the correct earlier one. I have fixed this in PCRE by forcing
1919 any group that contains a reference to itself to be an atomic group; that
1920 is, there cannot be any backtracking into it once it has completed. This is
1921 similar to recursive and subroutine calls.
1922
1923
1924 Version 8.00 19-Oct-09
1925 ----------------------
1926
1927 1. The table for translating pcre_compile() error codes into POSIX error codes
1928 was out-of-date, and there was no check on the pcre_compile() error code
1929 being within the table. This could lead to an OK return being given in
1930 error.
1931
1932 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1933 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1934 in a Windows environment.
1935
1936 3. The pcregrep --count option prints the count for each file even when it is
1937 zero, as does GNU grep. However, pcregrep was also printing all files when
1938 --files-with-matches was added. Now, when both options are given, it prints
1939 counts only for those files that have at least one match. (GNU grep just
1940 prints the file name in this circumstance, but including the count seems
1941 more useful - otherwise, why use --count?) Also ensured that the
1942 combination -clh just lists non-zero counts, with no names.
1943
1944 4. The long form of the pcregrep -F option was incorrectly implemented as
1945 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1946 but it seems right to fix it, and I didn't think it was worth preserving
1947 the old behaviour.
1948
1949 5. The command line items --regex=pattern and --regexp=pattern were not
1950 recognized by pcregrep, which required --regex pattern or --regexp pattern
1951 (with a space rather than an '='). The man page documented the '=' forms,
1952 which are compatible with GNU grep; these now work.
1953
1954 6. No libpcreposix.pc file was created for pkg-config; there was just
1955 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1956
1957 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1958 when UCP support is not needed, by modifying the Python script that
1959 generates it from Unicode data files. This should not matter if the module
1960 is correctly used as a library, but I received one complaint about 50K of
1961 unwanted data. My guess is that the person linked everything into his
1962 program rather than using a library. Anyway, it does no harm.
1963
1964 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1965 was a minimum greater than 1 for a wide character in a possessive
1966 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1967 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1968 character. Chaos in the form of incorrect output or a compiling loop could
1969 result.
1970
1971 9. The restrictions on what a pattern can contain when partial matching is
1972 requested for pcre_exec() have been removed. All patterns can now be
1973 partially matched by this function. In addition, if there are at least two
1974 slots in the offset vector, the offset of the earliest inspected character
1975 for the match and the offset of the end of the subject are set in them when
1976 PCRE_ERROR_PARTIAL is returned.
1977
1978 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1979 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1980 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1981 and may be more useful for multi-segment matching.
1982
1983 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1984 used to be given if ever the end of the subject was reached; now it is
1985 given only if matching could not proceed because another character was
1986 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1987 string "Z", which now yields "no match" instead of "partial match". In the
1988 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1989 final character ended with (*FAIL).
1990
1991 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1992 if the pattern had a "must contain" character that was already found in the
1993 earlier partial match, unless partial matching was again requested. For
1994 example, with the pattern /dog.(body)?/, the "must contain" character is
1995 "g". If the first part-match was for the string "dog", restarting with
1996 "sbody" failed. This bug has been fixed.
1997
1998 13. The string returned by pcre_dfa_exec() after a partial match has been
1999 changed so that it starts at the first inspected character rather than the
2000 first character of the match. This makes a difference only if the pattern
2001 starts with a lookbehind assertion or \b or \B (\K is not supported by
2002 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2003 matching functions compatible, and I think it's the right thing to do.
2004
2005 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2006 so that the demonstration program is easily available in environments where
2007 PCRE has not been installed from source.
2008
2009 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2010 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2011 library.
2012
2013 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2014 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2015 is not the first non-POSIX option to be added. Clearly some people find
2016 these options useful.
2017
2018 17. If a caller to the POSIX matching function regexec() passes a non-zero
2019 value for nmatch with a NULL value for pmatch, the value of
2020 nmatch is forced to zero.
2021
2022 18. RunGrepTest did not have a test for the availability of the -u option of
2023 the diff command, as RunTest does. It now checks in the same way as
2024 RunTest, and also checks for the -b option.
2025
2026 19. If an odd number of negated classes containing just a single character
2027 interposed, within parentheses, between a forward reference to a named
2028 subpattern and the definition of the subpattern, compilation crashed with
2029 an internal error, complaining that it could not find the referenced
2030 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2031 [The bug was that it was starting one character too far in when skipping
2032 over the character class, thus treating the ] as data rather than
2033 terminating the class. This meant it could skip too much.]
2034
2035 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2036 /g option in pcretest when the pattern contains \K, which makes it possible
2037 to have an empty string match not at the start, even when the pattern is
2038 anchored. Updated pcretest and pcredemo to use this option.
2039
2040 21. If the maximum number of capturing subpatterns in a recursion was greater
2041 than the maximum at the outer level, the higher number was returned, but
2042 with unset values at the outer level. The correct (outer level) value is
2043 now given.
2044
2045 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2046 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2047 make it do so. The string so far is captured, making this feature
2048 compatible with Perl.
2049
2050 23. The tests have been re-organized, adding tests 11 and 12, to make it
2051 possible to check the Perl 5.10 features against Perl 5.10.
2052
2053 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2054 pattern matches a fixed length string. PCRE did not allow this; now it
2055 does. Neither allows recursion.
2056
2057 25. I finally figured out how to implement a request to provide the minimum
2058 length of subject string that was needed in order to match a given pattern.
2059 (It was back references and recursion that I had previously got hung up
2060 on.) This code has now been added to pcre_study(); it finds a lower bound
2061 to the length of subject needed. It is not necessarily the greatest lower
2062 bound, but using it to avoid searching strings that are too short does give
2063 some useful speed-ups. The value is available to calling programs via
2064 pcre_fullinfo().
2065
2066 26. While implementing 25, I discovered to my embarrassment that pcretest had
2067 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2068 study optimizations had never been tested with that matching function.
2069 Oops. What is worse, even when it was passed study data, there was a bug in
2070 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2071 were also very few tests of studied patterns with pcre_dfa_exec().
2072
2073 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2074 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2075 on the other side of the coin, they are no longer allowed to have different
2076 names, because these cannot be distinguished in PCRE, and this has caused
2077 confusion. (This is a difference from Perl.)
2078
2079 28. When duplicate subpattern names are present (necessarily with different
2080 numbers, as required by 27 above), and a test is made by name in a
2081 conditional pattern, either for a subpattern having been matched, or for
2082 recursion in such a pattern, all the associated numbered subpatterns are
2083 tested, and the overall condition is true if the condition is true for any
2084 one of them. This is the way Perl works, and is also more like the way
2085 testing by number works.
2086
2087
2088 Version 7.9 11-Apr-09
2089 ---------------------
2090
2091 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2092 (pcretest), all targets were linked against these libraries. This included
2093 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2094 libraries. This caused unwanted dependencies to be created. This problem
2095 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2096 pcretest is linked with readline.
2097
2098 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2099 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2100 moved outside it again, because FALSE and TRUE are already defined in AIX,
2101 but BOOL is not.
2102
2103 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2104 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2105
2106 4. The pcregrep documentation said spaces were inserted as well as colons (or
2107 hyphens) following file names and line numbers when outputting matching
2108 lines. This is not true; no spaces are inserted. I have also clarified the
2109 wording for the --colour (or --color) option.
2110
2111 5. In pcregrep, when --colour was used with -o, the list of matching strings
2112 was not coloured; this is different to GNU grep, so I have changed it to be
2113 the same.
2114
2115 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2116 each matching line was coloured. Now it goes on to look for further matches
2117 of any of the test patterns, which is the same behaviour as GNU grep.
2118
2119 7. A pattern that could match an empty string could cause pcregrep to loop; it
2120 doesn't make sense to accept an empty string match in pcregrep, so I have
2121 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2122 seems to be how GNU grep behaves. [But see later change 40 for release
2123 8.33.]
2124
2125 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2126 start or after a newline", because the conditional assertion was not being
2127 correctly handled. The rule now is that both the assertion and what follows
2128 in the first alternative must satisfy the test.
2129
2130 9. If auto-callout was enabled in a pattern with a conditional group whose
2131 condition was an assertion, PCRE could crash during matching, both with
2132 pcre_exec() and pcre_dfa_exec().
2133
2134 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2135 used for matching.
2136
2137 11. Unicode property support in character classes was not working for
2138 characters (bytes) greater than 127 when not in UTF-8 mode.
2139
2140 12. Added the -M command line option to pcretest.
2141
2142 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2143
2144 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2145
2146 16. Added comments and documentation about mis-use of no_arg in the C++
2147 wrapper.
2148
2149 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2150 from Martin Jerabek that uses macro names for all relevant character and
2151 string constants.
2152
2153 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2154 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2155 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2156 these, but not everybody uses configure.
2157
2158 19. A conditional group that had only one branch was not being correctly
2159 recognized as an item that could match an empty string. This meant that an
2160 enclosing group might also not be so recognized, causing infinite looping
2161 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2162 with the subject "ab", where knowledge that the repeated group can match
2163 nothing is needed in order to break the loop.
2164
2165 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2166 exec(), but without supplying a callout function, matching went wrong.
2167
2168 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2169 leak if the size of the offset vector was greater than 30. When the vector
2170 is smaller, the saved offsets during recursion go onto a local stack
2171 vector, but for larger vectors malloc() is used. It was failing to free
2172 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2173 error, in fact).
2174
2175 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2176 heapframe that is used only when UTF-8 support is enabled. This caused no
2177 problem, but was untidy.
2178
2179 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2180 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2181 included within another project.
2182
2183 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2184 slightly modified by me:
2185
2186 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2187 not building pcregrep.
2188
2189 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2190 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2191
2192 25. Forward references, both numeric and by name, in patterns that made use of
2193 duplicate group numbers, could behave incorrectly or give incorrect errors,
2194 because when scanning forward to find the reference group, PCRE was not
2195 taking into account the duplicate group numbers. A pattern such as
2196 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2197
2198 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2199 the feature of a custom pointer more persuasive (as requested by a user).
2200
2201 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2202 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2203 pre-defined.
2204
2205 28. Added support for (*UTF8) at the start of a pattern.
2206
2207 29. Arrange for flags added by the "release type" setting in CMake to be shown
2208 in the configuration summary.
2209
2210
2211 Version 7.8 05-Sep-08
2212 ---------------------
2213
2214 1. Replaced UCP searching code with optimized version as implemented for Ad
2215 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2216 stage table and inline lookup instead of a function, giving speed ups of 2
2217 to 5 times on some simple patterns that I tested. Permission was given to
2218 distribute the MultiStage2.py script that generates the tables (it's not in
2219 the tarball, but is in the Subversion repository).
2220
2221 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2222 scripts.
2223
2224 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2225 a group with a zero qualifier. The result of the study could be incorrect,
2226 or the function might crash, depending on the pattern.
2227
2228 4. Caseless matching was not working for non-ASCII characters in back
2229 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2230 It now works when Unicode Property Support is available.
2231
2232 5. In pcretest, an escape such as \x{de} in the data was always generating
2233 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2234 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2235 truncation.
2236
2237 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2238
2239 7. Added two (int) casts to pcregrep when printing the difference of two
2240 pointers, in case they are 64-bit values.
2241
2242 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2243 test 2 if it fails.
2244
2245 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2246 and a #define of that name to empty if it is not externally set. This is to
2247 allow users of MSVC to set it if necessary.
2248
2249 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2250 the convenience functions in the pcre_get.c source file.
2251
2252 11. An option change at the start of a pattern that had top-level alternatives
2253 could cause overwriting and/or a crash. This command provoked a crash in
2254 some environments:
2255
2256 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2257
2258 This potential security problem was recorded as CVE-2008-2371.
2259
2260 12. For a pattern where the match had to start at the beginning or immediately
2261 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2262 pcre_dfa_exec() could read past the end of the passed subject if there was
2263 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2264 pcretest so that it places the subject at the end of its malloc-ed buffer.
2265
2266 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2267 exec() might read past the end of the data buffer in UTF-8 mode.
2268
2269 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2270 the data contained the byte 0x85 as part of a UTF-8 character within its
2271 first line. This applied both to normal and DFA matching.
2272
2273 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2274 /^[^d]*?$/8 failed to match "abc".
2275
2276 16. Added a missing copyright notice to pcrecpp_internal.h.
2277
2278 17. Make it more clear in the documentation that values returned from
2279 pcre_exec() in ovector are byte offsets, not character counts.
2280
2281 18. Tidied a few places to stop certain compilers from issuing warnings.
2282
2283 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2284 supplied by Stefan Weber. I made a further small update for 7.8 because
2285 there is a change of source arrangements: the pcre_searchfuncs.c module is
2286 replaced by pcre_ucd.c.
2287
2288
2289 Version 7.7 07-May-08
2290 ---------------------
2291
2292 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2293 a string to a long long, pretend we don't even have a long long." This is
2294 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2295
2296 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2297 pre-7.6 versions, which defined a global no_arg variable instead of putting
2298 it in the RE class. (See also #8 below.)
2299
2300 3. Remove a line of dead code, identified by coverity and reported by Nuno
2301 Lopes.
2302
2303 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2304
2305 (1) The include/exclude patterns were being applied to the whole pathnames
2306 of files, instead of just to the final components.
2307
2308 (2) If there was more than one level of directory, the subdirectories were
2309 skipped unless they satisfied the include/exclude conditions. This is
2310 inconsistent with GNU grep (and could even be seen as contrary to the
2311 pcregrep specification - which I improved to make it absolutely clear).
2312 The action now is always to scan all levels of directory, and just
2313 apply the include/exclude patterns to regular files.
2314
2315 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2316 --exclude_dir in the tests to avoid scanning .svn directories.
2317
2318 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2319 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2320 doesn't support NULs in patterns.
2321
2322 7. Added some missing "const"s to declarations of static tables in
2323 pcre_compile.c and pcre_dfa_exec.c.
2324
2325 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2326 caused by fix #2 above. (Subsequently also a second patch to fix the
2327 first patch. And a third patch - this was a messy problem.)
2328
2329 9. Applied Craig's patch to remove the use of push_back().
2330
2331 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2332 matching function regexec().
2333
2334 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2335 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2336 references. PCRE supports relative numbers with this syntax (I don't think
2337 Oniguruma does).
2338
2339 12. Previously, a group with a zero repeat such as (...){0} was completely
2340 omitted from the compiled regex. However, this means that if the group
2341 was called as a subroutine from elsewhere in the pattern, things went wrong
2342 (an internal error was given). Such groups are now left in the compiled
2343 pattern, with a new opcode that causes them to be skipped at execution
2344 time.
2345
2346 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2347 to the way PCRE behaves:
2348
2349 (a) A lone ] character is dis-allowed (Perl treats it as data).
2350
2351 (b) A back reference to an unmatched subpattern matches an empty string
2352 (Perl fails the current match path).
2353
2354 (c) A data ] in a character class must be notated as \] because if the
2355 first data character in a class is ], it defines an empty class. (In
2356 Perl it is not possible to have an empty class.) The empty class []
2357 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2358 The negative empty class [^] matches any one character, independently
2359 of the DOTALL setting.
2360
2361 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2362 non-existent subpattern following a character class starting with ']' and
2363 containing () gave an internal compiling error instead of "reference to
2364 non-existent subpattern". Fortunately, when the pattern did exist, the
2365 compiled code was correct. (When scanning forwards to check for the
2366 existence of the subpattern, it was treating the data ']' as terminating
2367 the class, so got the count wrong. When actually compiling, the reference
2368 was subsequently set up correctly.)
2369
2370 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2371 it was being rejected as not supported by pcre_dfa_exec(), even though
2372 other assertions are supported. I have made pcre_dfa_exec() support
2373 (*FAIL).
2374
2375 16. The implementation of 13c above involved the invention of a new opcode,
2376 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2377 cannot be changed at match time, I realized I could make a small
2378 improvement to matching performance by compiling OP_ALLANY instead of
2379 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2380 on the OP_ANY path.
2381
2382 17. Compiling pcretest on Windows with readline support failed without the
2383 following two fixes: (1) Make the unistd.h include conditional on
2384 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2385
2386 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2387 ncurses library to be included for pcretest when ReadLine support is
2388 requested, but also to allow for it to be overridden. This patch came from
2389 Daniel Bergström.
2390
2391 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2392 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2393 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2394 spotting this.
2395
2396
2397 Version 7.6 28-Jan-08
2398 ---------------------
2399
2400 1. A character class containing a very large number of characters with
2401 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2402 overflow.
2403
2404 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2405 HAVE_LONG_LONG is not defined.
2406
2407 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2408 bring them up to date and include new features. This patch includes:
2409
2410 - Fixed PH's badly added libz and libbz2 support.
2411 - Fixed a problem with static linking.
2412 - Added pcredemo. [But later removed - see 7 below.]
2413 - Fixed dftables problem and added an option.
2414 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2415 HAVE_LONG_LONG.
2416 - Added readline support for pcretest.
2417 - Added an listing of the option settings after cmake has run.
2418
2419 4. A user submitted a patch to Makefile that makes it easy to create
2420 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2421 Makefile.am that cause it to include this special target, without
2422 affecting anything else. Note that the same mingw target plus all
2423 the other distribution libraries and programs are now supported
2424 when configuring with CMake (see 6 below) instead of with
2425 Configure/Make.
2426
2427 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2428 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2429 exported in the Windows port". It has not yet been confirmed that the patch
2430 solves the problem, but it does no harm.
2431
2432 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2433 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2434 with CMake, and also correct the comment about stack recursion.
2435
2436 7. Remove the automatic building of pcredemo from the ./configure system and
2437 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2438 of a program that users should build themselves after PCRE is installed, so
2439 building it automatically is not really right. What is more, it gave
2440 trouble in some build environments.
2441
2442 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2443
2444
2445 Version 7.5 10-Jan-08
2446 ---------------------
2447
2448 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2449 values in parens when parsing an RE using the C++ wrapper."
2450
2451 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2452 Characters greater than 255 were excluded from the class instead of being
2453 included.
2454
2455 3. The same bug as (2) above applied to negated POSIX classes such as
2456 [:^space:].
2457
2458 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2459 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2460 I have changed it.
2461
2462 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2463 first named subpattern) and a construct such as (?&a) would reference the
2464 first named subpattern whose name started with "a" (in other words, the
2465 length check was missing). Both these problems are fixed. "Subpattern name
2466 expected" is now given for (?&) (a zero-length name), and this patch also
2467 makes it give the same error for \k'' (previously it complained that that
2468 was a reference to a non-existent subpattern).
2469
2470 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2471 this is right because (?- can be followed by option settings as well as by
2472 digits. I have, however, made the messages clearer.
2473
2474 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2475 than the number used in the conditional) now cause a compile-time error.
2476 This is actually not compatible with Perl, which accepts such patterns, but
2477 treats the conditional as always being FALSE (as PCRE used to), but it
2478 seems to me that giving a diagnostic is better.
2479
2480 8. Change "alphameric" to the more common word "alphanumeric" in comments
2481 and messages.
2482
2483 9. Fix two occurrences of "backslash" in comments that should have been
2484 "backspace".
2485
2486 10. Remove two redundant lines of code that can never be obeyed (their function
2487 was moved elsewhere).
2488
2489 11. The program that makes PCRE's Unicode character property table had a bug
2490 which caused it to generate incorrect table entries for sequences of
2491 characters that have the same character type, but are in different scripts.
2492 It amalgamated them into a single range, with the script of the first of
2493 them. In other words, some characters were in the wrong script. There were
2494 thirteen such cases, affecting characters in the following ranges:
2495
2496 U+002b0 - U+002c1
2497 U+0060c - U+0060d
2498 U+0061e - U+00612
2499 U+0064b - U+0065e
2500 U+0074d - U+0076d
2501 U+01800 - U+01805
2502 U+01d00 - U+01d77
2503 U+01d9b - U+01dbf
2504 U+0200b - U+0200f
2505 U+030fc - U+030fe
2506 U+03260 - U+0327f
2507 U+0fb46 - U+0fbb1
2508 U+10450 - U+1049d
2509
2510 12. The -o option (show only the matching part of a line) for pcregrep was not
2511 compatible with GNU grep in that, if there was more than one match in a
2512 line, it showed only the first of them. It now behaves in the same way as
2513 GNU grep.
2514
2515 13. If the -o and -v options were combined for pcregrep, it printed a blank
2516 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2517 does the same. The return code can be used to tell if there were any
2518 non-matching lines.
2519
2520 14. Added --file-offsets and --line-offsets to pcregrep.
2521
2522 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2523 infinitely looping recursion. The bug was that positive lookaheads were not
2524 being skipped when checking for a possible empty match (negative lookaheads
2525 and both kinds of lookbehind were skipped).
2526
2527 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2528 inclusion of <windows.h> to before rather than after the definition of
2529 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2530
2531 17. Specifying a possessive quantifier with a specific limit for a Unicode
2532 character property caused pcre_compile() to compile bad code, which led at
2533 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2534 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2535 caused the error; without that there was no problem.
2536
2537 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2538
2539 19. Added --enable-pcretest-libreadline.
2540
2541 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2542 RE::GlobalReplace(). As a result, the number of replacements returned was
2543 double what it should be. I removed one of the increments, but Craig sent a
2544 later patch that removed the other one (the right fix) and added unit tests
2545 that check the return values (which was not done before).
2546
2547 21. Several CMake things:
2548
2549 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2550 the names libpcre and libpcreposix, not just pcre and pcreposix.
2551
2552 (2) The above change means that pcretest and pcregrep are now correctly
2553 linked with the newly-built libraries, not previously installed ones.
2554
2555 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2556
2557 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2558 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2559 UTF-8 newline character). The key issue is that the pattern starts .*;
2560 this means that the match must be either at the beginning, or after a
2561 newline. The bug was in the code for advancing after a failed match and
2562 checking that the new position followed a newline. It was not taking
2563 account of UTF-8 characters correctly.
2564
2565 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2566 character classes. PCRE was not treating the sequence [:...:] as a
2567 character class unless the ... were all letters. Perl, however, seems to
2568 allow any characters between [: and :], though of course it rejects as
2569 unknown any "names" that contain non-letters, because all the known class
2570 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2571 for example, whereas PCRE did not - it did not recognize a POSIX character
2572 class. This seemed a bit dangerous, so the code has been changed to be
2573 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2574 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2575 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2576 Perl does, and where it didn't before.
2577
2578 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2579 Windows environments %n is disabled by default.
2580
2581
2582 Version 7.4 21-Sep-07
2583 ---------------------
2584
2585 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2586 means that a class such as [\s] counted as "explicit reference to CR or
2587 LF". That isn't really right - the whole point of the change was to try to
2588 help when there was an actual mention of one of the two characters. So now
2589 the change happens only if \r or \n (or a literal CR or LF) character is
2590 encountered.
2591
2592 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2593 of both had grown to the point where there were only 3 bits left.
2594 Fortunately, there was spare space in the data structure, and so I have
2595 moved the internal flags into a new 16-bit field to free up more option
2596 bits.
2597
2598 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2599 but did not set the internal JCHANGED flag - either of these is enough to
2600 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2601 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2602 start sets both bits.
2603
2604 4. Added options (at build time, compile time, exec time) to change \R from
2605 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2606
2607 5. doc/pcresyntax.html was missing from the distribution.
2608
2609 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2610 compatibility, even though it is no longer used.
2611
2612 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2613 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2614 windows.h file is present (where different names are used). [This was
2615 reversed later after testing - see 16 below.]
2616
2617 8. Changed all #include <config.h> to #include "config.h". There were also
2618 some further <pcre.h> cases that I changed to "pcre.h".
2619
2620 9. When pcregrep was used with the --colour option, it missed the line ending
2621 sequence off the lines that it output.
2622
2623 10. It was pointed out to me that arrays of string pointers cause lots of
2624 relocations when a shared library is dynamically loaded. A technique of
2625 using a single long string with a table of offsets can drastically reduce
2626 these. I have refactored PCRE in four places to do this. The result is
2627 dramatic:
2628
2629 Originally: 290
2630 After changing UCP table: 187
2631 After changing error message table: 43
2632 After changing table of "verbs" 36
2633 After changing table of Posix names 22
2634
2635 Thanks to the folks working on Gregex for glib for this insight.
2636
2637 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2638 unicode-properties was also set.
2639
2640 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2641
2642 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2643 checked only for CRLF.
2644
2645 14. Added casts to pcretest.c to avoid compiler warnings.
2646
2647 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2648
2649 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2650 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2651 entirely. This removes changes made in 7 above.
2652
2653 17. The CMake files have been updated, and there is now more information about
2654 building with CMake in the NON-UNIX-USE document.
2655
2656
2657 Version 7.3 28-Aug-07
2658 ---------------------
2659
2660 1. In the rejigging of the build system that eventually resulted in 7.1, the
2661 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2662 brackets there is not right, since it causes compilers to look for an
2663 installed pcre.h, not the version that is in the source that is being
2664 compiled (which of course may be different). I have changed it back to:
2665
2666 #include "pcre.h"
2667
2668 I have a vague recollection that the change was concerned with compiling in
2669 different directories, but in the new build system, that is taken care of
2670 by the VPATH setting the Makefile.
2671
2672 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2673 when the subject happened to end in the byte 0x85 (e.g. if the last
2674 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2675 characters but of course it shouldn't be taken as a newline when it is part
2676 of another character. The bug was that, for an unlimited repeat of . in
2677 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2678 characters when looking for a newline.
2679
2680 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2681
2682 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2683 in debug output.
2684
2685 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2686 long printing in the pcrecpp unittest when running under MinGW.
2687
2688 6. ESC_K was left out of the EBCDIC table.
2689
2690 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2691 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2692 limit also applies to "virtual nesting" when a pattern is recursive, and in
2693 this case 1000 isn't so big. I have been able to remove this limit at the
2694 expense of backing off one optimization in certain circumstances. Normally,
2695 when pcre_exec() would call its internal match() function recursively and
2696 immediately return the result unconditionally, it uses a "tail recursion"
2697 feature to save stack. However, when a subpattern that can match an empty
2698 string has an unlimited repetition quantifier, it no longer makes this
2699 optimization. That gives it a stack frame in which to save the data for
2700 checking that an empty string has been matched. Previously this was taken
2701 from the 1000-entry workspace that had been reserved. So now there is no
2702 explicit limit, but more stack is used.
2703
2704 8. Applied Daniel's patches to solve problems with the import/export magic
2705 syntax that is required for Windows, and which was going wrong for the
2706 pcreposix and pcrecpp parts of the library. These were overlooked when this
2707 problem was solved for the main library.
2708
2709 9. There were some crude static tests to avoid integer overflow when computing
2710 the size of patterns that contain repeated groups with explicit upper
2711 limits. As the maximum quantifier is 65535, the maximum group length was
2712 set at 30,000 so that the product of these two numbers did not overflow a
2713 32-bit integer. However, it turns out that people want to use groups that
2714 are longer than 30,000 bytes (though not repeat them that many times).
2715 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2716 made it possible to implement the integer overflow checks in a much more
2717 dynamic way, which I have now done. The artificial limitation on group
2718 length has been removed - we now have only the limit on the total length of
2719 the compiled pattern, which depends on the LINK_SIZE setting.
2720
2721 10. Fixed a bug in the documentation for get/copy named substring when
2722 duplicate names are permitted. If none of the named substrings are set, the
2723 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2724 empty string.
2725
2726 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2727 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2728 because the ] is interpreted as the first data character and the
2729 terminating ] is not found. PCRE has been made compatible with Perl in this
2730 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2731 cause memory overwriting.
2732
2733 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2734 string has been matched (to stop an infinite loop). It was not recognizing
2735 a conditional subpattern that could match an empty string if that
2736 subpattern was within another subpattern. For example, it looped when
2737 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2738 condition was not nested. This bug has been fixed.
2739
2740 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2741 past the start of the subject in the presence of bytes with the top bit
2742 set, for example "\x8aBCD".
2743
2744 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2745 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2746
2747 14. Optimized (?!) to (*FAIL).
2748
2749 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2750 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2751 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2752 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2753 does: it's just the validity check that is more restrictive.
2754
2755 16. Inserted checks for integer overflows during escape sequence (backslash)
2756 processing, and also fixed erroneous offset values for syntax errors during
2757 backslash processing.
2758
2759 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2760 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2761
2762 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2763 caused an overrun.
2764
2765 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2766 something other than just ASCII characters) inside a group that had an
2767 unlimited repeat caused a loop at compile time (while checking to see
2768 whether the group could match an empty string).
2769
2770 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2771 [\P{Any}] did so. (Error in the code for printing property names.)
2772
2773 21. An orphan \E inside a character class could cause a crash.
2774
2775 22. A repeated capturing bracket such as (A)? could cause a wild memory
2776 reference during compilation.
2777
2778 23. There are several functions in pcre_compile() that scan along a compiled
2779 expression for various reasons (e.g. to see if it's fixed length for look
2780 behind). There were bugs in these functions when a repeated \p or \P was
2781 present in the pattern. These operators have additional parameters compared
2782 with \d, etc, and these were not being taken into account when moving along
2783 the compiled data. Specifically:
2784
2785 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2786 length.
2787
2788 (b) An item such as \pL+ within a repeated group could cause crashes or
2789 loops.
2790
2791 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2792 "reference to non-existent subpattern" error.
2793
2794 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2795
2796 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2797 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2798
2799 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2800
2801 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2802 character were causing crashes (broken optimization).
2803
2804 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2805 \p or \P) caused a compile-time loop.
2806
2807 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2808 break. For example, the unstudied pattern [\r\n]A does not match the string
2809 "\r\nA" because change 7.0/46 below moves the current point on by two
2810 characters after failing to match at the start. However, the pattern \nA
2811 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2812 the same is true. There doesn't seem any very clean way out of this, but
2813 what I have chosen to do makes the common cases work: PCRE now takes note
2814 of whether there can be an explicit match for \r or \n anywhere in the
2815 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2816 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2817 pattern has explicit CR or LF references.
2818
2819 29. Added (*CR) etc for changing newline setting at start of pattern.
2820
2821
2822 Version 7.2 19-Jun-07
2823 ---------------------
2824
2825 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2826 which is apparently normally available under Windows.
2827
2828 2. Re-jig the pcregrep tests with different newline settings in an attempt
2829 to make them independent of the local environment's newline setting.
2830
2831 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2832
2833 4. Some of the "internals" tests were previously cut out when the link size
2834 was not 2, because the output contained actual offsets. The recent new
2835 "Z" feature of pcretest means that these can be cut out, making the tests
2836 usable with all link sizes.
2837
2838 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2839 stack recursion. This gives a massive performance boost under BSD, but just
2840 a small improvement under Linux. However, it saves one field in the frame
2841 in all cases.
2842
2843 6. Added more features from the forthcoming Perl 5.10:
2844
2845 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2846 recursion call. It refers to the nth most recently opened parentheses.
2847
2848 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2849 to be opened parentheses.
2850
2851 (c) Conditions that refer to capturing parentheses can be specified
2852 relatively, for example, (?(-2)... or (?(+3)...
2853
2854 (d) \K resets the start of the current match so that everything before
2855 is not part of it.
2856
2857 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2858
2859 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2860 reference syntax.
2861
2862 (g) (?| introduces a group in which the numbering of parentheses in each
2863 alternative starts with the same number.
2864
2865 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2866
2867 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2868 PCRE_INFO_JCHANGED.
2869
2870 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2871 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2872 for detecting groups that can match an empty string.
2873
2874 9. A pattern with a very large number of alternatives (more than several
2875 hundred) was running out of internal workspace during the pre-compile
2876 phase, where pcre_compile() figures out how much memory will be needed. A
2877 bit of new cunning has reduced the workspace needed for groups with
2878 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2879 workspace instead of running out of the 4096 that are available.
2880
2881 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2882
2883 11. Applied patch from Google to remove an optimization that didn't quite work.
2884 The report of the bug said:
2885
2886 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2887 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2888 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2889
2890 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2891 it matched the wrong number of bytes.
2892
2893
2894 Version 7.1 24-Apr-07
2895 ---------------------
2896
2897 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2898 that is more "standard", making use of automake and other Autotools. There
2899 is some re-arrangement of the files and adjustment of comments consequent
2900 on this.
2901
2902 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2903 for recursive directory scanning broke on some systems because the files
2904 are not scanned in any specific order and on different systems the order
2905 was different. A call to "sort" has been inserted into RunGrepTest for the
2906 approprate test as a short-term fix. In the longer term there may be an
2907 alternative.
2908
2909 3. I had an email from Eric Raymond about problems translating some of PCRE's
2910 man pages to HTML (despite the fact that I distribute HTML pages, some
2911 people do their own conversions for various reasons). The problems
2912 concerned the use of low-level troff macros .br and .in. I have therefore
2913 removed all such uses from the man pages (some were redundant, some could
2914 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2915 HTML has been updated to handle .nf/.fi and to complain if it encounters
2916 .br or .in.
2917
2918 4. Updated comments in configure.ac that get placed in config.h.in and also
2919 arranged for config.h to be included in the distribution, with the name
2920 config.h.generic, for the benefit of those who have to compile without
2921 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2922
2923 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2924 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2925 makevp.bat for the current PCRE, using the additional files
2926 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2927
2928 6. A Windows user reported a minor discrepancy with test 2, which turned out
2929 to be caused by a trailing space on an input line that had got lost in his
2930 copy. The trailing space was an accident, so I've just removed it.
2931
2932 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2933 that is needed.
2934
2935 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2936 as "const" (a) because they are and (b) because it helps the PHP
2937 maintainers who have recently made a script to detect big data structures
2938 in the php code that should be moved to the .rodata section. I remembered
2939 to update Builducptable as well, so it won't revert if ucptable.h is ever
2940 re-created.
2941
2942 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2943 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2944 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2945 support is not required. This saves 1.5-2K of code, which is important in
2946 some applications.
2947
2948 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2949 so as not to refer to the tables, even though these functions will never be
2950 called when UTF-8 support is disabled. Otherwise there are problems with a
2951 shared library.
2952
2953 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2954
2955 (a) It was defining its arguments as char * instead of void *.
2956
2957 (b) It was assuming that all moves were upwards in memory; this was true
2958 a long time ago when I wrote it, but is no longer the case.
2959
2960 The emulated memove() is provided for those environments that have neither
2961 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2962 is clearly not the case, as these two bugs were recently reported.
2963
2964 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2965 and Detrail to create the HTML documentation, the .txt form of the man
2966 pages, and it removes trailing spaces from listed files. It also creates
2967 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2968 case, it wraps all the #defines with #ifndefs. This script should be run
2969 before "make dist".
2970
2971 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2972 with Unicode property support.
2973
2974 (a) For a maximizing quantifier, if the two different cases of the
2975 character were of different lengths in their UTF-8 codings (there are
2976 some cases like this - I found 11), and the matching function had to
2977 back up over a mixture of the two cases, it incorrectly assumed they
2978 were both the same length.
2979
2980 (b) When PCRE was configured to use the heap rather than the stack for
2981 recursion during matching, it was not correctly preserving the data for
2982 the other case of a UTF-8 character when checking ahead for a match
2983 while processing a minimizing repeat. If the check also involved
2984 matching a wide character, but failed, corruption could cause an
2985 erroneous result when trying to check for a repeat of the original
2986 character.
2987
2988 13. Some tidying changes to the testing mechanism:
2989
2990 (a) The RunTest script now detects the internal link size and whether there
2991 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2992 values substituted by "configure". (The RunGrepTest script already did
2993 this for UTF-8.) The configure.ac script no longer substitutes the
2994 relevant variables.
2995
2996 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2997 with length and offset values. This means that the output is different
2998 for different internal link sizes. Test 2 is skipped for link sizes
2999 other than 2 because of this, bypassing the problem. Unfortunately,
3000 there was also a test in test 3 (the locale tests) that used /B and
3001 failed for link sizes other than 2. Rather than cut the whole test out,
3002 I have added a new /Z option to pcretest that replaces the length and
3003 offset values with spaces. This is now used to make test 3 independent
3004 of link size. (Test 2 will be tidied up later.)
3005
3006 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3007 segmentation fault instead of returning the appropriate error message.
3008
3009 15. In multiline mode when the newline sequence was set to "any", the pattern
3010 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3011 This doesn't seem right; it now treats the CRLF combination as the line
3012 ending, and so does not match in that case. It's only a pattern such as ^$
3013 that would hit this one: something like ^ABC$ would have failed after \r
3014 and then tried again after \r\n.
3015
3016 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3017 in an attempt to make files that differ only in their line terminators
3018 compare equal. This works on Linux.
3019
3020 17. Under certain error circumstances pcregrep might try to free random memory
3021 as it exited. This is now fixed, thanks to valgrind.
3022
3023 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3024 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3025 was because its rules for how to advance for /g after matching an empty
3026 string at the end of a line did not allow for this case. They now check for
3027 it specially.
3028
3029 20. pcretest is supposed to handle patterns and data of any length, by
3030 extending its buffers when necessary. It was getting this wrong when the
3031 buffer for a data line had to be extended.
3032
3033 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3034 CRLF as a newline sequence.
3035
3036 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3037 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3038 I have nevertheless tidied it up.
3039
3040 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3041
3042 24. Added a man page for pcre-config.
3043
3044
3045 Version 7.0 19-Dec-06
3046 ---------------------
3047
3048 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3049 moving to gcc 4.1.1.
3050
3051 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3052 sys/time.h, which is documented as needed for this function. It doesn't
3053 seem to matter on Linux, but it showed up on some releases of OS X.
3054
3055 3. It seems that there are systems where bytes whose values are greater than
3056 127 match isprint() in the "C" locale. The "C" locale should be the
3057 default when a C program starts up. In most systems, only ASCII printing
3058 characters match isprint(). This difference caused the output from pcretest
3059 to vary, making some of the tests fail. I have changed pcretest so that:
3060
3061 (a) When it is outputting text in the compiled version of a pattern, bytes
3062 other than 32-126 are always shown as hex escapes.
3063
3064 (b) When it is outputting text that is a matched part of a subject string,
3065 it does the same, unless a different locale has been set for the match
3066 (using the /L modifier). In this case, it uses isprint() to decide.
3067
3068 4. Fixed a major bug that caused incorrect computation of the amount of memory
3069 required for a compiled pattern when options that changed within the
3070 pattern affected the logic of the preliminary scan that determines the
3071 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3072 that the computed length was too small. The symptoms of this bug were
3073 either the PCRE error "internal error: code overflow" from pcre_compile(),
3074 or a glibc crash with a message such as "pcretest: free(): invalid next
3075 size (fast)". Examples of patterns that provoked this bug (shown in
3076 pcretest format) are:
3077
3078 /(?-x: )/x
3079 /(?x)(?-x: \s*#\s*)/
3080 /((?i)[\x{c0}])/8
3081 /(?i:[\x{c0}])/8
3082
3083 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3084 is now done differently.
3085
3086 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3087 wrapper classes; (b) implement a new function in the C++ scanner that is
3088 more efficient than the old way of doing things because it avoids levels of
3089 recursion in the regex matching; (c) add a paragraph to the documentation
3090 for the FullMatch() function.
3091
3092 6. The escape sequence \n was being treated as whatever was defined as
3093 "newline". Not only was this contrary to the documentation, which states
3094 that \n is character 10 (hex 0A), but it also went horribly wrong when
3095 "newline" was defined as CRLF. This has been fixed.
3096
3097 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3098 was being set to -1 for the "end of line" case (supposedly a value that no
3099 character can have). Though this value is never used (the check for end of
3100 line is "zero bytes in current character"), it caused compiler complaints.
3101 I've changed it to 0xffffffff.
3102
3103 8. In pcre_version.c, the version string was being built by a sequence of
3104 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3105 string (as it is for production releases) called a macro with an empty
3106 argument. The C standard says the result of this is undefined. The gcc
3107 compiler treats it as an empty string (which was what was wanted) but it is
3108 reported that Visual C gives an error. The source has been hacked around to
3109 avoid this problem.
3110
3111 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3112 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3113 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3114 of them did).
3115
3116 10. Originally, pcretest opened its input and output without "b"; then I was
3117 told that "b" was needed in some environments, so it was added for release
3118 5.0 to both the input and output. (It makes no difference on Unix-like
3119 systems.) Later I was told that it is wrong for the input on Windows. I've
3120 now abstracted the modes into two macros, to make it easier to fiddle with
3121 them, and removed "b" from the input mode under Windows.
3122
3123 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3124
3125 12. Added -help and --help to pcretest as an official way of being reminded
3126 of the options.
3127
3128 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3129 and pcrecpp.cc because they annoy compilers at high warning levels.
3130
3131 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3132
3133 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3134 scripts are not C programs :-) and which was not noticed because it works
3135 on Linux.
3136
3137 16. pcretest is supposed to handle any length of pattern and data line (as one
3138 line or as a continued sequence of lines) by extending its input buffer if
3139 necessary. This feature was broken for very long pattern lines, leading to
3140 a string of junk being passed to pcre_compile() if the pattern was longer
3141 than about 50K.
3142
3143 17. I have done a major re-factoring of the way pcre_compile() computes the
3144 amount of memory needed for a compiled pattern. Previously, there was code
3145 that made a preliminary scan of the pattern in order to do this. That was
3146 OK when PCRE was new, but as the facilities have expanded, it has become
3147 harder and harder to keep it in step with the real compile phase, and there
3148 have been a number of bugs (see for example, 4 above). I have now found a
3149 cunning way of running the real compile function in a "fake" mode that
3150 enables it to compute how much memory it would need, while actually only
3151 ever using a few hundred bytes of working memory and without too many
3152 tests of the mode. This should make future maintenance and development
3153 easier. A side effect of this work is that the limit of 200 on the nesting
3154 depth of parentheses has been removed (though this was never a serious
3155 limitation, I suspect). However, there is a downside: pcre_compile() now
3156 runs more slowly than before (30% or more, depending on the pattern). I
3157 hope this isn't a big issue. There is no effect on runtime performance.
3158
3159 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3160 newline (only possible for the last line of a file) and it was a
3161 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3162
3163 19. Added additional timing features to pcretest. (1) The -tm option now times
3164 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3165 separate command line item, by a number that specifies the number of
3166 repeats to use when timing. The default is 50000; this gives better
3167 precision, but takes uncomfortably long for very large patterns.
3168
3169 20. Extended pcre_study() to be more clever in cases where a branch of a
3170 subpattern has no definite first character. For example, (a*|b*)[cd] would
3171 previously give no result from pcre_study(). Now it recognizes that the
3172 first character must be a, b, c, or d.
3173
3174 21. There was an incorrect error "recursive call could loop indefinitely" if
3175 a subpattern (or the entire pattern) that was being tested for matching an
3176 empty string contained only one non-empty item after a nested subpattern.
3177 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3178 incorrectly, because the \d was being skipped in the check.
3179
3180 22. The pcretest program now has a new pattern option /B and a command line
3181 option -b, which is equivalent to adding /B to every pattern. This causes
3182 it to show the compiled bytecode, without the additional information that
3183 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3184 is the same as /B/I).
3185
3186 23. A new optimization is now able automatically to treat some sequences such
3187 as a*b as a*+b. More specifically, if something simple (such as a character
3188 or a simple class like \d) has an unlimited quantifier, and is followed by
3189 something that cannot possibly match the quantified thing, the quantifier
3190 is automatically "possessified".
3191
3192 24. A recursive reference to a subpattern whose number was greater than 39
3193 went wrong under certain circumstances in UTF-8 mode. This bug could also
3194 have affected the operation of pcre_study().
3195
3196 25. Realized that a little bit of performance could be had by replacing
3197 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3198
3199 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3200
3201 27. Possessive quantifiers such as a++ were previously implemented by turning
3202 them into atomic groups such as ($>a+). Now they have their own opcodes,
3203 which improves performance. This includes the automatically created ones
3204 from 23 above.
3205
3206 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3207 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3208 the first matched character to be a colon. This applied both to named and
3209 numbered groups.
3210
3211 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3212
3213 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3214 building PCRE on Windows easier, so I have included it in the distribution.
3215
3216 31. There is now a check in pcretest against a ridiculously large number being
3217 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3218 loop, the loop is abandoned.
3219
3220 32. Forward references to subpatterns in conditions such as (?(2)...) where
3221 subpattern 2 is defined later cause pcre_compile() to search forwards in
3222 the pattern for the relevant set of parentheses. This search went wrong
3223 when there were unescaped parentheses in a character class, parentheses
3224 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3225
3226 33. "Subroutine" calls and backreferences were previously restricted to
3227 referencing subpatterns earlier in the regex. This restriction has now
3228 been removed.
3229
3230 34. Added a number of extra features that are going to be in Perl 5.10. On the
3231 whole, these are just syntactic alternatives for features that PCRE had
3232 previously implemented using the Python syntax or my own invention. The
3233 other formats are all retained for compatibility.
3234
3235 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3236 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3237 also .NET compatible.
3238
3239 (b) A recursion or subroutine call to a named group can now be defined as
3240 (?&name) as well as (?P>name).
3241
3242 (c) A backreference to a named group can now be defined as \k<name> or
3243 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3244 5.10, are also .NET compatible.
3245
3246 (d) A conditional reference to a named group can now use the syntax
3247 (?(<name>) or (?('name') as well as (?(name).
3248
3249 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3250 groups (named and numbered) that are never evaluated inline, but can be
3251 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3252 is always false. There may be only one alternative in such a group.
3253
3254 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3255 as the simple (?(R). The condition is true only if the most recent
3256 recursion is that of the given number or name. It does not search out
3257 through the entire recursion stack.
3258
3259 (g) The escape \gN or \g{N} has been added, where N is a positive or
3260 negative number, specifying an absolute or relative reference.
3261
3262 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3263 some "unreachable code" warnings.
3264
3265 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3266 things, this adds five new scripts.
3267
3268 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3269 There were also incompatibilities regarding the handling of \Q..\E inside
3270 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3271 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3272
3273 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3274 matches an empty string, and forcibly breaks the loop. There were bugs in
3275 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3276 against aaaa the result was just "a" rather than "aaaa", for example. Two
3277 separate and independent bugs (that affected different cases) have been
3278 fixed.
3279
3280 39. Refactored the code to abolish the use of different opcodes for small
3281 capturing bracket numbers. This is a tidy that I avoided doing when I
3282 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3283 The new approach is not only tidier, it makes it possible to reduce the
3284 memory needed to fix the previous bug (38).
3285
3286 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3287 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3288 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3289 mode.
3290
3291 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3292 report.
3293
3294 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3295 copy construction and assignment in the C++ wrapper.
3296
3297 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3298 couple of bugs that could have given wrong results in the "--newline=crlf"
3299 case.
3300
3301 44. Added a number of casts and did some reorganization of signed/unsigned int
3302 variables following suggestions from Dair Grant. Also renamed the variable
3303 "this" as "item" because it is a C++ keyword.
3304
3305 45. Arranged for dftables to add
3306
3307 #include "pcre_internal.h"
3308
3309 to pcre_chartables.c because without it, gcc 4.x may remove the array
3310 definition from the final binary if PCRE is built into a static library and
3311 dead code stripping is activated.
3312
3313 46. For an unanchored pattern, if a match attempt fails at the start of a
3314 newline sequence, and the newline setting is CRLF or ANY, and the next two
3315 characters are CRLF, advance by two characters instead of one.
3316
3317
3318 Version 6.7 04-Jul-06
3319 ---------------------
3320
3321 1. In order to handle tests when input lines are enormously long, pcretest has
3322 been re-factored so that it automatically extends its buffers when
3323 necessary. The code is crude, but this _is_ just a test program. The
3324 default size has been increased from 32K to 50K.
3325
3326 2. The code in pcre_study() was using the value of the re argument before
3327 testing it for NULL. (Of course, in any sensible call of the function, it
3328 won't be NULL.)
3329
3330 3. The memmove() emulation function in pcre_internal.h, which is used on
3331 systems that lack both memmove() and bcopy() - that is, hardly ever -
3332 was missing a "static" storage class specifier.
3333
3334 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3335 containing an extended class (one that cannot be represented by a bitmap
3336 because it contains high-valued characters or Unicode property items, e.g.
3337 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3338 pattern, but PCRE should not loop if you do not (it no longer does).
3339 [Detail: two cases were found: (a) a repeated subpattern containing an
3340 extended class; (b) a recursive reference to a subpattern that followed a
3341 previous extended class. It wasn't skipping over the extended class
3342 correctly when UTF-8 mode was not set.]
3343
3344 5. A negated single-character class was not being recognized as fixed-length
3345 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3346 compile error "lookbehind assertion is not fixed length".
3347
3348 6. The RunPerlTest auxiliary script was showing an unexpected difference
3349 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3350 write a Perl script that can interpret lines of an input file either as
3351 byte characters or as UTF-8, which is what "perltest" was being required to
3352 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3353 can't do is switch easily at run time between having the "use utf8;" pragma
3354 or not. In the end, I fudged it by using the RunPerlTest script to insert
3355 "use utf8;" explicitly for the UTF-8 tests.
3356
3357 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3358 the end of the subject string, contrary to the documentation and to what
3359 Perl does. This was true of both matching functions. Now it matches only at
3360 the start of the subject and immediately after *internal* newlines.
3361
3362 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3363 a pointer to an int instead of a pointer to an unsigned long int. This
3364 caused problems on 64-bit systems.
3365
3366 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3367 instance of the 'standard' template library not being so standard".
3368
3369 10. There was no check on the number of named subpatterns nor the maximum
3370 length of a subpattern name. The product of these values is used to compute
3371 the size of the memory block for a compiled pattern. By supplying a very
3372 long subpattern name and a large number of named subpatterns, the size
3373 computation could be caused to overflow. This is now prevented by limiting
3374 the length of names to 32 characters, and the number of named subpatterns
3375 to 10,000.
3376
3377 11. Subpatterns that are repeated with specific counts have to be replicated in
3378 the compiled pattern. The size of memory for this was computed from the
3379 length of the subpattern and the repeat count. The latter is limited to
3380 65535, but there was no limit on the former, meaning that integer overflow
3381 could in principle occur. The compiled length of a repeated subpattern is
3382 now limited to 30,000 bytes in order to prevent this.
3383
3384 12. Added the optional facility to have named substrings with the same name.
3385
3386 13. Added the ability to use a named substring as a condition, using the
3387 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3388 are numbers (not recommended). Forward references are permitted.
3389
3390 14. Added forward references in named backreferences (if you see what I mean).
3391
3392 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3393 pattern could run off the end of the subject. For example, the pattern
3394 "(?s)(.{1,5})"8 did this with the subject "ab".
3395
3396 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3397 PCRE_CASELESS was set when matching characters that were quantified with ?
3398 or *.
3399
3400 17. A character class other than a single negated character that had a minimum
3401 but no maximum quantifier - for example [ab]{6,} - was not handled
3402 correctly by pce_dfa_exec(). It would match only one character.
3403
3404 18. A valid (though odd) pattern that looked like a POSIX character
3405 class but used an invalid character after [ (for example [[,abc,]]) caused
3406 pcre_compile() to give the error "Failed: internal error: code overflow" or
3407 in some cases to crash with a glibc free() error. This could even happen if
3408 the pattern terminated after [[ but there just happened to be a sequence of
3409 letters, a binary zero, and a closing ] in the memory that followed.
3410
3411 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3412 over the years. Originally (before any Unicode support), just the bottom 8
3413 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3414 output from "man perlunicode" includes this:
3415
3416 The regular expression compiler produces polymorphic opcodes. That
3417 is, the pattern adapts to the data and automatically switches to
3418 the Unicode character scheme when presented with Unicode data--or
3419 instead uses a traditional byte scheme when presented with byte
3420 data.
3421
3422 Sadly, a wide octal escape does not cause a switch, and in a string with
3423 no other multibyte characters, these octal escapes are treated as before.
3424 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3425 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3426 Unicode string.
3427
3428 I have not perpetrated such confusion in PCRE. Up till now, it took just
3429 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3430 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3431 translate to the appropriate multibyte character.
3432
3433 29. Applied some refactoring to reduce the number of warnings from Microsoft
3434 and Borland compilers. This has included removing the fudge introduced
3435 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3436 a warning about an unused variable.
3437
3438 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3439 characters since release 4.0, because Perl (from release 5.004) does not.
3440 [Or at least, is documented not to: some releases seem to be in conflict
3441 with the documentation.] However, when a pattern was studied with
3442 pcre_study() and all its branches started with \s, PCRE still included VT
3443 as a possible starting character. Of course, this did no harm; it just
3444 caused an unnecessary match attempt.
3445
3446 22. Removed a now-redundant internal flag bit that recorded the fact that case
3447 dependency changed within the pattern. This was once needed for "required
3448 byte" processing, but is no longer used. This recovers a now-scarce options
3449 bit. Also moved the least significant internal flag bit to the most-
3450 significant bit of the word, which was not previously used (hangover from
3451 the days when it was an int rather than a uint) to free up another bit for
3452 the future.
3453
3454 23. Added support for CRLF line endings as well as CR and LF. As well as the
3455 default being selectable at build time, it can now be changed at runtime
3456 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3457 specify that it is scanning data with non-default line endings.
3458
3459 24. Changed the definition of CXXLINK to make it agree with the definition of
3460 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3461
3462 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3463 recursions. This makes a big different to stack usage for some patterns.
3464
3465 26. If a subpattern containing a named recursion or subroutine reference such
3466 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3467 the space required for the compiled pattern went wrong and gave too small a
3468 value. Depending on the environment, this could lead to "Failed: internal
3469 error: code overflow at offset 49" or "glibc detected double free or
3470 corruption" errors.
3471
3472 27. Applied patches from Google (a) to support the new newline modes and (b) to
3473 advance over multibyte UTF-8 characters in GlobalReplace.
3474
3475 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3476 difference for some implementation of PCRE in some Windows version.
3477
3478 29. Added some extra testing facilities to pcretest:
3479
3480 \q<number> in a data line sets the "match limit" value
3481 \Q<number> in a data line sets the "match recursion limt" value
3482 -S <number> sets the stack size, where <number> is in megabytes
3483
3484 The -S option isn't available for Windows.
3485
3486
3487 Version 6.6 06-Feb-06
3488 ---------------------
3489
3490 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3491 in pcreposix.h. I have copied the definition from pcre.h.
3492
3493 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3494 because pcre.h is no longer a built file.
3495
3496 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3497 not normally included in the compiled code.
3498
3499
3500 Version 6.5 01-Feb-06
3501 ---------------------
3502
3503 1. When using the partial match feature with pcre_dfa_exec(), it was not
3504 anchoring the second and subsequent partial matches at the new starting
3505 point. This could lead to incorrect results. For example, with the pattern
3506 /1234/, partially matching against "123" and then "a4" gave a match.
3507
3508 2. Changes to pcregrep:
3509
3510 (a) All non-match returns from pcre_exec() were being treated as failures
3511 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3512 error message is output. Some extra information is given for the
3513 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3514 probably the only errors that are likely to be caused by users (by
3515 specifying a regex that has nested indefinite repeats, for instance).
3516 If there are more than 20 of these errors, pcregrep is abandoned.
3517
3518 (b) A binary zero was treated as data while matching, but terminated the
3519 output line if it was written out. This has been fixed: binary zeroes
3520 are now no different to any other data bytes.
3521
3522 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3523 used to set a locale for matching. The --locale=xxxx long option has
3524 been added (no short equivalent) to specify a locale explicitly on the
3525 pcregrep command, overriding the environment variables.
3526
3527 (d) When -B was used with -n, some line numbers in the output were one less
3528 than they should have been.
3529
3530 (e) Added the -o (--only-matching) option.
3531
3532 (f) If -A or -C was used with -c (count only), some lines of context were
3533 accidentally printed for the final match.
3534
3535 (g) Added the -H (--with-filename) option.
3536
3537 (h) The combination of options -rh failed to suppress file names for files
3538 that were found from directory arguments.
3539
3540 (i) Added the -D (--devices) and -d (--directories) options.
3541
3542 (j) Added the -F (--fixed-strings) option.
3543
3544 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3545
3546 (l) Added the --colo(u)r option.
3547
3548 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3549 is not present by default.
3550
3551 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3552 items such as (?R) or (?1), when the recursion could match a number of
3553 alternatives. If it matched one of the alternatives, but subsequently,
3554 outside the recursion, there was a failure, the code tried to back up into
3555 the recursion. However, because of the way PCRE is implemented, this is not
3556 possible, and the result was an incorrect result from the match.
3557
3558 In order to prevent this happening, the specification of recursion has
3559 been changed so that all such subpatterns are automatically treated as
3560 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3561
3562 4. I had overlooked the fact that, in some locales, there are characters for
3563 which isalpha() is true but neither isupper() nor islower() are true. In
3564 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3565 and ordfeminine) are like this. This affected the treatment of \w and \W
3566 when they appeared in character classes, but not when they appeared outside
3567 a character class. The bit map for "word" characters is now created
3568 separately from the results of isalnum() instead of just taking it from the
3569 upper, lower, and digit maps. (Plus the underscore character, of course.)
3570
3571 5. The above bug also affected the handling of POSIX character classes such as
3572 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3573 permanent tables. Instead, the bit maps for such a class were previously
3574 created as the appropriate unions of the upper, lower, and digit bitmaps.
3575 Now they are created by subtraction from the [[:word:]] class, which has
3576 its own bitmap.
3577
3578 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3579 It is created by subtracting the vertical space characters (\x09, \x0a,
3580 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3581 subtraction was done in the overall bitmap for a character class, meaning
3582 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3583 be recognized. This bug has been fixed.
3584
3585 7. Patches from the folks at Google:
3586
3587 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3588 real life, but is still worth protecting against".
3589
3590 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3591 regular expressions".
3592
3593 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3594 have it.
3595
3596 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3597 "configure" and the latter not, in order to fix a problem somebody had
3598 with compiling the Arg class on HP-UX.
3599
3600 (e) Improve the error-handling of the C++ wrapper a little bit.
3601
3602 (f) New tests for checking recursion limiting.
3603
3604 8. The pcre_memmove() function, which is used only if the environment does not
3605 have a standard memmove() function (and is therefore rarely compiled),
3606 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3607 returning a result (though PCRE never actually uses the result).
3608
3609 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3610 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3611 returned instead of calling malloc() with an overflowing number that would
3612 most likely cause subsequent chaos.
3613
3614 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3615
3616 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3617 with this option is matched, the nmatch and pmatch options of regexec() are
3618 ignored.
3619
3620 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3621 provided in case anyone wants to the the POSIX interface with UTF-8
3622 strings.
3623
3624 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3625 C++ linking (needed for some HP-UX environments).
3626
3627 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3628 (unused parameter) and in the pcre_printint() function (omitted "default"
3629 switch label when the default is to do nothing).
3630
3631 15. Added some code to make it possible, when PCRE is compiled as a C++
3632 library, to replace subject pointers for pcre_exec() with a smart pointer
3633 class, thus making it possible to process discontinuous strings.
3634
3635 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3636 much the same function. They were added by different people who were trying
3637 to make PCRE easy to compile on non-Unix systems. It has been suggested
3638 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3639 for compiling on Windows systems. I have therefore replaced it with
3640 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3641 defaults to "extern" for C or "extern C" for C++, which works fine on
3642 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3643 SCOPE with something explicit in config.h. In addition:
3644
3645 (a) pcreposix.h still had just "extern" instead of either of these macros;
3646 I have replaced it with PCRE_DATA_SCOPE.
3647
3648 (b) Functions such as _pcre_xclass(), which are internal to the library,
3649 but external in the C sense, all had PCRE_EXPORT in their definitions.
3650 This is apparently wrong for the Windows case, so I have removed it.
3651 (It makes no difference on Unix-like systems.)
3652
3653 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3654 of recursive calls to match(). This is different to MATCH_LIMIT because
3655 that limits the total number of calls to match(), not all of which increase
3656 the depth of recursion. Limiting the recursion depth limits the amount of
3657 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3658 when PCRE is compiled, and changed at run time. A patch from Google adds
3659 this functionality to the C++ interface.
3660
3661 18. Changes to the handling of Unicode character properties:
3662
3663 (a) Updated the table to Unicode 4.1.0.
3664
3665 (b) Recognize characters that are not in the table as "Cn" (undefined).
3666
3667 (c) I revised the way the table is implemented to a much improved format
3668 which includes recognition of ranges. It now supports the ranges that
3669 are defined in UnicodeData.txt, and it also amalgamates other
3670 characters into ranges. This has reduced the number of entries in the
3671 table from around 16,000 to around 3,000, thus reducing its size
3672 considerably. I realized I did not need to use a tree structure after
3673 all - a binary chop search is just as efficient. Having reduced the
3674 number of entries, I extended their size from 6 bytes to 8 bytes to
3675 allow for more data.
3676
3677 (d) Added support for Unicode script names via properties such as \p{Han}.
3678
3679 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3680 matching that character.
3681
3682 20. When matching a repeated Unicode property with a minimum greater than zero,
3683 (for example \pL{2,}), PCRE could look past the end of the subject if it
3684 reached it while seeking the minimum number of characters. This could
3685 happen only if some of the characters were more than one byte long, because
3686 there is a check for at least the minimum number of bytes.
3687
3688 21. Refactored the implementation of \p and \P so as to be more general, to
3689 allow for more different types of property in future. This has changed the
3690 compiled form incompatibly. Anybody with saved compiled patterns that use
3691 \p or \P will have to recompile them.
3692
3693 22. Added "Any" and "L&" to the supported property types.
3694
3695 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3696 but give a compile time error if the value is greater than 0xff.
3697
3698 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3699 accidentally not being installed or uninstalled.
3700
3701 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3702 made were to insert the current release number. This seemed silly, because
3703 it made things harder for people building PCRE on systems that don't run
3704 "configure". I have turned pcre.h into a distributed file, no longer built
3705 by "configure", with the version identification directly included. There is
3706 no longer a pcre.h.in file.
3707
3708 However, this change necessitated a change to the pcre-config script as
3709 well. It is built from pcre-config.in, and one of the substitutions was the
3710 release number. I have updated configure.ac so that ./configure now finds
3711 the release number by grepping pcre.h.
3712
3713 26. Added the ability to run the tests under valgrind.
3714
3715
3716 Version 6.4 05-Sep-05
3717 ---------------------
3718
3719 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3720 "--" to be printed when multiple files were scanned, even when none of the
3721 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3722 consider it to be a bug, and have restored the previous behaviour.
3723
3724 2. A couple of code tidies to get rid of compiler warnings.
3725
3726 3. The pcretest program used to cheat by referring to symbols in the library
3727 whose names begin with _pcre_. These are internal symbols that are not
3728 really supposed to be visible externally, and in some environments it is
3729 possible to suppress them. The cheating is now confined to including
3730 certain files from the library's source, which is a bit cleaner.
3731
3732 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3733 file's purpose clearer.
3734
3735 5. Reorganized pcre_ucp_findchar().
3736
3737
3738 Version 6.3 15-Aug-05
3739 ---------------------
3740
3741 1. The file libpcre.pc.in did not have general read permission in the tarball.
3742
3743 2. There were some problems when building without C++ support:
3744
3745 (a) If C++ support was not built, "make install" and "make test" still
3746 tried to test it.
3747
3748 (b) There were problems when the value of CXX was explicitly set. Some
3749 changes have been made to try to fix these, and ...
3750
3751 (c) --disable-cpp can now be used to explicitly disable C++ support.
3752
3753 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3754 backslash in a target when C++ was disabled. This confuses some
3755 versions of "make", apparently. Using an intermediate variable solves
3756 this. (Same for CPP_LOBJ.)
3757
3758 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3759 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3760 necessary on certain architectures.
3761
3762 4. Added a setting of -export-symbols-regex to the link command to remove
3763 those symbols that are exported in the C sense, but actually are local
3764 within the library, and not documented. Their names all begin with
3765 "_pcre_". This is not a perfect job, because (a) we have to except some
3766 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3767 available (and never for static libraries). I have made a note to try to
3768 find a way round (a) in the future.
3769
3770
3771 Version 6.2 01-Aug-05
3772 ---------------------
3773
3774 1. There was no test for integer overflow of quantifier values. A construction
3775 such as {1111111111111111} would give undefined results. What is worse, if
3776 a minimum quantifier for a parenthesized subpattern overflowed and became
3777 negative, the calculation of the memory size went wrong. This could have
3778 led to memory overwriting.
3779
3780 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3781
3782 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3783 operating environments where this matters.
3784
3785 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3786 PCRE options from within the C++ wrapper.
3787
3788 5. Named capturing subpatterns were not being correctly counted when a pattern
3789 was compiled. This caused two problems: (a) If there were more than 100
3790 such subpatterns, the calculation of the memory needed for the whole
3791 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3792 back references of the form \12, where the number was greater than 9, were
3793 not recognized as back references, even though there were sufficient
3794 previous subpatterns.
3795
3796 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3797 versions of gcc, e.g. 2.95.4.
3798
3799
3800 Version 6.1 21-Jun-05
3801 ---------------------
3802
3803 1. There was one reference to the variable "posix" in pcretest.c that was not
3804 surrounded by "#if !defined NOPOSIX".
3805
3806 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3807 the cross-check on the old pcre_info() function, for the benefit of the
3808 cut-down version of PCRE that is currently imported into Exim.
3809
3810 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3811 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3812 patterns that start (?i) but I don't think that matters. The use of (?i) is
3813 just an example; this all applies to the other options as well.
3814
3815 4. Since libtool seems to echo the compile commands it is issuing, the output
3816 from "make" can be reduced a bit by putting "@" in front of each libtool
3817 compile command.
3818
3819 5. Patch from the folks at Google for configure.in to be a bit more thorough
3820 in checking for a suitable C++ installation before trying to compile the
3821 C++ stuff. This should fix a reported problem when a compiler was present,
3822 but no suitable headers.
3823
3824 6. The man pages all had just "PCRE" as their title. I have changed them to
3825 be the relevant file name. I have also arranged that these names are
3826 retained in the file doc/pcre.txt, which is a concatenation in text format
3827 of all the man pages except the little individual ones for each function.
3828
3829 7. The NON-UNIX-USE file had not been updated for the different set of source
3830 files that come with release 6. I also added a few comments about the C++
3831 wrapper.
3832
3833
3834 Version 6.0 07-Jun-05
3835 ---------------------
3836
3837 1. Some minor internal re-organization to help with my DFA experiments.
3838
3839 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3840 didn't matter for the library itself when fully configured, but did matter
3841 when compiling without UCP support, or within Exim, where the ucp files are
3842 not imported.
3843
3844 3. Refactoring of the library code to split up the various functions into
3845 different source modules. The addition of the new DFA matching code (see
3846 below) to a single monolithic source would have made it really too
3847 unwieldy, quite apart from causing all the code to be include in a
3848 statically linked application, when only some functions are used. This is
3849 relevant even without the DFA addition now that patterns can be compiled in
3850 one application and matched in another.
3851
3852 The downside of splitting up is that there have to be some external
3853 functions and data tables that are used internally in different modules of
3854 the library but which are not part of the API. These have all had their
3855 names changed to start with "_pcre_" so that they are unlikely to clash
3856 with other external names.
3857
3858 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3859 a different (DFA) algorithm. Although it is slower than the original
3860 function, it does have some advantages for certain types of matching
3861 problem.
3862
3863 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3864 including restarting after a partial match.
3865
3866 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3867 defined when compiling for Windows was sent to me. I have put it into the
3868 code, though I have no means of testing or verifying it.
3869
3870 7. Added the pcre_refcount() auxiliary function.
3871
3872 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3873 match before or at the first newline in the subject string. In pcretest,
3874 the /f option on a pattern can be used to set this.
3875
3876 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3877 would behave wrongly. This has been present in PCRE since release 4.0.
3878
3879 10. A number of changes to the pcregrep command:
3880
3881 (a) Refactored how -x works; insert ^(...)$ instead of setting
3882 PCRE_ANCHORED and checking the length, in preparation for adding
3883 something similar for -w.
3884
3885 (b) Added the -w (match as a word) option.
3886
3887 (c) Refactored the way lines are read and buffered so as to have more
3888 than one at a time available.
3889
3890 (d) Implemented a pcregrep test script.
3891
3892 (e) Added the -M (multiline match) option. This allows patterns to match
3893 over several lines of the subject. The buffering ensures that at least
3894 8K, or the rest of the document (whichever is the shorter) is available
3895 for matching (and similarly the previous 8K for lookbehind assertions).
3896
3897 (f) Changed the --help output so that it now says
3898
3899 -w, --word-regex(p)
3900
3901 instead of two lines, one with "regex" and the other with "regexp"
3902 because that confused at least one person since the short forms are the
3903 same. (This required a bit of code, as the output is generated
3904 automatically from a table. It wasn't just a text change.)
3905
3906 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3907 option but starts with a hyphen. Could be a pattern or a path name
3908 starting with a hyphen, for instance.
3909
3910 (h) "-" can be given as a file name to represent stdin.
3911
3912 (i) When file names are being printed, "(standard input)" is used for
3913 the standard input, for compatibility with GNU grep. Previously
3914 "<stdin>" was used.
3915
3916 (j) The option --label=xxx can be used to supply a name to be used for
3917 stdin when file names are being printed. There is no short form.
3918
3919 (k) Re-factored the options decoding logic because we are going to add
3920 two more options that take data. Such options can now be given in four
3921 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3922
3923 (l) Added the -A, -B, and -C options for requesting that lines of context
3924 around matches be printed.
3925
3926 (m) Added the -L option to print the names of files that do not contain
3927 any matching lines, that is, the complement of -l.
3928
3929 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3930 continue to scan other files.
3931
3932 (o) The -s option was incorrectly implemented. For compatibility with other
3933 greps, it now suppresses the error message for a non-existent or non-
3934 accessible file (but not the return code). There is a new option called
3935 -q that suppresses the output of matching lines, which was what -s was
3936 previously doing.
3937
3938 (p) Added --include and --exclude options to specify files for inclusion
3939 and exclusion when recursing.
3940
3941 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3942 Hopefully, it now does.
3943
3944 12. Missing cast in pcre_study().
3945
3946 13. Added an "uninstall" target to the makefile.
3947
3948 14. Replaced "extern" in the function prototypes in Makefile.in with
3949 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3950 world, but is set differently for Windows.
3951
3952 15. Added a second compiling function called pcre_compile2(). The only
3953 difference is that it has an extra argument, which is a pointer to an
3954 integer error code. When there is a compile-time failure, this is set
3955 non-zero, in addition to the error test pointer being set to point to an
3956 error message. The new argument may be NULL if no error number is required
3957 (but then you may as well call pcre_compile(), which is now just a
3958 wrapper). This facility is provided because some applications need a
3959 numeric error indication, but it has also enabled me to tidy up the way
3960 compile-time errors are handled in the POSIX wrapper.
3961
3962 16. Added VPATH=.libs to the makefile; this should help when building with one
3963 prefix path and installing with another. (Or so I'm told by someone who
3964 knows more about this stuff than I do.)
3965
3966 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3967 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3968 match everything, including newlines. This is not POSIX-compatible, but
3969 somebody wanted the feature. From pcretest it can be activated by using
3970 both the P and the s flags.
3971
3972 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3973
3974 19. libpcre.pc was being incorrectly installed as executable.
3975
3976 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3977 it now also looks for '\r' so that it will work unmodified on Windows.
3978
3979 21. Added Google's contributed C++ wrapper to the distribution.
3980
3981 22. Added some untidy missing memory free() calls in pcretest, to keep
3982 Electric Fence happy when testing.
3983
3984
3985
3986 Version 5.0 13-Sep-04
3987 ---------------------
3988
3989 1. Internal change: literal characters are no longer packed up into items
3990 containing multiple characters in a single byte-string. Each character
3991 is now matched using a separate opcode. However, there may be more than one
3992 byte in the character in UTF-8 mode.
3993
3994 2. The pcre_callout_block structure has two new fields: pattern_position and
3995 next_item_length. These contain the offset in the pattern to the next match
3996 item, and its length, respectively.
3997
3998 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3999 insertion of callouts before each pattern item. Added the /C option to
4000 pcretest to make use of this.
4001
4002 4. On the advice of a Windows user, the lines
4003
4004 #if defined(_WIN32) || defined(WIN32)
4005 _setmode( _fileno( stdout ), 0x8000 );
4006 #endif /* defined(_WIN32) || defined(WIN32) */
4007
4008 have been added to the source of pcretest. This apparently does useful
4009 magic in relation to line terminators.
4010
4011 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4012 for the benefit of those environments where the "b" makes a difference.
4013
4014 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4015 to know about it. I have put a hack into configure.in that adds in code
4016 to set GCC=yes if CC=icc. This seems to end up at a point in the
4017 generated configure script that is early enough to affect the setting of
4018 compiler options, which is what is needed, but I have no means of testing
4019 whether it really works. (The user who reported this had patched the
4020 generated configure script, which of course I cannot do.)
4021
4022 LATER: After change 22 below (new libtool files), the configure script
4023 seems to know about icc (and also ecc). Therefore, I have commented out
4024 this hack in configure.in.
4025
4026 7. Added support for pkg-config (2 patches were sent in).
4027
4028 8. Negated POSIX character classes that used a combination of internal tables
4029 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4030 [[:^ascii]]. Typically, they would match almost any characters. The other
4031 POSIX classes were not broken in this way.
4032
4033 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4034 to find the match, as PCRE was deluded into thinking that the match had to
4035 start at the start point or following a newline. The same bug applied to
4036 patterns with negative forward assertions or any backward assertions
4037 preceding ".*" at the start, unless the pattern required a fixed first
4038 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4039
4040 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4041 starting at the last subject character, bytes beyond the end of the subject
4042 string were read.
4043
4044 11. Renamed the variable "class" as "classbits" to make life easier for C++
4045 users. (Previously there was a macro definition, but it apparently wasn't
4046 enough.)
4047
4048 12. Added the new field "tables" to the extra data so that tables can be passed
4049 in at exec time, or the internal tables can be re-selected. This allows
4050 a compiled regex to be saved and re-used at a later time by a different
4051 program that might have everything at different addresses.
4052
4053 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4054 -R library as well as a -L library.
4055
4056 14. The debugging options of pcretest (-d on the command line or D on a
4057 pattern) showed incorrect output for anything following an extended class
4058 that contained multibyte characters and which was followed by a quantifier.
4059
4060 15. Added optional support for general category Unicode character properties
4061 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4062 support. It adds about 90K to the size of the library. The meanings of the
4063 inbuilt class escapes such as \d and \s have NOT been changed.
4064
4065 16. Updated pcredemo.c to include calls to free() to release the memory for the
4066 compiled pattern.
4067
4068 17. The generated file chartables.c was being created in the source directory
4069 instead of in the building directory. This caused the build to fail if the
4070 source directory was different from the building directory, and was
4071 read-only.
4072
4073 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4074 file. No doubt somebody will tell me if they don't make sense... Also added
4075 Dan Mooney's comments about building on OpenVMS.
4076
4077 19. Added support for partial matching via the PCRE_PARTIAL option for
4078 pcre_exec() and the \P data escape in pcretest.
4079
4080 20. Extended pcretest with 3 new pattern features:
4081
4082 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4083 write the compiled pattern to the file whose name is "rest-of-line".
4084 This is a straight binary dump of the data, with the saved pointer to
4085 the character tables forced to be NULL. The study data, if any, is
4086 written too. After writing, pcretest reads a new pattern.
4087
4088 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4089 compiled pattern from the given file. There must not be any
4090 occurrences of "<" in the file name (pretty unlikely); if there are,
4091 pcretest will instead treat the initial "<" as a pattern delimiter.
4092 After reading in the pattern, pcretest goes on to read data lines as
4093 usual.
4094
4095 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4096 and 16-bit fields in a compiled pattern, to simulate a pattern that
4097 was compiled on a host of opposite endianness.
4098
4099 21. The pcre-exec() function can now cope with patterns that were compiled on
4100 hosts of opposite endianness, with this restriction:
4101
4102 As for any compiled expression that is saved and used later, the tables
4103 pointer field cannot be preserved; the extra_data field in the arguments
4104 to pcre_exec() should be used to pass in a tables address if a value
4105 other than the default internal tables were used at compile time.
4106
4107 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4108 now diagnosed as an error. Previously, most of the time, a negative number
4109 would have been treated as zero, but if in addition "ovector" was passed as
4110 NULL, a crash could occur.
4111
4112 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4113 new versions from the libtool 1.5 distribution (the last one is a copy of
4114 a file called libtool.m4). This seems to have fixed the need to patch
4115 "configure" to support Darwin 1.3 (which I used to do). However, I still
4116 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4117 workstation).
4118
4119 24. Changed the PCRE licence to be the more standard "BSD" licence.
4120
4121
4122 Version 4.5 01-Dec-03
4123 ---------------------
4124
4125 1. There has been some re-arrangement of the code for the match() function so
4126 that it can be compiled in a version that does not call itself recursively.
4127 Instead, it keeps those local variables that need separate instances for
4128 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4129 needs to "recurse". Keeping track of where control must go is done by means
4130 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4131 hide most of the details from the main code, and operates only if
4132 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4133 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4134 operating.
4135
4136 To make it easier for callers to provide specially tailored get/free
4137 functions for this usage, two new functions, pcre_stack_malloc, and
4138 pcre_stack_free, are used. They are always called in strict stacking order,
4139 and the size of block requested is always the same.
4140
4141 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4142 PCRE has been compiled to use the stack or the heap for recursion. The
4143 -C option of pcretest uses this to show which version is compiled.
4144
4145 A new data escape \S, is added to pcretest; it causes the amounts of store
4146 obtained and freed by both kinds of malloc/free at match time to be added
4147 to the output.
4148
4149 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4150 what's available on my current Linux desktop machine.
4151
4152 3. When matching a UTF-8 string, the test for a valid string at the start has
4153 been extended. If start_offset is not zero, PCRE now checks that it points
4154 to a byte that is the start of a UTF-8 character. If not, it returns
4155 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4156 this is necessary because there may be backward assertions in the pattern.
4157 When matching the same subject several times, it may save resources to use
4158 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4159
4160 4. The code for checking the validity of UTF-8 strings has been tightened so
4161 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4162 containing "overlong sequences".
4163
4164 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4165 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4166 should have been "&", but it just so happened that all the cases this let
4167 through by mistake were picked up later in the function.
4168
4169 6. I had used a variable called "isblank" - this is a C99 function, causing
4170 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4171
4172 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4173 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4174 so the version is shown; (c) stop "make test" echoing "./RunTest".
4175
4176 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4177
4178 9. The prototype for memmove() for systems that don't have it was using
4179 size_t, but the inclusion of the header that defines size_t was later. I've
4180 moved the #includes for the C headers earlier to avoid this.
4181
4182 10. Added some adjustments to the code to make it easier to compiler on certain
4183 special systems:
4184
4185 (a) Some "const" qualifiers were missing.
4186 (b) Added the macro EXPORT before all exported functions; by default this
4187 is defined to be empty.
4188 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4189 that it reads its output file name as an argument instead of writing
4190 to the standard output and assuming this can be redirected.
4191
4192 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4193 class containing characters with values greater than 255, PCRE compilation
4194 went into a loop.
4195
4196 12. A recursive reference to a subpattern that was within another subpattern
4197 that had a minimum quantifier of zero caused PCRE to crash. For example,
4198 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4199 recursion. If the recursively-called subpattern itself had a zero repeat,
4200 that was OK.
4201
4202 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4203 buffer into which it was copied (for escape processing) was still set at
4204 1024, so long lines caused crashes.
4205
4206 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4207 "internal error: code overflow...". This applied to any character class
4208 that was followed by a possessive quantifier.
4209
4210 15. Modified the Makefile to add libpcre.la as a prerequisite for
4211 libpcreposix.la because I was told this is needed for a parallel build to
4212 work.
4213
4214 16. If a pattern that contained .* following optional items at the start was
4215 studied, the wrong optimizing data was generated, leading to matching
4216 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4217 matching string must start with a or b or c. The correct conclusion for
4218 this pattern is that a match can start with any character.
4219
4220
4221 Version 4.4 13-Aug-03
4222 ---------------------
4223
4224 1. In UTF-8 mode, a character class containing characters with values between
4225 127 and 255 was not handled correctly if the compiled pattern was studied.
4226 In fixing this, I have also improved the studying algorithm for such
4227 classes (slightly).
4228
4229 2. Three internal functions had redundant arguments passed to them. Removal
4230 might give a very teeny performance improvement.
4231
4232 3. Documentation bug: the value of the capture_top field in a callout is *one
4233 more than* the number of the hightest numbered captured substring.
4234
4235 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4236 in incorrectly linking with a previously installed version. They now link
4237 explicitly with libpcre.la.
4238
4239 5. configure.in no longer needs to recognize Cygwin specially.
4240
4241 6. A problem in pcre.in for Windows platforms is fixed.
4242
4243 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4244 pcretest, it used to include the size of the study block as part of its
4245 output. Unfortunately, the structure contains a field that has a different
4246 size on different hardware architectures. This meant that the tests that
4247 showed this size failed. As the block is currently always of a fixed size,
4248 this information isn't actually particularly useful in pcretest output, so
4249 I have just removed it.
4250
4251 8. Three pre-processor statements accidentally did not start in column 1.
4252 Sadly, there are *still* compilers around that complain, even though
4253 standard C has not required this for well over a decade. Sigh.
4254
4255 9. In pcretest, the code for checking callouts passed small integers in the
4256 callout_data field, which is a void * field. However, some picky compilers
4257 complained about the casts involved for this on 64-bit systems. Now
4258 pcretest passes the address of the small integer instead, which should get
4259 rid of the warnings.
4260
4261 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4262 both compile and run time, and gives an error if an invalid UTF-8 sequence
4263 is found. There is a option for disabling this check in cases where the
4264 string is known to be correct and/or the maximum performance is wanted.
4265
4266 11. In response to a bug report, I changed one line in Makefile.in from
4267
4268 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4269 to
4270 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4271
4272 to look similar to other lines, but I have no way of telling whether this
4273 is the right thing to do, as I do not use Windows. No doubt I'll get told
4274 if it's wrong...
4275
4276
4277 Version 4.3 21-May-03
4278 ---------------------
4279
4280 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4281 Makefile.
4282
4283 2. Some refactoring to improve the quality of the code:
4284
4285 (i) The utf8_table... variables are now declared "const".
4286
4287 (ii) The code for \cx, which used the "case flipping" table to upper case
4288 lower case letters, now just substracts 32. This is ASCII-specific,
4289 but the whole concept of \cx is ASCII-specific, so it seems
4290 reasonable.
4291
4292 (iii) PCRE was using its character types table to recognize decimal and
4293 hexadecimal digits in the pattern. This is silly, because it handles
4294 only 0-9, a-f, and A-F, but the character types table is locale-
4295 specific, which means strange things might happen. A private
4296 table is now used for this - though it costs 256 bytes, a table is
4297 much faster than multiple explicit tests. Of course, the standard
4298 character types table is still used for matching digits in subject
4299 strings against \d.
4300
4301 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4302 ending in _t are). So I've renamed it as ESC_tee.
4303
4304 3. The first argument for regexec() in the POSIX wrapper should have been
4305 defined as "const".
4306
4307 4. Changed pcretest to use malloc() for its buffers so that they can be
4308 Electric Fenced for debugging.
4309
4310 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4311 to read one or more bytes before the start of the subject string. Often this
4312 had no effect on PCRE's behaviour, but in some circumstances it could
4313 provoke a segmentation fault.
4314
4315 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4316 to try to read one or more bytes before the start of the subject string.
4317
4318 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4319 UTF-8 support could misbehave in various ways if the subject string
4320 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4321 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4322 back over UTF-8 characters.)
4323
4324
4325 Version 4.2 14-Apr-03
4326 ---------------------
4327
4328 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4329
4330 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4331 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4332 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4333 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4334 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4335 and BUILD_EXEEXT
4336 Note: automatic setting of the BUILD variables is not yet working
4337 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4338 compile-time but not at link-time
4339 [LINK]: use for linking executables only
4340 make different versions for Windows and non-Windows
4341 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4342 libraries
4343 [LINK_FOR_BUILD]: new variable
4344 [OBJEXT]: use throughout
4345 [EXEEXT]: use throughout
4346 <winshared>: new target
4347 <wininstall>: new target
4348 <dftables.o>: use native compiler
4349 <dftables>: use native linker
4350 <install>: handle Windows platform correctly
4351 <clean>: ditto
4352 <check>: ditto
4353 copy DLL to top builddir before testing
4354
4355 As part of these changes, -no-undefined was removed again. This was reported
4356 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4357 in any case.
4358
4359 3. Some tidies to get rid of compiler warnings:
4360
4361 . In the match_data structure, match_limit was an unsigned long int, whereas
4362 match_call_count was an int. I've made them both unsigned long ints.
4363
4364 . In pcretest the fact that a const uschar * doesn't automatically cast to
4365 a void * provoked a warning.
4366
4367 . Turning on some more compiler warnings threw up some "shadow" variables
4368 and a few more missing casts.
4369
4370 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4371 option, a class that contained a single character with a value between 128
4372 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4373
4374 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4375 option, a class that contained several characters, but with at least one
4376 whose value was between 128 and 255 caused PCRE to crash.
4377
4378
4379 Version 4.1 12-Mar-03
4380 ---------------------
4381
4382 1. Compiling with gcc -pedantic found a couple of places where casts were
4383 needed, and a string in dftables.c that was longer than standard compilers are
4384 required to support.
4385
4386 2. Compiling with Sun's compiler found a few more places where the code could
4387 be tidied up in order to avoid warnings.
4388
4389 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4390 first of these names is deprecated in the latest Autoconf in favour of the name
4391 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4392 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4393 analogy I have changed it to CFLAGS_FOR_BUILD.
4394
4395 4. Added -no-undefined to the linking command in the Makefile, because this is
4396 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4397 linking step for the pcreposix library.
4398
4399 5. PCRE was failing to diagnose the case of two named groups with the same
4400 name.
4401
4402 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4403 literal character that is needed in the subject for a match, and scans along to
4404 ensure that it is present before embarking on the full matching process. This
4405 saves time in cases of nested unlimited repeats that are never going to match.
4406 Problem: the scan can take a lot of time if the subject is very long (e.g.
4407 megabytes), thus penalizing straightforward matches. It is now done only if the
4408 amount of subject to be scanned is less than 1000 bytes.
4409
4410 7. A lesser problem with the same optimization is that it was recording the
4411 first character of an anchored pattern as "needed", thus provoking a search
4412 right along the subject, even when the first match of the pattern was going to
4413 fail. The "needed" character is now not set for anchored patterns, unless it
4414 follows something in the pattern that is of non-fixed length. Thus, it still
4415 fulfils its original purpose of finding quick non-matches in cases of nested
4416 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4417
4418
4419 Version 4.0 17-Feb-03
4420 ---------------------
4421
4422 1. If a comment in an extended regex that started immediately after a meta-item
4423 extended to the end of string, PCRE compiled incorrect data. This could lead to
4424 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4425
4426 2. Moved to autoconf 2.53 and libtool 1.4.2.
4427
4428 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4429 the special perltest8 script is no longer needed - all the tests can be run
4430 from a single perltest script.
4431
4432 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4433 by \s. It has now been removed in PCRE. This means it isn't recognized as
4434 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4435 class [:space:] *does* include VT, thereby creating a mess.
4436
4437 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4438 space and tab.
4439
4440 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4441 its new features into the main test script, reducing the number of scripts.
4442
4443 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4444 were backward compatible, and made the (?i) apply to the whole pattern, as if
4445 /i were given. Now it behaves more logically, and applies the option setting
4446 only to what follows. PCRE has been changed to follow suit. However, if it
4447 finds options settings right at the start of the pattern, it extracts them into
4448 the global options, as before. Thus, they show up in the info data.
4449
4450 8. Added support for the \Q...\E escape sequence. Characters in between are
4451 treated as literals. This is slightly different from Perl in that $ and @ are
4452 also handled as literals inside the quotes. In Perl, they will cause variable
4453 interpolation. Note the following examples:
4454
4455 Pattern PCRE matches Perl matches
4456
4457 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4458 \Qabc\$xyz\E abc\$xyz abc\$xyz
4459 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4460
4461 For compatibility with Perl, \Q...\E sequences are recognized inside character
4462 classes as well as outside them.
4463
4464 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4465 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4466 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4467 signed/unsigned warnings.
4468
4469 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4470 option for pcretest, so I've replaced it by a simple function that does just
4471 that job.
4472
4473 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4474 "pcregrep -".
4475
4476 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4477 Java package. This provides some syntactic sugar for simple cases of what my
4478 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4479 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4480 item, you can use this simplified notation. Note that only makes sense with
4481 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4482 greediness, whatever the setting of the PCRE_UNGREEDY option.
4483
4484 13. A change of greediness default within a pattern was not taking effect at
4485 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4486 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4487 was abstracted outside.
4488
4489 14. PCRE now supports the \G assertion. It is true when the current matching
4490 position is at the start point of the match. This differs from \A when the
4491 starting offset is non-zero. Used with the /g option of pcretest (or similar
4492 code), it works in the same way as it does for Perl's /g option. If all
4493 alternatives of a regex begin with \G, the expression is anchored to the start
4494 match position, and the "anchored" flag is set in the compiled expression.
4495
4496 15. Some bugs concerning the handling of certain option changes within patterns
4497 have been fixed. These applied to options other than (?ims). For example,
4498 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4499 been the other way round. Some of this was related to change 7 above.
4500
4501 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4502 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4503 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4504 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4505
4506 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4507 mode. Unlike ".", it always matches newline, whatever the setting of
4508 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4509 assertions. Perl allows it, but it doesn't (in general) work because it can't
4510 calculate the length of the lookbehind. At least, that's the case for Perl
4511 5.8.0 - I've been told they are going to document that it doesn't work in
4512 future.
4513
4514 18. Added an error diagnosis for escapes that PCRE does not support: these are
4515 \L, \l, \N, \P, \p, \U, \u, and \X.
4516
4517 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4518 reading past the end of the pattern in cases such as /[abcd/.
4519
4520 20. PCRE was getting more memory than necessary for patterns with classes that
4521 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4522
4523 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4524 compiling PCRE for use with Virtual Pascal.
4525
4526 22. Small fix to the Makefile to make it work properly if the build is done
4527 outside the source tree.
4528
4529 23. Added a new extension: a condition to go with recursion. If a conditional
4530 subpattern starts with (?(R) the "true" branch is used if recursion has
4531 happened, whereas the "false" branch is used only at the top level.
4532
4533 24. When there was a very long string of literal characters (over 255 bytes
4534 without UTF support, over 250 bytes with UTF support), the computation of how
4535 much memory was required could be incorrect, leading to segfaults or other
4536 strange effects.
4537
4538 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4539 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4540 there was a subsequent back reference to those brackets. This meant that, for
4541 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4542 possible to check for precisely this case. All we can do is abandon the
4543 optimization if .* occurs inside capturing brackets when there are any back
4544 references whatsoever. (See below for a better fix that came later.)
4545
4546 26. The handling of the optimization for finding the first character of a
4547 non-anchored pattern, and for finding a character that is required later in the
4548 match were failing in some cases. This didn't break the matching; it just
4549 failed to optimize when it could. The way this is done has been re-implemented.
4550
4551 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4552
4553 28. Added a new feature that provides some of the functionality that Perl
4554 provides with (?{...}). The facility is termed a "callout". The way it is done
4555 in PCRE is for the caller to provide an optional function, by setting
4556 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4557 global variable. By default it is unset, which disables all calling out. To get
4558 the function called, the regex must include (?C) at appropriate points. This
4559 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4560 This provides a means of identifying different callout points. When PCRE
4561 reaches such a point in the regex, if pcre_callout has been set, the external
4562 function is called. It is provided with data in a structure called
4563 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4564 matching continues; if it returns a non-zero value, the match at the current
4565 point fails. However, backtracking will occur if possible. [This was changed
4566 later and other features added - see item 49 below.]
4567
4568 29. pcretest is upgraded to test the callout functionality. It provides a
4569 callout function that displays information. By default, it shows the start of
4570 the match and the current position in the text. There are some new data escapes
4571 to vary what happens:
4572
4573 \C+ in addition, show current contents of captured substrings
4574 \C- do not supply a callout function
4575 \C!n return 1 when callout number n is reached
4576 \C!n!m return 1 when callout number n is reached for the mth time
4577
4578 30. If pcregrep was called with the -l option and just a single file name, it
4579 output "<stdin>" if a match was found, instead of the file name.
4580
4581 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4582 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4583 pcre_exec(). This saves a malloc/free per call. The default value of
4584 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4585 when configuring.
4586
4587 32. The default maximum size of a compiled pattern is 64K. There have been a
4588 few cases of people hitting this limit. The code now uses macros to handle the
4589 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4590 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4591 configuring. Tests 2 and 5 work only with 2-byte links because they output
4592 debugging information about compiled patterns.
4593
4594 33. Internal code re-arrangements:
4595
4596 (a) Moved the debugging function for printing out a compiled regex into
4597 its own source file (printint.c) and used #include to pull it into
4598 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4599 separate copies.
4600
4601 (b) Defined the list of op-code names for debugging as a macro in
4602 internal.h so that it is next to the definition of the opcodes.
4603
4604 (c) Defined a table of op-code lengths for simpler skipping along compiled
4605 code. This is again a macro in internal.h so that it is next to the
4606 definition of the opcodes.
4607
4608 34. Added support for recursive calls to individual subpatterns, along the
4609 lines of Robin Houston's patch (but implemented somewhat differently).
4610
4611 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4612 allow it to read and process whole directories in Win32. This code was
4613 contributed by Lionel Fourquaux; it has not been tested by me.
4614
4615 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4616 used to name a group. Names consist of alphanumerics and underscores, and must
4617 be unique. Back references use the syntax (?P=name) and recursive calls use
4618 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4619 numbers. The function pcre_fullinfo() can be used after compilation to extract
4620 a name/number map. There are three relevant calls:
4621
4622 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4623 PCRE_INFO_NAMECOUNT yields the number of entries
4624 PCRE_INFO_NAMETABLE yields a pointer to the map.
4625
4626 The map is a vector of fixed-size entries. The size of each entry depends on
4627 the length of the longest name used. The first two bytes of each entry are the
4628 group number, most significant byte first. There follows the corresponding
4629 name, zero terminated. The names are in alphabetical order.
4630
4631 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4632 case instead of 255. Making it the same both with and without UTF-8 support
4633 means that the same test output works with both.
4634
4635 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4636 calling malloc() with a zero argument.
4637
4638 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4639 optimization. I've improved things by keeping a bitmap of backreferences with
4640 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4641 fact referenced, the optimization can be applied. It is unlikely that a
4642 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4643 the match to follow \n) will appear inside brackets with a number greater than
4644 31, but if it does, any back reference > 31 suppresses the optimization.
4645
4646 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4647 of disabling numbered capturing parentheses. Any opening parenthesis that is
4648 not followed by ? behaves as if it were followed by ?: but named parentheses
4649 can still be used for capturing (and they will acquire numbers in the usual
4650 way).
4651
4652 41. Redesigned the return codes from the match() function into yes/no/error so
4653 that errors can be passed back from deep inside the nested calls. A malloc
4654 failure while inside a recursive subpattern call now causes the
4655 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4656
4657 42. It is now possible to set a limit on the number of times the match()
4658 function is called in a call to pcre_exec(). This facility makes it possible to
4659 limit the amount of recursion and backtracking, though not in a directly
4660 obvious way, because the match() function is used in a number of different
4661 circumstances. The count starts from zero for each position in the subject
4662 string (for non-anchored patterns). The default limit is, for compatibility, a
4663 large number, namely 10 000 000. You can change this in two ways:
4664
4665 (a) When configuring PCRE before making, you can use --with-match-limit=n
4666 to set a default value for the compiled library.
4667
4668 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4669 a different value is set. See 45 below.
4670
4671 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4672
4673 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4674 of things that can be changed at compile time. The first argument specifies
4675 what is wanted and the second points to where the information is to be placed.
4676 The current list of available information is:
4677
4678 PCRE_CONFIG_UTF8
4679
4680 The output is an integer that is set to one if UTF-8 support is available;
4681 otherwise it is set to zero.
4682
4683 PCRE_CONFIG_NEWLINE
4684
4685 The output is an integer that it set to the value of the code that is used for
4686 newline. It is either LF (10) or CR (13).
4687
4688 PCRE_CONFIG_LINK_SIZE
4689
4690 The output is an integer that contains the number of bytes used for internal
4691 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4692
4693 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4694
4695 The output is an integer that contains the threshold above which the POSIX
4696 interface uses malloc() for output vectors. See item 31 above.
4697
4698 PCRE_CONFIG_MATCH_LIMIT
4699
4700 The output is an unsigned integer that contains the default limit of the number
4701 of match() calls in a pcre_exec() execution. See 42 above.
4702
4703 44. pcretest has been upgraded by the addition of the -C option. This causes it
4704 to extract all the available output from the new pcre_config() function, and to
4705 output it. The program then exits immediately.
4706
4707 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4708 order to support additional features. One way would have been to define
4709 pcre_exec2() (for example) with extra arguments, but this would not have been
4710 extensible, and would also have required all calls to the original function to
4711 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4712 is used for passing in "extra" data from pcre_study().
4713
4714 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4715 contains the following fields:
4716
4717 flags a bitmap indicating which of the following fields are set
4718 study_data opaque data from pcre_study()
4719 match_limit a way of specifying a limit on match() calls for a specific
4720 call to pcre_exec()
4721 callout_data data for callouts (see 49 below)
4722
4723 The flag bits are also defined in pcre.h, and are
4724
4725 PCRE_EXTRA_STUDY_DATA
4726 PCRE_EXTRA_MATCH_LIMIT
4727 PCRE_EXTRA_CALLOUT_DATA
4728
4729 The pcre_study() function now returns one of these new pcre_extra blocks, with
4730 the actual study data pointed to by the study_data field, and the
4731 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4732 before. That is, this change is entirely upwards-compatible and requires no
4733 change to existing code.
4734
4735 If you want to pass in additional data to pcre_exec(), you can either place it
4736 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4737 block.
4738
4739 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4740 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4741 times with different match limits, until it finds the minimum value needed for
4742 pcre_exec() to complete. The value is then output. This can be instructive; for
4743 most simple matches the number is quite small, but for pathological cases it
4744 gets very large very quickly.
4745
4746 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4747 returns the size of the data block pointed to by the study_data field in a
4748 pcre_extra block, that is, the value that was passed as the argument to
4749 pcre_malloc() when PCRE was getting memory in which to place the information
4750 created by pcre_study(). The fourth argument should point to a size_t variable.
4751 pcretest has been extended so that this information is shown after a successful
4752 pcre_study() call when information about the compiled regex is being displayed.
4753
4754 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4755 because what follows is always an absolute path. (Later: it turns out that this
4756 is more than cosmetic for MinGW, because it doesn't like empty path
4757 components.)
4758
4759 49. Some changes have been made to the callout feature (see 28 above):
4760
4761 (i) A callout function now has three choices for what it returns:
4762
4763 0 => success, carry on matching
4764 > 0 => failure at this point, but backtrack if possible
4765 < 0 => serious error, return this value from pcre_exec()
4766
4767 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4768 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4769 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4770 use by callout functions. It will never be used by PCRE itself.
4771
4772 (ii) The pcre_extra structure (see 45 above) has a void * field called
4773 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4774 pcre_callout_block structure has a field of the same name. The contents of
4775 the field passed in the pcre_extra structure are passed to the callout
4776 function in the corresponding field in the callout block. This makes it
4777 easier to use the same callout-containing regex from multiple threads. For
4778 testing, the pcretest program has a new data escape
4779
4780 \C*n pass the number n (may be negative) as callout_data
4781
4782 If the callout function in pcretest receives a non-zero value as
4783 callout_data, it returns that value.
4784
4785 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4786 there were some redundant $(CFLAGS) in commands that are now specified as
4787 $(LINK), which already includes $(CFLAGS).
4788
4789 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4790 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4791 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4792 one-byte characters throughout. Note that case-insensitive matching applies
4793 only to characters whose values are less than 256. PCRE doesn't support the
4794 notion of cases for higher-valued characters.
4795
4796 (i) A character class whose characters are all within 0-255 is handled as
4797 a bit map, and the map is inverted for negative classes. Previously, a
4798 character > 255 always failed to match such a class; however it should
4799 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4800
4801 (ii) A negated character class with a single character < 255 is coded as
4802 "not this character" (OP_NOT). This wasn't working properly when the test
4803 character was multibyte, either singly or repeated.
4804
4805 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4806 mode, for example: \x{100}{2,3}.
4807
4808 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4809 singly or repeated) now correctly test multibyte characters. However,
4810 PCRE doesn't recognize any characters with values greater than 255 as
4811 digits, spaces, or word characters. Such characters always match \D, \S,
4812 and \W, and never match \d, \s, or \w.
4813
4814 (v) Classes may now contain characters and character ranges with values
4815 greater than 255. For example: [ab\x{100}-\x{400}].
4816
4817 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4818 PCRE in UTF-8 mode.
4819
4820 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4821 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4822 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4823 value.)
4824
4825 53. The single man page has become too large. I have therefore split it up into
4826 a number of separate man pages. These also give rise to individual HTML pages;
4827 these are now put in a separate directory, and there is an index.html page that
4828 lists them all. Some hyperlinking between the pages has been installed.
4829
4830 54. Added convenience functions for handling named capturing parentheses.
4831
4832 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4833 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4834 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4835 are faulted.
4836
4837 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4838 calling configure. These values are used when compiling the dftables.c program
4839 which is run to generate the source of the default character tables. They
4840 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4841 you will need to set these values.
4842
4843 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4844
4845
4846 Version 3.9 02-Jan-02
4847 ---------------------
4848
4849 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4850
4851 2. If --disable-static was given, the building process failed when trying to
4852 build pcretest and pcregrep. (For some reason it was using libtool to compile
4853 them, which is not right, as they aren't part of the library.)
4854
4855
4856 Version 3.8 18-Dec-01
4857 ---------------------
4858
4859 1. The experimental UTF-8 code was completely screwed up. It was packing the
4860 bytes in the wrong order. How dumb can you get?
4861
4862
4863 Version 3.7 29-Oct-01
4864 ---------------------
4865
4866 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4867 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4868 this didn't happen under Solaris 8, where I normally test things.
4869
4870 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4871 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4872 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4873 this fix an hour or so after the initial 3.7 release.)
4874
4875
4876 Version 3.6 23-Oct-01
4877 ---------------------
4878
4879 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4880 offsets passed as NULL with zero offset count.
4881
4882 2. The config.guess and config.sub files had not been updated when I moved to
4883 the latest autoconf.
4884
4885
4886 Version 3.5 15-Aug-01
4887 ---------------------
4888
4889 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4890 had been forgotten.
4891
4892 2. By using declared but undefined structures, we can avoid using "void"
4893 definitions in pcre.h while keeping the internal definitions of the structures
4894 private.
4895
4896 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4897 user point of view, this means that both static and shared libraries are built
4898 by default, but this can be individually controlled. More of the work of
4899 handling this static/shared cases is now inside libtool instead of PCRE's make
4900 file.
4901
4902 4. The pcretest utility is now installed along with pcregrep because it is
4903 useful for users (to test regexs) and by doing this, it automatically gets
4904 relinked by libtool. The documentation has been turned into a man page, so
4905 there are now .1, .txt, and .html versions in /doc.
4906
4907 5. Upgrades to pcregrep:
4908 (i) Added long-form option names like gnu grep.
4909 (ii) Added --help to list all options with an explanatory phrase.
4910 (iii) Added -r, --recursive to recurse into sub-directories.
4911 (iv) Added -f, --file to read patterns from a file.
4912
4913 6. pcre_exec() was referring to its "code" argument before testing that
4914 argument for NULL (and giving an error if it was NULL).
4915
4916 7. Upgraded Makefile.in to allow for compiling in a different directory from
4917 the source directory.
4918
4919 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4920 options bits, the pointer it was passed was to an int instead of to an unsigned
4921 long int. This mattered only on 64-bit systems.
4922
4923 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4924 generated) instead of pcre.in, which it its source. Also made the same change
4925 in several of the .c files.
4926
4927 10. A new release of gcc defines printf() as a macro, which broke pcretest
4928 because it had an ifdef in the middle of a string argument for printf(). Fixed
4929 by using separate calls to printf().
4930
4931 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4932 script, to force use of CR or LF instead of \n in the source. On non-Unix
4933 systems, the value can be set in config.h.
4934
4935 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4936 absolute limit. Changed the text of the error message to make this clear, and
4937 likewise updated the man page.
4938
4939 13. The limit of 99 on the number of capturing subpatterns has been removed.
4940 The new limit is 65535, which I hope will not be a "real" limit.
4941
4942
4943 Version 3.4 22-Aug-00
4944 ---------------------
4945
4946 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4947
4948 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4949
4950
4951 Version 3.3 01-Aug-00
4952 ---------------------
4953
4954 1. If an octal character was given, but the value was greater than \377, it
4955 was not getting masked to the least significant bits, as documented. This could
4956 lead to crashes in some systems.
4957
4958 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4959 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4960
4961 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4962 These just pass their arguments on to (pcre_free)(), but they are provided
4963 because some uses of PCRE bind it to non-C systems that can call its functions,
4964 but cannot call free() or pcre_free() directly.
4965
4966 4. Add "make test" as a synonym for "make check". Corrected some comments in
4967 the Makefile.
4968
4969 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4970 Makefile.
4971
4972 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4973 command called pgrep for grepping around the active processes.
4974
4975 7. Added the beginnings of support for UTF-8 character strings.
4976
4977 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4978 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4979 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4980 out for the ar command.)
4981
4982
4983 Version 3.2 12-May-00
4984 ---------------------
4985
4986 This is purely a bug fixing release.
4987
4988 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4989 of ZA. This was just one example of several cases that could provoke this bug,
4990 which was introduced by change 9 of version 2.00. The code for breaking
4991 infinite loops after an iteration that matches an empty string was't working
4992 correctly.
4993
4994 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4995 when matched against abbab (for example). After matching an empty string, it
4996 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4997 caused it to match further down the string than it should.
4998
4999 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5000 was there because it doesn't seem to be needed, and it causes trouble on some
5001 systems, as it is not a Standard C header. It has been removed.
5002
5003 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5004 were reported on the Macintosh. The changes were from
5005
5006 while ((c = *(++ptr)) != 0 && c != '\n');
5007 to
5008 while ((c = *(++ptr)) != 0 && c != '\n') ;
5009
5010 Totally extraordinary, but if that's what it takes...
5011
5012 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5013 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5014 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5015 assumes the way PCRE uses memmove() (always moving upwards).
5016
5017 6. PCRE is being used in one environment where strchr() is not available. There
5018 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5019 faster code anyway.
5020
5021
5022 Version 3.1 09-Feb-00
5023 ---------------------
5024
5025 The only change in this release is the fixing of some bugs in Makefile.in for
5026 the "install" target:
5027
5028 (1) It was failing to install pcreposix.h.
5029
5030 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5031
5032
5033 Version 3.0 01-Feb-00
5034 ---------------------
5035
5036 1. Add support for the /+ modifier to perltest (to output $` like it does in
5037 pcretest).
5038
5039 2. Add support for the /g modifier to perltest.
5040
5041 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5042 matches null strings.
5043
5044 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5045 pattern. Perl treats empty patterns specially - it reuses the most recent
5046 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5047 effect.
5048
5049 5. The POSIX interface was broken in that it was just handing over the POSIX
5050 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5051 required a bigger vector, with some working space on the end. This means that
5052 the POSIX wrapper now has to get and free some memory, and copy the results.
5053
5054 6. Added some simple autoconf support, placing the test data and the
5055 documentation in separate directories, re-organizing some of the
5056 information files, and making it build pcre-config (a GNU standard). Also added
5057 libtool support for building PCRE as a shared library, which is now the
5058 default.
5059
5060 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5061 09 are not valid octal constants. Single digits will be used for minor values
5062 less than 10.
5063
5064 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5065 existing programs that set these in the POSIX interface can use PCRE without
5066 modification.
5067
5068 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5069 return all that pcre_info() returns, plus additional data. The pcre_info()
5070 function is retained for compatibility, but is considered to be obsolete.
5071
5072 10. Added experimental recursion feature (?R) to handle one common case that
5073 Perl 5.6 will be able to do with (?p{...}).
5074
5075 11. Added support for POSIX character classes like [:alpha:], which Perl is
5076 adopting.
5077
5078
5079 Version 2.08 31-Aug-99
5080 ----------------------
5081
5082 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5083 trying to match at the startoffset position, but instead was moving forward to
5084 the next newline as if a previous match had failed.
5085
5086 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5087 and could get into a loop if a null string was matched other than at the start
5088 of the subject.
5089
5090 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5091 be distinguished at compile time, and for completeness also added PCRE_DATE.
5092
5093 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5094 in GnuWin32 environments.
5095
5096
5097 Version 2.07 29-Jul-99
5098 ----------------------
5099
5100 1. The documentation is now supplied in plain text form and HTML as well as in
5101 the form of man page sources.
5102
5103 2. C++ compilers don't like assigning (void *) values to other pointer types.
5104 In particular this affects malloc(). Although there is no problem in Standard
5105 C, I've put in casts to keep C++ compilers happy.
5106
5107 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5108 should be (const char *).
5109
5110 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5111 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5112 However, I haven't made this a standard facility. The documentation doesn't
5113 mention it, and the Makefile doesn't support it.
5114
5115 5. The Makefile now contains an "install" target, with editable destinations at
5116 the top of the file. The pcretest program is not installed.
5117
5118 6. pgrep -V now gives the PCRE version number and date.
5119
5120 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5121 causing the entire string to be ignored, instead of just the last character.
5122
5123 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5124 non-matching string, it can take a very, very long time, even for strings of
5125 quite modest length, because of the nested recursion. PCRE now does better in
5126 some of these cases. It does this by remembering the last required literal
5127 character in the pattern, and pre-searching the subject to ensure it is present
5128 before running the real match. In other words, it applies a heuristic to detect
5129 some types of certain failure quickly, and in the above example, if presented
5130 with a string that has no trailing " it gives "no match" very quickly.
5131
5132 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5133 other alternatives are tried instead.
5134
5135
5136 Version 2.06 09-Jun-99
5137 ----------------------
5138
5139 1. Change pcretest's output for amount of store used to show just the code
5140 space, because the remainder (the data block) varies in size between 32-bit and
5141 64-bit systems.
5142
5143 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5144 start matching at. This allows lookbehinds to work when searching for multiple
5145 occurrences in a string.
5146
5147 3. Added additional options to pcretest for testing multiple occurrences:
5148
5149 /+ outputs the rest of the string that follows a match
5150 /g loops for multiple occurrences, using the new startoffset argument
5151 /G loops for multiple occurrences by passing an incremented pointer
5152
5153 4. PCRE wasn't doing the "first character" optimization for patterns starting
5154 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5155 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5156 the letter 'x'. On long subject strings, this gives a significant speed-up.
5157
5158
5159 Version 2.05 21-Apr-99
5160 ----------------------
5161
5162 1. Changed the type of magic_number from int to long int so that it works
5163 properly on 16-bit systems.
5164
5165 2. Fixed a bug which caused patterns starting with .* not to work correctly
5166 when the subject string contained newline characters. PCRE was assuming
5167 anchoring for such patterns in all cases, which is not correct because .* will
5168 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5169 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5170 must be retried after every newline in the subject.
5171
5172
5173 Version 2.04 18-Feb-99
5174 ----------------------
5175
5176 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5177 computation of the store needed to hold the pattern was incorrect (too large).
5178 If such patterns were nested a few deep, this could multiply and become a real
5179 problem.
5180
5181 2. Added /M option to pcretest to show the memory requirement of a specific
5182 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5183
5184 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5185 compiled in such a way that the backtracking after subsequent failure was
5186 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5187 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5188
5189
5190 Version 2.03 02-Feb-99
5191 ----------------------
5192
5193 1. Fixed typo and small mistake in man page.
5194
5195 2. Added 4th condition (GPL supersedes if conflict) and created separate
5196 LICENCE file containing the conditions.
5197
5198 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5199 Perl, that is the internal \ allows the delimiter to be included in the
5200 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5201 the final delimiter, add \ to the end of the pattern (to test the error).
5202
5203 4. Added the convenience functions for extracting substrings after a successful
5204 match. Updated pcretest to make it able to test these functions.
5205
5206
5207 Version 2.02 14-Jan-99
5208 ----------------------
5209
5210 1. Initialized the working variables associated with each extraction so that
5211 their saving and restoring doesn't refer to uninitialized store.
5212
5213 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5214 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5215 fix the problem.
5216
5217 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5218 calls, and wasn't printing the correct value for compiling calls. Increased the
5219 default value of LOOPREPEAT, and the number of significant figures in the
5220 times.
5221
5222 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5223
5224 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5225 a building problem on Windows NT with a FAT file system.
5226
5227
5228 Version 2.01 21-Oct-98
5229 ----------------------
5230
5231 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5232 to character tables built by pcre_maketables() in the current locale. If NULL
5233 is passed, the default tables are used.
5234
5235
5236 Version 2.00 24-Sep-98
5237 ----------------------
5238
5239 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5240 it any more.
5241
5242 2. Allow quantification of (?>) groups, and make it work correctly.
5243
5244 3. The first character computation wasn't working for (?>) groups.
5245
5246 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5247 end of the subject) and add 5.005's \z, which really does match only at the
5248 very end of the subject.
5249
5250 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5251
5252 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5253 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5254 localized options. All options to pcre_study() were also removed.
5255
5256 7. Add other new features from 5.005:
5257
5258 $(?<= positive lookbehind
5259 $(?<! negative lookbehind
5260 (?imsx-imsx) added the unsetting capability
5261 such a setting is global if at outer level; local otherwise
5262 (?imsx-imsx:) non-capturing groups with option setting
5263 (?(cond)re|re) conditional pattern matching
5264
5265 A backreference to itself in a repeated group matches the previous
5266 captured string.
5267
5268 8. General tidying up of studying (both automatic and via "study")
5269 consequential on the addition of new assertions.
5270
5271 9. As in 5.005, unlimited repeated groups that could match an empty substring
5272 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5273 runtime if any iteration does actually match an empty substring.
5274
5275 10. Include the RunTest script in the distribution.
5276
5277 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5278 discrepancies, some of which were old and were also with respect to 5.004. They
5279 have now been fixed.
5280
5281
5282 Version 1.09 28-Apr-98
5283 ----------------------
5284
5285 1. A negated single character class followed by a quantifier with a minimum
5286 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5287 program crashes, or just wrong answers. This did not apply to negated classes
5288 containing more than one character, or to minima other than one.
5289
5290
5291 Version 1.08 27-Mar-98
5292 ----------------------
5293
5294 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5295
5296 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5297 latter must appear before anything that relies on it in the pattern.
5298
5299
5300 Version 1.07 16-Feb-98
5301 ----------------------
5302
5303 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5304 repeat of a potentially empty string).
5305
5306
5307 Version 1.06 23-Jan-98
5308 ----------------------
5309
5310 1. Added Markus Oberhumer's little patches for C++.
5311
5312 2. Literal strings longer than 255 characters were broken.
5313
5314
5315 Version 1.05 23-Dec-97
5316 ----------------------
5317
5318 1. Negated character classes containing more than one character were failing if
5319 PCRE_CASELESS was set at run time.
5320
5321
5322 Version 1.04 19-Dec-97
5323 ----------------------
5324
5325 1. Corrected the man page, where some "const" qualifiers had been omitted.
5326
5327 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5328 input syntax.
5329
5330 3. Fixed memory leak which occurred when a regex with back references was
5331 matched with an offsets vector that wasn't big enough. The temporary memory
5332 that is used in this case wasn't being freed if the match failed.
5333
5334 4. Tidied pcretest to ensure it frees memory that it gets.
5335
5336 5. Temporary memory was being obtained in the case where the passed offsets
5337 vector was exactly big enough.
5338
5339 6. Corrected definition of offsetof() from change 5 below.
5340
5341 7. I had screwed up change 6 below and broken the rules for the use of
5342 setjmp(). Now fixed.
5343
5344
5345 Version 1.03 18-Dec-97
5346 ----------------------
5347
5348 1. A erroneous regex with a missing opening parenthesis was correctly
5349 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5350 on some systems.
5351
5352 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5353 it was reported that one broken compiler failed on the former because "code" is
5354 also an independent variable.
5355
5356 3. The erroneous regex a[]b caused an array overrun reference.
5357
5358 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5359 fail on data ending with that character. (It was going on too far, and checking
5360 the next character, typically a binary zero.) This was specific to the
5361 optimized code for single-character negative classes.
5362
5363 5. Added a contributed patch from the TIN world which does the following:
5364
5365 + Add an undef for memmove, in case the the system defines a macro for it.
5366
5367 + Add a definition of offsetof(), in case there isn't one. (I don't know
5368 the reason behind this - offsetof() is part of the ANSI standard - but
5369 it does no harm).
5370
5371 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5372 most of the places where whitespace preceded '#'. I have given up and
5373 allowed the remaining 2 cases to be at the margin.
5374
5375 + Rename some variables in pcre to eliminate shadowing. This seems very
5376 pedantic, but does no harm, of course.
5377
5378 6. Moved the call to setjmp() into its own function, to get rid of warnings
5379 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5380
5381 7. Constructs such as \d{8,} were compiling into the equivalent of
5382 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5383 outcome, but in this particular case used more store than had been allocated,
5384 which caused the bug to be discovered because it threw up an internal error.
5385
5386 8. The debugging code in both pcre and pcretest for outputting the compiled
5387 form of a regex was going wrong in the case of back references followed by
5388 curly-bracketed repeats.
5389
5390
5391 Version 1.02 12-Dec-97
5392 ----------------------
5393
5394 1. Typos in pcre.3 and comments in the source fixed.
5395
5396 2. Applied a contributed patch to get rid of places where it used to remove
5397 'const' from variables, and fixed some signed/unsigned and uninitialized
5398 variable warnings.
5399
5400 3. Added the "runtest" target to Makefile.
5401
5402 4. Set default compiler flag to -O2 rather than just -O.
5403
5404
5405 Version 1.01 19-Nov-97
5406 ----------------------
5407
5408 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5409 like /([ab]*)*/, that is, for classes with more than one character in them.
5410
5411 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5412 as /((?>a*))*/ (a PCRE_EXTRA facility).
5413
5414
5415 Version 1.00 18-Nov-97
5416 ----------------------
5417
5418 1. Added compile-time macros to support systems such as SunOS4 which don't have
5419 memmove() or strerror() but have other things that can be used instead.
5420
5421 2. Arranged that "make clean" removes the executables.
5422
5423
5424 Version 0.99 27-Oct-97
5425 ----------------------
5426
5427 1. Fixed bug in code for optimizing classes with only one character. It was
5428 initializing a 32-byte map regardless, which could cause it to run off the end
5429 of the memory it had got.
5430
5431 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5432
5433
5434 Version 0.98 22-Oct-97
5435 ----------------------
5436
5437 1. Fixed bug in code for handling temporary memory usage when there are more
5438 back references than supplied space in the ovector. This could cause segfaults.
5439
5440
5441 Version 0.97 21-Oct-97
5442 ----------------------
5443
5444 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5445
5446 2. Optimized negated single characters not to use a bit map.
5447
5448 3. Brought error texts together as macro definitions; clarified some of them;
5449 fixed one that was wrong - it said "range out of order" when it meant "invalid
5450 escape sequence".
5451
5452 4. Changed some char * arguments to const char *.
5453
5454 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5455
5456 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5457 pcretest.
5458
5459
5460 Version 0.96 16-Oct-97
5461 ----------------------
5462
5463 1. Added a simple "pgrep" utility to the distribution.
5464
5465 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5466 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5467 where "ddd" means "one or more decimal digits".
5468
5469 3. Fixed serious bug. If a pattern had a back reference, but the call to
5470 pcre_exec() didn't supply a large enough ovector to record the related
5471 identifying subpattern, the match always failed. PCRE now remembers the number
5472 of the largest back reference, and gets some temporary memory in which to save
5473 the offsets during matching if necessary, in order to ensure that
5474 backreferences always work.
5475
5476 4. Increased the compatibility with Perl in a number of ways:
5477
5478 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5479 to request this handling. The option can be set at compile or exec time.
5480
5481 (b) $ matches before a terminating newline by default; an option
5482 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5483 mode). The option can be set at compile or exec time.
5484
5485 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5486 the same as Perl's. If the decimal number it represents is less than 10
5487 or there aren't that many previous left capturing parentheses, an octal
5488 escape is read. Inside a character class, it's always an octal escape,
5489 even if it is a single digit.
5490
5491 (d) An escaped but undefined alphabetic character is taken as a literal,
5492 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5493 escapes.
5494
5495 (e) {0} is now permitted. (The previous item is removed from the compiled
5496 pattern).
5497
5498 5. Changed all the names of code files so that the basic parts are no longer
5499 than 10 characters, and abolished the teeny "globals.c" file.
5500
5501 6. Changed the handling of character classes; they are now done with a 32-byte
5502 bit map always.
5503
5504 7. Added the -d and /D options to pcretest to make it possible to look at the
5505 internals of compilation without having to recompile pcre.
5506
5507
5508 Version 0.95 23-Sep-97
5509 ----------------------
5510
5511 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5512 \x20 at the start of a run of normal characters. These were being treated as
5513 real characters, instead of the source characters being re-checked.
5514
5515
5516 Version 0.94 18-Sep-97
5517 ----------------------
5518
5519 1. The functions are now thread-safe, with the caveat that the global variables
5520 containing pointers to malloc() and free() or alternative functions are the
5521 same for all threads.
5522
5523 2. Get pcre_study() to generate a bitmap of initial characters for non-
5524 anchored patterns when this is possible, and use it if passed to pcre_exec().
5525
5526
5527 Version 0.93 15-Sep-97
5528 ----------------------
5529
5530 1. /(b)|(:+)/ was computing an incorrect first character.
5531
5532 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5533 but not actually doing anything yet.
5534
5535 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5536 as Perl does (e.g. [-az] or [az-]).
5537
5538 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5539 all possible positions.
5540
5541 5. Split up into different modules to avoid including unneeded functions in a
5542 compiled binary. However, compile and exec are still in one module. The "study"
5543 function is split off.
5544
5545 6. The character tables are now in a separate module whose source is generated
5546 by an auxiliary program - but can then be edited by hand if required. There are
5547 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5548 toupper() in the code.
5549
5550 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5551 make them global. Abolish the function for setting them, as the caller can now
5552 set them directly.
5553
5554
5555 Version 0.92 11-Sep-97
5556 ----------------------
5557
5558 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5559 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5560
5561 2. Caseless matching was not working in character classes if the characters in
5562 the pattern were in upper case.
5563
5564 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5565
5566 4. Make PCRE_ANCHORED public and accept as a compile option.
5567
5568 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5569 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5570 pass them.
5571
5572 6. Give an error if bad option bits passed at compile or run time.
5573
5574 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5575 pcretest to cause it to pass that flag.
5576
5577 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5578 options, and the first character, if set.
5579
5580 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5581
5582
5583 Version 0.91 10-Sep-97
5584 ----------------------
5585
5586 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5587 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5588
5589 2. PCRE was looping on encountering an indefinitely repeated back reference to
5590 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5591 Perl does - treats the match as successful.
5592
5593 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5