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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1521 - (show annotations)
Sun Feb 8 16:43:13 2015 UTC (4 years, 9 months ago) by ph10
File size: 271077 byte(s)
Error occurred while calculating annotation data.
Fix duplicate names memory calculation error.
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
52 Version 8.36 26-September-2014
53 ------------------------------
54
55 1. Got rid of some compiler warnings in the C++ modules that were shown up by
56 -Wmissing-field-initializers and -Wunused-parameter.
57
58 2. The tests for quantifiers being too big (greater than 65535) were being
59 applied after reading the number, and stupidly assuming that integer
60 overflow would give a negative number. The tests are now applied as the
61 numbers are read.
62
63 3. Tidy code in pcre_exec.c where two branches that used to be different are
64 now the same.
65
66 4. The JIT compiler did not generate match limit checks for certain
67 bracketed expressions with quantifiers. This may lead to exponential
68 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
69 issue should be resolved now.
70
71 5. Fixed an issue, which occures when nested alternatives are optimized
72 with table jumps.
73
74 6. Inserted two casts and changed some ints to size_t in the light of some
75 reported 64-bit compiler warnings (Bugzilla 1477).
76
77 7. Fixed a bug concerned with zero-minimum possessive groups that could match
78 an empty string, which sometimes were behaving incorrectly in the
79 interpreter (though correctly in the JIT matcher). This pcretest input is
80 an example:
81
82 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
83 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
84
85 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
86 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
87 for an empty string was breaking the inner loop and carrying on at a lower
88 level, when possessive repeated groups should always return to a higher
89 level as they have no backtrack points in them. The empty string test now
90 occurs at the outer level.
91
92 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
93 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
94
95 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
96 doesn't).
97
98 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
99 to hold a pattern to be incorrectly computed (too small) when there were
100 named back references to duplicated names. This could cause "internal
101 error: code overflow" or "double free or corruption" or other memory
102 handling errors.
103
104 11. When named subpatterns had the same prefixes, back references could be
105 confused. For example, in this pattern:
106
107 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
108
109 the reference to 'Name' was incorrectly treated as a reference to a
110 duplicate name.
111
112 12. A pattern such as /^s?c/mi8 where the optional character has more than
113 one "other case" was incorrectly compiled such that it would only try to
114 match starting at "c".
115
116 13. When a pattern starting with \s was studied, VT was not included in the
117 list of possible starting characters; this should have been part of the
118 8.34/18 patch.
119
120 14. If a character class started [\Qx]... where x is any character, the class
121 was incorrectly terminated at the ].
122
123 15. If a pattern that started with a caseless match for a character with more
124 than one "other case" was studied, PCRE did not set up the starting code
125 unit bit map for the list of possible characters. Now it does. This is an
126 optimization improvement, not a bug fix.
127
128 16. The Unicode data tables have been updated to Unicode 7.0.0.
129
130 17. Fixed a number of memory leaks in pcregrep.
131
132 18. Avoid a compiler warning (from some compilers) for a function call with
133 a cast that removes "const" from an lvalue by using an intermediate
134 variable (to which the compiler does not object).
135
136 19. Incorrect code was compiled if a group that contained an internal recursive
137 back reference was optional (had quantifier with a minimum of zero). This
138 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
139 caused segmentation faults because of stack overflows at compile time.
140
141 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
142 group that is quantified with an indefinite repeat, caused a compile-time
143 loop which used up all the system stack and provoked a segmentation fault.
144 This was not the same bug as 19 above.
145
146 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
147 Patch by Mike Frysinger.
148
149
150 Version 8.35 04-April-2014
151 --------------------------
152
153 1. A new flag is set, when property checks are present in an XCLASS.
154 When this flag is not set, PCRE can perform certain optimizations
155 such as studying these XCLASS-es.
156
157 2. The auto-possessification of character sets were improved: a normal
158 and an extended character set can be compared now. Furthermore
159 the JIT compiler optimizes more character set checks.
160
161 3. Got rid of some compiler warnings for potentially uninitialized variables
162 that show up only when compiled with -O2.
163
164 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
165 of a match later then the end of the match. The pcretest program was not
166 handling the case sensibly - it was outputting from the start to the next
167 binary zero. It now reports this situation in a message, and outputs the
168 text from the end to the start.
169
170 5. Fast forward search is improved in JIT. Instead of the first three
171 characters, any three characters with fixed position can be searched.
172 Search order: first, last, middle.
173
174 6. Improve character range checks in JIT. Characters are read by an inprecise
175 function now, which returns with an unknown value if the character code is
176 above a certain threshold (e.g: 256). The only limitation is that the value
177 must be bigger than the threshold as well. This function is useful when
178 the characters above the threshold are handled in the same way.
179
180 7. The macros whose names start with RAWUCHAR are placeholders for a future
181 mode in which only the bottom 21 bits of 32-bit data items are used. To
182 make this more memorable for those maintaining the code, the names have
183 been changed to start with UCHAR21, and an extensive comment has been added
184 to their definition.
185
186 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
187 to the export list in Makefile.am (they were accidentally omitted from the
188 8.34 tarball).
189
190 9. The informational output from pcretest used the phrase "starting byte set"
191 which is inappropriate for the 16-bit and 32-bit libraries. As the output
192 for "first char" and "need char" really means "non-UTF-char", I've changed
193 "byte" to "char", and slightly reworded the output. The documentation about
194 these values has also been (I hope) clarified.
195
196 10. Another JIT related optimization: use table jumps for selecting the correct
197 backtracking path, when more than four alternatives are present inside a
198 bracket.
199
200 11. Empty match is not possible, when the minimum length is greater than zero,
201 and there is no \K in the pattern. JIT should avoid empty match checks in
202 such cases.
203
204 12. In a caseless character class with UCP support, when a character with more
205 than one alternative case was not the first character of a range, not all
206 the alternative cases were added to the class. For example, s and \x{17f}
207 are both alternative cases for S: the class [RST] was handled correctly,
208 but [R-T] was not.
209
210 13. The configure.ac file always checked for pthread support when JIT was
211 enabled. This is not used in Windows, so I have put this test inside a
212 check for the presence of windows.h (which was already tested for).
213
214 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
215 The algorithm provides a way to skip certain starting offsets, and usually
216 faster than linear prefix searches.
217
218 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
219 as for 'fr_FR' and 'french'. For some reason, however, it then used the
220 Windows-specific input and output files, which have 'french' screwed in.
221 So this could never have worked. One of the problems with locales is that
222 they aren't always the same. I have now updated RunTest so that it checks
223 the output of the locale test (test 3) against three different output
224 files, and it allows the test to pass if any one of them matches. With luck
225 this should make the test pass on some versions of Solaris where it was
226 failing. Because of the uncertainty, the script did not used to stop if
227 test 3 failed; it now does. If further versions of a French locale ever
228 come to light, they can now easily be added.
229
230 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
231 there was a message during ./configure, but it did not stop. This now
232 provokes an error. The invalid example in README has been corrected.
233 If a value less than the minimum is given, the minimum value has always
234 been used, but now a warning is given.
235
236 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
237 was a bug in the test system, which is now fixed. Also, the list of various
238 configurations that are tested for each release did not have one with both
239 16/32 bits and --enable-bar-anycrlf. It now does.
240
241 18. pcretest was missing "-C bsr" for displaying the \R default setting.
242
243 19. Little endian PowerPC systems are supported now by the JIT compiler.
244
245 20. The fast forward newline mechanism could enter to an infinite loop on
246 certain invalid UTF-8 input. Although we don't support these cases
247 this issue can be fixed by a performance optimization.
248
249 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
250 not take account if existing stack usage. There is now a new global
251 variable called pcre_stack_guard that can be set to point to an external
252 function to check stack availability. It is called at the start of
253 processing every parenthesized group.
254
255 22. A typo in the code meant that in ungreedy mode the max/min qualifier
256 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
257 not match "ab".
258
259 23. When UTF was disabled, the JIT program reported some incorrect compile
260 errors. These messages are silenced now.
261
262 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
263 compiler.
264
265 25. Change all the temporary files used in RunGrepTest to be different to those
266 used by RunTest so that the tests can be run simultaneously, for example by
267 "make -j check".
268
269
270 Version 8.34 15-December-2013
271 -----------------------------
272
273 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
274 executable memory. Patch inspired by Carsten Klein.
275
276 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
277 this macro is never tested and has no effect, because the work to support
278 coverage involves only compiling and linking options and special targets in
279 the Makefile. The comment in config.h implied that defining the macro would
280 enable coverage support, which is totally false. There was also support for
281 setting this macro in the CMake files (my fault, I just copied it from
282 configure). SUPPORT_GCOV has now been removed.
283
284 3. Make a small performance improvement in strlen16() and strlen32() in
285 pcretest.
286
287 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
288 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
289 these cases). There was also one in pcretest.c.
290
291 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
292
293 6. In UTF mode, the code for checking whether a group could match an empty
294 string (which is used for indefinitely repeated groups to allow for
295 breaking an infinite loop) was broken when the group contained a repeated
296 negated single-character class with a character that occupied more than one
297 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
298 UTF-8 mode). The effect was undefined: the group might or might not be
299 deemed as matching an empty string, or the program might have crashed.
300
301 7. The code for checking whether a group could match an empty string was not
302 recognizing that \h, \H, \v, \V, and \R must match a character.
303
304 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
305 an empty string. If it can, pcretest shows this in its information output.
306
307 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
308 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
309 matched by pcre_exec() without using JIT:
310
311 (a) If the rest of the pattern did not match after a maximal run of
312 grapheme clusters, the code for backing up to try with fewer of them
313 did not always back up over a full grapheme when characters that do not
314 have the modifier quality were involved, e.g. Hangul syllables.
315
316 (b) If the match point in a subject started with modifier character, and
317 there was no match, the code could incorrectly back up beyond the match
318 point, and potentially beyond the first character in the subject,
319 leading to a segfault or an incorrect match result.
320
321 10. A conditional group with an assertion condition could lead to PCRE
322 recording an incorrect first data item for a match if no other first data
323 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
324 first data item, and therefore matched "ca" after "c" instead of at the
325 start.
326
327 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
328 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
329
330 12. The source of pcregrep now includes z/OS-specific code so that it can be
331 compiled for z/OS as part of the special z/OS distribution.
332
333 13. Added the -T and -TM options to pcretest.
334
335 14. The code in pcre_compile.c for creating the table of named capturing groups
336 has been refactored. Instead of creating the table dynamically during the
337 actual compiling pass, the information is remembered during the pre-compile
338 pass (on the stack unless there are more than 20 named groups, in which
339 case malloc() is used) and the whole table is created before the actual
340 compile happens. This has simplified the code (it is now nearly 150 lines
341 shorter) and prepared the way for better handling of references to groups
342 with duplicate names.
343
344 15. A back reference to a named subpattern when there is more than one of the
345 same name now checks them in the order in which they appear in the pattern.
346 The first one that is set is used for the reference. Previously only the
347 first one was inspected. This change makes PCRE more compatible with Perl.
348
349 16. Unicode character properties were updated from Unicode 6.3.0.
350
351 17. The compile-time code for auto-possessification has been refactored, based
352 on a patch by Zoltan Herczeg. It now happens after instead of during
353 compilation. The code is cleaner, and more cases are handled. The option
354 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
355 options in pcretest are provided to set it. It can also be set by
356 (*NO_AUTO_POSSESS) at the start of a pattern.
357
358 18. The character VT has been added to the default ("C" locale) set of
359 characters that match \s and are generally treated as white space,
360 following this same change in Perl 5.18. There is now no difference between
361 "Perl space" and "POSIX space". Whether VT is treated as white space in
362 other locales depends on the locale.
363
364 19. The code for checking named groups as conditions, either for being set or
365 for being recursed, has been refactored (this is related to 14 and 15
366 above). Processing unduplicated named groups should now be as fast at
367 numerical groups, and processing duplicated groups should be faster than
368 before.
369
370 20. Two patches to the CMake build system, by Alexander Barkov:
371
372 (1) Replace the "source" command by "." in CMakeLists.txt because
373 "source" is a bash-ism.
374
375 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
376 without these the CMake build does not work on Solaris.
377
378 21. Perl has changed its handling of \8 and \9. If there is no previously
379 encountered capturing group of those numbers, they are treated as the
380 literal characters 8 and 9 instead of a binary zero followed by the
381 literals. PCRE now does the same.
382
383 22. Following Perl, added \o{} to specify codepoints in octal, making it
384 possible to specify values greater than 0777 and also making them
385 unambiguous.
386
387 23. Perl now gives an error for missing closing braces after \x{... instead of
388 treating the string as literal. PCRE now does the same.
389
390 24. RunTest used to grumble if an inappropriate test was selected explicitly,
391 but just skip it when running all tests. This make it awkward to run ranges
392 of tests when one of them was inappropriate. Now it just skips any
393 inappropriate tests, as it always did when running all tests.
394
395 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
396 character types such as \d or \w, too many callouts were inserted, and the
397 data that they returned was rubbish.
398
399 26. In UCP mode, \s was not matching two of the characters that Perl matches,
400 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
401 were matched by \h. The code has now been refactored so that the lists of
402 the horizontal and vertical whitespace characters used for \h and \v (which
403 are defined only in one place) are now also used for \s.
404
405 27. Add JIT support for the 64 bit TileGX architecture.
406 Patch by Jiong Wang (Tilera Corporation).
407
408 28. Possessive quantifiers for classes (both explicit and automatically
409 generated) now use special opcodes instead of wrapping in ONCE brackets.
410
411 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
412 (because it's meaningless), this was not happening when PCRE_CASELESS was
413 set. Not wrong, but inefficient.
414
415 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
416 properties for \w, \d, etc) is present in a test regex. Otherwise if the
417 test contains no characters greater than 255, Perl doesn't realise it
418 should be using Unicode semantics.
419
420 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
421 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
422 does in Unicode mode.
423
424 32. Added the "forbid" facility to pcretest so that putting tests into the
425 wrong test files can sometimes be quickly detected.
426
427 33. There is now a limit (default 250) on the depth of nesting of parentheses.
428 This limit is imposed to control the amount of system stack used at compile
429 time. It can be changed at build time by --with-parens-nest-limit=xxx or
430 the equivalent in CMake.
431
432 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
433 errors. Perl warns for these when in warning mode, but PCRE has no facility
434 for giving warnings.
435
436 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
437 However, this was not working for (?!) because it is optimized to (*FAIL),
438 for which PCRE does not allow quantifiers. The optimization is now disabled
439 when a quantifier follows (?!). I can't see any use for this, but it makes
440 things uniform.
441
442 36. Perl no longer allows group names to start with digits, so I have made this
443 change also in PCRE. It simplifies the code a bit.
444
445 37. In extended mode, Perl ignores spaces before a + that indicates a
446 possessive quantifier. PCRE allowed a space before the quantifier, but not
447 before the possessive +. It now does.
448
449 38. The use of \K (reset reported match start) within a repeated possessive
450 group such as (a\Kb)*+ was not working.
451
452 40. Document that the same character tables must be used at compile time and
453 run time, and that the facility to pass tables to pcre_exec() and
454 pcre_dfa_exec() is for use only with saved/restored patterns.
455
456 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
457 features for Builds with MSVC:
458
459 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
460 builds). The .rc files can be used to set FileDescription and many other
461 attributes.
462
463 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
464 This allows higher-level build scripts which want .pdb files to avoid
465 hard-coding the exact files needed."
466
467 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
468 mean "start of word" and "end of word", respectively, as a transition aid.
469
470 43. A minimizing repeat of a class containing codepoints greater than 255 in
471 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
472 compiled to use the heap for recursion.
473
474 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
475 is configured.
476
477
478 Version 8.33 28-May-2013
479 ------------------------
480
481 1. Added 'U' to some constants that are compared to unsigned integers, to
482 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
483 variables that are added to signed variables, to ensure the result is
484 signed and can be negated.
485
486 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
487 CMake config files.
488
489 3. Revise the creation of config.h.generic so that all boolean macros are
490 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
491 overriding via -D on the command line possible.
492
493 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
494 to unsigned int is reported to make a quite noticeable speed difference in
495 a specific Windows environment. Testing on Linux did also appear to show
496 some benefit (and it is clearly not harmful). Also fixed the definition of
497 Xop which should be unsigned.
498
499 5. Related to (4), changing the definition of the intermediate variable cc
500 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
501 improvements.
502
503 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
504 unnecessary spaces.
505
506 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
507 and later.
508
509 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
510
511 9. Optimizing fast_forward_start_bits in JIT.
512
513 10. Adding support for callouts in JIT, and fixing some issues revealed
514 during this work. Namely:
515
516 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
517
518 (b) Minimum length was not checked before the matching is started.
519
520 11. The value of capture_last that is passed to callouts was incorrect in some
521 cases when there was a capture on one path that was subsequently abandoned
522 after a backtrack. Also, the capture_last value is now reset after a
523 recursion, since all captures are also reset in this case.
524
525 12. The interpreter no longer returns the "too many substrings" error in the
526 case when an overflowing capture is in a branch that is subsequently
527 abandoned after a backtrack.
528
529 13. In the pathological case when an offset vector of size 2 is used, pcretest
530 now prints out the matched string after a yield of 0 or 1.
531
532 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
533 Only supported by the JIT compiler at the moment.
534
535 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
536
537 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
538 offset of the starting point of the matching process, provided the offsets
539 vector is large enough.
540
541 17. The \A escape now records a lookbehind value of 1, though its execution
542 does not actually inspect the previous character. This is to ensure that,
543 in partial multi-segment matching, at least one character from the old
544 segment is retained when a new segment is processed. Otherwise, if there
545 are no lookbehinds in the pattern, \A might match incorrectly at the start
546 of a new segment.
547
548 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
549
550 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
551 gives some modest performance improvement in 8-bit mode.
552
553 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
554 be expressed in certain programming languages using Universal Character
555 Names.
556
557 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
558 which points out that "non characters" are not "characters that may not
559 appear in Unicode strings" but rather "characters that are reserved for
560 internal use and have only local meaning".
561
562 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
563 there was a conditional group that depended on an assertion, if the
564 assertion was false, the callout that immediately followed the alternation
565 in the condition was skipped when pcre_exec() was used for matching.
566
567 23. Allow an explicit callout to be inserted before an assertion that is the
568 condition for a conditional group, for compatibility with automatic
569 callouts, which always insert a callout at this point.
570
571 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
572 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
573
574 25. (*PRUNE) is now supported by the JIT compiler.
575
576 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
577
578 27. Fix the case where there are two or more SKIPs with arguments that may be
579 ignored.
580
581 28. (*SKIP) is now supported by the JIT compiler.
582
583 29. (*THEN) is now supported by the JIT compiler.
584
585 30. Update RunTest with additional test selector options.
586
587 31. The way PCRE handles backtracking verbs has been changed in two ways.
588
589 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
590 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
591 backtracking. In some cases this makes it more Perl-compatible, but Perl's
592 rather obscure rules do not always do the same thing.
593
594 (2) Previously, backtracking verbs were confined within assertions. This is
595 no longer the case for positive assertions, except for (*ACCEPT). Again,
596 this sometimes improves Perl compatibility, and sometimes does not.
597
598 32. A number of tests that were in test 2 because Perl did things differently
599 have been moved to test 1, because either Perl or PCRE has changed, and
600 these tests are now compatible.
601
602 32. Backtracking control verbs are now handled in the same way in JIT and
603 interpreter.
604
605 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
606 contained a forward subroutine reference caused a compile error.
607
608 34. Auto-detect and optimize limited repetitions in JIT.
609
610 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
611 blocking (*UTF) etc.
612
613 36. In the interpreter, maximizing pattern repetitions for characters and
614 character types now use tail recursion, which reduces stack usage.
615
616 37. The value of the max lookbehind was not correctly preserved if a compiled
617 and saved regex was reloaded on a host of different endianness.
618
619 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
620 of the compiled pattern block, expand the flags field from 16 to 32 bits
621 because it was almost full.
622
623 39. Try madvise first before posix_madvise.
624
625 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
626 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
627 original change locked out all matches of empty strings. This has been
628 changed so that one match of an empty string per line is recognized.
629 Subsequent searches on the same line (for colouring or for --only-matching,
630 for example) do not recognize empty strings.
631
632 41. Applied a user patch to fix a number of spelling mistakes in comments.
633
634 42. Data lines longer than 65536 caused pcretest to crash.
635
636 43. Clarified the data type for length and startoffset arguments for pcre_exec
637 and pcre_dfa_exec in the function-specific man pages, where they were
638 explicitly stated to be in bytes, never having been updated. I also added
639 some clarification to the pcreapi man page.
640
641 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
642 a segmentation fault.
643
644
645 Version 8.32 30-November-2012
646 -----------------------------
647
648 1. Improved JIT compiler optimizations for first character search and single
649 character iterators.
650
651 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
652 Patch by Daniel Richard G.
653
654 3. Single character iterator optimizations in the JIT compiler.
655
656 4. Improved JIT compiler optimizations for character ranges.
657
658 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
659 Reported by Giuseppe D'Angelo.
660
661 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
662 of a line, was being set incorrectly in cases where .* appeared inside
663 atomic brackets at the start of a pattern, or where there was a subsequent
664 *PRUNE or *SKIP.
665
666 7. Improved instruction cache flush for POWER/PowerPC.
667 Patch by Daniel Richard G.
668
669 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
670 grep:
671
672 (a) There is now no limit to the number of patterns to be matched.
673
674 (b) An error is given if a pattern is too long.
675
676 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
677 are now supported.
678
679 (d) --exclude-from and --include-from (multiple use) have been added.
680
681 (e) Exclusions and inclusions now apply to all files and directories, not
682 just to those obtained from scanning a directory recursively.
683
684 (f) Multiple uses of -f and --file-list are now supported.
685
686 (g) In a Windows environment, the default for -d has been changed from
687 "read" (the GNU grep default) to "skip", because otherwise the presence
688 of a directory in the file list provokes an error.
689
690 (h) The documentation has been revised and clarified in places.
691
692 9. Improve the matching speed of capturing brackets.
693
694 10. Changed the meaning of \X so that it now matches a Unicode extended
695 grapheme cluster.
696
697 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
698 out POSIX threads when JIT support is configured.
699
700 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
701
702 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
703 structure could go wrong in environments where size_t is not the same size
704 as int.
705
706 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
707 set.
708
709 15. The EBCDIC support had decayed; later updates to the code had included
710 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
711 general tidy up of EBCDIC-related issues, and the documentation was also
712 not quite right. There is now a test that can be run on ASCII systems to
713 check some of the EBCDIC-related things (but is it not a full test).
714
715 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
716 in a small tidy to the code.
717
718 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
719
720 18. If the --only-matching (-o) option in pcregrep is specified multiple
721 times, each one causes appropriate output. For example, -o1 -o2 outputs the
722 substrings matched by the 1st and 2nd capturing parentheses. A separating
723 string can be specified by --om-separator (default empty).
724
725 19. Improving the first n character searches.
726
727 20. Turn case lists for horizontal and vertical white space into macros so that
728 they are defined only once.
729
730 21. This set of changes together give more compatible Unicode case-folding
731 behaviour for characters that have more than one other case when UCP
732 support is available.
733
734 (a) The Unicode property table now has offsets into a new table of sets of
735 three or more characters that are case-equivalent. The MultiStage2.py
736 script that generates these tables (the pcre_ucd.c file) now scans
737 CaseFolding.txt instead of UnicodeData.txt for character case
738 information.
739
740 (b) The code for adding characters or ranges of characters to a character
741 class has been abstracted into a generalized function that also handles
742 case-independence. In UTF-mode with UCP support, this uses the new data
743 to handle characters with more than one other case.
744
745 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
746 whose other case is greater than 256 are now correctly matched
747 caselessly. Previously, the high codepoint matched the low one, but not
748 vice versa.
749
750 (d) The processing of \h, \H, \v, and \ in character classes now makes use
751 of the new class addition function, using character lists defined as
752 macros alongside the case definitions of 20 above.
753
754 (e) Caseless back references now work with characters that have more than
755 one other case.
756
757 (f) General caseless matching of characters with more than one other case
758 is supported.
759
760 22. Unicode character properties were updated from Unicode 6.2.0
761
762 23. Improved CMake support under Windows. Patch by Daniel Richard G.
763
764 24. Add support for 32-bit character strings, and UTF-32
765
766 25. Major JIT compiler update (code refactoring and bugfixing).
767 Experimental Sparc 32 support is added.
768
769 26. Applied a modified version of Daniel Richard G's patch to create
770 pcre.h.generic and config.h.generic by "make" instead of in the
771 PrepareRelease script.
772
773 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
774 pcre_compile.c when checking for a zero character.
775
776 28. Introducing a native interface for JIT. Through this interface, the compiled
777 machine code can be directly executed. The purpose of this interface is to
778 provide fast pattern matching, so several sanity checks are not performed.
779 However, feature tests are still performed. The new interface provides
780 1.4x speedup compared to the old one.
781
782 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
783 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
784 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
785
786 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
787 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
788 the "old" RFC 2279). Instead, it ended up passing a negative length to
789 pcre_exec().
790
791 31. Add support for GCC's visibility feature to hide internal functions.
792
793 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
794 "unknown -C option" after outputting 0 or 1.
795
796 33. There is now support for generating a code coverage report for the test
797 suite in environments where gcc is the compiler and lcov is installed. This
798 is mainly for the benefit of the developers.
799
800 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
801 unaddressable using valgrind annotations, allowing valgrind to detect
802 invalid memory accesses. This is mainly for the benefit of the developers.
803
804 25. (*UTF) can now be used to start a pattern in any of the three libraries.
805
806 26. Give configure error if --enable-cpp but no C++ compiler found.
807
808
809 Version 8.31 06-July-2012
810 -------------------------
811
812 1. Fixing a wrong JIT test case and some compiler warnings.
813
814 2. Removed a bashism from the RunTest script.
815
816 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
817 to unsigned type, result still unsigned" that was given by an MS compiler
818 on encountering the code "-sizeof(xxx)".
819
820 4. Partial matching support is added to the JIT compiler.
821
822 5. Fixed several bugs concerned with partial matching of items that consist
823 of more than one character:
824
825 (a) /^(..)\1/ did not partially match "aba" because checking references was
826 done on an "all or nothing" basis. This also applied to repeated
827 references.
828
829 (b) \R did not give a hard partial match if \r was found at the end of the
830 subject.
831
832 (c) \X did not give a hard partial match after matching one or more
833 characters at the end of the subject.
834
835 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
836 a partial match for the string "\r".
837
838 (e) When newline was set to CRLF, the metacharacter "." did not recognize
839 a partial match for a CR character at the end of the subject string.
840
841 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
842 running pcretest, the text "(JIT)" added to the output whenever JIT is
843 actually used to run the match.
844
845 7. Individual JIT compile options can be set in pcretest by following -s+[+]
846 or /S+[+] with a digit between 1 and 7.
847
848 8. OP_NOT now supports any UTF character not just single-byte ones.
849
850 9. (*MARK) control verb is now supported by the JIT compiler.
851
852 10. The command "./RunTest list" lists the available tests without actually
853 running any of them. (Because I keep forgetting what they all are.)
854
855 11. Add PCRE_INFO_MAXLOOKBEHIND.
856
857 12. Applied a (slightly modified) user-supplied patch that improves performance
858 when the heap is used for recursion (compiled with --disable-stack-for-
859 recursion). Instead of malloc and free for each heap frame each time a
860 logical recursion happens, frames are retained on a chain and re-used where
861 possible. This sometimes gives as much as 30% improvement.
862
863 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
864 call.
865
866 14. As documented, (*COMMIT) is now confined to within a positive assertion.
867
868 15. It is now possible to link pcretest with libedit as an alternative to
869 libreadline.
870
871 16. (*COMMIT) control verb is now supported by the JIT compiler.
872
873 17. The Unicode data tables have been updated to Unicode 6.1.0.
874
875 18. Added --file-list option to pcregrep.
876
877 19. Added binary file support to pcregrep, including the -a, --binary-files,
878 -I, and --text options.
879
880 20. The madvise function is renamed for posix_madvise for QNX compatibility
881 reasons. Fixed by Giuseppe D'Angelo.
882
883 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
884
885 22. Changed the option for creating symbolic links for 16-bit man pages from
886 -s to -sf so that re-installing does not cause issues.
887
888 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
889
890 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
891 matches in certain environments (the workspace was not being correctly
892 retained). Also added to pcre_dfa_exec() a simple plausibility check on
893 some of the workspace data at the beginning of a restart.
894
895 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
896 was not doing so when it should - probably a typo introduced by SVN 528
897 (change 8.10/14).
898
899 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
900 \w+ when the character tables indicated that \x{c4} was a word character.
901 There were several related cases, all because the tests for doing a table
902 lookup were testing for characters less than 127 instead of 255.
903
904 27. If a pattern contains capturing parentheses that are not used in a match,
905 their slots in the ovector are set to -1. For those that are higher than
906 any matched groups, this happens at the end of processing. In the case when
907 there were back references that the ovector was too small to contain
908 (causing temporary malloc'd memory to be used during matching), and the
909 highest capturing number was not used, memory off the end of the ovector
910 was incorrectly being set to -1. (It was using the size of the temporary
911 memory instead of the true size.)
912
913 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
914 ovector size, it uses memory at the end of the block that it has got.
915
916 29. Check for an overlong MARK name and give an error at compile time. The
917 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
918
919 30. JIT compiler update.
920
921 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
922 Rill for the patch.
923
924 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
925
926 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
927
928 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
929 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
930 was enabled.
931
932 35. Improve JIT code generation for greedy plus quantifier.
933
934 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
935 1 to "aa" instead of to an empty string. The bug affected repeated groups
936 that could potentially match an empty string.
937
938 37. Optimizing single character iterators in JIT.
939
940 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
941 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
942 codepoints that are too big for the mode are faulted, and in a UTF mode,
943 disallowed codepoints are also faulted.
944
945 39. If PCRE was compiled with UTF support, in three places in the DFA
946 matcher there was code that should only have been obeyed in UTF mode, but
947 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
948 processing when bytes with values greater than 127 were present. In 16-bit
949 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
950 both cases the values are those that cannot be the first data item in a UTF
951 character. The three items that might have provoked this were recursions,
952 possessively repeated groups, and atomic groups.
953
954 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
955 and pcregrep, because some OS require shared objects to be explicitly
956 passed to ld, causing the link step to fail if they are not.
957
958 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
959 mode, patterns that started with \h* or \R* might be incorrectly matched.
960
961
962 Version 8.30 04-February-2012
963 -----------------------------
964
965 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
966 name is defined in ctype.h.
967
968 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
969 only in quite long subpatterns.
970
971 3. Removed the function pcre_info(), which has been obsolete and deprecated
972 since it was replaced by pcre_fullinfo() in February 2000.
973
974 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
975 match a (*MARK), and the match failed at the start of the subject, a
976 reference to memory before the start of the subject could occur. This bug
977 was introduced by fix 17 of release 8.21.
978
979 5. A reference to an unset group with zero minimum repetition was giving
980 totally wrong answers (in non-JavaScript-compatibility mode). For example,
981 /(another)?(\1?)test/ matched against "hello world test". This bug was
982 introduced in release 8.13.
983
984 6. Add support for 16-bit character strings (a large amount of work involving
985 many changes and refactorings).
986
987 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
988 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
989 from a file.
990
991 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
992 rounding is not applied in this particular case).
993
994 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
995 if they appear, or are escaped, in patterns.
996
997 10. Get rid of a number of -Wunused-but-set-variable warnings.
998
999 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1000 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1001 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1002 also returns the mark "x". This bug applied to capturing parentheses,
1003 non-capturing parentheses, and atomic parentheses. It also applied to some
1004 assertions.
1005
1006 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1007 information out of configure.ac instead of relying on pcre.h.generic, which
1008 is not stored in the repository.
1009
1010 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1011 -lreadline.
1012
1013 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1014
1015 15. Applied Graycode's patch to put the top-level frame on the stack rather
1016 than the heap when not using the stack for recursion. This gives a
1017 performance improvement in many cases when recursion is not deep.
1018
1019 16. Experimental code added to "pcretest -C" to output the stack frame size.
1020
1021
1022 Version 8.21 12-Dec-2011
1023 ------------------------
1024
1025 1. Updating the JIT compiler.
1026
1027 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1028 are added as well.
1029
1030 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1031 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1032 calling _pcre_jit_exec. Some extra comments are added.
1033
1034 4. (*MARK) settings inside atomic groups that do not contain any capturing
1035 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1036 was introduced by change 18 for 8.20.
1037
1038 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1039 ECMA-262 standard.
1040
1041 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1042 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1043 This bug was probably introduced by change 9 of 8.13.
1044
1045 7. While fixing 6 above, I noticed that a number of other items were being
1046 incorrectly rejected as "not fixed length". This arose partly because newer
1047 opcodes had not been added to the fixed-length checking code. I have (a)
1048 corrected the bug and added tests for these items, and (b) arranged for an
1049 error to occur if an unknown opcode is encountered while checking for fixed
1050 length instead of just assuming "not fixed length". The items that were
1051 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1052 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1053 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1054
1055 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1056 being incorrectly compiled and would have given unpredicatble results.
1057
1058 9. A possessively repeated subpattern with minimum repeat count greater than
1059 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1060 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1061 the first (A) could occur when it should not.
1062
1063 10. Add a cast and remove a redundant test from the code.
1064
1065 11. JIT should use pcre_malloc/pcre_free for allocation.
1066
1067 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1068 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1069 is anything other than /usr, -L is still shown).
1070
1071 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1072
1073 14. Perl does not support \N without a following name in a [] class; PCRE now
1074 also gives an error.
1075
1076 15. If a forward reference was repeated with an upper limit of around 2000,
1077 it caused the error "internal error: overran compiling workspace". The
1078 maximum number of forward references (including repeats) was limited by the
1079 internal workspace, and dependent on the LINK_SIZE. The code has been
1080 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1081 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1082 of around 200,000 forward references. While doing this, I also speeded up
1083 the filling in of repeated forward references.
1084
1085 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1086 incorrectly expecting the subject to contain another "a" after the start.
1087
1088 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1089 in the match, the SKIP should be ignored. This was not happening; instead
1090 the SKIP was being treated as NOMATCH. For patterns such as
1091 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1092 tested.
1093
1094 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1095 now much more compatible with Perl, in particular in cases where the result
1096 is a non-match for a non-anchored pattern. For example, if
1097 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1098 "m", where previously it did not return a name. A side effect of this
1099 change is that for partial matches, the last encountered mark name is
1100 returned, as for non matches. A number of tests that were previously not
1101 Perl-compatible have been moved into the Perl-compatible test files. The
1102 refactoring has had the pleasing side effect of removing one argument from
1103 the match() function, thus reducing its stack requirements.
1104
1105 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1106 subsequent uses of /S (without +) incorrectly behaved like /S+.
1107
1108 21. Retrieve executable code size support for the JIT compiler and fixing
1109 some warnings.
1110
1111 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1112 not work when the shorter character appeared right at the end of the
1113 subject string.
1114
1115 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1116 systems.
1117
1118 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1119 output it when the /M option is used in pcretest.
1120
1121 25. The CheckMan script was not being included in the distribution. Also, added
1122 an explicit "perl" to run Perl scripts from the PrepareRelease script
1123 because this is reportedly needed in Windows.
1124
1125 26. If study data was being save in a file and studying had not found a set of
1126 "starts with" bytes for the pattern, the data written to the file (though
1127 never used) was taken from uninitialized memory and so caused valgrind to
1128 complain.
1129
1130 27. Updated RunTest.bat as provided by Sheri Pierce.
1131
1132 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1133
1134 29. Computation of memory usage for the table of capturing group names was
1135 giving an unnecessarily large value.
1136
1137
1138 Version 8.20 21-Oct-2011
1139 ------------------------
1140
1141 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1142 a POSIX class. After further experiments with Perl, which convinced me that
1143 Perl has bugs and confusions, a closing square bracket is no longer allowed
1144 in a POSIX name. This bug also affected patterns with classes that started
1145 with full stops.
1146
1147 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1148 captured substring, but while checking the failing first alternative,
1149 substring 1 is temporarily captured. If the output vector supplied to
1150 pcre_exec() was not big enough for this capture, the yield of the function
1151 was still zero ("insufficient space for captured substrings"). This cannot
1152 be totally fixed without adding another stack variable, which seems a lot
1153 of expense for a edge case. However, I have improved the situation in cases
1154 such as /(a)(b)x|abc/ matched against "abc", where the return code
1155 indicates that fewer than the maximum number of slots in the ovector have
1156 been set.
1157
1158 3. Related to (2) above: when there are more back references in a pattern than
1159 slots in the output vector, pcre_exec() uses temporary memory during
1160 matching, and copies in the captures as far as possible afterwards. It was
1161 using the entire output vector, but this conflicts with the specification
1162 that only 2/3 is used for passing back captured substrings. Now it uses
1163 only the first 2/3, for compatibility. This is, of course, another edge
1164 case.
1165
1166 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1167 main code base, and can be used by building with --enable-jit. When this is
1168 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1169 runtime --no-jit option is given.
1170
1171 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1172 ovector, the return from the function was zero, implying that there were
1173 other matches that did not fit. The correct "exactly full" value is now
1174 returned.
1175
1176 6. If a subpattern that was called recursively or as a subroutine contained
1177 (*PRUNE) or any other control that caused it to give a non-standard return,
1178 invalid errors such as "Error -26 (nested recursion at the same subject
1179 position)" or even infinite loops could occur.
1180
1181 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1182 computing the minimum length on reaching *ACCEPT, and so ended up with the
1183 wrong value of 1 rather than 0. Further investigation indicates that
1184 computing a minimum subject length in the presence of *ACCEPT is difficult
1185 (think back references, subroutine calls), and so I have changed the code
1186 so that no minimum is registered for a pattern that contains *ACCEPT.
1187
1188 8. If (*THEN) was present in the first (true) branch of a conditional group,
1189 it was not handled as intended. [But see 16 below.]
1190
1191 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1192 Sheri Pierce.
1193
1194 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1195 the first byte in a match must be "a".
1196
1197 11. Change 17 for 8.13 increased the recursion depth for patterns like
1198 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1199 pattern contains any instances of (*THEN). If it does not, the old
1200 optimizations are restored. It would be nice to do this on a per-group
1201 basis, but at the moment that is not feasible.
1202
1203 12. In some environments, the output of pcretest -C is CRLF terminated. This
1204 broke RunTest's code that checks for the link size. A single white space
1205 character after the value is now allowed for.
1206
1207 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1208 For "fr", it uses the Windows-specific input and output files.
1209
1210 14. If (*THEN) appeared in a group that was called recursively or as a
1211 subroutine, it did not work as intended. [But see next item.]
1212
1213 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1214 pattern fragments (but not containing any | characters). If A and B are
1215 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1216 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1217 D. In other words, Perl considers parentheses that do not contain any |
1218 characters to be part of a surrounding alternative, whereas PCRE was
1219 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1220 differently. PCRE now behaves in the same way as Perl, except in the case
1221 of subroutine/recursion calls such as (?1) which have in any case always
1222 been different (but PCRE had them first :-).
1223
1224 16. Related to 15 above: Perl does not treat the | in a conditional group as
1225 creating alternatives. Such a group is treated in the same way as an
1226 ordinary group without any | characters when processing (*THEN). PCRE has
1227 been changed to match Perl's behaviour.
1228
1229 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1230 RunGrepTest script failed.
1231
1232 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1233 inevitable for groups that contain captures, but it can lead to a lot of
1234 stack use in large patterns. The old behaviour has been restored for atomic
1235 groups that do not contain any capturing parentheses.
1236
1237 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1238 suppress the check for a minimum subject length at run time. (If it was
1239 given to pcre_exec() or pcre_dfa_exec() it did work.)
1240
1241 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1242 fail to work when decoding hex characters in data strings in EBCDIC
1243 environments.
1244
1245 21. It appears that in at least one Mac OS environment, the isxdigit() function
1246 is implemented as a macro that evaluates to its argument more than once,
1247 contravening the C 90 Standard (I haven't checked a later standard). There
1248 was an instance in pcretest which caused it to go wrong when processing
1249 \x{...} escapes in subject strings. The has been rewritten to avoid using
1250 things like p++ in the argument of isxdigit().
1251
1252
1253 Version 8.13 16-Aug-2011
1254 ------------------------
1255
1256 1. The Unicode data tables have been updated to Unicode 6.0.0.
1257
1258 2. Two minor typos in pcre_internal.h have been fixed.
1259
1260 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1261 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1262 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1263
1264 4. There were a number of related bugs in the code for matching backrefences
1265 caselessly in UTF-8 mode when codes for the characters concerned were
1266 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1267 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1268 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1269 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1270 2-byte code at the end of the subject (it thought there wasn't enough data
1271 left).
1272
1273 5. Comprehensive information about what went wrong is now returned by
1274 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1275 as the output vector has at least 2 elements. The offset of the start of
1276 the failing character and a reason code are placed in the vector.
1277
1278 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1279 now returned is for the first byte of the failing character, instead of the
1280 last byte inspected. This is an incompatible change, but I hope it is small
1281 enough not to be a problem. It makes the returned offset consistent with
1282 pcre_exec() and pcre_dfa_exec().
1283
1284 7. pcretest now gives a text phrase as well as the error number when
1285 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1286 failure, the offset and reason code are output.
1287
1288 8. When \R was used with a maximizing quantifier it failed to skip backwards
1289 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1290 back over a single character (\n). This seems wrong (because it treated the
1291 two characters as a single entity when going forwards), conflicts with the
1292 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1293 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1294 has been changed.
1295
1296 9. Some internal refactoring has changed the processing so that the handling
1297 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1298 time (the PCRE_DOTALL option was changed this way some time ago: version
1299 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1300 which was always a bit of a fudge. It also means that there is one less
1301 argument for the match() function, which reduces its stack requirements
1302 slightly. This change also fixes an incompatibility with Perl: the pattern
1303 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1304
1305 10. More internal refactoring has drastically reduced the number of recursive
1306 calls to match() for possessively repeated groups such as (abc)++ when
1307 using pcre_exec().
1308
1309 11. While implementing 10, a number of bugs in the handling of groups were
1310 discovered and fixed:
1311
1312 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1313 (a|)*(?1) gave a compile-time internal error.
1314 ((a|)+)+ did not notice that the outer group could match an empty string.
1315 (^a|^)+ was not marked as anchored.
1316 (.*a|.*)+ was not marked as matching at start or after a newline.
1317
1318 12. Yet more internal refactoring has removed another argument from the match()
1319 function. Special calls to this function are now indicated by setting a
1320 value in a variable in the "match data" data block.
1321
1322 13. Be more explicit in pcre_study() instead of relying on "default" for
1323 opcodes that mean there is no starting character; this means that when new
1324 ones are added and accidentally left out of pcre_study(), testing should
1325 pick them up.
1326
1327 14. The -s option of pcretest has been documented for ages as being an old
1328 synonym of -m (show memory usage). I have changed it to mean "force study
1329 for every regex", that is, assume /S for every regex. This is similar to -i
1330 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1331 using it. It makes it easier to run collections of tests with and without
1332 study enabled, and thereby test pcre_study() more easily. All the standard
1333 tests are now run with and without -s (but some patterns can be marked as
1334 "never study" - see 20 below).
1335
1336 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1337 restoration of the capturing data to the outer values was not happening
1338 correctly.
1339
1340 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1341 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1342 pattern had matched an empty string, and so incorrectly returned a no
1343 match.
1344
1345 17. There was optimizing code for the last branch of non-capturing parentheses,
1346 and also for the obeyed branch of a conditional subexpression, which used
1347 tail recursion to cut down on stack usage. Unfortunately, now that there is
1348 the possibility of (*THEN) occurring in these branches, tail recursion is
1349 no longer possible because the return has to be checked for (*THEN). These
1350 two optimizations have therefore been removed. [But see 8.20/11 above.]
1351
1352 18. If a pattern containing \R was studied, it was assumed that \R always
1353 matched two bytes, thus causing the minimum subject length to be
1354 incorrectly computed because \R can also match just one byte.
1355
1356 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1357 was incorrectly computed.
1358
1359 20. If /S is present twice on a test pattern in pcretest input, it now
1360 *disables* studying, thereby overriding the use of -s on the command line
1361 (see 14 above). This is necessary for one or two tests to keep the output
1362 identical in both cases.
1363
1364 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1365 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1366
1367 22. When an atomic group that contained a capturing parenthesis was
1368 successfully matched, but the branch in which it appeared failed, the
1369 capturing was not being forgotten if a higher numbered group was later
1370 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1371 group 1 to "a", when in fact it should be unset. This applied to multi-
1372 branched capturing and non-capturing groups, repeated or not, and also to
1373 positive assertions (capturing in negative assertions does not happen
1374 in PCRE) and also to nested atomic groups.
1375
1376 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1377 subject after a captured substring, to make it easier to tell which of a
1378 number of identical substrings has been captured.
1379
1380 24. The way atomic groups are processed by pcre_exec() has been changed so that
1381 if they are repeated, backtracking one repetition now resets captured
1382 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1383 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1384 "aaa". Previously, it would have been "a". As part of this code
1385 refactoring, the way recursive calls are handled has also been changed.
1386
1387 25. If an assertion condition captured any substrings, they were not passed
1388 back unless some other capturing happened later. For example, if
1389 (?(?=(a))a) was matched against "a", no capturing was returned.
1390
1391 26. When studying a pattern that contained subroutine calls or assertions,
1392 the code for finding the minimum length of a possible match was handling
1393 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1394 group 1 called group 2 while simultaneously a separate group 2 called group
1395 1). A stack overflow occurred in this case. I have fixed this by limiting
1396 the recursion depth to 10.
1397
1398 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1399 Fortmann. This supports explicit test numbers on the command line, and has
1400 argument validation and error reporting.
1401
1402 28. An instance of \X with an unlimited repeat could fail if at any point the
1403 first character it looked at was a mark character.
1404
1405 29. Some minor code refactoring concerning Unicode properties and scripts
1406 should reduce the stack requirement of match() slightly.
1407
1408 30. Added the '=' option to pcretest to check the setting of unused capturing
1409 slots at the end of the pattern, which are documented as being -1, but are
1410 not included in the return count.
1411
1412 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1413 compiled something random. Now it gives a compile-time error (as does
1414 Perl).
1415
1416 32. A *MARK encountered during the processing of a positive assertion is now
1417 recorded and passed back (compatible with Perl).
1418
1419 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1420 had alternative anchored branches, the search for a second match in a line
1421 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1422 matched the line "0102" twice. The same bug affected patterns that started
1423 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1424 twice.
1425
1426 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1427 does, and because of capturing effects, quantifying parenthesized
1428 assertions may at times be useful. Quantifiers are now allowed for
1429 parenthesized assertions.
1430
1431 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1432
1433 36. \g was being checked for fancy things in a character class, when it should
1434 just be a literal "g".
1435
1436 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1437 appearance of a nested POSIX class supersedes an apparent external class.
1438 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1439 unescaped square brackets may also appear as part of class names. For
1440 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1441 more like Perl. (But see 8.20/1 above.)
1442
1443 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1444 was because it thought it was \N{name}, which is not supported).
1445
1446 39. Add minix to OS list not supporting the -S option in pcretest.
1447
1448 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1449 cannot analyze patterns in sufficient detail to catch mutual recursions
1450 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1451 subgroup is called recursively as a subpattern for a second time at the
1452 same position in the subject string. In previous releases this might have
1453 been caught by the recursion limit, or it might have run out of stack.
1454
1455 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1456 happen only once. PCRE was, however incorrectly giving a compile time error
1457 "recursive call could loop indefinitely" because it cannot analyze the
1458 pattern in sufficient detail. The compile time test no longer happens when
1459 PCRE is compiling a conditional subpattern, but actual runaway loops are
1460 now caught at runtime (see 40 above).
1461
1462 42. It seems that Perl allows any characters other than a closing parenthesis
1463 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1464 has been changed to be the same.
1465
1466 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1467 as not to get warnings when autogen.sh is called. Also changed
1468 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1469
1470 44. To help people who use pcregrep to scan files containing exceedingly long
1471 lines, the following changes have been made:
1472
1473 (a) The default value of the buffer size parameter has been increased from
1474 8K to 20K. (The actual buffer used is three times this size.)
1475
1476 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1477 PCRE is built.
1478
1479 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1480 to be set at run time.
1481
1482 (d) Numerical values in pcregrep options can be followed by K or M, for
1483 example --buffer-size=50K.
1484
1485 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1486 given and the return code is set to 2.
1487
1488 45. Add a pointer to the latest mark to the callout data block.
1489
1490 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1491 partial match of an empty string instead of no match. This was specific to
1492 the use of ".".
1493
1494 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1495 complete match instead of a partial match. This bug was dependent on both
1496 the PCRE_UTF8 and PCRE_DOTALL options being set.
1497
1498 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1499 starting byte set, because \b was not being ignored.
1500
1501
1502 Version 8.12 15-Jan-2011
1503 ------------------------
1504
1505 1. Fixed some typos in the markup of the man pages, and wrote a script that
1506 checks for such things as part of the documentation building process.
1507
1508 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1509 --match-limit and --recursion-limit options (added for 8.11). In
1510 particular, this made one of the standard tests fail. (The integer value
1511 went into the wrong half of a long int.)
1512
1513 3. If the --colour option was given to pcregrep with -v (invert match), it
1514 did strange things, either producing crazy output, or crashing. It should,
1515 of course, ignore a request for colour when reporting lines that do not
1516 match.
1517
1518 4. Another pcregrep bug caused similar problems if --colour was specified with
1519 -M (multiline) and the pattern match finished with a line ending.
1520
1521 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1522 matched in multiline mode, the following line was shown as part of the
1523 match. This seems wrong, so I have changed it.
1524
1525 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1526 the check for further matches in the same line (so they could be coloured)
1527 to overrun the end of the current line. If another match was found, it was
1528 incorrectly shown (and then shown again when found in the next line).
1529
1530 7. If pcregrep was compiled under Windows, there was a reference to the
1531 function pcregrep_exit() before it was defined. I am assuming this was
1532 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1533 reported by a user. I've moved the definition above the reference.
1534
1535
1536 Version 8.11 10-Dec-2010
1537 ------------------------
1538
1539 1. (*THEN) was not working properly if there were untried alternatives prior
1540 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1541 backtracked to try for "b" instead of moving to the next alternative branch
1542 at the same level (in this case, to look for "c"). The Perl documentation
1543 is clear that when (*THEN) is backtracked onto, it goes to the "next
1544 alternative in the innermost enclosing group".
1545
1546 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1547 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1548 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1549 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1550 (*THEN).
1551
1552 3. If \s appeared in a character class, it removed the VT character from
1553 the class, even if it had been included by some previous item, for example
1554 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1555 of \s, but is part of the POSIX "space" class.)
1556
1557 4. A partial match never returns an empty string (because you can always
1558 match an empty string at the end of the subject); however the checking for
1559 an empty string was starting at the "start of match" point. This has been
1560 changed to the "earliest inspected character" point, because the returned
1561 data for a partial match starts at this character. This means that, for
1562 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1563 (previously it gave "no match").
1564
1565 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1566 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1567 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1568 has an implication that the given string is incomplete (because a partial
1569 match is preferred over a full match). For this reason, these items now
1570 give a partial match in this situation. [Aside: previously, the one case
1571 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1572 match rather than a full match, which was wrong by the old rules, but is
1573 now correct.]
1574
1575 6. There was a bug in the handling of #-introduced comments, recognized when
1576 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1577 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1578 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1579 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1580 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1581 places in pcre_compile().
1582
1583 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1584 comments when looking ahead for named forward references to subpatterns,
1585 the only newline sequence it recognized was NL. It now handles newlines
1586 according to the set newline convention.
1587
1588 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1589 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1590 cater for a lack of strerror(). These oversights have been fixed.
1591
1592 9. Added --match-limit and --recursion-limit to pcregrep.
1593
1594 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1595
1596 11. When the -o option was used, pcregrep was setting a return code of 1, even
1597 when matches were found, and --line-buffered was not being honoured.
1598
1599 12. Added an optional parentheses number to the -o and --only-matching options
1600 of pcregrep.
1601
1602 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1603 can match an empty string. The code to do it in pcretest and pcredemo
1604 needed fixing:
1605
1606 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1607 only one byte after an empty string match just before CRLF (this case
1608 just got forgotten; "any" and "anycrlf" were OK).
1609
1610 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1611 mode when an empty string match preceded an ASCII character followed by
1612 a non-ASCII character. (The code for advancing by one character rather
1613 than one byte was nonsense.)
1614
1615 (c) The pcredemo.c sample program did not have any code at all to handle
1616 the cases when CRLF is a valid newline sequence.
1617
1618 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1619 as a starting offset was within the subject string. There is now a new
1620 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1621 negative or greater than the length of the string. In order to test this,
1622 pcretest is extended to allow the setting of negative starting offsets.
1623
1624 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1625 starting offset points to the beginning of a UTF-8 character was
1626 unnecessarily clumsy. I tidied it up.
1627
1628 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1629 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1630
1631 17. Nobody had reported that the --include_dir option, which was added in
1632 release 7.7 should have been called --include-dir (hyphen, not underscore)
1633 for compatibility with GNU grep. I have changed it to --include-dir, but
1634 left --include_dir as an undocumented synonym, and the same for
1635 --exclude-dir, though that is not available in GNU grep, at least as of
1636 release 2.5.4.
1637
1638 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1639 characters from a string of bytes) have been redefined so as not to use
1640 loops, in order to improve performance in some environments. At the same
1641 time, I abstracted some of the common code into auxiliary macros to save
1642 repetition (this should not affect the compiled code).
1643
1644 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1645 compile-time error is now given if \c is not followed by an ASCII
1646 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1647 different, and any byte value is allowed.)
1648
1649 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1650 START_OPTIMIZE option, which is now allowed at compile time - but just
1651 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1652 to pcregrep and other applications that have no direct access to PCRE
1653 options. The new /Y option in pcretest sets this option when calling
1654 pcre_compile().
1655
1656 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1657 back references. Groups containing recursive back references were forced to
1658 be atomic by that change, but in the case of named groups, the amount of
1659 memory required was incorrectly computed, leading to "Failed: internal
1660 error: code overflow". This has been fixed.
1661
1662 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1663 pcretest.c, to avoid build problems in some Borland environments.
1664
1665
1666 Version 8.10 25-Jun-2010
1667 ------------------------
1668
1669 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1670 THEN.
1671
1672 2. (*ACCEPT) was not working when inside an atomic group.
1673
1674 3. Inside a character class, \B is treated as a literal by default, but
1675 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1676 causes the error). The code is unchanged, but I tidied the documentation.
1677
1678 4. Inside a character class, PCRE always treated \R and \X as literals,
1679 whereas Perl faults them if its -w option is set. I have changed PCRE so
1680 that it faults them when PCRE_EXTRA is set.
1681
1682 5. Added support for \N, which always matches any character other than
1683 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1684
1685 6. When compiling pcregrep with newer versions of gcc which may have
1686 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1687 declared with attribute warn_unused_result" were given. Just casting the
1688 result to (void) does not stop the warnings; a more elaborate fudge is
1689 needed. I've used a macro to implement this.
1690
1691 7. Minor change to pcretest.c to avoid a compiler warning.
1692
1693 8. Added four artifical Unicode properties to help with an option to make
1694 \s etc use properties (see next item). The new properties are: Xan
1695 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1696
1697 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1698 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1699 this option. Modified pcretest to add /W to test this facility. Added
1700 REG_UCP to make it available via the POSIX interface.
1701
1702 10. Added --line-buffered to pcregrep.
1703
1704 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1705 studied, and the match started with a letter with a code point greater than
1706 127 whose first byte was different to the first byte of the other case of
1707 the letter, the other case of this starting letter was not recognized
1708 (#976).
1709
1710 12. If a pattern that was studied started with a repeated Unicode property
1711 test, for example, \p{Nd}+, there was the theoretical possibility of
1712 setting up an incorrect bitmap of starting bytes, but fortunately it could
1713 not have actually happened in practice until change 8 above was made (it
1714 added property types that matched character-matching opcodes).
1715
1716 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1717 possible starting bytes for non-anchored patterns.
1718
1719 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1720 \R, and also a number of cases that involve Unicode properties, both
1721 explicit and implicit when PCRE_UCP is set.
1722
1723 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1724 input, it could crash or give wrong results if characters with values
1725 greater than 0xc0 were present in the subject string. (Detail: it assumed
1726 UTF-8 input when processing these items.)
1727
1728 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1729 size_t is 64-bit (#991).
1730
1731 17. Added a check for running out of memory when PCRE is compiled with
1732 --disable-stack-for-recursion (#990).
1733
1734 18. If the last data line in a file for pcretest does not have a newline on
1735 the end, a newline was missing in the output.
1736
1737 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1738 less than 128) in its various bitmaps. However, there is a facility for
1739 generating tables according to the current locale when PCRE is compiled. It
1740 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1741 characters, are recognized by isspace() and therefore were getting set in
1742 these tables, and indeed these tables seem to approximate to ISO 8859. This
1743 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1744 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1745 which of course cannot start UTF-8 characters. I have changed the code so
1746 that only real ASCII characters (less than 128) and the correct starting
1747 bytes for UTF-8 encodings are set for characters greater than 127 when in
1748 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1749 altogether.)
1750
1751 20. Added the /T option to pcretest so as to be able to run tests with non-
1752 standard character tables, thus making it possible to include the tests
1753 used for 19 above in the standard set of tests.
1754
1755 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1756 reference to a subpattern the other side of a comment that contains an
1757 opening parenthesis caused either an internal compiling error, or a
1758 reference to the wrong subpattern.
1759
1760
1761 Version 8.02 19-Mar-2010
1762 ------------------------
1763
1764 1. The Unicode data tables have been updated to Unicode 5.2.0.
1765
1766 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1767 configured.
1768
1769 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1770 original author of that file, following a query about its status.
1771
1772 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1773 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1774
1775 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1776 quantifier applied to a forward-referencing subroutine call, could compile
1777 incorrect code or give the error "internal error: previously-checked
1778 referenced subpattern not found".
1779
1780 6. Both MS Visual Studio and Symbian OS have problems with initializing
1781 variables to point to external functions. For these systems, therefore,
1782 pcre_malloc etc. are now initialized to local functions that call the
1783 relevant global functions.
1784
1785 7. There were two entries missing in the vectors called coptable and poptable
1786 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1787 I've fixed the data, and added a kludgy way of testing at compile time that
1788 the lengths are correct (equal to the number of opcodes).
1789
1790 8. Following on from 7, I added a similar kludge to check the length of the
1791 eint vector in pcreposix.c.
1792
1793 9. Error texts for pcre_compile() are held as one long string to avoid too
1794 much relocation at load time. To find a text, the string is searched,
1795 counting zeros. There was no check for running off the end of the string,
1796 which could happen if a new error number was added without updating the
1797 string.
1798
1799 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1800
1801 11. \K was not working if it appeared in an atomic group or in a group that
1802 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1803 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1804 the assertion is positive, but not if it is negative.
1805
1806 12. Change 11 fortuitously reduced the size of the stack frame used in the
1807 "match()" function of pcre_exec.c by one pointer. Forthcoming
1808 implementation of support for (*MARK) will need an extra pointer on the
1809 stack; I have reserved it now, so that the stack frame size does not
1810 decrease.
1811
1812 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1813 item in branch that calls a recursion is a subroutine call - as in the
1814 second branch in the above example - was incorrectly given the compile-
1815 time error "recursive call could loop indefinitely" because pcre_compile()
1816 was not correctly checking the subroutine for matching a non-empty string.
1817
1818 14. The checks for overrunning compiling workspace could trigger after an
1819 overrun had occurred. This is a "should never occur" error, but it can be
1820 triggered by pathological patterns such as hundreds of nested parentheses.
1821 The checks now trigger 100 bytes before the end of the workspace.
1822
1823 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1824
1825
1826 Version 8.01 19-Jan-2010
1827 ------------------------
1828
1829 1. If a pattern contained a conditional subpattern with only one branch (in
1830 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1831 computed the wrong minimum data length (which is of course zero for such
1832 subpatterns). This could cause incorrect "no match" results.
1833
1834 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1835 the pattern is reset in the first branch, pcre_compile() failed with
1836 "internal error: code overflow at offset...". This happened only when
1837 the reset was to the original external option setting. (An optimization
1838 abstracts leading options settings into an external setting, which was the
1839 cause of this.)
1840
1841 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1842 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1843 assertion pattern did not match (meaning that the assertion was true), it
1844 was incorrectly treated as false if the SKIP had been reached during the
1845 matching. This also applied to assertions used as conditions.
1846
1847 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1848 assertion subpattern, including such a pattern used as a condition,
1849 unpredictable results occurred, instead of the error return
1850 PCRE_ERROR_DFA_UITEM.
1851
1852 5. The C++ GlobalReplace function was not working like Perl for the special
1853 situation when an empty string is matched. It now does the fancy magic
1854 stuff that is necessary.
1855
1856 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1857 removed. (These were left over from very, very early versions of PCRE.)
1858
1859 7. Some cosmetic changes to the code to make life easier when compiling it
1860 as part of something else:
1861
1862 (a) Change DEBUG to PCRE_DEBUG.
1863
1864 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1865 called "current" as "current_branch", to prevent a collision with the
1866 Linux macro when compiled as a kernel module.
1867
1868 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1869 prevent a collision with the Linux macro when compiled as a kernel
1870 module.
1871
1872 8. In pcre_compile() there are some checks for integer overflows that used to
1873 cast potentially large values to (double). This has been changed to that
1874 when building, a check for int64_t is made, and if it is found, it is used
1875 instead, thus avoiding the use of floating point arithmetic. (There is no
1876 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1877 double.
1878
1879 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1880 2005 (difference between two addresses compared to an unsigned value).
1881
1882 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1883 custom one, because of the following reported problem in Windows:
1884
1885 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1886 under Win32.
1887 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1888 therefore missing the function definition.
1889 - The compiler thus generates a "C" signature for the test function.
1890 - The linker fails to find the "C" function.
1891 - PCRE fails to configure if asked to do so against libbz2.
1892
1893 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1894 messages were output:
1895
1896 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1897 rerunning libtoolize, to keep the correct libtool macros in-tree.
1898 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1899
1900 I have done both of these things.
1901
1902 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1903 most of the time, it *can* run out if it is given a pattern that contains a
1904 runaway infinite recursion. I updated the discussion in the pcrestack man
1905 page.
1906
1907 13. Now that we have gone to the x.xx style of version numbers, the minor
1908 version may start with zero. Using 08 or 09 is a bad idea because users
1909 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1910 interpreted as invalid octal numbers. I've updated the previous comment in
1911 configure.ac, and also added a check that gives an error if 08 or 09 are
1912 used.
1913
1914 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1915 causing partial matching to fail when the end of the subject matched \W
1916 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1917
1918 15. There were some discrepancies between the declarations in pcre_internal.h
1919 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1920 their definitions. The declarations used "const uschar *" and the
1921 definitions used USPTR. Even though USPTR is normally defined as "const
1922 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1923 reported that: "This difference in casting confuses some C++ compilers, for
1924 example, SunCC recognizes above declarations as different functions and
1925 generates broken code for hbpcre." I have changed the declarations to use
1926 USPTR.
1927
1928 16. GNU libtool is named differently on some systems. The autogen.sh script now
1929 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1930 (FreeBSD).
1931
1932 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1933 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1934 comment: "Figure out how to create a longlong from a string: strtoll and
1935 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1936 instance, but it only takes 2 args instead of 3!"
1937
1938 18. A subtle bug concerned with back references has been fixed by a change of
1939 specification, with a corresponding code fix. A pattern such as
1940 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1941 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1942 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1943 same bug. Such groups have to be quantified to be useful, or contained
1944 inside another quantified group. (If there's no repetition, the reference
1945 can never match.) The problem arises because, having left the group and
1946 moved on to the rest of the pattern, a later failure that backtracks into
1947 the group uses the captured value from the final iteration of the group
1948 rather than the correct earlier one. I have fixed this in PCRE by forcing
1949 any group that contains a reference to itself to be an atomic group; that
1950 is, there cannot be any backtracking into it once it has completed. This is
1951 similar to recursive and subroutine calls.
1952
1953
1954 Version 8.00 19-Oct-09
1955 ----------------------
1956
1957 1. The table for translating pcre_compile() error codes into POSIX error codes
1958 was out-of-date, and there was no check on the pcre_compile() error code
1959 being within the table. This could lead to an OK return being given in
1960 error.
1961
1962 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1963 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1964 in a Windows environment.
1965
1966 3. The pcregrep --count option prints the count for each file even when it is
1967 zero, as does GNU grep. However, pcregrep was also printing all files when
1968 --files-with-matches was added. Now, when both options are given, it prints
1969 counts only for those files that have at least one match. (GNU grep just
1970 prints the file name in this circumstance, but including the count seems
1971 more useful - otherwise, why use --count?) Also ensured that the
1972 combination -clh just lists non-zero counts, with no names.
1973
1974 4. The long form of the pcregrep -F option was incorrectly implemented as
1975 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1976 but it seems right to fix it, and I didn't think it was worth preserving
1977 the old behaviour.
1978
1979 5. The command line items --regex=pattern and --regexp=pattern were not
1980 recognized by pcregrep, which required --regex pattern or --regexp pattern
1981 (with a space rather than an '='). The man page documented the '=' forms,
1982 which are compatible with GNU grep; these now work.
1983
1984 6. No libpcreposix.pc file was created for pkg-config; there was just
1985 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1986
1987 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1988 when UCP support is not needed, by modifying the Python script that
1989 generates it from Unicode data files. This should not matter if the module
1990 is correctly used as a library, but I received one complaint about 50K of
1991 unwanted data. My guess is that the person linked everything into his
1992 program rather than using a library. Anyway, it does no harm.
1993
1994 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1995 was a minimum greater than 1 for a wide character in a possessive
1996 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1997 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1998 character. Chaos in the form of incorrect output or a compiling loop could
1999 result.
2000
2001 9. The restrictions on what a pattern can contain when partial matching is
2002 requested for pcre_exec() have been removed. All patterns can now be
2003 partially matched by this function. In addition, if there are at least two
2004 slots in the offset vector, the offset of the earliest inspected character
2005 for the match and the offset of the end of the subject are set in them when
2006 PCRE_ERROR_PARTIAL is returned.
2007
2008 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2009 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2010 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2011 and may be more useful for multi-segment matching.
2012
2013 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2014 used to be given if ever the end of the subject was reached; now it is
2015 given only if matching could not proceed because another character was
2016 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2017 string "Z", which now yields "no match" instead of "partial match". In the
2018 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2019 final character ended with (*FAIL).
2020
2021 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2022 if the pattern had a "must contain" character that was already found in the
2023 earlier partial match, unless partial matching was again requested. For
2024 example, with the pattern /dog.(body)?/, the "must contain" character is
2025 "g". If the first part-match was for the string "dog", restarting with
2026 "sbody" failed. This bug has been fixed.
2027
2028 13. The string returned by pcre_dfa_exec() after a partial match has been
2029 changed so that it starts at the first inspected character rather than the
2030 first character of the match. This makes a difference only if the pattern
2031 starts with a lookbehind assertion or \b or \B (\K is not supported by
2032 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2033 matching functions compatible, and I think it's the right thing to do.
2034
2035 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2036 so that the demonstration program is easily available in environments where
2037 PCRE has not been installed from source.
2038
2039 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2040 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2041 library.
2042
2043 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2044 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2045 is not the first non-POSIX option to be added. Clearly some people find
2046 these options useful.
2047
2048 17. If a caller to the POSIX matching function regexec() passes a non-zero
2049 value for nmatch with a NULL value for pmatch, the value of
2050 nmatch is forced to zero.
2051
2052 18. RunGrepTest did not have a test for the availability of the -u option of
2053 the diff command, as RunTest does. It now checks in the same way as
2054 RunTest, and also checks for the -b option.
2055
2056 19. If an odd number of negated classes containing just a single character
2057 interposed, within parentheses, between a forward reference to a named
2058 subpattern and the definition of the subpattern, compilation crashed with
2059 an internal error, complaining that it could not find the referenced
2060 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2061 [The bug was that it was starting one character too far in when skipping
2062 over the character class, thus treating the ] as data rather than
2063 terminating the class. This meant it could skip too much.]
2064
2065 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2066 /g option in pcretest when the pattern contains \K, which makes it possible
2067 to have an empty string match not at the start, even when the pattern is
2068 anchored. Updated pcretest and pcredemo to use this option.
2069
2070 21. If the maximum number of capturing subpatterns in a recursion was greater
2071 than the maximum at the outer level, the higher number was returned, but
2072 with unset values at the outer level. The correct (outer level) value is
2073 now given.
2074
2075 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2076 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2077 make it do so. The string so far is captured, making this feature
2078 compatible with Perl.
2079
2080 23. The tests have been re-organized, adding tests 11 and 12, to make it
2081 possible to check the Perl 5.10 features against Perl 5.10.
2082
2083 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2084 pattern matches a fixed length string. PCRE did not allow this; now it
2085 does. Neither allows recursion.
2086
2087 25. I finally figured out how to implement a request to provide the minimum
2088 length of subject string that was needed in order to match a given pattern.
2089 (It was back references and recursion that I had previously got hung up
2090 on.) This code has now been added to pcre_study(); it finds a lower bound
2091 to the length of subject needed. It is not necessarily the greatest lower
2092 bound, but using it to avoid searching strings that are too short does give
2093 some useful speed-ups. The value is available to calling programs via
2094 pcre_fullinfo().
2095
2096 26. While implementing 25, I discovered to my embarrassment that pcretest had
2097 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2098 study optimizations had never been tested with that matching function.
2099 Oops. What is worse, even when it was passed study data, there was a bug in
2100 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2101 were also very few tests of studied patterns with pcre_dfa_exec().
2102
2103 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2104 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2105 on the other side of the coin, they are no longer allowed to have different
2106 names, because these cannot be distinguished in PCRE, and this has caused
2107 confusion. (This is a difference from Perl.)
2108
2109 28. When duplicate subpattern names are present (necessarily with different
2110 numbers, as required by 27 above), and a test is made by name in a
2111 conditional pattern, either for a subpattern having been matched, or for
2112 recursion in such a pattern, all the associated numbered subpatterns are
2113 tested, and the overall condition is true if the condition is true for any
2114 one of them. This is the way Perl works, and is also more like the way
2115 testing by number works.
2116
2117
2118 Version 7.9 11-Apr-09
2119 ---------------------
2120
2121 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2122 (pcretest), all targets were linked against these libraries. This included
2123 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2124 libraries. This caused unwanted dependencies to be created. This problem
2125 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2126 pcretest is linked with readline.
2127
2128 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2129 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2130 moved outside it again, because FALSE and TRUE are already defined in AIX,
2131 but BOOL is not.
2132
2133 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2134 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2135
2136 4. The pcregrep documentation said spaces were inserted as well as colons (or
2137 hyphens) following file names and line numbers when outputting matching
2138 lines. This is not true; no spaces are inserted. I have also clarified the
2139 wording for the --colour (or --color) option.
2140
2141 5. In pcregrep, when --colour was used with -o, the list of matching strings
2142 was not coloured; this is different to GNU grep, so I have changed it to be
2143 the same.
2144
2145 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2146 each matching line was coloured. Now it goes on to look for further matches
2147 of any of the test patterns, which is the same behaviour as GNU grep.
2148
2149 7. A pattern that could match an empty string could cause pcregrep to loop; it
2150 doesn't make sense to accept an empty string match in pcregrep, so I have
2151 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2152 seems to be how GNU grep behaves. [But see later change 40 for release
2153 8.33.]
2154
2155 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2156 start or after a newline", because the conditional assertion was not being
2157 correctly handled. The rule now is that both the assertion and what follows
2158 in the first alternative must satisfy the test.
2159
2160 9. If auto-callout was enabled in a pattern with a conditional group whose
2161 condition was an assertion, PCRE could crash during matching, both with
2162 pcre_exec() and pcre_dfa_exec().
2163
2164 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2165 used for matching.
2166
2167 11. Unicode property support in character classes was not working for
2168 characters (bytes) greater than 127 when not in UTF-8 mode.
2169
2170 12. Added the -M command line option to pcretest.
2171
2172 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2173
2174 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2175
2176 16. Added comments and documentation about mis-use of no_arg in the C++
2177 wrapper.
2178
2179 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2180 from Martin Jerabek that uses macro names for all relevant character and
2181 string constants.
2182
2183 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2184 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2185 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2186 these, but not everybody uses configure.
2187
2188 19. A conditional group that had only one branch was not being correctly
2189 recognized as an item that could match an empty string. This meant that an
2190 enclosing group might also not be so recognized, causing infinite looping
2191 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2192 with the subject "ab", where knowledge that the repeated group can match
2193 nothing is needed in order to break the loop.
2194
2195 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2196 exec(), but without supplying a callout function, matching went wrong.
2197
2198 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2199 leak if the size of the offset vector was greater than 30. When the vector
2200 is smaller, the saved offsets during recursion go onto a local stack
2201 vector, but for larger vectors malloc() is used. It was failing to free
2202 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2203 error, in fact).
2204
2205 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2206 heapframe that is used only when UTF-8 support is enabled. This caused no
2207 problem, but was untidy.
2208
2209 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2210 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2211 included within another project.
2212
2213 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2214 slightly modified by me:
2215
2216 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2217 not building pcregrep.
2218
2219 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2220 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2221
2222 25. Forward references, both numeric and by name, in patterns that made use of
2223 duplicate group numbers, could behave incorrectly or give incorrect errors,
2224 because when scanning forward to find the reference group, PCRE was not
2225 taking into account the duplicate group numbers. A pattern such as
2226 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2227
2228 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2229 the feature of a custom pointer more persuasive (as requested by a user).
2230
2231 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2232 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2233 pre-defined.
2234
2235 28. Added support for (*UTF8) at the start of a pattern.
2236
2237 29. Arrange for flags added by the "release type" setting in CMake to be shown
2238 in the configuration summary.
2239
2240
2241 Version 7.8 05-Sep-08
2242 ---------------------
2243
2244 1. Replaced UCP searching code with optimized version as implemented for Ad
2245 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2246 stage table and inline lookup instead of a function, giving speed ups of 2
2247 to 5 times on some simple patterns that I tested. Permission was given to
2248 distribute the MultiStage2.py script that generates the tables (it's not in
2249 the tarball, but is in the Subversion repository).
2250
2251 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2252 scripts.
2253
2254 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2255 a group with a zero qualifier. The result of the study could be incorrect,
2256 or the function might crash, depending on the pattern.
2257
2258 4. Caseless matching was not working for non-ASCII characters in back
2259 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2260 It now works when Unicode Property Support is available.
2261
2262 5. In pcretest, an escape such as \x{de} in the data was always generating
2263 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2264 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2265 truncation.
2266
2267 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2268
2269 7. Added two (int) casts to pcregrep when printing the difference of two
2270 pointers, in case they are 64-bit values.
2271
2272 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2273 test 2 if it fails.
2274
2275 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2276 and a #define of that name to empty if it is not externally set. This is to
2277 allow users of MSVC to set it if necessary.
2278
2279 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2280 the convenience functions in the pcre_get.c source file.
2281
2282 11. An option change at the start of a pattern that had top-level alternatives
2283 could cause overwriting and/or a crash. This command provoked a crash in
2284 some environments:
2285
2286 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2287
2288 This potential security problem was recorded as CVE-2008-2371.
2289
2290 12. For a pattern where the match had to start at the beginning or immediately
2291 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2292 pcre_dfa_exec() could read past the end of the passed subject if there was
2293 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2294 pcretest so that it places the subject at the end of its malloc-ed buffer.
2295
2296 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2297 exec() might read past the end of the data buffer in UTF-8 mode.
2298
2299 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2300 the data contained the byte 0x85 as part of a UTF-8 character within its
2301 first line. This applied both to normal and DFA matching.
2302
2303 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2304 /^[^d]*?$/8 failed to match "abc".
2305
2306 16. Added a missing copyright notice to pcrecpp_internal.h.
2307
2308 17. Make it more clear in the documentation that values returned from
2309 pcre_exec() in ovector are byte offsets, not character counts.
2310
2311 18. Tidied a few places to stop certain compilers from issuing warnings.
2312
2313 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2314 supplied by Stefan Weber. I made a further small update for 7.8 because
2315 there is a change of source arrangements: the pcre_searchfuncs.c module is
2316 replaced by pcre_ucd.c.
2317
2318
2319 Version 7.7 07-May-08
2320 ---------------------
2321
2322 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2323 a string to a long long, pretend we don't even have a long long." This is
2324 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2325
2326 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2327 pre-7.6 versions, which defined a global no_arg variable instead of putting
2328 it in the RE class. (See also #8 below.)
2329
2330 3. Remove a line of dead code, identified by coverity and reported by Nuno
2331 Lopes.
2332
2333 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2334
2335 (1) The include/exclude patterns were being applied to the whole pathnames
2336 of files, instead of just to the final components.
2337
2338 (2) If there was more than one level of directory, the subdirectories were
2339 skipped unless they satisfied the include/exclude conditions. This is
2340 inconsistent with GNU grep (and could even be seen as contrary to the
2341 pcregrep specification - which I improved to make it absolutely clear).
2342 The action now is always to scan all levels of directory, and just
2343 apply the include/exclude patterns to regular files.
2344
2345 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2346 --exclude_dir in the tests to avoid scanning .svn directories.
2347
2348 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2349 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2350 doesn't support NULs in patterns.
2351
2352 7. Added some missing "const"s to declarations of static tables in
2353 pcre_compile.c and pcre_dfa_exec.c.
2354
2355 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2356 caused by fix #2 above. (Subsequently also a second patch to fix the
2357 first patch. And a third patch - this was a messy problem.)
2358
2359 9. Applied Craig's patch to remove the use of push_back().
2360
2361 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2362 matching function regexec().
2363
2364 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2365 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2366 references. PCRE supports relative numbers with this syntax (I don't think
2367 Oniguruma does).
2368
2369 12. Previously, a group with a zero repeat such as (...){0} was completely
2370 omitted from the compiled regex. However, this means that if the group
2371 was called as a subroutine from elsewhere in the pattern, things went wrong
2372 (an internal error was given). Such groups are now left in the compiled
2373 pattern, with a new opcode that causes them to be skipped at execution
2374 time.
2375
2376 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2377 to the way PCRE behaves:
2378
2379 (a) A lone ] character is dis-allowed (Perl treats it as data).
2380
2381 (b) A back reference to an unmatched subpattern matches an empty string
2382 (Perl fails the current match path).
2383
2384 (c) A data ] in a character class must be notated as \] because if the
2385 first data character in a class is ], it defines an empty class. (In
2386 Perl it is not possible to have an empty class.) The empty class []
2387 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2388 The negative empty class [^] matches any one character, independently
2389 of the DOTALL setting.
2390
2391 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2392 non-existent subpattern following a character class starting with ']' and
2393 containing () gave an internal compiling error instead of "reference to
2394 non-existent subpattern". Fortunately, when the pattern did exist, the
2395 compiled code was correct. (When scanning forwards to check for the
2396 existence of the subpattern, it was treating the data ']' as terminating
2397 the class, so got the count wrong. When actually compiling, the reference
2398 was subsequently set up correctly.)
2399
2400 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2401 it was being rejected as not supported by pcre_dfa_exec(), even though
2402 other assertions are supported. I have made pcre_dfa_exec() support
2403 (*FAIL).
2404
2405 16. The implementation of 13c above involved the invention of a new opcode,
2406 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2407 cannot be changed at match time, I realized I could make a small
2408 improvement to matching performance by compiling OP_ALLANY instead of
2409 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2410 on the OP_ANY path.
2411
2412 17. Compiling pcretest on Windows with readline support failed without the
2413 following two fixes: (1) Make the unistd.h include conditional on
2414 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2415
2416 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2417 ncurses library to be included for pcretest when ReadLine support is
2418 requested, but also to allow for it to be overridden. This patch came from
2419 Daniel Bergström.
2420
2421 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2422 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2423 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2424 spotting this.
2425
2426
2427 Version 7.6 28-Jan-08
2428 ---------------------
2429
2430 1. A character class containing a very large number of characters with
2431 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2432 overflow.
2433
2434 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2435 HAVE_LONG_LONG is not defined.
2436
2437 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2438 bring them up to date and include new features. This patch includes:
2439
2440 - Fixed PH's badly added libz and libbz2 support.
2441 - Fixed a problem with static linking.
2442 - Added pcredemo. [But later removed - see 7 below.]
2443 - Fixed dftables problem and added an option.
2444 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2445 HAVE_LONG_LONG.
2446 - Added readline support for pcretest.
2447 - Added an listing of the option settings after cmake has run.
2448
2449 4. A user submitted a patch to Makefile that makes it easy to create
2450 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2451 Makefile.am that cause it to include this special target, without
2452 affecting anything else. Note that the same mingw target plus all
2453 the other distribution libraries and programs are now supported
2454 when configuring with CMake (see 6 below) instead of with
2455 Configure/Make.
2456
2457 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2458 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2459 exported in the Windows port". It has not yet been confirmed that the patch
2460 solves the problem, but it does no harm.
2461
2462 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2463 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2464 with CMake, and also correct the comment about stack recursion.
2465
2466 7. Remove the automatic building of pcredemo from the ./configure system and
2467 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2468 of a program that users should build themselves after PCRE is installed, so
2469 building it automatically is not really right. What is more, it gave
2470 trouble in some build environments.
2471
2472 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2473
2474
2475 Version 7.5 10-Jan-08
2476 ---------------------
2477
2478 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2479 values in parens when parsing an RE using the C++ wrapper."
2480
2481 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2482 Characters greater than 255 were excluded from the class instead of being
2483 included.
2484
2485 3. The same bug as (2) above applied to negated POSIX classes such as
2486 [:^space:].
2487
2488 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2489 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2490 I have changed it.
2491
2492 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2493 first named subpattern) and a construct such as (?&a) would reference the
2494 first named subpattern whose name started with "a" (in other words, the
2495 length check was missing). Both these problems are fixed. "Subpattern name
2496 expected" is now given for (?&) (a zero-length name), and this patch also
2497 makes it give the same error for \k'' (previously it complained that that
2498 was a reference to a non-existent subpattern).
2499
2500 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2501 this is right because (?- can be followed by option settings as well as by
2502 digits. I have, however, made the messages clearer.
2503
2504 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2505 than the number used in the conditional) now cause a compile-time error.
2506 This is actually not compatible with Perl, which accepts such patterns, but
2507 treats the conditional as always being FALSE (as PCRE used to), but it
2508 seems to me that giving a diagnostic is better.
2509
2510 8. Change "alphameric" to the more common word "alphanumeric" in comments
2511 and messages.
2512
2513 9. Fix two occurrences of "backslash" in comments that should have been
2514 "backspace".
2515
2516 10. Remove two redundant lines of code that can never be obeyed (their function
2517 was moved elsewhere).
2518
2519 11. The program that makes PCRE's Unicode character property table had a bug
2520 which caused it to generate incorrect table entries for sequences of
2521 characters that have the same character type, but are in different scripts.
2522 It amalgamated them into a single range, with the script of the first of
2523 them. In other words, some characters were in the wrong script. There were
2524 thirteen such cases, affecting characters in the following ranges:
2525
2526 U+002b0 - U+002c1
2527 U+0060c - U+0060d
2528 U+0061e - U+00612
2529 U+0064b - U+0065e
2530 U+0074d - U+0076d
2531 U+01800 - U+01805
2532 U+01d00 - U+01d77
2533 U+01d9b - U+01dbf
2534 U+0200b - U+0200f
2535 U+030fc - U+030fe
2536 U+03260 - U+0327f
2537 U+0fb46 - U+0fbb1
2538 U+10450 - U+1049d
2539
2540 12. The -o option (show only the matching part of a line) for pcregrep was not
2541 compatible with GNU grep in that, if there was more than one match in a
2542 line, it showed only the first of them. It now behaves in the same way as
2543 GNU grep.
2544
2545 13. If the -o and -v options were combined for pcregrep, it printed a blank
2546 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2547 does the same. The return code can be used to tell if there were any
2548 non-matching lines.
2549
2550 14. Added --file-offsets and --line-offsets to pcregrep.
2551
2552 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2553 infinitely looping recursion. The bug was that positive lookaheads were not
2554 being skipped when checking for a possible empty match (negative lookaheads
2555 and both kinds of lookbehind were skipped).
2556
2557 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2558 inclusion of <windows.h> to before rather than after the definition of
2559 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2560
2561 17. Specifying a possessive quantifier with a specific limit for a Unicode
2562 character property caused pcre_compile() to compile bad code, which led at
2563 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2564 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2565 caused the error; without that there was no problem.
2566
2567 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2568
2569 19. Added --enable-pcretest-libreadline.
2570
2571 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2572 RE::GlobalReplace(). As a result, the number of replacements returned was
2573 double what it should be. I removed one of the increments, but Craig sent a
2574 later patch that removed the other one (the right fix) and added unit tests
2575 that check the return values (which was not done before).
2576
2577 21. Several CMake things:
2578
2579 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2580 the names libpcre and libpcreposix, not just pcre and pcreposix.
2581
2582 (2) The above change means that pcretest and pcregrep are now correctly
2583 linked with the newly-built libraries, not previously installed ones.
2584
2585 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2586
2587 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2588 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2589 UTF-8 newline character). The key issue is that the pattern starts .*;
2590 this means that the match must be either at the beginning, or after a
2591 newline. The bug was in the code for advancing after a failed match and
2592 checking that the new position followed a newline. It was not taking
2593 account of UTF-8 characters correctly.
2594
2595 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2596 character classes. PCRE was not treating the sequence [:...:] as a
2597 character class unless the ... were all letters. Perl, however, seems to
2598 allow any characters between [: and :], though of course it rejects as
2599 unknown any "names" that contain non-letters, because all the known class
2600 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2601 for example, whereas PCRE did not - it did not recognize a POSIX character
2602 class. This seemed a bit dangerous, so the code has been changed to be
2603 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2604 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2605 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2606 Perl does, and where it didn't before.
2607
2608 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2609 Windows environments %n is disabled by default.
2610
2611
2612 Version 7.4 21-Sep-07
2613 ---------------------
2614
2615 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2616 means that a class such as [\s] counted as "explicit reference to CR or
2617 LF". That isn't really right - the whole point of the change was to try to
2618 help when there was an actual mention of one of the two characters. So now
2619 the change happens only if \r or \n (or a literal CR or LF) character is
2620 encountered.
2621
2622 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2623 of both had grown to the point where there were only 3 bits left.
2624 Fortunately, there was spare space in the data structure, and so I have
2625 moved the internal flags into a new 16-bit field to free up more option
2626 bits.
2627
2628 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2629 but did not set the internal JCHANGED flag - either of these is enough to
2630 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2631 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2632 start sets both bits.
2633
2634 4. Added options (at build time, compile time, exec time) to change \R from
2635 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2636
2637 5. doc/pcresyntax.html was missing from the distribution.
2638
2639 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2640 compatibility, even though it is no longer used.
2641
2642 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2643 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2644 windows.h file is present (where different names are used). [This was
2645 reversed later after testing - see 16 below.]
2646
2647 8. Changed all #include <config.h> to #include "config.h". There were also
2648 some further <pcre.h> cases that I changed to "pcre.h".
2649
2650 9. When pcregrep was used with the --colour option, it missed the line ending
2651 sequence off the lines that it output.
2652
2653 10. It was pointed out to me that arrays of string pointers cause lots of
2654 relocations when a shared library is dynamically loaded. A technique of
2655 using a single long string with a table of offsets can drastically reduce
2656 these. I have refactored PCRE in four places to do this. The result is
2657 dramatic:
2658
2659 Originally: 290
2660 After changing UCP table: 187
2661 After changing error message table: 43
2662 After changing table of "verbs" 36
2663 After changing table of Posix names 22
2664
2665 Thanks to the folks working on Gregex for glib for this insight.
2666
2667 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2668 unicode-properties was also set.
2669
2670 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2671
2672 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2673 checked only for CRLF.
2674
2675 14. Added casts to pcretest.c to avoid compiler warnings.
2676
2677 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2678
2679 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2680 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2681 entirely. This removes changes made in 7 above.
2682
2683 17. The CMake files have been updated, and there is now more information about
2684 building with CMake in the NON-UNIX-USE document.
2685
2686
2687 Version 7.3 28-Aug-07
2688 ---------------------
2689
2690 1. In the rejigging of the build system that eventually resulted in 7.1, the
2691 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2692 brackets there is not right, since it causes compilers to look for an
2693 installed pcre.h, not the version that is in the source that is being
2694 compiled (which of course may be different). I have changed it back to:
2695
2696 #include "pcre.h"
2697
2698 I have a vague recollection that the change was concerned with compiling in
2699 different directories, but in the new build system, that is taken care of
2700 by the VPATH setting the Makefile.
2701
2702 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2703 when the subject happened to end in the byte 0x85 (e.g. if the last
2704 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2705 characters but of course it shouldn't be taken as a newline when it is part
2706 of another character. The bug was that, for an unlimited repeat of . in
2707 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2708 characters when looking for a newline.
2709
2710 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2711
2712 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2713 in debug output.
2714
2715 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2716 long printing in the pcrecpp unittest when running under MinGW.
2717
2718 6. ESC_K was left out of the EBCDIC table.
2719
2720 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2721 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2722 limit also applies to "virtual nesting" when a pattern is recursive, and in
2723 this case 1000 isn't so big. I have been able to remove this limit at the
2724 expense of backing off one optimization in certain circumstances. Normally,
2725 when pcre_exec() would call its internal match() function recursively and
2726 immediately return the result unconditionally, it uses a "tail recursion"
2727 feature to save stack. However, when a subpattern that can match an empty
2728 string has an unlimited repetition quantifier, it no longer makes this
2729 optimization. That gives it a stack frame in which to save the data for
2730 checking that an empty string has been matched. Previously this was taken
2731 from the 1000-entry workspace that had been reserved. So now there is no
2732 explicit limit, but more stack is used.
2733
2734 8. Applied Daniel's patches to solve problems with the import/export magic
2735 syntax that is required for Windows, and which was going wrong for the
2736 pcreposix and pcrecpp parts of the library. These were overlooked when this
2737 problem was solved for the main library.
2738
2739 9. There were some crude static tests to avoid integer overflow when computing
2740 the size of patterns that contain repeated groups with explicit upper
2741 limits. As the maximum quantifier is 65535, the maximum group length was
2742 set at 30,000 so that the product of these two numbers did not overflow a
2743 32-bit integer. However, it turns out that people want to use groups that
2744 are longer than 30,000 bytes (though not repeat them that many times).
2745 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2746 made it possible to implement the integer overflow checks in a much more
2747 dynamic way, which I have now done. The artificial limitation on group
2748 length has been removed - we now have only the limit on the total length of
2749 the compiled pattern, which depends on the LINK_SIZE setting.
2750
2751 10. Fixed a bug in the documentation for get/copy named substring when
2752 duplicate names are permitted. If none of the named substrings are set, the
2753 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2754 empty string.
2755
2756 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2757 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2758 because the ] is interpreted as the first data character and the
2759 terminating ] is not found. PCRE has been made compatible with Perl in this
2760 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2761 cause memory overwriting.
2762
2763 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2764 string has been matched (to stop an infinite loop). It was not recognizing
2765 a conditional subpattern that could match an empty string if that
2766 subpattern was within another subpattern. For example, it looped when
2767 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2768 condition was not nested. This bug has been fixed.
2769
2770 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2771 past the start of the subject in the presence of bytes with the top bit
2772 set, for example "\x8aBCD".
2773
2774 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2775 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2776
2777 14. Optimized (?!) to (*FAIL).
2778
2779 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2780 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2781 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2782 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2783 does: it's just the validity check that is more restrictive.
2784
2785 16. Inserted checks for integer overflows during escape sequence (backslash)
2786 processing, and also fixed erroneous offset values for syntax errors during
2787 backslash processing.
2788
2789 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2790 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2791
2792 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2793 caused an overrun.
2794
2795 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2796 something other than just ASCII characters) inside a group that had an
2797 unlimited repeat caused a loop at compile time (while checking to see
2798 whether the group could match an empty string).
2799
2800 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2801 [\P{Any}] did so. (Error in the code for printing property names.)
2802
2803 21. An orphan \E inside a character class could cause a crash.
2804
2805 22. A repeated capturing bracket such as (A)? could cause a wild memory
2806 reference during compilation.
2807
2808 23. There are several functions in pcre_compile() that scan along a compiled
2809 expression for various reasons (e.g. to see if it's fixed length for look
2810 behind). There were bugs in these functions when a repeated \p or \P was
2811 present in the pattern. These operators have additional parameters compared
2812 with \d, etc, and these were not being taken into account when moving along
2813 the compiled data. Specifically:
2814
2815 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2816 length.
2817
2818 (b) An item such as \pL+ within a repeated group could cause crashes or
2819 loops.
2820
2821 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2822 "reference to non-existent subpattern" error.
2823
2824 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2825
2826 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2827 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2828
2829 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2830
2831 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2832 character were causing crashes (broken optimization).
2833
2834 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2835 \p or \P) caused a compile-time loop.
2836
2837 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2838 break. For example, the unstudied pattern [\r\n]A does not match the string
2839 "\r\nA" because change 7.0/46 below moves the current point on by two
2840 characters after failing to match at the start. However, the pattern \nA
2841 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2842 the same is true. There doesn't seem any very clean way out of this, but
2843 what I have chosen to do makes the common cases work: PCRE now takes note
2844 of whether there can be an explicit match for \r or \n anywhere in the
2845 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2846 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2847 pattern has explicit CR or LF references.
2848
2849 29. Added (*CR) etc for changing newline setting at start of pattern.
2850
2851
2852 Version 7.2 19-Jun-07
2853 ---------------------
2854
2855 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2856 which is apparently normally available under Windows.
2857
2858 2. Re-jig the pcregrep tests with different newline settings in an attempt
2859 to make them independent of the local environment's newline setting.
2860
2861 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2862
2863 4. Some of the "internals" tests were previously cut out when the link size
2864 was not 2, because the output contained actual offsets. The recent new
2865 "Z" feature of pcretest means that these can be cut out, making the tests
2866 usable with all link sizes.
2867
2868 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2869 stack recursion. This gives a massive performance boost under BSD, but just
2870 a small improvement under Linux. However, it saves one field in the frame
2871 in all cases.
2872
2873 6. Added more features from the forthcoming Perl 5.10:
2874
2875 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2876 recursion call. It refers to the nth most recently opened parentheses.
2877
2878 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2879 to be opened parentheses.
2880
2881 (c) Conditions that refer to capturing parentheses can be specified
2882 relatively, for example, (?(-2)... or (?(+3)...
2883
2884 (d) \K resets the start of the current match so that everything before
2885 is not part of it.
2886
2887 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2888
2889 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2890 reference syntax.
2891
2892 (g) (?| introduces a group in which the numbering of parentheses in each
2893 alternative starts with the same number.
2894
2895 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2896
2897 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2898 PCRE_INFO_JCHANGED.
2899
2900 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2901 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2902 for detecting groups that can match an empty string.
2903
2904 9. A pattern with a very large number of alternatives (more than several
2905 hundred) was running out of internal workspace during the pre-compile
2906 phase, where pcre_compile() figures out how much memory will be needed. A
2907 bit of new cunning has reduced the workspace needed for groups with
2908 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2909 workspace instead of running out of the 4096 that are available.
2910
2911 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2912
2913 11. Applied patch from Google to remove an optimization that didn't quite work.
2914 The report of the bug said:
2915
2916 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2917 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2918 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2919
2920 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2921 it matched the wrong number of bytes.
2922
2923
2924 Version 7.1 24-Apr-07
2925 ---------------------
2926
2927 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2928 that is more "standard", making use of automake and other Autotools. There
2929 is some re-arrangement of the files and adjustment of comments consequent
2930 on this.
2931
2932 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2933 for recursive directory scanning broke on some systems because the files
2934 are not scanned in any specific order and on different systems the order
2935 was different. A call to "sort" has been inserted into RunGrepTest for the
2936 approprate test as a short-term fix. In the longer term there may be an
2937 alternative.
2938
2939 3. I had an email from Eric Raymond about problems translating some of PCRE's
2940 man pages to HTML (despite the fact that I distribute HTML pages, some
2941 people do their own conversions for various reasons). The problems
2942 concerned the use of low-level troff macros .br and .in. I have therefore
2943 removed all such uses from the man pages (some were redundant, some could
2944 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2945 HTML has been updated to handle .nf/.fi and to complain if it encounters
2946 .br or .in.
2947
2948 4. Updated comments in configure.ac that get placed in config.h.in and also
2949 arranged for config.h to be included in the distribution, with the name
2950 config.h.generic, for the benefit of those who have to compile without
2951 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2952
2953 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2954 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2955 makevp.bat for the current PCRE, using the additional files
2956 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2957
2958 6. A Windows user reported a minor discrepancy with test 2, which turned out
2959 to be caused by a trailing space on an input line that had got lost in his
2960 copy. The trailing space was an accident, so I've just removed it.
2961
2962 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2963 that is needed.
2964
2965 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2966 as "const" (a) because they are and (b) because it helps the PHP
2967 maintainers who have recently made a script to detect big data structures
2968 in the php code that should be moved to the .rodata section. I remembered
2969 to update Builducptable as well, so it won't revert if ucptable.h is ever
2970 re-created.
2971
2972 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2973 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2974 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2975 support is not required. This saves 1.5-2K of code, which is important in
2976 some applications.
2977
2978 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2979 so as not to refer to the tables, even though these functions will never be
2980 called when UTF-8 support is disabled. Otherwise there are problems with a
2981 shared library.
2982
2983 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2984
2985 (a) It was defining its arguments as char * instead of void *.
2986
2987 (b) It was assuming that all moves were upwards in memory; this was true
2988 a long time ago when I wrote it, but is no longer the case.
2989
2990 The emulated memove() is provided for those environments that have neither
2991 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2992 is clearly not the case, as these two bugs were recently reported.
2993
2994 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2995 and Detrail to create the HTML documentation, the .txt form of the man
2996 pages, and it removes trailing spaces from listed files. It also creates
2997 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2998 case, it wraps all the #defines with #ifndefs. This script should be run
2999 before "make dist".
3000
3001 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3002 with Unicode property support.
3003
3004 (a) For a maximizing quantifier, if the two different cases of the
3005 character were of different lengths in their UTF-8 codings (there are
3006 some cases like this - I found 11), and the matching function had to
3007 back up over a mixture of the two cases, it incorrectly assumed they
3008 were both the same length.
3009
3010 (b) When PCRE was configured to use the heap rather than the stack for
3011 recursion during matching, it was not correctly preserving the data for
3012 the other case of a UTF-8 character when checking ahead for a match
3013 while processing a minimizing repeat. If the check also involved
3014 matching a wide character, but failed, corruption could cause an
3015 erroneous result when trying to check for a repeat of the original
3016 character.
3017
3018 13. Some tidying changes to the testing mechanism:
3019
3020 (a) The RunTest script now detects the internal link size and whether there
3021 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3022 values substituted by "configure". (The RunGrepTest script already did
3023 this for UTF-8.) The configure.ac script no longer substitutes the
3024 relevant variables.
3025
3026 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3027 with length and offset values. This means that the output is different
3028 for different internal link sizes. Test 2 is skipped for link sizes
3029 other than 2 because of this, bypassing the problem. Unfortunately,
3030 there was also a test in test 3 (the locale tests) that used /B and
3031 failed for link sizes other than 2. Rather than cut the whole test out,
3032 I have added a new /Z option to pcretest that replaces the length and
3033 offset values with spaces. This is now used to make test 3 independent
3034 of link size. (Test 2 will be tidied up later.)
3035
3036 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3037 segmentation fault instead of returning the appropriate error message.
3038
3039 15. In multiline mode when the newline sequence was set to "any", the pattern
3040 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3041 This doesn't seem right; it now treats the CRLF combination as the line
3042 ending, and so does not match in that case. It's only a pattern such as ^$
3043 that would hit this one: something like ^ABC$ would have failed after \r
3044 and then tried again after \r\n.
3045
3046 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3047 in an attempt to make files that differ only in their line terminators
3048 compare equal. This works on Linux.
3049
3050 17. Under certain error circumstances pcregrep might try to free random memory
3051 as it exited. This is now fixed, thanks to valgrind.
3052
3053 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3054 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3055 was because its rules for how to advance for /g after matching an empty
3056 string at the end of a line did not allow for this case. They now check for
3057 it specially.
3058
3059 20. pcretest is supposed to handle patterns and data of any length, by
3060 extending its buffers when necessary. It was getting this wrong when the
3061 buffer for a data line had to be extended.
3062
3063 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3064 CRLF as a newline sequence.
3065
3066 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3067 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3068 I have nevertheless tidied it up.
3069
3070 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3071
3072 24. Added a man page for pcre-config.
3073
3074
3075 Version 7.0 19-Dec-06
3076 ---------------------
3077
3078 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3079 moving to gcc 4.1.1.
3080
3081 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3082 sys/time.h, which is documented as needed for this function. It doesn't
3083 seem to matter on Linux, but it showed up on some releases of OS X.
3084
3085 3. It seems that there are systems where bytes whose values are greater than
3086 127 match isprint() in the "C" locale. The "C" locale should be the
3087 default when a C program starts up. In most systems, only ASCII printing
3088 characters match isprint(). This difference caused the output from pcretest
3089 to vary, making some of the tests fail. I have changed pcretest so that:
3090
3091 (a) When it is outputting text in the compiled version of a pattern, bytes
3092 other than 32-126 are always shown as hex escapes.
3093
3094 (b) When it is outputting text that is a matched part of a subject string,
3095 it does the same, unless a different locale has been set for the match
3096 (using the /L modifier). In this case, it uses isprint() to decide.
3097
3098 4. Fixed a major bug that caused incorrect computation of the amount of memory
3099 required for a compiled pattern when options that changed within the
3100 pattern affected the logic of the preliminary scan that determines the
3101 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3102 that the computed length was too small. The symptoms of this bug were
3103 either the PCRE error "internal error: code overflow" from pcre_compile(),
3104 or a glibc crash with a message such as "pcretest: free(): invalid next
3105 size (fast)". Examples of patterns that provoked this bug (shown in
3106 pcretest format) are:
3107
3108 /(?-x: )/x
3109 /(?x)(?-x: \s*#\s*)/
3110 /((?i)[\x{c0}])/8
3111 /(?i:[\x{c0}])/8
3112
3113 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3114 is now done differently.
3115
3116 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3117 wrapper classes; (b) implement a new function in the C++ scanner that is
3118 more efficient than the old way of doing things because it avoids levels of
3119 recursion in the regex matching; (c) add a paragraph to the documentation
3120 for the FullMatch() function.
3121
3122 6. The escape sequence \n was being treated as whatever was defined as
3123 "newline". Not only was this contrary to the documentation, which states
3124 that \n is character 10 (hex 0A), but it also went horribly wrong when
3125 "newline" was defined as CRLF. This has been fixed.
3126
3127 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3128 was being set to -1 for the "end of line" case (supposedly a value that no
3129 character can have). Though this value is never used (the check for end of
3130 line is "zero bytes in current character"), it caused compiler complaints.
3131 I've changed it to 0xffffffff.
3132
3133 8. In pcre_version.c, the version string was being built by a sequence of
3134 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3135 string (as it is for production releases) called a macro with an empty
3136 argument. The C standard says the result of this is undefined. The gcc
3137 compiler treats it as an empty string (which was what was wanted) but it is
3138 reported that Visual C gives an error. The source has been hacked around to
3139 avoid this problem.
3140
3141 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3142 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3143 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3144 of them did).
3145
3146 10. Originally, pcretest opened its input and output without "b"; then I was
3147 told that "b" was needed in some environments, so it was added for release
3148 5.0 to both the input and output. (It makes no difference on Unix-like
3149 systems.) Later I was told that it is wrong for the input on Windows. I've
3150 now abstracted the modes into two macros, to make it easier to fiddle with
3151 them, and removed "b" from the input mode under Windows.
3152
3153 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3154
3155 12. Added -help and --help to pcretest as an official way of being reminded
3156 of the options.
3157
3158 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3159 and pcrecpp.cc because they annoy compilers at high warning levels.
3160
3161 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3162
3163 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3164 scripts are not C programs :-) and which was not noticed because it works
3165 on Linux.
3166
3167 16. pcretest is supposed to handle any length of pattern and data line (as one
3168 line or as a continued sequence of lines) by extending its input buffer if
3169 necessary. This feature was broken for very long pattern lines, leading to
3170 a string of junk being passed to pcre_compile() if the pattern was longer
3171 than about 50K.
3172
3173 17. I have done a major re-factoring of the way pcre_compile() computes the
3174 amount of memory needed for a compiled pattern. Previously, there was code
3175 that made a preliminary scan of the pattern in order to do this. That was
3176 OK when PCRE was new, but as the facilities have expanded, it has become
3177 harder and harder to keep it in step with the real compile phase, and there
3178 have been a number of bugs (see for example, 4 above). I have now found a
3179 cunning way of running the real compile function in a "fake" mode that
3180 enables it to compute how much memory it would need, while actually only
3181 ever using a few hundred bytes of working memory and without too many
3182 tests of the mode. This should make future maintenance and development
3183 easier. A side effect of this work is that the limit of 200 on the nesting
3184 depth of parentheses has been removed (though this was never a serious
3185 limitation, I suspect). However, there is a downside: pcre_compile() now
3186 runs more slowly than before (30% or more, depending on the pattern). I
3187 hope this isn't a big issue. There is no effect on runtime performance.
3188
3189 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3190 newline (only possible for the last line of a file) and it was a
3191 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3192
3193 19. Added additional timing features to pcretest. (1) The -tm option now times
3194 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3195 separate command line item, by a number that specifies the number of
3196 repeats to use when timing. The default is 50000; this gives better
3197 precision, but takes uncomfortably long for very large patterns.
3198
3199 20. Extended pcre_study() to be more clever in cases where a branch of a
3200 subpattern has no definite first character. For example, (a*|b*)[cd] would
3201 previously give no result from pcre_study(). Now it recognizes that the
3202 first character must be a, b, c, or d.
3203
3204 21. There was an incorrect error "recursive call could loop indefinitely" if
3205 a subpattern (or the entire pattern) that was being tested for matching an
3206 empty string contained only one non-empty item after a nested subpattern.
3207 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3208 incorrectly, because the \d was being skipped in the check.
3209
3210 22. The pcretest program now has a new pattern option /B and a command line
3211 option -b, which is equivalent to adding /B to every pattern. This causes
3212 it to show the compiled bytecode, without the additional information that
3213 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3214 is the same as /B/I).
3215
3216 23. A new optimization is now able automatically to treat some sequences such
3217 as a*b as a*+b. More specifically, if something simple (such as a character
3218 or a simple class like \d) has an unlimited quantifier, and is followed by
3219 something that cannot possibly match the quantified thing, the quantifier
3220 is automatically "possessified".
3221
3222 24. A recursive reference to a subpattern whose number was greater than 39
3223 went wrong under certain circumstances in UTF-8 mode. This bug could also
3224 have affected the operation of pcre_study().
3225
3226 25. Realized that a little bit of performance could be had by replacing
3227 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3228
3229 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3230
3231 27. Possessive quantifiers such as a++ were previously implemented by turning
3232 them into atomic groups such as ($>a+). Now they have their own opcodes,
3233 which improves performance. This includes the automatically created ones
3234 from 23 above.
3235
3236 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3237 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3238 the first matched character to be a colon. This applied both to named and
3239 numbered groups.
3240
3241 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3242
3243 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3244 building PCRE on Windows easier, so I have included it in the distribution.
3245
3246 31. There is now a check in pcretest against a ridiculously large number being
3247 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3248 loop, the loop is abandoned.
3249
3250 32. Forward references to subpatterns in conditions such as (?(2)...) where
3251 subpattern 2 is defined later cause pcre_compile() to search forwards in
3252 the pattern for the relevant set of parentheses. This search went wrong
3253 when there were unescaped parentheses in a character class, parentheses
3254 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3255
3256 33. "Subroutine" calls and backreferences were previously restricted to
3257 referencing subpatterns earlier in the regex. This restriction has now
3258 been removed.
3259
3260 34. Added a number of extra features that are going to be in Perl 5.10. On the
3261 whole, these are just syntactic alternatives for features that PCRE had
3262 previously implemented using the Python syntax or my own invention. The
3263 other formats are all retained for compatibility.
3264
3265 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3266 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3267 also .NET compatible.
3268
3269 (b) A recursion or subroutine call to a named group can now be defined as
3270 (?&name) as well as (?P>name).
3271
3272 (c) A backreference to a named group can now be defined as \k<name> or
3273 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3274 5.10, are also .NET compatible.
3275
3276 (d) A conditional reference to a named group can now use the syntax
3277 (?(<name>) or (?('name') as well as (?(name).
3278
3279 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3280 groups (named and numbered) that are never evaluated inline, but can be
3281 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3282 is always false. There may be only one alternative in such a group.
3283
3284 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3285 as the simple (?(R). The condition is true only if the most recent
3286 recursion is that of the given number or name. It does not search out
3287 through the entire recursion stack.
3288
3289 (g) The escape \gN or \g{N} has been added, where N is a positive or
3290 negative number, specifying an absolute or relative reference.
3291
3292 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3293 some "unreachable code" warnings.
3294
3295 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3296 things, this adds five new scripts.
3297
3298 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3299 There were also incompatibilities regarding the handling of \Q..\E inside
3300 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3301 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3302
3303 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3304 matches an empty string, and forcibly breaks the loop. There were bugs in
3305 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3306 against aaaa the result was just "a" rather than "aaaa", for example. Two
3307 separate and independent bugs (that affected different cases) have been
3308 fixed.
3309
3310 39. Refactored the code to abolish the use of different opcodes for small
3311 capturing bracket numbers. This is a tidy that I avoided doing when I
3312 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3313 The new approach is not only tidier, it makes it possible to reduce the
3314 memory needed to fix the previous bug (38).
3315
3316 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3317 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3318 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3319 mode.
3320
3321 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3322 report.
3323
3324 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3325 copy construction and assignment in the C++ wrapper.
3326
3327 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3328 couple of bugs that could have given wrong results in the "--newline=crlf"
3329 case.
3330
3331 44. Added a number of casts and did some reorganization of signed/unsigned int
3332 variables following suggestions from Dair Grant. Also renamed the variable
3333 "this" as "item" because it is a C++ keyword.
3334
3335 45. Arranged for dftables to add
3336
3337 #include "pcre_internal.h"
3338
3339 to pcre_chartables.c because without it, gcc 4.x may remove the array
3340 definition from the final binary if PCRE is built into a static library and
3341 dead code stripping is activated.
3342
3343 46. For an unanchored pattern, if a match attempt fails at the start of a
3344 newline sequence, and the newline setting is CRLF or ANY, and the next two
3345 characters are CRLF, advance by two characters instead of one.
3346
3347
3348 Version 6.7 04-Jul-06
3349 ---------------------
3350
3351 1. In order to handle tests when input lines are enormously long, pcretest has
3352 been re-factored so that it automatically extends its buffers when
3353 necessary. The code is crude, but this _is_ just a test program. The
3354 default size has been increased from 32K to 50K.
3355
3356 2. The code in pcre_study() was using the value of the re argument before
3357 testing it for NULL. (Of course, in any sensible call of the function, it
3358 won't be NULL.)
3359
3360 3. The memmove() emulation function in pcre_internal.h, which is used on
3361 systems that lack both memmove() and bcopy() - that is, hardly ever -
3362 was missing a "static" storage class specifier.
3363
3364 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3365 containing an extended class (one that cannot be represented by a bitmap
3366 because it contains high-valued characters or Unicode property items, e.g.
3367 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3368 pattern, but PCRE should not loop if you do not (it no longer does).
3369 [Detail: two cases were found: (a) a repeated subpattern containing an
3370 extended class; (b) a recursive reference to a subpattern that followed a
3371 previous extended class. It wasn't skipping over the extended class
3372 correctly when UTF-8 mode was not set.]
3373
3374 5. A negated single-character class was not being recognized as fixed-length
3375 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3376 compile error "lookbehind assertion is not fixed length".
3377
3378 6. The RunPerlTest auxiliary script was showing an unexpected difference
3379 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3380 write a Perl script that can interpret lines of an input file either as
3381 byte characters or as UTF-8, which is what "perltest" was being required to
3382 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3383 can't do is switch easily at run time between having the "use utf8;" pragma
3384 or not. In the end, I fudged it by using the RunPerlTest script to insert
3385 "use utf8;" explicitly for the UTF-8 tests.
3386
3387 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3388 the end of the subject string, contrary to the documentation and to what
3389 Perl does. This was true of both matching functions. Now it matches only at
3390 the start of the subject and immediately after *internal* newlines.
3391
3392 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3393 a pointer to an int instead of a pointer to an unsigned long int. This
3394 caused problems on 64-bit systems.
3395
3396 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3397 instance of the 'standard' template library not being so standard".
3398
3399 10. There was no check on the number of named subpatterns nor the maximum
3400 length of a subpattern name. The product of these values is used to compute
3401 the size of the memory block for a compiled pattern. By supplying a very
3402 long subpattern name and a large number of named subpatterns, the size
3403 computation could be caused to overflow. This is now prevented by limiting
3404 the length of names to 32 characters, and the number of named subpatterns
3405 to 10,000.
3406
3407 11. Subpatterns that are repeated with specific counts have to be replicated in
3408 the compiled pattern. The size of memory for this was computed from the
3409 length of the subpattern and the repeat count. The latter is limited to
3410 65535, but there was no limit on the former, meaning that integer overflow
3411 could in principle occur. The compiled length of a repeated subpattern is
3412 now limited to 30,000 bytes in order to prevent this.
3413
3414 12. Added the optional facility to have named substrings with the same name.
3415
3416 13. Added the ability to use a named substring as a condition, using the
3417 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3418 are numbers (not recommended). Forward references are permitted.
3419
3420 14. Added forward references in named backreferences (if you see what I mean).
3421
3422 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3423 pattern could run off the end of the subject. For example, the pattern
3424 "(?s)(.{1,5})"8 did this with the subject "ab".
3425
3426 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3427 PCRE_CASELESS was set when matching characters that were quantified with ?
3428 or *.
3429
3430 17. A character class other than a single negated character that had a minimum
3431 but no maximum quantifier - for example [ab]{6,} - was not handled
3432 correctly by pce_dfa_exec(). It would match only one character.
3433
3434 18. A valid (though odd) pattern that looked like a POSIX character
3435 class but used an invalid character after [ (for example [[,abc,]]) caused
3436 pcre_compile() to give the error "Failed: internal error: code overflow" or
3437 in some cases to crash with a glibc free() error. This could even happen if
3438 the pattern terminated after [[ but there just happened to be a sequence of
3439 letters, a binary zero, and a closing ] in the memory that followed.
3440
3441 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3442 over the years. Originally (before any Unicode support), just the bottom 8
3443 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3444 output from "man perlunicode" includes this:
3445
3446 The regular expression compiler produces polymorphic opcodes. That
3447 is, the pattern adapts to the data and automatically switches to
3448 the Unicode character scheme when presented with Unicode data--or
3449 instead uses a traditional byte scheme when presented with byte
3450 data.
3451
3452 Sadly, a wide octal escape does not cause a switch, and in a string with
3453 no other multibyte characters, these octal escapes are treated as before.
3454 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3455 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3456 Unicode string.
3457
3458 I have not perpetrated such confusion in PCRE. Up till now, it took just
3459 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3460 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3461 translate to the appropriate multibyte character.
3462
3463 29. Applied some refactoring to reduce the number of warnings from Microsoft
3464 and Borland compilers. This has included removing the fudge introduced
3465 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3466 a warning about an unused variable.
3467
3468 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3469 characters since release 4.0, because Perl (from release 5.004) does not.
3470 [Or at least, is documented not to: some releases seem to be in conflict
3471 with the documentation.] However, when a pattern was studied with
3472 pcre_study() and all its branches started with \s, PCRE still included VT
3473 as a possible starting character. Of course, this did no harm; it just
3474 caused an unnecessary match attempt.
3475
3476 22. Removed a now-redundant internal flag bit that recorded the fact that case
3477 dependency changed within the pattern. This was once needed for "required
3478 byte" processing, but is no longer used. This recovers a now-scarce options
3479 bit. Also moved the least significant internal flag bit to the most-
3480 significant bit of the word, which was not previously used (hangover from
3481 the days when it was an int rather than a uint) to free up another bit for
3482 the future.
3483
3484 23. Added support for CRLF line endings as well as CR and LF. As well as the
3485 default being selectable at build time, it can now be changed at runtime
3486 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3487 specify that it is scanning data with non-default line endings.
3488
3489 24. Changed the definition of CXXLINK to make it agree with the definition of
3490 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3491
3492 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3493 recursions. This makes a big different to stack usage for some patterns.
3494
3495 26. If a subpattern containing a named recursion or subroutine reference such
3496 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3497 the space required for the compiled pattern went wrong and gave too small a
3498 value. Depending on the environment, this could lead to "Failed: internal
3499 error: code overflow at offset 49" or "glibc detected double free or
3500 corruption" errors.
3501
3502 27. Applied patches from Google (a) to support the new newline modes and (b) to
3503 advance over multibyte UTF-8 characters in GlobalReplace.
3504
3505 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3506 difference for some implementation of PCRE in some Windows version.
3507
3508 29. Added some extra testing facilities to pcretest:
3509
3510 \q<number> in a data line sets the "match limit" value
3511 \Q<number> in a data line sets the "match recursion limt" value
3512 -S <number> sets the stack size, where <number> is in megabytes
3513
3514 The -S option isn't available for Windows.
3515
3516
3517 Version 6.6 06-Feb-06
3518 ---------------------
3519
3520 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3521 in pcreposix.h. I have copied the definition from pcre.h.
3522
3523 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3524 because pcre.h is no longer a built file.
3525
3526 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3527 not normally included in the compiled code.
3528
3529
3530 Version 6.5 01-Feb-06
3531 ---------------------
3532
3533 1. When using the partial match feature with pcre_dfa_exec(), it was not
3534 anchoring the second and subsequent partial matches at the new starting
3535 point. This could lead to incorrect results. For example, with the pattern
3536 /1234/, partially matching against "123" and then "a4" gave a match.
3537
3538 2. Changes to pcregrep:
3539
3540 (a) All non-match returns from pcre_exec() were being treated as failures
3541 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3542 error message is output. Some extra information is given for the
3543 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3544 probably the only errors that are likely to be caused by users (by
3545 specifying a regex that has nested indefinite repeats, for instance).
3546 If there are more than 20 of these errors, pcregrep is abandoned.
3547
3548 (b) A binary zero was treated as data while matching, but terminated the
3549 output line if it was written out. This has been fixed: binary zeroes
3550 are now no different to any other data bytes.
3551
3552 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3553 used to set a locale for matching. The --locale=xxxx long option has
3554 been added (no short equivalent) to specify a locale explicitly on the
3555 pcregrep command, overriding the environment variables.
3556
3557 (d) When -B was used with -n, some line numbers in the output were one less
3558 than they should have been.
3559
3560 (e) Added the -o (--only-matching) option.
3561
3562 (f) If -A or -C was used with -c (count only), some lines of context were
3563 accidentally printed for the final match.
3564
3565 (g) Added the -H (--with-filename) option.
3566
3567 (h) The combination of options -rh failed to suppress file names for files
3568 that were found from directory arguments.
3569
3570 (i) Added the -D (--devices) and -d (--directories) options.
3571
3572 (j) Added the -F (--fixed-strings) option.
3573
3574 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3575
3576 (l) Added the --colo(u)r option.
3577
3578 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3579 is not present by default.
3580
3581 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3582 items such as (?R) or (?1), when the recursion could match a number of
3583 alternatives. If it matched one of the alternatives, but subsequently,
3584 outside the recursion, there was a failure, the code tried to back up into
3585 the recursion. However, because of the way PCRE is implemented, this is not
3586 possible, and the result was an incorrect result from the match.
3587
3588 In order to prevent this happening, the specification of recursion has
3589 been changed so that all such subpatterns are automatically treated as
3590 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3591
3592 4. I had overlooked the fact that, in some locales, there are characters for
3593 which isalpha() is true but neither isupper() nor islower() are true. In
3594 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3595 and ordfeminine) are like this. This affected the treatment of \w and \W
3596 when they appeared in character classes, but not when they appeared outside
3597 a character class. The bit map for "word" characters is now created
3598 separately from the results of isalnum() instead of just taking it from the
3599 upper, lower, and digit maps. (Plus the underscore character, of course.)
3600
3601 5. The above bug also affected the handling of POSIX character classes such as
3602 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3603 permanent tables. Instead, the bit maps for such a class were previously
3604 created as the appropriate unions of the upper, lower, and digit bitmaps.
3605 Now they are created by subtraction from the [[:word:]] class, which has
3606 its own bitmap.
3607
3608 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3609 It is created by subtracting the vertical space characters (\x09, \x0a,
3610 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3611 subtraction was done in the overall bitmap for a character class, meaning
3612 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3613 be recognized. This bug has been fixed.
3614
3615 7. Patches from the folks at Google:
3616
3617 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3618 real life, but is still worth protecting against".
3619
3620 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3621 regular expressions".
3622
3623 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3624 have it.
3625
3626 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3627 "configure" and the latter not, in order to fix a problem somebody had
3628 with compiling the Arg class on HP-UX.
3629
3630 (e) Improve the error-handling of the C++ wrapper a little bit.
3631
3632 (f) New tests for checking recursion limiting.
3633
3634 8. The pcre_memmove() function, which is used only if the environment does not
3635 have a standard memmove() function (and is therefore rarely compiled),
3636 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3637 returning a result (though PCRE never actually uses the result).
3638
3639 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3640 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3641 returned instead of calling malloc() with an overflowing number that would
3642 most likely cause subsequent chaos.
3643
3644 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3645
3646 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3647 with this option is matched, the nmatch and pmatch options of regexec() are
3648 ignored.
3649
3650 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3651 provided in case anyone wants to the the POSIX interface with UTF-8
3652 strings.
3653
3654 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3655 C++ linking (needed for some HP-UX environments).
3656
3657 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3658 (unused parameter) and in the pcre_printint() function (omitted "default"
3659 switch label when the default is to do nothing).
3660
3661 15. Added some code to make it possible, when PCRE is compiled as a C++
3662 library, to replace subject pointers for pcre_exec() with a smart pointer
3663 class, thus making it possible to process discontinuous strings.
3664
3665 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3666 much the same function. They were added by different people who were trying
3667 to make PCRE easy to compile on non-Unix systems. It has been suggested
3668 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3669 for compiling on Windows systems. I have therefore replaced it with
3670 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3671 defaults to "extern" for C or "extern C" for C++, which works fine on
3672 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3673 SCOPE with something explicit in config.h. In addition:
3674
3675 (a) pcreposix.h still had just "extern" instead of either of these macros;
3676 I have replaced it with PCRE_DATA_SCOPE.
3677
3678 (b) Functions such as _pcre_xclass(), which are internal to the library,
3679 but external in the C sense, all had PCRE_EXPORT in their definitions.
3680 This is apparently wrong for the Windows case, so I have removed it.
3681 (It makes no difference on Unix-like systems.)
3682
3683 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3684 of recursive calls to match(). This is different to MATCH_LIMIT because
3685 that limits the total number of calls to match(), not all of which increase
3686 the depth of recursion. Limiting the recursion depth limits the amount of
3687 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3688 when PCRE is compiled, and changed at run time. A patch from Google adds
3689 this functionality to the C++ interface.
3690
3691 18. Changes to the handling of Unicode character properties:
3692
3693 (a) Updated the table to Unicode 4.1.0.
3694
3695 (b) Recognize characters that are not in the table as "Cn" (undefined).
3696
3697 (c) I revised the way the table is implemented to a much improved format
3698 which includes recognition of ranges. It now supports the ranges that
3699 are defined in UnicodeData.txt, and it also amalgamates other
3700 characters into ranges. This has reduced the number of entries in the
3701 table from around 16,000 to around 3,000, thus reducing its size
3702 considerably. I realized I did not need to use a tree structure after
3703 all - a binary chop search is just as efficient. Having reduced the
3704 number of entries, I extended their size from 6 bytes to 8 bytes to
3705 allow for more data.
3706
3707 (d) Added support for Unicode script names via properties such as \p{Han}.
3708
3709 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3710 matching that character.
3711
3712 20. When matching a repeated Unicode property with a minimum greater than zero,
3713 (for example \pL{2,}), PCRE could look past the end of the subject if it
3714 reached it while seeking the minimum number of characters. This could
3715 happen only if some of the characters were more than one byte long, because
3716 there is a check for at least the minimum number of bytes.
3717
3718 21. Refactored the implementation of \p and \P so as to be more general, to
3719 allow for more different types of property in future. This has changed the
3720 compiled form incompatibly. Anybody with saved compiled patterns that use
3721 \p or \P will have to recompile them.
3722
3723 22. Added "Any" and "L&" to the supported property types.
3724
3725 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3726 but give a compile time error if the value is greater than 0xff.
3727
3728 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3729 accidentally not being installed or uninstalled.
3730
3731 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3732 made were to insert the current release number. This seemed silly, because
3733 it made things harder for people building PCRE on systems that don't run
3734 "configure". I have turned pcre.h into a distributed file, no longer built
3735 by "configure", with the version identification directly included. There is
3736 no longer a pcre.h.in file.
3737
3738 However, this change necessitated a change to the pcre-config script as
3739 well. It is built from pcre-config.in, and one of the substitutions was the
3740 release number. I have updated configure.ac so that ./configure now finds
3741 the release number by grepping pcre.h.
3742
3743 26. Added the ability to run the tests under valgrind.
3744
3745
3746 Version 6.4 05-Sep-05
3747 ---------------------
3748
3749 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3750 "--" to be printed when multiple files were scanned, even when none of the
3751 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3752 consider it to be a bug, and have restored the previous behaviour.
3753
3754 2. A couple of code tidies to get rid of compiler warnings.
3755
3756 3. The pcretest program used to cheat by referring to symbols in the library
3757 whose names begin with _pcre_. These are internal symbols that are not
3758 really supposed to be visible externally, and in some environments it is
3759 possible to suppress them. The cheating is now confined to including
3760 certain files from the library's source, which is a bit cleaner.
3761
3762 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3763 file's purpose clearer.
3764
3765 5. Reorganized pcre_ucp_findchar().
3766
3767
3768 Version 6.3 15-Aug-05
3769 ---------------------
3770
3771 1. The file libpcre.pc.in did not have general read permission in the tarball.
3772
3773 2. There were some problems when building without C++ support:
3774
3775 (a) If C++ support was not built, "make install" and "make test" still
3776 tried to test it.
3777
3778 (b) There were problems when the value of CXX was explicitly set. Some
3779 changes have been made to try to fix these, and ...
3780
3781 (c) --disable-cpp can now be used to explicitly disable C++ support.
3782
3783 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3784 backslash in a target when C++ was disabled. This confuses some
3785 versions of "make", apparently. Using an intermediate variable solves
3786 this. (Same for CPP_LOBJ.)
3787
3788 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3789 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3790 necessary on certain architectures.
3791
3792 4. Added a setting of -export-symbols-regex to the link command to remove
3793 those symbols that are exported in the C sense, but actually are local
3794 within the library, and not documented. Their names all begin with
3795 "_pcre_". This is not a perfect job, because (a) we have to except some
3796 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3797 available (and never for static libraries). I have made a note to try to
3798 find a way round (a) in the future.
3799
3800
3801 Version 6.2 01-Aug-05
3802 ---------------------
3803
3804 1. There was no test for integer overflow of quantifier values. A construction
3805 such as {1111111111111111} would give undefined results. What is worse, if
3806 a minimum quantifier for a parenthesized subpattern overflowed and became
3807 negative, the calculation of the memory size went wrong. This could have
3808 led to memory overwriting.
3809
3810 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3811
3812 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3813 operating environments where this matters.
3814
3815 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3816 PCRE options from within the C++ wrapper.
3817
3818 5. Named capturing subpatterns were not being correctly counted when a pattern
3819 was compiled. This caused two problems: (a) If there were more than 100
3820 such subpatterns, the calculation of the memory needed for the whole
3821 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3822 back references of the form \12, where the number was greater than 9, were
3823 not recognized as back references, even though there were sufficient
3824 previous subpatterns.
3825
3826 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3827 versions of gcc, e.g. 2.95.4.
3828
3829
3830 Version 6.1 21-Jun-05
3831 ---------------------
3832
3833 1. There was one reference to the variable "posix" in pcretest.c that was not
3834 surrounded by "#if !defined NOPOSIX".
3835
3836 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3837 the cross-check on the old pcre_info() function, for the benefit of the
3838 cut-down version of PCRE that is currently imported into Exim.
3839
3840 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3841 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3842 patterns that start (?i) but I don't think that matters. The use of (?i) is
3843 just an example; this all applies to the other options as well.
3844
3845 4. Since libtool seems to echo the compile commands it is issuing, the output
3846 from "make" can be reduced a bit by putting "@" in front of each libtool
3847 compile command.
3848
3849 5. Patch from the folks at Google for configure.in to be a bit more thorough
3850 in checking for a suitable C++ installation before trying to compile the
3851 C++ stuff. This should fix a reported problem when a compiler was present,
3852 but no suitable headers.
3853
3854 6. The man pages all had just "PCRE" as their title. I have changed them to
3855 be the relevant file name. I have also arranged that these names are
3856 retained in the file doc/pcre.txt, which is a concatenation in text format
3857 of all the man pages except the little individual ones for each function.
3858
3859 7. The NON-UNIX-USE file had not been updated for the different set of source
3860 files that come with release 6. I also added a few comments about the C++
3861 wrapper.
3862
3863
3864 Version 6.0 07-Jun-05
3865 ---------------------
3866
3867 1. Some minor internal re-organization to help with my DFA experiments.
3868
3869 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3870 didn't matter for the library itself when fully configured, but did matter
3871 when compiling without UCP support, or within Exim, where the ucp files are
3872 not imported.
3873
3874 3. Refactoring of the library code to split up the various functions into
3875 different source modules. The addition of the new DFA matching code (see
3876 below) to a single monolithic source would have made it really too
3877 unwieldy, quite apart from causing all the code to be include in a
3878 statically linked application, when only some functions are used. This is
3879 relevant even without the DFA addition now that patterns can be compiled in
3880 one application and matched in another.
3881
3882 The downside of splitting up is that there have to be some external
3883 functions and data tables that are used internally in different modules of
3884 the library but which are not part of the API. These have all had their
3885 names changed to start with "_pcre_" so that they are unlikely to clash
3886 with other external names.
3887
3888 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3889 a different (DFA) algorithm. Although it is slower than the original
3890 function, it does have some advantages for certain types of matching
3891 problem.
3892
3893 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3894 including restarting after a partial match.
3895
3896 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3897 defined when compiling for Windows was sent to me. I have put it into the
3898 code, though I have no means of testing or verifying it.
3899
3900 7. Added the pcre_refcount() auxiliary function.
3901
3902 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3903 match before or at the first newline in the subject string. In pcretest,
3904 the /f option on a pattern can be used to set this.
3905
3906 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3907 would behave wrongly. This has been present in PCRE since release 4.0.
3908
3909 10. A number of changes to the pcregrep command:
3910
3911 (a) Refactored how -x works; insert ^(...)$ instead of setting
3912 PCRE_ANCHORED and checking the length, in preparation for adding
3913 something similar for -w.
3914
3915 (b) Added the -w (match as a word) option.
3916
3917 (c) Refactored the way lines are read and buffered so as to have more
3918 than one at a time available.
3919
3920 (d) Implemented a pcregrep test script.
3921
3922 (e) Added the -M (multiline match) option. This allows patterns to match
3923 over several lines of the subject. The buffering ensures that at least
3924 8K, or the rest of the document (whichever is the shorter) is available
3925 for matching (and similarly the previous 8K for lookbehind assertions).
3926
3927 (f) Changed the --help output so that it now says
3928
3929 -w, --word-regex(p)
3930
3931 instead of two lines, one with "regex" and the other with "regexp"
3932 because that confused at least one person since the short forms are the
3933 same. (This required a bit of code, as the output is generated
3934 automatically from a table. It wasn't just a text change.)
3935
3936 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3937 option but starts with a hyphen. Could be a pattern or a path name
3938 starting with a hyphen, for instance.
3939
3940 (h) "-" can be given as a file name to represent stdin.
3941
3942 (i) When file names are being printed, "(standard input)" is used for
3943 the standard input, for compatibility with GNU grep. Previously
3944 "<stdin>" was used.
3945
3946 (j) The option --label=xxx can be used to supply a name to be used for
3947 stdin when file names are being printed. There is no short form.
3948
3949 (k) Re-factored the options decoding logic because we are going to add
3950 two more options that take data. Such options can now be given in four
3951 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3952
3953 (l) Added the -A, -B, and -C options for requesting that lines of context
3954 around matches be printed.
3955
3956 (m) Added the -L option to print the names of files that do not contain
3957 any matching lines, that is, the complement of -l.
3958
3959 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3960 continue to scan other files.
3961
3962 (o) The -s option was incorrectly implemented. For compatibility with other
3963 greps, it now suppresses the error message for a non-existent or non-
3964 accessible file (but not the return code). There is a new option called
3965 -q that suppresses the output of matching lines, which was what -s was
3966 previously doing.
3967
3968 (p) Added --include and --exclude options to specify files for inclusion
3969 and exclusion when recursing.
3970
3971 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3972 Hopefully, it now does.
3973
3974 12. Missing cast in pcre_study().
3975
3976 13. Added an "uninstall" target to the makefile.
3977
3978 14. Replaced "extern" in the function prototypes in Makefile.in with
3979 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3980 world, but is set differently for Windows.
3981
3982 15. Added a second compiling function called pcre_compile2(). The only
3983 difference is that it has an extra argument, which is a pointer to an
3984 integer error code. When there is a compile-time failure, this is set
3985 non-zero, in addition to the error test pointer being set to point to an
3986 error message. The new argument may be NULL if no error number is required
3987 (but then you may as well call pcre_compile(), which is now just a
3988 wrapper). This facility is provided because some applications need a
3989 numeric error indication, but it has also enabled me to tidy up the way
3990 compile-time errors are handled in the POSIX wrapper.
3991
3992 16. Added VPATH=.libs to the makefile; this should help when building with one
3993 prefix path and installing with another. (Or so I'm told by someone who
3994 knows more about this stuff than I do.)
3995
3996 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3997 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3998 match everything, including newlines. This is not POSIX-compatible, but
3999 somebody wanted the feature. From pcretest it can be activated by using
4000 both the P and the s flags.
4001
4002 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4003
4004 19. libpcre.pc was being incorrectly installed as executable.
4005
4006 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4007 it now also looks for '\r' so that it will work unmodified on Windows.
4008
4009 21. Added Google's contributed C++ wrapper to the distribution.
4010
4011 22. Added some untidy missing memory free() calls in pcretest, to keep
4012 Electric Fence happy when testing.
4013
4014
4015
4016 Version 5.0 13-Sep-04
4017 ---------------------
4018
4019 1. Internal change: literal characters are no longer packed up into items
4020 containing multiple characters in a single byte-string. Each character
4021 is now matched using a separate opcode. However, there may be more than one
4022 byte in the character in UTF-8 mode.
4023
4024 2. The pcre_callout_block structure has two new fields: pattern_position and
4025 next_item_length. These contain the offset in the pattern to the next match
4026 item, and its length, respectively.
4027
4028 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4029 insertion of callouts before each pattern item. Added the /C option to
4030 pcretest to make use of this.
4031
4032 4. On the advice of a Windows user, the lines
4033
4034 #if defined(_WIN32) || defined(WIN32)
4035 _setmode( _fileno( stdout ), 0x8000 );
4036 #endif /* defined(_WIN32) || defined(WIN32) */
4037
4038 have been added to the source of pcretest. This apparently does useful
4039 magic in relation to line terminators.
4040
4041 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4042 for the benefit of those environments where the "b" makes a difference.
4043
4044 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4045 to know about it. I have put a hack into configure.in that adds in code
4046 to set GCC=yes if CC=icc. This seems to end up at a point in the
4047 generated configure script that is early enough to affect the setting of
4048 compiler options, which is what is needed, but I have no means of testing
4049 whether it really works. (The user who reported this had patched the
4050 generated configure script, which of course I cannot do.)
4051
4052 LATER: After change 22 below (new libtool files), the configure script
4053 seems to know about icc (and also ecc). Therefore, I have commented out
4054 this hack in configure.in.
4055
4056 7. Added support for pkg-config (2 patches were sent in).
4057
4058 8. Negated POSIX character classes that used a combination of internal tables
4059 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4060 [[:^ascii]]. Typically, they would match almost any characters. The other
4061 POSIX classes were not broken in this way.
4062
4063 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4064 to find the match, as PCRE was deluded into thinking that the match had to
4065 start at the start point or following a newline. The same bug applied to
4066 patterns with negative forward assertions or any backward assertions
4067 preceding ".*" at the start, unless the pattern required a fixed first
4068 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4069
4070 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4071 starting at the last subject character, bytes beyond the end of the subject
4072 string were read.
4073
4074 11. Renamed the variable "class" as "classbits" to make life easier for C++
4075 users. (Previously there was a macro definition, but it apparently wasn't
4076 enough.)
4077
4078 12. Added the new field "tables" to the extra data so that tables can be passed
4079 in at exec time, or the internal tables can be re-selected. This allows
4080 a compiled regex to be saved and re-used at a later time by a different
4081 program that might have everything at different addresses.
4082
4083 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4084 -R library as well as a -L library.
4085
4086 14. The debugging options of pcretest (-d on the command line or D on a
4087 pattern) showed incorrect output for anything following an extended class
4088 that contained multibyte characters and which was followed by a quantifier.
4089
4090 15. Added optional support for general category Unicode character properties
4091 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4092 support. It adds about 90K to the size of the library. The meanings of the
4093 inbuilt class escapes such as \d and \s have NOT been changed.
4094
4095 16. Updated pcredemo.c to include calls to free() to release the memory for the
4096 compiled pattern.
4097
4098 17. The generated file chartables.c was being created in the source directory
4099 instead of in the building directory. This caused the build to fail if the
4100 source directory was different from the building directory, and was
4101 read-only.
4102
4103 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4104 file. No doubt somebody will tell me if they don't make sense... Also added
4105 Dan Mooney's comments about building on OpenVMS.
4106
4107 19. Added support for partial matching via the PCRE_PARTIAL option for
4108 pcre_exec() and the \P data escape in pcretest.
4109
4110 20. Extended pcretest with 3 new pattern features:
4111
4112 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4113 write the compiled pattern to the file whose name is "rest-of-line".
4114 This is a straight binary dump of the data, with the saved pointer to
4115 the character tables forced to be NULL. The study data, if any, is
4116 written too. After writing, pcretest reads a new pattern.
4117
4118 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4119 compiled pattern from the given file. There must not be any
4120 occurrences of "<" in the file name (pretty unlikely); if there are,
4121 pcretest will instead treat the initial "<" as a pattern delimiter.
4122 After reading in the pattern, pcretest goes on to read data lines as
4123 usual.
4124
4125 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4126 and 16-bit fields in a compiled pattern, to simulate a pattern that
4127 was compiled on a host of opposite endianness.
4128
4129 21. The pcre-exec() function can now cope with patterns that were compiled on
4130 hosts of opposite endianness, with this restriction:
4131
4132 As for any compiled expression that is saved and used later, the tables
4133 pointer field cannot be preserved; the extra_data field in the arguments
4134 to pcre_exec() should be used to pass in a tables address if a value
4135 other than the default internal tables were used at compile time.
4136
4137 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4138 now diagnosed as an error. Previously, most of the time, a negative number
4139 would have been treated as zero, but if in addition "ovector" was passed as
4140 NULL, a crash could occur.
4141
4142 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4143 new versions from the libtool 1.5 distribution (the last one is a copy of
4144 a file called libtool.m4). This seems to have fixed the need to patch
4145 "configure" to support Darwin 1.3 (which I used to do). However, I still
4146 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4147 workstation).
4148
4149 24. Changed the PCRE licence to be the more standard "BSD" licence.
4150
4151
4152 Version 4.5 01-Dec-03
4153 ---------------------
4154
4155 1. There has been some re-arrangement of the code for the match() function so
4156 that it can be compiled in a version that does not call itself recursively.
4157 Instead, it keeps those local variables that need separate instances for
4158 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4159 needs to "recurse". Keeping track of where control must go is done by means
4160 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4161 hide most of the details from the main code, and operates only if
4162 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4163 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4164 operating.
4165
4166 To make it easier for callers to provide specially tailored get/free
4167 functions for this usage, two new functions, pcre_stack_malloc, and
4168 pcre_stack_free, are used. They are always called in strict stacking order,
4169 and the size of block requested is always the same.
4170
4171 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4172 PCRE has been compiled to use the stack or the heap for recursion. The
4173 -C option of pcretest uses this to show which version is compiled.
4174
4175 A new data escape \S, is added to pcretest; it causes the amounts of store
4176 obtained and freed by both kinds of malloc/free at match time to be added
4177 to the output.
4178
4179 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4180 what's available on my current Linux desktop machine.
4181
4182 3. When matching a UTF-8 string, the test for a valid string at the start has
4183 been extended. If start_offset is not zero, PCRE now checks that it points
4184 to a byte that is the start of a UTF-8 character. If not, it returns
4185 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4186 this is necessary because there may be backward assertions in the pattern.
4187 When matching the same subject several times, it may save resources to use
4188 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4189
4190 4. The code for checking the validity of UTF-8 strings has been tightened so
4191 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4192 containing "overlong sequences".
4193
4194 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4195 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4196 should have been "&", but it just so happened that all the cases this let
4197 through by mistake were picked up later in the function.
4198
4199 6. I had used a variable called "isblank" - this is a C99 function, causing
4200 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4201
4202 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4203 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4204 so the version is shown; (c) stop "make test" echoing "./RunTest".
4205
4206 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4207
4208 9. The prototype for memmove() for systems that don't have it was using
4209 size_t, but the inclusion of the header that defines size_t was later. I've
4210 moved the #includes for the C headers earlier to avoid this.
4211
4212 10. Added some adjustments to the code to make it easier to compiler on certain
4213 special systems:
4214
4215 (a) Some "const" qualifiers were missing.
4216 (b) Added the macro EXPORT before all exported functions; by default this
4217 is defined to be empty.
4218 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4219 that it reads its output file name as an argument instead of writing
4220 to the standard output and assuming this can be redirected.
4221
4222 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4223 class containing characters with values greater than 255, PCRE compilation
4224 went into a loop.
4225
4226 12. A recursive reference to a subpattern that was within another subpattern
4227 that had a minimum quantifier of zero caused PCRE to crash. For example,
4228 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4229 recursion. If the recursively-called subpattern itself had a zero repeat,
4230 that was OK.
4231
4232 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4233 buffer into which it was copied (for escape processing) was still set at
4234 1024, so long lines caused crashes.
4235
4236 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4237 "internal error: code overflow...". This applied to any character class
4238 that was followed by a possessive quantifier.
4239
4240 15. Modified the Makefile to add libpcre.la as a prerequisite for
4241 libpcreposix.la because I was told this is needed for a parallel build to
4242 work.
4243
4244 16. If a pattern that contained .* following optional items at the start was
4245 studied, the wrong optimizing data was generated, leading to matching
4246 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4247 matching string must start with a or b or c. The correct conclusion for
4248 this pattern is that a match can start with any character.
4249
4250
4251 Version 4.4 13-Aug-03
4252 ---------------------
4253
4254 1. In UTF-8 mode, a character class containing characters with values between
4255 127 and 255 was not handled correctly if the compiled pattern was studied.
4256 In fixing this, I have also improved the studying algorithm for such
4257 classes (slightly).
4258
4259 2. Three internal functions had redundant arguments passed to them. Removal
4260 might give a very teeny performance improvement.
4261
4262 3. Documentation bug: the value of the capture_top field in a callout is *one
4263 more than* the number of the hightest numbered captured substring.
4264
4265 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4266 in incorrectly linking with a previously installed version. They now link
4267 explicitly with libpcre.la.
4268
4269 5. configure.in no longer needs to recognize Cygwin specially.
4270
4271 6. A problem in pcre.in for Windows platforms is fixed.
4272
4273 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4274 pcretest, it used to include the size of the study block as part of its
4275 output. Unfortunately, the structure contains a field that has a different
4276 size on different hardware architectures. This meant that the tests that
4277 showed this size failed. As the block is currently always of a fixed size,
4278 this information isn't actually particularly useful in pcretest output, so
4279 I have just removed it.
4280
4281 8. Three pre-processor statements accidentally did not start in column 1.
4282 Sadly, there are *still* compilers around that complain, even though
4283 standard C has not required this for well over a decade. Sigh.
4284
4285 9. In pcretest, the code for checking callouts passed small integers in the
4286 callout_data field, which is a void * field. However, some picky compilers
4287 complained about the casts involved for this on 64-bit systems. Now
4288 pcretest passes the address of the small integer instead, which should get
4289 rid of the warnings.
4290
4291 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4292 both compile and run time, and gives an error if an invalid UTF-8 sequence
4293 is found. There is a option for disabling this check in cases where the
4294 string is known to be correct and/or the maximum performance is wanted.
4295
4296 11. In response to a bug report, I changed one line in Makefile.in from
4297
4298 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4299 to
4300 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4301
4302 to look similar to other lines, but I have no way of telling whether this
4303 is the right thing to do, as I do not use Windows. No doubt I'll get told
4304 if it's wrong...
4305
4306
4307 Version 4.3 21-May-03
4308 ---------------------
4309
4310 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4311 Makefile.
4312
4313 2. Some refactoring to improve the quality of the code:
4314
4315 (i) The utf8_table... variables are now declared "const".
4316
4317 (ii) The code for \cx, which used the "case flipping" table to upper case
4318 lower case letters, now just substracts 32. This is ASCII-specific,
4319 but the whole concept of \cx is ASCII-specific, so it seems
4320 reasonable.
4321
4322 (iii) PCRE was using its character types table to recognize decimal and
4323 hexadecimal digits in the pattern. This is silly, because it handles
4324 only 0-9, a-f, and A-F, but the character types table is locale-
4325 specific, which means strange things might happen. A private
4326 table is now used for this - though it costs 256 bytes, a table is
4327 much faster than multiple explicit tests. Of course, the standard
4328 character types table is still used for matching digits in subject
4329 strings against \d.
4330
4331 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4332 ending in _t are). So I've renamed it as ESC_tee.
4333
4334 3. The first argument for regexec() in the POSIX wrapper should have been
4335 defined as "const".
4336
4337 4. Changed pcretest to use malloc() for its buffers so that they can be
4338 Electric Fenced for debugging.
4339
4340 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4341 to read one or more bytes before the start of the subject string. Often this
4342 had no effect on PCRE's behaviour, but in some circumstances it could
4343 provoke a segmentation fault.
4344
4345 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4346 to try to read one or more bytes before the start of the subject string.
4347
4348 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4349 UTF-8 support could misbehave in various ways if the subject string
4350 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4351 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4352 back over UTF-8 characters.)
4353
4354
4355 Version 4.2 14-Apr-03
4356 ---------------------
4357
4358 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4359
4360 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4361 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4362 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4363 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4364 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4365 and BUILD_EXEEXT
4366 Note: automatic setting of the BUILD variables is not yet working
4367 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4368 compile-time but not at link-time
4369 [LINK]: use for linking executables only
4370 make different versions for Windows and non-Windows
4371 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4372 libraries
4373 [LINK_FOR_BUILD]: new variable
4374 [OBJEXT]: use throughout
4375 [EXEEXT]: use throughout
4376 <winshared>: new target
4377 <wininstall>: new target
4378 <dftables.o>: use native compiler
4379 <dftables>: use native linker
4380 <install>: handle Windows platform correctly
4381 <clean>: ditto
4382 <check>: ditto
4383 copy DLL to top builddir before testing
4384
4385 As part of these changes, -no-undefined was removed again. This was reported
4386 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4387 in any case.
4388
4389 3. Some tidies to get rid of compiler warnings:
4390
4391 . In the match_data structure, match_limit was an unsigned long int, whereas
4392 match_call_count was an int. I've made them both unsigned long ints.
4393
4394 . In pcretest the fact that a const uschar * doesn't automatically cast to
4395 a void * provoked a warning.
4396
4397 . Turning on some more compiler warnings threw up some "shadow" variables
4398 and a few more missing casts.
4399
4400 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4401 option, a class that contained a single character with a value between 128
4402 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4403
4404 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4405 option, a class that contained several characters, but with at least one
4406 whose value was between 128 and 255 caused PCRE to crash.
4407
4408
4409 Version 4.1 12-Mar-03
4410 ---------------------
4411
4412 1. Compiling with gcc -pedantic found a couple of places where casts were
4413 needed, and a string in dftables.c that was longer than standard compilers are
4414 required to support.
4415
4416 2. Compiling with Sun's compiler found a few more places where the code could
4417 be tidied up in order to avoid warnings.
4418
4419 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4420 first of these names is deprecated in the latest Autoconf in favour of the name
4421 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4422 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4423 analogy I have changed it to CFLAGS_FOR_BUILD.
4424
4425 4. Added -no-undefined to the linking command in the Makefile, because this is
4426 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4427 linking step for the pcreposix library.
4428
4429 5. PCRE was failing to diagnose the case of two named groups with the same
4430 name.
4431
4432 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4433 literal character that is needed in the subject for a match, and scans along to
4434 ensure that it is present before embarking on the full matching process. This
4435 saves time in cases of nested unlimited repeats that are never going to match.
4436 Problem: the scan can take a lot of time if the subject is very long (e.g.
4437 megabytes), thus penalizing straightforward matches. It is now done only if the
4438 amount of subject to be scanned is less than 1000 bytes.
4439
4440 7. A lesser problem with the same optimization is that it was recording the
4441 first character of an anchored pattern as "needed", thus provoking a search
4442 right along the subject, even when the first match of the pattern was going to
4443 fail. The "needed" character is now not set for anchored patterns, unless it
4444 follows something in the pattern that is of non-fixed length. Thus, it still
4445 fulfils its original purpose of finding quick non-matches in cases of nested
4446 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4447
4448
4449 Version 4.0 17-Feb-03
4450 ---------------------
4451
4452 1. If a comment in an extended regex that started immediately after a meta-item
4453 extended to the end of string, PCRE compiled incorrect data. This could lead to
4454 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4455
4456 2. Moved to autoconf 2.53 and libtool 1.4.2.
4457
4458 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4459 the special perltest8 script is no longer needed - all the tests can be run
4460 from a single perltest script.
4461
4462 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4463 by \s. It has now been removed in PCRE. This means it isn't recognized as
4464 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4465 class [:space:] *does* include VT, thereby creating a mess.
4466
4467 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4468 space and tab.
4469
4470 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4471 its new features into the main test script, reducing the number of scripts.
4472
4473 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4474 were backward compatible, and made the (?i) apply to the whole pattern, as if
4475 /i were given. Now it behaves more logically, and applies the option setting
4476 only to what follows. PCRE has been changed to follow suit. However, if it
4477 finds options settings right at the start of the pattern, it extracts them into
4478 the global options, as before. Thus, they show up in the info data.
4479
4480 8. Added support for the \Q...\E escape sequence. Characters in between are
4481 treated as literals. This is slightly different from Perl in that $ and @ are
4482 also handled as literals inside the quotes. In Perl, they will cause variable
4483 interpolation. Note the following examples:
4484
4485 Pattern PCRE matches Perl matches
4486
4487 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4488 \Qabc\$xyz\E abc\$xyz abc\$xyz
4489 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4490
4491 For compatibility with Perl, \Q...\E sequences are recognized inside character
4492 classes as well as outside them.
4493
4494 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4495 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4496 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4497 signed/unsigned warnings.
4498
4499 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4500 option for pcretest, so I've replaced it by a simple function that does just
4501 that job.
4502
4503 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4504 "pcregrep -".
4505
4506 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4507 Java package. This provides some syntactic sugar for simple cases of what my
4508 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4509 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4510 item, you can use this simplified notation. Note that only makes sense with
4511 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4512 greediness, whatever the setting of the PCRE_UNGREEDY option.
4513
4514 13. A change of greediness default within a pattern was not taking effect at
4515 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4516 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4517 was abstracted outside.
4518
4519 14. PCRE now supports the \G assertion. It is true when the current matching
4520 position is at the start point of the match. This differs from \A when the
4521 starting offset is non-zero. Used with the /g option of pcretest (or similar
4522 code), it works in the same way as it does for Perl's /g option. If all
4523 alternatives of a regex begin with \G, the expression is anchored to the start
4524 match position, and the "anchored" flag is set in the compiled expression.
4525
4526 15. Some bugs concerning the handling of certain option changes within patterns
4527 have been fixed. These applied to options other than (?ims). For example,
4528 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4529 been the other way round. Some of this was related to change 7 above.
4530
4531 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4532 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4533 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4534 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4535
4536 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4537 mode. Unlike ".", it always matches newline, whatever the setting of
4538 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4539 assertions. Perl allows it, but it doesn't (in general) work because it can't
4540 calculate the length of the lookbehind. At least, that's the case for Perl
4541 5.8.0 - I've been told they are going to document that it doesn't work in
4542 future.
4543
4544 18. Added an error diagnosis for escapes that PCRE does not support: these are
4545 \L, \l, \N, \P, \p, \U, \u, and \X.
4546
4547 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4548 reading past the end of the pattern in cases such as /[abcd/.
4549
4550 20. PCRE was getting more memory than necessary for patterns with classes that
4551 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4552
4553 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4554 compiling PCRE for use with Virtual Pascal.
4555
4556 22. Small fix to the Makefile to make it work properly if the build is done
4557 outside the source tree.
4558
4559 23. Added a new extension: a condition to go with recursion. If a conditional
4560 subpattern starts with (?(R) the "true" branch is used if recursion has
4561 happened, whereas the "false" branch is used only at the top level.
4562
4563 24. When there was a very long string of literal characters (over 255 bytes
4564 without UTF support, over 250 bytes with UTF support), the computation of how
4565 much memory was required could be incorrect, leading to segfaults or other
4566 strange effects.
4567
4568 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4569 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4570 there was a subsequent back reference to those brackets. This meant that, for
4571 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4572 possible to check for precisely this case. All we can do is abandon the
4573 optimization if .* occurs inside capturing brackets when there are any back
4574 references whatsoever. (See below for a better fix that came later.)
4575
4576 26. The handling of the optimization for finding the first character of a
4577 non-anchored pattern, and for finding a character that is required later in the
4578 match were failing in some cases. This didn't break the matching; it just
4579 failed to optimize when it could. The way this is done has been re-implemented.
4580
4581 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4582
4583 28. Added a new feature that provides some of the functionality that Perl
4584 provides with (?{...}). The facility is termed a "callout". The way it is done
4585 in PCRE is for the caller to provide an optional function, by setting
4586 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4587 global variable. By default it is unset, which disables all calling out. To get
4588 the function called, the regex must include (?C) at appropriate points. This
4589 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4590 This provides a means of identifying different callout points. When PCRE
4591 reaches such a point in the regex, if pcre_callout has been set, the external
4592 function is called. It is provided with data in a structure called
4593 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4594 matching continues; if it returns a non-zero value, the match at the current
4595 point fails. However, backtracking will occur if possible. [This was changed
4596 later and other features added - see item 49 below.]
4597
4598 29. pcretest is upgraded to test the callout functionality. It provides a
4599 callout function that displays information. By default, it shows the start of
4600 the match and the current position in the text. There are some new data escapes
4601 to vary what happens:
4602
4603 \C+ in addition, show current contents of captured substrings
4604 \C- do not supply a callout function
4605 \C!n return 1 when callout number n is reached
4606 \C!n!m return 1 when callout number n is reached for the mth time
4607
4608 30. If pcregrep was called with the -l option and just a single file name, it
4609 output "<stdin>" if a match was found, instead of the file name.
4610
4611 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4612 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4613 pcre_exec(). This saves a malloc/free per call. The default value of
4614 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4615 when configuring.
4616
4617 32. The default maximum size of a compiled pattern is 64K. There have been a
4618 few cases of people hitting this limit. The code now uses macros to handle the
4619 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4620 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4621 configuring. Tests 2 and 5 work only with 2-byte links because they output
4622 debugging information about compiled patterns.
4623
4624 33. Internal code re-arrangements:
4625
4626 (a) Moved the debugging function for printing out a compiled regex into
4627 its own source file (printint.c) and used #include to pull it into
4628 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4629 separate copies.
4630
4631 (b) Defined the list of op-code names for debugging as a macro in
4632 internal.h so that it is next to the definition of the opcodes.
4633
4634 (c) Defined a table of op-code lengths for simpler skipping along compiled
4635 code. This is again a macro in internal.h so that it is next to the
4636 definition of the opcodes.
4637
4638 34. Added support for recursive calls to individual subpatterns, along the
4639 lines of Robin Houston's patch (but implemented somewhat differently).
4640
4641 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4642 allow it to read and process whole directories in Win32. This code was
4643 contributed by Lionel Fourquaux; it has not been tested by me.
4644
4645 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4646 used to name a group. Names consist of alphanumerics and underscores, and must
4647 be unique. Back references use the syntax (?P=name) and recursive calls use
4648 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4649 numbers. The function pcre_fullinfo() can be used after compilation to extract
4650 a name/number map. There are three relevant calls:
4651
4652 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4653 PCRE_INFO_NAMECOUNT yields the number of entries
4654 PCRE_INFO_NAMETABLE yields a pointer to the map.
4655
4656 The map is a vector of fixed-size entries. The size of each entry depends on
4657 the length of the longest name used. The first two bytes of each entry are the
4658 group number, most significant byte first. There follows the corresponding
4659 name, zero terminated. The names are in alphabetical order.
4660
4661 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4662 case instead of 255. Making it the same both with and without UTF-8 support
4663 means that the same test output works with both.
4664
4665 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4666 calling malloc() with a zero argument.
4667
4668 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4669 optimization. I've improved things by keeping a bitmap of backreferences with
4670 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4671 fact referenced, the optimization can be applied. It is unlikely that a
4672 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4673 the match to follow \n) will appear inside brackets with a number greater than
4674 31, but if it does, any back reference > 31 suppresses the optimization.
4675
4676 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4677 of disabling numbered capturing parentheses. Any opening parenthesis that is
4678 not followed by ? behaves as if it were followed by ?: but named parentheses
4679 can still be used for capturing (and they will acquire numbers in the usual
4680 way).
4681
4682 41. Redesigned the return codes from the match() function into yes/no/error so
4683 that errors can be passed back from deep inside the nested calls. A malloc
4684 failure while inside a recursive subpattern call now causes the
4685 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4686
4687 42. It is now possible to set a limit on the number of times the match()
4688 function is called in a call to pcre_exec(). This facility makes it possible to
4689 limit the amount of recursion and backtracking, though not in a directly
4690 obvious way, because the match() function is used in a number of different
4691 circumstances. The count starts from zero for each position in the subject
4692 string (for non-anchored patterns). The default limit is, for compatibility, a
4693 large number, namely 10 000 000. You can change this in two ways:
4694
4695 (a) When configuring PCRE before making, you can use --with-match-limit=n
4696 to set a default value for the compiled library.
4697
4698 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4699 a different value is set. See 45 below.
4700
4701 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4702
4703 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4704 of things that can be changed at compile time. The first argument specifies
4705 what is wanted and the second points to where the information is to be placed.
4706 The current list of available information is:
4707
4708 PCRE_CONFIG_UTF8
4709
4710 The output is an integer that is set to one if UTF-8 support is available;
4711 otherwise it is set to zero.
4712
4713 PCRE_CONFIG_NEWLINE
4714
4715 The output is an integer that it set to the value of the code that is used for
4716 newline. It is either LF (10) or CR (13).
4717
4718 PCRE_CONFIG_LINK_SIZE
4719
4720 The output is an integer that contains the number of bytes used for internal
4721 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4722
4723 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4724
4725 The output is an integer that contains the threshold above which the POSIX
4726 interface uses malloc() for output vectors. See item 31 above.
4727
4728 PCRE_CONFIG_MATCH_LIMIT
4729
4730 The output is an unsigned integer that contains the default limit of the number
4731 of match() calls in a pcre_exec() execution. See 42 above.
4732
4733 44. pcretest has been upgraded by the addition of the -C option. This causes it
4734 to extract all the available output from the new pcre_config() function, and to
4735 output it. The program then exits immediately.
4736
4737 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4738 order to support additional features. One way would have been to define
4739 pcre_exec2() (for example) with extra arguments, but this would not have been
4740 extensible, and would also have required all calls to the original function to
4741 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4742 is used for passing in "extra" data from pcre_study().
4743
4744 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4745 contains the following fields:
4746
4747 flags a bitmap indicating which of the following fields are set
4748 study_data opaque data from pcre_study()
4749 match_limit a way of specifying a limit on match() calls for a specific
4750 call to pcre_exec()
4751 callout_data data for callouts (see 49 below)
4752
4753 The flag bits are also defined in pcre.h, and are
4754
4755 PCRE_EXTRA_STUDY_DATA
4756 PCRE_EXTRA_MATCH_LIMIT
4757 PCRE_EXTRA_CALLOUT_DATA
4758
4759 The pcre_study() function now returns one of these new pcre_extra blocks, with
4760 the actual study data pointed to by the study_data field, and the
4761 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4762 before. That is, this change is entirely upwards-compatible and requires no
4763 change to existing code.
4764
4765 If you want to pass in additional data to pcre_exec(), you can either place it
4766 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4767 block.
4768
4769 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4770 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4771 times with different match limits, until it finds the minimum value needed for
4772 pcre_exec() to complete. The value is then output. This can be instructive; for
4773 most simple matches the number is quite small, but for pathological cases it
4774 gets very large very quickly.
4775
4776 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4777 returns the size of the data block pointed to by the study_data field in a
4778 pcre_extra block, that is, the value that was passed as the argument to
4779 pcre_malloc() when PCRE was getting memory in which to place the information
4780 created by pcre_study(). The fourth argument should point to a size_t variable.
4781 pcretest has been extended so that this information is shown after a successful
4782 pcre_study() call when information about the compiled regex is being displayed.
4783
4784 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4785 because what follows is always an absolute path. (Later: it turns out that this
4786 is more than cosmetic for MinGW, because it doesn't like empty path
4787 components.)
4788
4789 49. Some changes have been made to the callout feature (see 28 above):
4790
4791 (i) A callout function now has three choices for what it returns:
4792
4793 0 => success, carry on matching
4794 > 0 => failure at this point, but backtrack if possible
4795 < 0 => serious error, return this value from pcre_exec()
4796
4797 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4798 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4799 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4800 use by callout functions. It will never be used by PCRE itself.
4801
4802 (ii) The pcre_extra structure (see 45 above) has a void * field called
4803 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4804 pcre_callout_block structure has a field of the same name. The contents of
4805 the field passed in the pcre_extra structure are passed to the callout
4806 function in the corresponding field in the callout block. This makes it
4807 easier to use the same callout-containing regex from multiple threads. For
4808 testing, the pcretest program has a new data escape
4809
4810 \C*n pass the number n (may be negative) as callout_data
4811
4812 If the callout function in pcretest receives a non-zero value as
4813 callout_data, it returns that value.
4814
4815 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4816 there were some redundant $(CFLAGS) in commands that are now specified as
4817 $(LINK), which already includes $(CFLAGS).
4818
4819 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4820 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4821 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4822 one-byte characters throughout. Note that case-insensitive matching applies
4823 only to characters whose values are less than 256. PCRE doesn't support the
4824 notion of cases for higher-valued characters.
4825
4826 (i) A character class whose characters are all within 0-255 is handled as
4827 a bit map, and the map is inverted for negative classes. Previously, a
4828 character > 255 always failed to match such a class; however it should
4829 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4830
4831 (ii) A negated character class with a single character < 255 is coded as
4832 "not this character" (OP_NOT). This wasn't working properly when the test
4833 character was multibyte, either singly or repeated.
4834
4835 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4836 mode, for example: \x{100}{2,3}.
4837
4838 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4839 singly or repeated) now correctly test multibyte characters. However,
4840 PCRE doesn't recognize any characters with values greater than 255 as
4841 digits, spaces, or word characters. Such characters always match \D, \S,
4842 and \W, and never match \d, \s, or \w.
4843
4844 (v) Classes may now contain characters and character ranges with values
4845 greater than 255. For example: [ab\x{100}-\x{400}].
4846
4847 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4848 PCRE in UTF-8 mode.
4849
4850 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4851 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4852 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4853 value.)
4854
4855 53. The single man page has become too large. I have therefore split it up into
4856 a number of separate man pages. These also give rise to individual HTML pages;
4857 these are now put in a separate directory, and there is an index.html page that
4858 lists them all. Some hyperlinking between the pages has been installed.
4859
4860 54. Added convenience functions for handling named capturing parentheses.
4861
4862 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4863 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4864 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4865 are faulted.
4866
4867 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4868 calling configure. These values are used when compiling the dftables.c program
4869 which is run to generate the source of the default character tables. They
4870 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4871 you will need to set these values.
4872
4873 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4874
4875
4876 Version 3.9 02-Jan-02
4877 ---------------------
4878
4879 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4880
4881 2. If --disable-static was given, the building process failed when trying to
4882 build pcretest and pcregrep. (For some reason it was using libtool to compile
4883 them, which is not right, as they aren't part of the library.)
4884
4885
4886 Version 3.8 18-Dec-01
4887 ---------------------
4888
4889 1. The experimental UTF-8 code was completely screwed up. It was packing the
4890 bytes in the wrong order. How dumb can you get?
4891
4892
4893 Version 3.7 29-Oct-01
4894 ---------------------
4895
4896 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4897 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4898 this didn't happen under Solaris 8, where I normally test things.
4899
4900 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4901 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4902 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4903 this fix an hour or so after the initial 3.7 release.)
4904
4905
4906 Version 3.6 23-Oct-01
4907 ---------------------
4908
4909 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4910 offsets passed as NULL with zero offset count.
4911
4912 2. The config.guess and config.sub files had not been updated when I moved to
4913 the latest autoconf.
4914
4915
4916 Version 3.5 15-Aug-01
4917 ---------------------
4918
4919 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4920 had been forgotten.
4921
4922 2. By using declared but undefined structures, we can avoid using "void"
4923 definitions in pcre.h while keeping the internal definitions of the structures
4924 private.
4925
4926 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4927 user point of view, this means that both static and shared libraries are built
4928 by default, but this can be individually controlled. More of the work of
4929 handling this static/shared cases is now inside libtool instead of PCRE's make
4930 file.
4931
4932 4. The pcretest utility is now installed along with pcregrep because it is
4933 useful for users (to test regexs) and by doing this, it automatically gets
4934 relinked by libtool. The documentation has been turned into a man page, so
4935 there are now .1, .txt, and .html versions in /doc.
4936
4937 5. Upgrades to pcregrep:
4938 (i) Added long-form option names like gnu grep.
4939 (ii) Added --help to list all options with an explanatory phrase.
4940 (iii) Added -r, --recursive to recurse into sub-directories.
4941 (iv) Added -f, --file to read patterns from a file.
4942
4943 6. pcre_exec() was referring to its "code" argument before testing that
4944 argument for NULL (and giving an error if it was NULL).
4945
4946 7. Upgraded Makefile.in to allow for compiling in a different directory from
4947 the source directory.
4948
4949 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4950 options bits, the pointer it was passed was to an int instead of to an unsigned
4951 long int. This mattered only on 64-bit systems.
4952
4953 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4954 generated) instead of pcre.in, which it its source. Also made the same change
4955 in several of the .c files.
4956
4957 10. A new release of gcc defines printf() as a macro, which broke pcretest
4958 because it had an ifdef in the middle of a string argument for printf(). Fixed
4959 by using separate calls to printf().
4960
4961 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4962 script, to force use of CR or LF instead of \n in the source. On non-Unix
4963 systems, the value can be set in config.h.
4964
4965 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4966 absolute limit. Changed the text of the error message to make this clear, and
4967 likewise updated the man page.
4968
4969 13. The limit of 99 on the number of capturing subpatterns has been removed.
4970 The new limit is 65535, which I hope will not be a "real" limit.
4971
4972
4973 Version 3.4 22-Aug-00
4974 ---------------------
4975
4976 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4977
4978 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4979
4980
4981 Version 3.3 01-Aug-00
4982 ---------------------
4983
4984 1. If an octal character was given, but the value was greater than \377, it
4985 was not getting masked to the least significant bits, as documented. This could
4986 lead to crashes in some systems.
4987
4988 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4989 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4990
4991 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4992 These just pass their arguments on to (pcre_free)(), but they are provided
4993 because some uses of PCRE bind it to non-C systems that can call its functions,
4994 but cannot call free() or pcre_free() directly.
4995
4996 4. Add "make test" as a synonym for "make check". Corrected some comments in
4997 the Makefile.
4998
4999 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5000 Makefile.
5001
5002 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5003 command called pgrep for grepping around the active processes.
5004
5005 7. Added the beginnings of support for UTF-8 character strings.
5006
5007 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5008 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5009 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5010 out for the ar command.)
5011
5012
5013 Version 3.2 12-May-00
5014 ---------------------
5015
5016 This is purely a bug fixing release.
5017
5018 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5019 of ZA. This was just one example of several cases that could provoke this bug,
5020 which was introduced by change 9 of version 2.00. The code for breaking
5021 infinite loops after an iteration that matches an empty string was't working
5022 correctly.
5023
5024 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5025 when matched against abbab (for example). After matching an empty string, it
5026 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5027 caused it to match further down the string than it should.
5028
5029 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5030 was there because it doesn't seem to be needed, and it causes trouble on some
5031 systems, as it is not a Standard C header. It has been removed.
5032
5033 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5034 were reported on the Macintosh. The changes were from
5035
5036 while ((c = *(++ptr)) != 0 && c != '\n');
5037 to
5038 while ((c = *(++ptr)) != 0 && c != '\n') ;
5039
5040 Totally extraordinary, but if that's what it takes...
5041
5042 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5043 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5044 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5045 assumes the way PCRE uses memmove() (always moving upwards).
5046
5047 6. PCRE is being used in one environment where strchr() is not available. There
5048 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5049 faster code anyway.
5050
5051
5052 Version 3.1 09-Feb-00
5053 ---------------------
5054
5055 The only change in this release is the fixing of some bugs in Makefile.in for
5056 the "install" target:
5057
5058 (1) It was failing to install pcreposix.h.
5059
5060 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5061
5062
5063 Version 3.0 01-Feb-00
5064 ---------------------
5065
5066 1. Add support for the /+ modifier to perltest (to output $` like it does in
5067 pcretest).
5068
5069 2. Add support for the /g modifier to perltest.
5070
5071 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5072 matches null strings.
5073
5074 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5075 pattern. Perl treats empty patterns specially - it reuses the most recent
5076 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5077 effect.
5078
5079 5. The POSIX interface was broken in that it was just handing over the POSIX
5080 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5081 required a bigger vector, with some working space on the end. This means that
5082 the POSIX wrapper now has to get and free some memory, and copy the results.
5083
5084 6. Added some simple autoconf support, placing the test data and the
5085 documentation in separate directories, re-organizing some of the
5086 information files, and making it build pcre-config (a GNU standard). Also added
5087 libtool support for building PCRE as a shared library, which is now the
5088 default.
5089
5090 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5091 09 are not valid octal constants. Single digits will be used for minor values
5092 less than 10.
5093
5094 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5095 existing programs that set these in the POSIX interface can use PCRE without
5096 modification.
5097
5098 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5099 return all that pcre_info() returns, plus additional data. The pcre_info()
5100 function is retained for compatibility, but is considered to be obsolete.
5101
5102 10. Added experimental recursion feature (?R) to handle one common case that
5103 Perl 5.6 will be able to do with (?p{...}).
5104
5105 11. Added support for POSIX character classes like [:alpha:], which Perl is
5106 adopting.
5107
5108
5109 Version 2.08 31-Aug-99
5110 ----------------------
5111
5112 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5113 trying to match at the startoffset position, but instead was moving forward to
5114 the next newline as if a previous match had failed.
5115
5116 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5117 and could get into a loop if a null string was matched other than at the start
5118 of the subject.
5119
5120 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5121 be distinguished at compile time, and for completeness also added PCRE_DATE.
5122
5123 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5124 in GnuWin32 environments.
5125
5126
5127 Version 2.07 29-Jul-99
5128 ----------------------
5129
5130 1. The documentation is now supplied in plain text form and HTML as well as in
5131 the form of man page sources.
5132
5133 2. C++ compilers don't like assigning (void *) values to other pointer types.
5134 In particular this affects malloc(). Although there is no problem in Standard
5135 C, I've put in casts to keep C++ compilers happy.
5136
5137 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5138 should be (const char *).
5139
5140 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5141 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5142 However, I haven't made this a standard facility. The documentation doesn't
5143 mention it, and the Makefile doesn't support it.
5144
5145 5. The Makefile now contains an "install" target, with editable destinations at
5146 the top of the file. The pcretest program is not installed.
5147
5148 6. pgrep -V now gives the PCRE version number and date.
5149
5150 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5151 causing the entire string to be ignored, instead of just the last character.
5152
5153 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5154 non-matching string, it can take a very, very long time, even for strings of
5155 quite modest length, because of the nested recursion. PCRE now does better in
5156 some of these cases. It does this by remembering the last required literal
5157 character in the pattern, and pre-searching the subject to ensure it is present
5158 before running the real match. In other words, it applies a heuristic to detect
5159 some types of certain failure quickly, and in the above example, if presented
5160 with a string that has no trailing " it gives "no match" very quickly.
5161
5162 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5163 other alternatives are tried instead.
5164
5165
5166 Version 2.06 09-Jun-99
5167 ----------------------
5168
5169 1. Change pcretest's output for amount of store used to show just the code
5170 space, because the remainder (the data block) varies in size between 32-bit and
5171 64-bit systems.
5172
5173 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5174 start matching at. This allows lookbehinds to work when searching for multiple
5175 occurrences in a string.
5176
5177 3. Added additional options to pcretest for testing multiple occurrences:
5178
5179 /+ outputs the rest of the string that follows a match
5180 /g loops for multiple occurrences, using the new startoffset argument
5181 /G loops for multiple occurrences by passing an incremented pointer
5182
5183 4. PCRE wasn't doing the "first character" optimization for patterns starting
5184 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5185 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5186 the letter 'x'. On long subject strings, this gives a significant speed-up.
5187
5188
5189 Version 2.05 21-Apr-99
5190 ----------------------
5191
5192 1. Changed the type of magic_number from int to long int so that it works
5193 properly on 16-bit systems.
5194
5195 2. Fixed a bug which caused patterns starting with .* not to work correctly
5196 when the subject string contained newline characters. PCRE was assuming
5197 anchoring for such patterns in all cases, which is not correct because .* will
5198 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5199 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5200 must be retried after every newline in the subject.
5201
5202
5203 Version 2.04 18-Feb-99
5204 ----------------------
5205
5206 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5207 computation of the store needed to hold the pattern was incorrect (too large).
5208 If such patterns were nested a few deep, this could multiply and become a real
5209 problem.
5210
5211 2. Added /M option to pcretest to show the memory requirement of a specific
5212 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5213
5214 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5215 compiled in such a way that the backtracking after subsequent failure was
5216 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5217 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5218
5219
5220 Version 2.03 02-Feb-99
5221 ----------------------
5222
5223 1. Fixed typo and small mistake in man page.
5224
5225 2. Added 4th condition (GPL supersedes if conflict) and created separate
5226 LICENCE file containing the conditions.
5227
5228 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5229 Perl, that is the internal \ allows the delimiter to be included in the
5230 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5231 the final delimiter, add \ to the end of the pattern (to test the error).
5232
5233 4. Added the convenience functions for extracting substrings after a successful
5234 match. Updated pcretest to make it able to test these functions.
5235
5236
5237 Version 2.02 14-Jan-99
5238 ----------------------
5239
5240 1. Initialized the working variables associated with each extraction so that
5241 their saving and restoring doesn't refer to uninitialized store.
5242
5243 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5244 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5245 fix the problem.
5246
5247 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5248 calls, and wasn't printing the correct value for compiling calls. Increased the
5249 default value of LOOPREPEAT, and the number of significant figures in the
5250 times.
5251
5252 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5253
5254 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5255 a building problem on Windows NT with a FAT file system.
5256
5257
5258 Version 2.01 21-Oct-98
5259 ----------------------
5260
5261 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5262 to character tables built by pcre_maketables() in the current locale. If NULL
5263 is passed, the default tables are used.
5264
5265
5266 Version 2.00 24-Sep-98
5267 ----------------------
5268
5269 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5270 it any more.
5271
5272 2. Allow quantification of (?>) groups, and make it work correctly.
5273
5274 3. The first character computation wasn't working for (?>) groups.
5275
5276 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5277 end of the subject) and add 5.005's \z, which really does match only at the
5278 very end of the subject.
5279
5280 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5281
5282 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5283 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5284 localized options. All options to pcre_study() were also removed.
5285
5286 7. Add other new features from 5.005:
5287
5288 $(?<= positive lookbehind
5289 $(?<! negative lookbehind
5290 (?imsx-imsx) added the unsetting capability
5291 such a setting is global if at outer level; local otherwise
5292 (?imsx-imsx:) non-capturing groups with option setting
5293 (?(cond)re|re) conditional pattern matching
5294
5295 A backreference to itself in a repeated group matches the previous
5296 captured string.
5297
5298 8. General tidying up of studying (both automatic and via "study")
5299 consequential on the addition of new assertions.
5300
5301 9. As in 5.005, unlimited repeated groups that could match an empty substring
5302 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5303 runtime if any iteration does actually match an empty substring.
5304
5305 10. Include the RunTest script in the distribution.
5306
5307 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5308 discrepancies, some of which were old and were also with respect to 5.004. They
5309 have now been fixed.
5310
5311
5312 Version 1.09 28-Apr-98
5313 ----------------------
5314
5315 1. A negated single character class followed by a quantifier with a minimum
5316 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5317 program crashes, or just wrong answers. This did not apply to negated classes
5318 containing more than one character, or to minima other than one.
5319
5320
5321 Version 1.08 27-Mar-98
5322 ----------------------
5323
5324 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5325
5326 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5327 latter must appear before anything that relies on it in the pattern.
5328
5329
5330 Version 1.07 16-Feb-98
5331 ----------------------
5332
5333 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5334 repeat of a potentially empty string).
5335
5336
5337 Version 1.06 23-Jan-98
5338 ----------------------
5339
5340 1. Added Markus Oberhumer's little patches for C++.
5341
5342 2. Literal strings longer than 255 characters were broken.
5343
5344
5345 Version 1.05 23-Dec-97
5346 ----------------------
5347
5348 1. Negated character classes containing more than one character were failing if
5349 PCRE_CASELESS was set at run time.
5350
5351
5352 Version 1.04 19-Dec-97
5353 ----------------------
5354
5355 1. Corrected the man page, where some "const" qualifiers had been omitted.
5356
5357 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5358 input syntax.
5359
5360 3. Fixed memory leak which occurred when a regex with back references was
5361 matched with an offsets vector that wasn't big enough. The temporary memory
5362 that is used in this case wasn't being freed if the match failed.
5363
5364 4. Tidied pcretest to ensure it frees memory that it gets.
5365
5366 5. Temporary memory was being obtained in the case where the passed offsets
5367 vector was exactly big enough.
5368
5369 6. Corrected definition of offsetof() from change 5 below.
5370
5371 7. I had screwed up change 6 below and broken the rules for the use of
5372 setjmp(). Now fixed.
5373
5374
5375 Version 1.03 18-Dec-97
5376 ----------------------
5377
5378 1. A erroneous regex with a missing opening parenthesis was correctly
5379 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5380 on some systems.
5381
5382 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5383 it was reported that one broken compiler failed on the former because "code" is
5384 also an independent variable.
5385
5386 3. The erroneous regex a[]b caused an array overrun reference.
5387
5388 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5389 fail on data ending with that character. (It was going on too far, and checking
5390 the next character, typically a binary zero.) This was specific to the
5391 optimized code for single-character negative classes.
5392
5393 5. Added a contributed patch from the TIN world which does the following:
5394
5395 + Add an undef for memmove, in case the the system defines a macro for it.
5396
5397 + Add a definition of offsetof(), in case there isn't one. (I don't know
5398 the reason behind this - offsetof() is part of the ANSI standard - but
5399 it does no harm).
5400
5401 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5402 most of the places where whitespace preceded '#'. I have given up and
5403 allowed the remaining 2 cases to be at the margin.
5404
5405 + Rename some variables in pcre to eliminate shadowing. This seems very
5406 pedantic, but does no harm, of course.
5407
5408 6. Moved the call to setjmp() into its own function, to get rid of warnings
5409 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5410
5411 7. Constructs such as \d{8,} were compiling into the equivalent of
5412 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5413 outcome, but in this particular case used more store than had been allocated,
5414 which caused the bug to be discovered because it threw up an internal error.
5415
5416 8. The debugging code in both pcre and pcretest for outputting the compiled
5417 form of a regex was going wrong in the case of back references followed by
5418 curly-bracketed repeats.
5419
5420
5421 Version 1.02 12-Dec-97
5422 ----------------------
5423
5424 1. Typos in pcre.3 and comments in the source fixed.
5425
5426 2. Applied a contributed patch to get rid of places where it used to remove
5427 'const' from variables, and fixed some signed/unsigned and uninitialized
5428 variable warnings.
5429
5430 3. Added the "runtest" target to Makefile.
5431
5432 4. Set default compiler flag to -O2 rather than just -O.
5433
5434
5435 Version 1.01 19-Nov-97
5436 ----------------------
5437
5438 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5439 like /([ab]*)*/, that is, for classes with more than one character in them.
5440
5441 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5442 as /((?>a*))*/ (a PCRE_EXTRA facility).
5443
5444
5445 Version 1.00 18-Nov-97
5446 ----------------------
5447
5448 1. Added compile-time macros to support systems such as SunOS4 which don't have
5449 memmove() or strerror() but have other things that can be used instead.
5450
5451 2. Arranged that "make clean" removes the executables.
5452
5453
5454 Version 0.99 27-Oct-97
5455 ----------------------
5456
5457 1. Fixed bug in code for optimizing classes with only one character. It was
5458 initializing a 32-byte map regardless, which could cause it to run off the end
5459 of the memory it had got.
5460
5461 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5462
5463
5464 Version 0.98 22-Oct-97
5465 ----------------------
5466
5467 1. Fixed bug in code for handling temporary memory usage when there are more
5468 back references than supplied space in the ovector. This could cause segfaults.
5469
5470
5471 Version 0.97 21-Oct-97
5472 ----------------------
5473
5474 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5475
5476 2. Optimized negated single characters not to use a bit map.
5477
5478 3. Brought error texts together as macro definitions; clarified some of them;
5479 fixed one that was wrong - it said "range out of order" when it meant "invalid
5480 escape sequence".
5481
5482 4. Changed some char * arguments to const char *.
5483
5484 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5485
5486 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5487 pcretest.
5488
5489
5490 Version 0.96 16-Oct-97
5491 ----------------------
5492
5493 1. Added a simple "pgrep" utility to the distribution.
5494
5495 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5496 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5497 where "ddd" means "one or more decimal digits".
5498
5499 3. Fixed serious bug. If a pattern had a back reference, but the call to
5500 pcre_exec() didn't supply a large enough ovector to record the related
5501 identifying subpattern, the match always failed. PCRE now remembers the number
5502 of the largest back reference, and gets some temporary memory in which to save
5503 the offsets during matching if necessary, in order to ensure that
5504 backreferences always work.
5505
5506 4. Increased the compatibility with Perl in a number of ways:
5507
5508 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5509 to request this handling. The option can be set at compile or exec time.
5510
5511 (b) $ matches before a terminating newline by default; an option
5512 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5513 mode). The option can be set at compile or exec time.
5514
5515 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5516 the same as Perl's. If the decimal number it represents is less than 10
5517 or there aren't that many previous left capturing parentheses, an octal
5518 escape is read. Inside a character class, it's always an octal escape,
5519 even if it is a single digit.
5520
5521 (d) An escaped but undefined alphabetic character is taken as a literal,
5522 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5523 escapes.
5524
5525 (e) {0} is now permitted. (The previous item is removed from the compiled
5526 pattern).
5527
5528 5. Changed all the names of code files so that the basic parts are no longer
5529 than 10 characters, and abolished the teeny "globals.c" file.
5530
5531 6. Changed the handling of character classes; they are now done with a 32-byte
5532 bit map always.
5533
5534 7. Added the -d and /D options to pcretest to make it possible to look at the
5535 internals of compilation without having to recompile pcre.
5536
5537
5538 Version 0.95 23-Sep-97
5539 ----------------------
5540
5541 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5542 \x20 at the start of a run of normal characters. These were being treated as
5543 real characters, instead of the source characters being re-checked.
5544
5545
5546 Version 0.94 18-Sep-97
5547 ----------------------
5548
5549 1. The functions are now thread-safe, with the caveat that the global variables
5550 containing pointers to malloc() and free() or alternative functions are the
5551 same for all threads.
5552
5553 2. Get pcre_study() to generate a bitmap of initial characters for non-
5554 anchored patterns when this is possible, and use it if passed to pcre_exec().
5555
5556
5557 Version 0.93 15-Sep-97
5558 ----------------------
5559
5560 1. /(b)|(:+)/ was computing an incorrect first character.
5561
5562 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5563 but not actually doing anything yet.
5564
5565 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5566 as Perl does (e.g. [-az] or [az-]).
5567
5568 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5569 all possible positions.
5570
5571 5. Split up into different modules to avoid including unneeded functions in a
5572 compiled binary. However, compile and exec are still in one module. The "study"
5573 function is split off.
5574
5575 6. The character tables are now in a separate module whose source is generated
5576 by an auxiliary program - but can then be edited by hand if required. There are
5577 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5578 toupper() in the code.
5579
5580 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5581 make them global. Abolish the function for setting them, as the caller can now
5582 set them directly.
5583
5584
5585 Version 0.92 11-Sep-97
5586 ----------------------
5587
5588 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5589 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5590
5591 2. Caseless matching was not working in character classes if the characters in
5592 the pattern were in upper case.
5593
5594 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5595
5596 4. Make PCRE_ANCHORED public and accept as a compile option.
5597
5598 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5599 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5600 pass them.
5601
5602 6. Give an error if bad option bits passed at compile or run time.
5603
5604 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5605 pcretest to cause it to pass that flag.
5606
5607 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5608 options, and the first character, if set.
5609
5610 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5611
5612
5613 Version 0.91 10-Sep-97
5614 ----------------------
5615
5616 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5617 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5618
5619 2. PCRE was looping on encountering an indefinitely repeated back reference to
5620 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5621 Perl does - treats the match as successful.
5622
5623 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5