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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1528 - (show annotations) (download)
Sat Feb 28 11:48:09 2015 UTC (33 hours, 17 minutes ago) by ph10
File size: 272851 byte(s)
Fix bad compile for groups like "((?2){0,1999}())?".

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.37 xx-xxx-2015
5 ------------------------
6
7 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8 for those parentheses to be closed with whatever has been captured so far.
9 However, it was failing to mark any other groups between the hightest
10 capture so far and the currrent group as "unset". Thus, the ovector for
11 those groups contained whatever was previously there. An example is the
12 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
13
14 2. If an assertion condition was quantified with a minimum of zero (an odd
15 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
16
17 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
18 unrecognized modifier, a crash could occur.
19
20 4. An attempt to do global matching in pcretest with a zero-length ovector
21 caused a crash.
22
23 5. Fixed a memory leak during matching that could occur for a subpattern
24 subroutine call (recursive or otherwise) if the number of captured groups
25 that had to be saved was greater than ten.
26
27 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
28 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
29 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
30
31 7. A UTF pattern containing a "not" match of a non-ASCII character and a
32 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
33
34 8. When a pattern is compiled, it remembers the highest back reference so that
35 when matching, if the ovector is too small, extra memory can be obtained to
36 use instead. A conditional subpattern whose condition is a check on a
37 capture having happened, such as, for example in the pattern
38 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
39 setting the highest backreference number. This mattered only if pcre_exec()
40 was called with an ovector that was too small to hold the capture, and there
41 was no other kind of back reference (a situation which is probably quite
42 rare). The effect of the bug was that the condition was always treated as
43 FALSE when the capture could not be consulted, leading to a incorrect
44 behaviour by pcre2_match(). This bug has been fixed.
45
46 9. A reference to a duplicated named group (either a back reference or a test
47 for being set in a conditional) that occurred in a part of the pattern where
48 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
49 to be incorrectly calculated, leading to overwriting.
50
51 10. A mutually recursive set of back references such as (\2)(\1) caused a
52 segfault at study time (while trying to find the minimum matching length).
53 The infinite loop is now broken (with the minimum length unset, that is,
54 zero).
55
56 11. If an assertion that was used as a condition was quantified with a minimum
57 of zero, matching went wrong. In particular, if the whole group had
58 unlimited repetition and could match an empty string, a segfault was
59 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
60 assertions to be quantified, but not if they are being used as conditions,
61 so the above pattern is faulted by Perl. PCRE has now been changed so that
62 it also rejects such patterns.
63
64 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
65 failed to allow the zero-repeat case if pcre2_exec() was called with an
66 ovector too small to capture the group.
67
68 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
69 Red Hat Product Security:
70
71 (a) A crash if /K and /F were both set with the option to save the compiled
72 pattern.
73
74 (b) Another crash if the option to print captured substrings in a callout
75 was combined with setting a null ovector, for example \O\C+ as a subject
76 string.
77
78 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
79 forward reference repeated a large (but limited) number of times within a
80 repeated outer group that has a zero minimum quantifier, caused incorrect
81 code to be compiled, leading to the error "internal error:
82 previously-checked referenced subpattern not found" when an incorrect
83 memory address was read. This bug was reported as "heap overflow",
84 discovered by Kai Lu of Fortinet's FortiGuard Labs.
85
86
87 Version 8.36 26-September-2014
88 ------------------------------
89
90 1. Got rid of some compiler warnings in the C++ modules that were shown up by
91 -Wmissing-field-initializers and -Wunused-parameter.
92
93 2. The tests for quantifiers being too big (greater than 65535) were being
94 applied after reading the number, and stupidly assuming that integer
95 overflow would give a negative number. The tests are now applied as the
96 numbers are read.
97
98 3. Tidy code in pcre_exec.c where two branches that used to be different are
99 now the same.
100
101 4. The JIT compiler did not generate match limit checks for certain
102 bracketed expressions with quantifiers. This may lead to exponential
103 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
104 issue should be resolved now.
105
106 5. Fixed an issue, which occures when nested alternatives are optimized
107 with table jumps.
108
109 6. Inserted two casts and changed some ints to size_t in the light of some
110 reported 64-bit compiler warnings (Bugzilla 1477).
111
112 7. Fixed a bug concerned with zero-minimum possessive groups that could match
113 an empty string, which sometimes were behaving incorrectly in the
114 interpreter (though correctly in the JIT matcher). This pcretest input is
115 an example:
116
117 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
118 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
119
120 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
121 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
122 for an empty string was breaking the inner loop and carrying on at a lower
123 level, when possessive repeated groups should always return to a higher
124 level as they have no backtrack points in them. The empty string test now
125 occurs at the outer level.
126
127 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
128 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
129
130 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
131 doesn't).
132
133 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
134 to hold a pattern to be incorrectly computed (too small) when there were
135 named back references to duplicated names. This could cause "internal
136 error: code overflow" or "double free or corruption" or other memory
137 handling errors.
138
139 11. When named subpatterns had the same prefixes, back references could be
140 confused. For example, in this pattern:
141
142 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
143
144 the reference to 'Name' was incorrectly treated as a reference to a
145 duplicate name.
146
147 12. A pattern such as /^s?c/mi8 where the optional character has more than
148 one "other case" was incorrectly compiled such that it would only try to
149 match starting at "c".
150
151 13. When a pattern starting with \s was studied, VT was not included in the
152 list of possible starting characters; this should have been part of the
153 8.34/18 patch.
154
155 14. If a character class started [\Qx]... where x is any character, the class
156 was incorrectly terminated at the ].
157
158 15. If a pattern that started with a caseless match for a character with more
159 than one "other case" was studied, PCRE did not set up the starting code
160 unit bit map for the list of possible characters. Now it does. This is an
161 optimization improvement, not a bug fix.
162
163 16. The Unicode data tables have been updated to Unicode 7.0.0.
164
165 17. Fixed a number of memory leaks in pcregrep.
166
167 18. Avoid a compiler warning (from some compilers) for a function call with
168 a cast that removes "const" from an lvalue by using an intermediate
169 variable (to which the compiler does not object).
170
171 19. Incorrect code was compiled if a group that contained an internal recursive
172 back reference was optional (had quantifier with a minimum of zero). This
173 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
174 caused segmentation faults because of stack overflows at compile time.
175
176 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
177 group that is quantified with an indefinite repeat, caused a compile-time
178 loop which used up all the system stack and provoked a segmentation fault.
179 This was not the same bug as 19 above.
180
181 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
182 Patch by Mike Frysinger.
183
184
185 Version 8.35 04-April-2014
186 --------------------------
187
188 1. A new flag is set, when property checks are present in an XCLASS.
189 When this flag is not set, PCRE can perform certain optimizations
190 such as studying these XCLASS-es.
191
192 2. The auto-possessification of character sets were improved: a normal
193 and an extended character set can be compared now. Furthermore
194 the JIT compiler optimizes more character set checks.
195
196 3. Got rid of some compiler warnings for potentially uninitialized variables
197 that show up only when compiled with -O2.
198
199 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
200 of a match later then the end of the match. The pcretest program was not
201 handling the case sensibly - it was outputting from the start to the next
202 binary zero. It now reports this situation in a message, and outputs the
203 text from the end to the start.
204
205 5. Fast forward search is improved in JIT. Instead of the first three
206 characters, any three characters with fixed position can be searched.
207 Search order: first, last, middle.
208
209 6. Improve character range checks in JIT. Characters are read by an inprecise
210 function now, which returns with an unknown value if the character code is
211 above a certain threshold (e.g: 256). The only limitation is that the value
212 must be bigger than the threshold as well. This function is useful when
213 the characters above the threshold are handled in the same way.
214
215 7. The macros whose names start with RAWUCHAR are placeholders for a future
216 mode in which only the bottom 21 bits of 32-bit data items are used. To
217 make this more memorable for those maintaining the code, the names have
218 been changed to start with UCHAR21, and an extensive comment has been added
219 to their definition.
220
221 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
222 to the export list in Makefile.am (they were accidentally omitted from the
223 8.34 tarball).
224
225 9. The informational output from pcretest used the phrase "starting byte set"
226 which is inappropriate for the 16-bit and 32-bit libraries. As the output
227 for "first char" and "need char" really means "non-UTF-char", I've changed
228 "byte" to "char", and slightly reworded the output. The documentation about
229 these values has also been (I hope) clarified.
230
231 10. Another JIT related optimization: use table jumps for selecting the correct
232 backtracking path, when more than four alternatives are present inside a
233 bracket.
234
235 11. Empty match is not possible, when the minimum length is greater than zero,
236 and there is no \K in the pattern. JIT should avoid empty match checks in
237 such cases.
238
239 12. In a caseless character class with UCP support, when a character with more
240 than one alternative case was not the first character of a range, not all
241 the alternative cases were added to the class. For example, s and \x{17f}
242 are both alternative cases for S: the class [RST] was handled correctly,
243 but [R-T] was not.
244
245 13. The configure.ac file always checked for pthread support when JIT was
246 enabled. This is not used in Windows, so I have put this test inside a
247 check for the presence of windows.h (which was already tested for).
248
249 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
250 The algorithm provides a way to skip certain starting offsets, and usually
251 faster than linear prefix searches.
252
253 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
254 as for 'fr_FR' and 'french'. For some reason, however, it then used the
255 Windows-specific input and output files, which have 'french' screwed in.
256 So this could never have worked. One of the problems with locales is that
257 they aren't always the same. I have now updated RunTest so that it checks
258 the output of the locale test (test 3) against three different output
259 files, and it allows the test to pass if any one of them matches. With luck
260 this should make the test pass on some versions of Solaris where it was
261 failing. Because of the uncertainty, the script did not used to stop if
262 test 3 failed; it now does. If further versions of a French locale ever
263 come to light, they can now easily be added.
264
265 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
266 there was a message during ./configure, but it did not stop. This now
267 provokes an error. The invalid example in README has been corrected.
268 If a value less than the minimum is given, the minimum value has always
269 been used, but now a warning is given.
270
271 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
272 was a bug in the test system, which is now fixed. Also, the list of various
273 configurations that are tested for each release did not have one with both
274 16/32 bits and --enable-bar-anycrlf. It now does.
275
276 18. pcretest was missing "-C bsr" for displaying the \R default setting.
277
278 19. Little endian PowerPC systems are supported now by the JIT compiler.
279
280 20. The fast forward newline mechanism could enter to an infinite loop on
281 certain invalid UTF-8 input. Although we don't support these cases
282 this issue can be fixed by a performance optimization.
283
284 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
285 not take account if existing stack usage. There is now a new global
286 variable called pcre_stack_guard that can be set to point to an external
287 function to check stack availability. It is called at the start of
288 processing every parenthesized group.
289
290 22. A typo in the code meant that in ungreedy mode the max/min qualifier
291 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
292 not match "ab".
293
294 23. When UTF was disabled, the JIT program reported some incorrect compile
295 errors. These messages are silenced now.
296
297 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
298 compiler.
299
300 25. Change all the temporary files used in RunGrepTest to be different to those
301 used by RunTest so that the tests can be run simultaneously, for example by
302 "make -j check".
303
304
305 Version 8.34 15-December-2013
306 -----------------------------
307
308 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
309 executable memory. Patch inspired by Carsten Klein.
310
311 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
312 this macro is never tested and has no effect, because the work to support
313 coverage involves only compiling and linking options and special targets in
314 the Makefile. The comment in config.h implied that defining the macro would
315 enable coverage support, which is totally false. There was also support for
316 setting this macro in the CMake files (my fault, I just copied it from
317 configure). SUPPORT_GCOV has now been removed.
318
319 3. Make a small performance improvement in strlen16() and strlen32() in
320 pcretest.
321
322 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
323 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
324 these cases). There was also one in pcretest.c.
325
326 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
327
328 6. In UTF mode, the code for checking whether a group could match an empty
329 string (which is used for indefinitely repeated groups to allow for
330 breaking an infinite loop) was broken when the group contained a repeated
331 negated single-character class with a character that occupied more than one
332 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
333 UTF-8 mode). The effect was undefined: the group might or might not be
334 deemed as matching an empty string, or the program might have crashed.
335
336 7. The code for checking whether a group could match an empty string was not
337 recognizing that \h, \H, \v, \V, and \R must match a character.
338
339 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
340 an empty string. If it can, pcretest shows this in its information output.
341
342 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
343 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
344 matched by pcre_exec() without using JIT:
345
346 (a) If the rest of the pattern did not match after a maximal run of
347 grapheme clusters, the code for backing up to try with fewer of them
348 did not always back up over a full grapheme when characters that do not
349 have the modifier quality were involved, e.g. Hangul syllables.
350
351 (b) If the match point in a subject started with modifier character, and
352 there was no match, the code could incorrectly back up beyond the match
353 point, and potentially beyond the first character in the subject,
354 leading to a segfault or an incorrect match result.
355
356 10. A conditional group with an assertion condition could lead to PCRE
357 recording an incorrect first data item for a match if no other first data
358 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
359 first data item, and therefore matched "ca" after "c" instead of at the
360 start.
361
362 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
363 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
364
365 12. The source of pcregrep now includes z/OS-specific code so that it can be
366 compiled for z/OS as part of the special z/OS distribution.
367
368 13. Added the -T and -TM options to pcretest.
369
370 14. The code in pcre_compile.c for creating the table of named capturing groups
371 has been refactored. Instead of creating the table dynamically during the
372 actual compiling pass, the information is remembered during the pre-compile
373 pass (on the stack unless there are more than 20 named groups, in which
374 case malloc() is used) and the whole table is created before the actual
375 compile happens. This has simplified the code (it is now nearly 150 lines
376 shorter) and prepared the way for better handling of references to groups
377 with duplicate names.
378
379 15. A back reference to a named subpattern when there is more than one of the
380 same name now checks them in the order in which they appear in the pattern.
381 The first one that is set is used for the reference. Previously only the
382 first one was inspected. This change makes PCRE more compatible with Perl.
383
384 16. Unicode character properties were updated from Unicode 6.3.0.
385
386 17. The compile-time code for auto-possessification has been refactored, based
387 on a patch by Zoltan Herczeg. It now happens after instead of during
388 compilation. The code is cleaner, and more cases are handled. The option
389 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
390 options in pcretest are provided to set it. It can also be set by
391 (*NO_AUTO_POSSESS) at the start of a pattern.
392
393 18. The character VT has been added to the default ("C" locale) set of
394 characters that match \s and are generally treated as white space,
395 following this same change in Perl 5.18. There is now no difference between
396 "Perl space" and "POSIX space". Whether VT is treated as white space in
397 other locales depends on the locale.
398
399 19. The code for checking named groups as conditions, either for being set or
400 for being recursed, has been refactored (this is related to 14 and 15
401 above). Processing unduplicated named groups should now be as fast at
402 numerical groups, and processing duplicated groups should be faster than
403 before.
404
405 20. Two patches to the CMake build system, by Alexander Barkov:
406
407 (1) Replace the "source" command by "." in CMakeLists.txt because
408 "source" is a bash-ism.
409
410 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
411 without these the CMake build does not work on Solaris.
412
413 21. Perl has changed its handling of \8 and \9. If there is no previously
414 encountered capturing group of those numbers, they are treated as the
415 literal characters 8 and 9 instead of a binary zero followed by the
416 literals. PCRE now does the same.
417
418 22. Following Perl, added \o{} to specify codepoints in octal, making it
419 possible to specify values greater than 0777 and also making them
420 unambiguous.
421
422 23. Perl now gives an error for missing closing braces after \x{... instead of
423 treating the string as literal. PCRE now does the same.
424
425 24. RunTest used to grumble if an inappropriate test was selected explicitly,
426 but just skip it when running all tests. This make it awkward to run ranges
427 of tests when one of them was inappropriate. Now it just skips any
428 inappropriate tests, as it always did when running all tests.
429
430 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
431 character types such as \d or \w, too many callouts were inserted, and the
432 data that they returned was rubbish.
433
434 26. In UCP mode, \s was not matching two of the characters that Perl matches,
435 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
436 were matched by \h. The code has now been refactored so that the lists of
437 the horizontal and vertical whitespace characters used for \h and \v (which
438 are defined only in one place) are now also used for \s.
439
440 27. Add JIT support for the 64 bit TileGX architecture.
441 Patch by Jiong Wang (Tilera Corporation).
442
443 28. Possessive quantifiers for classes (both explicit and automatically
444 generated) now use special opcodes instead of wrapping in ONCE brackets.
445
446 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
447 (because it's meaningless), this was not happening when PCRE_CASELESS was
448 set. Not wrong, but inefficient.
449
450 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
451 properties for \w, \d, etc) is present in a test regex. Otherwise if the
452 test contains no characters greater than 255, Perl doesn't realise it
453 should be using Unicode semantics.
454
455 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
456 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
457 does in Unicode mode.
458
459 32. Added the "forbid" facility to pcretest so that putting tests into the
460 wrong test files can sometimes be quickly detected.
461
462 33. There is now a limit (default 250) on the depth of nesting of parentheses.
463 This limit is imposed to control the amount of system stack used at compile
464 time. It can be changed at build time by --with-parens-nest-limit=xxx or
465 the equivalent in CMake.
466
467 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
468 errors. Perl warns for these when in warning mode, but PCRE has no facility
469 for giving warnings.
470
471 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
472 However, this was not working for (?!) because it is optimized to (*FAIL),
473 for which PCRE does not allow quantifiers. The optimization is now disabled
474 when a quantifier follows (?!). I can't see any use for this, but it makes
475 things uniform.
476
477 36. Perl no longer allows group names to start with digits, so I have made this
478 change also in PCRE. It simplifies the code a bit.
479
480 37. In extended mode, Perl ignores spaces before a + that indicates a
481 possessive quantifier. PCRE allowed a space before the quantifier, but not
482 before the possessive +. It now does.
483
484 38. The use of \K (reset reported match start) within a repeated possessive
485 group such as (a\Kb)*+ was not working.
486
487 40. Document that the same character tables must be used at compile time and
488 run time, and that the facility to pass tables to pcre_exec() and
489 pcre_dfa_exec() is for use only with saved/restored patterns.
490
491 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
492 features for Builds with MSVC:
493
494 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
495 builds). The .rc files can be used to set FileDescription and many other
496 attributes.
497
498 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
499 This allows higher-level build scripts which want .pdb files to avoid
500 hard-coding the exact files needed."
501
502 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
503 mean "start of word" and "end of word", respectively, as a transition aid.
504
505 43. A minimizing repeat of a class containing codepoints greater than 255 in
506 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
507 compiled to use the heap for recursion.
508
509 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
510 is configured.
511
512
513 Version 8.33 28-May-2013
514 ------------------------
515
516 1. Added 'U' to some constants that are compared to unsigned integers, to
517 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
518 variables that are added to signed variables, to ensure the result is
519 signed and can be negated.
520
521 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
522 CMake config files.
523
524 3. Revise the creation of config.h.generic so that all boolean macros are
525 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
526 overriding via -D on the command line possible.
527
528 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
529 to unsigned int is reported to make a quite noticeable speed difference in
530 a specific Windows environment. Testing on Linux did also appear to show
531 some benefit (and it is clearly not harmful). Also fixed the definition of
532 Xop which should be unsigned.
533
534 5. Related to (4), changing the definition of the intermediate variable cc
535 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
536 improvements.
537
538 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
539 unnecessary spaces.
540
541 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
542 and later.
543
544 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
545
546 9. Optimizing fast_forward_start_bits in JIT.
547
548 10. Adding support for callouts in JIT, and fixing some issues revealed
549 during this work. Namely:
550
551 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
552
553 (b) Minimum length was not checked before the matching is started.
554
555 11. The value of capture_last that is passed to callouts was incorrect in some
556 cases when there was a capture on one path that was subsequently abandoned
557 after a backtrack. Also, the capture_last value is now reset after a
558 recursion, since all captures are also reset in this case.
559
560 12. The interpreter no longer returns the "too many substrings" error in the
561 case when an overflowing capture is in a branch that is subsequently
562 abandoned after a backtrack.
563
564 13. In the pathological case when an offset vector of size 2 is used, pcretest
565 now prints out the matched string after a yield of 0 or 1.
566
567 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
568 Only supported by the JIT compiler at the moment.
569
570 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
571
572 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
573 offset of the starting point of the matching process, provided the offsets
574 vector is large enough.
575
576 17. The \A escape now records a lookbehind value of 1, though its execution
577 does not actually inspect the previous character. This is to ensure that,
578 in partial multi-segment matching, at least one character from the old
579 segment is retained when a new segment is processed. Otherwise, if there
580 are no lookbehinds in the pattern, \A might match incorrectly at the start
581 of a new segment.
582
583 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
584
585 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
586 gives some modest performance improvement in 8-bit mode.
587
588 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
589 be expressed in certain programming languages using Universal Character
590 Names.
591
592 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
593 which points out that "non characters" are not "characters that may not
594 appear in Unicode strings" but rather "characters that are reserved for
595 internal use and have only local meaning".
596
597 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
598 there was a conditional group that depended on an assertion, if the
599 assertion was false, the callout that immediately followed the alternation
600 in the condition was skipped when pcre_exec() was used for matching.
601
602 23. Allow an explicit callout to be inserted before an assertion that is the
603 condition for a conditional group, for compatibility with automatic
604 callouts, which always insert a callout at this point.
605
606 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
607 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
608
609 25. (*PRUNE) is now supported by the JIT compiler.
610
611 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
612
613 27. Fix the case where there are two or more SKIPs with arguments that may be
614 ignored.
615
616 28. (*SKIP) is now supported by the JIT compiler.
617
618 29. (*THEN) is now supported by the JIT compiler.
619
620 30. Update RunTest with additional test selector options.
621
622 31. The way PCRE handles backtracking verbs has been changed in two ways.
623
624 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
625 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
626 backtracking. In some cases this makes it more Perl-compatible, but Perl's
627 rather obscure rules do not always do the same thing.
628
629 (2) Previously, backtracking verbs were confined within assertions. This is
630 no longer the case for positive assertions, except for (*ACCEPT). Again,
631 this sometimes improves Perl compatibility, and sometimes does not.
632
633 32. A number of tests that were in test 2 because Perl did things differently
634 have been moved to test 1, because either Perl or PCRE has changed, and
635 these tests are now compatible.
636
637 32. Backtracking control verbs are now handled in the same way in JIT and
638 interpreter.
639
640 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
641 contained a forward subroutine reference caused a compile error.
642
643 34. Auto-detect and optimize limited repetitions in JIT.
644
645 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
646 blocking (*UTF) etc.
647
648 36. In the interpreter, maximizing pattern repetitions for characters and
649 character types now use tail recursion, which reduces stack usage.
650
651 37. The value of the max lookbehind was not correctly preserved if a compiled
652 and saved regex was reloaded on a host of different endianness.
653
654 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
655 of the compiled pattern block, expand the flags field from 16 to 32 bits
656 because it was almost full.
657
658 39. Try madvise first before posix_madvise.
659
660 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
661 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
662 original change locked out all matches of empty strings. This has been
663 changed so that one match of an empty string per line is recognized.
664 Subsequent searches on the same line (for colouring or for --only-matching,
665 for example) do not recognize empty strings.
666
667 41. Applied a user patch to fix a number of spelling mistakes in comments.
668
669 42. Data lines longer than 65536 caused pcretest to crash.
670
671 43. Clarified the data type for length and startoffset arguments for pcre_exec
672 and pcre_dfa_exec in the function-specific man pages, where they were
673 explicitly stated to be in bytes, never having been updated. I also added
674 some clarification to the pcreapi man page.
675
676 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
677 a segmentation fault.
678
679
680 Version 8.32 30-November-2012
681 -----------------------------
682
683 1. Improved JIT compiler optimizations for first character search and single
684 character iterators.
685
686 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
687 Patch by Daniel Richard G.
688
689 3. Single character iterator optimizations in the JIT compiler.
690
691 4. Improved JIT compiler optimizations for character ranges.
692
693 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
694 Reported by Giuseppe D'Angelo.
695
696 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
697 of a line, was being set incorrectly in cases where .* appeared inside
698 atomic brackets at the start of a pattern, or where there was a subsequent
699 *PRUNE or *SKIP.
700
701 7. Improved instruction cache flush for POWER/PowerPC.
702 Patch by Daniel Richard G.
703
704 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
705 grep:
706
707 (a) There is now no limit to the number of patterns to be matched.
708
709 (b) An error is given if a pattern is too long.
710
711 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
712 are now supported.
713
714 (d) --exclude-from and --include-from (multiple use) have been added.
715
716 (e) Exclusions and inclusions now apply to all files and directories, not
717 just to those obtained from scanning a directory recursively.
718
719 (f) Multiple uses of -f and --file-list are now supported.
720
721 (g) In a Windows environment, the default for -d has been changed from
722 "read" (the GNU grep default) to "skip", because otherwise the presence
723 of a directory in the file list provokes an error.
724
725 (h) The documentation has been revised and clarified in places.
726
727 9. Improve the matching speed of capturing brackets.
728
729 10. Changed the meaning of \X so that it now matches a Unicode extended
730 grapheme cluster.
731
732 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
733 out POSIX threads when JIT support is configured.
734
735 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
736
737 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
738 structure could go wrong in environments where size_t is not the same size
739 as int.
740
741 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
742 set.
743
744 15. The EBCDIC support had decayed; later updates to the code had included
745 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
746 general tidy up of EBCDIC-related issues, and the documentation was also
747 not quite right. There is now a test that can be run on ASCII systems to
748 check some of the EBCDIC-related things (but is it not a full test).
749
750 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
751 in a small tidy to the code.
752
753 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
754
755 18. If the --only-matching (-o) option in pcregrep is specified multiple
756 times, each one causes appropriate output. For example, -o1 -o2 outputs the
757 substrings matched by the 1st and 2nd capturing parentheses. A separating
758 string can be specified by --om-separator (default empty).
759
760 19. Improving the first n character searches.
761
762 20. Turn case lists for horizontal and vertical white space into macros so that
763 they are defined only once.
764
765 21. This set of changes together give more compatible Unicode case-folding
766 behaviour for characters that have more than one other case when UCP
767 support is available.
768
769 (a) The Unicode property table now has offsets into a new table of sets of
770 three or more characters that are case-equivalent. The MultiStage2.py
771 script that generates these tables (the pcre_ucd.c file) now scans
772 CaseFolding.txt instead of UnicodeData.txt for character case
773 information.
774
775 (b) The code for adding characters or ranges of characters to a character
776 class has been abstracted into a generalized function that also handles
777 case-independence. In UTF-mode with UCP support, this uses the new data
778 to handle characters with more than one other case.
779
780 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
781 whose other case is greater than 256 are now correctly matched
782 caselessly. Previously, the high codepoint matched the low one, but not
783 vice versa.
784
785 (d) The processing of \h, \H, \v, and \ in character classes now makes use
786 of the new class addition function, using character lists defined as
787 macros alongside the case definitions of 20 above.
788
789 (e) Caseless back references now work with characters that have more than
790 one other case.
791
792 (f) General caseless matching of characters with more than one other case
793 is supported.
794
795 22. Unicode character properties were updated from Unicode 6.2.0
796
797 23. Improved CMake support under Windows. Patch by Daniel Richard G.
798
799 24. Add support for 32-bit character strings, and UTF-32
800
801 25. Major JIT compiler update (code refactoring and bugfixing).
802 Experimental Sparc 32 support is added.
803
804 26. Applied a modified version of Daniel Richard G's patch to create
805 pcre.h.generic and config.h.generic by "make" instead of in the
806 PrepareRelease script.
807
808 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
809 pcre_compile.c when checking for a zero character.
810
811 28. Introducing a native interface for JIT. Through this interface, the compiled
812 machine code can be directly executed. The purpose of this interface is to
813 provide fast pattern matching, so several sanity checks are not performed.
814 However, feature tests are still performed. The new interface provides
815 1.4x speedup compared to the old one.
816
817 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
818 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
819 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
820
821 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
822 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
823 the "old" RFC 2279). Instead, it ended up passing a negative length to
824 pcre_exec().
825
826 31. Add support for GCC's visibility feature to hide internal functions.
827
828 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
829 "unknown -C option" after outputting 0 or 1.
830
831 33. There is now support for generating a code coverage report for the test
832 suite in environments where gcc is the compiler and lcov is installed. This
833 is mainly for the benefit of the developers.
834
835 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
836 unaddressable using valgrind annotations, allowing valgrind to detect
837 invalid memory accesses. This is mainly for the benefit of the developers.
838
839 25. (*UTF) can now be used to start a pattern in any of the three libraries.
840
841 26. Give configure error if --enable-cpp but no C++ compiler found.
842
843
844 Version 8.31 06-July-2012
845 -------------------------
846
847 1. Fixing a wrong JIT test case and some compiler warnings.
848
849 2. Removed a bashism from the RunTest script.
850
851 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
852 to unsigned type, result still unsigned" that was given by an MS compiler
853 on encountering the code "-sizeof(xxx)".
854
855 4. Partial matching support is added to the JIT compiler.
856
857 5. Fixed several bugs concerned with partial matching of items that consist
858 of more than one character:
859
860 (a) /^(..)\1/ did not partially match "aba" because checking references was
861 done on an "all or nothing" basis. This also applied to repeated
862 references.
863
864 (b) \R did not give a hard partial match if \r was found at the end of the
865 subject.
866
867 (c) \X did not give a hard partial match after matching one or more
868 characters at the end of the subject.
869
870 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
871 a partial match for the string "\r".
872
873 (e) When newline was set to CRLF, the metacharacter "." did not recognize
874 a partial match for a CR character at the end of the subject string.
875
876 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
877 running pcretest, the text "(JIT)" added to the output whenever JIT is
878 actually used to run the match.
879
880 7. Individual JIT compile options can be set in pcretest by following -s+[+]
881 or /S+[+] with a digit between 1 and 7.
882
883 8. OP_NOT now supports any UTF character not just single-byte ones.
884
885 9. (*MARK) control verb is now supported by the JIT compiler.
886
887 10. The command "./RunTest list" lists the available tests without actually
888 running any of them. (Because I keep forgetting what they all are.)
889
890 11. Add PCRE_INFO_MAXLOOKBEHIND.
891
892 12. Applied a (slightly modified) user-supplied patch that improves performance
893 when the heap is used for recursion (compiled with --disable-stack-for-
894 recursion). Instead of malloc and free for each heap frame each time a
895 logical recursion happens, frames are retained on a chain and re-used where
896 possible. This sometimes gives as much as 30% improvement.
897
898 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
899 call.
900
901 14. As documented, (*COMMIT) is now confined to within a positive assertion.
902
903 15. It is now possible to link pcretest with libedit as an alternative to
904 libreadline.
905
906 16. (*COMMIT) control verb is now supported by the JIT compiler.
907
908 17. The Unicode data tables have been updated to Unicode 6.1.0.
909
910 18. Added --file-list option to pcregrep.
911
912 19. Added binary file support to pcregrep, including the -a, --binary-files,
913 -I, and --text options.
914
915 20. The madvise function is renamed for posix_madvise for QNX compatibility
916 reasons. Fixed by Giuseppe D'Angelo.
917
918 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
919
920 22. Changed the option for creating symbolic links for 16-bit man pages from
921 -s to -sf so that re-installing does not cause issues.
922
923 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
924
925 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
926 matches in certain environments (the workspace was not being correctly
927 retained). Also added to pcre_dfa_exec() a simple plausibility check on
928 some of the workspace data at the beginning of a restart.
929
930 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
931 was not doing so when it should - probably a typo introduced by SVN 528
932 (change 8.10/14).
933
934 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
935 \w+ when the character tables indicated that \x{c4} was a word character.
936 There were several related cases, all because the tests for doing a table
937 lookup were testing for characters less than 127 instead of 255.
938
939 27. If a pattern contains capturing parentheses that are not used in a match,
940 their slots in the ovector are set to -1. For those that are higher than
941 any matched groups, this happens at the end of processing. In the case when
942 there were back references that the ovector was too small to contain
943 (causing temporary malloc'd memory to be used during matching), and the
944 highest capturing number was not used, memory off the end of the ovector
945 was incorrectly being set to -1. (It was using the size of the temporary
946 memory instead of the true size.)
947
948 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
949 ovector size, it uses memory at the end of the block that it has got.
950
951 29. Check for an overlong MARK name and give an error at compile time. The
952 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
953
954 30. JIT compiler update.
955
956 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
957 Rill for the patch.
958
959 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
960
961 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
962
963 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
964 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
965 was enabled.
966
967 35. Improve JIT code generation for greedy plus quantifier.
968
969 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
970 1 to "aa" instead of to an empty string. The bug affected repeated groups
971 that could potentially match an empty string.
972
973 37. Optimizing single character iterators in JIT.
974
975 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
976 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
977 codepoints that are too big for the mode are faulted, and in a UTF mode,
978 disallowed codepoints are also faulted.
979
980 39. If PCRE was compiled with UTF support, in three places in the DFA
981 matcher there was code that should only have been obeyed in UTF mode, but
982 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
983 processing when bytes with values greater than 127 were present. In 16-bit
984 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
985 both cases the values are those that cannot be the first data item in a UTF
986 character. The three items that might have provoked this were recursions,
987 possessively repeated groups, and atomic groups.
988
989 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
990 and pcregrep, because some OS require shared objects to be explicitly
991 passed to ld, causing the link step to fail if they are not.
992
993 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
994 mode, patterns that started with \h* or \R* might be incorrectly matched.
995
996
997 Version 8.30 04-February-2012
998 -----------------------------
999
1000 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1001 name is defined in ctype.h.
1002
1003 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1004 only in quite long subpatterns.
1005
1006 3. Removed the function pcre_info(), which has been obsolete and deprecated
1007 since it was replaced by pcre_fullinfo() in February 2000.
1008
1009 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1010 match a (*MARK), and the match failed at the start of the subject, a
1011 reference to memory before the start of the subject could occur. This bug
1012 was introduced by fix 17 of release 8.21.
1013
1014 5. A reference to an unset group with zero minimum repetition was giving
1015 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1016 /(another)?(\1?)test/ matched against "hello world test". This bug was
1017 introduced in release 8.13.
1018
1019 6. Add support for 16-bit character strings (a large amount of work involving
1020 many changes and refactorings).
1021
1022 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1023 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1024 from a file.
1025
1026 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1027 rounding is not applied in this particular case).
1028
1029 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1030 if they appear, or are escaped, in patterns.
1031
1032 10. Get rid of a number of -Wunused-but-set-variable warnings.
1033
1034 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1035 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1036 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1037 also returns the mark "x". This bug applied to capturing parentheses,
1038 non-capturing parentheses, and atomic parentheses. It also applied to some
1039 assertions.
1040
1041 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1042 information out of configure.ac instead of relying on pcre.h.generic, which
1043 is not stored in the repository.
1044
1045 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1046 -lreadline.
1047
1048 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1049
1050 15. Applied Graycode's patch to put the top-level frame on the stack rather
1051 than the heap when not using the stack for recursion. This gives a
1052 performance improvement in many cases when recursion is not deep.
1053
1054 16. Experimental code added to "pcretest -C" to output the stack frame size.
1055
1056
1057 Version 8.21 12-Dec-2011
1058 ------------------------
1059
1060 1. Updating the JIT compiler.
1061
1062 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1063 are added as well.
1064
1065 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1066 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1067 calling _pcre_jit_exec. Some extra comments are added.
1068
1069 4. (*MARK) settings inside atomic groups that do not contain any capturing
1070 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1071 was introduced by change 18 for 8.20.
1072
1073 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1074 ECMA-262 standard.
1075
1076 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1077 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1078 This bug was probably introduced by change 9 of 8.13.
1079
1080 7. While fixing 6 above, I noticed that a number of other items were being
1081 incorrectly rejected as "not fixed length". This arose partly because newer
1082 opcodes had not been added to the fixed-length checking code. I have (a)
1083 corrected the bug and added tests for these items, and (b) arranged for an
1084 error to occur if an unknown opcode is encountered while checking for fixed
1085 length instead of just assuming "not fixed length". The items that were
1086 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1087 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1088 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1089
1090 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1091 being incorrectly compiled and would have given unpredicatble results.
1092
1093 9. A possessively repeated subpattern with minimum repeat count greater than
1094 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1095 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1096 the first (A) could occur when it should not.
1097
1098 10. Add a cast and remove a redundant test from the code.
1099
1100 11. JIT should use pcre_malloc/pcre_free for allocation.
1101
1102 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1103 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1104 is anything other than /usr, -L is still shown).
1105
1106 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1107
1108 14. Perl does not support \N without a following name in a [] class; PCRE now
1109 also gives an error.
1110
1111 15. If a forward reference was repeated with an upper limit of around 2000,
1112 it caused the error "internal error: overran compiling workspace". The
1113 maximum number of forward references (including repeats) was limited by the
1114 internal workspace, and dependent on the LINK_SIZE. The code has been
1115 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1116 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1117 of around 200,000 forward references. While doing this, I also speeded up
1118 the filling in of repeated forward references.
1119
1120 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1121 incorrectly expecting the subject to contain another "a" after the start.
1122
1123 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1124 in the match, the SKIP should be ignored. This was not happening; instead
1125 the SKIP was being treated as NOMATCH. For patterns such as
1126 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1127 tested.
1128
1129 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1130 now much more compatible with Perl, in particular in cases where the result
1131 is a non-match for a non-anchored pattern. For example, if
1132 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1133 "m", where previously it did not return a name. A side effect of this
1134 change is that for partial matches, the last encountered mark name is
1135 returned, as for non matches. A number of tests that were previously not
1136 Perl-compatible have been moved into the Perl-compatible test files. The
1137 refactoring has had the pleasing side effect of removing one argument from
1138 the match() function, thus reducing its stack requirements.
1139
1140 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1141 subsequent uses of /S (without +) incorrectly behaved like /S+.
1142
1143 21. Retrieve executable code size support for the JIT compiler and fixing
1144 some warnings.
1145
1146 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1147 not work when the shorter character appeared right at the end of the
1148 subject string.
1149
1150 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1151 systems.
1152
1153 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1154 output it when the /M option is used in pcretest.
1155
1156 25. The CheckMan script was not being included in the distribution. Also, added
1157 an explicit "perl" to run Perl scripts from the PrepareRelease script
1158 because this is reportedly needed in Windows.
1159
1160 26. If study data was being save in a file and studying had not found a set of
1161 "starts with" bytes for the pattern, the data written to the file (though
1162 never used) was taken from uninitialized memory and so caused valgrind to
1163 complain.
1164
1165 27. Updated RunTest.bat as provided by Sheri Pierce.
1166
1167 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1168
1169 29. Computation of memory usage for the table of capturing group names was
1170 giving an unnecessarily large value.
1171
1172
1173 Version 8.20 21-Oct-2011
1174 ------------------------
1175
1176 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1177 a POSIX class. After further experiments with Perl, which convinced me that
1178 Perl has bugs and confusions, a closing square bracket is no longer allowed
1179 in a POSIX name. This bug also affected patterns with classes that started
1180 with full stops.
1181
1182 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1183 captured substring, but while checking the failing first alternative,
1184 substring 1 is temporarily captured. If the output vector supplied to
1185 pcre_exec() was not big enough for this capture, the yield of the function
1186 was still zero ("insufficient space for captured substrings"). This cannot
1187 be totally fixed without adding another stack variable, which seems a lot
1188 of expense for a edge case. However, I have improved the situation in cases
1189 such as /(a)(b)x|abc/ matched against "abc", where the return code
1190 indicates that fewer than the maximum number of slots in the ovector have
1191 been set.
1192
1193 3. Related to (2) above: when there are more back references in a pattern than
1194 slots in the output vector, pcre_exec() uses temporary memory during
1195 matching, and copies in the captures as far as possible afterwards. It was
1196 using the entire output vector, but this conflicts with the specification
1197 that only 2/3 is used for passing back captured substrings. Now it uses
1198 only the first 2/3, for compatibility. This is, of course, another edge
1199 case.
1200
1201 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1202 main code base, and can be used by building with --enable-jit. When this is
1203 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1204 runtime --no-jit option is given.
1205
1206 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1207 ovector, the return from the function was zero, implying that there were
1208 other matches that did not fit. The correct "exactly full" value is now
1209 returned.
1210
1211 6. If a subpattern that was called recursively or as a subroutine contained
1212 (*PRUNE) or any other control that caused it to give a non-standard return,
1213 invalid errors such as "Error -26 (nested recursion at the same subject
1214 position)" or even infinite loops could occur.
1215
1216 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1217 computing the minimum length on reaching *ACCEPT, and so ended up with the
1218 wrong value of 1 rather than 0. Further investigation indicates that
1219 computing a minimum subject length in the presence of *ACCEPT is difficult
1220 (think back references, subroutine calls), and so I have changed the code
1221 so that no minimum is registered for a pattern that contains *ACCEPT.
1222
1223 8. If (*THEN) was present in the first (true) branch of a conditional group,
1224 it was not handled as intended. [But see 16 below.]
1225
1226 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1227 Sheri Pierce.
1228
1229 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1230 the first byte in a match must be "a".
1231
1232 11. Change 17 for 8.13 increased the recursion depth for patterns like
1233 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1234 pattern contains any instances of (*THEN). If it does not, the old
1235 optimizations are restored. It would be nice to do this on a per-group
1236 basis, but at the moment that is not feasible.
1237
1238 12. In some environments, the output of pcretest -C is CRLF terminated. This
1239 broke RunTest's code that checks for the link size. A single white space
1240 character after the value is now allowed for.
1241
1242 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1243 For "fr", it uses the Windows-specific input and output files.
1244
1245 14. If (*THEN) appeared in a group that was called recursively or as a
1246 subroutine, it did not work as intended. [But see next item.]
1247
1248 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1249 pattern fragments (but not containing any | characters). If A and B are
1250 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1251 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1252 D. In other words, Perl considers parentheses that do not contain any |
1253 characters to be part of a surrounding alternative, whereas PCRE was
1254 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1255 differently. PCRE now behaves in the same way as Perl, except in the case
1256 of subroutine/recursion calls such as (?1) which have in any case always
1257 been different (but PCRE had them first :-).
1258
1259 16. Related to 15 above: Perl does not treat the | in a conditional group as
1260 creating alternatives. Such a group is treated in the same way as an
1261 ordinary group without any | characters when processing (*THEN). PCRE has
1262 been changed to match Perl's behaviour.
1263
1264 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1265 RunGrepTest script failed.
1266
1267 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1268 inevitable for groups that contain captures, but it can lead to a lot of
1269 stack use in large patterns. The old behaviour has been restored for atomic
1270 groups that do not contain any capturing parentheses.
1271
1272 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1273 suppress the check for a minimum subject length at run time. (If it was
1274 given to pcre_exec() or pcre_dfa_exec() it did work.)
1275
1276 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1277 fail to work when decoding hex characters in data strings in EBCDIC
1278 environments.
1279
1280 21. It appears that in at least one Mac OS environment, the isxdigit() function
1281 is implemented as a macro that evaluates to its argument more than once,
1282 contravening the C 90 Standard (I haven't checked a later standard). There
1283 was an instance in pcretest which caused it to go wrong when processing
1284 \x{...} escapes in subject strings. The has been rewritten to avoid using
1285 things like p++ in the argument of isxdigit().
1286
1287
1288 Version 8.13 16-Aug-2011
1289 ------------------------
1290
1291 1. The Unicode data tables have been updated to Unicode 6.0.0.
1292
1293 2. Two minor typos in pcre_internal.h have been fixed.
1294
1295 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1296 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1297 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1298
1299 4. There were a number of related bugs in the code for matching backrefences
1300 caselessly in UTF-8 mode when codes for the characters concerned were
1301 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1302 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1303 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1304 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1305 2-byte code at the end of the subject (it thought there wasn't enough data
1306 left).
1307
1308 5. Comprehensive information about what went wrong is now returned by
1309 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1310 as the output vector has at least 2 elements. The offset of the start of
1311 the failing character and a reason code are placed in the vector.
1312
1313 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1314 now returned is for the first byte of the failing character, instead of the
1315 last byte inspected. This is an incompatible change, but I hope it is small
1316 enough not to be a problem. It makes the returned offset consistent with
1317 pcre_exec() and pcre_dfa_exec().
1318
1319 7. pcretest now gives a text phrase as well as the error number when
1320 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1321 failure, the offset and reason code are output.
1322
1323 8. When \R was used with a maximizing quantifier it failed to skip backwards
1324 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1325 back over a single character (\n). This seems wrong (because it treated the
1326 two characters as a single entity when going forwards), conflicts with the
1327 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1328 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1329 has been changed.
1330
1331 9. Some internal refactoring has changed the processing so that the handling
1332 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1333 time (the PCRE_DOTALL option was changed this way some time ago: version
1334 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1335 which was always a bit of a fudge. It also means that there is one less
1336 argument for the match() function, which reduces its stack requirements
1337 slightly. This change also fixes an incompatibility with Perl: the pattern
1338 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1339
1340 10. More internal refactoring has drastically reduced the number of recursive
1341 calls to match() for possessively repeated groups such as (abc)++ when
1342 using pcre_exec().
1343
1344 11. While implementing 10, a number of bugs in the handling of groups were
1345 discovered and fixed:
1346
1347 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1348 (a|)*(?1) gave a compile-time internal error.
1349 ((a|)+)+ did not notice that the outer group could match an empty string.
1350 (^a|^)+ was not marked as anchored.
1351 (.*a|.*)+ was not marked as matching at start or after a newline.
1352
1353 12. Yet more internal refactoring has removed another argument from the match()
1354 function. Special calls to this function are now indicated by setting a
1355 value in a variable in the "match data" data block.
1356
1357 13. Be more explicit in pcre_study() instead of relying on "default" for
1358 opcodes that mean there is no starting character; this means that when new
1359 ones are added and accidentally left out of pcre_study(), testing should
1360 pick them up.
1361
1362 14. The -s option of pcretest has been documented for ages as being an old
1363 synonym of -m (show memory usage). I have changed it to mean "force study
1364 for every regex", that is, assume /S for every regex. This is similar to -i
1365 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1366 using it. It makes it easier to run collections of tests with and without
1367 study enabled, and thereby test pcre_study() more easily. All the standard
1368 tests are now run with and without -s (but some patterns can be marked as
1369 "never study" - see 20 below).
1370
1371 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1372 restoration of the capturing data to the outer values was not happening
1373 correctly.
1374
1375 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1376 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1377 pattern had matched an empty string, and so incorrectly returned a no
1378 match.
1379
1380 17. There was optimizing code for the last branch of non-capturing parentheses,
1381 and also for the obeyed branch of a conditional subexpression, which used
1382 tail recursion to cut down on stack usage. Unfortunately, now that there is
1383 the possibility of (*THEN) occurring in these branches, tail recursion is
1384 no longer possible because the return has to be checked for (*THEN). These
1385 two optimizations have therefore been removed. [But see 8.20/11 above.]
1386
1387 18. If a pattern containing \R was studied, it was assumed that \R always
1388 matched two bytes, thus causing the minimum subject length to be
1389 incorrectly computed because \R can also match just one byte.
1390
1391 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1392 was incorrectly computed.
1393
1394 20. If /S is present twice on a test pattern in pcretest input, it now
1395 *disables* studying, thereby overriding the use of -s on the command line
1396 (see 14 above). This is necessary for one or two tests to keep the output
1397 identical in both cases.
1398
1399 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1400 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1401
1402 22. When an atomic group that contained a capturing parenthesis was
1403 successfully matched, but the branch in which it appeared failed, the
1404 capturing was not being forgotten if a higher numbered group was later
1405 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1406 group 1 to "a", when in fact it should be unset. This applied to multi-
1407 branched capturing and non-capturing groups, repeated or not, and also to
1408 positive assertions (capturing in negative assertions does not happen
1409 in PCRE) and also to nested atomic groups.
1410
1411 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1412 subject after a captured substring, to make it easier to tell which of a
1413 number of identical substrings has been captured.
1414
1415 24. The way atomic groups are processed by pcre_exec() has been changed so that
1416 if they are repeated, backtracking one repetition now resets captured
1417 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1418 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1419 "aaa". Previously, it would have been "a". As part of this code
1420 refactoring, the way recursive calls are handled has also been changed.
1421
1422 25. If an assertion condition captured any substrings, they were not passed
1423 back unless some other capturing happened later. For example, if
1424 (?(?=(a))a) was matched against "a", no capturing was returned.
1425
1426 26. When studying a pattern that contained subroutine calls or assertions,
1427 the code for finding the minimum length of a possible match was handling
1428 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1429 group 1 called group 2 while simultaneously a separate group 2 called group
1430 1). A stack overflow occurred in this case. I have fixed this by limiting
1431 the recursion depth to 10.
1432
1433 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1434 Fortmann. This supports explicit test numbers on the command line, and has
1435 argument validation and error reporting.
1436
1437 28. An instance of \X with an unlimited repeat could fail if at any point the
1438 first character it looked at was a mark character.
1439
1440 29. Some minor code refactoring concerning Unicode properties and scripts
1441 should reduce the stack requirement of match() slightly.
1442
1443 30. Added the '=' option to pcretest to check the setting of unused capturing
1444 slots at the end of the pattern, which are documented as being -1, but are
1445 not included in the return count.
1446
1447 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1448 compiled something random. Now it gives a compile-time error (as does
1449 Perl).
1450
1451 32. A *MARK encountered during the processing of a positive assertion is now
1452 recorded and passed back (compatible with Perl).
1453
1454 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1455 had alternative anchored branches, the search for a second match in a line
1456 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1457 matched the line "0102" twice. The same bug affected patterns that started
1458 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1459 twice.
1460
1461 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1462 does, and because of capturing effects, quantifying parenthesized
1463 assertions may at times be useful. Quantifiers are now allowed for
1464 parenthesized assertions.
1465
1466 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1467
1468 36. \g was being checked for fancy things in a character class, when it should
1469 just be a literal "g".
1470
1471 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1472 appearance of a nested POSIX class supersedes an apparent external class.
1473 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1474 unescaped square brackets may also appear as part of class names. For
1475 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1476 more like Perl. (But see 8.20/1 above.)
1477
1478 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1479 was because it thought it was \N{name}, which is not supported).
1480
1481 39. Add minix to OS list not supporting the -S option in pcretest.
1482
1483 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1484 cannot analyze patterns in sufficient detail to catch mutual recursions
1485 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1486 subgroup is called recursively as a subpattern for a second time at the
1487 same position in the subject string. In previous releases this might have
1488 been caught by the recursion limit, or it might have run out of stack.
1489
1490 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1491 happen only once. PCRE was, however incorrectly giving a compile time error
1492 "recursive call could loop indefinitely" because it cannot analyze the
1493 pattern in sufficient detail. The compile time test no longer happens when
1494 PCRE is compiling a conditional subpattern, but actual runaway loops are
1495 now caught at runtime (see 40 above).
1496
1497 42. It seems that Perl allows any characters other than a closing parenthesis
1498 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1499 has been changed to be the same.
1500
1501 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1502 as not to get warnings when autogen.sh is called. Also changed
1503 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1504
1505 44. To help people who use pcregrep to scan files containing exceedingly long
1506 lines, the following changes have been made:
1507
1508 (a) The default value of the buffer size parameter has been increased from
1509 8K to 20K. (The actual buffer used is three times this size.)
1510
1511 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1512 PCRE is built.
1513
1514 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1515 to be set at run time.
1516
1517 (d) Numerical values in pcregrep options can be followed by K or M, for
1518 example --buffer-size=50K.
1519
1520 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1521 given and the return code is set to 2.
1522
1523 45. Add a pointer to the latest mark to the callout data block.
1524
1525 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1526 partial match of an empty string instead of no match. This was specific to
1527 the use of ".".
1528
1529 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1530 complete match instead of a partial match. This bug was dependent on both
1531 the PCRE_UTF8 and PCRE_DOTALL options being set.
1532
1533 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1534 starting byte set, because \b was not being ignored.
1535
1536
1537 Version 8.12 15-Jan-2011
1538 ------------------------
1539
1540 1. Fixed some typos in the markup of the man pages, and wrote a script that
1541 checks for such things as part of the documentation building process.
1542
1543 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1544 --match-limit and --recursion-limit options (added for 8.11). In
1545 particular, this made one of the standard tests fail. (The integer value
1546 went into the wrong half of a long int.)
1547
1548 3. If the --colour option was given to pcregrep with -v (invert match), it
1549 did strange things, either producing crazy output, or crashing. It should,
1550 of course, ignore a request for colour when reporting lines that do not
1551 match.
1552
1553 4. Another pcregrep bug caused similar problems if --colour was specified with
1554 -M (multiline) and the pattern match finished with a line ending.
1555
1556 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1557 matched in multiline mode, the following line was shown as part of the
1558 match. This seems wrong, so I have changed it.
1559
1560 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1561 the check for further matches in the same line (so they could be coloured)
1562 to overrun the end of the current line. If another match was found, it was
1563 incorrectly shown (and then shown again when found in the next line).
1564
1565 7. If pcregrep was compiled under Windows, there was a reference to the
1566 function pcregrep_exit() before it was defined. I am assuming this was
1567 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1568 reported by a user. I've moved the definition above the reference.
1569
1570
1571 Version 8.11 10-Dec-2010
1572 ------------------------
1573
1574 1. (*THEN) was not working properly if there were untried alternatives prior
1575 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1576 backtracked to try for "b" instead of moving to the next alternative branch
1577 at the same level (in this case, to look for "c"). The Perl documentation
1578 is clear that when (*THEN) is backtracked onto, it goes to the "next
1579 alternative in the innermost enclosing group".
1580
1581 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1582 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1583 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1584 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1585 (*THEN).
1586
1587 3. If \s appeared in a character class, it removed the VT character from
1588 the class, even if it had been included by some previous item, for example
1589 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1590 of \s, but is part of the POSIX "space" class.)
1591
1592 4. A partial match never returns an empty string (because you can always
1593 match an empty string at the end of the subject); however the checking for
1594 an empty string was starting at the "start of match" point. This has been
1595 changed to the "earliest inspected character" point, because the returned
1596 data for a partial match starts at this character. This means that, for
1597 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1598 (previously it gave "no match").
1599
1600 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1601 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1602 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1603 has an implication that the given string is incomplete (because a partial
1604 match is preferred over a full match). For this reason, these items now
1605 give a partial match in this situation. [Aside: previously, the one case
1606 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1607 match rather than a full match, which was wrong by the old rules, but is
1608 now correct.]
1609
1610 6. There was a bug in the handling of #-introduced comments, recognized when
1611 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1612 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1613 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1614 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1615 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1616 places in pcre_compile().
1617
1618 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1619 comments when looking ahead for named forward references to subpatterns,
1620 the only newline sequence it recognized was NL. It now handles newlines
1621 according to the set newline convention.
1622
1623 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1624 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1625 cater for a lack of strerror(). These oversights have been fixed.
1626
1627 9. Added --match-limit and --recursion-limit to pcregrep.
1628
1629 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1630
1631 11. When the -o option was used, pcregrep was setting a return code of 1, even
1632 when matches were found, and --line-buffered was not being honoured.
1633
1634 12. Added an optional parentheses number to the -o and --only-matching options
1635 of pcregrep.
1636
1637 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1638 can match an empty string. The code to do it in pcretest and pcredemo
1639 needed fixing:
1640
1641 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1642 only one byte after an empty string match just before CRLF (this case
1643 just got forgotten; "any" and "anycrlf" were OK).
1644
1645 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1646 mode when an empty string match preceded an ASCII character followed by
1647 a non-ASCII character. (The code for advancing by one character rather
1648 than one byte was nonsense.)
1649
1650 (c) The pcredemo.c sample program did not have any code at all to handle
1651 the cases when CRLF is a valid newline sequence.
1652
1653 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1654 as a starting offset was within the subject string. There is now a new
1655 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1656 negative or greater than the length of the string. In order to test this,
1657 pcretest is extended to allow the setting of negative starting offsets.
1658
1659 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1660 starting offset points to the beginning of a UTF-8 character was
1661 unnecessarily clumsy. I tidied it up.
1662
1663 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1664 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1665
1666 17. Nobody had reported that the --include_dir option, which was added in
1667 release 7.7 should have been called --include-dir (hyphen, not underscore)
1668 for compatibility with GNU grep. I have changed it to --include-dir, but
1669 left --include_dir as an undocumented synonym, and the same for
1670 --exclude-dir, though that is not available in GNU grep, at least as of
1671 release 2.5.4.
1672
1673 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1674 characters from a string of bytes) have been redefined so as not to use
1675 loops, in order to improve performance in some environments. At the same
1676 time, I abstracted some of the common code into auxiliary macros to save
1677 repetition (this should not affect the compiled code).
1678
1679 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1680 compile-time error is now given if \c is not followed by an ASCII
1681 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1682 different, and any byte value is allowed.)
1683
1684 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1685 START_OPTIMIZE option, which is now allowed at compile time - but just
1686 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1687 to pcregrep and other applications that have no direct access to PCRE
1688 options. The new /Y option in pcretest sets this option when calling
1689 pcre_compile().
1690
1691 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1692 back references. Groups containing recursive back references were forced to
1693 be atomic by that change, but in the case of named groups, the amount of
1694 memory required was incorrectly computed, leading to "Failed: internal
1695 error: code overflow". This has been fixed.
1696
1697 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1698 pcretest.c, to avoid build problems in some Borland environments.
1699
1700
1701 Version 8.10 25-Jun-2010
1702 ------------------------
1703
1704 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1705 THEN.
1706
1707 2. (*ACCEPT) was not working when inside an atomic group.
1708
1709 3. Inside a character class, \B is treated as a literal by default, but
1710 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1711 causes the error). The code is unchanged, but I tidied the documentation.
1712
1713 4. Inside a character class, PCRE always treated \R and \X as literals,
1714 whereas Perl faults them if its -w option is set. I have changed PCRE so
1715 that it faults them when PCRE_EXTRA is set.
1716
1717 5. Added support for \N, which always matches any character other than
1718 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1719
1720 6. When compiling pcregrep with newer versions of gcc which may have
1721 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1722 declared with attribute warn_unused_result" were given. Just casting the
1723 result to (void) does not stop the warnings; a more elaborate fudge is
1724 needed. I've used a macro to implement this.
1725
1726 7. Minor change to pcretest.c to avoid a compiler warning.
1727
1728 8. Added four artifical Unicode properties to help with an option to make
1729 \s etc use properties (see next item). The new properties are: Xan
1730 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1731
1732 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1733 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1734 this option. Modified pcretest to add /W to test this facility. Added
1735 REG_UCP to make it available via the POSIX interface.
1736
1737 10. Added --line-buffered to pcregrep.
1738
1739 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1740 studied, and the match started with a letter with a code point greater than
1741 127 whose first byte was different to the first byte of the other case of
1742 the letter, the other case of this starting letter was not recognized
1743 (#976).
1744
1745 12. If a pattern that was studied started with a repeated Unicode property
1746 test, for example, \p{Nd}+, there was the theoretical possibility of
1747 setting up an incorrect bitmap of starting bytes, but fortunately it could
1748 not have actually happened in practice until change 8 above was made (it
1749 added property types that matched character-matching opcodes).
1750
1751 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1752 possible starting bytes for non-anchored patterns.
1753
1754 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1755 \R, and also a number of cases that involve Unicode properties, both
1756 explicit and implicit when PCRE_UCP is set.
1757
1758 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1759 input, it could crash or give wrong results if characters with values
1760 greater than 0xc0 were present in the subject string. (Detail: it assumed
1761 UTF-8 input when processing these items.)
1762
1763 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1764 size_t is 64-bit (#991).
1765
1766 17. Added a check for running out of memory when PCRE is compiled with
1767 --disable-stack-for-recursion (#990).
1768
1769 18. If the last data line in a file for pcretest does not have a newline on
1770 the end, a newline was missing in the output.
1771
1772 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1773 less than 128) in its various bitmaps. However, there is a facility for
1774 generating tables according to the current locale when PCRE is compiled. It
1775 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1776 characters, are recognized by isspace() and therefore were getting set in
1777 these tables, and indeed these tables seem to approximate to ISO 8859. This
1778 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1779 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1780 which of course cannot start UTF-8 characters. I have changed the code so
1781 that only real ASCII characters (less than 128) and the correct starting
1782 bytes for UTF-8 encodings are set for characters greater than 127 when in
1783 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1784 altogether.)
1785
1786 20. Added the /T option to pcretest so as to be able to run tests with non-
1787 standard character tables, thus making it possible to include the tests
1788 used for 19 above in the standard set of tests.
1789
1790 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1791 reference to a subpattern the other side of a comment that contains an
1792 opening parenthesis caused either an internal compiling error, or a
1793 reference to the wrong subpattern.
1794
1795
1796 Version 8.02 19-Mar-2010
1797 ------------------------
1798
1799 1. The Unicode data tables have been updated to Unicode 5.2.0.
1800
1801 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1802 configured.
1803
1804 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1805 original author of that file, following a query about its status.
1806
1807 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1808 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1809
1810 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1811 quantifier applied to a forward-referencing subroutine call, could compile
1812 incorrect code or give the error "internal error: previously-checked
1813 referenced subpattern not found".
1814
1815 6. Both MS Visual Studio and Symbian OS have problems with initializing
1816 variables to point to external functions. For these systems, therefore,
1817 pcre_malloc etc. are now initialized to local functions that call the
1818 relevant global functions.
1819
1820 7. There were two entries missing in the vectors called coptable and poptable
1821 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1822 I've fixed the data, and added a kludgy way of testing at compile time that
1823 the lengths are correct (equal to the number of opcodes).
1824
1825 8. Following on from 7, I added a similar kludge to check the length of the
1826 eint vector in pcreposix.c.
1827
1828 9. Error texts for pcre_compile() are held as one long string to avoid too
1829 much relocation at load time. To find a text, the string is searched,
1830 counting zeros. There was no check for running off the end of the string,
1831 which could happen if a new error number was added without updating the
1832 string.
1833
1834 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1835
1836 11. \K was not working if it appeared in an atomic group or in a group that
1837 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1838 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1839 the assertion is positive, but not if it is negative.
1840
1841 12. Change 11 fortuitously reduced the size of the stack frame used in the
1842 "match()" function of pcre_exec.c by one pointer. Forthcoming
1843 implementation of support for (*MARK) will need an extra pointer on the
1844 stack; I have reserved it now, so that the stack frame size does not
1845 decrease.
1846
1847 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1848 item in branch that calls a recursion is a subroutine call - as in the
1849 second branch in the above example - was incorrectly given the compile-
1850 time error "recursive call could loop indefinitely" because pcre_compile()
1851 was not correctly checking the subroutine for matching a non-empty string.
1852
1853 14. The checks for overrunning compiling workspace could trigger after an
1854 overrun had occurred. This is a "should never occur" error, but it can be
1855 triggered by pathological patterns such as hundreds of nested parentheses.
1856 The checks now trigger 100 bytes before the end of the workspace.
1857
1858 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1859
1860
1861 Version 8.01 19-Jan-2010
1862 ------------------------
1863
1864 1. If a pattern contained a conditional subpattern with only one branch (in
1865 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1866 computed the wrong minimum data length (which is of course zero for such
1867 subpatterns). This could cause incorrect "no match" results.
1868
1869 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1870 the pattern is reset in the first branch, pcre_compile() failed with
1871 "internal error: code overflow at offset...". This happened only when
1872 the reset was to the original external option setting. (An optimization
1873 abstracts leading options settings into an external setting, which was the
1874 cause of this.)
1875
1876 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1877 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1878 assertion pattern did not match (meaning that the assertion was true), it
1879 was incorrectly treated as false if the SKIP had been reached during the
1880 matching. This also applied to assertions used as conditions.
1881
1882 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1883 assertion subpattern, including such a pattern used as a condition,
1884 unpredictable results occurred, instead of the error return
1885 PCRE_ERROR_DFA_UITEM.
1886
1887 5. The C++ GlobalReplace function was not working like Perl for the special
1888 situation when an empty string is matched. It now does the fancy magic
1889 stuff that is necessary.
1890
1891 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1892 removed. (These were left over from very, very early versions of PCRE.)
1893
1894 7. Some cosmetic changes to the code to make life easier when compiling it
1895 as part of something else:
1896
1897 (a) Change DEBUG to PCRE_DEBUG.
1898
1899 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1900 called "current" as "current_branch", to prevent a collision with the
1901 Linux macro when compiled as a kernel module.
1902
1903 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1904 prevent a collision with the Linux macro when compiled as a kernel
1905 module.
1906
1907 8. In pcre_compile() there are some checks for integer overflows that used to
1908 cast potentially large values to (double). This has been changed to that
1909 when building, a check for int64_t is made, and if it is found, it is used
1910 instead, thus avoiding the use of floating point arithmetic. (There is no
1911 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1912 double.
1913
1914 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1915 2005 (difference between two addresses compared to an unsigned value).
1916
1917 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1918 custom one, because of the following reported problem in Windows:
1919
1920 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1921 under Win32.
1922 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1923 therefore missing the function definition.
1924 - The compiler thus generates a "C" signature for the test function.
1925 - The linker fails to find the "C" function.
1926 - PCRE fails to configure if asked to do so against libbz2.
1927
1928 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1929 messages were output:
1930
1931 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1932 rerunning libtoolize, to keep the correct libtool macros in-tree.
1933 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1934
1935 I have done both of these things.
1936
1937 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1938 most of the time, it *can* run out if it is given a pattern that contains a
1939 runaway infinite recursion. I updated the discussion in the pcrestack man
1940 page.
1941
1942 13. Now that we have gone to the x.xx style of version numbers, the minor
1943 version may start with zero. Using 08 or 09 is a bad idea because users
1944 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1945 interpreted as invalid octal numbers. I've updated the previous comment in
1946 configure.ac, and also added a check that gives an error if 08 or 09 are
1947 used.
1948
1949 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1950 causing partial matching to fail when the end of the subject matched \W
1951 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1952
1953 15. There were some discrepancies between the declarations in pcre_internal.h
1954 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1955 their definitions. The declarations used "const uschar *" and the
1956 definitions used USPTR. Even though USPTR is normally defined as "const
1957 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1958 reported that: "This difference in casting confuses some C++ compilers, for
1959 example, SunCC recognizes above declarations as different functions and
1960 generates broken code for hbpcre." I have changed the declarations to use
1961 USPTR.
1962
1963 16. GNU libtool is named differently on some systems. The autogen.sh script now
1964 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1965 (FreeBSD).
1966
1967 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1968 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1969 comment: "Figure out how to create a longlong from a string: strtoll and
1970 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1971 instance, but it only takes 2 args instead of 3!"
1972
1973 18. A subtle bug concerned with back references has been fixed by a change of
1974 specification, with a corresponding code fix. A pattern such as
1975 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1976 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1977 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1978 same bug. Such groups have to be quantified to be useful, or contained
1979 inside another quantified group. (If there's no repetition, the reference
1980 can never match.) The problem arises because, having left the group and
1981 moved on to the rest of the pattern, a later failure that backtracks into
1982 the group uses the captured value from the final iteration of the group
1983 rather than the correct earlier one. I have fixed this in PCRE by forcing
1984 any group that contains a reference to itself to be an atomic group; that
1985 is, there cannot be any backtracking into it once it has completed. This is
1986 similar to recursive and subroutine calls.
1987
1988
1989 Version 8.00 19-Oct-09
1990 ----------------------
1991
1992 1. The table for translating pcre_compile() error codes into POSIX error codes
1993 was out-of-date, and there was no check on the pcre_compile() error code
1994 being within the table. This could lead to an OK return being given in
1995 error.
1996
1997 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1998 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1999 in a Windows environment.
2000
2001 3. The pcregrep --count option prints the count for each file even when it is
2002 zero, as does GNU grep. However, pcregrep was also printing all files when
2003 --files-with-matches was added. Now, when both options are given, it prints
2004 counts only for those files that have at least one match. (GNU grep just
2005 prints the file name in this circumstance, but including the count seems
2006 more useful - otherwise, why use --count?) Also ensured that the
2007 combination -clh just lists non-zero counts, with no names.
2008
2009 4. The long form of the pcregrep -F option was incorrectly implemented as
2010 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2011 but it seems right to fix it, and I didn't think it was worth preserving
2012 the old behaviour.
2013
2014 5. The command line items --regex=pattern and --regexp=pattern were not
2015 recognized by pcregrep, which required --regex pattern or --regexp pattern
2016 (with a space rather than an '='). The man page documented the '=' forms,
2017 which are compatible with GNU grep; these now work.
2018
2019 6. No libpcreposix.pc file was created for pkg-config; there was just
2020 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2021
2022 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2023 when UCP support is not needed, by modifying the Python script that
2024 generates it from Unicode data files. This should not matter if the module
2025 is correctly used as a library, but I received one complaint about 50K of
2026 unwanted data. My guess is that the person linked everything into his
2027 program rather than using a library. Anyway, it does no harm.
2028
2029 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2030 was a minimum greater than 1 for a wide character in a possessive
2031 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2032 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2033 character. Chaos in the form of incorrect output or a compiling loop could
2034 result.
2035
2036 9. The restrictions on what a pattern can contain when partial matching is
2037 requested for pcre_exec() have been removed. All patterns can now be
2038 partially matched by this function. In addition, if there are at least two
2039 slots in the offset vector, the offset of the earliest inspected character
2040 for the match and the offset of the end of the subject are set in them when
2041 PCRE_ERROR_PARTIAL is returned.
2042
2043 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2044 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2045 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2046 and may be more useful for multi-segment matching.
2047
2048 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2049 used to be given if ever the end of the subject was reached; now it is
2050 given only if matching could not proceed because another character was
2051 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2052 string "Z", which now yields "no match" instead of "partial match". In the
2053 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2054 final character ended with (*FAIL).
2055
2056 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2057 if the pattern had a "must contain" character that was already found in the
2058 earlier partial match, unless partial matching was again requested. For
2059 example, with the pattern /dog.(body)?/, the "must contain" character is
2060 "g". If the first part-match was for the string "dog", restarting with
2061 "sbody" failed. This bug has been fixed.
2062
2063 13. The string returned by pcre_dfa_exec() after a partial match has been
2064 changed so that it starts at the first inspected character rather than the
2065 first character of the match. This makes a difference only if the pattern
2066 starts with a lookbehind assertion or \b or \B (\K is not supported by
2067 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2068 matching functions compatible, and I think it's the right thing to do.
2069
2070 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2071 so that the demonstration program is easily available in environments where
2072 PCRE has not been installed from source.
2073
2074 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2075 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2076 library.
2077
2078 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2079 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2080 is not the first non-POSIX option to be added. Clearly some people find
2081 these options useful.
2082
2083 17. If a caller to the POSIX matching function regexec() passes a non-zero
2084 value for nmatch with a NULL value for pmatch, the value of
2085 nmatch is forced to zero.
2086
2087 18. RunGrepTest did not have a test for the availability of the -u option of
2088 the diff command, as RunTest does. It now checks in the same way as
2089 RunTest, and also checks for the -b option.
2090
2091 19. If an odd number of negated classes containing just a single character
2092 interposed, within parentheses, between a forward reference to a named
2093 subpattern and the definition of the subpattern, compilation crashed with
2094 an internal error, complaining that it could not find the referenced
2095 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2096 [The bug was that it was starting one character too far in when skipping
2097 over the character class, thus treating the ] as data rather than
2098 terminating the class. This meant it could skip too much.]
2099
2100 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2101 /g option in pcretest when the pattern contains \K, which makes it possible
2102 to have an empty string match not at the start, even when the pattern is
2103 anchored. Updated pcretest and pcredemo to use this option.
2104
2105 21. If the maximum number of capturing subpatterns in a recursion was greater
2106 than the maximum at the outer level, the higher number was returned, but
2107 with unset values at the outer level. The correct (outer level) value is
2108 now given.
2109
2110 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2111 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2112 make it do so. The string so far is captured, making this feature
2113 compatible with Perl.
2114
2115 23. The tests have been re-organized, adding tests 11 and 12, to make it
2116 possible to check the Perl 5.10 features against Perl 5.10.
2117
2118 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2119 pattern matches a fixed length string. PCRE did not allow this; now it
2120 does. Neither allows recursion.
2121
2122 25. I finally figured out how to implement a request to provide the minimum
2123 length of subject string that was needed in order to match a given pattern.
2124 (It was back references and recursion that I had previously got hung up
2125 on.) This code has now been added to pcre_study(); it finds a lower bound
2126 to the length of subject needed. It is not necessarily the greatest lower
2127 bound, but using it to avoid searching strings that are too short does give
2128 some useful speed-ups. The value is available to calling programs via
2129 pcre_fullinfo().
2130
2131 26. While implementing 25, I discovered to my embarrassment that pcretest had
2132 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2133 study optimizations had never been tested with that matching function.
2134 Oops. What is worse, even when it was passed study data, there was a bug in
2135 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2136 were also very few tests of studied patterns with pcre_dfa_exec().
2137
2138 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2139 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2140 on the other side of the coin, they are no longer allowed to have different
2141 names, because these cannot be distinguished in PCRE, and this has caused
2142 confusion. (This is a difference from Perl.)
2143
2144 28. When duplicate subpattern names are present (necessarily with different
2145 numbers, as required by 27 above), and a test is made by name in a
2146 conditional pattern, either for a subpattern having been matched, or for
2147 recursion in such a pattern, all the associated numbered subpatterns are
2148 tested, and the overall condition is true if the condition is true for any
2149 one of them. This is the way Perl works, and is also more like the way
2150 testing by number works.
2151
2152
2153 Version 7.9 11-Apr-09
2154 ---------------------
2155
2156 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2157 (pcretest), all targets were linked against these libraries. This included
2158 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2159 libraries. This caused unwanted dependencies to be created. This problem
2160 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2161 pcretest is linked with readline.
2162
2163 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2164 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2165 moved outside it again, because FALSE and TRUE are already defined in AIX,
2166 but BOOL is not.
2167
2168 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2169 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2170
2171 4. The pcregrep documentation said spaces were inserted as well as colons (or
2172 hyphens) following file names and line numbers when outputting matching
2173 lines. This is not true; no spaces are inserted. I have also clarified the
2174 wording for the --colour (or --color) option.
2175
2176 5. In pcregrep, when --colour was used with -o, the list of matching strings
2177 was not coloured; this is different to GNU grep, so I have changed it to be
2178 the same.
2179
2180 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2181 each matching line was coloured. Now it goes on to look for further matches
2182 of any of the test patterns, which is the same behaviour as GNU grep.
2183
2184 7. A pattern that could match an empty string could cause pcregrep to loop; it
2185 doesn't make sense to accept an empty string match in pcregrep, so I have
2186 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2187 seems to be how GNU grep behaves. [But see later change 40 for release
2188 8.33.]
2189
2190 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2191 start or after a newline", because the conditional assertion was not being
2192 correctly handled. The rule now is that both the assertion and what follows
2193 in the first alternative must satisfy the test.
2194
2195 9. If auto-callout was enabled in a pattern with a conditional group whose
2196 condition was an assertion, PCRE could crash during matching, both with
2197 pcre_exec() and pcre_dfa_exec().
2198
2199 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2200 used for matching.
2201
2202 11. Unicode property support in character classes was not working for
2203 characters (bytes) greater than 127 when not in UTF-8 mode.
2204
2205 12. Added the -M command line option to pcretest.
2206
2207 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2208
2209 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2210
2211 16. Added comments and documentation about mis-use of no_arg in the C++
2212 wrapper.
2213
2214 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2215 from Martin Jerabek that uses macro names for all relevant character and
2216 string constants.
2217
2218 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2219 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2220 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2221 these, but not everybody uses configure.
2222
2223 19. A conditional group that had only one branch was not being correctly
2224 recognized as an item that could match an empty string. This meant that an
2225 enclosing group might also not be so recognized, causing infinite looping
2226 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2227 with the subject "ab", where knowledge that the repeated group can match
2228 nothing is needed in order to break the loop.
2229
2230 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2231 exec(), but without supplying a callout function, matching went wrong.
2232
2233 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2234 leak if the size of the offset vector was greater than 30. When the vector
2235 is smaller, the saved offsets during recursion go onto a local stack
2236 vector, but for larger vectors malloc() is used. It was failing to free
2237 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2238 error, in fact).
2239
2240 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2241 heapframe that is used only when UTF-8 support is enabled. This caused no
2242 problem, but was untidy.
2243
2244 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2245 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2246 included within another project.
2247
2248 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2249 slightly modified by me:
2250
2251 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2252 not building pcregrep.
2253
2254 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2255 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2256
2257 25. Forward references, both numeric and by name, in patterns that made use of
2258 duplicate group numbers, could behave incorrectly or give incorrect errors,
2259 because when scanning forward to find the reference group, PCRE was not
2260 taking into account the duplicate group numbers. A pattern such as
2261 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2262
2263 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2264 the feature of a custom pointer more persuasive (as requested by a user).
2265
2266 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2267 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2268 pre-defined.
2269
2270 28. Added support for (*UTF8) at the start of a pattern.
2271
2272 29. Arrange for flags added by the "release type" setting in CMake to be shown
2273 in the configuration summary.
2274
2275
2276 Version 7.8 05-Sep-08
2277 ---------------------
2278
2279 1. Replaced UCP searching code with optimized version as implemented for Ad
2280 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2281 stage table and inline lookup instead of a function, giving speed ups of 2
2282 to 5 times on some simple patterns that I tested. Permission was given to
2283 distribute the MultiStage2.py script that generates the tables (it's not in
2284 the tarball, but is in the Subversion repository).
2285
2286 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2287 scripts.
2288
2289 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2290 a group with a zero qualifier. The result of the study could be incorrect,
2291 or the function might crash, depending on the pattern.
2292
2293 4. Caseless matching was not working for non-ASCII characters in back
2294 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2295 It now works when Unicode Property Support is available.
2296
2297 5. In pcretest, an escape such as \x{de} in the data was always generating
2298 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2299 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2300 truncation.
2301
2302 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2303
2304 7. Added two (int) casts to pcregrep when printing the difference of two
2305 pointers, in case they are 64-bit values.
2306
2307 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2308 test 2 if it fails.
2309
2310 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2311 and a #define of that name to empty if it is not externally set. This is to
2312 allow users of MSVC to set it if necessary.
2313
2314 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2315 the convenience functions in the pcre_get.c source file.
2316
2317 11. An option change at the start of a pattern that had top-level alternatives
2318 could cause overwriting and/or a crash. This command provoked a crash in
2319 some environments:
2320
2321 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2322
2323 This potential security problem was recorded as CVE-2008-2371.
2324
2325 12. For a pattern where the match had to start at the beginning or immediately
2326 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2327 pcre_dfa_exec() could read past the end of the passed subject if there was
2328 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2329 pcretest so that it places the subject at the end of its malloc-ed buffer.
2330
2331 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2332 exec() might read past the end of the data buffer in UTF-8 mode.
2333
2334 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2335 the data contained the byte 0x85 as part of a UTF-8 character within its
2336 first line. This applied both to normal and DFA matching.
2337
2338 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2339 /^[^d]*?$/8 failed to match "abc".
2340
2341 16. Added a missing copyright notice to pcrecpp_internal.h.
2342
2343 17. Make it more clear in the documentation that values returned from
2344 pcre_exec() in ovector are byte offsets, not character counts.
2345
2346 18. Tidied a few places to stop certain compilers from issuing warnings.
2347
2348 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2349 supplied by Stefan Weber. I made a further small update for 7.8 because
2350 there is a change of source arrangements: the pcre_searchfuncs.c module is
2351 replaced by pcre_ucd.c.
2352
2353
2354 Version 7.7 07-May-08
2355 ---------------------
2356
2357 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2358 a string to a long long, pretend we don't even have a long long." This is
2359 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2360
2361 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2362 pre-7.6 versions, which defined a global no_arg variable instead of putting
2363 it in the RE class. (See also #8 below.)
2364
2365 3. Remove a line of dead code, identified by coverity and reported by Nuno
2366 Lopes.
2367
2368 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2369
2370 (1) The include/exclude patterns were being applied to the whole pathnames
2371 of files, instead of just to the final components.
2372
2373 (2) If there was more than one level of directory, the subdirectories were
2374 skipped unless they satisfied the include/exclude conditions. This is
2375 inconsistent with GNU grep (and could even be seen as contrary to the
2376 pcregrep specification - which I improved to make it absolutely clear).
2377 The action now is always to scan all levels of directory, and just
2378 apply the include/exclude patterns to regular files.
2379
2380 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2381 --exclude_dir in the tests to avoid scanning .svn directories.
2382
2383 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2384 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2385 doesn't support NULs in patterns.
2386
2387 7. Added some missing "const"s to declarations of static tables in
2388 pcre_compile.c and pcre_dfa_exec.c.
2389
2390 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2391 caused by fix #2 above. (Subsequently also a second patch to fix the
2392 first patch. And a third patch - this was a messy problem.)
2393
2394 9. Applied Craig's patch to remove the use of push_back().
2395
2396 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2397 matching function regexec().
2398
2399 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2400 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2401 references. PCRE supports relative numbers with this syntax (I don't think
2402 Oniguruma does).
2403
2404 12. Previously, a group with a zero repeat such as (...){0} was completely
2405 omitted from the compiled regex. However, this means that if the group
2406 was called as a subroutine from elsewhere in the pattern, things went wrong
2407 (an internal error was given). Such groups are now left in the compiled
2408 pattern, with a new opcode that causes them to be skipped at execution
2409 time.
2410
2411 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2412 to the way PCRE behaves:
2413
2414 (a) A lone ] character is dis-allowed (Perl treats it as data).
2415
2416 (b) A back reference to an unmatched subpattern matches an empty string
2417 (Perl fails the current match path).
2418
2419 (c) A data ] in a character class must be notated as \] because if the
2420 first data character in a class is ], it defines an empty class. (In
2421 Perl it is not possible to have an empty class.) The empty class []
2422 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2423 The negative empty class [^] matches any one character, independently
2424 of the DOTALL setting.
2425
2426 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2427 non-existent subpattern following a character class starting with ']' and
2428 containing () gave an internal compiling error instead of "reference to
2429 non-existent subpattern". Fortunately, when the pattern did exist, the
2430 compiled code was correct. (When scanning forwards to check for the
2431 existence of the subpattern, it was treating the data ']' as terminating
2432 the class, so got the count wrong. When actually compiling, the reference
2433 was subsequently set up correctly.)
2434
2435 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2436 it was being rejected as not supported by pcre_dfa_exec(), even though
2437 other assertions are supported. I have made pcre_dfa_exec() support
2438 (*FAIL).
2439
2440 16. The implementation of 13c above involved the invention of a new opcode,
2441 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2442 cannot be changed at match time, I realized I could make a small
2443 improvement to matching performance by compiling OP_ALLANY instead of
2444 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2445 on the OP_ANY path.
2446
2447 17. Compiling pcretest on Windows with readline support failed without the
2448 following two fixes: (1) Make the unistd.h include conditional on
2449 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2450
2451 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2452 ncurses library to be included for pcretest when ReadLine support is
2453 requested, but also to allow for it to be overridden. This patch came from
2454 Daniel Bergström.
2455
2456 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2457 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2458 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2459 spotting this.
2460
2461
2462 Version 7.6 28-Jan-08
2463 ---------------------
2464
2465 1. A character class containing a very large number of characters with
2466 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2467 overflow.
2468
2469 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2470 HAVE_LONG_LONG is not defined.
2471
2472 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2473 bring them up to date and include new features. This patch includes:
2474
2475 - Fixed PH's badly added libz and libbz2 support.
2476 - Fixed a problem with static linking.
2477 - Added pcredemo. [But later removed - see 7 below.]
2478 - Fixed dftables problem and added an option.
2479 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2480 HAVE_LONG_LONG.
2481 - Added readline support for pcretest.
2482 - Added an listing of the option settings after cmake has run.
2483
2484 4. A user submitted a patch to Makefile that makes it easy to create
2485 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2486 Makefile.am that cause it to include this special target, without
2487 affecting anything else. Note that the same mingw target plus all
2488 the other distribution libraries and programs are now supported
2489 when configuring with CMake (see 6 below) instead of with
2490 Configure/Make.
2491
2492 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2493 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2494 exported in the Windows port". It has not yet been confirmed that the patch
2495 solves the problem, but it does no harm.
2496
2497 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2498 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2499 with CMake, and also correct the comment about stack recursion.
2500
2501 7. Remove the automatic building of pcredemo from the ./configure system and
2502 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2503 of a program that users should build themselves after PCRE is installed, so
2504 building it automatically is not really right. What is more, it gave
2505 trouble in some build environments.
2506
2507 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2508
2509
2510 Version 7.5 10-Jan-08
2511 ---------------------
2512
2513 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2514 values in parens when parsing an RE using the C++ wrapper."
2515
2516 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2517 Characters greater than 255 were excluded from the class instead of being
2518 included.
2519
2520 3. The same bug as (2) above applied to negated POSIX classes such as
2521 [:^space:].
2522
2523 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2524 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2525 I have changed it.
2526
2527 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2528 first named subpattern) and a construct such as (?&a) would reference the
2529 first named subpattern whose name started with "a" (in other words, the
2530 length check was missing). Both these problems are fixed. "Subpattern name
2531 expected" is now given for (?&) (a zero-length name), and this patch also
2532 makes it give the same error for \k'' (previously it complained that that
2533 was a reference to a non-existent subpattern).
2534
2535 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2536 this is right because (?- can be followed by option settings as well as by
2537 digits. I have, however, made the messages clearer.
2538
2539 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2540 than the number used in the conditional) now cause a compile-time error.
2541 This is actually not compatible with Perl, which accepts such patterns, but
2542 treats the conditional as always being FALSE (as PCRE used to), but it
2543 seems to me that giving a diagnostic is better.
2544
2545 8. Change "alphameric" to the more common word "alphanumeric" in comments
2546 and messages.
2547
2548 9. Fix two occurrences of "backslash" in comments that should have been
2549 "backspace".
2550
2551 10. Remove two redundant lines of code that can never be obeyed (their function
2552 was moved elsewhere).
2553
2554 11. The program that makes PCRE's Unicode character property table had a bug
2555 which caused it to generate incorrect table entries for sequences of
2556 characters that have the same character type, but are in different scripts.
2557 It amalgamated them into a single range, with the script of the first of
2558 them. In other words, some characters were in the wrong script. There were
2559 thirteen such cases, affecting characters in the following ranges:
2560
2561 U+002b0 - U+002c1
2562 U+0060c - U+0060d
2563 U+0061e - U+00612
2564 U+0064b - U+0065e
2565 U+0074d - U+0076d
2566 U+01800 - U+01805
2567 U+01d00 - U+01d77
2568 U+01d9b - U+01dbf
2569 U+0200b - U+0200f
2570 U+030fc - U+030fe
2571 U+03260 - U+0327f
2572 U+0fb46 - U+0fbb1
2573 U+10450 - U+1049d
2574
2575 12. The -o option (show only the matching part of a line) for pcregrep was not
2576 compatible with GNU grep in that, if there was more than one match in a
2577 line, it showed only the first of them. It now behaves in the same way as
2578 GNU grep.
2579
2580 13. If the -o and -v options were combined for pcregrep, it printed a blank
2581 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2582 does the same. The return code can be used to tell if there were any
2583 non-matching lines.
2584
2585 14. Added --file-offsets and --line-offsets to pcregrep.
2586
2587 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2588 infinitely looping recursion. The bug was that positive lookaheads were not
2589 being skipped when checking for a possible empty match (negative lookaheads
2590 and both kinds of lookbehind were skipped).
2591
2592 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2593 inclusion of <windows.h> to before rather than after the definition of
2594 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2595
2596 17. Specifying a possessive quantifier with a specific limit for a Unicode
2597 character property caused pcre_compile() to compile bad code, which led at
2598 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2599 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2600 caused the error; without that there was no problem.
2601
2602 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2603
2604 19. Added --enable-pcretest-libreadline.
2605
2606 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2607 RE::GlobalReplace(). As a result, the number of replacements returned was
2608 double what it should be. I removed one of the increments, but Craig sent a
2609 later patch that removed the other one (the right fix) and added unit tests
2610 that check the return values (which was not done before).
2611
2612 21. Several CMake things:
2613
2614 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2615 the names libpcre and libpcreposix, not just pcre and pcreposix.
2616
2617 (2) The above change means that pcretest and pcregrep are now correctly
2618 linked with the newly-built libraries, not previously installed ones.
2619
2620 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2621
2622 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2623 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2624 UTF-8 newline character). The key issue is that the pattern starts .*;
2625 this means that the match must be either at the beginning, or after a
2626 newline. The bug was in the code for advancing after a failed match and
2627 checking that the new position followed a newline. It was not taking
2628 account of UTF-8 characters correctly.
2629
2630 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2631 character classes. PCRE was not treating the sequence [:...:] as a
2632 character class unless the ... were all letters. Perl, however, seems to
2633 allow any characters between [: and :], though of course it rejects as
2634 unknown any "names" that contain non-letters, because all the known class
2635 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2636 for example, whereas PCRE did not - it did not recognize a POSIX character
2637 class. This seemed a bit dangerous, so the code has been changed to be
2638 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2639 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2640 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2641 Perl does, and where it didn't before.
2642
2643 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2644 Windows environments %n is disabled by default.
2645
2646
2647 Version 7.4 21-Sep-07
2648 ---------------------
2649
2650 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2651 means that a class such as [\s] counted as "explicit reference to CR or
2652 LF". That isn't really right - the whole point of the change was to try to
2653 help when there was an actual mention of one of the two characters. So now
2654 the change happens only if \r or \n (or a literal CR or LF) character is
2655 encountered.
2656
2657 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2658 of both had grown to the point where there were only 3 bits left.
2659 Fortunately, there was spare space in the data structure, and so I have
2660 moved the internal flags into a new 16-bit field to free up more option
2661 bits.
2662
2663 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2664 but did not set the internal JCHANGED flag - either of these is enough to
2665 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2666 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2667 start sets both bits.
2668
2669 4. Added options (at build time, compile time, exec time) to change \R from
2670 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2671
2672 5. doc/pcresyntax.html was missing from the distribution.
2673
2674 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2675 compatibility, even though it is no longer used.
2676
2677 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2678 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2679 windows.h file is present (where different names are used). [This was
2680 reversed later after testing - see 16 below.]
2681
2682 8. Changed all #include <config.h> to #include "config.h". There were also
2683 some further <pcre.h> cases that I changed to "pcre.h".
2684
2685 9. When pcregrep was used with the --colour option, it missed the line ending
2686 sequence off the lines that it output.
2687
2688 10. It was pointed out to me that arrays of string pointers cause lots of
2689 relocations when a shared library is dynamically loaded. A technique of
2690 using a single long string with a table of offsets can drastically reduce
2691 these. I have refactored PCRE in four places to do this. The result is
2692 dramatic:
2693
2694 Originally: 290
2695 After changing UCP table: 187
2696 After changing error message table: 43
2697 After changing table of "verbs" 36
2698 After changing table of Posix names 22
2699
2700 Thanks to the folks working on Gregex for glib for this insight.
2701
2702 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2703 unicode-properties was also set.
2704
2705 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2706
2707 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2708 checked only for CRLF.
2709
2710 14. Added casts to pcretest.c to avoid compiler warnings.
2711
2712 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2713
2714 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2715 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2716 entirely. This removes changes made in 7 above.
2717
2718 17. The CMake files have been updated, and there is now more information about
2719 building with CMake in the NON-UNIX-USE document.
2720
2721
2722 Version 7.3 28-Aug-07
2723 ---------------------
2724
2725 1. In the rejigging of the build system that eventually resulted in 7.1, the
2726 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2727 brackets there is not right, since it causes compilers to look for an
2728 installed pcre.h, not the version that is in the source that is being
2729 compiled (which of course may be different). I have changed it back to:
2730
2731 #include "pcre.h"
2732
2733 I have a vague recollection that the change was concerned with compiling in
2734 different directories, but in the new build system, that is taken care of
2735 by the VPATH setting the Makefile.
2736
2737 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2738 when the subject happened to end in the byte 0x85 (e.g. if the last
2739 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2740 characters but of course it shouldn't be taken as a newline when it is part
2741 of another character. The bug was that, for an unlimited repeat of . in
2742 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2743 characters when looking for a newline.
2744
2745 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2746
2747 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2748 in debug output.
2749
2750 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2751 long printing in the pcrecpp unittest when running under MinGW.
2752
2753 6. ESC_K was left out of the EBCDIC table.
2754
2755 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2756 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2757 limit also applies to "virtual nesting" when a pattern is recursive, and in
2758 this case 1000 isn't so big. I have been able to remove this limit at the
2759 expense of backing off one optimization in certain circumstances. Normally,
2760 when pcre_exec() would call its internal match() function recursively and
2761 immediately return the result unconditionally, it uses a "tail recursion"
2762 feature to save stack. However, when a subpattern that can match an empty
2763 string has an unlimited repetition quantifier, it no longer makes this
2764 optimization. That gives it a stack frame in which to save the data for
2765 checking that an empty string has been matched. Previously this was taken
2766 from the 1000-entry workspace that had been reserved. So now there is no
2767 explicit limit, but more stack is used.
2768
2769 8. Applied Daniel's patches to solve problems with the import/export magic
2770 syntax that is required for Windows, and which was going wrong for the
2771 pcreposix and pcrecpp parts of the library. These were overlooked when this
2772 problem was solved for the main library.
2773
2774 9. There were some crude static tests to avoid integer overflow when computing
2775 the size of patterns that contain repeated groups with explicit upper
2776 limits. As the maximum quantifier is 65535, the maximum group length was
2777 set at 30,000 so that the product of these two numbers did not overflow a
2778 32-bit integer. However, it turns out that people want to use groups that
2779 are longer than 30,000 bytes (though not repeat them that many times).
2780 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2781 made it possible to implement the integer overflow checks in a much more
2782 dynamic way, which I have now done. The artificial limitation on group
2783 length has been removed - we now have only the limit on the total length of
2784 the compiled pattern, which depends on the LINK_SIZE setting.
2785
2786 10. Fixed a bug in the documentation for get/copy named substring when
2787 duplicate names are permitted. If none of the named substrings are set, the
2788 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2789 empty string.
2790
2791 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2792 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2793 because the ] is interpreted as the first data character and the
2794 terminating ] is not found. PCRE has been made compatible with Perl in this
2795 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2796 cause memory overwriting.
2797
2798 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2799 string has been matched (to stop an infinite loop). It was not recognizing
2800 a conditional subpattern that could match an empty string if that
2801 subpattern was within another subpattern. For example, it looped when
2802 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2803 condition was not nested. This bug has been fixed.
2804
2805 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2806 past the start of the subject in the presence of bytes with the top bit
2807 set, for example "\x8aBCD".
2808
2809 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2810 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2811
2812 14. Optimized (?!) to (*FAIL).
2813
2814 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2815 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2816 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2817 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2818 does: it's just the validity check that is more restrictive.
2819
2820 16. Inserted checks for integer overflows during escape sequence (backslash)
2821 processing, and also fixed erroneous offset values for syntax errors during
2822 backslash processing.
2823
2824 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2825 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2826
2827 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2828 caused an overrun.
2829
2830 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2831 something other than just ASCII characters) inside a group that had an
2832 unlimited repeat caused a loop at compile time (while checking to see
2833 whether the group could match an empty string).
2834
2835 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2836 [\P{Any}] did so. (Error in the code for printing property names.)
2837
2838 21. An orphan \E inside a character class could cause a crash.
2839
2840 22. A repeated capturing bracket such as (A)? could cause a wild memory
2841 reference during compilation.
2842
2843 23. There are several functions in pcre_compile() that scan along a compiled
2844 expression for various reasons (e.g. to see if it's fixed length for look
2845 behind). There were bugs in these functions when a repeated \p or \P was
2846 present in the pattern. These operators have additional parameters compared
2847 with \d, etc, and these were not being taken into account when moving along
2848 the compiled data. Specifically:
2849
2850 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2851 length.
2852
2853 (b) An item such as \pL+ within a repeated group could cause crashes or
2854 loops.
2855
2856 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2857 "reference to non-existent subpattern" error.
2858
2859 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2860
2861 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2862 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2863
2864 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2865
2866 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2867 character were causing crashes (broken optimization).
2868
2869 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2870 \p or \P) caused a compile-time loop.
2871
2872 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2873 break. For example, the unstudied pattern [\r\n]A does not match the string
2874 "\r\nA" because change 7.0/46 below moves the current point on by two
2875 characters after failing to match at the start. However, the pattern \nA
2876 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2877 the same is true. There doesn't seem any very clean way out of this, but
2878 what I have chosen to do makes the common cases work: PCRE now takes note
2879 of whether there can be an explicit match for \r or \n anywhere in the
2880 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2881 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2882 pattern has explicit CR or LF references.
2883
2884 29. Added (*CR) etc for changing newline setting at start of pattern.
2885
2886
2887 Version 7.2 19-Jun-07
2888 ---------------------
2889
2890 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2891 which is apparently normally available under Windows.
2892
2893 2. Re-jig the pcregrep tests with different newline settings in an attempt
2894 to make them independent of the local environment's newline setting.
2895
2896 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2897
2898 4. Some of the "internals" tests were previously cut out when the link size
2899 was not 2, because the output contained actual offsets. The recent new
2900 "Z" feature of pcretest means that these can be cut out, making the tests
2901 usable with all link sizes.
2902
2903 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2904 stack recursion. This gives a massive performance boost under BSD, but just
2905 a small improvement under Linux. However, it saves one field in the frame
2906 in all cases.
2907
2908 6. Added more features from the forthcoming Perl 5.10:
2909
2910 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2911 recursion call. It refers to the nth most recently opened parentheses.
2912
2913 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2914 to be opened parentheses.
2915
2916 (c) Conditions that refer to capturing parentheses can be specified
2917 relatively, for example, (?(-2)... or (?(+3)...
2918
2919 (d) \K resets the start of the current match so that everything before
2920 is not part of it.
2921
2922 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2923
2924 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2925 reference syntax.
2926
2927 (g) (?| introduces a group in which the numbering of parentheses in each
2928 alternative starts with the same number.
2929
2930 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2931
2932 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2933 PCRE_INFO_JCHANGED.
2934
2935 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2936 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2937 for detecting groups that can match an empty string.
2938
2939 9. A pattern with a very large number of alternatives (more than several
2940 hundred) was running out of internal workspace during the pre-compile
2941 phase, where pcre_compile() figures out how much memory will be needed. A
2942 bit of new cunning has reduced the workspace needed for groups with
2943 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2944 workspace instead of running out of the 4096 that are available.
2945
2946 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2947
2948 11. Applied patch from Google to remove an optimization that didn't quite work.
2949 The report of the bug said:
2950
2951 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2952 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2953 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2954
2955 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2956 it matched the wrong number of bytes.
2957
2958
2959 Version 7.1 24-Apr-07
2960 ---------------------
2961
2962 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2963 that is more "standard", making use of automake and other Autotools. There
2964 is some re-arrangement of the files and adjustment of comments consequent
2965 on this.
2966
2967 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2968 for recursive directory scanning broke on some systems because the files
2969 are not scanned in any specific order and on different systems the order
2970 was different. A call to "sort" has been inserted into RunGrepTest for the
2971 approprate test as a short-term fix. In the longer term there may be an
2972 alternative.
2973
2974 3. I had an email from Eric Raymond about problems translating some of PCRE's
2975 man pages to HTML (despite the fact that I distribute HTML pages, some
2976 people do their own conversions for various reasons). The problems
2977 concerned the use of low-level troff macros .br and .in. I have therefore
2978 removed all such uses from the man pages (some were redundant, some could
2979 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2980 HTML has been updated to handle .nf/.fi and to complain if it encounters
2981 .br or .in.
2982
2983 4. Updated comments in configure.ac that get placed in config.h.in and also
2984 arranged for config.h to be included in the distribution, with the name
2985 config.h.generic, for the benefit of those who have to compile without
2986 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2987
2988 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2989 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2990 makevp.bat for the current PCRE, using the additional files
2991 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2992
2993 6. A Windows user reported a minor discrepancy with test 2, which turned out
2994 to be caused by a trailing space on an input line that had got lost in his
2995 copy. The trailing space was an accident, so I've just removed it.
2996
2997 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2998 that is needed.
2999
3000 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3001 as "const" (a) because they are and (b) because it helps the PHP
3002 maintainers who have recently made a script to detect big data structures
3003 in the php code that should be moved to the .rodata section. I remembered
3004 to update Builducptable as well, so it won't revert if ucptable.h is ever
3005 re-created.
3006
3007 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3008 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3009 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3010 support is not required. This saves 1.5-2K of code, which is important in
3011 some applications.
3012
3013 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3014 so as not to refer to the tables, even though these functions will never be
3015 called when UTF-8 support is disabled. Otherwise there are problems with a
3016 shared library.
3017
3018 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3019
3020 (a) It was defining its arguments as char * instead of void *.
3021
3022 (b) It was assuming that all moves were upwards in memory; this was true
3023 a long time ago when I wrote it, but is no longer the case.
3024
3025 The emulated memove() is provided for those environments that have neither
3026 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3027 is clearly not the case, as these two bugs were recently reported.
3028
3029 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3030 and Detrail to create the HTML documentation, the .txt form of the man
3031 pages, and it removes trailing spaces from listed files. It also creates
3032 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3033 case, it wraps all the #defines with #ifndefs. This script should be run
3034 before "make dist".
3035
3036 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3037 with Unicode property support.
3038
3039 (a) For a maximizing quantifier, if the two different cases of the
3040 character were of different lengths in their UTF-8 codings (there are
3041 some cases like this - I found 11), and the matching function had to
3042 back up over a mixture of the two cases, it incorrectly assumed they
3043 were both the same length.
3044
3045 (b) When PCRE was configured to use the heap rather than the stack for
3046 recursion during matching, it was not correctly preserving the data for
3047 the other case of a UTF-8 character when checking ahead for a match
3048 while processing a minimizing repeat. If the check also involved
3049 matching a wide character, but failed, corruption could cause an
3050 erroneous result when trying to check for a repeat of the original
3051 character.
3052
3053 13. Some tidying changes to the testing mechanism:
3054
3055 (a) The RunTest script now detects the internal link size and whether there
3056 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3057 values substituted by "configure". (The RunGrepTest script already did
3058 this for UTF-8.) The configure.ac script no longer substitutes the
3059 relevant variables.
3060
3061 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3062 with length and offset values. This means that the output is different
3063 for different internal link sizes. Test 2 is skipped for link sizes
3064 other than 2 because of this, bypassing the problem. Unfortunately,
3065 there was also a test in test 3 (the locale tests) that used /B and
3066 failed for link sizes other than 2. Rather than cut the whole test out,
3067 I have added a new /Z option to pcretest that replaces the length and
3068 offset values with spaces. This is now used to make test 3 independent
3069 of link size. (Test 2 will be tidied up later.)
3070
3071 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3072 segmentation fault instead of returning the appropriate error message.
3073
3074 15. In multiline mode when the newline sequence was set to "any", the pattern
3075 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3076 This doesn't seem right; it now treats the CRLF combination as the line
3077 ending, and so does not match in that case. It's only a pattern such as ^$
3078 that would hit this one: something like ^ABC$ would have failed after \r
3079 and then tried again after \r\n.
3080
3081 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3082 in an attempt to make files that differ only in their line terminators
3083 compare equal. This works on Linux.
3084
3085 17. Under certain error circumstances pcregrep might try to free random memory
3086 as it exited. This is now fixed, thanks to valgrind.
3087
3088 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3089 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3090 was because its rules for how to advance for /g after matching an empty
3091 string at the end of a line did not allow for this case. They now check for
3092 it specially.
3093
3094 20. pcretest is supposed to handle patterns and data of any length, by
3095 extending its buffers when necessary. It was getting this wrong when the
3096 buffer for a data line had to be extended.
3097
3098 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3099 CRLF as a newline sequence.
3100
3101 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3102 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3103 I have nevertheless tidied it up.
3104
3105 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3106
3107 24. Added a man page for pcre-config.
3108
3109
3110 Version 7.0 19-Dec-06
3111 ---------------------
3112
3113 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3114 moving to gcc 4.1.1.
3115
3116 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3117 sys/time.h, which is documented as needed for this function. It doesn't
3118 seem to matter on Linux, but it showed up on some releases of OS X.
3119
3120 3. It seems that there are systems where bytes whose values are greater than
3121 127 match isprint() in the "C" locale. The "C" locale should be the
3122 default when a C program starts up. In most systems, only ASCII printing
3123 characters match isprint(). This difference caused the output from pcretest
3124 to vary, making some of the tests fail. I have changed pcretest so that:
3125
3126 (a) When it is outputting text in the compiled version of a pattern, bytes
3127 other than 32-126 are always shown as hex escapes.
3128
3129 (b) When it is outputting text that is a matched part of a subject string,
3130 it does the same, unless a different locale has been set for the match
3131 (using the /L modifier). In this case, it uses isprint() to decide.
3132
3133 4. Fixed a major bug that caused incorrect computation of the amount of memory
3134 required for a compiled pattern when options that changed within the
3135 pattern affected the logic of the preliminary scan that determines the
3136 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3137 that the computed length was too small. The symptoms of this bug were
3138 either the PCRE error "internal error: code overflow" from pcre_compile(),
3139 or a glibc crash with a message such as "pcretest: free(): invalid next
3140 size (fast)". Examples of patterns that provoked this bug (shown in
3141 pcretest format) are:
3142
3143 /(?-x: )/x
3144 /(?x)(?-x: \s*#\s*)/
3145 /((?i)[\x{c0}])/8
3146 /(?i:[\x{c0}])/8
3147
3148 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3149 is now done differently.
3150
3151 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3152 wrapper classes; (b) implement a new function in the C++ scanner that is
3153 more efficient than the old way of doing things because it avoids levels of
3154 recursion in the regex matching; (c) add a paragraph to the documentation
3155 for the FullMatch() function.
3156
3157 6. The escape sequence \n was being treated as whatever was defined as
3158 "newline". Not only was this contrary to the documentation, which states
3159 that \n is character 10 (hex 0A), but it also went horribly wrong when
3160 "newline" was defined as CRLF. This has been fixed.
3161
3162 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3163 was being set to -1 for the "end of line" case (supposedly a value that no
3164 character can have). Though this value is never used (the check for end of
3165 line is "zero bytes in current character"), it caused compiler complaints.
3166 I've changed it to 0xffffffff.
3167
3168 8. In pcre_version.c, the version string was being built by a sequence of
3169 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3170 string (as it is for production releases) called a macro with an empty
3171 argument. The C standard says the result of this is undefined. The gcc
3172 compiler treats it as an empty string (which was what was wanted) but it is
3173 reported that Visual C gives an error. The source has been hacked around to
3174 avoid this problem.
3175
3176 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3177 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3178 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3179 of them did).
3180
3181 10. Originally, pcretest opened its input and output without "b"; then I was
3182 told that "b" was needed in some environments, so it was added for release
3183 5.0 to both the input and output. (It makes no difference on Unix-like
3184 systems.) Later I was told that it is wrong for the input on Windows. I've
3185 now abstracted the modes into two macros, to make it easier to fiddle with
3186 them, and removed "b" from the input mode under Windows.
3187
3188 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3189
3190 12. Added -help and --help to pcretest as an official way of being reminded
3191 of the options.
3192
3193 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3194 and pcrecpp.cc because they annoy compilers at high warning levels.
3195
3196 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3197
3198 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3199 scripts are not C programs :-) and which was not noticed because it works
3200 on Linux.
3201
3202 16. pcretest is supposed to handle any length of pattern and data line (as one
3203 line or as a continued sequence of lines) by extending its input buffer if
3204 necessary. This feature was broken for very long pattern lines, leading to
3205 a string of junk being passed to pcre_compile() if the pattern was longer
3206 than about 50K.
3207
3208 17. I have done a major re-factoring of the way pcre_compile() computes the
3209 amount of memory needed for a compiled pattern. Previously, there was code
3210 that made a preliminary scan of the pattern in order to do this. That was
3211 OK when PCRE was new, but as the facilities have expanded, it has become
3212 harder and harder to keep it in step with the real compile phase, and there
3213 have been a number of bugs (see for example, 4 above). I have now found a
3214 cunning way of running the real compile function in a "fake" mode that
3215 enables it to compute how much memory it would need, while actually only
3216 ever using a few hundred bytes of working memory and without too many
3217 tests of the mode. This should make future maintenance and development
3218 easier. A side effect of this work is that the limit of 200 on the nesting
3219 depth of parentheses has been removed (though this was never a serious
3220 limitation, I suspect). However, there is a downside: pcre_compile() now
3221 runs more slowly than before (30% or more, depending on the pattern). I
3222 hope this isn't a big issue. There is no effect on runtime performance.
3223
3224 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3225 newline (only possible for the last line of a file) and it was a
3226 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3227
3228 19. Added additional timing features to pcretest. (1) The -tm option now times
3229 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3230 separate command line item, by a number that specifies the number of
3231 repeats to use when timing. The default is 50000; this gives better
3232 precision, but takes uncomfortably long for very large patterns.
3233
3234 20. Extended pcre_study() to be more clever in cases where a branch of a
3235 subpattern has no definite first character. For example, (a*|b*)[cd] would
3236 previously give no result from pcre_study(). Now it recognizes that the
3237 first character must be a, b, c, or d.
3238
3239 21. There was an incorrect error "recursive call could loop indefinitely" if
3240 a subpattern (or the entire pattern) that was being tested for matching an
3241 empty string contained only one non-empty item after a nested subpattern.
3242 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3243 incorrectly, because the \d was being skipped in the check.
3244
3245 22. The pcretest program now has a new pattern option /B and a command line
3246 option -b, which is equivalent to adding /B to every pattern. This causes
3247 it to show the compiled bytecode, without the additional information that
3248 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3249 is the same as /B/I).
3250
3251 23. A new optimization is now able automatically to treat some sequences such
3252 as a*b as a*+b. More specifically, if something simple (such as a character
3253 or a simple class like \d) has an unlimited quantifier, and is followed by
3254 something that cannot possibly match the quantified thing, the quantifier
3255 is automatically "possessified".
3256
3257 24. A recursive reference to a subpattern whose number was greater than 39
3258 went wrong under certain circumstances in UTF-8 mode. This bug could also
3259 have affected the operation of pcre_study().
3260
3261 25. Realized that a little bit of performance could be had by replacing
3262 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3263
3264 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3265
3266 27. Possessive quantifiers such as a++ were previously implemented by turning
3267 them into atomic groups such as ($>a+). Now they have their own opcodes,
3268 which improves performance. This includes the automatically created ones
3269 from 23 above.
3270
3271 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3272 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3273 the first matched character to be a colon. This applied both to named and
3274 numbered groups.
3275
3276 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3277
3278 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3279 building PCRE on Windows easier, so I have included it in the distribution.
3280
3281 31. There is now a check in pcretest against a ridiculously large number being
3282 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3283 loop, the loop is abandoned.
3284
3285 32. Forward references to subpatterns in conditions such as (?(2)...) where
3286 subpattern 2 is defined later cause pcre_compile() to search forwards in
3287 the pattern for the relevant set of parentheses. This search went wrong
3288 when there were unescaped parentheses in a character class, parentheses
3289 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3290
3291 33. "Subroutine" calls and backreferences were previously restricted to
3292 referencing subpatterns earlier in the regex. This restriction has now
3293 been removed.
3294
3295 34. Added a number of extra features that are going to be in Perl 5.10. On the
3296 whole, these are just syntactic alternatives for features that PCRE had
3297 previously implemented using the Python syntax or my own invention. The
3298 other formats are all retained for compatibility.
3299
3300 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3301 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3302 also .NET compatible.
3303
3304 (b) A recursion or subroutine call to a named group can now be defined as
3305 (?&name) as well as (?P>name).
3306
3307 (c) A backreference to a named group can now be defined as \k<name> or
3308 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3309 5.10, are also .NET compatible.
3310
3311 (d) A conditional reference to a named group can now use the syntax
3312 (?(<name>) or (?('name') as well as (?(name).
3313
3314 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3315 groups (named and numbered) that are never evaluated inline, but can be
3316 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3317 is always false. There may be only one alternative in such a group.
3318
3319 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3320 as the simple (?(R). The condition is true only if the most recent
3321 recursion is that of the given number or name. It does not search out
3322 through the entire recursion stack.
3323
3324 (g) The escape \gN or \g{N} has been added, where N is a positive or
3325 negative number, specifying an absolute or relative reference.
3326
3327 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3328 some "unreachable code" warnings.
3329
3330 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3331 things, this adds five new scripts.
3332
3333 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3334 There were also incompatibilities regarding the handling of \Q..\E inside
3335 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3336 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3337
3338 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3339 matches an empty string, and forcibly breaks the loop. There were bugs in
3340 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3341 against aaaa the result was just "a" rather than "aaaa", for example. Two
3342 separate and independent bugs (that affected different cases) have been
3343 fixed.
3344
3345 39. Refactored the code to abolish the use of different opcodes for small
3346 capturing bracket numbers. This is a tidy that I avoided doing when I
3347 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3348 The new approach is not only tidier, it makes it possible to reduce the
3349 memory needed to fix the previous bug (38).
3350
3351 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3352 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3353 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3354 mode.
3355
3356 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3357 report.
3358
3359 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3360 copy construction and assignment in the C++ wrapper.
3361
3362 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3363 couple of bugs that could have given wrong results in the "--newline=crlf"
3364 case.
3365
3366 44. Added a number of casts and did some reorganization of signed/unsigned int
3367 variables following suggestions from Dair Grant. Also renamed the variable
3368 "this" as "item" because it is a C++ keyword.
3369
3370 45. Arranged for dftables to add
3371
3372 #include "pcre_internal.h"
3373
3374 to pcre_chartables.c because without it, gcc 4.x may remove the array
3375 definition from the final binary if PCRE is built into a static library and
3376 dead code stripping is activated.
3377
3378 46. For an unanchored pattern, if a match attempt fails at the start of a
3379 newline sequence, and the newline setting is CRLF or ANY, and the next two
3380 characters are CRLF, advance by two characters instead of one.
3381
3382
3383 Version 6.7 04-Jul-06
3384 ---------------------
3385
3386 1. In order to handle tests when input lines are enormously long, pcretest has
3387 been re-factored so that it automatically extends its buffers when
3388 necessary. The code is crude, but this _is_ just a test program. The
3389 default size has been increased from 32K to 50K.
3390
3391 2. The code in pcre_study() was using the value of the re argument before
3392 testing it for NULL. (Of course, in any sensible call of the function, it
3393 won't be NULL.)
3394
3395 3. The memmove() emulation function in pcre_internal.h, which is used on
3396 systems that lack both memmove() and bcopy() - that is, hardly ever -
3397 was missing a "static" storage class specifier.
3398
3399 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3400 containing an extended class (one that cannot be represented by a bitmap
3401 because it contains high-valued characters or Unicode property items, e.g.
3402 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3403 pattern, but PCRE should not loop if you do not (it no longer does).
3404 [Detail: two cases were found: (a) a repeated subpattern containing an
3405 extended class; (b) a recursive reference to a subpattern that followed a
3406 previous extended class. It wasn't skipping over the extended class
3407 correctly when UTF-8 mode was not set.]
3408
3409 5. A negated single-character class was not being recognized as fixed-length
3410 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3411 compile error "lookbehind assertion is not fixed length".
3412
3413 6. The RunPerlTest auxiliary script was showing an unexpected difference
3414 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3415 write a Perl script that can interpret lines of an input file either as
3416 byte characters or as UTF-8, which is what "perltest" was being required to
3417 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3418 can't do is switch easily at run time between having the "use utf8;" pragma
3419 or not. In the end, I fudged it by using the RunPerlTest script to insert
3420 "use utf8;" explicitly for the UTF-8 tests.
3421
3422 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3423 the end of the subject string, contrary to the documentation and to what
3424 Perl does. This was true of both matching functions. Now it matches only at
3425 the start of the subject and immediately after *internal* newlines.
3426
3427 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3428 a pointer to an int instead of a pointer to an unsigned long int. This
3429 caused problems on 64-bit systems.
3430
3431 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3432 instance of the 'standard' template library not being so standard".
3433
3434 10. There was no check on the number of named subpatterns nor the maximum
3435 length of a subpattern name. The product of these values is used to compute
3436 the size of the memory block for a compiled pattern. By supplying a very
3437 long subpattern name and a large number of named subpatterns, the size
3438 computation could be caused to overflow. This is now prevented by limiting
3439 the length of names to 32 characters, and the number of named subpatterns
3440 to 10,000.
3441
3442 11. Subpatterns that are repeated with specific counts have to be replicated in
3443 the compiled pattern. The size of memory for this was computed from the
3444 length of the subpattern and the repeat count. The latter is limited to
3445 65535, but there was no limit on the former, meaning that integer overflow
3446 could in principle occur. The compiled length of a repeated subpattern is
3447 now limited to 30,000 bytes in order to prevent this.
3448
3449 12. Added the optional facility to have named substrings with the same name.
3450
3451 13. Added the ability to use a named substring as a condition, using the
3452 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3453 are numbers (not recommended). Forward references are permitted.
3454
3455 14. Added forward references in named backreferences (if you see what I mean).
3456
3457 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3458 pattern could run off the end of the subject. For example, the pattern
3459 "(?s)(.{1,5})"8 did this with the subject "ab".
3460
3461 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3462 PCRE_CASELESS was set when matching characters that were quantified with ?
3463 or *.
3464
3465 17. A character class other than a single negated character that had a minimum
3466 but no maximum quantifier - for example [ab]{6,} - was not handled
3467 correctly by pce_dfa_exec(). It would match only one character.
3468
3469 18. A valid (though odd) pattern that looked like a POSIX character
3470 class but used an invalid character after [ (for example [[,abc,]]) caused
3471 pcre_compile() to give the error "Failed: internal error: code overflow" or
3472 in some cases to crash with a glibc free() error. This could even happen if
3473 the pattern terminated after [[ but there just happened to be a sequence of
3474 letters, a binary zero, and a closing ] in the memory that followed.
3475
3476 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3477 over the years. Originally (before any Unicode support), just the bottom 8
3478 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3479 output from "man perlunicode" includes this:
3480
3481 The regular expression compiler produces polymorphic opcodes. That
3482 is, the pattern adapts to the data and automatically switches to
3483 the Unicode character scheme when presented with Unicode data--or
3484 instead uses a traditional byte scheme when presented with byte
3485 data.
3486
3487 Sadly, a wide octal escape does not cause a switch, and in a string with
3488 no other multibyte characters, these octal escapes are treated as before.
3489 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3490 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3491 Unicode string.
3492
3493 I have not perpetrated such confusion in PCRE. Up till now, it took just
3494 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3495 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3496 translate to the appropriate multibyte character.
3497
3498 29. Applied some refactoring to reduce the number of warnings from Microsoft
3499 and Borland compilers. This has included removing the fudge introduced
3500 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3501 a warning about an unused variable.
3502
3503 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3504 characters since release 4.0, because Perl (from release 5.004) does not.
3505 [Or at least, is documented not to: some releases seem to be in conflict
3506 with the documentation.] However, when a pattern was studied with
3507 pcre_study() and all its branches started with \s, PCRE still included VT
3508 as a possible starting character. Of course, this did no harm; it just
3509 caused an unnecessary match attempt.
3510
3511 22. Removed a now-redundant internal flag bit that recorded the fact that case
3512 dependency changed within the pattern. This was once needed for "required
3513 byte" processing, but is no longer used. This recovers a now-scarce options
3514 bit. Also moved the least significant internal flag bit to the most-
3515 significant bit of the word, which was not previously used (hangover from
3516 the days when it was an int rather than a uint) to free up another bit for
3517 the future.
3518
3519 23. Added support for CRLF line endings as well as CR and LF. As well as the
3520 default being selectable at build time, it can now be changed at runtime
3521 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3522 specify that it is scanning data with non-default line endings.
3523
3524 24. Changed the definition of CXXLINK to make it agree with the definition of
3525 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3526
3527 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3528 recursions. This makes a big different to stack usage for some patterns.
3529
3530 26. If a subpattern containing a named recursion or subroutine reference such
3531 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3532 the space required for the compiled pattern went wrong and gave too small a
3533 value. Depending on the environment, this could lead to "Failed: internal
3534 error: code overflow at offset 49" or "glibc detected double free or
3535 corruption" errors.
3536
3537 27. Applied patches from Google (a) to support the new newline modes and (b) to
3538 advance over multibyte UTF-8 characters in GlobalReplace.
3539
3540 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3541 difference for some implementation of PCRE in some Windows version.
3542
3543 29. Added some extra testing facilities to pcretest:
3544
3545 \q<number> in a data line sets the "match limit" value
3546 \Q<number> in a data line sets the "match recursion limt" value
3547 -S <number> sets the stack size, where <number> is in megabytes
3548
3549 The -S option isn't available for Windows.
3550
3551
3552 Version 6.6 06-Feb-06
3553 ---------------------
3554
3555 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3556 in pcreposix.h. I have copied the definition from pcre.h.
3557
3558 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3559 because pcre.h is no longer a built file.
3560
3561 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3562 not normally included in the compiled code.
3563
3564
3565 Version 6.5 01-Feb-06
3566 ---------------------
3567
3568 1. When using the partial match feature with pcre_dfa_exec(), it was not
3569 anchoring the second and subsequent partial matches at the new starting
3570 point. This could lead to incorrect results. For example, with the pattern
3571 /1234/, partially matching against "123" and then "a4" gave a match.
3572
3573 2. Changes to pcregrep:
3574
3575 (a) All non-match returns from pcre_exec() were being treated as failures
3576 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3577 error message is output. Some extra information is given for the
3578 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3579 probably the only errors that are likely to be caused by users (by
3580 specifying a regex that has nested indefinite repeats, for instance).
3581 If there are more than 20 of these errors, pcregrep is abandoned.
3582
3583 (b) A binary zero was treated as data while matching, but terminated the
3584 output line if it was written out. This has been fixed: binary zeroes
3585 are now no different to any other data bytes.
3586
3587 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3588 used to set a locale for matching. The --locale=xxxx long option has
3589 been added (no short equivalent) to specify a locale explicitly on the
3590 pcregrep command, overriding the environment variables.
3591
3592 (d) When -B was used with -n, some line numbers in the output were one less
3593 than they should have been.
3594
3595 (e) Added the -o (--only-matching) option.
3596
3597 (f) If -A or -C was used with -c (count only), some lines of context were
3598 accidentally printed for the final match.
3599
3600 (g) Added the -H (--with-filename) option.
3601
3602 (h) The combination of options -rh failed to suppress file names for files
3603 that were found from directory arguments.
3604
3605 (i) Added the -D (--devices) and -d (--directories) options.
3606
3607 (j) Added the -F (--fixed-strings) option.
3608
3609 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3610
3611 (l) Added the --colo(u)r option.
3612
3613 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3614 is not present by default.
3615
3616 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3617 items such as (?R) or (?1), when the recursion could match a number of
3618 alternatives. If it matched one of the alternatives, but subsequently,
3619 outside the recursion, there was a failure, the code tried to back up into
3620 the recursion. However, because of the way PCRE is implemented, this is not
3621 possible, and the result was an incorrect result from the match.
3622
3623 In order to prevent this happening, the specification of recursion has
3624 been changed so that all such subpatterns are automatically treated as
3625 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3626
3627 4. I had overlooked the fact that, in some locales, there are characters for
3628 which isalpha() is true but neither isupper() nor islower() are true. In
3629 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3630 and ordfeminine) are like this. This affected the treatment of \w and \W
3631 when they appeared in character classes, but not when they appeared outside
3632 a character class. The bit map for "word" characters is now created
3633 separately from the results of isalnum() instead of just taking it from the
3634 upper, lower, and digit maps. (Plus the underscore character, of course.)
3635
3636 5. The above bug also affected the handling of POSIX character classes such as
3637 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3638 permanent tables. Instead, the bit maps for such a class were previously
3639 created as the appropriate unions of the upper, lower, and digit bitmaps.
3640 Now they are created by subtraction from the [[:word:]] class, which has
3641 its own bitmap.
3642
3643 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3644 It is created by subtracting the vertical space characters (\x09, \x0a,
3645 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3646 subtraction was done in the overall bitmap for a character class, meaning
3647 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3648 be recognized. This bug has been fixed.
3649
3650 7. Patches from the folks at Google:
3651
3652 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3653 real life, but is still worth protecting against".
3654
3655 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3656 regular expressions".
3657
3658 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3659 have it.
3660
3661 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3662 "configure" and the latter not, in order to fix a problem somebody had
3663 with compiling the Arg class on HP-UX.
3664
3665 (e) Improve the error-handling of the C++ wrapper a little bit.
3666
3667 (f) New tests for checking recursion limiting.
3668
3669 8. The pcre_memmove() function, which is used only if the environment does not
3670 have a standard memmove() function (and is therefore rarely compiled),
3671 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3672 returning a result (though PCRE never actually uses the result).
3673
3674 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3675 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3676 returned instead of calling malloc() with an overflowing number that would
3677 most likely cause subsequent chaos.
3678
3679 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3680
3681 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3682 with this option is matched, the nmatch and pmatch options of regexec() are
3683 ignored.
3684
3685 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3686 provided in case anyone wants to the the POSIX interface with UTF-8
3687 strings.
3688
3689 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3690 C++ linking (needed for some HP-UX environments).
3691
3692 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3693 (unused parameter) and in the pcre_printint() function (omitted "default"
3694 switch label when the default is to do nothing).
3695
3696 15. Added some code to make it possible, when PCRE is compiled as a C++
3697 library, to replace subject pointers for pcre_exec() with a smart pointer
3698 class, thus making it possible to process discontinuous strings.
3699
3700 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3701 much the same function. They were added by different people who were trying
3702 to make PCRE easy to compile on non-Unix systems. It has been suggested
3703 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3704 for compiling on Windows systems. I have therefore replaced it with
3705 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3706 defaults to "extern" for C or "extern C" for C++, which works fine on
3707 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3708 SCOPE with something explicit in config.h. In addition:
3709
3710 (a) pcreposix.h still had just "extern" instead of either of these macros;
3711 I have replaced it with PCRE_DATA_SCOPE.
3712
3713 (b) Functions such as _pcre_xclass(), which are internal to the library,
3714 but external in the C sense, all had PCRE_EXPORT in their definitions.
3715 This is apparently wrong for the Windows case, so I have removed it.
3716 (It makes no difference on Unix-like systems.)
3717
3718 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3719 of recursive calls to match(). This is different to MATCH_LIMIT because
3720 that limits the total number of calls to match(), not all of which increase
3721 the depth of recursion. Limiting the recursion depth limits the amount of
3722 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3723 when PCRE is compiled, and changed at run time. A patch from Google adds
3724 this functionality to the C++ interface.
3725
3726 18. Changes to the handling of Unicode character properties:
3727
3728 (a) Updated the table to Unicode 4.1.0.
3729
3730 (b) Recognize characters that are not in the table as "Cn" (undefined).
3731
3732 (c) I revised the way the table is implemented to a much improved format
3733 which includes recognition of ranges. It now supports the ranges that
3734 are defined in UnicodeData.txt, and it also amalgamates other
3735 characters into ranges. This has reduced the number of entries in the
3736 table from around 16,000 to around 3,000, thus reducing its size
3737 considerably. I realized I did not need to use a tree structure after
3738 all - a binary chop search is just as efficient. Having reduced the
3739 number of entries, I extended their size from 6 bytes to 8 bytes to
3740 allow for more data.
3741
3742 (d) Added support for Unicode script names via properties such as \p{Han}.
3743
3744 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3745 matching that character.
3746
3747 20. When matching a repeated Unicode property with a minimum greater than zero,
3748 (for example \pL{2,}), PCRE could look past the end of the subject if it
3749 reached it while seeking the minimum number of characters. This could
3750 happen only if some of the characters were more than one byte long, because
3751 there is a check for at least the minimum number of bytes.
3752
3753 21. Refactored the implementation of \p and \P so as to be more general, to
3754 allow for more different types of property in future. This has changed the
3755 compiled form incompatibly. Anybody with saved compiled patterns that use
3756 \p or \P will have to recompile them.
3757
3758 22. Added "Any" and "L&" to the supported property types.
3759
3760 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3761 but give a compile time error if the value is greater than 0xff.
3762
3763 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3764 accidentally not being installed or uninstalled.
3765
3766 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3767 made were to insert the current release number. This seemed silly, because
3768 it made things harder for people building PCRE on systems that don't run
3769 "configure". I have turned pcre.h into a distributed file, no longer built
3770 by "configure", with the version identification directly included. There is
3771 no longer a pcre.h.in file.
3772
3773 However, this change necessitated a change to the pcre-config script as
3774 well. It is built from pcre-config.in, and one of the substitutions was the
3775 release number. I have updated configure.ac so that ./configure now finds
3776 the release number by grepping pcre.h.
3777
3778 26. Added the ability to run the tests under valgrind.
3779
3780
3781 Version 6.4 05-Sep-05
3782 ---------------------
3783
3784 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3785 "--" to be printed when multiple files were scanned, even when none of the
3786 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3787 consider it to be a bug, and have restored the previous behaviour.
3788
3789 2. A couple of code tidies to get rid of compiler warnings.
3790
3791 3. The pcretest program used to cheat by referring to symbols in the library
3792 whose names begin with _pcre_. These are internal symbols that are not
3793 really supposed to be visible externally, and in some environments it is
3794 possible to suppress them. The cheating is now confined to including
3795 certain files from the library's source, which is a bit cleaner.
3796
3797 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3798 file's purpose clearer.
3799
3800 5. Reorganized pcre_ucp_findchar().
3801
3802
3803 Version 6.3 15-Aug-05
3804 ---------------------
3805
3806 1. The file libpcre.pc.in did not have general read permission in the tarball.
3807
3808 2. There were some problems when building without C++ support:
3809
3810 (a) If C++ support was not built, "make install" and "make test" still
3811 tried to test it.
3812
3813 (b) There were problems when the value of CXX was explicitly set. Some
3814 changes have been made to try to fix these, and ...
3815
3816 (c) --disable-cpp can now be used to explicitly disable C++ support.
3817
3818 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3819 backslash in a target when C++ was disabled. This confuses some
3820 versions of "make", apparently. Using an intermediate variable solves
3821 this. (Same for CPP_LOBJ.)
3822
3823 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3824 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3825 necessary on certain architectures.
3826
3827 4. Added a setting of -export-symbols-regex to the link command to remove
3828 those symbols that are exported in the C sense, but actually are local
3829 within the library, and not documented. Their names all begin with
3830 "_pcre_". This is not a perfect job, because (a) we have to except some
3831 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3832 available (and never for static libraries). I have made a note to try to
3833 find a way round (a) in the future.
3834
3835
3836 Version 6.2 01-Aug-05
3837 ---------------------
3838
3839 1. There was no test for integer overflow of quantifier values. A construction
3840 such as {1111111111111111} would give undefined results. What is worse, if
3841 a minimum quantifier for a parenthesized subpattern overflowed and became
3842 negative, the calculation of the memory size went wrong. This could have
3843 led to memory overwriting.
3844
3845 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3846
3847 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3848 operating environments where this matters.
3849
3850 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3851 PCRE options from within the C++ wrapper.
3852
3853 5. Named capturing subpatterns were not being correctly counted when a pattern
3854 was compiled. This caused two problems: (a) If there were more than 100
3855 such subpatterns, the calculation of the memory needed for the whole
3856 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3857 back references of the form \12, where the number was greater than 9, were
3858 not recognized as back references, even though there were sufficient
3859 previous subpatterns.
3860
3861 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3862 versions of gcc, e.g. 2.95.4.
3863
3864
3865 Version 6.1 21-Jun-05
3866 ---------------------
3867
3868 1. There was one reference to the variable "posix" in pcretest.c that was not
3869 surrounded by "#if !defined NOPOSIX".
3870
3871 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3872 the cross-check on the old pcre_info() function, for the benefit of the
3873 cut-down version of PCRE that is currently imported into Exim.
3874
3875 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3876 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3877 patterns that start (?i) but I don't think that matters. The use of (?i) is
3878 just an example; this all applies to the other options as well.
3879
3880 4. Since libtool seems to echo the compile commands it is issuing, the output
3881 from "make" can be reduced a bit by putting "@" in front of each libtool
3882 compile command.
3883
3884 5. Patch from the folks at Google for configure.in to be a bit more thorough
3885 in checking for a suitable C++ installation before trying to compile the
3886 C++ stuff. This should fix a reported problem when a compiler was present,
3887 but no suitable headers.
3888
3889 6. The man pages all had just "PCRE" as their title. I have changed them to
3890 be the relevant file name. I have also arranged that these names are
3891 retained in the file doc/pcre.txt, which is a concatenation in text format
3892 of all the man pages except the little individual ones for each function.
3893
3894 7. The NON-UNIX-USE file had not been updated for the different set of source
3895 files that come with release 6. I also added a few comments about the C++
3896 wrapper.
3897
3898
3899 Version 6.0 07-Jun-05
3900 ---------------------
3901
3902 1. Some minor internal re-organization to help with my DFA experiments.
3903
3904 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3905 didn't matter for the library itself when fully configured, but did matter
3906 when compiling without UCP support, or within Exim, where the ucp files are
3907 not imported.
3908
3909 3. Refactoring of the library code to split up the various functions into
3910 different source modules. The addition of the new DFA matching code (see
3911 below) to a single monolithic source would have made it really too
3912 unwieldy, quite apart from causing all the code to be include in a
3913 statically linked application, when only some functions are used. This is
3914 relevant even without the DFA addition now that patterns can be compiled in
3915 one application and matched in another.
3916
3917 The downside of splitting up is that there have to be some external
3918 functions and data tables that are used internally in different modules of
3919 the library but which are not part of the API. These have all had their
3920 names changed to start with "_pcre_" so that they are unlikely to clash
3921 with other external names.
3922
3923 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3924 a different (DFA) algorithm. Although it is slower than the original
3925 function, it does have some advantages for certain types of matching
3926 problem.
3927
3928 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3929 including restarting after a partial match.
3930
3931 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3932 defined when compiling for Windows was sent to me. I have put it into the
3933 code, though I have no means of testing or verifying it.
3934
3935 7. Added the pcre_refcount() auxiliary function.
3936
3937 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3938 match before or at the first newline in the subject string. In pcretest,
3939 the /f option on a pattern can be used to set this.
3940
3941 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3942 would behave wrongly. This has been present in PCRE since release 4.0.
3943
3944 10. A number of changes to the pcregrep command:
3945
3946 (a) Refactored how -x works; insert ^(...)$ instead of setting
3947 PCRE_ANCHORED and checking the length, in preparation for adding
3948 something similar for -w.
3949
3950 (b) Added the -w (match as a word) option.
3951
3952 (c) Refactored the way lines are read and buffered so as to have more
3953 than one at a time available.
3954
3955 (d) Implemented a pcregrep test script.
3956
3957 (e) Added the -M (multiline match) option. This allows patterns to match
3958 over several lines of the subject. The buffering ensures that at least
3959 8K, or the rest of the document (whichever is the shorter) is available
3960 for matching (and similarly the previous 8K for lookbehind assertions).
3961
3962 (f) Changed the --help output so that it now says
3963
3964 -w, --word-regex(p)
3965
3966 instead of two lines, one with "regex" and the other with "regexp"
3967 because that confused at least one person since the short forms are the
3968 same. (This required a bit of code, as the output is generated
3969 automatically from a table. It wasn't just a text change.)
3970
3971 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3972 option but starts with a hyphen. Could be a pattern or a path name
3973 starting with a hyphen, for instance.
3974
3975 (h) "-" can be given as a file name to represent stdin.
3976
3977 (i) When file names are being printed, "(standard input)" is used for
3978 the standard input, for compatibility with GNU grep. Previously
3979 "<stdin>" was used.
3980
3981 (j) The option --label=xxx can be used to supply a name to be used for
3982 stdin when file names are being printed. There is no short form.
3983
3984 (k) Re-factored the options decoding logic because we are going to add
3985 two more options that take data. Such options can now be given in four
3986 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3987
3988 (l) Added the -A, -B, and -C options for requesting that lines of context
3989 around matches be printed.
3990
3991 (m) Added the -L option to print the names of files that do not contain
3992 any matching lines, that is, the complement of -l.
3993
3994 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3995 continue to scan other files.
3996
3997 (o) The -s option was incorrectly implemented. For compatibility with other
3998 greps, it now suppresses the error message for a non-existent or non-
3999 accessible file (but not the return code). There is a new option called
4000 -q that suppresses the output of matching lines, which was what -s was
4001 previously doing.
4002
4003 (p) Added --include and --exclude options to specify files for inclusion
4004 and exclusion when recursing.
4005
4006 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4007 Hopefully, it now does.
4008
4009 12. Missing cast in pcre_study().
4010
4011 13. Added an "uninstall" target to the makefile.
4012
4013 14. Replaced "extern" in the function prototypes in Makefile.in with
4014 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4015 world, but is set differently for Windows.
4016
4017 15. Added a second compiling function called pcre_compile2(). The only
4018 difference is that it has an extra argument, which is a pointer to an
4019 integer error code. When there is a compile-time failure, this is set
4020 non-zero, in addition to the error test pointer being set to point to an
4021 error message. The new argument may be NULL if no error number is required
4022 (but then you may as well call pcre_compile(), which is now just a
4023 wrapper). This facility is provided because some applications need a
4024 numeric error indication, but it has also enabled me to tidy up the way
4025 compile-time errors are handled in the POSIX wrapper.
4026
4027 16. Added VPATH=.libs to the makefile; this should help when building with one
4028 prefix path and installing with another. (Or so I'm told by someone who
4029 knows more about this stuff than I do.)
4030
4031 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4032 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4033 match everything, including newlines. This is not POSIX-compatible, but
4034 somebody wanted the feature. From pcretest it can be activated by using
4035 both the P and the s flags.
4036
4037 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4038
4039 19. libpcre.pc was being incorrectly installed as executable.
4040
4041 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4042 it now also looks for '\r' so that it will work unmodified on Windows.
4043
4044 21. Added Google's contributed C++ wrapper to the distribution.
4045
4046 22. Added some untidy missing memory free() calls in pcretest, to keep
4047 Electric Fence happy when testing.
4048
4049
4050
4051 Version 5.0 13-Sep-04
4052 ---------------------
4053
4054 1. Internal change: literal characters are no longer packed up into items
4055 containing multiple characters in a single byte-string. Each character
4056 is now matched using a separate opcode. However, there may be more than one
4057 byte in the character in UTF-8 mode.
4058
4059 2. The pcre_callout_block structure has two new fields: pattern_position and
4060 next_item_length. These contain the offset in the pattern to the next match
4061 item, and its length, respectively.
4062
4063 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4064 insertion of callouts before each pattern item. Added the /C option to
4065 pcretest to make use of this.
4066
4067 4. On the advice of a Windows user, the lines
4068
4069 #if defined(_WIN32) || defined(WIN32)
4070 _setmode( _fileno( stdout ), 0x8000 );
4071 #endif /* defined(_WIN32) || defined(WIN32) */
4072
4073 have been added to the source of pcretest. This apparently does useful
4074 magic in relation to line terminators.
4075
4076 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4077 for the benefit of those environments where the "b" makes a difference.
4078
4079 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4080 to know about it. I have put a hack into configure.in that adds in code
4081 to set GCC=yes if CC=icc. This seems to end up at a point in the
4082 generated configure script that is early enough to affect the setting of
4083 compiler options, which is what is needed, but I have no means of testing
4084 whether it really works. (The user who reported this had patched the
4085 generated configure script, which of course I cannot do.)
4086
4087 LATER: After change 22 below (new libtool files), the configure script
4088 seems to know about icc (and also ecc). Therefore, I have commented out
4089 this hack in configure.in.
4090
4091 7. Added support for pkg-config (2 patches were sent in).
4092
4093 8. Negated POSIX character classes that used a combination of internal tables
4094 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4095 [[:^ascii]]. Typically, they would match almost any characters. The other
4096 POSIX classes were not broken in this way.
4097
4098 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4099 to find the match, as PCRE was deluded into thinking that the match had to
4100 start at the start point or following a newline. The same bug applied to
4101 patterns with negative forward assertions or any backward assertions
4102 preceding ".*" at the start, unless the pattern required a fixed first
4103 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4104
4105 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4106 starting at the last subject character, bytes beyond the end of the subject
4107 string were read.
4108
4109 11. Renamed the variable "class" as "classbits" to make life easier for C++
4110 users. (Previously there was a macro definition, but it apparently wasn't
4111 enough.)
4112
4113 12. Added the new field "tables" to the extra data so that tables can be passed
4114 in at exec time, or the internal tables can be re-selected. This allows
4115 a compiled regex to be saved and re-used at a later time by a different
4116 program that might have everything at different addresses.
4117
4118 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4119 -R library as well as a -L library.
4120
4121 14. The debugging options of pcretest (-d on the command line or D on a
4122 pattern) showed incorrect output for anything following an extended class
4123 that contained multibyte characters and which was followed by a quantifier.
4124
4125 15. Added optional support for general category Unicode character properties
4126 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4127 support. It adds about 90K to the size of the library. The meanings of the
4128 inbuilt class escapes such as \d and \s have NOT been changed.
4129
4130 16. Updated pcredemo.c to include calls to free() to release the memory for the
4131 compiled pattern.
4132
4133 17. The generated file chartables.c was being created in the source directory
4134 instead of in the building directory. This caused the build to fail if the
4135 source directory was different from the building directory, and was
4136 read-only.
4137
4138 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4139 file. No doubt somebody will tell me if they don't make sense... Also added
4140 Dan Mooney's comments about building on OpenVMS.
4141
4142 19. Added support for partial matching via the PCRE_PARTIAL option for
4143 pcre_exec() and the \P data escape in pcretest.
4144
4145 20. Extended pcretest with 3 new pattern features:
4146
4147 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4148 write the compiled pattern to the file whose name is "rest-of-line".
4149 This is a straight binary dump of the data, with the saved pointer to
4150 the character tables forced to be NULL. The study data, if any, is
4151 written too. After writing, pcretest reads a new pattern.
4152
4153 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4154 compiled pattern from the given file. There must not be any
4155 occurrences of "<" in the file name (pretty unlikely); if there are,
4156 pcretest will instead treat the initial "<" as a pattern delimiter.
4157 After reading in the pattern, pcretest goes on to read data lines as
4158 usual.
4159
4160 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4161 and 16-bit fields in a compiled pattern, to simulate a pattern that
4162 was compiled on a host of opposite endianness.
4163
4164 21. The pcre-exec() function can now cope with patterns that were compiled on
4165 hosts of opposite endianness, with this restriction:
4166
4167 As for any compiled expression that is saved and used later, the tables
4168 pointer field cannot be preserved; the extra_data field in the arguments
4169 to pcre_exec() should be used to pass in a tables address if a value
4170 other than the default internal tables were used at compile time.
4171
4172 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4173 now diagnosed as an error. Previously, most of the time, a negative number
4174 would have been treated as zero, but if in addition "ovector" was passed as
4175 NULL, a crash could occur.
4176
4177 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4178 new versions from the libtool 1.5 distribution (the last one is a copy of
4179 a file called libtool.m4). This seems to have fixed the need to patch
4180 "configure" to support Darwin 1.3 (which I used to do). However, I still
4181 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4182 workstation).
4183
4184 24. Changed the PCRE licence to be the more standard "BSD" licence.
4185
4186
4187 Version 4.5 01-Dec-03
4188 ---------------------
4189
4190 1. There has been some re-arrangement of the code for the match() function so
4191 that it can be compiled in a version that does not call itself recursively.
4192 Instead, it keeps those local variables that need separate instances for
4193 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4194 needs to "recurse". Keeping track of where control must go is done by means
4195 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4196 hide most of the details from the main code, and operates only if
4197 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4198 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4199 operating.
4200
4201 To make it easier for callers to provide specially tailored get/free
4202 functions for this usage, two new functions, pcre_stack_malloc, and
4203 pcre_stack_free, are used. They are always called in strict stacking order,
4204 and the size of block requested is always the same.
4205
4206 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4207 PCRE has been compiled to use the stack or the heap for recursion. The
4208 -C option of pcretest uses this to show which version is compiled.
4209
4210 A new data escape \S, is added to pcretest; it causes the amounts of store
4211 obtained and freed by both kinds of malloc/free at match time to be added
4212 to the output.
4213
4214 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4215 what's available on my current Linux desktop machine.
4216
4217 3. When matching a UTF-8 string, the test for a valid string at the start has
4218 been extended. If start_offset is not zero, PCRE now checks that it points
4219 to a byte that is the start of a UTF-8 character. If not, it returns
4220 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4221 this is necessary because there may be backward assertions in the pattern.
4222 When matching the same subject several times, it may save resources to use
4223 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4224
4225 4. The code for checking the validity of UTF-8 strings has been tightened so
4226 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4227 containing "overlong sequences".
4228
4229 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4230 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4231 should have been "&", but it just so happened that all the cases this let
4232 through by mistake were picked up later in the function.
4233
4234 6. I had used a variable called "isblank" - this is a C99 function, causing
4235 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4236
4237 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4238 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4239 so the version is shown; (c) stop "make test" echoing "./RunTest".
4240
4241 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4242
4243 9. The prototype for memmove() for systems that don't have it was using
4244 size_t, but the inclusion of the header that defines size_t was later. I've
4245 moved the #includes for the C headers earlier to avoid this.
4246
4247 10. Added some adjustments to the code to make it easier to compiler on certain
4248 special systems:
4249
4250 (a) Some "const" qualifiers were missing.
4251 (b) Added the macro EXPORT before all exported functions; by default this
4252 is defined to be empty.
4253 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4254 that it reads its output file name as an argument instead of writing
4255 to the standard output and assuming this can be redirected.
4256
4257 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4258 class containing characters with values greater than 255, PCRE compilation
4259 went into a loop.
4260
4261 12. A recursive reference to a subpattern that was within another subpattern
4262 that had a minimum quantifier of zero caused PCRE to crash. For example,
4263 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4264 recursion. If the recursively-called subpattern itself had a zero repeat,
4265 that was OK.
4266
4267 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4268 buffer into which it was copied (for escape processing) was still set at
4269 1024, so long lines caused crashes.
4270
4271 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4272 "internal error: code overflow...". This applied to any character class
4273 that was followed by a possessive quantifier.
4274
4275 15. Modified the Makefile to add libpcre.la as a prerequisite for
4276 libpcreposix.la because I was told this is needed for a parallel build to
4277 work.
4278
4279 16. If a pattern that contained .* following optional items at the start was
4280 studied, the wrong optimizing data was generated, leading to matching
4281 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4282 matching string must start with a or b or c. The correct conclusion for
4283 this pattern is that a match can start with any character.
4284
4285
4286 Version 4.4 13-Aug-03
4287 ---------------------
4288
4289 1. In UTF-8 mode, a character class containing characters with values between
4290 127 and 255 was not handled correctly if the compiled pattern was studied.
4291 In fixing this, I have also improved the studying algorithm for such
4292 classes (slightly).
4293
4294 2. Three internal functions had redundant arguments passed to them. Removal
4295 might give a very teeny performance improvement.
4296
4297 3. Documentation bug: the value of the capture_top field in a callout is *one
4298 more than* the number of the hightest numbered captured substring.
4299
4300 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4301 in incorrectly linking with a previously installed version. They now link
4302 explicitly with libpcre.la.
4303
4304 5. configure.in no longer needs to recognize Cygwin specially.
4305
4306 6. A problem in pcre.in for Windows platforms is fixed.
4307
4308 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4309 pcretest, it used to include the size of the study block as part of its
4310 output. Unfortunately, the structure contains a field that has a different
4311 size on different hardware architectures. This meant that the tests that
4312 showed this size failed. As the block is currently always of a fixed size,
4313 this information isn't actually particularly useful in pcretest output, so
4314 I have just removed it.
4315
4316 8. Three pre-processor statements accidentally did not start in column 1.
4317 Sadly, there are *still* compilers around that complain, even though
4318 standard C has not required this for well over a decade. Sigh.
4319
4320 9. In pcretest, the code for checking callouts passed small integers in the
4321 callout_data field, which is a void * field. However, some picky compilers
4322 complained about the casts involved for this on 64-bit systems. Now
4323 pcretest passes the address of the small integer instead, which should get
4324 rid of the warnings.
4325
4326 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4327 both compile and run time, and gives an error if an invalid UTF-8 sequence
4328 is found. There is a option for disabling this check in cases where the
4329 string is known to be correct and/or the maximum performance is wanted.
4330
4331 11. In response to a bug report, I changed one line in Makefile.in from
4332
4333 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4334 to
4335 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4336
4337 to look similar to other lines, but I have no way of telling whether this
4338 is the right thing to do, as I do not use Windows. No doubt I'll get told
4339 if it's wrong...
4340
4341
4342 Version 4.3 21-May-03
4343 ---------------------
4344
4345 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4346 Makefile.
4347
4348 2. Some refactoring to improve the quality of the code:
4349
4350 (i) The utf8_table... variables are now declared "const".
4351
4352 (ii) The code for \cx, which used the "case flipping" table to upper case
4353 lower case letters, now just substracts 32. This is ASCII-specific,
4354 but the whole concept of \cx is ASCII-specific, so it seems
4355 reasonable.
4356
4357 (iii) PCRE was using its character types table to recognize decimal and
4358 hexadecimal digits in the pattern. This is silly, because it handles
4359 only 0-9, a-f, and A-F, but the character types table is locale-
4360 specific, which means strange things might happen. A private
4361 table is now used for this - though it costs 256 bytes, a table is
4362 much faster than multiple explicit tests. Of course, the standard
4363 character types table is still used for matching digits in subject
4364 strings against \d.
4365
4366 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4367 ending in _t are). So I've renamed it as ESC_tee.
4368
4369 3. The first argument for regexec() in the POSIX wrapper should have been
4370 defined as "const".
4371
4372 4. Changed pcretest to use malloc() for its buffers so that they can be
4373 Electric Fenced for debugging.
4374
4375 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4376 to read one or more bytes before the start of the subject string. Often this
4377 had no effect on PCRE's behaviour, but in some circumstances it could
4378 provoke a segmentation fault.
4379
4380 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4381 to try to read one or more bytes before the start of the subject string.
4382
4383 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4384 UTF-8 support could misbehave in various ways if the subject string
4385 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4386 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4387 back over UTF-8 characters.)
4388
4389
4390 Version 4.2 14-Apr-03
4391 ---------------------
4392
4393 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4394
4395 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4396 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4397 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4398 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4399 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4400 and BUILD_EXEEXT
4401 Note: automatic setting of the BUILD variables is not yet working
4402 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4403 compile-time but not at link-time
4404 [LINK]: use for linking executables only
4405 make different versions for Windows and non-Windows
4406 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4407 libraries
4408 [LINK_FOR_BUILD]: new variable
4409 [OBJEXT]: use throughout
4410 [EXEEXT]: use throughout
4411 <winshared>: new target
4412 <wininstall>: new target
4413 <dftables.o>: use native compiler
4414 <dftables>: use native linker
4415 <install>: handle Windows platform correctly
4416 <clean>: ditto
4417 <check>: ditto
4418 copy DLL to top builddir before testing
4419
4420 As part of these changes, -no-undefined was removed again. This was reported
4421 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4422 in any case.
4423
4424 3. Some tidies to get rid of compiler warnings:
4425
4426 . In the match_data structure, match_limit was an unsigned long int, whereas
4427 match_call_count was an int. I've made them both unsigned long ints.
4428
4429 . In pcretest the fact that a const uschar * doesn't automatically cast to
4430 a void * provoked a warning.
4431
4432 . Turning on some more compiler warnings threw up some "shadow" variables
4433 and a few more missing casts.
4434
4435 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4436 option, a class that contained a single character with a value between 128
4437 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4438
4439 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4440 option, a class that contained several characters, but with at least one
4441 whose value was between 128 and 255 caused PCRE to crash.
4442
4443
4444 Version 4.1 12-Mar-03
4445 ---------------------
4446
4447 1. Compiling with gcc -pedantic found a couple of places where casts were
4448 needed, and a string in dftables.c that was longer than standard compilers are
4449 required to support.
4450
4451 2. Compiling with Sun's compiler found a few more places where the code could
4452 be tidied up in order to avoid warnings.
4453
4454 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4455 first of these names is deprecated in the latest Autoconf in favour of the name
4456 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4457 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4458 analogy I have changed it to CFLAGS_FOR_BUILD.
4459
4460 4. Added -no-undefined to the linking command in the Makefile, because this is
4461 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4462 linking step for the pcreposix library.
4463
4464 5. PCRE was failing to diagnose the case of two named groups with the same
4465 name.
4466
4467 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4468 literal character that is needed in the subject for a match, and scans along to
4469 ensure that it is present before embarking on the full matching process. This
4470 saves time in cases of nested unlimited repeats that are never going to match.
4471 Problem: the scan can take a lot of time if the subject is very long (e.g.
4472 megabytes), thus penalizing straightforward matches. It is now done only if the
4473 amount of subject to be scanned is less than 1000 bytes.
4474
4475 7. A lesser problem with the same optimization is that it was recording the
4476 first character of an anchored pattern as "needed", thus provoking a search
4477 right along the subject, even when the first match of the pattern was going to
4478 fail. The "needed" character is now not set for anchored patterns, unless it
4479 follows something in the pattern that is of non-fixed length. Thus, it still
4480 fulfils its original purpose of finding quick non-matches in cases of nested
4481 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4482
4483
4484 Version 4.0 17-Feb-03
4485 ---------------------
4486
4487 1. If a comment in an extended regex that started immediately after a meta-item
4488 extended to the end of string, PCRE compiled incorrect data. This could lead to
4489 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4490
4491 2. Moved to autoconf 2.53 and libtool 1.4.2.
4492
4493 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4494 the special perltest8 script is no longer needed - all the tests can be run
4495 from a single perltest script.
4496
4497 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4498 by \s. It has now been removed in PCRE. This means it isn't recognized as
4499 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4500 class [:space:] *does* include VT, thereby creating a mess.
4501
4502 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4503 space and tab.
4504
4505 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4506 its new features into the main test script, reducing the number of scripts.
4507
4508 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4509 were backward compatible, and made the (?i) apply to the whole pattern, as if
4510 /i were given. Now it behaves more logically, and applies the option setting
4511 only to what follows. PCRE has been changed to follow suit. However, if it
4512 finds options settings right at the start of the pattern, it extracts them into
4513 the global options, as before. Thus, they show up in the info data.
4514
4515 8. Added support for the \Q...\E escape sequence. Characters in between are
4516 treated as literals. This is slightly different from Perl in that $ and @ are
4517 also handled as literals inside the quotes. In Perl, they will cause variable
4518 interpolation. Note the following examples:
4519
4520 Pattern PCRE matches Perl matches
4521
4522 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4523 \Qabc\$xyz\E abc\$xyz abc\$xyz
4524 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4525
4526 For compatibility with Perl, \Q...\E sequences are recognized inside character
4527 classes as well as outside them.
4528
4529 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4530 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4531 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4532 signed/unsigned warnings.
4533
4534 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4535 option for pcretest, so I've replaced it by a simple function that does just
4536 that job.
4537
4538 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4539 "pcregrep -".
4540
4541 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4542 Java package. This provides some syntactic sugar for simple cases of what my
4543 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4544 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4545 item, you can use this simplified notation. Note that only makes sense with
4546 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4547 greediness, whatever the setting of the PCRE_UNGREEDY option.
4548
4549 13. A change of greediness default within a pattern was not taking effect at
4550 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4551 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4552 was abstracted outside.
4553
4554 14. PCRE now supports the \G assertion. It is true when the current matching
4555 position is at the start point of the match. This differs from \A when the
4556 starting offset is non-zero. Used with the /g option of pcretest (or similar
4557 code), it works in the same way as it does for Perl's /g option. If all
4558 alternatives of a regex begin with \G, the expression is anchored to the start
4559 match position, and the "anchored" flag is set in the compiled expression.
4560
4561 15. Some bugs concerning the handling of certain option changes within patterns
4562 have been fixed. These applied to options other than (?ims). For example,
4563 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4564 been the other way round. Some of this was related to change 7 above.
4565
4566 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4567 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4568 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4569 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4570
4571 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4572 mode. Unlike ".", it always matches newline, whatever the setting of
4573 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4574 assertions. Perl allows it, but it doesn't (in general) work because it can't
4575 calculate the length of the lookbehind. At least, that's the case for Perl
4576 5.8.0 - I've been told they are going to document that it doesn't work in
4577 future.
4578
4579 18. Added an error diagnosis for escapes that PCRE does not support: these are
4580 \L, \l, \N, \P, \p, \U, \u, and \X.
4581
4582 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4583 reading past the end of the pattern in cases such as /[abcd/.
4584
4585 20. PCRE was getting more memory than necessary for patterns with classes that
4586 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4587
4588 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4589 compiling PCRE for use with Virtual Pascal.
4590
4591 22. Small fix to the Makefile to make it work properly if the build is done
4592 outside the source tree.
4593
4594 23. Added a new extension: a condition to go with recursion. If a conditional
4595 subpattern starts with (?(R) the "true" branch is used if recursion has
4596 happened, whereas the "false" branch is used only at the top level.
4597
4598 24. When there was a very long string of literal characters (over 255 bytes
4599 without UTF support, over 250 bytes with UTF support), the computation of how
4600 much memory was required could be incorrect, leading to segfaults or other
4601 strange effects.
4602
4603 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4604 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4605 there was a subsequent back reference to those brackets. This meant that, for
4606 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4607 possible to check for precisely this case. All we can do is abandon the
4608 optimization if .* occurs inside capturing brackets when there are any back
4609 references whatsoever. (See below for a better fix that came later.)
4610
4611 26. The handling of the optimization for finding the first character of a
4612 non-anchored pattern, and for finding a character that is required later in the
4613 match were failing in some cases. This didn't break the matching; it just
4614 failed to optimize when it could. The way this is done has been re-implemented.
4615
4616 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4617
4618 28. Added a new feature that provides some of the functionality that Perl
4619 provides with (?{...}). The facility is termed a "callout". The way it is done
4620 in PCRE is for the caller to provide an optional function, by setting
4621 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4622 global variable. By default it is unset, which disables all calling out. To get
4623 the function called, the regex must include (?C) at appropriate points. This
4624 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4625 This provides a means of identifying different callout points. When PCRE
4626 reaches such a point in the regex, if pcre_callout has been set, the external
4627 function is called. It is provided with data in a structure called
4628 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4629 matching continues; if it returns a non-zero value, the match at the current
4630 point fails. However, backtracking will occur if possible. [This was changed
4631 later and other features added - see item 49 below.]
4632
4633 29. pcretest is upgraded to test the callout functionality. It provides a
4634 callout function that displays information. By default, it shows the start of
4635 the match and the current position in the text. There are some new data escapes
4636 to vary what happens:
4637
4638 \C+ in addition, show current contents of captured substrings
4639 \C- do not supply a callout function
4640 \C!n return 1 when callout number n is reached
4641 \C!n!m return 1 when callout number n is reached for the mth time
4642
4643 30. If pcregrep was called with the -l option and just a single file name, it
4644 output "<stdin>" if a match was found, instead of the file name.
4645
4646 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4647 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4648 pcre_exec(). This saves a malloc/free per call. The default value of
4649 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4650 when configuring.
4651
4652 32. The default maximum size of a compiled pattern is 64K. There have been a
4653 few cases of people hitting this limit. The code now uses macros to handle the
4654 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4655 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4656 configuring. Tests 2 and 5 work only with 2-byte links because they output
4657 debugging information about compiled patterns.
4658
4659 33. Internal code re-arrangements:
4660
4661 (a) Moved the debugging function for printing out a compiled regex into
4662 its own source file (printint.c) and used #include to pull it into
4663 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4664 separate copies.
4665
4666 (b) Defined the list of op-code names for debugging as a macro in
4667 internal.h so that it is next to the definition of the opcodes.
4668
4669 (c) Defined a table of op-code lengths for simpler skipping along compiled
4670 code. This is again a macro in internal.h so that it is next to the
4671 definition of the opcodes.
4672
4673 34. Added support for recursive calls to individual subpatterns, along the
4674 lines of Robin Houston's patch (but implemented somewhat differently).
4675
4676 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4677 allow it to read and process whole directories in Win32. This code was
4678 contributed by Lionel Fourquaux; it has not been tested by me.
4679
4680 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4681 used to name a group. Names consist of alphanumerics and underscores, and must
4682 be unique. Back references use the syntax (?P=name) and recursive calls use
4683 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4684 numbers. The function pcre_fullinfo() can be used after compilation to extract
4685 a name/number map. There are three relevant calls:
4686
4687 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4688 PCRE_INFO_NAMECOUNT yields the number of entries
4689 PCRE_INFO_NAMETABLE yields a pointer to the map.
4690
4691 The map is a vector of fixed-size entries. The size of each entry depends on
4692 the length of the longest name used. The first two bytes of each entry are the
4693 group number, most significant byte first. There follows the corresponding
4694 name, zero terminated. The names are in alphabetical order.
4695
4696 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4697 case instead of 255. Making it the same both with and without UTF-8 support
4698 means that the same test output works with both.
4699
4700 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4701 calling malloc() with a zero argument.
4702
4703 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4704 optimization. I've improved things by keeping a bitmap of backreferences with
4705 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4706 fact referenced, the optimization can be applied. It is unlikely that a
4707 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4708 the match to follow \n) will appear inside brackets with a number greater than
4709 31, but if it does, any back reference > 31 suppresses the optimization.
4710
4711 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4712 of disabling numbered capturing parentheses. Any opening parenthesis that is
4713 not followed by ? behaves as if it were followed by ?: but named parentheses
4714 can still be used for capturing (and they will acquire numbers in the usual
4715 way).
4716
4717 41. Redesigned the return codes from the match() function into yes/no/error so
4718 that errors can be passed back from deep inside the nested calls. A malloc
4719 failure while inside a recursive subpattern call now causes the
4720 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4721
4722 42. It is now possible to set a limit on the number of times the match()
4723 function is called in a call to pcre_exec(). This facility makes it possible to
4724 limit the amount of recursion and backtracking, though not in a directly
4725 obvious way, because the match() function is used in a number of different
4726 circumstances. The count starts from zero for each position in the subject
4727 string (for non-anchored patterns). The default limit is, for compatibility, a
4728 large number, namely 10 000 000. You can change this in two ways:
4729
4730 (a) When configuring PCRE before making, you can use --with-match-limit=n
4731 to set a default value for the compiled library.
4732
4733 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4734 a different value is set. See 45 below.
4735
4736 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4737
4738 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4739 of things that can be changed at compile time. The first argument specifies
4740 what is wanted and the second points to where the information is to be placed.
4741 The current list of available information is:
4742
4743 PCRE_CONFIG_UTF8
4744
4745 The output is an integer that is set to one if UTF-8 support is available;
4746 otherwise it is set to zero.
4747
4748 PCRE_CONFIG_NEWLINE
4749
4750 The output is an integer that it set to the value of the code that is used for
4751 newline. It is either LF (10) or CR (13).
4752
4753 PCRE_CONFIG_LINK_SIZE
4754
4755 The output is an integer that contains the number of bytes used for internal
4756 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4757
4758 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4759
4760 The output is an integer that contains the threshold above which the POSIX
4761 interface uses malloc() for output vectors. See item 31 above.
4762
4763 PCRE_CONFIG_MATCH_LIMIT
4764
4765 The output is an unsigned integer that contains the default limit of the number
4766 of match() calls in a pcre_exec() execution. See 42 above.
4767
4768 44. pcretest has been upgraded by the addition of the -C option. This causes it
4769 to extract all the available output from the new pcre_config() function, and to
4770 output it. The program then exits immediately.
4771
4772 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4773 order to support additional features. One way would have been to define
4774 pcre_exec2() (for example) with extra arguments, but this would not have been
4775 extensible, and would also have required all calls to the original function to
4776 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4777 is used for passing in "extra" data from pcre_study().
4778
4779 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4780 contains the following fields:
4781
4782 flags a bitmap indicating which of the following fields are set
4783 study_data opaque data from pcre_study()
4784 match_limit a way of specifying a limit on match() calls for a specific
4785 call to pcre_exec()
4786 callout_data data for callouts (see 49 below)
4787
4788 The flag bits are also defined in pcre.h, and are
4789
4790 PCRE_EXTRA_STUDY_DATA
4791 PCRE_EXTRA_MATCH_LIMIT
4792 PCRE_EXTRA_CALLOUT_DATA
4793
4794 The pcre_study() function now returns one of these new pcre_extra blocks, with
4795 the actual study data pointed to by the study_data field, and the
4796 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4797 before. That is, this change is entirely upwards-compatible and requires no
4798 change to existing code.
4799
4800 If you want to pass in additional data to pcre_exec(), you can either place it
4801 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4802 block.
4803
4804 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4805 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4806 times with different match limits, until it finds the minimum value needed for
4807 pcre_exec() to complete. The value is then output. This can be instructive; for
4808 most simple matches the number is quite small, but for pathological cases it
4809 gets very large very quickly.
4810
4811 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4812 returns the size of the data block pointed to by the study_data field in a
4813 pcre_extra block, that is, the value that was passed as the argument to
4814 pcre_malloc() when PCRE was getting memory in which to place the information
4815 created by pcre_study(). The fourth argument should point to a size_t variable.
4816 pcretest has been extended so that this information is shown after a successful
4817 pcre_study() call when information about the compiled regex is being displayed.
4818
4819 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4820 because what follows is always an absolute path. (Later: it turns out that this
4821 is more than cosmetic for MinGW, because it doesn't like empty path
4822 components.)
4823
4824 49. Some changes have been made to the callout feature (see 28 above):
4825
4826 (i) A callout function now has three choices for what it returns:
4827
4828 0 => success, carry on matching
4829 > 0 => failure at this point, but backtrack if possible
4830 < 0 => serious error, return this value from pcre_exec()
4831
4832 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4833 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4834 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4835 use by callout functions. It will never be used by PCRE itself.
4836
4837 (ii) The pcre_extra structure (see 45 above) has a void * field called
4838 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4839 pcre_callout_block structure has a field of the same name. The contents of
4840 the field passed in the pcre_extra structure are passed to the callout
4841 function in the corresponding field in the callout block. This makes it
4842 easier to use the same callout-containing regex from multiple threads. For
4843 testing, the pcretest program has a new data escape
4844
4845 \C*n pass the number n (may be negative) as callout_data
4846
4847 If the callout function in pcretest receives a non-zero value as
4848 callout_data, it returns that value.
4849
4850 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4851 there were some redundant $(CFLAGS) in commands that are now specified as
4852 $(LINK), which already includes $(CFLAGS).
4853
4854 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4855 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4856 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4857 one-byte characters throughout. Note that case-insensitive matching applies
4858 only to characters whose values are less than 256. PCRE doesn't support the
4859 notion of cases for higher-valued characters.
4860
4861 (i) A character class whose characters are all within 0-255 is handled as
4862 a bit map, and the map is inverted for negative classes. Previously, a
4863 character > 255 always failed to match such a class; however it should
4864 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4865
4866 (ii) A negated character class with a single character < 255 is coded as
4867 "not this character" (OP_NOT). This wasn't working properly when the test
4868 character was multibyte, either singly or repeated.
4869
4870 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4871 mode, for example: \x{100}{2,3}.
4872
4873 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4874 singly or repeated) now correctly test multibyte characters. However,
4875 PCRE doesn't recognize any characters with values greater than 255 as
4876 digits, spaces, or word characters. Such characters always match \D, \S,
4877 and \W, and never match \d, \s, or \w.
4878
4879 (v) Classes may now contain characters and character ranges with values
4880 greater than 255. For example: [ab\x{100}-\x{400}].
4881
4882 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4883 PCRE in UTF-8 mode.
4884
4885 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4886 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4887 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4888 value.)
4889
4890 53. The single man page has become too large. I have therefore split it up into
4891 a number of separate man pages. These also give rise to individual HTML pages;
4892 these are now put in a separate directory, and there is an index.html page that
4893 lists them all. Some hyperlinking between the pages has been installed.
4894
4895 54. Added convenience functions for handling named capturing parentheses.
4896
4897 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4898 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4899 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4900 are faulted.
4901
4902 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4903 calling configure. These values are used when compiling the dftables.c program
4904 which is run to generate the source of the default character tables. They
4905 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4906 you will need to set these values.
4907
4908 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4909
4910
4911 Version 3.9 02-Jan-02
4912 ---------------------
4913
4914 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4915
4916 2. If --disable-static was given, the building process failed when trying to
4917 build pcretest and pcregrep. (For some reason it was using libtool to compile
4918 them, which is not right, as they aren't part of the library.)
4919
4920
4921 Version 3.8 18-Dec-01
4922 ---------------------
4923
4924 1. The experimental UTF-8 code was completely screwed up. It was packing the
4925 bytes in the wrong order. How dumb can you get?
4926
4927
4928 Version 3.7 29-Oct-01
4929 ---------------------
4930
4931 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4932 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4933 this didn't happen under Solaris 8, where I normally test things.
4934
4935 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4936 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4937 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4938 this fix an hour or so after the initial 3.7 release.)
4939
4940
4941 Version 3.6 23-Oct-01
4942 ---------------------
4943
4944 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4945 offsets passed as NULL with zero offset count.
4946
4947 2. The config.guess and config.sub files had not been updated when I moved to
4948 the latest autoconf.
4949
4950
4951 Version 3.5 15-Aug-01
4952 ---------------------
4953
4954 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4955 had been forgotten.
4956
4957 2. By using declared but undefined structures, we can avoid using "void"
4958 definitions in pcre.h while keeping the internal definitions of the structures
4959 private.
4960
4961 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4962 user point of view, this means that both static and shared libraries are built
4963 by default, but this can be individually controlled. More of the work of
4964 handling this static/shared cases is now inside libtool instead of PCRE's make
4965 file.
4966
4967 4. The pcretest utility is now installed along with pcregrep because it is
4968 useful for users (to test regexs) and by doing this, it automatically gets
4969 relinked by libtool. The documentation has been turned into a man page, so
4970 there are now .1, .txt, and .html versions in /doc.
4971
4972 5. Upgrades to pcregrep:
4973 (i) Added long-form option names like gnu grep.
4974 (ii) Added --help to list all options with an explanatory phrase.
4975 (iii) Added -r, --recursive to recurse into sub-directories.
4976 (iv) Added -f, --file to read patterns from a file.
4977
4978 6. pcre_exec() was referring to its "code" argument before testing that
4979 argument for NULL (and giving an error if it was NULL).
4980
4981 7. Upgraded Makefile.in to allow for compiling in a different directory from
4982 the source directory.
4983
4984 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4985 options bits, the pointer it was passed was to an int instead of to an unsigned
4986 long int. This mattered only on 64-bit systems.
4987
4988 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4989 generated) instead of pcre.in, which it its source. Also made the same change
4990 in several of the .c files.
4991
4992 10. A new release of gcc defines printf() as a macro, which broke pcretest
4993 because it had an ifdef in the middle of a string argument for printf(). Fixed
4994 by using separate calls to printf().
4995
4996 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4997 script, to force use of CR or LF instead of \n in the source. On non-Unix
4998 systems, the value can be set in config.h.
4999
5000 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5001 absolute limit. Changed the text of the error message to make this clear, and
5002 likewise updated the man page.
5003
5004 13. The limit of 99 on the number of capturing subpatterns has been removed.
5005 The new limit is 65535, which I hope will not be a "real" limit.
5006
5007
5008 Version 3.4 22-Aug-00
5009 ---------------------
5010
5011 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5012
5013 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5014
5015
5016 Version 3.3 01-Aug-00
5017 ---------------------
5018
5019 1. If an octal character was given, but the value was greater than \377, it
5020 was not getting masked to the least significant bits, as documented. This could
5021 lead to crashes in some systems.
5022
5023 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5024 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5025
5026 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5027 These just pass their arguments on to (pcre_free)(), but they are provided
5028 because some uses of PCRE bind it to non-C systems that can call its functions,
5029 but cannot call free() or pcre_free() directly.
5030
5031 4. Add "make test" as a synonym for "make check". Corrected some comments in
5032 the Makefile.
5033
5034 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5035 Makefile.
5036
5037 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5038 command called pgrep for grepping around the active processes.
5039
5040 7. Added the beginnings of support for UTF-8 character strings.
5041
5042 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5043 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5044 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5045 out for the ar command.)
5046
5047
5048 Version 3.2 12-May-00
5049 ---------------------
5050
5051 This is purely a bug fixing release.
5052
5053 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5054 of ZA. This was just one example of several cases that could provoke this bug,
5055 which was introduced by change 9 of version 2.00. The code for breaking
5056 infinite loops after an iteration that matches an empty string was't working
5057 correctly.
5058
5059 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5060 when matched against abbab (for example). After matching an empty string, it
5061 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5062 caused it to match further down the string than it should.
5063
5064 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5065 was there because it doesn't seem to be needed, and it causes trouble on some
5066 systems, as it is not a Standard C header. It has been removed.
5067
5068 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5069 were reported on the Macintosh. The changes were from
5070
5071 while ((c = *(++ptr)) != 0 && c != '\n');
5072 to
5073 while ((c = *(++ptr)) != 0 && c != '\n') ;
5074
5075 Totally extraordinary, but if that's what it takes...
5076
5077 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5078 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5079 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5080 assumes the way PCRE uses memmove() (always moving upwards).
5081
5082 6. PCRE is being used in one environment where strchr() is not available. There
5083 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5084 faster code anyway.
5085
5086
5087 Version 3.1 09-Feb-00
5088 ---------------------
5089
5090 The only change in this release is the fixing of some bugs in Makefile.in for
5091 the "install" target:
5092
5093 (1) It was failing to install pcreposix.h.
5094
5095 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5096
5097
5098 Version 3.0 01-Feb-00
5099 ---------------------
5100
5101 1. Add support for the /+ modifier to perltest (to output $` like it does in
5102 pcretest).
5103
5104 2. Add support for the /g modifier to perltest.
5105
5106 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5107 matches null strings.
5108
5109 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5110 pattern. Perl treats empty patterns specially - it reuses the most recent
5111 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5112 effect.
5113
5114 5. The POSIX interface was broken in that it was just handing over the POSIX
5115 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5116 required a bigger vector, with some working space on the end. This means that
5117 the POSIX wrapper now has to get and free some memory, and copy the results.
5118
5119 6. Added some simple autoconf support, placing the test data and the
5120 documentation in separate directories, re-organizing some of the
5121 information files, and making it build pcre-config (a GNU standard). Also added
5122 libtool support for building PCRE as a shared library, which is now the
5123 default.
5124
5125 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5126 09 are not valid octal constants. Single digits will be used for minor values
5127 less than 10.
5128
5129 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5130 existing programs that set these in the POSIX interface can use PCRE without
5131 modification.
5132
5133 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5134 return all that pcre_info() returns, plus additional data. The pcre_info()
5135 function is retained for compatibility, but is considered to be obsolete.
5136
5137 10. Added experimental recursion feature (?R) to handle one common case that
5138 Perl 5.6 will be able to do with (?p{...}).
5139
5140 11. Added support for POSIX character classes like [:alpha:], which Perl is
5141 adopting.
5142
5143
5144 Version 2.08 31-Aug-99
5145 ----------------------
5146
5147 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5148 trying to match at the startoffset position, but instead was moving forward to
5149 the next newline as if a previous match had failed.
5150
5151 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5152 and could get into a loop if a null string was matched other than at the start
5153 of the subject.
5154
5155 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5156 be distinguished at compile time, and for completeness also added PCRE_DATE.
5157
5158 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5159 in GnuWin32 environments.
5160
5161
5162 Version 2.07 29-Jul-99
5163 ----------------------
5164
5165 1. The documentation is now supplied in plain text form and HTML as well as in
5166 the form of man page sources.
5167
5168 2. C++ compilers don't like assigning (void *) values to other pointer types.
5169 In particular this affects malloc(). Although there is no problem in Standard
5170 C, I've put in casts to keep C++ compilers happy.
5171
5172 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5173 should be (const char *).
5174
5175 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5176 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5177 However, I haven't made this a standard facility. The documentation doesn't
5178 mention it, and the Makefile doesn't support it.
5179
5180 5. The Makefile now contains an "install" target, with editable destinations at
5181 the top of the file. The pcretest program is not installed.
5182
5183 6. pgrep -V now gives the PCRE version number and date.
5184
5185 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5186 causing the entire string to be ignored, instead of just the last character.
5187
5188 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5189 non-matching string, it can take a very, very long time, even for strings of
5190 quite modest length, because of the nested recursion. PCRE now does better in
5191 some of these cases. It does this by remembering the last required literal
5192 character in the pattern, and pre-searching the subject to ensure it is present
5193 before running the real match. In other words, it applies a heuristic to detect
5194 some types of certain failure quickly, and in the above example, if presented
5195 with a string that has no trailing " it gives "no match" very quickly.
5196
5197 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5198 other alternatives are tried instead.
5199
5200
5201 Version 2.06 09-Jun-99
5202 ----------------------
5203
5204 1. Change pcretest's output for amount of store used to show just the code
5205 space, because the remainder (the data block) varies in size between 32-bit and
5206 64-bit systems.
5207
5208 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5209 start matching at. This allows lookbehinds to work when searching for multiple
5210 occurrences in a string.
5211
5212 3. Added additional options to pcretest for testing multiple occurrences:
5213
5214 /+ outputs the rest of the string that follows a match
5215 /g loops for multiple occurrences, using the new startoffset argument
5216 /G loops for multiple occurrences by passing an incremented pointer
5217
5218 4. PCRE wasn't doing the "first character" optimization for patterns starting
5219 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5220 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5221 the letter 'x'. On long subject strings, this gives a significant speed-up.
5222
5223
5224 Version 2.05 21-Apr-99
5225 ----------------------
5226
5227 1. Changed the type of magic_number from int to long int so that it works
5228 properly on 16-bit systems.
5229
5230 2. Fixed a bug which caused patterns starting with .* not to work correctly
5231 when the subject string contained newline characters. PCRE was assuming
5232 anchoring for such patterns in all cases, which is not correct because .* will
5233 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5234 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5235 must be retried after every newline in the subject.
5236
5237
5238 Version 2.04 18-Feb-99
5239 ----------------------
5240
5241 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5242 computation of the store needed to hold the pattern was incorrect (too large).
5243 If such patterns were nested a few deep, this could multiply and become a real
5244 problem.
5245
5246 2. Added /M option to pcretest to show the memory requirement of a specific
5247 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5248
5249 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5250 compiled in such a way that the backtracking after subsequent failure was
5251 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5252 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5253
5254
5255 Version 2.03 02-Feb-99
5256 ----------------------
5257
5258 1. Fixed typo and small mistake in man page.
5259
5260 2. Added 4th condition (GPL supersedes if conflict) and created separate
5261 LICENCE file containing the conditions.
5262
5263 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5264 Perl, that is the internal \ allows the delimiter to be included in the
5265 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5266 the final delimiter, add \ to the end of the pattern (to test the error).
5267
5268 4. Added the convenience functions for extracting substrings after a successful
5269 match. Updated pcretest to make it able to test these functions.
5270
5271
5272 Version 2.02 14-Jan-99
5273 ----------------------
5274
5275 1. Initialized the working variables associated with each extraction so that
5276 their saving and restoring doesn't refer to uninitialized store.
5277
5278 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5279 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5280 fix the problem.
5281
5282 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5283 calls, and wasn't printing the correct value for compiling calls. Increased the
5284 default value of LOOPREPEAT, and the number of significant figures in the
5285 times.
5286
5287 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5288
5289 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5290 a building problem on Windows NT with a FAT file system.
5291
5292
5293 Version 2.01 21-Oct-98
5294 ----------------------
5295
5296 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5297 to character tables built by pcre_maketables() in the current locale. If NULL
5298 is passed, the default tables are used.
5299
5300
5301 Version 2.00 24-Sep-98
5302 ----------------------
5303
5304 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5305 it any more.
5306
5307 2. Allow quantification of (?>) groups, and make it work correctly.
5308
5309 3. The first character computation wasn't working for (?>) groups.
5310
5311 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5312 end of the subject) and add 5.005's \z, which really does match only at the
5313 very end of the subject.
5314
5315 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5316
5317 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5318 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5319 localized options. All options to pcre_study() were also removed.
5320
5321 7. Add other new features from 5.005:
5322
5323 $(?<= positive lookbehind
5324 $(?<! negative lookbehind
5325 (?imsx-imsx) added the unsetting capability
5326 such a setting is global if at outer level; local otherwise
5327 (?imsx-imsx:) non-capturing groups with option setting
5328 (?(cond)re|re) conditional pattern matching
5329
5330 A backreference to itself in a repeated group matches the previous
5331 captured string.
5332
5333 8. General tidying up of studying (both automatic and via "study")
5334 consequential on the addition of new assertions.
5335
5336 9. As in 5.005, unlimited repeated groups that could match an empty substring
5337 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5338 runtime if any iteration does actually match an empty substring.
5339
5340 10. Include the RunTest script in the distribution.
5341
5342 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5343 discrepancies, some of which were old and were also with respect to 5.004. They
5344 have now been fixed.
5345
5346
5347 Version 1.09 28-Apr-98
5348 ----------------------
5349
5350 1. A negated single character class followed by a quantifier with a minimum
5351 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5352 program crashes, or just wrong answers. This did not apply to negated classes
5353 containing more than one character, or to minima other than one.
5354
5355
5356 Version 1.08 27-Mar-98
5357 ----------------------
5358
5359 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5360
5361 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5362 latter must appear before anything that relies on it in the pattern.
5363
5364
5365 Version 1.07 16-Feb-98
5366 ----------------------
5367
5368 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5369 repeat of a potentially empty string).
5370
5371
5372 Version 1.06 23-Jan-98
5373 ----------------------
5374
5375 1. Added Markus Oberhumer's little patches for C++.
5376
5377 2. Literal strings longer than 255 characters were broken.
5378
5379
5380 Version 1.05 23-Dec-97
5381 ----------------------
5382
5383 1. Negated character classes containing more than one character were failing if
5384 PCRE_CASELESS was set at run time.
5385
5386
5387 Version 1.04 19-Dec-97
5388 ----------------------
5389
5390 1. Corrected the man page, where some "const" qualifiers had been omitted.
5391
5392 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5393 input syntax.
5394
5395 3. Fixed memory leak which occurred when a regex with back references was
5396 matched with an offsets vector that wasn't big enough. The temporary memory
5397 that is used in this case wasn't being freed if the match failed.
5398
5399 4. Tidied pcretest to ensure it frees memory that it gets.
5400
5401 5. Temporary memory was being obtained in the case where the passed offsets
5402 vector was exactly big enough.
5403
5404 6. Corrected definition of offsetof() from change 5 below.
5405
5406 7. I had screwed up change 6 below and broken the rules for the use of
5407 setjmp(). Now fixed.
5408
5409
5410 Version 1.03 18-Dec-97
5411 ----------------------
5412
5413 1. A erroneous regex with a missing opening parenthesis was correctly
5414 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5415 on some systems.
5416
5417 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5418 it was reported that one broken compiler failed on the former because "code" is
5419 also an independent variable.
5420
5421 3. The erroneous regex a[]b caused an array overrun reference.
5422
5423 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5424 fail on data ending with that character. (It was going on too far, and checking
5425 the next character, typically a binary zero.) This was specific to the
5426 optimized code for single-character negative classes.
5427
5428 5. Added a contributed patch from the TIN world which does the following:
5429
5430 + Add an undef for memmove, in case the the system defines a macro for it.
5431
5432 + Add a definition of offsetof(), in case there isn't one. (I don't know
5433 the reason behind this - offsetof() is part of the ANSI standard - but
5434 it does no harm).
5435
5436 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5437 most of the places where whitespace preceded '#'. I have given up and
5438 allowed the remaining 2 cases to be at the margin.
5439
5440 + Rename some variables in pcre to eliminate shadowing. This seems very
5441 pedantic, but does no harm, of course.
5442
5443 6. Moved the call to setjmp() into its own function, to get rid of warnings
5444 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5445
5446 7. Constructs such as \d{8,} were compiling into the equivalent of
5447 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5448 outcome, but in this particular case used more store than had been allocated,
5449 which caused the bug to be discovered because it threw up an internal error.
5450
5451 8. The debugging code in both pcre and pcretest for outputting the compiled
5452 form of a regex was going wrong in the case of back references followed by
5453 curly-bracketed repeats.
5454
5455
5456 Version 1.02 12-Dec-97
5457 ----------------------
5458
5459 1. Typos in pcre.3 and comments in the source fixed.
5460
5461 2. Applied a contributed patch to get rid of places where it used to remove
5462 'const' from variables, and fixed some signed/unsigned and uninitialized
5463 variable warnings.
5464
5465 3. Added the "runtest" target to Makefile.
5466
5467 4. Set default compiler flag to -O2 rather than just -O.
5468
5469
5470 Version 1.01 19-Nov-97
5471 ----------------------
5472
5473 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5474 like /([ab]*)*/, that is, for classes with more than one character in them.
5475
5476 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5477 as /((?>a*))*/ (a PCRE_EXTRA facility).
5478
5479
5480 Version 1.00 18-Nov-97
5481 ----------------------
5482
5483 1. Added compile-time macros to support systems such as SunOS4 which don't have
5484 memmove() or strerror() but have other things that can be used instead.
5485
5486 2. Arranged that "make clean" removes the executables.
5487
5488
5489 Version 0.99 27-Oct-97
5490 ----------------------
5491
5492 1. Fixed bug in code for optimizing classes with only one character. It was
5493 initializing a 32-byte map regardless, which could cause it to run off the end
5494 of the memory it had got.
5495
5496 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5497
5498
5499 Version 0.98 22-Oct-97
5500 ----------------------
5501
5502 1. Fixed bug in code for handling temporary memory usage when there are more
5503 back references than supplied space in the ovector. This could cause segfaults.
5504
5505
5506 Version 0.97 21-Oct-97
5507 ----------------------
5508
5509 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5510
5511 2. Optimized negated single characters not to use a bit map.
5512
5513 3. Brought error texts together as macro definitions; clarified some of them;
5514 fixed one that was wrong - it said "range out of order" when it meant "invalid
5515 escape sequence".
5516
5517 4. Changed some char * arguments to const char *.
5518
5519 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5520
5521 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5522 pcretest.
5523
5524
5525 Version 0.96 16-Oct-97
5526 ----------------------
5527
5528 1. Added a simple "pgrep" utility to the distribution.
5529
5530 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5531 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5532 where "ddd" means "one or more decimal digits".
5533
5534 3. Fixed serious bug. If a pattern had a back reference, but the call to
5535 pcre_exec() didn't supply a large enough ovector to record the related
5536 identifying subpattern, the match always failed. PCRE now remembers the number
5537 of the largest back reference, and gets some temporary memory in which to save
5538 the offsets during matching if necessary, in order to ensure that
5539 backreferences always work.
5540
5541 4. Increased the compatibility with Perl in a number of ways:
5542
5543 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5544 to request this handling. The option can be set at compile or exec time.
5545
5546 (b) $ matches before a terminating newline by default; an option
5547 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5548 mode). The option can be set at compile or exec time.
5549
5550 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5551 the same as Perl's. If the decimal number it represents is less than 10
5552 or there aren't that many previous left capturing parentheses, an octal
5553 escape is read. Inside a character class, it's always an octal escape,
5554 even if it is a single digit.
5555
5556 (d) An escaped but undefined alphabetic character is taken as a literal,
5557 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5558 escapes.
5559
5560 (e) {0} is now permitted. (The previous item is removed from the compiled
5561 pattern).
5562
5563 5. Changed all the names of code files so that the basic parts are no longer
5564 than 10 characters, and abolished the teeny "globals.c" file.
5565
5566 6. Changed the handling of character classes; they are now done with a 32-byte
5567 bit map always.
5568
5569 7. Added the -d and /D options to pcretest to make it possible to look at the
5570 internals of compilation without having to recompile pcre.
5571
5572
5573 Version 0.95 23-Sep-97
5574 ----------------------
5575
5576 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5577 \x20 at the start of a run of normal characters. These were being treated as
5578 real characters, instead of the source characters being re-checked.
5579
5580
5581 Version 0.94 18-Sep-97
5582 ----------------------
5583
5584 1. The functions are now thread-safe, with the caveat that the global variables
5585 containing pointers to malloc() and free() or alternative functions are the
5586 same for all threads.
5587
5588 2. Get pcre_study() to generate a bitmap of initial characters for non-
5589 anchored patterns when this is possible, and use it if passed to pcre_exec().
5590
5591
5592 Version 0.93 15-Sep-97
5593 ----------------------
5594
5595 1. /(b)|(:+)/ was computing an incorrect first character.
5596
5597 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5598 but not actually doing anything yet.
5599
5600 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5601 as Perl does (e.g. [-az] or [az-]).
5602
5603 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5604 all possible positions.
5605
5606 5. Split up into different modules to avoid including unneeded functions in a
5607 compiled binary. However, compile and exec are still in one module. The "study"
5608 function is split off.
5609
5610 6. The character tables are now in a separate module whose source is generated
5611 by an auxiliary program - but can then be edited by hand if required. There are
5612 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5613 toupper() in the code.
5614
5615 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5616 make them global. Abolish the function for setting them, as the caller can now
5617 set them directly.
5618
5619
5620 Version 0.92 11-Sep-97
5621 ----------------------
5622
5623 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5624 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5625
5626 2. Caseless matching was not working in character classes if the characters in
5627 the pattern were in upper case.
5628
5629 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5630
5631 4. Make PCRE_ANCHORED public and accept as a compile option.
5632
5633 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5634 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5635 pass them.
5636
5637 6. Give an error if bad option bits passed at compile or run time.
5638
5639 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5640 pcretest to cause it to pass that flag.
5641
5642 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5643 options, and the first character, if set.
5644
5645 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5646
5647
5648 Version 0.91 10-Sep-97
5649 ----------------------
5650
5651 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5652 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5653
5654 2. PCRE was looping on encountering an indefinitely repeated back reference to
5655 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5656 Perl does - treats the match as successful.
5657
5658 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12