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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1522 - (show annotations)
Sun Feb 8 17:02:05 2015 UTC (4 years, 8 months ago) by ph10
File size: 271319 byte(s)
Fix crash when mutual recursion such as (\2)(\1) is studied.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.37 xx-xxx-2015
5 ------------------------
6
7 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8 for those parentheses to be closed with whatever has been captured so far.
9 However, it was failing to mark any other groups between the hightest
10 capture so far and the currrent group as "unset". Thus, the ovector for
11 those groups contained whatever was previously there. An example is the
12 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
13
14 2. If an assertion condition was quantified with a minimum of zero (an odd
15 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
16
17 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
18 unrecognized modifier, a crash could occur.
19
20 4. An attempt to do global matching in pcretest with a zero-length ovector
21 caused a crash.
22
23 5. Fixed a memory leak during matching that could occur for a subpattern
24 subroutine call (recursive or otherwise) if the number of captured groups
25 that had to be saved was greater than ten.
26
27 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
28 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
29 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
30
31 7. A UTF pattern containing a "not" match of a non-ASCII character and a
32 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
33
34 8. When a pattern is compiled, it remembers the highest back reference so that
35 when matching, if the ovector is too small, extra memory can be obtained to
36 use instead. A conditional subpattern whose condition is a check on a
37 capture having happened, such as, for example in the pattern
38 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
39 setting the highest backreference number. This mattered only if pcre_exec()
40 was called with an ovector that was too small to hold the capture, and there
41 was no other kind of back reference (a situation which is probably quite
42 rare). The effect of the bug was that the condition was always treated as
43 FALSE when the capture could not be consulted, leading to a incorrect
44 behaviour by pcre2_match(). This bug has been fixed.
45
46 9. A reference to a duplicated named group (either a back reference or a test
47 for being set in a conditional) that occurred in a part of the pattern where
48 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
49 to be incorrectly calculated, leading to overwriting.
50
51 10. A mutually recursive set of back references such as (\2)(\1) caused a
52 segfault at study time (while trying to find the minimum matching length).
53 The infinite loop is now broken (with the minimum length unset, that is,
54 zero).
55
56
57 Version 8.36 26-September-2014
58 ------------------------------
59
60 1. Got rid of some compiler warnings in the C++ modules that were shown up by
61 -Wmissing-field-initializers and -Wunused-parameter.
62
63 2. The tests for quantifiers being too big (greater than 65535) were being
64 applied after reading the number, and stupidly assuming that integer
65 overflow would give a negative number. The tests are now applied as the
66 numbers are read.
67
68 3. Tidy code in pcre_exec.c where two branches that used to be different are
69 now the same.
70
71 4. The JIT compiler did not generate match limit checks for certain
72 bracketed expressions with quantifiers. This may lead to exponential
73 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
74 issue should be resolved now.
75
76 5. Fixed an issue, which occures when nested alternatives are optimized
77 with table jumps.
78
79 6. Inserted two casts and changed some ints to size_t in the light of some
80 reported 64-bit compiler warnings (Bugzilla 1477).
81
82 7. Fixed a bug concerned with zero-minimum possessive groups that could match
83 an empty string, which sometimes were behaving incorrectly in the
84 interpreter (though correctly in the JIT matcher). This pcretest input is
85 an example:
86
87 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
88 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
89
90 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
91 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
92 for an empty string was breaking the inner loop and carrying on at a lower
93 level, when possessive repeated groups should always return to a higher
94 level as they have no backtrack points in them. The empty string test now
95 occurs at the outer level.
96
97 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
98 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
99
100 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
101 doesn't).
102
103 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
104 to hold a pattern to be incorrectly computed (too small) when there were
105 named back references to duplicated names. This could cause "internal
106 error: code overflow" or "double free or corruption" or other memory
107 handling errors.
108
109 11. When named subpatterns had the same prefixes, back references could be
110 confused. For example, in this pattern:
111
112 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
113
114 the reference to 'Name' was incorrectly treated as a reference to a
115 duplicate name.
116
117 12. A pattern such as /^s?c/mi8 where the optional character has more than
118 one "other case" was incorrectly compiled such that it would only try to
119 match starting at "c".
120
121 13. When a pattern starting with \s was studied, VT was not included in the
122 list of possible starting characters; this should have been part of the
123 8.34/18 patch.
124
125 14. If a character class started [\Qx]... where x is any character, the class
126 was incorrectly terminated at the ].
127
128 15. If a pattern that started with a caseless match for a character with more
129 than one "other case" was studied, PCRE did not set up the starting code
130 unit bit map for the list of possible characters. Now it does. This is an
131 optimization improvement, not a bug fix.
132
133 16. The Unicode data tables have been updated to Unicode 7.0.0.
134
135 17. Fixed a number of memory leaks in pcregrep.
136
137 18. Avoid a compiler warning (from some compilers) for a function call with
138 a cast that removes "const" from an lvalue by using an intermediate
139 variable (to which the compiler does not object).
140
141 19. Incorrect code was compiled if a group that contained an internal recursive
142 back reference was optional (had quantifier with a minimum of zero). This
143 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
144 caused segmentation faults because of stack overflows at compile time.
145
146 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
147 group that is quantified with an indefinite repeat, caused a compile-time
148 loop which used up all the system stack and provoked a segmentation fault.
149 This was not the same bug as 19 above.
150
151 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
152 Patch by Mike Frysinger.
153
154
155 Version 8.35 04-April-2014
156 --------------------------
157
158 1. A new flag is set, when property checks are present in an XCLASS.
159 When this flag is not set, PCRE can perform certain optimizations
160 such as studying these XCLASS-es.
161
162 2. The auto-possessification of character sets were improved: a normal
163 and an extended character set can be compared now. Furthermore
164 the JIT compiler optimizes more character set checks.
165
166 3. Got rid of some compiler warnings for potentially uninitialized variables
167 that show up only when compiled with -O2.
168
169 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
170 of a match later then the end of the match. The pcretest program was not
171 handling the case sensibly - it was outputting from the start to the next
172 binary zero. It now reports this situation in a message, and outputs the
173 text from the end to the start.
174
175 5. Fast forward search is improved in JIT. Instead of the first three
176 characters, any three characters with fixed position can be searched.
177 Search order: first, last, middle.
178
179 6. Improve character range checks in JIT. Characters are read by an inprecise
180 function now, which returns with an unknown value if the character code is
181 above a certain threshold (e.g: 256). The only limitation is that the value
182 must be bigger than the threshold as well. This function is useful when
183 the characters above the threshold are handled in the same way.
184
185 7. The macros whose names start with RAWUCHAR are placeholders for a future
186 mode in which only the bottom 21 bits of 32-bit data items are used. To
187 make this more memorable for those maintaining the code, the names have
188 been changed to start with UCHAR21, and an extensive comment has been added
189 to their definition.
190
191 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
192 to the export list in Makefile.am (they were accidentally omitted from the
193 8.34 tarball).
194
195 9. The informational output from pcretest used the phrase "starting byte set"
196 which is inappropriate for the 16-bit and 32-bit libraries. As the output
197 for "first char" and "need char" really means "non-UTF-char", I've changed
198 "byte" to "char", and slightly reworded the output. The documentation about
199 these values has also been (I hope) clarified.
200
201 10. Another JIT related optimization: use table jumps for selecting the correct
202 backtracking path, when more than four alternatives are present inside a
203 bracket.
204
205 11. Empty match is not possible, when the minimum length is greater than zero,
206 and there is no \K in the pattern. JIT should avoid empty match checks in
207 such cases.
208
209 12. In a caseless character class with UCP support, when a character with more
210 than one alternative case was not the first character of a range, not all
211 the alternative cases were added to the class. For example, s and \x{17f}
212 are both alternative cases for S: the class [RST] was handled correctly,
213 but [R-T] was not.
214
215 13. The configure.ac file always checked for pthread support when JIT was
216 enabled. This is not used in Windows, so I have put this test inside a
217 check for the presence of windows.h (which was already tested for).
218
219 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
220 The algorithm provides a way to skip certain starting offsets, and usually
221 faster than linear prefix searches.
222
223 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
224 as for 'fr_FR' and 'french'. For some reason, however, it then used the
225 Windows-specific input and output files, which have 'french' screwed in.
226 So this could never have worked. One of the problems with locales is that
227 they aren't always the same. I have now updated RunTest so that it checks
228 the output of the locale test (test 3) against three different output
229 files, and it allows the test to pass if any one of them matches. With luck
230 this should make the test pass on some versions of Solaris where it was
231 failing. Because of the uncertainty, the script did not used to stop if
232 test 3 failed; it now does. If further versions of a French locale ever
233 come to light, they can now easily be added.
234
235 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
236 there was a message during ./configure, but it did not stop. This now
237 provokes an error. The invalid example in README has been corrected.
238 If a value less than the minimum is given, the minimum value has always
239 been used, but now a warning is given.
240
241 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
242 was a bug in the test system, which is now fixed. Also, the list of various
243 configurations that are tested for each release did not have one with both
244 16/32 bits and --enable-bar-anycrlf. It now does.
245
246 18. pcretest was missing "-C bsr" for displaying the \R default setting.
247
248 19. Little endian PowerPC systems are supported now by the JIT compiler.
249
250 20. The fast forward newline mechanism could enter to an infinite loop on
251 certain invalid UTF-8 input. Although we don't support these cases
252 this issue can be fixed by a performance optimization.
253
254 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
255 not take account if existing stack usage. There is now a new global
256 variable called pcre_stack_guard that can be set to point to an external
257 function to check stack availability. It is called at the start of
258 processing every parenthesized group.
259
260 22. A typo in the code meant that in ungreedy mode the max/min qualifier
261 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
262 not match "ab".
263
264 23. When UTF was disabled, the JIT program reported some incorrect compile
265 errors. These messages are silenced now.
266
267 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
268 compiler.
269
270 25. Change all the temporary files used in RunGrepTest to be different to those
271 used by RunTest so that the tests can be run simultaneously, for example by
272 "make -j check".
273
274
275 Version 8.34 15-December-2013
276 -----------------------------
277
278 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
279 executable memory. Patch inspired by Carsten Klein.
280
281 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
282 this macro is never tested and has no effect, because the work to support
283 coverage involves only compiling and linking options and special targets in
284 the Makefile. The comment in config.h implied that defining the macro would
285 enable coverage support, which is totally false. There was also support for
286 setting this macro in the CMake files (my fault, I just copied it from
287 configure). SUPPORT_GCOV has now been removed.
288
289 3. Make a small performance improvement in strlen16() and strlen32() in
290 pcretest.
291
292 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
293 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
294 these cases). There was also one in pcretest.c.
295
296 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
297
298 6. In UTF mode, the code for checking whether a group could match an empty
299 string (which is used for indefinitely repeated groups to allow for
300 breaking an infinite loop) was broken when the group contained a repeated
301 negated single-character class with a character that occupied more than one
302 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
303 UTF-8 mode). The effect was undefined: the group might or might not be
304 deemed as matching an empty string, or the program might have crashed.
305
306 7. The code for checking whether a group could match an empty string was not
307 recognizing that \h, \H, \v, \V, and \R must match a character.
308
309 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
310 an empty string. If it can, pcretest shows this in its information output.
311
312 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
313 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
314 matched by pcre_exec() without using JIT:
315
316 (a) If the rest of the pattern did not match after a maximal run of
317 grapheme clusters, the code for backing up to try with fewer of them
318 did not always back up over a full grapheme when characters that do not
319 have the modifier quality were involved, e.g. Hangul syllables.
320
321 (b) If the match point in a subject started with modifier character, and
322 there was no match, the code could incorrectly back up beyond the match
323 point, and potentially beyond the first character in the subject,
324 leading to a segfault or an incorrect match result.
325
326 10. A conditional group with an assertion condition could lead to PCRE
327 recording an incorrect first data item for a match if no other first data
328 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
329 first data item, and therefore matched "ca" after "c" instead of at the
330 start.
331
332 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
333 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
334
335 12. The source of pcregrep now includes z/OS-specific code so that it can be
336 compiled for z/OS as part of the special z/OS distribution.
337
338 13. Added the -T and -TM options to pcretest.
339
340 14. The code in pcre_compile.c for creating the table of named capturing groups
341 has been refactored. Instead of creating the table dynamically during the
342 actual compiling pass, the information is remembered during the pre-compile
343 pass (on the stack unless there are more than 20 named groups, in which
344 case malloc() is used) and the whole table is created before the actual
345 compile happens. This has simplified the code (it is now nearly 150 lines
346 shorter) and prepared the way for better handling of references to groups
347 with duplicate names.
348
349 15. A back reference to a named subpattern when there is more than one of the
350 same name now checks them in the order in which they appear in the pattern.
351 The first one that is set is used for the reference. Previously only the
352 first one was inspected. This change makes PCRE more compatible with Perl.
353
354 16. Unicode character properties were updated from Unicode 6.3.0.
355
356 17. The compile-time code for auto-possessification has been refactored, based
357 on a patch by Zoltan Herczeg. It now happens after instead of during
358 compilation. The code is cleaner, and more cases are handled. The option
359 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
360 options in pcretest are provided to set it. It can also be set by
361 (*NO_AUTO_POSSESS) at the start of a pattern.
362
363 18. The character VT has been added to the default ("C" locale) set of
364 characters that match \s and are generally treated as white space,
365 following this same change in Perl 5.18. There is now no difference between
366 "Perl space" and "POSIX space". Whether VT is treated as white space in
367 other locales depends on the locale.
368
369 19. The code for checking named groups as conditions, either for being set or
370 for being recursed, has been refactored (this is related to 14 and 15
371 above). Processing unduplicated named groups should now be as fast at
372 numerical groups, and processing duplicated groups should be faster than
373 before.
374
375 20. Two patches to the CMake build system, by Alexander Barkov:
376
377 (1) Replace the "source" command by "." in CMakeLists.txt because
378 "source" is a bash-ism.
379
380 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
381 without these the CMake build does not work on Solaris.
382
383 21. Perl has changed its handling of \8 and \9. If there is no previously
384 encountered capturing group of those numbers, they are treated as the
385 literal characters 8 and 9 instead of a binary zero followed by the
386 literals. PCRE now does the same.
387
388 22. Following Perl, added \o{} to specify codepoints in octal, making it
389 possible to specify values greater than 0777 and also making them
390 unambiguous.
391
392 23. Perl now gives an error for missing closing braces after \x{... instead of
393 treating the string as literal. PCRE now does the same.
394
395 24. RunTest used to grumble if an inappropriate test was selected explicitly,
396 but just skip it when running all tests. This make it awkward to run ranges
397 of tests when one of them was inappropriate. Now it just skips any
398 inappropriate tests, as it always did when running all tests.
399
400 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
401 character types such as \d or \w, too many callouts were inserted, and the
402 data that they returned was rubbish.
403
404 26. In UCP mode, \s was not matching two of the characters that Perl matches,
405 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
406 were matched by \h. The code has now been refactored so that the lists of
407 the horizontal and vertical whitespace characters used for \h and \v (which
408 are defined only in one place) are now also used for \s.
409
410 27. Add JIT support for the 64 bit TileGX architecture.
411 Patch by Jiong Wang (Tilera Corporation).
412
413 28. Possessive quantifiers for classes (both explicit and automatically
414 generated) now use special opcodes instead of wrapping in ONCE brackets.
415
416 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
417 (because it's meaningless), this was not happening when PCRE_CASELESS was
418 set. Not wrong, but inefficient.
419
420 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
421 properties for \w, \d, etc) is present in a test regex. Otherwise if the
422 test contains no characters greater than 255, Perl doesn't realise it
423 should be using Unicode semantics.
424
425 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
426 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
427 does in Unicode mode.
428
429 32. Added the "forbid" facility to pcretest so that putting tests into the
430 wrong test files can sometimes be quickly detected.
431
432 33. There is now a limit (default 250) on the depth of nesting of parentheses.
433 This limit is imposed to control the amount of system stack used at compile
434 time. It can be changed at build time by --with-parens-nest-limit=xxx or
435 the equivalent in CMake.
436
437 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
438 errors. Perl warns for these when in warning mode, but PCRE has no facility
439 for giving warnings.
440
441 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
442 However, this was not working for (?!) because it is optimized to (*FAIL),
443 for which PCRE does not allow quantifiers. The optimization is now disabled
444 when a quantifier follows (?!). I can't see any use for this, but it makes
445 things uniform.
446
447 36. Perl no longer allows group names to start with digits, so I have made this
448 change also in PCRE. It simplifies the code a bit.
449
450 37. In extended mode, Perl ignores spaces before a + that indicates a
451 possessive quantifier. PCRE allowed a space before the quantifier, but not
452 before the possessive +. It now does.
453
454 38. The use of \K (reset reported match start) within a repeated possessive
455 group such as (a\Kb)*+ was not working.
456
457 40. Document that the same character tables must be used at compile time and
458 run time, and that the facility to pass tables to pcre_exec() and
459 pcre_dfa_exec() is for use only with saved/restored patterns.
460
461 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
462 features for Builds with MSVC:
463
464 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
465 builds). The .rc files can be used to set FileDescription and many other
466 attributes.
467
468 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
469 This allows higher-level build scripts which want .pdb files to avoid
470 hard-coding the exact files needed."
471
472 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
473 mean "start of word" and "end of word", respectively, as a transition aid.
474
475 43. A minimizing repeat of a class containing codepoints greater than 255 in
476 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
477 compiled to use the heap for recursion.
478
479 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
480 is configured.
481
482
483 Version 8.33 28-May-2013
484 ------------------------
485
486 1. Added 'U' to some constants that are compared to unsigned integers, to
487 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
488 variables that are added to signed variables, to ensure the result is
489 signed and can be negated.
490
491 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
492 CMake config files.
493
494 3. Revise the creation of config.h.generic so that all boolean macros are
495 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
496 overriding via -D on the command line possible.
497
498 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
499 to unsigned int is reported to make a quite noticeable speed difference in
500 a specific Windows environment. Testing on Linux did also appear to show
501 some benefit (and it is clearly not harmful). Also fixed the definition of
502 Xop which should be unsigned.
503
504 5. Related to (4), changing the definition of the intermediate variable cc
505 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
506 improvements.
507
508 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
509 unnecessary spaces.
510
511 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
512 and later.
513
514 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
515
516 9. Optimizing fast_forward_start_bits in JIT.
517
518 10. Adding support for callouts in JIT, and fixing some issues revealed
519 during this work. Namely:
520
521 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
522
523 (b) Minimum length was not checked before the matching is started.
524
525 11. The value of capture_last that is passed to callouts was incorrect in some
526 cases when there was a capture on one path that was subsequently abandoned
527 after a backtrack. Also, the capture_last value is now reset after a
528 recursion, since all captures are also reset in this case.
529
530 12. The interpreter no longer returns the "too many substrings" error in the
531 case when an overflowing capture is in a branch that is subsequently
532 abandoned after a backtrack.
533
534 13. In the pathological case when an offset vector of size 2 is used, pcretest
535 now prints out the matched string after a yield of 0 or 1.
536
537 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
538 Only supported by the JIT compiler at the moment.
539
540 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
541
542 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
543 offset of the starting point of the matching process, provided the offsets
544 vector is large enough.
545
546 17. The \A escape now records a lookbehind value of 1, though its execution
547 does not actually inspect the previous character. This is to ensure that,
548 in partial multi-segment matching, at least one character from the old
549 segment is retained when a new segment is processed. Otherwise, if there
550 are no lookbehinds in the pattern, \A might match incorrectly at the start
551 of a new segment.
552
553 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
554
555 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
556 gives some modest performance improvement in 8-bit mode.
557
558 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
559 be expressed in certain programming languages using Universal Character
560 Names.
561
562 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
563 which points out that "non characters" are not "characters that may not
564 appear in Unicode strings" but rather "characters that are reserved for
565 internal use and have only local meaning".
566
567 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
568 there was a conditional group that depended on an assertion, if the
569 assertion was false, the callout that immediately followed the alternation
570 in the condition was skipped when pcre_exec() was used for matching.
571
572 23. Allow an explicit callout to be inserted before an assertion that is the
573 condition for a conditional group, for compatibility with automatic
574 callouts, which always insert a callout at this point.
575
576 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
577 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
578
579 25. (*PRUNE) is now supported by the JIT compiler.
580
581 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
582
583 27. Fix the case where there are two or more SKIPs with arguments that may be
584 ignored.
585
586 28. (*SKIP) is now supported by the JIT compiler.
587
588 29. (*THEN) is now supported by the JIT compiler.
589
590 30. Update RunTest with additional test selector options.
591
592 31. The way PCRE handles backtracking verbs has been changed in two ways.
593
594 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
595 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
596 backtracking. In some cases this makes it more Perl-compatible, but Perl's
597 rather obscure rules do not always do the same thing.
598
599 (2) Previously, backtracking verbs were confined within assertions. This is
600 no longer the case for positive assertions, except for (*ACCEPT). Again,
601 this sometimes improves Perl compatibility, and sometimes does not.
602
603 32. A number of tests that were in test 2 because Perl did things differently
604 have been moved to test 1, because either Perl or PCRE has changed, and
605 these tests are now compatible.
606
607 32. Backtracking control verbs are now handled in the same way in JIT and
608 interpreter.
609
610 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
611 contained a forward subroutine reference caused a compile error.
612
613 34. Auto-detect and optimize limited repetitions in JIT.
614
615 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
616 blocking (*UTF) etc.
617
618 36. In the interpreter, maximizing pattern repetitions for characters and
619 character types now use tail recursion, which reduces stack usage.
620
621 37. The value of the max lookbehind was not correctly preserved if a compiled
622 and saved regex was reloaded on a host of different endianness.
623
624 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
625 of the compiled pattern block, expand the flags field from 16 to 32 bits
626 because it was almost full.
627
628 39. Try madvise first before posix_madvise.
629
630 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
631 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
632 original change locked out all matches of empty strings. This has been
633 changed so that one match of an empty string per line is recognized.
634 Subsequent searches on the same line (for colouring or for --only-matching,
635 for example) do not recognize empty strings.
636
637 41. Applied a user patch to fix a number of spelling mistakes in comments.
638
639 42. Data lines longer than 65536 caused pcretest to crash.
640
641 43. Clarified the data type for length and startoffset arguments for pcre_exec
642 and pcre_dfa_exec in the function-specific man pages, where they were
643 explicitly stated to be in bytes, never having been updated. I also added
644 some clarification to the pcreapi man page.
645
646 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
647 a segmentation fault.
648
649
650 Version 8.32 30-November-2012
651 -----------------------------
652
653 1. Improved JIT compiler optimizations for first character search and single
654 character iterators.
655
656 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
657 Patch by Daniel Richard G.
658
659 3. Single character iterator optimizations in the JIT compiler.
660
661 4. Improved JIT compiler optimizations for character ranges.
662
663 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
664 Reported by Giuseppe D'Angelo.
665
666 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
667 of a line, was being set incorrectly in cases where .* appeared inside
668 atomic brackets at the start of a pattern, or where there was a subsequent
669 *PRUNE or *SKIP.
670
671 7. Improved instruction cache flush for POWER/PowerPC.
672 Patch by Daniel Richard G.
673
674 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
675 grep:
676
677 (a) There is now no limit to the number of patterns to be matched.
678
679 (b) An error is given if a pattern is too long.
680
681 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
682 are now supported.
683
684 (d) --exclude-from and --include-from (multiple use) have been added.
685
686 (e) Exclusions and inclusions now apply to all files and directories, not
687 just to those obtained from scanning a directory recursively.
688
689 (f) Multiple uses of -f and --file-list are now supported.
690
691 (g) In a Windows environment, the default for -d has been changed from
692 "read" (the GNU grep default) to "skip", because otherwise the presence
693 of a directory in the file list provokes an error.
694
695 (h) The documentation has been revised and clarified in places.
696
697 9. Improve the matching speed of capturing brackets.
698
699 10. Changed the meaning of \X so that it now matches a Unicode extended
700 grapheme cluster.
701
702 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
703 out POSIX threads when JIT support is configured.
704
705 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
706
707 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
708 structure could go wrong in environments where size_t is not the same size
709 as int.
710
711 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
712 set.
713
714 15. The EBCDIC support had decayed; later updates to the code had included
715 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
716 general tidy up of EBCDIC-related issues, and the documentation was also
717 not quite right. There is now a test that can be run on ASCII systems to
718 check some of the EBCDIC-related things (but is it not a full test).
719
720 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
721 in a small tidy to the code.
722
723 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
724
725 18. If the --only-matching (-o) option in pcregrep is specified multiple
726 times, each one causes appropriate output. For example, -o1 -o2 outputs the
727 substrings matched by the 1st and 2nd capturing parentheses. A separating
728 string can be specified by --om-separator (default empty).
729
730 19. Improving the first n character searches.
731
732 20. Turn case lists for horizontal and vertical white space into macros so that
733 they are defined only once.
734
735 21. This set of changes together give more compatible Unicode case-folding
736 behaviour for characters that have more than one other case when UCP
737 support is available.
738
739 (a) The Unicode property table now has offsets into a new table of sets of
740 three or more characters that are case-equivalent. The MultiStage2.py
741 script that generates these tables (the pcre_ucd.c file) now scans
742 CaseFolding.txt instead of UnicodeData.txt for character case
743 information.
744
745 (b) The code for adding characters or ranges of characters to a character
746 class has been abstracted into a generalized function that also handles
747 case-independence. In UTF-mode with UCP support, this uses the new data
748 to handle characters with more than one other case.
749
750 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
751 whose other case is greater than 256 are now correctly matched
752 caselessly. Previously, the high codepoint matched the low one, but not
753 vice versa.
754
755 (d) The processing of \h, \H, \v, and \ in character classes now makes use
756 of the new class addition function, using character lists defined as
757 macros alongside the case definitions of 20 above.
758
759 (e) Caseless back references now work with characters that have more than
760 one other case.
761
762 (f) General caseless matching of characters with more than one other case
763 is supported.
764
765 22. Unicode character properties were updated from Unicode 6.2.0
766
767 23. Improved CMake support under Windows. Patch by Daniel Richard G.
768
769 24. Add support for 32-bit character strings, and UTF-32
770
771 25. Major JIT compiler update (code refactoring and bugfixing).
772 Experimental Sparc 32 support is added.
773
774 26. Applied a modified version of Daniel Richard G's patch to create
775 pcre.h.generic and config.h.generic by "make" instead of in the
776 PrepareRelease script.
777
778 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
779 pcre_compile.c when checking for a zero character.
780
781 28. Introducing a native interface for JIT. Through this interface, the compiled
782 machine code can be directly executed. The purpose of this interface is to
783 provide fast pattern matching, so several sanity checks are not performed.
784 However, feature tests are still performed. The new interface provides
785 1.4x speedup compared to the old one.
786
787 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
788 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
789 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
790
791 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
792 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
793 the "old" RFC 2279). Instead, it ended up passing a negative length to
794 pcre_exec().
795
796 31. Add support for GCC's visibility feature to hide internal functions.
797
798 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
799 "unknown -C option" after outputting 0 or 1.
800
801 33. There is now support for generating a code coverage report for the test
802 suite in environments where gcc is the compiler and lcov is installed. This
803 is mainly for the benefit of the developers.
804
805 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
806 unaddressable using valgrind annotations, allowing valgrind to detect
807 invalid memory accesses. This is mainly for the benefit of the developers.
808
809 25. (*UTF) can now be used to start a pattern in any of the three libraries.
810
811 26. Give configure error if --enable-cpp but no C++ compiler found.
812
813
814 Version 8.31 06-July-2012
815 -------------------------
816
817 1. Fixing a wrong JIT test case and some compiler warnings.
818
819 2. Removed a bashism from the RunTest script.
820
821 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
822 to unsigned type, result still unsigned" that was given by an MS compiler
823 on encountering the code "-sizeof(xxx)".
824
825 4. Partial matching support is added to the JIT compiler.
826
827 5. Fixed several bugs concerned with partial matching of items that consist
828 of more than one character:
829
830 (a) /^(..)\1/ did not partially match "aba" because checking references was
831 done on an "all or nothing" basis. This also applied to repeated
832 references.
833
834 (b) \R did not give a hard partial match if \r was found at the end of the
835 subject.
836
837 (c) \X did not give a hard partial match after matching one or more
838 characters at the end of the subject.
839
840 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
841 a partial match for the string "\r".
842
843 (e) When newline was set to CRLF, the metacharacter "." did not recognize
844 a partial match for a CR character at the end of the subject string.
845
846 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
847 running pcretest, the text "(JIT)" added to the output whenever JIT is
848 actually used to run the match.
849
850 7. Individual JIT compile options can be set in pcretest by following -s+[+]
851 or /S+[+] with a digit between 1 and 7.
852
853 8. OP_NOT now supports any UTF character not just single-byte ones.
854
855 9. (*MARK) control verb is now supported by the JIT compiler.
856
857 10. The command "./RunTest list" lists the available tests without actually
858 running any of them. (Because I keep forgetting what they all are.)
859
860 11. Add PCRE_INFO_MAXLOOKBEHIND.
861
862 12. Applied a (slightly modified) user-supplied patch that improves performance
863 when the heap is used for recursion (compiled with --disable-stack-for-
864 recursion). Instead of malloc and free for each heap frame each time a
865 logical recursion happens, frames are retained on a chain and re-used where
866 possible. This sometimes gives as much as 30% improvement.
867
868 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
869 call.
870
871 14. As documented, (*COMMIT) is now confined to within a positive assertion.
872
873 15. It is now possible to link pcretest with libedit as an alternative to
874 libreadline.
875
876 16. (*COMMIT) control verb is now supported by the JIT compiler.
877
878 17. The Unicode data tables have been updated to Unicode 6.1.0.
879
880 18. Added --file-list option to pcregrep.
881
882 19. Added binary file support to pcregrep, including the -a, --binary-files,
883 -I, and --text options.
884
885 20. The madvise function is renamed for posix_madvise for QNX compatibility
886 reasons. Fixed by Giuseppe D'Angelo.
887
888 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
889
890 22. Changed the option for creating symbolic links for 16-bit man pages from
891 -s to -sf so that re-installing does not cause issues.
892
893 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
894
895 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
896 matches in certain environments (the workspace was not being correctly
897 retained). Also added to pcre_dfa_exec() a simple plausibility check on
898 some of the workspace data at the beginning of a restart.
899
900 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
901 was not doing so when it should - probably a typo introduced by SVN 528
902 (change 8.10/14).
903
904 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
905 \w+ when the character tables indicated that \x{c4} was a word character.
906 There were several related cases, all because the tests for doing a table
907 lookup were testing for characters less than 127 instead of 255.
908
909 27. If a pattern contains capturing parentheses that are not used in a match,
910 their slots in the ovector are set to -1. For those that are higher than
911 any matched groups, this happens at the end of processing. In the case when
912 there were back references that the ovector was too small to contain
913 (causing temporary malloc'd memory to be used during matching), and the
914 highest capturing number was not used, memory off the end of the ovector
915 was incorrectly being set to -1. (It was using the size of the temporary
916 memory instead of the true size.)
917
918 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
919 ovector size, it uses memory at the end of the block that it has got.
920
921 29. Check for an overlong MARK name and give an error at compile time. The
922 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
923
924 30. JIT compiler update.
925
926 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
927 Rill for the patch.
928
929 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
930
931 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
932
933 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
934 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
935 was enabled.
936
937 35. Improve JIT code generation for greedy plus quantifier.
938
939 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
940 1 to "aa" instead of to an empty string. The bug affected repeated groups
941 that could potentially match an empty string.
942
943 37. Optimizing single character iterators in JIT.
944
945 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
946 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
947 codepoints that are too big for the mode are faulted, and in a UTF mode,
948 disallowed codepoints are also faulted.
949
950 39. If PCRE was compiled with UTF support, in three places in the DFA
951 matcher there was code that should only have been obeyed in UTF mode, but
952 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
953 processing when bytes with values greater than 127 were present. In 16-bit
954 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
955 both cases the values are those that cannot be the first data item in a UTF
956 character. The three items that might have provoked this were recursions,
957 possessively repeated groups, and atomic groups.
958
959 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
960 and pcregrep, because some OS require shared objects to be explicitly
961 passed to ld, causing the link step to fail if they are not.
962
963 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
964 mode, patterns that started with \h* or \R* might be incorrectly matched.
965
966
967 Version 8.30 04-February-2012
968 -----------------------------
969
970 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
971 name is defined in ctype.h.
972
973 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
974 only in quite long subpatterns.
975
976 3. Removed the function pcre_info(), which has been obsolete and deprecated
977 since it was replaced by pcre_fullinfo() in February 2000.
978
979 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
980 match a (*MARK), and the match failed at the start of the subject, a
981 reference to memory before the start of the subject could occur. This bug
982 was introduced by fix 17 of release 8.21.
983
984 5. A reference to an unset group with zero minimum repetition was giving
985 totally wrong answers (in non-JavaScript-compatibility mode). For example,
986 /(another)?(\1?)test/ matched against "hello world test". This bug was
987 introduced in release 8.13.
988
989 6. Add support for 16-bit character strings (a large amount of work involving
990 many changes and refactorings).
991
992 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
993 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
994 from a file.
995
996 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
997 rounding is not applied in this particular case).
998
999 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1000 if they appear, or are escaped, in patterns.
1001
1002 10. Get rid of a number of -Wunused-but-set-variable warnings.
1003
1004 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1005 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1006 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1007 also returns the mark "x". This bug applied to capturing parentheses,
1008 non-capturing parentheses, and atomic parentheses. It also applied to some
1009 assertions.
1010
1011 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1012 information out of configure.ac instead of relying on pcre.h.generic, which
1013 is not stored in the repository.
1014
1015 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1016 -lreadline.
1017
1018 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1019
1020 15. Applied Graycode's patch to put the top-level frame on the stack rather
1021 than the heap when not using the stack for recursion. This gives a
1022 performance improvement in many cases when recursion is not deep.
1023
1024 16. Experimental code added to "pcretest -C" to output the stack frame size.
1025
1026
1027 Version 8.21 12-Dec-2011
1028 ------------------------
1029
1030 1. Updating the JIT compiler.
1031
1032 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1033 are added as well.
1034
1035 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1036 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1037 calling _pcre_jit_exec. Some extra comments are added.
1038
1039 4. (*MARK) settings inside atomic groups that do not contain any capturing
1040 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1041 was introduced by change 18 for 8.20.
1042
1043 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1044 ECMA-262 standard.
1045
1046 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1047 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1048 This bug was probably introduced by change 9 of 8.13.
1049
1050 7. While fixing 6 above, I noticed that a number of other items were being
1051 incorrectly rejected as "not fixed length". This arose partly because newer
1052 opcodes had not been added to the fixed-length checking code. I have (a)
1053 corrected the bug and added tests for these items, and (b) arranged for an
1054 error to occur if an unknown opcode is encountered while checking for fixed
1055 length instead of just assuming "not fixed length". The items that were
1056 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1057 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1058 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1059
1060 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1061 being incorrectly compiled and would have given unpredicatble results.
1062
1063 9. A possessively repeated subpattern with minimum repeat count greater than
1064 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1065 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1066 the first (A) could occur when it should not.
1067
1068 10. Add a cast and remove a redundant test from the code.
1069
1070 11. JIT should use pcre_malloc/pcre_free for allocation.
1071
1072 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1073 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1074 is anything other than /usr, -L is still shown).
1075
1076 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1077
1078 14. Perl does not support \N without a following name in a [] class; PCRE now
1079 also gives an error.
1080
1081 15. If a forward reference was repeated with an upper limit of around 2000,
1082 it caused the error "internal error: overran compiling workspace". The
1083 maximum number of forward references (including repeats) was limited by the
1084 internal workspace, and dependent on the LINK_SIZE. The code has been
1085 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1086 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1087 of around 200,000 forward references. While doing this, I also speeded up
1088 the filling in of repeated forward references.
1089
1090 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1091 incorrectly expecting the subject to contain another "a" after the start.
1092
1093 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1094 in the match, the SKIP should be ignored. This was not happening; instead
1095 the SKIP was being treated as NOMATCH. For patterns such as
1096 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1097 tested.
1098
1099 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1100 now much more compatible with Perl, in particular in cases where the result
1101 is a non-match for a non-anchored pattern. For example, if
1102 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1103 "m", where previously it did not return a name. A side effect of this
1104 change is that for partial matches, the last encountered mark name is
1105 returned, as for non matches. A number of tests that were previously not
1106 Perl-compatible have been moved into the Perl-compatible test files. The
1107 refactoring has had the pleasing side effect of removing one argument from
1108 the match() function, thus reducing its stack requirements.
1109
1110 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1111 subsequent uses of /S (without +) incorrectly behaved like /S+.
1112
1113 21. Retrieve executable code size support for the JIT compiler and fixing
1114 some warnings.
1115
1116 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1117 not work when the shorter character appeared right at the end of the
1118 subject string.
1119
1120 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1121 systems.
1122
1123 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1124 output it when the /M option is used in pcretest.
1125
1126 25. The CheckMan script was not being included in the distribution. Also, added
1127 an explicit "perl" to run Perl scripts from the PrepareRelease script
1128 because this is reportedly needed in Windows.
1129
1130 26. If study data was being save in a file and studying had not found a set of
1131 "starts with" bytes for the pattern, the data written to the file (though
1132 never used) was taken from uninitialized memory and so caused valgrind to
1133 complain.
1134
1135 27. Updated RunTest.bat as provided by Sheri Pierce.
1136
1137 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1138
1139 29. Computation of memory usage for the table of capturing group names was
1140 giving an unnecessarily large value.
1141
1142
1143 Version 8.20 21-Oct-2011
1144 ------------------------
1145
1146 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1147 a POSIX class. After further experiments with Perl, which convinced me that
1148 Perl has bugs and confusions, a closing square bracket is no longer allowed
1149 in a POSIX name. This bug also affected patterns with classes that started
1150 with full stops.
1151
1152 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1153 captured substring, but while checking the failing first alternative,
1154 substring 1 is temporarily captured. If the output vector supplied to
1155 pcre_exec() was not big enough for this capture, the yield of the function
1156 was still zero ("insufficient space for captured substrings"). This cannot
1157 be totally fixed without adding another stack variable, which seems a lot
1158 of expense for a edge case. However, I have improved the situation in cases
1159 such as /(a)(b)x|abc/ matched against "abc", where the return code
1160 indicates that fewer than the maximum number of slots in the ovector have
1161 been set.
1162
1163 3. Related to (2) above: when there are more back references in a pattern than
1164 slots in the output vector, pcre_exec() uses temporary memory during
1165 matching, and copies in the captures as far as possible afterwards. It was
1166 using the entire output vector, but this conflicts with the specification
1167 that only 2/3 is used for passing back captured substrings. Now it uses
1168 only the first 2/3, for compatibility. This is, of course, another edge
1169 case.
1170
1171 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1172 main code base, and can be used by building with --enable-jit. When this is
1173 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1174 runtime --no-jit option is given.
1175
1176 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1177 ovector, the return from the function was zero, implying that there were
1178 other matches that did not fit. The correct "exactly full" value is now
1179 returned.
1180
1181 6. If a subpattern that was called recursively or as a subroutine contained
1182 (*PRUNE) or any other control that caused it to give a non-standard return,
1183 invalid errors such as "Error -26 (nested recursion at the same subject
1184 position)" or even infinite loops could occur.
1185
1186 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1187 computing the minimum length on reaching *ACCEPT, and so ended up with the
1188 wrong value of 1 rather than 0. Further investigation indicates that
1189 computing a minimum subject length in the presence of *ACCEPT is difficult
1190 (think back references, subroutine calls), and so I have changed the code
1191 so that no minimum is registered for a pattern that contains *ACCEPT.
1192
1193 8. If (*THEN) was present in the first (true) branch of a conditional group,
1194 it was not handled as intended. [But see 16 below.]
1195
1196 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1197 Sheri Pierce.
1198
1199 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1200 the first byte in a match must be "a".
1201
1202 11. Change 17 for 8.13 increased the recursion depth for patterns like
1203 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1204 pattern contains any instances of (*THEN). If it does not, the old
1205 optimizations are restored. It would be nice to do this on a per-group
1206 basis, but at the moment that is not feasible.
1207
1208 12. In some environments, the output of pcretest -C is CRLF terminated. This
1209 broke RunTest's code that checks for the link size. A single white space
1210 character after the value is now allowed for.
1211
1212 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1213 For "fr", it uses the Windows-specific input and output files.
1214
1215 14. If (*THEN) appeared in a group that was called recursively or as a
1216 subroutine, it did not work as intended. [But see next item.]
1217
1218 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1219 pattern fragments (but not containing any | characters). If A and B are
1220 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1221 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1222 D. In other words, Perl considers parentheses that do not contain any |
1223 characters to be part of a surrounding alternative, whereas PCRE was
1224 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1225 differently. PCRE now behaves in the same way as Perl, except in the case
1226 of subroutine/recursion calls such as (?1) which have in any case always
1227 been different (but PCRE had them first :-).
1228
1229 16. Related to 15 above: Perl does not treat the | in a conditional group as
1230 creating alternatives. Such a group is treated in the same way as an
1231 ordinary group without any | characters when processing (*THEN). PCRE has
1232 been changed to match Perl's behaviour.
1233
1234 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1235 RunGrepTest script failed.
1236
1237 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1238 inevitable for groups that contain captures, but it can lead to a lot of
1239 stack use in large patterns. The old behaviour has been restored for atomic
1240 groups that do not contain any capturing parentheses.
1241
1242 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1243 suppress the check for a minimum subject length at run time. (If it was
1244 given to pcre_exec() or pcre_dfa_exec() it did work.)
1245
1246 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1247 fail to work when decoding hex characters in data strings in EBCDIC
1248 environments.
1249
1250 21. It appears that in at least one Mac OS environment, the isxdigit() function
1251 is implemented as a macro that evaluates to its argument more than once,
1252 contravening the C 90 Standard (I haven't checked a later standard). There
1253 was an instance in pcretest which caused it to go wrong when processing
1254 \x{...} escapes in subject strings. The has been rewritten to avoid using
1255 things like p++ in the argument of isxdigit().
1256
1257
1258 Version 8.13 16-Aug-2011
1259 ------------------------
1260
1261 1. The Unicode data tables have been updated to Unicode 6.0.0.
1262
1263 2. Two minor typos in pcre_internal.h have been fixed.
1264
1265 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1266 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1267 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1268
1269 4. There were a number of related bugs in the code for matching backrefences
1270 caselessly in UTF-8 mode when codes for the characters concerned were
1271 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1272 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1273 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1274 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1275 2-byte code at the end of the subject (it thought there wasn't enough data
1276 left).
1277
1278 5. Comprehensive information about what went wrong is now returned by
1279 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1280 as the output vector has at least 2 elements. The offset of the start of
1281 the failing character and a reason code are placed in the vector.
1282
1283 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1284 now returned is for the first byte of the failing character, instead of the
1285 last byte inspected. This is an incompatible change, but I hope it is small
1286 enough not to be a problem. It makes the returned offset consistent with
1287 pcre_exec() and pcre_dfa_exec().
1288
1289 7. pcretest now gives a text phrase as well as the error number when
1290 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1291 failure, the offset and reason code are output.
1292
1293 8. When \R was used with a maximizing quantifier it failed to skip backwards
1294 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1295 back over a single character (\n). This seems wrong (because it treated the
1296 two characters as a single entity when going forwards), conflicts with the
1297 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1298 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1299 has been changed.
1300
1301 9. Some internal refactoring has changed the processing so that the handling
1302 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1303 time (the PCRE_DOTALL option was changed this way some time ago: version
1304 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1305 which was always a bit of a fudge. It also means that there is one less
1306 argument for the match() function, which reduces its stack requirements
1307 slightly. This change also fixes an incompatibility with Perl: the pattern
1308 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1309
1310 10. More internal refactoring has drastically reduced the number of recursive
1311 calls to match() for possessively repeated groups such as (abc)++ when
1312 using pcre_exec().
1313
1314 11. While implementing 10, a number of bugs in the handling of groups were
1315 discovered and fixed:
1316
1317 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1318 (a|)*(?1) gave a compile-time internal error.
1319 ((a|)+)+ did not notice that the outer group could match an empty string.
1320 (^a|^)+ was not marked as anchored.
1321 (.*a|.*)+ was not marked as matching at start or after a newline.
1322
1323 12. Yet more internal refactoring has removed another argument from the match()
1324 function. Special calls to this function are now indicated by setting a
1325 value in a variable in the "match data" data block.
1326
1327 13. Be more explicit in pcre_study() instead of relying on "default" for
1328 opcodes that mean there is no starting character; this means that when new
1329 ones are added and accidentally left out of pcre_study(), testing should
1330 pick them up.
1331
1332 14. The -s option of pcretest has been documented for ages as being an old
1333 synonym of -m (show memory usage). I have changed it to mean "force study
1334 for every regex", that is, assume /S for every regex. This is similar to -i
1335 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1336 using it. It makes it easier to run collections of tests with and without
1337 study enabled, and thereby test pcre_study() more easily. All the standard
1338 tests are now run with and without -s (but some patterns can be marked as
1339 "never study" - see 20 below).
1340
1341 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1342 restoration of the capturing data to the outer values was not happening
1343 correctly.
1344
1345 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1346 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1347 pattern had matched an empty string, and so incorrectly returned a no
1348 match.
1349
1350 17. There was optimizing code for the last branch of non-capturing parentheses,
1351 and also for the obeyed branch of a conditional subexpression, which used
1352 tail recursion to cut down on stack usage. Unfortunately, now that there is
1353 the possibility of (*THEN) occurring in these branches, tail recursion is
1354 no longer possible because the return has to be checked for (*THEN). These
1355 two optimizations have therefore been removed. [But see 8.20/11 above.]
1356
1357 18. If a pattern containing \R was studied, it was assumed that \R always
1358 matched two bytes, thus causing the minimum subject length to be
1359 incorrectly computed because \R can also match just one byte.
1360
1361 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1362 was incorrectly computed.
1363
1364 20. If /S is present twice on a test pattern in pcretest input, it now
1365 *disables* studying, thereby overriding the use of -s on the command line
1366 (see 14 above). This is necessary for one or two tests to keep the output
1367 identical in both cases.
1368
1369 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1370 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1371
1372 22. When an atomic group that contained a capturing parenthesis was
1373 successfully matched, but the branch in which it appeared failed, the
1374 capturing was not being forgotten if a higher numbered group was later
1375 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1376 group 1 to "a", when in fact it should be unset. This applied to multi-
1377 branched capturing and non-capturing groups, repeated or not, and also to
1378 positive assertions (capturing in negative assertions does not happen
1379 in PCRE) and also to nested atomic groups.
1380
1381 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1382 subject after a captured substring, to make it easier to tell which of a
1383 number of identical substrings has been captured.
1384
1385 24. The way atomic groups are processed by pcre_exec() has been changed so that
1386 if they are repeated, backtracking one repetition now resets captured
1387 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1388 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1389 "aaa". Previously, it would have been "a". As part of this code
1390 refactoring, the way recursive calls are handled has also been changed.
1391
1392 25. If an assertion condition captured any substrings, they were not passed
1393 back unless some other capturing happened later. For example, if
1394 (?(?=(a))a) was matched against "a", no capturing was returned.
1395
1396 26. When studying a pattern that contained subroutine calls or assertions,
1397 the code for finding the minimum length of a possible match was handling
1398 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1399 group 1 called group 2 while simultaneously a separate group 2 called group
1400 1). A stack overflow occurred in this case. I have fixed this by limiting
1401 the recursion depth to 10.
1402
1403 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1404 Fortmann. This supports explicit test numbers on the command line, and has
1405 argument validation and error reporting.
1406
1407 28. An instance of \X with an unlimited repeat could fail if at any point the
1408 first character it looked at was a mark character.
1409
1410 29. Some minor code refactoring concerning Unicode properties and scripts
1411 should reduce the stack requirement of match() slightly.
1412
1413 30. Added the '=' option to pcretest to check the setting of unused capturing
1414 slots at the end of the pattern, which are documented as being -1, but are
1415 not included in the return count.
1416
1417 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1418 compiled something random. Now it gives a compile-time error (as does
1419 Perl).
1420
1421 32. A *MARK encountered during the processing of a positive assertion is now
1422 recorded and passed back (compatible with Perl).
1423
1424 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1425 had alternative anchored branches, the search for a second match in a line
1426 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1427 matched the line "0102" twice. The same bug affected patterns that started
1428 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1429 twice.
1430
1431 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1432 does, and because of capturing effects, quantifying parenthesized
1433 assertions may at times be useful. Quantifiers are now allowed for
1434 parenthesized assertions.
1435
1436 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1437
1438 36. \g was being checked for fancy things in a character class, when it should
1439 just be a literal "g".
1440
1441 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1442 appearance of a nested POSIX class supersedes an apparent external class.
1443 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1444 unescaped square brackets may also appear as part of class names. For
1445 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1446 more like Perl. (But see 8.20/1 above.)
1447
1448 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1449 was because it thought it was \N{name}, which is not supported).
1450
1451 39. Add minix to OS list not supporting the -S option in pcretest.
1452
1453 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1454 cannot analyze patterns in sufficient detail to catch mutual recursions
1455 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1456 subgroup is called recursively as a subpattern for a second time at the
1457 same position in the subject string. In previous releases this might have
1458 been caught by the recursion limit, or it might have run out of stack.
1459
1460 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1461 happen only once. PCRE was, however incorrectly giving a compile time error
1462 "recursive call could loop indefinitely" because it cannot analyze the
1463 pattern in sufficient detail. The compile time test no longer happens when
1464 PCRE is compiling a conditional subpattern, but actual runaway loops are
1465 now caught at runtime (see 40 above).
1466
1467 42. It seems that Perl allows any characters other than a closing parenthesis
1468 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1469 has been changed to be the same.
1470
1471 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1472 as not to get warnings when autogen.sh is called. Also changed
1473 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1474
1475 44. To help people who use pcregrep to scan files containing exceedingly long
1476 lines, the following changes have been made:
1477
1478 (a) The default value of the buffer size parameter has been increased from
1479 8K to 20K. (The actual buffer used is three times this size.)
1480
1481 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1482 PCRE is built.
1483
1484 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1485 to be set at run time.
1486
1487 (d) Numerical values in pcregrep options can be followed by K or M, for
1488 example --buffer-size=50K.
1489
1490 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1491 given and the return code is set to 2.
1492
1493 45. Add a pointer to the latest mark to the callout data block.
1494
1495 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1496 partial match of an empty string instead of no match. This was specific to
1497 the use of ".".
1498
1499 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1500 complete match instead of a partial match. This bug was dependent on both
1501 the PCRE_UTF8 and PCRE_DOTALL options being set.
1502
1503 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1504 starting byte set, because \b was not being ignored.
1505
1506
1507 Version 8.12 15-Jan-2011
1508 ------------------------
1509
1510 1. Fixed some typos in the markup of the man pages, and wrote a script that
1511 checks for such things as part of the documentation building process.
1512
1513 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1514 --match-limit and --recursion-limit options (added for 8.11). In
1515 particular, this made one of the standard tests fail. (The integer value
1516 went into the wrong half of a long int.)
1517
1518 3. If the --colour option was given to pcregrep with -v (invert match), it
1519 did strange things, either producing crazy output, or crashing. It should,
1520 of course, ignore a request for colour when reporting lines that do not
1521 match.
1522
1523 4. Another pcregrep bug caused similar problems if --colour was specified with
1524 -M (multiline) and the pattern match finished with a line ending.
1525
1526 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1527 matched in multiline mode, the following line was shown as part of the
1528 match. This seems wrong, so I have changed it.
1529
1530 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1531 the check for further matches in the same line (so they could be coloured)
1532 to overrun the end of the current line. If another match was found, it was
1533 incorrectly shown (and then shown again when found in the next line).
1534
1535 7. If pcregrep was compiled under Windows, there was a reference to the
1536 function pcregrep_exit() before it was defined. I am assuming this was
1537 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1538 reported by a user. I've moved the definition above the reference.
1539
1540
1541 Version 8.11 10-Dec-2010
1542 ------------------------
1543
1544 1. (*THEN) was not working properly if there were untried alternatives prior
1545 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1546 backtracked to try for "b" instead of moving to the next alternative branch
1547 at the same level (in this case, to look for "c"). The Perl documentation
1548 is clear that when (*THEN) is backtracked onto, it goes to the "next
1549 alternative in the innermost enclosing group".
1550
1551 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1552 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1553 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1554 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1555 (*THEN).
1556
1557 3. If \s appeared in a character class, it removed the VT character from
1558 the class, even if it had been included by some previous item, for example
1559 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1560 of \s, but is part of the POSIX "space" class.)
1561
1562 4. A partial match never returns an empty string (because you can always
1563 match an empty string at the end of the subject); however the checking for
1564 an empty string was starting at the "start of match" point. This has been
1565 changed to the "earliest inspected character" point, because the returned
1566 data for a partial match starts at this character. This means that, for
1567 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1568 (previously it gave "no match").
1569
1570 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1571 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1572 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1573 has an implication that the given string is incomplete (because a partial
1574 match is preferred over a full match). For this reason, these items now
1575 give a partial match in this situation. [Aside: previously, the one case
1576 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1577 match rather than a full match, which was wrong by the old rules, but is
1578 now correct.]
1579
1580 6. There was a bug in the handling of #-introduced comments, recognized when
1581 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1582 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1583 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1584 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1585 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1586 places in pcre_compile().
1587
1588 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1589 comments when looking ahead for named forward references to subpatterns,
1590 the only newline sequence it recognized was NL. It now handles newlines
1591 according to the set newline convention.
1592
1593 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1594 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1595 cater for a lack of strerror(). These oversights have been fixed.
1596
1597 9. Added --match-limit and --recursion-limit to pcregrep.
1598
1599 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1600
1601 11. When the -o option was used, pcregrep was setting a return code of 1, even
1602 when matches were found, and --line-buffered was not being honoured.
1603
1604 12. Added an optional parentheses number to the -o and --only-matching options
1605 of pcregrep.
1606
1607 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1608 can match an empty string. The code to do it in pcretest and pcredemo
1609 needed fixing:
1610
1611 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1612 only one byte after an empty string match just before CRLF (this case
1613 just got forgotten; "any" and "anycrlf" were OK).
1614
1615 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1616 mode when an empty string match preceded an ASCII character followed by
1617 a non-ASCII character. (The code for advancing by one character rather
1618 than one byte was nonsense.)
1619
1620 (c) The pcredemo.c sample program did not have any code at all to handle
1621 the cases when CRLF is a valid newline sequence.
1622
1623 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1624 as a starting offset was within the subject string. There is now a new
1625 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1626 negative or greater than the length of the string. In order to test this,
1627 pcretest is extended to allow the setting of negative starting offsets.
1628
1629 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1630 starting offset points to the beginning of a UTF-8 character was
1631 unnecessarily clumsy. I tidied it up.
1632
1633 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1634 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1635
1636 17. Nobody had reported that the --include_dir option, which was added in
1637 release 7.7 should have been called --include-dir (hyphen, not underscore)
1638 for compatibility with GNU grep. I have changed it to --include-dir, but
1639 left --include_dir as an undocumented synonym, and the same for
1640 --exclude-dir, though that is not available in GNU grep, at least as of
1641 release 2.5.4.
1642
1643 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1644 characters from a string of bytes) have been redefined so as not to use
1645 loops, in order to improve performance in some environments. At the same
1646 time, I abstracted some of the common code into auxiliary macros to save
1647 repetition (this should not affect the compiled code).
1648
1649 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1650 compile-time error is now given if \c is not followed by an ASCII
1651 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1652 different, and any byte value is allowed.)
1653
1654 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1655 START_OPTIMIZE option, which is now allowed at compile time - but just
1656 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1657 to pcregrep and other applications that have no direct access to PCRE
1658 options. The new /Y option in pcretest sets this option when calling
1659 pcre_compile().
1660
1661 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1662 back references. Groups containing recursive back references were forced to
1663 be atomic by that change, but in the case of named groups, the amount of
1664 memory required was incorrectly computed, leading to "Failed: internal
1665 error: code overflow". This has been fixed.
1666
1667 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1668 pcretest.c, to avoid build problems in some Borland environments.
1669
1670
1671 Version 8.10 25-Jun-2010
1672 ------------------------
1673
1674 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1675 THEN.
1676
1677 2. (*ACCEPT) was not working when inside an atomic group.
1678
1679 3. Inside a character class, \B is treated as a literal by default, but
1680 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1681 causes the error). The code is unchanged, but I tidied the documentation.
1682
1683 4. Inside a character class, PCRE always treated \R and \X as literals,
1684 whereas Perl faults them if its -w option is set. I have changed PCRE so
1685 that it faults them when PCRE_EXTRA is set.
1686
1687 5. Added support for \N, which always matches any character other than
1688 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1689
1690 6. When compiling pcregrep with newer versions of gcc which may have
1691 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1692 declared with attribute warn_unused_result" were given. Just casting the
1693 result to (void) does not stop the warnings; a more elaborate fudge is
1694 needed. I've used a macro to implement this.
1695
1696 7. Minor change to pcretest.c to avoid a compiler warning.
1697
1698 8. Added four artifical Unicode properties to help with an option to make
1699 \s etc use properties (see next item). The new properties are: Xan
1700 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1701
1702 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1703 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1704 this option. Modified pcretest to add /W to test this facility. Added
1705 REG_UCP to make it available via the POSIX interface.
1706
1707 10. Added --line-buffered to pcregrep.
1708
1709 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1710 studied, and the match started with a letter with a code point greater than
1711 127 whose first byte was different to the first byte of the other case of
1712 the letter, the other case of this starting letter was not recognized
1713 (#976).
1714
1715 12. If a pattern that was studied started with a repeated Unicode property
1716 test, for example, \p{Nd}+, there was the theoretical possibility of
1717 setting up an incorrect bitmap of starting bytes, but fortunately it could
1718 not have actually happened in practice until change 8 above was made (it
1719 added property types that matched character-matching opcodes).
1720
1721 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1722 possible starting bytes for non-anchored patterns.
1723
1724 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1725 \R, and also a number of cases that involve Unicode properties, both
1726 explicit and implicit when PCRE_UCP is set.
1727
1728 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1729 input, it could crash or give wrong results if characters with values
1730 greater than 0xc0 were present in the subject string. (Detail: it assumed
1731 UTF-8 input when processing these items.)
1732
1733 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1734 size_t is 64-bit (#991).
1735
1736 17. Added a check for running out of memory when PCRE is compiled with
1737 --disable-stack-for-recursion (#990).
1738
1739 18. If the last data line in a file for pcretest does not have a newline on
1740 the end, a newline was missing in the output.
1741
1742 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1743 less than 128) in its various bitmaps. However, there is a facility for
1744 generating tables according to the current locale when PCRE is compiled. It
1745 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1746 characters, are recognized by isspace() and therefore were getting set in
1747 these tables, and indeed these tables seem to approximate to ISO 8859. This
1748 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1749 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1750 which of course cannot start UTF-8 characters. I have changed the code so
1751 that only real ASCII characters (less than 128) and the correct starting
1752 bytes for UTF-8 encodings are set for characters greater than 127 when in
1753 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1754 altogether.)
1755
1756 20. Added the /T option to pcretest so as to be able to run tests with non-
1757 standard character tables, thus making it possible to include the tests
1758 used for 19 above in the standard set of tests.
1759
1760 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1761 reference to a subpattern the other side of a comment that contains an
1762 opening parenthesis caused either an internal compiling error, or a
1763 reference to the wrong subpattern.
1764
1765
1766 Version 8.02 19-Mar-2010
1767 ------------------------
1768
1769 1. The Unicode data tables have been updated to Unicode 5.2.0.
1770
1771 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1772 configured.
1773
1774 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1775 original author of that file, following a query about its status.
1776
1777 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1778 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1779
1780 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1781 quantifier applied to a forward-referencing subroutine call, could compile
1782 incorrect code or give the error "internal error: previously-checked
1783 referenced subpattern not found".
1784
1785 6. Both MS Visual Studio and Symbian OS have problems with initializing
1786 variables to point to external functions. For these systems, therefore,
1787 pcre_malloc etc. are now initialized to local functions that call the
1788 relevant global functions.
1789
1790 7. There were two entries missing in the vectors called coptable and poptable
1791 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1792 I've fixed the data, and added a kludgy way of testing at compile time that
1793 the lengths are correct (equal to the number of opcodes).
1794
1795 8. Following on from 7, I added a similar kludge to check the length of the
1796 eint vector in pcreposix.c.
1797
1798 9. Error texts for pcre_compile() are held as one long string to avoid too
1799 much relocation at load time. To find a text, the string is searched,
1800 counting zeros. There was no check for running off the end of the string,
1801 which could happen if a new error number was added without updating the
1802 string.
1803
1804 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1805
1806 11. \K was not working if it appeared in an atomic group or in a group that
1807 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1808 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1809 the assertion is positive, but not if it is negative.
1810
1811 12. Change 11 fortuitously reduced the size of the stack frame used in the
1812 "match()" function of pcre_exec.c by one pointer. Forthcoming
1813 implementation of support for (*MARK) will need an extra pointer on the
1814 stack; I have reserved it now, so that the stack frame size does not
1815 decrease.
1816
1817 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1818 item in branch that calls a recursion is a subroutine call - as in the
1819 second branch in the above example - was incorrectly given the compile-
1820 time error "recursive call could loop indefinitely" because pcre_compile()
1821 was not correctly checking the subroutine for matching a non-empty string.
1822
1823 14. The checks for overrunning compiling workspace could trigger after an
1824 overrun had occurred. This is a "should never occur" error, but it can be
1825 triggered by pathological patterns such as hundreds of nested parentheses.
1826 The checks now trigger 100 bytes before the end of the workspace.
1827
1828 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1829
1830
1831 Version 8.01 19-Jan-2010
1832 ------------------------
1833
1834 1. If a pattern contained a conditional subpattern with only one branch (in
1835 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1836 computed the wrong minimum data length (which is of course zero for such
1837 subpatterns). This could cause incorrect "no match" results.
1838
1839 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1840 the pattern is reset in the first branch, pcre_compile() failed with
1841 "internal error: code overflow at offset...". This happened only when
1842 the reset was to the original external option setting. (An optimization
1843 abstracts leading options settings into an external setting, which was the
1844 cause of this.)
1845
1846 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1847 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1848 assertion pattern did not match (meaning that the assertion was true), it
1849 was incorrectly treated as false if the SKIP had been reached during the
1850 matching. This also applied to assertions used as conditions.
1851
1852 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1853 assertion subpattern, including such a pattern used as a condition,
1854 unpredictable results occurred, instead of the error return
1855 PCRE_ERROR_DFA_UITEM.
1856
1857 5. The C++ GlobalReplace function was not working like Perl for the special
1858 situation when an empty string is matched. It now does the fancy magic
1859 stuff that is necessary.
1860
1861 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1862 removed. (These were left over from very, very early versions of PCRE.)
1863
1864 7. Some cosmetic changes to the code to make life easier when compiling it
1865 as part of something else:
1866
1867 (a) Change DEBUG to PCRE_DEBUG.
1868
1869 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1870 called "current" as "current_branch", to prevent a collision with the
1871 Linux macro when compiled as a kernel module.
1872
1873 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1874 prevent a collision with the Linux macro when compiled as a kernel
1875 module.
1876
1877 8. In pcre_compile() there are some checks for integer overflows that used to
1878 cast potentially large values to (double). This has been changed to that
1879 when building, a check for int64_t is made, and if it is found, it is used
1880 instead, thus avoiding the use of floating point arithmetic. (There is no
1881 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1882 double.
1883
1884 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1885 2005 (difference between two addresses compared to an unsigned value).
1886
1887 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1888 custom one, because of the following reported problem in Windows:
1889
1890 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1891 under Win32.
1892 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1893 therefore missing the function definition.
1894 - The compiler thus generates a "C" signature for the test function.
1895 - The linker fails to find the "C" function.
1896 - PCRE fails to configure if asked to do so against libbz2.
1897
1898 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1899 messages were output:
1900
1901 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1902 rerunning libtoolize, to keep the correct libtool macros in-tree.
1903 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1904
1905 I have done both of these things.
1906
1907 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1908 most of the time, it *can* run out if it is given a pattern that contains a
1909 runaway infinite recursion. I updated the discussion in the pcrestack man
1910 page.
1911
1912 13. Now that we have gone to the x.xx style of version numbers, the minor
1913 version may start with zero. Using 08 or 09 is a bad idea because users
1914 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1915 interpreted as invalid octal numbers. I've updated the previous comment in
1916 configure.ac, and also added a check that gives an error if 08 or 09 are
1917 used.
1918
1919 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1920 causing partial matching to fail when the end of the subject matched \W
1921 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1922
1923 15. There were some discrepancies between the declarations in pcre_internal.h
1924 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1925 their definitions. The declarations used "const uschar *" and the
1926 definitions used USPTR. Even though USPTR is normally defined as "const
1927 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1928 reported that: "This difference in casting confuses some C++ compilers, for
1929 example, SunCC recognizes above declarations as different functions and
1930 generates broken code for hbpcre." I have changed the declarations to use
1931 USPTR.
1932
1933 16. GNU libtool is named differently on some systems. The autogen.sh script now
1934 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1935 (FreeBSD).
1936
1937 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1938 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1939 comment: "Figure out how to create a longlong from a string: strtoll and
1940 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1941 instance, but it only takes 2 args instead of 3!"
1942
1943 18. A subtle bug concerned with back references has been fixed by a change of
1944 specification, with a corresponding code fix. A pattern such as
1945 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1946 refers, was giving matches when it shouldn't. For example, xa=xaaa would
1947 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1948 same bug. Such groups have to be quantified to be useful, or contained
1949 inside another quantified group. (If there's no repetition, the reference
1950 can never match.) The problem arises because, having left the group and
1951 moved on to the rest of the pattern, a later failure that backtracks into
1952 the group uses the captured value from the final iteration of the group
1953 rather than the correct earlier one. I have fixed this in PCRE by forcing
1954 any group that contains a reference to itself to be an atomic group; that
1955 is, there cannot be any backtracking into it once it has completed. This is
1956 similar to recursive and subroutine calls.
1957
1958
1959 Version 8.00 19-Oct-09
1960 ----------------------
1961
1962 1. The table for translating pcre_compile() error codes into POSIX error codes
1963 was out-of-date, and there was no check on the pcre_compile() error code
1964 being within the table. This could lead to an OK return being given in
1965 error.
1966
1967 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1968 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1969 in a Windows environment.
1970
1971 3. The pcregrep --count option prints the count for each file even when it is
1972 zero, as does GNU grep. However, pcregrep was also printing all files when
1973 --files-with-matches was added. Now, when both options are given, it prints
1974 counts only for those files that have at least one match. (GNU grep just
1975 prints the file name in this circumstance, but including the count seems
1976 more useful - otherwise, why use --count?) Also ensured that the
1977 combination -clh just lists non-zero counts, with no names.
1978
1979 4. The long form of the pcregrep -F option was incorrectly implemented as
1980 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1981 but it seems right to fix it, and I didn't think it was worth preserving
1982 the old behaviour.
1983
1984 5. The command line items --regex=pattern and --regexp=pattern were not
1985 recognized by pcregrep, which required --regex pattern or --regexp pattern
1986 (with a space rather than an '='). The man page documented the '=' forms,
1987 which are compatible with GNU grep; these now work.
1988
1989 6. No libpcreposix.pc file was created for pkg-config; there was just
1990 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1991
1992 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1993 when UCP support is not needed, by modifying the Python script that
1994 generates it from Unicode data files. This should not matter if the module
1995 is correctly used as a library, but I received one complaint about 50K of
1996 unwanted data. My guess is that the person linked everything into his
1997 program rather than using a library. Anyway, it does no harm.
1998
1999 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2000 was a minimum greater than 1 for a wide character in a possessive
2001 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2002 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2003 character. Chaos in the form of incorrect output or a compiling loop could
2004 result.
2005
2006 9. The restrictions on what a pattern can contain when partial matching is
2007 requested for pcre_exec() have been removed. All patterns can now be
2008 partially matched by this function. In addition, if there are at least two
2009 slots in the offset vector, the offset of the earliest inspected character
2010 for the match and the offset of the end of the subject are set in them when
2011 PCRE_ERROR_PARTIAL is returned.
2012
2013 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2014 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2015 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2016 and may be more useful for multi-segment matching.
2017
2018 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2019 used to be given if ever the end of the subject was reached; now it is
2020 given only if matching could not proceed because another character was
2021 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2022 string "Z", which now yields "no match" instead of "partial match". In the
2023 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2024 final character ended with (*FAIL).
2025
2026 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2027 if the pattern had a "must contain" character that was already found in the
2028 earlier partial match, unless partial matching was again requested. For
2029 example, with the pattern /dog.(body)?/, the "must contain" character is
2030 "g". If the first part-match was for the string "dog", restarting with
2031 "sbody" failed. This bug has been fixed.
2032
2033 13. The string returned by pcre_dfa_exec() after a partial match has been
2034 changed so that it starts at the first inspected character rather than the
2035 first character of the match. This makes a difference only if the pattern
2036 starts with a lookbehind assertion or \b or \B (\K is not supported by
2037 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2038 matching functions compatible, and I think it's the right thing to do.
2039
2040 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2041 so that the demonstration program is easily available in environments where
2042 PCRE has not been installed from source.
2043
2044 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2045 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2046 library.
2047
2048 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2049 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2050 is not the first non-POSIX option to be added. Clearly some people find
2051 these options useful.
2052
2053 17. If a caller to the POSIX matching function regexec() passes a non-zero
2054 value for nmatch with a NULL value for pmatch, the value of
2055 nmatch is forced to zero.
2056
2057 18. RunGrepTest did not have a test for the availability of the -u option of
2058 the diff command, as RunTest does. It now checks in the same way as
2059 RunTest, and also checks for the -b option.
2060
2061 19. If an odd number of negated classes containing just a single character
2062 interposed, within parentheses, between a forward reference to a named
2063 subpattern and the definition of the subpattern, compilation crashed with
2064 an internal error, complaining that it could not find the referenced
2065 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2066 [The bug was that it was starting one character too far in when skipping
2067 over the character class, thus treating the ] as data rather than
2068 terminating the class. This meant it could skip too much.]
2069
2070 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2071 /g option in pcretest when the pattern contains \K, which makes it possible
2072 to have an empty string match not at the start, even when the pattern is
2073 anchored. Updated pcretest and pcredemo to use this option.
2074
2075 21. If the maximum number of capturing subpatterns in a recursion was greater
2076 than the maximum at the outer level, the higher number was returned, but
2077 with unset values at the outer level. The correct (outer level) value is
2078 now given.
2079
2080 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2081 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2082 make it do so. The string so far is captured, making this feature
2083 compatible with Perl.
2084
2085 23. The tests have been re-organized, adding tests 11 and 12, to make it
2086 possible to check the Perl 5.10 features against Perl 5.10.
2087
2088 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2089 pattern matches a fixed length string. PCRE did not allow this; now it
2090 does. Neither allows recursion.
2091
2092 25. I finally figured out how to implement a request to provide the minimum
2093 length of subject string that was needed in order to match a given pattern.
2094 (It was back references and recursion that I had previously got hung up
2095 on.) This code has now been added to pcre_study(); it finds a lower bound
2096 to the length of subject needed. It is not necessarily the greatest lower
2097 bound, but using it to avoid searching strings that are too short does give
2098 some useful speed-ups. The value is available to calling programs via
2099 pcre_fullinfo().
2100
2101 26. While implementing 25, I discovered to my embarrassment that pcretest had
2102 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2103 study optimizations had never been tested with that matching function.
2104 Oops. What is worse, even when it was passed study data, there was a bug in
2105 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2106 were also very few tests of studied patterns with pcre_dfa_exec().
2107
2108 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2109 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2110 on the other side of the coin, they are no longer allowed to have different
2111 names, because these cannot be distinguished in PCRE, and this has caused
2112 confusion. (This is a difference from Perl.)
2113
2114 28. When duplicate subpattern names are present (necessarily with different
2115 numbers, as required by 27 above), and a test is made by name in a
2116 conditional pattern, either for a subpattern having been matched, or for
2117 recursion in such a pattern, all the associated numbered subpatterns are
2118 tested, and the overall condition is true if the condition is true for any
2119 one of them. This is the way Perl works, and is also more like the way
2120 testing by number works.
2121
2122
2123 Version 7.9 11-Apr-09
2124 ---------------------
2125
2126 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2127 (pcretest), all targets were linked against these libraries. This included
2128 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2129 libraries. This caused unwanted dependencies to be created. This problem
2130 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2131 pcretest is linked with readline.
2132
2133 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2134 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2135 moved outside it again, because FALSE and TRUE are already defined in AIX,
2136 but BOOL is not.
2137
2138 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2139 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2140
2141 4. The pcregrep documentation said spaces were inserted as well as colons (or
2142 hyphens) following file names and line numbers when outputting matching
2143 lines. This is not true; no spaces are inserted. I have also clarified the
2144 wording for the --colour (or --color) option.
2145
2146 5. In pcregrep, when --colour was used with -o, the list of matching strings
2147 was not coloured; this is different to GNU grep, so I have changed it to be
2148 the same.
2149
2150 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2151 each matching line was coloured. Now it goes on to look for further matches
2152 of any of the test patterns, which is the same behaviour as GNU grep.
2153
2154 7. A pattern that could match an empty string could cause pcregrep to loop; it
2155 doesn't make sense to accept an empty string match in pcregrep, so I have
2156 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2157 seems to be how GNU grep behaves. [But see later change 40 for release
2158 8.33.]
2159
2160 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2161 start or after a newline", because the conditional assertion was not being
2162 correctly handled. The rule now is that both the assertion and what follows
2163 in the first alternative must satisfy the test.
2164
2165 9. If auto-callout was enabled in a pattern with a conditional group whose
2166 condition was an assertion, PCRE could crash during matching, both with
2167 pcre_exec() and pcre_dfa_exec().
2168
2169 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2170 used for matching.
2171
2172 11. Unicode property support in character classes was not working for
2173 characters (bytes) greater than 127 when not in UTF-8 mode.
2174
2175 12. Added the -M command line option to pcretest.
2176
2177 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2178
2179 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2180
2181 16. Added comments and documentation about mis-use of no_arg in the C++
2182 wrapper.
2183
2184 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2185 from Martin Jerabek that uses macro names for all relevant character and
2186 string constants.
2187
2188 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2189 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2190 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2191 these, but not everybody uses configure.
2192
2193 19. A conditional group that had only one branch was not being correctly
2194 recognized as an item that could match an empty string. This meant that an
2195 enclosing group might also not be so recognized, causing infinite looping
2196 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2197 with the subject "ab", where knowledge that the repeated group can match
2198 nothing is needed in order to break the loop.
2199
2200 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2201 exec(), but without supplying a callout function, matching went wrong.
2202
2203 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2204 leak if the size of the offset vector was greater than 30. When the vector
2205 is smaller, the saved offsets during recursion go onto a local stack
2206 vector, but for larger vectors malloc() is used. It was failing to free
2207 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2208 error, in fact).
2209
2210 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2211 heapframe that is used only when UTF-8 support is enabled. This caused no
2212 problem, but was untidy.
2213
2214 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2215 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2216 included within another project.
2217
2218 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2219 slightly modified by me:
2220
2221 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2222 not building pcregrep.
2223
2224 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2225 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2226
2227 25. Forward references, both numeric and by name, in patterns that made use of
2228 duplicate group numbers, could behave incorrectly or give incorrect errors,
2229 because when scanning forward to find the reference group, PCRE was not
2230 taking into account the duplicate group numbers. A pattern such as
2231 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2232
2233 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2234 the feature of a custom pointer more persuasive (as requested by a user).
2235
2236 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2237 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2238 pre-defined.
2239
2240 28. Added support for (*UTF8) at the start of a pattern.
2241
2242 29. Arrange for flags added by the "release type" setting in CMake to be shown
2243 in the configuration summary.
2244
2245
2246 Version 7.8 05-Sep-08
2247 ---------------------
2248
2249 1. Replaced UCP searching code with optimized version as implemented for Ad
2250 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2251 stage table and inline lookup instead of a function, giving speed ups of 2
2252 to 5 times on some simple patterns that I tested. Permission was given to
2253 distribute the MultiStage2.py script that generates the tables (it's not in
2254 the tarball, but is in the Subversion repository).
2255
2256 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2257 scripts.
2258
2259 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2260 a group with a zero qualifier. The result of the study could be incorrect,
2261 or the function might crash, depending on the pattern.
2262
2263 4. Caseless matching was not working for non-ASCII characters in back
2264 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2265 It now works when Unicode Property Support is available.
2266
2267 5. In pcretest, an escape such as \x{de} in the data was always generating
2268 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2269 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2270 truncation.
2271
2272 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2273
2274 7. Added two (int) casts to pcregrep when printing the difference of two
2275 pointers, in case they are 64-bit values.
2276
2277 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2278 test 2 if it fails.
2279
2280 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2281 and a #define of that name to empty if it is not externally set. This is to
2282 allow users of MSVC to set it if necessary.
2283
2284 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2285 the convenience functions in the pcre_get.c source file.
2286
2287 11. An option change at the start of a pattern that had top-level alternatives
2288 could cause overwriting and/or a crash. This command provoked a crash in
2289 some environments:
2290
2291 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2292
2293 This potential security problem was recorded as CVE-2008-2371.
2294
2295 12. For a pattern where the match had to start at the beginning or immediately
2296 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2297 pcre_dfa_exec() could read past the end of the passed subject if there was
2298 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2299 pcretest so that it places the subject at the end of its malloc-ed buffer.
2300
2301 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2302 exec() might read past the end of the data buffer in UTF-8 mode.
2303
2304 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2305 the data contained the byte 0x85 as part of a UTF-8 character within its
2306 first line. This applied both to normal and DFA matching.
2307
2308 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2309 /^[^d]*?$/8 failed to match "abc".
2310
2311 16. Added a missing copyright notice to pcrecpp_internal.h.
2312
2313 17. Make it more clear in the documentation that values returned from
2314 pcre_exec() in ovector are byte offsets, not character counts.
2315
2316 18. Tidied a few places to stop certain compilers from issuing warnings.
2317
2318 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2319 supplied by Stefan Weber. I made a further small update for 7.8 because
2320 there is a change of source arrangements: the pcre_searchfuncs.c module is
2321 replaced by pcre_ucd.c.
2322
2323
2324 Version 7.7 07-May-08
2325 ---------------------
2326
2327 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2328 a string to a long long, pretend we don't even have a long long." This is
2329 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2330
2331 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2332 pre-7.6 versions, which defined a global no_arg variable instead of putting
2333 it in the RE class. (See also #8 below.)
2334
2335 3. Remove a line of dead code, identified by coverity and reported by Nuno
2336 Lopes.
2337
2338 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2339
2340 (1) The include/exclude patterns were being applied to the whole pathnames
2341 of files, instead of just to the final components.
2342
2343 (2) If there was more than one level of directory, the subdirectories were
2344 skipped unless they satisfied the include/exclude conditions. This is
2345 inconsistent with GNU grep (and could even be seen as contrary to the
2346 pcregrep specification - which I improved to make it absolutely clear).
2347 The action now is always to scan all levels of directory, and just
2348 apply the include/exclude patterns to regular files.
2349
2350 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2351 --exclude_dir in the tests to avoid scanning .svn directories.
2352
2353 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2354 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2355 doesn't support NULs in patterns.
2356
2357 7. Added some missing "const"s to declarations of static tables in
2358 pcre_compile.c and pcre_dfa_exec.c.
2359
2360 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2361 caused by fix #2 above. (Subsequently also a second patch to fix the
2362 first patch. And a third patch - this was a messy problem.)
2363
2364 9. Applied Craig's patch to remove the use of push_back().
2365
2366 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2367 matching function regexec().
2368
2369 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2370 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2371 references. PCRE supports relative numbers with this syntax (I don't think
2372 Oniguruma does).
2373
2374 12. Previously, a group with a zero repeat such as (...){0} was completely
2375 omitted from the compiled regex. However, this means that if the group
2376 was called as a subroutine from elsewhere in the pattern, things went wrong
2377 (an internal error was given). Such groups are now left in the compiled
2378 pattern, with a new opcode that causes them to be skipped at execution
2379 time.
2380
2381 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2382 to the way PCRE behaves:
2383
2384 (a) A lone ] character is dis-allowed (Perl treats it as data).
2385
2386 (b) A back reference to an unmatched subpattern matches an empty string
2387 (Perl fails the current match path).
2388
2389 (c) A data ] in a character class must be notated as \] because if the
2390 first data character in a class is ], it defines an empty class. (In
2391 Perl it is not possible to have an empty class.) The empty class []
2392 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2393 The negative empty class [^] matches any one character, independently
2394 of the DOTALL setting.
2395
2396 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2397 non-existent subpattern following a character class starting with ']' and
2398 containing () gave an internal compiling error instead of "reference to
2399 non-existent subpattern". Fortunately, when the pattern did exist, the
2400 compiled code was correct. (When scanning forwards to check for the
2401 existence of the subpattern, it was treating the data ']' as terminating
2402 the class, so got the count wrong. When actually compiling, the reference
2403 was subsequently set up correctly.)
2404
2405 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2406 it was being rejected as not supported by pcre_dfa_exec(), even though
2407 other assertions are supported. I have made pcre_dfa_exec() support
2408 (*FAIL).
2409
2410 16. The implementation of 13c above involved the invention of a new opcode,
2411 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2412 cannot be changed at match time, I realized I could make a small
2413 improvement to matching performance by compiling OP_ALLANY instead of
2414 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2415 on the OP_ANY path.
2416
2417 17. Compiling pcretest on Windows with readline support failed without the
2418 following two fixes: (1) Make the unistd.h include conditional on
2419 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2420
2421 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2422 ncurses library to be included for pcretest when ReadLine support is
2423 requested, but also to allow for it to be overridden. This patch came from
2424 Daniel Bergström.
2425
2426 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2427 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2428 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2429 spotting this.
2430
2431
2432 Version 7.6 28-Jan-08
2433 ---------------------
2434
2435 1. A character class containing a very large number of characters with
2436 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2437 overflow.
2438
2439 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2440 HAVE_LONG_LONG is not defined.
2441
2442 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2443 bring them up to date and include new features. This patch includes:
2444
2445 - Fixed PH's badly added libz and libbz2 support.
2446 - Fixed a problem with static linking.
2447 - Added pcredemo. [But later removed - see 7 below.]
2448 - Fixed dftables problem and added an option.
2449 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2450 HAVE_LONG_LONG.
2451 - Added readline support for pcretest.
2452 - Added an listing of the option settings after cmake has run.
2453
2454 4. A user submitted a patch to Makefile that makes it easy to create
2455 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2456 Makefile.am that cause it to include this special target, without
2457 affecting anything else. Note that the same mingw target plus all
2458 the other distribution libraries and programs are now supported
2459 when configuring with CMake (see 6 below) instead of with
2460 Configure/Make.
2461
2462 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2463 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2464 exported in the Windows port". It has not yet been confirmed that the patch
2465 solves the problem, but it does no harm.
2466
2467 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2468 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2469 with CMake, and also correct the comment about stack recursion.
2470
2471 7. Remove the automatic building of pcredemo from the ./configure system and
2472 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2473 of a program that users should build themselves after PCRE is installed, so
2474 building it automatically is not really right. What is more, it gave
2475 trouble in some build environments.
2476
2477 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2478
2479
2480 Version 7.5 10-Jan-08
2481 ---------------------
2482
2483 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2484 values in parens when parsing an RE using the C++ wrapper."
2485
2486 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2487 Characters greater than 255 were excluded from the class instead of being
2488 included.
2489
2490 3. The same bug as (2) above applied to negated POSIX classes such as
2491 [:^space:].
2492
2493 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2494 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2495 I have changed it.
2496
2497 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2498 first named subpattern) and a construct such as (?&a) would reference the
2499 first named subpattern whose name started with "a" (in other words, the
2500 length check was missing). Both these problems are fixed. "Subpattern name
2501 expected" is now given for (?&) (a zero-length name), and this patch also
2502 makes it give the same error for \k'' (previously it complained that that
2503 was a reference to a non-existent subpattern).
2504
2505 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2506 this is right because (?- can be followed by option settings as well as by
2507 digits. I have, however, made the messages clearer.
2508
2509 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2510 than the number used in the conditional) now cause a compile-time error.
2511 This is actually not compatible with Perl, which accepts such patterns, but
2512 treats the conditional as always being FALSE (as PCRE used to), but it
2513 seems to me that giving a diagnostic is better.
2514
2515 8. Change "alphameric" to the more common word "alphanumeric" in comments
2516 and messages.
2517
2518 9. Fix two occurrences of "backslash" in comments that should have been
2519 "backspace".
2520
2521 10. Remove two redundant lines of code that can never be obeyed (their function
2522 was moved elsewhere).
2523
2524 11. The program that makes PCRE's Unicode character property table had a bug
2525 which caused it to generate incorrect table entries for sequences of
2526 characters that have the same character type, but are in different scripts.
2527 It amalgamated them into a single range, with the script of the first of
2528 them. In other words, some characters were in the wrong script. There were
2529 thirteen such cases, affecting characters in the following ranges:
2530
2531 U+002b0 - U+002c1
2532 U+0060c - U+0060d
2533 U+0061e - U+00612
2534 U+0064b - U+0065e
2535 U+0074d - U+0076d
2536 U+01800 - U+01805
2537 U+01d00 - U+01d77
2538 U+01d9b - U+01dbf
2539 U+0200b - U+0200f
2540 U+030fc - U+030fe
2541 U+03260 - U+0327f
2542 U+0fb46 - U+0fbb1
2543 U+10450 - U+1049d
2544
2545 12. The -o option (show only the matching part of a line) for pcregrep was not
2546 compatible with GNU grep in that, if there was more than one match in a
2547 line, it showed only the first of them. It now behaves in the same way as
2548 GNU grep.
2549
2550 13. If the -o and -v options were combined for pcregrep, it printed a blank
2551 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2552 does the same. The return code can be used to tell if there were any
2553 non-matching lines.
2554
2555 14. Added --file-offsets and --line-offsets to pcregrep.
2556
2557 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2558 infinitely looping recursion. The bug was that positive lookaheads were not
2559 being skipped when checking for a possible empty match (negative lookaheads
2560 and both kinds of lookbehind were skipped).
2561
2562 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2563 inclusion of <windows.h> to before rather than after the definition of
2564 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2565
2566 17. Specifying a possessive quantifier with a specific limit for a Unicode
2567 character property caused pcre_compile() to compile bad code, which led at
2568 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2569 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2570 caused the error; without that there was no problem.
2571
2572 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2573
2574 19. Added --enable-pcretest-libreadline.
2575
2576 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2577 RE::GlobalReplace(). As a result, the number of replacements returned was
2578 double what it should be. I removed one of the increments, but Craig sent a
2579 later patch that removed the other one (the right fix) and added unit tests
2580 that check the return values (which was not done before).
2581
2582 21. Several CMake things:
2583
2584 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2585 the names libpcre and libpcreposix, not just pcre and pcreposix.
2586
2587 (2) The above change means that pcretest and pcregrep are now correctly
2588 linked with the newly-built libraries, not previously installed ones.
2589
2590 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2591
2592 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2593 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2594 UTF-8 newline character). The key issue is that the pattern starts .*;
2595 this means that the match must be either at the beginning, or after a
2596 newline. The bug was in the code for advancing after a failed match and
2597 checking that the new position followed a newline. It was not taking
2598 account of UTF-8 characters correctly.
2599
2600 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2601 character classes. PCRE was not treating the sequence [:...:] as a
2602 character class unless the ... were all letters. Perl, however, seems to
2603 allow any characters between [: and :], though of course it rejects as
2604 unknown any "names" that contain non-letters, because all the known class
2605 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2606 for example, whereas PCRE did not - it did not recognize a POSIX character
2607 class. This seemed a bit dangerous, so the code has been changed to be
2608 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2609 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2610 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2611 Perl does, and where it didn't before.
2612
2613 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2614 Windows environments %n is disabled by default.
2615
2616
2617 Version 7.4 21-Sep-07
2618 ---------------------
2619
2620 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2621 means that a class such as [\s] counted as "explicit reference to CR or
2622 LF". That isn't really right - the whole point of the change was to try to
2623 help when there was an actual mention of one of the two characters. So now
2624 the change happens only if \r or \n (or a literal CR or LF) character is
2625 encountered.
2626
2627 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2628 of both had grown to the point where there were only 3 bits left.
2629 Fortunately, there was spare space in the data structure, and so I have
2630 moved the internal flags into a new 16-bit field to free up more option
2631 bits.
2632
2633 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2634 but did not set the internal JCHANGED flag - either of these is enough to
2635 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2636 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2637 start sets both bits.
2638
2639 4. Added options (at build time, compile time, exec time) to change \R from
2640 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2641
2642 5. doc/pcresyntax.html was missing from the distribution.
2643
2644 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2645 compatibility, even though it is no longer used.
2646
2647 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2648 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2649 windows.h file is present (where different names are used). [This was
2650 reversed later after testing - see 16 below.]
2651
2652 8. Changed all #include <config.h> to #include "config.h". There were also
2653 some further <pcre.h> cases that I changed to "pcre.h".
2654
2655 9. When pcregrep was used with the --colour option, it missed the line ending
2656 sequence off the lines that it output.
2657
2658 10. It was pointed out to me that arrays of string pointers cause lots of
2659 relocations when a shared library is dynamically loaded. A technique of
2660 using a single long string with a table of offsets can drastically reduce
2661 these. I have refactored PCRE in four places to do this. The result is
2662 dramatic:
2663
2664 Originally: 290
2665 After changing UCP table: 187
2666 After changing error message table: 43
2667 After changing table of "verbs" 36
2668 After changing table of Posix names 22
2669
2670 Thanks to the folks working on Gregex for glib for this insight.
2671
2672 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2673 unicode-properties was also set.
2674
2675 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2676
2677 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2678 checked only for CRLF.
2679
2680 14. Added casts to pcretest.c to avoid compiler warnings.
2681
2682 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2683
2684 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2685 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2686 entirely. This removes changes made in 7 above.
2687
2688 17. The CMake files have been updated, and there is now more information about
2689 building with CMake in the NON-UNIX-USE document.
2690
2691
2692 Version 7.3 28-Aug-07
2693 ---------------------
2694
2695 1. In the rejigging of the build system that eventually resulted in 7.1, the
2696 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2697 brackets there is not right, since it causes compilers to look for an
2698 installed pcre.h, not the version that is in the source that is being
2699 compiled (which of course may be different). I have changed it back to:
2700
2701 #include "pcre.h"
2702
2703 I have a vague recollection that the change was concerned with compiling in
2704 different directories, but in the new build system, that is taken care of
2705 by the VPATH setting the Makefile.
2706
2707 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2708 when the subject happened to end in the byte 0x85 (e.g. if the last
2709 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2710 characters but of course it shouldn't be taken as a newline when it is part
2711 of another character. The bug was that, for an unlimited repeat of . in
2712 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2713 characters when looking for a newline.
2714
2715 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2716
2717 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2718 in debug output.
2719
2720 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2721 long printing in the pcrecpp unittest when running under MinGW.
2722
2723 6. ESC_K was left out of the EBCDIC table.
2724
2725 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2726 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2727 limit also applies to "virtual nesting" when a pattern is recursive, and in
2728 this case 1000 isn't so big. I have been able to remove this limit at the
2729 expense of backing off one optimization in certain circumstances. Normally,
2730 when pcre_exec() would call its internal match() function recursively and
2731 immediately return the result unconditionally, it uses a "tail recursion"
2732 feature to save stack. However, when a subpattern that can match an empty
2733 string has an unlimited repetition quantifier, it no longer makes this
2734 optimization. That gives it a stack frame in which to save the data for
2735 checking that an empty string has been matched. Previously this was taken
2736 from the 1000-entry workspace that had been reserved. So now there is no
2737 explicit limit, but more stack is used.
2738
2739 8. Applied Daniel's patches to solve problems with the import/export magic
2740 syntax that is required for Windows, and which was going wrong for the
2741 pcreposix and pcrecpp parts of the library. These were overlooked when this
2742 problem was solved for the main library.
2743
2744 9. There were some crude static tests to avoid integer overflow when computing
2745 the size of patterns that contain repeated groups with explicit upper
2746 limits. As the maximum quantifier is 65535, the maximum group length was
2747 set at 30,000 so that the product of these two numbers did not overflow a
2748 32-bit integer. However, it turns out that people want to use groups that
2749 are longer than 30,000 bytes (though not repeat them that many times).
2750 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2751 made it possible to implement the integer overflow checks in a much more
2752 dynamic way, which I have now done. The artificial limitation on group
2753 length has been removed - we now have only the limit on the total length of
2754 the compiled pattern, which depends on the LINK_SIZE setting.
2755
2756 10. Fixed a bug in the documentation for get/copy named substring when
2757 duplicate names are permitted. If none of the named substrings are set, the
2758 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2759 empty string.
2760
2761 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2762 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2763 because the ] is interpreted as the first data character and the
2764 terminating ] is not found. PCRE has been made compatible with Perl in this
2765 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2766 cause memory overwriting.
2767
2768 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2769 string has been matched (to stop an infinite loop). It was not recognizing
2770 a conditional subpattern that could match an empty string if that
2771 subpattern was within another subpattern. For example, it looped when
2772 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2773 condition was not nested. This bug has been fixed.
2774
2775 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2776 past the start of the subject in the presence of bytes with the top bit
2777 set, for example "\x8aBCD".
2778
2779 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2780 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2781
2782 14. Optimized (?!) to (*FAIL).
2783
2784 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2785 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2786 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2787 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2788 does: it's just the validity check that is more restrictive.
2789
2790 16. Inserted checks for integer overflows during escape sequence (backslash)
2791 processing, and also fixed erroneous offset values for syntax errors during
2792 backslash processing.
2793
2794 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2795 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2796
2797 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2798 caused an overrun.
2799
2800 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2801 something other than just ASCII characters) inside a group that had an
2802 unlimited repeat caused a loop at compile time (while checking to see
2803 whether the group could match an empty string).
2804
2805 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2806 [\P{Any}] did so. (Error in the code for printing property names.)
2807
2808 21. An orphan \E inside a character class could cause a crash.
2809
2810 22. A repeated capturing bracket such as (A)? could cause a wild memory
2811 reference during compilation.
2812
2813 23. There are several functions in pcre_compile() that scan along a compiled
2814 expression for various reasons (e.g. to see if it's fixed length for look
2815 behind). There were bugs in these functions when a repeated \p or \P was
2816 present in the pattern. These operators have additional parameters compared
2817 with \d, etc, and these were not being taken into account when moving along
2818 the compiled data. Specifically:
2819
2820 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2821 length.
2822
2823 (b) An item such as \pL+ within a repeated group could cause crashes or
2824 loops.
2825
2826 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2827 "reference to non-existent subpattern" error.
2828
2829 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2830
2831 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2832 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2833
2834 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2835
2836 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2837 character were causing crashes (broken optimization).
2838
2839 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2840 \p or \P) caused a compile-time loop.
2841
2842 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2843 break. For example, the unstudied pattern [\r\n]A does not match the string
2844 "\r\nA" because change 7.0/46 below moves the current point on by two
2845 characters after failing to match at the start. However, the pattern \nA
2846 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2847 the same is true. There doesn't seem any very clean way out of this, but
2848 what I have chosen to do makes the common cases work: PCRE now takes note
2849 of whether there can be an explicit match for \r or \n anywhere in the
2850 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2851 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2852 pattern has explicit CR or LF references.
2853
2854 29. Added (*CR) etc for changing newline setting at start of pattern.
2855
2856
2857 Version 7.2 19-Jun-07
2858 ---------------------
2859
2860 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2861 which is apparently normally available under Windows.
2862
2863 2. Re-jig the pcregrep tests with different newline settings in an attempt
2864 to make them independent of the local environment's newline setting.
2865
2866 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2867
2868 4. Some of the "internals" tests were previously cut out when the link size
2869 was not 2, because the output contained actual offsets. The recent new
2870 "Z" feature of pcretest means that these can be cut out, making the tests
2871 usable with all link sizes.
2872
2873 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2874 stack recursion. This gives a massive performance boost under BSD, but just
2875 a small improvement under Linux. However, it saves one field in the frame
2876 in all cases.
2877
2878 6. Added more features from the forthcoming Perl 5.10:
2879
2880 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2881 recursion call. It refers to the nth most recently opened parentheses.
2882
2883 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2884 to be opened parentheses.
2885
2886 (c) Conditions that refer to capturing parentheses can be specified
2887 relatively, for example, (?(-2)... or (?(+3)...
2888
2889 (d) \K resets the start of the current match so that everything before
2890 is not part of it.
2891
2892 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2893
2894 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2895 reference syntax.
2896
2897 (g) (?| introduces a group in which the numbering of parentheses in each
2898 alternative starts with the same number.
2899
2900 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2901
2902 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2903 PCRE_INFO_JCHANGED.
2904
2905 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2906 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2907 for detecting groups that can match an empty string.
2908
2909 9. A pattern with a very large number of alternatives (more than several
2910 hundred) was running out of internal workspace during the pre-compile
2911 phase, where pcre_compile() figures out how much memory will be needed. A
2912 bit of new cunning has reduced the workspace needed for groups with
2913 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2914 workspace instead of running out of the 4096 that are available.
2915
2916 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2917
2918 11. Applied patch from Google to remove an optimization that didn't quite work.
2919 The report of the bug said:
2920
2921 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2922 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2923 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2924
2925 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2926 it matched the wrong number of bytes.
2927
2928
2929 Version 7.1 24-Apr-07
2930 ---------------------
2931
2932 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2933 that is more "standard", making use of automake and other Autotools. There
2934 is some re-arrangement of the files and adjustment of comments consequent
2935 on this.
2936
2937 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2938 for recursive directory scanning broke on some systems because the files
2939 are not scanned in any specific order and on different systems the order
2940 was different. A call to "sort" has been inserted into RunGrepTest for the
2941 approprate test as a short-term fix. In the longer term there may be an
2942 alternative.
2943
2944 3. I had an email from Eric Raymond about problems translating some of PCRE's
2945 man pages to HTML (despite the fact that I distribute HTML pages, some
2946 people do their own conversions for various reasons). The problems
2947 concerned the use of low-level troff macros .br and .in. I have therefore
2948 removed all such uses from the man pages (some were redundant, some could
2949 be replaced by .nf/.fi pairs). The 132html script that I use to generate
2950 HTML has been updated to handle .nf/.fi and to complain if it encounters
2951 .br or .in.
2952
2953 4. Updated comments in configure.ac that get placed in config.h.in and also
2954 arranged for config.h to be included in the distribution, with the name
2955 config.h.generic, for the benefit of those who have to compile without
2956 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2957
2958 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2959 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2960 makevp.bat for the current PCRE, using the additional files
2961 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2962
2963 6. A Windows user reported a minor discrepancy with test 2, which turned out
2964 to be caused by a trailing space on an input line that had got lost in his
2965 copy. The trailing space was an accident, so I've just removed it.
2966
2967 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2968 that is needed.
2969
2970 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2971 as "const" (a) because they are and (b) because it helps the PHP
2972 maintainers who have recently made a script to detect big data structures
2973 in the php code that should be moved to the .rodata section. I remembered
2974 to update Builducptable as well, so it won't revert if ucptable.h is ever
2975 re-created.
2976
2977 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2978 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2979 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2980 support is not required. This saves 1.5-2K of code, which is important in
2981 some applications.
2982
2983 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2984 so as not to refer to the tables, even though these functions will never be
2985 called when UTF-8 support is disabled. Otherwise there are problems with a
2986 shared library.
2987
2988 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2989
2990 (a) It was defining its arguments as char * instead of void *.
2991
2992 (b) It was assuming that all moves were upwards in memory; this was true
2993 a long time ago when I wrote it, but is no longer the case.
2994
2995 The emulated memove() is provided for those environments that have neither
2996 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2997 is clearly not the case, as these two bugs were recently reported.
2998
2999 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3000 and Detrail to create the HTML documentation, the .txt form of the man
3001 pages, and it removes trailing spaces from listed files. It also creates
3002 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3003 case, it wraps all the #defines with #ifndefs. This script should be run
3004 before "make dist".
3005
3006 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3007 with Unicode property support.
3008
3009 (a) For a maximizing quantifier, if the two different cases of the
3010 character were of different lengths in their UTF-8 codings (there are
3011 some cases like this - I found 11), and the matching function had to
3012 back up over a mixture of the two cases, it incorrectly assumed they
3013 were both the same length.
3014
3015 (b) When PCRE was configured to use the heap rather than the stack for
3016 recursion during matching, it was not correctly preserving the data for
3017 the other case of a UTF-8 character when checking ahead for a match
3018 while processing a minimizing repeat. If the check also involved
3019 matching a wide character, but failed, corruption could cause an
3020 erroneous result when trying to check for a repeat of the original
3021 character.
3022
3023 13. Some tidying changes to the testing mechanism:
3024
3025 (a) The RunTest script now detects the internal link size and whether there
3026 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3027 values substituted by "configure". (The RunGrepTest script already did
3028 this for UTF-8.) The configure.ac script no longer substitutes the
3029 relevant variables.
3030
3031 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3032 with length and offset values. This means that the output is different
3033 for different internal link sizes. Test 2 is skipped for link sizes
3034 other than 2 because of this, bypassing the problem. Unfortunately,
3035 there was also a test in test 3 (the locale tests) that used /B and
3036 failed for link sizes other than 2. Rather than cut the whole test out,
3037 I have added a new /Z option to pcretest that replaces the length and
3038 offset values with spaces. This is now used to make test 3 independent
3039 of link size. (Test 2 will be tidied up later.)
3040
3041 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3042 segmentation fault instead of returning the appropriate error message.
3043
3044 15. In multiline mode when the newline sequence was set to "any", the pattern
3045 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3046 This doesn't seem right; it now treats the CRLF combination as the line
3047 ending, and so does not match in that case. It's only a pattern such as ^$
3048 that would hit this one: something like ^ABC$ would have failed after \r
3049 and then tried again after \r\n.
3050
3051 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3052 in an attempt to make files that differ only in their line terminators
3053 compare equal. This works on Linux.
3054
3055 17. Under certain error circumstances pcregrep might try to free random memory
3056 as it exited. This is now fixed, thanks to valgrind.
3057
3058 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3059 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3060 was because its rules for how to advance for /g after matching an empty
3061 string at the end of a line did not allow for this case. They now check for
3062 it specially.
3063
3064 20. pcretest is supposed to handle patterns and data of any length, by
3065 extending its buffers when necessary. It was getting this wrong when the
3066 buffer for a data line had to be extended.
3067
3068 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3069 CRLF as a newline sequence.
3070
3071 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3072 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3073 I have nevertheless tidied it up.
3074
3075 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3076
3077 24. Added a man page for pcre-config.
3078
3079
3080 Version 7.0 19-Dec-06
3081 ---------------------
3082
3083 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3084 moving to gcc 4.1.1.
3085
3086 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3087 sys/time.h, which is documented as needed for this function. It doesn't
3088 seem to matter on Linux, but it showed up on some releases of OS X.
3089
3090 3. It seems that there are systems where bytes whose values are greater than
3091 127 match isprint() in the "C" locale. The "C" locale should be the
3092 default when a C program starts up. In most systems, only ASCII printing
3093 characters match isprint(). This difference caused the output from pcretest
3094 to vary, making some of the tests fail. I have changed pcretest so that:
3095
3096 (a) When it is outputting text in the compiled version of a pattern, bytes
3097 other than 32-126 are always shown as hex escapes.
3098
3099 (b) When it is outputting text that is a matched part of a subject string,
3100 it does the same, unless a different locale has been set for the match
3101 (using the /L modifier). In this case, it uses isprint() to decide.
3102
3103 4. Fixed a major bug that caused incorrect computation of the amount of memory
3104 required for a compiled pattern when options that changed within the
3105 pattern affected the logic of the preliminary scan that determines the
3106 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3107 that the computed length was too small. The symptoms of this bug were
3108 either the PCRE error "internal error: code overflow" from pcre_compile(),
3109 or a glibc crash with a message such as "pcretest: free(): invalid next
3110 size (fast)". Examples of patterns that provoked this bug (shown in
3111 pcretest format) are:
3112
3113 /(?-x: )/x
3114 /(?x)(?-x: \s*#\s*)/
3115 /((?i)[\x{c0}])/8
3116 /(?i:[\x{c0}])/8
3117
3118 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3119 is now done differently.
3120
3121 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3122 wrapper classes; (b) implement a new function in the C++ scanner that is
3123 more efficient than the old way of doing things because it avoids levels of
3124 recursion in the regex matching; (c) add a paragraph to the documentation
3125 for the FullMatch() function.
3126
3127 6. The escape sequence \n was being treated as whatever was defined as
3128 "newline". Not only was this contrary to the documentation, which states
3129 that \n is character 10 (hex 0A), but it also went horribly wrong when
3130 "newline" was defined as CRLF. This has been fixed.
3131
3132 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3133 was being set to -1 for the "end of line" case (supposedly a value that no
3134 character can have). Though this value is never used (the check for end of
3135 line is "zero bytes in current character"), it caused compiler complaints.
3136 I've changed it to 0xffffffff.
3137
3138 8. In pcre_version.c, the version string was being built by a sequence of
3139 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3140 string (as it is for production releases) called a macro with an empty
3141 argument. The C standard says the result of this is undefined. The gcc
3142 compiler treats it as an empty string (which was what was wanted) but it is
3143 reported that Visual C gives an error. The source has been hacked around to
3144 avoid this problem.
3145
3146 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3147 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3148 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3149 of them did).
3150
3151 10. Originally, pcretest opened its input and output without "b"; then I was
3152 told that "b" was needed in some environments, so it was added for release
3153 5.0 to both the input and output. (It makes no difference on Unix-like
3154 systems.) Later I was told that it is wrong for the input on Windows. I've
3155 now abstracted the modes into two macros, to make it easier to fiddle with
3156 them, and removed "b" from the input mode under Windows.
3157
3158 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3159
3160 12. Added -help and --help to pcretest as an official way of being reminded
3161 of the options.
3162
3163 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3164 and pcrecpp.cc because they annoy compilers at high warning levels.
3165
3166 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3167
3168 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3169 scripts are not C programs :-) and which was not noticed because it works
3170 on Linux.
3171
3172 16. pcretest is supposed to handle any length of pattern and data line (as one
3173 line or as a continued sequence of lines) by extending its input buffer if
3174 necessary. This feature was broken for very long pattern lines, leading to
3175 a string of junk being passed to pcre_compile() if the pattern was longer
3176 than about 50K.
3177
3178 17. I have done a major re-factoring of the way pcre_compile() computes the
3179 amount of memory needed for a compiled pattern. Previously, there was code
3180 that made a preliminary scan of the pattern in order to do this. That was
3181 OK when PCRE was new, but as the facilities have expanded, it has become
3182 harder and harder to keep it in step with the real compile phase, and there
3183 have been a number of bugs (see for example, 4 above). I have now found a
3184 cunning way of running the real compile function in a "fake" mode that
3185 enables it to compute how much memory it would need, while actually only
3186 ever using a few hundred bytes of working memory and without too many
3187 tests of the mode. This should make future maintenance and development
3188 easier. A side effect of this work is that the limit of 200 on the nesting
3189 depth of parentheses has been removed (though this was never a serious
3190 limitation, I suspect). However, there is a downside: pcre_compile() now
3191 runs more slowly than before (30% or more, depending on the pattern). I
3192 hope this isn't a big issue. There is no effect on runtime performance.
3193
3194 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3195 newline (only possible for the last line of a file) and it was a
3196 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3197
3198 19. Added additional timing features to pcretest. (1) The -tm option now times
3199 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3200 separate command line item, by a number that specifies the number of
3201 repeats to use when timing. The default is 50000; this gives better
3202 precision, but takes uncomfortably long for very large patterns.
3203
3204 20. Extended pcre_study() to be more clever in cases where a branch of a
3205 subpattern has no definite first character. For example, (a*|b*)[cd] would
3206 previously give no result from pcre_study(). Now it recognizes that the
3207 first character must be a, b, c, or d.
3208
3209 21. There was an incorrect error "recursive call could loop indefinitely" if
3210 a subpattern (or the entire pattern) that was being tested for matching an
3211 empty string contained only one non-empty item after a nested subpattern.
3212 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3213 incorrectly, because the \d was being skipped in the check.
3214
3215 22. The pcretest program now has a new pattern option /B and a command line
3216 option -b, which is equivalent to adding /B to every pattern. This causes
3217 it to show the compiled bytecode, without the additional information that
3218 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3219 is the same as /B/I).
3220
3221 23. A new optimization is now able automatically to treat some sequences such
3222 as a*b as a*+b. More specifically, if something simple (such as a character
3223 or a simple class like \d) has an unlimited quantifier, and is followed by
3224 something that cannot possibly match the quantified thing, the quantifier
3225 is automatically "possessified".
3226
3227 24. A recursive reference to a subpattern whose number was greater than 39
3228 went wrong under certain circumstances in UTF-8 mode. This bug could also
3229 have affected the operation of pcre_study().
3230
3231 25. Realized that a little bit of performance could be had by replacing
3232 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3233
3234 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3235
3236 27. Possessive quantifiers such as a++ were previously implemented by turning
3237 them into atomic groups such as ($>a+). Now they have their own opcodes,
3238 which improves performance. This includes the automatically created ones
3239 from 23 above.
3240
3241 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3242 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3243 the first matched character to be a colon. This applied both to named and
3244 numbered groups.
3245
3246 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3247
3248 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3249 building PCRE on Windows easier, so I have included it in the distribution.
3250
3251 31. There is now a check in pcretest against a ridiculously large number being
3252 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3253 loop, the loop is abandoned.
3254
3255 32. Forward references to subpatterns in conditions such as (?(2)...) where
3256 subpattern 2 is defined later cause pcre_compile() to search forwards in
3257 the pattern for the relevant set of parentheses. This search went wrong
3258 when there were unescaped parentheses in a character class, parentheses
3259 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3260
3261 33. "Subroutine" calls and backreferences were previously restricted to
3262 referencing subpatterns earlier in the regex. This restriction has now
3263 been removed.
3264
3265 34. Added a number of extra features that are going to be in Perl 5.10. On the
3266 whole, these are just syntactic alternatives for features that PCRE had
3267 previously implemented using the Python syntax or my own invention. The
3268 other formats are all retained for compatibility.
3269
3270 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3271 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3272 also .NET compatible.
3273
3274 (b) A recursion or subroutine call to a named group can now be defined as
3275 (?&name) as well as (?P>name).
3276
3277 (c) A backreference to a named group can now be defined as \k<name> or
3278 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3279 5.10, are also .NET compatible.
3280
3281 (d) A conditional reference to a named group can now use the syntax
3282 (?(<name>) or (?('name') as well as (?(name).
3283
3284 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3285 groups (named and numbered) that are never evaluated inline, but can be
3286 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3287 is always false. There may be only one alternative in such a group.
3288
3289 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3290 as the simple (?(R). The condition is true only if the most recent
3291 recursion is that of the given number or name. It does not search out
3292 through the entire recursion stack.
3293
3294 (g) The escape \gN or \g{N} has been added, where N is a positive or
3295 negative number, specifying an absolute or relative reference.
3296
3297 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3298 some "unreachable code" warnings.
3299
3300 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3301 things, this adds five new scripts.
3302
3303 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3304 There were also incompatibilities regarding the handling of \Q..\E inside
3305 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3306 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3307
3308 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3309 matches an empty string, and forcibly breaks the loop. There were bugs in
3310 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3311 against aaaa the result was just "a" rather than "aaaa", for example. Two
3312 separate and independent bugs (that affected different cases) have been
3313 fixed.
3314
3315 39. Refactored the code to abolish the use of different opcodes for small
3316 capturing bracket numbers. This is a tidy that I avoided doing when I
3317 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3318 The new approach is not only tidier, it makes it possible to reduce the
3319 memory needed to fix the previous bug (38).
3320
3321 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3322 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3323 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3324 mode.
3325
3326 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3327 report.
3328
3329 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3330 copy construction and assignment in the C++ wrapper.
3331
3332 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3333 couple of bugs that could have given wrong results in the "--newline=crlf"
3334 case.
3335
3336 44. Added a number of casts and did some reorganization of signed/unsigned int
3337 variables following suggestions from Dair Grant. Also renamed the variable
3338 "this" as "item" because it is a C++ keyword.
3339
3340 45. Arranged for dftables to add
3341
3342 #include "pcre_internal.h"
3343
3344 to pcre_chartables.c because without it, gcc 4.x may remove the array
3345 definition from the final binary if PCRE is built into a static library and
3346 dead code stripping is activated.
3347
3348 46. For an unanchored pattern, if a match attempt fails at the start of a
3349 newline sequence, and the newline setting is CRLF or ANY, and the next two
3350 characters are CRLF, advance by two characters instead of one.
3351
3352
3353 Version 6.7 04-Jul-06
3354 ---------------------
3355
3356 1. In order to handle tests when input lines are enormously long, pcretest has
3357 been re-factored so that it automatically extends its buffers when
3358 necessary. The code is crude, but this _is_ just a test program. The
3359 default size has been increased from 32K to 50K.
3360
3361 2. The code in pcre_study() was using the value of the re argument before
3362 testing it for NULL. (Of course, in any sensible call of the function, it
3363 won't be NULL.)
3364
3365 3. The memmove() emulation function in pcre_internal.h, which is used on
3366 systems that lack both memmove() and bcopy() - that is, hardly ever -
3367 was missing a "static" storage class specifier.
3368
3369 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3370 containing an extended class (one that cannot be represented by a bitmap
3371 because it contains high-valued characters or Unicode property items, e.g.
3372 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3373 pattern, but PCRE should not loop if you do not (it no longer does).
3374 [Detail: two cases were found: (a) a repeated subpattern containing an
3375 extended class; (b) a recursive reference to a subpattern that followed a
3376 previous extended class. It wasn't skipping over the extended class
3377 correctly when UTF-8 mode was not set.]
3378
3379 5. A negated single-character class was not being recognized as fixed-length
3380 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3381 compile error "lookbehind assertion is not fixed length".
3382
3383 6. The RunPerlTest auxiliary script was showing an unexpected difference
3384 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3385 write a Perl script that can interpret lines of an input file either as
3386 byte characters or as UTF-8, which is what "perltest" was being required to
3387 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3388 can't do is switch easily at run time between having the "use utf8;" pragma
3389 or not. In the end, I fudged it by using the RunPerlTest script to insert
3390 "use utf8;" explicitly for the UTF-8 tests.
3391
3392 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3393 the end of the subject string, contrary to the documentation and to what
3394 Perl does. This was true of both matching functions. Now it matches only at
3395 the start of the subject and immediately after *internal* newlines.
3396
3397 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3398 a pointer to an int instead of a pointer to an unsigned long int. This
3399 caused problems on 64-bit systems.
3400
3401 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3402 instance of the 'standard' template library not being so standard".
3403
3404 10. There was no check on the number of named subpatterns nor the maximum
3405 length of a subpattern name. The product of these values is used to compute
3406 the size of the memory block for a compiled pattern. By supplying a very
3407 long subpattern name and a large number of named subpatterns, the size
3408 computation could be caused to overflow. This is now prevented by limiting
3409 the length of names to 32 characters, and the number of named subpatterns
3410 to 10,000.
3411
3412 11. Subpatterns that are repeated with specific counts have to be replicated in
3413 the compiled pattern. The size of memory for this was computed from the
3414 length of the subpattern and the repeat count. The latter is limited to
3415 65535, but there was no limit on the former, meaning that integer overflow
3416 could in principle occur. The compiled length of a repeated subpattern is
3417 now limited to 30,000 bytes in order to prevent this.
3418
3419 12. Added the optional facility to have named substrings with the same name.
3420
3421 13. Added the ability to use a named substring as a condition, using the
3422 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3423 are numbers (not recommended). Forward references are permitted.
3424
3425 14. Added forward references in named backreferences (if you see what I mean).
3426
3427 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3428 pattern could run off the end of the subject. For example, the pattern
3429 "(?s)(.{1,5})"8 did this with the subject "ab".
3430
3431 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3432 PCRE_CASELESS was set when matching characters that were quantified with ?
3433 or *.
3434
3435 17. A character class other than a single negated character that had a minimum
3436 but no maximum quantifier - for example [ab]{6,} - was not handled
3437 correctly by pce_dfa_exec(). It would match only one character.
3438
3439 18. A valid (though odd) pattern that looked like a POSIX character
3440 class but used an invalid character after [ (for example [[,abc,]]) caused
3441 pcre_compile() to give the error "Failed: internal error: code overflow" or
3442 in some cases to crash with a glibc free() error. This could even happen if
3443 the pattern terminated after [[ but there just happened to be a sequence of
3444 letters, a binary zero, and a closing ] in the memory that followed.
3445
3446 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3447 over the years. Originally (before any Unicode support), just the bottom 8
3448 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3449 output from "man perlunicode" includes this:
3450
3451 The regular expression compiler produces polymorphic opcodes. That
3452 is, the pattern adapts to the data and automatically switches to
3453 the Unicode character scheme when presented with Unicode data--or
3454 instead uses a traditional byte scheme when presented with byte
3455 data.
3456
3457 Sadly, a wide octal escape does not cause a switch, and in a string with
3458 no other multibyte characters, these octal escapes are treated as before.
3459 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3460 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3461 Unicode string.
3462
3463 I have not perpetrated such confusion in PCRE. Up till now, it took just
3464 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3465 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3466 translate to the appropriate multibyte character.
3467
3468 29. Applied some refactoring to reduce the number of warnings from Microsoft
3469 and Borland compilers. This has included removing the fudge introduced
3470 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3471 a warning about an unused variable.
3472
3473 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3474 characters since release 4.0, because Perl (from release 5.004) does not.
3475 [Or at least, is documented not to: some releases seem to be in conflict
3476 with the documentation.] However, when a pattern was studied with
3477 pcre_study() and all its branches started with \s, PCRE still included VT
3478 as a possible starting character. Of course, this did no harm; it just
3479 caused an unnecessary match attempt.
3480
3481 22. Removed a now-redundant internal flag bit that recorded the fact that case
3482 dependency changed within the pattern. This was once needed for "required
3483 byte" processing, but is no longer used. This recovers a now-scarce options
3484 bit. Also moved the least significant internal flag bit to the most-
3485 significant bit of the word, which was not previously used (hangover from
3486 the days when it was an int rather than a uint) to free up another bit for
3487 the future.
3488
3489 23. Added support for CRLF line endings as well as CR and LF. As well as the
3490 default being selectable at build time, it can now be changed at runtime
3491 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3492 specify that it is scanning data with non-default line endings.
3493
3494 24. Changed the definition of CXXLINK to make it agree with the definition of
3495 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3496
3497 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3498 recursions. This makes a big different to stack usage for some patterns.
3499
3500 26. If a subpattern containing a named recursion or subroutine reference such
3501 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3502 the space required for the compiled pattern went wrong and gave too small a
3503 value. Depending on the environment, this could lead to "Failed: internal
3504 error: code overflow at offset 49" or "glibc detected double free or
3505 corruption" errors.
3506
3507 27. Applied patches from Google (a) to support the new newline modes and (b) to
3508 advance over multibyte UTF-8 characters in GlobalReplace.
3509
3510 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3511 difference for some implementation of PCRE in some Windows version.
3512
3513 29. Added some extra testing facilities to pcretest:
3514
3515 \q<number> in a data line sets the "match limit" value
3516 \Q<number> in a data line sets the "match recursion limt" value
3517 -S <number> sets the stack size, where <number> is in megabytes
3518
3519 The -S option isn't available for Windows.
3520
3521
3522 Version 6.6 06-Feb-06
3523 ---------------------
3524
3525 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3526 in pcreposix.h. I have copied the definition from pcre.h.
3527
3528 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3529 because pcre.h is no longer a built file.
3530
3531 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3532 not normally included in the compiled code.
3533
3534
3535 Version 6.5 01-Feb-06
3536 ---------------------
3537
3538 1. When using the partial match feature with pcre_dfa_exec(), it was not
3539 anchoring the second and subsequent partial matches at the new starting
3540 point. This could lead to incorrect results. For example, with the pattern
3541 /1234/, partially matching against "123" and then "a4" gave a match.
3542
3543 2. Changes to pcregrep:
3544
3545 (a) All non-match returns from pcre_exec() were being treated as failures
3546 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3547 error message is output. Some extra information is given for the
3548 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3549 probably the only errors that are likely to be caused by users (by
3550 specifying a regex that has nested indefinite repeats, for instance).
3551 If there are more than 20 of these errors, pcregrep is abandoned.
3552
3553 (b) A binary zero was treated as data while matching, but terminated the
3554 output line if it was written out. This has been fixed: binary zeroes
3555 are now no different to any other data bytes.
3556
3557 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3558 used to set a locale for matching. The --locale=xxxx long option has
3559 been added (no short equivalent) to specify a locale explicitly on the
3560 pcregrep command, overriding the environment variables.
3561
3562 (d) When -B was used with -n, some line numbers in the output were one less
3563 than they should have been.
3564
3565 (e) Added the -o (--only-matching) option.
3566
3567 (f) If -A or -C was used with -c (count only), some lines of context were
3568 accidentally printed for the final match.
3569
3570 (g) Added the -H (--with-filename) option.
3571
3572 (h) The combination of options -rh failed to suppress file names for files
3573 that were found from directory arguments.
3574
3575 (i) Added the -D (--devices) and -d (--directories) options.
3576
3577 (j) Added the -F (--fixed-strings) option.
3578
3579 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3580
3581 (l) Added the --colo(u)r option.
3582
3583 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3584 is not present by default.
3585
3586 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3587 items such as (?R) or (?1), when the recursion could match a number of
3588 alternatives. If it matched one of the alternatives, but subsequently,
3589 outside the recursion, there was a failure, the code tried to back up into
3590 the recursion. However, because of the way PCRE is implemented, this is not
3591 possible, and the result was an incorrect result from the match.
3592
3593 In order to prevent this happening, the specification of recursion has
3594 been changed so that all such subpatterns are automatically treated as
3595 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3596
3597 4. I had overlooked the fact that, in some locales, there are characters for
3598 which isalpha() is true but neither isupper() nor islower() are true. In
3599 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3600 and ordfeminine) are like this. This affected the treatment of \w and \W
3601 when they appeared in character classes, but not when they appeared outside
3602 a character class. The bit map for "word" characters is now created
3603 separately from the results of isalnum() instead of just taking it from the
3604 upper, lower, and digit maps. (Plus the underscore character, of course.)
3605
3606 5. The above bug also affected the handling of POSIX character classes such as
3607 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3608 permanent tables. Instead, the bit maps for such a class were previously
3609 created as the appropriate unions of the upper, lower, and digit bitmaps.
3610 Now they are created by subtraction from the [[:word:]] class, which has
3611 its own bitmap.
3612
3613 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3614 It is created by subtracting the vertical space characters (\x09, \x0a,
3615 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3616 subtraction was done in the overall bitmap for a character class, meaning
3617 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3618 be recognized. This bug has been fixed.
3619
3620 7. Patches from the folks at Google:
3621
3622 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3623 real life, but is still worth protecting against".
3624
3625 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3626 regular expressions".
3627
3628 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3629 have it.
3630
3631 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3632 "configure" and the latter not, in order to fix a problem somebody had
3633 with compiling the Arg class on HP-UX.
3634
3635 (e) Improve the error-handling of the C++ wrapper a little bit.
3636
3637 (f) New tests for checking recursion limiting.
3638
3639 8. The pcre_memmove() function, which is used only if the environment does not
3640 have a standard memmove() function (and is therefore rarely compiled),
3641 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3642 returning a result (though PCRE never actually uses the result).
3643
3644 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3645 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3646 returned instead of calling malloc() with an overflowing number that would
3647 most likely cause subsequent chaos.
3648
3649 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3650
3651 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3652 with this option is matched, the nmatch and pmatch options of regexec() are
3653 ignored.
3654
3655 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3656 provided in case anyone wants to the the POSIX interface with UTF-8
3657 strings.
3658
3659 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3660 C++ linking (needed for some HP-UX environments).
3661
3662 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3663 (unused parameter) and in the pcre_printint() function (omitted "default"
3664 switch label when the default is to do nothing).
3665
3666 15. Added some code to make it possible, when PCRE is compiled as a C++
3667 library, to replace subject pointers for pcre_exec() with a smart pointer
3668 class, thus making it possible to process discontinuous strings.
3669
3670 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3671 much the same function. They were added by different people who were trying
3672 to make PCRE easy to compile on non-Unix systems. It has been suggested
3673 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3674 for compiling on Windows systems. I have therefore replaced it with
3675 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3676 defaults to "extern" for C or "extern C" for C++, which works fine on
3677 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3678 SCOPE with something explicit in config.h. In addition:
3679
3680 (a) pcreposix.h still had just "extern" instead of either of these macros;
3681 I have replaced it with PCRE_DATA_SCOPE.
3682
3683 (b) Functions such as _pcre_xclass(), which are internal to the library,
3684 but external in the C sense, all had PCRE_EXPORT in their definitions.
3685 This is apparently wrong for the Windows case, so I have removed it.
3686 (It makes no difference on Unix-like systems.)
3687
3688 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3689 of recursive calls to match(). This is different to MATCH_LIMIT because
3690 that limits the total number of calls to match(), not all of which increase
3691 the depth of recursion. Limiting the recursion depth limits the amount of
3692 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3693 when PCRE is compiled, and changed at run time. A patch from Google adds
3694 this functionality to the C++ interface.
3695
3696 18. Changes to the handling of Unicode character properties:
3697
3698 (a) Updated the table to Unicode 4.1.0.
3699
3700 (b) Recognize characters that are not in the table as "Cn" (undefined).
3701
3702 (c) I revised the way the table is implemented to a much improved format
3703 which includes recognition of ranges. It now supports the ranges that
3704 are defined in UnicodeData.txt, and it also amalgamates other
3705 characters into ranges. This has reduced the number of entries in the
3706 table from around 16,000 to around 3,000, thus reducing its size
3707 considerably. I realized I did not need to use a tree structure after
3708 all - a binary chop search is just as efficient. Having reduced the
3709 number of entries, I extended their size from 6 bytes to 8 bytes to
3710 allow for more data.
3711
3712 (d) Added support for Unicode script names via properties such as \p{Han}.
3713
3714 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3715 matching that character.
3716
3717 20. When matching a repeated Unicode property with a minimum greater than zero,
3718 (for example \pL{2,}), PCRE could look past the end of the subject if it
3719 reached it while seeking the minimum number of characters. This could
3720 happen only if some of the characters were more than one byte long, because
3721 there is a check for at least the minimum number of bytes.
3722
3723 21. Refactored the implementation of \p and \P so as to be more general, to
3724 allow for more different types of property in future. This has changed the
3725 compiled form incompatibly. Anybody with saved compiled patterns that use
3726 \p or \P will have to recompile them.
3727
3728 22. Added "Any" and "L&" to the supported property types.
3729
3730 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3731 but give a compile time error if the value is greater than 0xff.
3732
3733 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3734 accidentally not being installed or uninstalled.
3735
3736 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3737 made were to insert the current release number. This seemed silly, because
3738 it made things harder for people building PCRE on systems that don't run
3739 "configure". I have turned pcre.h into a distributed file, no longer built
3740 by "configure", with the version identification directly included. There is
3741 no longer a pcre.h.in file.
3742
3743 However, this change necessitated a change to the pcre-config script as
3744 well. It is built from pcre-config.in, and one of the substitutions was the
3745 release number. I have updated configure.ac so that ./configure now finds
3746 the release number by grepping pcre.h.
3747
3748 26. Added the ability to run the tests under valgrind.
3749
3750
3751 Version 6.4 05-Sep-05
3752 ---------------------
3753
3754 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3755 "--" to be printed when multiple files were scanned, even when none of the
3756 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3757 consider it to be a bug, and have restored the previous behaviour.
3758
3759 2. A couple of code tidies to get rid of compiler warnings.
3760
3761 3. The pcretest program used to cheat by referring to symbols in the library
3762 whose names begin with _pcre_. These are internal symbols that are not
3763 really supposed to be visible externally, and in some environments it is
3764 possible to suppress them. The cheating is now confined to including
3765 certain files from the library's source, which is a bit cleaner.
3766
3767 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3768 file's purpose clearer.
3769
3770 5. Reorganized pcre_ucp_findchar().
3771
3772
3773 Version 6.3 15-Aug-05
3774 ---------------------
3775
3776 1. The file libpcre.pc.in did not have general read permission in the tarball.
3777
3778 2. There were some problems when building without C++ support:
3779
3780 (a) If C++ support was not built, "make install" and "make test" still
3781 tried to test it.
3782
3783 (b) There were problems when the value of CXX was explicitly set. Some
3784 changes have been made to try to fix these, and ...
3785
3786 (c) --disable-cpp can now be used to explicitly disable C++ support.
3787
3788 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3789 backslash in a target when C++ was disabled. This confuses some
3790 versions of "make", apparently. Using an intermediate variable solves
3791 this. (Same for CPP_LOBJ.)
3792
3793 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3794 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3795 necessary on certain architectures.
3796
3797 4. Added a setting of -export-symbols-regex to the link command to remove
3798 those symbols that are exported in the C sense, but actually are local
3799 within the library, and not documented. Their names all begin with
3800 "_pcre_". This is not a perfect job, because (a) we have to except some
3801 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3802 available (and never for static libraries). I have made a note to try to
3803 find a way round (a) in the future.
3804
3805
3806 Version 6.2 01-Aug-05
3807 ---------------------
3808
3809 1. There was no test for integer overflow of quantifier values. A construction
3810 such as {1111111111111111} would give undefined results. What is worse, if
3811 a minimum quantifier for a parenthesized subpattern overflowed and became
3812 negative, the calculation of the memory size went wrong. This could have
3813 led to memory overwriting.
3814
3815 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3816
3817 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3818 operating environments where this matters.
3819
3820 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3821 PCRE options from within the C++ wrapper.
3822
3823 5. Named capturing subpatterns were not being correctly counted when a pattern
3824 was compiled. This caused two problems: (a) If there were more than 100
3825 such subpatterns, the calculation of the memory needed for the whole
3826 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3827 back references of the form \12, where the number was greater than 9, were
3828 not recognized as back references, even though there were sufficient
3829 previous subpatterns.
3830
3831 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3832 versions of gcc, e.g. 2.95.4.
3833
3834
3835 Version 6.1 21-Jun-05
3836 ---------------------
3837
3838 1. There was one reference to the variable "posix" in pcretest.c that was not
3839 surrounded by "#if !defined NOPOSIX".
3840
3841 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3842 the cross-check on the old pcre_info() function, for the benefit of the
3843 cut-down version of PCRE that is currently imported into Exim.
3844
3845 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3846 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3847 patterns that start (?i) but I don't think that matters. The use of (?i) is
3848 just an example; this all applies to the other options as well.
3849
3850 4. Since libtool seems to echo the compile commands it is issuing, the output
3851 from "make" can be reduced a bit by putting "@" in front of each libtool
3852 compile command.
3853
3854 5. Patch from the folks at Google for configure.in to be a bit more thorough
3855 in checking for a suitable C++ installation before trying to compile the
3856 C++ stuff. This should fix a reported problem when a compiler was present,
3857 but no suitable headers.
3858
3859 6. The man pages all had just "PCRE" as their title. I have changed them to
3860 be the relevant file name. I have also arranged that these names are
3861 retained in the file doc/pcre.txt, which is a concatenation in text format
3862 of all the man pages except the little individual ones for each function.
3863
3864 7. The NON-UNIX-USE file had not been updated for the different set of source
3865 files that come with release 6. I also added a few comments about the C++
3866 wrapper.
3867
3868
3869 Version 6.0 07-Jun-05
3870 ---------------------
3871
3872 1. Some minor internal re-organization to help with my DFA experiments.
3873
3874 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3875 didn't matter for the library itself when fully configured, but did matter
3876 when compiling without UCP support, or within Exim, where the ucp files are
3877 not imported.
3878
3879 3. Refactoring of the library code to split up the various functions into
3880 different source modules. The addition of the new DFA matching code (see
3881 below) to a single monolithic source would have made it really too
3882 unwieldy, quite apart from causing all the code to be include in a
3883 statically linked application, when only some functions are used. This is
3884 relevant even without the DFA addition now that patterns can be compiled in
3885 one application and matched in another.
3886
3887 The downside of splitting up is that there have to be some external
3888 functions and data tables that are used internally in different modules of
3889 the library but which are not part of the API. These have all had their
3890 names changed to start with "_pcre_" so that they are unlikely to clash
3891 with other external names.
3892
3893 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3894 a different (DFA) algorithm. Although it is slower than the original
3895 function, it does have some advantages for certain types of matching
3896 problem.
3897
3898 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3899 including restarting after a partial match.
3900
3901 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3902 defined when compiling for Windows was sent to me. I have put it into the
3903 code, though I have no means of testing or verifying it.
3904
3905 7. Added the pcre_refcount() auxiliary function.
3906
3907 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3908 match before or at the first newline in the subject string. In pcretest,
3909 the /f option on a pattern can be used to set this.
3910
3911 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3912 would behave wrongly. This has been present in PCRE since release 4.0.
3913
3914 10. A number of changes to the pcregrep command:
3915
3916 (a) Refactored how -x works; insert ^(...)$ instead of setting
3917 PCRE_ANCHORED and checking the length, in preparation for adding
3918 something similar for -w.
3919
3920 (b) Added the -w (match as a word) option.
3921
3922 (c) Refactored the way lines are read and buffered so as to have more
3923 than one at a time available.
3924
3925 (d) Implemented a pcregrep test script.
3926
3927 (e) Added the -M (multiline match) option. This allows patterns to match
3928 over several lines of the subject. The buffering ensures that at least
3929 8K, or the rest of the document (whichever is the shorter) is available
3930 for matching (and similarly the previous 8K for lookbehind assertions).
3931
3932 (f) Changed the --help output so that it now says
3933
3934 -w, --word-regex(p)
3935
3936 instead of two lines, one with "regex" and the other with "regexp"
3937 because that confused at least one person since the short forms are the
3938 same. (This required a bit of code, as the output is generated
3939 automatically from a table. It wasn't just a text change.)
3940
3941 (g) -- can be used to terminate pcregrep options if the next thing isn't an
3942 option but starts with a hyphen. Could be a pattern or a path name
3943 starting with a hyphen, for instance.
3944
3945 (h) "-" can be given as a file name to represent stdin.
3946
3947 (i) When file names are being printed, "(standard input)" is used for
3948 the standard input, for compatibility with GNU grep. Previously
3949 "<stdin>" was used.
3950
3951 (j) The option --label=xxx can be used to supply a name to be used for
3952 stdin when file names are being printed. There is no short form.
3953
3954 (k) Re-factored the options decoding logic because we are going to add
3955 two more options that take data. Such options can now be given in four
3956 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3957
3958 (l) Added the -A, -B, and -C options for requesting that lines of context
3959 around matches be printed.
3960
3961 (m) Added the -L option to print the names of files that do not contain
3962 any matching lines, that is, the complement of -l.
3963
3964 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3965 continue to scan other files.
3966
3967 (o) The -s option was incorrectly implemented. For compatibility with other
3968 greps, it now suppresses the error message for a non-existent or non-
3969 accessible file (but not the return code). There is a new option called
3970 -q that suppresses the output of matching lines, which was what -s was
3971 previously doing.
3972
3973 (p) Added --include and --exclude options to specify files for inclusion
3974 and exclusion when recursing.
3975
3976 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3977 Hopefully, it now does.
3978
3979 12. Missing cast in pcre_study().
3980
3981 13. Added an "uninstall" target to the makefile.
3982
3983 14. Replaced "extern" in the function prototypes in Makefile.in with
3984 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3985 world, but is set differently for Windows.
3986
3987 15. Added a second compiling function called pcre_compile2(). The only
3988 difference is that it has an extra argument, which is a pointer to an
3989 integer error code. When there is a compile-time failure, this is set
3990 non-zero, in addition to the error test pointer being set to point to an
3991 error message. The new argument may be NULL if no error number is required
3992 (but then you may as well call pcre_compile(), which is now just a
3993 wrapper). This facility is provided because some applications need a
3994 numeric error indication, but it has also enabled me to tidy up the way
3995 compile-time errors are handled in the POSIX wrapper.
3996
3997 16. Added VPATH=.libs to the makefile; this should help when building with one
3998 prefix path and installing with another. (Or so I'm told by someone who
3999 knows more about this stuff than I do.)
4000
4001 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4002 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4003 match everything, including newlines. This is not POSIX-compatible, but
4004 somebody wanted the feature. From pcretest it can be activated by using
4005 both the P and the s flags.
4006
4007 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4008
4009 19. libpcre.pc was being incorrectly installed as executable.
4010
4011 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4012 it now also looks for '\r' so that it will work unmodified on Windows.
4013
4014 21. Added Google's contributed C++ wrapper to the distribution.
4015
4016 22. Added some untidy missing memory free() calls in pcretest, to keep
4017 Electric Fence happy when testing.
4018
4019
4020
4021 Version 5.0 13-Sep-04
4022 ---------------------
4023
4024 1. Internal change: literal characters are no longer packed up into items
4025 containing multiple characters in a single byte-string. Each character
4026 is now matched using a separate opcode. However, there may be more than one
4027 byte in the character in UTF-8 mode.
4028
4029 2. The pcre_callout_block structure has two new fields: pattern_position and
4030 next_item_length. These contain the offset in the pattern to the next match
4031 item, and its length, respectively.
4032
4033 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4034 insertion of callouts before each pattern item. Added the /C option to
4035 pcretest to make use of this.
4036
4037 4. On the advice of a Windows user, the lines
4038
4039 #if defined(_WIN32) || defined(WIN32)
4040 _setmode( _fileno( stdout ), 0x8000 );
4041 #endif /* defined(_WIN32) || defined(WIN32) */
4042
4043 have been added to the source of pcretest. This apparently does useful
4044 magic in relation to line terminators.
4045
4046 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4047 for the benefit of those environments where the "b" makes a difference.
4048
4049 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4050 to know about it. I have put a hack into configure.in that adds in code
4051 to set GCC=yes if CC=icc. This seems to end up at a point in the
4052 generated configure script that is early enough to affect the setting of
4053 compiler options, which is what is needed, but I have no means of testing
4054 whether it really works. (The user who reported this had patched the
4055 generated configure script, which of course I cannot do.)
4056
4057 LATER: After change 22 below (new libtool files), the configure script
4058 seems to know about icc (and also ecc). Therefore, I have commented out
4059 this hack in configure.in.
4060
4061 7. Added support for pkg-config (2 patches were sent in).
4062
4063 8. Negated POSIX character classes that used a combination of internal tables
4064 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4065 [[:^ascii]]. Typically, they would match almost any characters. The other
4066 POSIX classes were not broken in this way.
4067
4068 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4069 to find the match, as PCRE was deluded into thinking that the match had to
4070 start at the start point or following a newline. The same bug applied to
4071 patterns with negative forward assertions or any backward assertions
4072 preceding ".*" at the start, unless the pattern required a fixed first
4073 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4074
4075 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4076 starting at the last subject character, bytes beyond the end of the subject
4077 string were read.
4078
4079 11. Renamed the variable "class" as "classbits" to make life easier for C++
4080 users. (Previously there was a macro definition, but it apparently wasn't
4081 enough.)
4082
4083 12. Added the new field "tables" to the extra data so that tables can be passed
4084 in at exec time, or the internal tables can be re-selected. This allows
4085 a compiled regex to be saved and re-used at a later time by a different
4086 program that might have everything at different addresses.
4087
4088 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4089 -R library as well as a -L library.
4090
4091 14. The debugging options of pcretest (-d on the command line or D on a
4092 pattern) showed incorrect output for anything following an extended class
4093 that contained multibyte characters and which was followed by a quantifier.
4094
4095 15. Added optional support for general category Unicode character properties
4096 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4097 support. It adds about 90K to the size of the library. The meanings of the
4098 inbuilt class escapes such as \d and \s have NOT been changed.
4099
4100 16. Updated pcredemo.c to include calls to free() to release the memory for the
4101 compiled pattern.
4102
4103 17. The generated file chartables.c was being created in the source directory
4104 instead of in the building directory. This caused the build to fail if the
4105 source directory was different from the building directory, and was
4106 read-only.
4107
4108 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4109 file. No doubt somebody will tell me if they don't make sense... Also added
4110 Dan Mooney's comments about building on OpenVMS.
4111
4112 19. Added support for partial matching via the PCRE_PARTIAL option for
4113 pcre_exec() and the \P data escape in pcretest.
4114
4115 20. Extended pcretest with 3 new pattern features:
4116
4117 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4118 write the compiled pattern to the file whose name is "rest-of-line".
4119 This is a straight binary dump of the data, with the saved pointer to
4120 the character tables forced to be NULL. The study data, if any, is
4121 written too. After writing, pcretest reads a new pattern.
4122
4123 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4124 compiled pattern from the given file. There must not be any
4125 occurrences of "<" in the file name (pretty unlikely); if there are,
4126 pcretest will instead treat the initial "<" as a pattern delimiter.
4127 After reading in the pattern, pcretest goes on to read data lines as
4128 usual.
4129
4130 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4131 and 16-bit fields in a compiled pattern, to simulate a pattern that
4132 was compiled on a host of opposite endianness.
4133
4134 21. The pcre-exec() function can now cope with patterns that were compiled on
4135 hosts of opposite endianness, with this restriction:
4136
4137 As for any compiled expression that is saved and used later, the tables
4138 pointer field cannot be preserved; the extra_data field in the arguments
4139 to pcre_exec() should be used to pass in a tables address if a value
4140 other than the default internal tables were used at compile time.
4141
4142 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4143 now diagnosed as an error. Previously, most of the time, a negative number
4144 would have been treated as zero, but if in addition "ovector" was passed as
4145 NULL, a crash could occur.
4146
4147 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4148 new versions from the libtool 1.5 distribution (the last one is a copy of
4149 a file called libtool.m4). This seems to have fixed the need to patch
4150 "configure" to support Darwin 1.3 (which I used to do). However, I still
4151 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4152 workstation).
4153
4154 24. Changed the PCRE licence to be the more standard "BSD" licence.
4155
4156
4157 Version 4.5 01-Dec-03
4158 ---------------------
4159
4160 1. There has been some re-arrangement of the code for the match() function so
4161 that it can be compiled in a version that does not call itself recursively.
4162 Instead, it keeps those local variables that need separate instances for
4163 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4164 needs to "recurse". Keeping track of where control must go is done by means
4165 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4166 hide most of the details from the main code, and operates only if
4167 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4168 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4169 operating.
4170
4171 To make it easier for callers to provide specially tailored get/free
4172 functions for this usage, two new functions, pcre_stack_malloc, and
4173 pcre_stack_free, are used. They are always called in strict stacking order,
4174 and the size of block requested is always the same.
4175
4176 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4177 PCRE has been compiled to use the stack or the heap for recursion. The
4178 -C option of pcretest uses this to show which version is compiled.
4179
4180 A new data escape \S, is added to pcretest; it causes the amounts of store
4181 obtained and freed by both kinds of malloc/free at match time to be added
4182 to the output.
4183
4184 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4185 what's available on my current Linux desktop machine.
4186
4187 3. When matching a UTF-8 string, the test for a valid string at the start has
4188 been extended. If start_offset is not zero, PCRE now checks that it points
4189 to a byte that is the start of a UTF-8 character. If not, it returns
4190 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4191 this is necessary because there may be backward assertions in the pattern.
4192 When matching the same subject several times, it may save resources to use
4193 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4194
4195 4. The code for checking the validity of UTF-8 strings has been tightened so
4196 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4197 containing "overlong sequences".
4198
4199 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4200 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4201 should have been "&", but it just so happened that all the cases this let
4202 through by mistake were picked up later in the function.
4203
4204 6. I had used a variable called "isblank" - this is a C99 function, causing
4205 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4206
4207 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4208 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4209 so the version is shown; (c) stop "make test" echoing "./RunTest".
4210
4211 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4212
4213 9. The prototype for memmove() for systems that don't have it was using
4214 size_t, but the inclusion of the header that defines size_t was later. I've
4215 moved the #includes for the C headers earlier to avoid this.
4216
4217 10. Added some adjustments to the code to make it easier to compiler on certain
4218 special systems:
4219
4220 (a) Some "const" qualifiers were missing.
4221 (b) Added the macro EXPORT before all exported functions; by default this
4222 is defined to be empty.
4223 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4224 that it reads its output file name as an argument instead of writing
4225 to the standard output and assuming this can be redirected.
4226
4227 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4228 class containing characters with values greater than 255, PCRE compilation
4229 went into a loop.
4230
4231 12. A recursive reference to a subpattern that was within another subpattern
4232 that had a minimum quantifier of zero caused PCRE to crash. For example,
4233 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4234 recursion. If the recursively-called subpattern itself had a zero repeat,
4235 that was OK.
4236
4237 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4238 buffer into which it was copied (for escape processing) was still set at
4239 1024, so long lines caused crashes.
4240
4241 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4242 "internal error: code overflow...". This applied to any character class
4243 that was followed by a possessive quantifier.
4244
4245 15. Modified the Makefile to add libpcre.la as a prerequisite for
4246 libpcreposix.la because I was told this is needed for a parallel build to
4247 work.
4248
4249 16. If a pattern that contained .* following optional items at the start was
4250 studied, the wrong optimizing data was generated, leading to matching
4251 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4252 matching string must start with a or b or c. The correct conclusion for
4253 this pattern is that a match can start with any character.
4254
4255
4256 Version 4.4 13-Aug-03
4257 ---------------------
4258
4259 1. In UTF-8 mode, a character class containing characters with values between
4260 127 and 255 was not handled correctly if the compiled pattern was studied.
4261 In fixing this, I have also improved the studying algorithm for such
4262 classes (slightly).
4263
4264 2. Three internal functions had redundant arguments passed to them. Removal
4265 might give a very teeny performance improvement.
4266
4267 3. Documentation bug: the value of the capture_top field in a callout is *one
4268 more than* the number of the hightest numbered captured substring.
4269
4270 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4271 in incorrectly linking with a previously installed version. They now link
4272 explicitly with libpcre.la.
4273
4274 5. configure.in no longer needs to recognize Cygwin specially.
4275
4276 6. A problem in pcre.in for Windows platforms is fixed.
4277
4278 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4279 pcretest, it used to include the size of the study block as part of its
4280 output. Unfortunately, the structure contains a field that has a different
4281 size on different hardware architectures. This meant that the tests that
4282 showed this size failed. As the block is currently always of a fixed size,
4283 this information isn't actually particularly useful in pcretest output, so
4284 I have just removed it.
4285
4286 8. Three pre-processor statements accidentally did not start in column 1.
4287 Sadly, there are *still* compilers around that complain, even though
4288 standard C has not required this for well over a decade. Sigh.
4289
4290 9. In pcretest, the code for checking callouts passed small integers in the
4291 callout_data field, which is a void * field. However, some picky compilers
4292 complained about the casts involved for this on 64-bit systems. Now
4293 pcretest passes the address of the small integer instead, which should get
4294 rid of the warnings.
4295
4296 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4297 both compile and run time, and gives an error if an invalid UTF-8 sequence
4298 is found. There is a option for disabling this check in cases where the
4299 string is known to be correct and/or the maximum performance is wanted.
4300
4301 11. In response to a bug report, I changed one line in Makefile.in from
4302
4303 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4304 to
4305 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4306
4307 to look similar to other lines, but I have no way of telling whether this
4308 is the right thing to do, as I do not use Windows. No doubt I'll get told
4309 if it's wrong...
4310
4311
4312 Version 4.3 21-May-03
4313 ---------------------
4314
4315 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4316 Makefile.
4317
4318 2. Some refactoring to improve the quality of the code:
4319
4320 (i) The utf8_table... variables are now declared "const".
4321
4322 (ii) The code for \cx, which used the "case flipping" table to upper case
4323 lower case letters, now just substracts 32. This is ASCII-specific,
4324 but the whole concept of \cx is ASCII-specific, so it seems
4325 reasonable.
4326
4327 (iii) PCRE was using its character types table to recognize decimal and
4328 hexadecimal digits in the pattern. This is silly, because it handles
4329 only 0-9, a-f, and A-F, but the character types table is locale-
4330 specific, which means strange things might happen. A private
4331 table is now used for this - though it costs 256 bytes, a table is
4332 much faster than multiple explicit tests. Of course, the standard
4333 character types table is still used for matching digits in subject
4334 strings against \d.
4335
4336 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4337 ending in _t are). So I've renamed it as ESC_tee.
4338
4339 3. The first argument for regexec() in the POSIX wrapper should have been
4340 defined as "const".
4341
4342 4. Changed pcretest to use malloc() for its buffers so that they can be
4343 Electric Fenced for debugging.
4344
4345 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4346 to read one or more bytes before the start of the subject string. Often this
4347 had no effect on PCRE's behaviour, but in some circumstances it could
4348 provoke a segmentation fault.
4349
4350 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4351 to try to read one or more bytes before the start of the subject string.
4352
4353 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4354 UTF-8 support could misbehave in various ways if the subject string
4355 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4356 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4357 back over UTF-8 characters.)
4358
4359
4360 Version 4.2 14-Apr-03
4361 ---------------------
4362
4363 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4364
4365 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4366 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4367 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4368 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4369 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4370 and BUILD_EXEEXT
4371 Note: automatic setting of the BUILD variables is not yet working
4372 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4373 compile-time but not at link-time
4374 [LINK]: use for linking executables only
4375 make different versions for Windows and non-Windows
4376 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4377 libraries
4378 [LINK_FOR_BUILD]: new variable
4379 [OBJEXT]: use throughout
4380 [EXEEXT]: use throughout
4381 <winshared>: new target
4382 <wininstall>: new target
4383 <dftables.o>: use native compiler
4384 <dftables>: use native linker
4385 <install>: handle Windows platform correctly
4386 <clean>: ditto
4387 <check>: ditto
4388 copy DLL to top builddir before testing
4389
4390 As part of these changes, -no-undefined was removed again. This was reported
4391 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4392 in any case.
4393
4394 3. Some tidies to get rid of compiler warnings:
4395
4396 . In the match_data structure, match_limit was an unsigned long int, whereas
4397 match_call_count was an int. I've made them both unsigned long ints.
4398
4399 . In pcretest the fact that a const uschar * doesn't automatically cast to
4400 a void * provoked a warning.
4401
4402 . Turning on some more compiler warnings threw up some "shadow" variables
4403 and a few more missing casts.
4404
4405 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4406 option, a class that contained a single character with a value between 128
4407 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4408
4409 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4410 option, a class that contained several characters, but with at least one
4411 whose value was between 128 and 255 caused PCRE to crash.
4412
4413
4414 Version 4.1 12-Mar-03
4415 ---------------------
4416
4417 1. Compiling with gcc -pedantic found a couple of places where casts were
4418 needed, and a string in dftables.c that was longer than standard compilers are
4419 required to support.
4420
4421 2. Compiling with Sun's compiler found a few more places where the code could
4422 be tidied up in order to avoid warnings.
4423
4424 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4425 first of these names is deprecated in the latest Autoconf in favour of the name
4426 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4427 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4428 analogy I have changed it to CFLAGS_FOR_BUILD.
4429
4430 4. Added -no-undefined to the linking command in the Makefile, because this is
4431 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4432 linking step for the pcreposix library.
4433
4434 5. PCRE was failing to diagnose the case of two named groups with the same
4435 name.
4436
4437 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4438 literal character that is needed in the subject for a match, and scans along to
4439 ensure that it is present before embarking on the full matching process. This
4440 saves time in cases of nested unlimited repeats that are never going to match.
4441 Problem: the scan can take a lot of time if the subject is very long (e.g.
4442 megabytes), thus penalizing straightforward matches. It is now done only if the
4443 amount of subject to be scanned is less than 1000 bytes.
4444
4445 7. A lesser problem with the same optimization is that it was recording the
4446 first character of an anchored pattern as "needed", thus provoking a search
4447 right along the subject, even when the first match of the pattern was going to
4448 fail. The "needed" character is now not set for anchored patterns, unless it
4449 follows something in the pattern that is of non-fixed length. Thus, it still
4450 fulfils its original purpose of finding quick non-matches in cases of nested
4451 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4452
4453
4454 Version 4.0 17-Feb-03
4455 ---------------------
4456
4457 1. If a comment in an extended regex that started immediately after a meta-item
4458 extended to the end of string, PCRE compiled incorrect data. This could lead to
4459 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4460
4461 2. Moved to autoconf 2.53 and libtool 1.4.2.
4462
4463 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4464 the special perltest8 script is no longer needed - all the tests can be run
4465 from a single perltest script.
4466
4467 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4468 by \s. It has now been removed in PCRE. This means it isn't recognized as
4469 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4470 class [:space:] *does* include VT, thereby creating a mess.
4471
4472 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4473 space and tab.
4474
4475 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4476 its new features into the main test script, reducing the number of scripts.
4477
4478 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4479 were backward compatible, and made the (?i) apply to the whole pattern, as if
4480 /i were given. Now it behaves more logically, and applies the option setting
4481 only to what follows. PCRE has been changed to follow suit. However, if it
4482 finds options settings right at the start of the pattern, it extracts them into
4483 the global options, as before. Thus, they show up in the info data.
4484
4485 8. Added support for the \Q...\E escape sequence. Characters in between are
4486 treated as literals. This is slightly different from Perl in that $ and @ are
4487 also handled as literals inside the quotes. In Perl, they will cause variable
4488 interpolation. Note the following examples:
4489
4490 Pattern PCRE matches Perl matches
4491
4492 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4493 \Qabc\$xyz\E abc\$xyz abc\$xyz
4494 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4495
4496 For compatibility with Perl, \Q...\E sequences are recognized inside character
4497 classes as well as outside them.
4498
4499 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4500 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4501 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4502 signed/unsigned warnings.
4503
4504 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4505 option for pcretest, so I've replaced it by a simple function that does just
4506 that job.
4507
4508 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4509 "pcregrep -".
4510
4511 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4512 Java package. This provides some syntactic sugar for simple cases of what my
4513 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4514 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4515 item, you can use this simplified notation. Note that only makes sense with
4516 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4517 greediness, whatever the setting of the PCRE_UNGREEDY option.
4518
4519 13. A change of greediness default within a pattern was not taking effect at
4520 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4521 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4522 was abstracted outside.
4523
4524 14. PCRE now supports the \G assertion. It is true when the current matching
4525 position is at the start point of the match. This differs from \A when the
4526 starting offset is non-zero. Used with the /g option of pcretest (or similar
4527 code), it works in the same way as it does for Perl's /g option. If all
4528 alternatives of a regex begin with \G, the expression is anchored to the start
4529 match position, and the "anchored" flag is set in the compiled expression.
4530
4531 15. Some bugs concerning the handling of certain option changes within patterns
4532 have been fixed. These applied to options other than (?ims). For example,
4533 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4534 been the other way round. Some of this was related to change 7 above.
4535
4536 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4537 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4538 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4539 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4540
4541 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4542 mode. Unlike ".", it always matches newline, whatever the setting of
4543 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4544 assertions. Perl allows it, but it doesn't (in general) work because it can't
4545 calculate the length of the lookbehind. At least, that's the case for Perl
4546 5.8.0 - I've been told they are going to document that it doesn't work in
4547 future.
4548
4549 18. Added an error diagnosis for escapes that PCRE does not support: these are
4550 \L, \l, \N, \P, \p, \U, \u, and \X.
4551
4552 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4553 reading past the end of the pattern in cases such as /[abcd/.
4554
4555 20. PCRE was getting more memory than necessary for patterns with classes that
4556 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4557
4558 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4559 compiling PCRE for use with Virtual Pascal.
4560
4561 22. Small fix to the Makefile to make it work properly if the build is done
4562 outside the source tree.
4563
4564 23. Added a new extension: a condition to go with recursion. If a conditional
4565 subpattern starts with (?(R) the "true" branch is used if recursion has
4566 happened, whereas the "false" branch is used only at the top level.
4567
4568 24. When there was a very long string of literal characters (over 255 bytes
4569 without UTF support, over 250 bytes with UTF support), the computation of how
4570 much memory was required could be incorrect, leading to segfaults or other
4571 strange effects.
4572
4573 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4574 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4575 there was a subsequent back reference to those brackets. This meant that, for
4576 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4577 possible to check for precisely this case. All we can do is abandon the
4578 optimization if .* occurs inside capturing brackets when there are any back
4579 references whatsoever. (See below for a better fix that came later.)
4580
4581 26. The handling of the optimization for finding the first character of a
4582 non-anchored pattern, and for finding a character that is required later in the
4583 match were failing in some cases. This didn't break the matching; it just
4584 failed to optimize when it could. The way this is done has been re-implemented.
4585
4586 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4587
4588 28. Added a new feature that provides some of the functionality that Perl
4589 provides with (?{...}). The facility is termed a "callout". The way it is done
4590 in PCRE is for the caller to provide an optional function, by setting
4591 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4592 global variable. By default it is unset, which disables all calling out. To get
4593 the function called, the regex must include (?C) at appropriate points. This
4594 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4595 This provides a means of identifying different callout points. When PCRE
4596 reaches such a point in the regex, if pcre_callout has been set, the external
4597 function is called. It is provided with data in a structure called
4598 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4599 matching continues; if it returns a non-zero value, the match at the current
4600 point fails. However, backtracking will occur if possible. [This was changed
4601 later and other features added - see item 49 below.]
4602
4603 29. pcretest is upgraded to test the callout functionality. It provides a
4604 callout function that displays information. By default, it shows the start of
4605 the match and the current position in the text. There are some new data escapes
4606 to vary what happens:
4607
4608 \C+ in addition, show current contents of captured substrings
4609 \C- do not supply a callout function
4610 \C!n return 1 when callout number n is reached
4611 \C!n!m return 1 when callout number n is reached for the mth time
4612
4613 30. If pcregrep was called with the -l option and just a single file name, it
4614 output "<stdin>" if a match was found, instead of the file name.
4615
4616 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4617 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4618 pcre_exec(). This saves a malloc/free per call. The default value of
4619 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4620 when configuring.
4621
4622 32. The default maximum size of a compiled pattern is 64K. There have been a
4623 few cases of people hitting this limit. The code now uses macros to handle the
4624 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4625 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4626 configuring. Tests 2 and 5 work only with 2-byte links because they output
4627 debugging information about compiled patterns.
4628
4629 33. Internal code re-arrangements:
4630
4631 (a) Moved the debugging function for printing out a compiled regex into
4632 its own source file (printint.c) and used #include to pull it into
4633 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4634 separate copies.
4635
4636 (b) Defined the list of op-code names for debugging as a macro in
4637 internal.h so that it is next to the definition of the opcodes.
4638
4639 (c) Defined a table of op-code lengths for simpler skipping along compiled
4640 code. This is again a macro in internal.h so that it is next to the
4641 definition of the opcodes.
4642
4643 34. Added support for recursive calls to individual subpatterns, along the
4644 lines of Robin Houston's patch (but implemented somewhat differently).
4645
4646 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4647 allow it to read and process whole directories in Win32. This code was
4648 contributed by Lionel Fourquaux; it has not been tested by me.
4649
4650 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4651 used to name a group. Names consist of alphanumerics and underscores, and must
4652 be unique. Back references use the syntax (?P=name) and recursive calls use
4653 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4654 numbers. The function pcre_fullinfo() can be used after compilation to extract
4655 a name/number map. There are three relevant calls:
4656
4657 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4658 PCRE_INFO_NAMECOUNT yields the number of entries
4659 PCRE_INFO_NAMETABLE yields a pointer to the map.
4660
4661 The map is a vector of fixed-size entries. The size of each entry depends on
4662 the length of the longest name used. The first two bytes of each entry are the
4663 group number, most significant byte first. There follows the corresponding
4664 name, zero terminated. The names are in alphabetical order.
4665
4666 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4667 case instead of 255. Making it the same both with and without UTF-8 support
4668 means that the same test output works with both.
4669
4670 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4671 calling malloc() with a zero argument.
4672
4673 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4674 optimization. I've improved things by keeping a bitmap of backreferences with
4675 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4676 fact referenced, the optimization can be applied. It is unlikely that a
4677 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4678 the match to follow \n) will appear inside brackets with a number greater than
4679 31, but if it does, any back reference > 31 suppresses the optimization.
4680
4681 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4682 of disabling numbered capturing parentheses. Any opening parenthesis that is
4683 not followed by ? behaves as if it were followed by ?: but named parentheses
4684 can still be used for capturing (and they will acquire numbers in the usual
4685 way).
4686
4687 41. Redesigned the return codes from the match() function into yes/no/error so
4688 that errors can be passed back from deep inside the nested calls. A malloc
4689 failure while inside a recursive subpattern call now causes the
4690 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4691
4692 42. It is now possible to set a limit on the number of times the match()
4693 function is called in a call to pcre_exec(). This facility makes it possible to
4694 limit the amount of recursion and backtracking, though not in a directly
4695 obvious way, because the match() function is used in a number of different
4696 circumstances. The count starts from zero for each position in the subject
4697 string (for non-anchored patterns). The default limit is, for compatibility, a
4698 large number, namely 10 000 000. You can change this in two ways:
4699
4700 (a) When configuring PCRE before making, you can use --with-match-limit=n
4701 to set a default value for the compiled library.
4702
4703 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4704 a different value is set. See 45 below.
4705
4706 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4707
4708 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4709 of things that can be changed at compile time. The first argument specifies
4710 what is wanted and the second points to where the information is to be placed.
4711 The current list of available information is:
4712
4713 PCRE_CONFIG_UTF8
4714
4715 The output is an integer that is set to one if UTF-8 support is available;
4716 otherwise it is set to zero.
4717
4718 PCRE_CONFIG_NEWLINE
4719
4720 The output is an integer that it set to the value of the code that is used for
4721 newline. It is either LF (10) or CR (13).
4722
4723 PCRE_CONFIG_LINK_SIZE
4724
4725 The output is an integer that contains the number of bytes used for internal
4726 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4727
4728 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4729
4730 The output is an integer that contains the threshold above which the POSIX
4731 interface uses malloc() for output vectors. See item 31 above.
4732
4733 PCRE_CONFIG_MATCH_LIMIT
4734
4735 The output is an unsigned integer that contains the default limit of the number
4736 of match() calls in a pcre_exec() execution. See 42 above.
4737
4738 44. pcretest has been upgraded by the addition of the -C option. This causes it
4739 to extract all the available output from the new pcre_config() function, and to
4740 output it. The program then exits immediately.
4741
4742 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4743 order to support additional features. One way would have been to define
4744 pcre_exec2() (for example) with extra arguments, but this would not have been
4745 extensible, and would also have required all calls to the original function to
4746 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4747 is used for passing in "extra" data from pcre_study().
4748
4749 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4750 contains the following fields:
4751
4752 flags a bitmap indicating which of the following fields are set
4753 study_data opaque data from pcre_study()
4754 match_limit a way of specifying a limit on match() calls for a specific
4755 call to pcre_exec()
4756 callout_data data for callouts (see 49 below)
4757
4758 The flag bits are also defined in pcre.h, and are
4759
4760 PCRE_EXTRA_STUDY_DATA
4761 PCRE_EXTRA_MATCH_LIMIT
4762 PCRE_EXTRA_CALLOUT_DATA
4763
4764 The pcre_study() function now returns one of these new pcre_extra blocks, with
4765 the actual study data pointed to by the study_data field, and the
4766 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4767 before. That is, this change is entirely upwards-compatible and requires no
4768 change to existing code.
4769
4770 If you want to pass in additional data to pcre_exec(), you can either place it
4771 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4772 block.
4773
4774 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4775 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4776 times with different match limits, until it finds the minimum value needed for
4777 pcre_exec() to complete. The value is then output. This can be instructive; for
4778 most simple matches the number is quite small, but for pathological cases it
4779 gets very large very quickly.
4780
4781 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4782 returns the size of the data block pointed to by the study_data field in a
4783 pcre_extra block, that is, the value that was passed as the argument to
4784 pcre_malloc() when PCRE was getting memory in which to place the information
4785 created by pcre_study(). The fourth argument should point to a size_t variable.
4786 pcretest has been extended so that this information is shown after a successful
4787 pcre_study() call when information about the compiled regex is being displayed.
4788
4789 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4790 because what follows is always an absolute path. (Later: it turns out that this
4791 is more than cosmetic for MinGW, because it doesn't like empty path
4792 components.)
4793
4794 49. Some changes have been made to the callout feature (see 28 above):
4795
4796 (i) A callout function now has three choices for what it returns:
4797
4798 0 => success, carry on matching
4799 > 0 => failure at this point, but backtrack if possible
4800 < 0 => serious error, return this value from pcre_exec()
4801
4802 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4803 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4804 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4805 use by callout functions. It will never be used by PCRE itself.
4806
4807 (ii) The pcre_extra structure (see 45 above) has a void * field called
4808 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4809 pcre_callout_block structure has a field of the same name. The contents of
4810 the field passed in the pcre_extra structure are passed to the callout
4811 function in the corresponding field in the callout block. This makes it
4812 easier to use the same callout-containing regex from multiple threads. For
4813 testing, the pcretest program has a new data escape
4814
4815 \C*n pass the number n (may be negative) as callout_data
4816
4817 If the callout function in pcretest receives a non-zero value as
4818 callout_data, it returns that value.
4819
4820 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4821 there were some redundant $(CFLAGS) in commands that are now specified as
4822 $(LINK), which already includes $(CFLAGS).
4823
4824 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4825 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4826 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4827 one-byte characters throughout. Note that case-insensitive matching applies
4828 only to characters whose values are less than 256. PCRE doesn't support the
4829 notion of cases for higher-valued characters.
4830
4831 (i) A character class whose characters are all within 0-255 is handled as
4832 a bit map, and the map is inverted for negative classes. Previously, a
4833 character > 255 always failed to match such a class; however it should
4834 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4835
4836 (ii) A negated character class with a single character < 255 is coded as
4837 "not this character" (OP_NOT). This wasn't working properly when the test
4838 character was multibyte, either singly or repeated.
4839
4840 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4841 mode, for example: \x{100}{2,3}.
4842
4843 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4844 singly or repeated) now correctly test multibyte characters. However,
4845 PCRE doesn't recognize any characters with values greater than 255 as
4846 digits, spaces, or word characters. Such characters always match \D, \S,
4847 and \W, and never match \d, \s, or \w.
4848
4849 (v) Classes may now contain characters and character ranges with values
4850 greater than 255. For example: [ab\x{100}-\x{400}].
4851
4852 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4853 PCRE in UTF-8 mode.
4854
4855 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4856 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4857 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4858 value.)
4859
4860 53. The single man page has become too large. I have therefore split it up into
4861 a number of separate man pages. These also give rise to individual HTML pages;
4862 these are now put in a separate directory, and there is an index.html page that
4863 lists them all. Some hyperlinking between the pages has been installed.
4864
4865 54. Added convenience functions for handling named capturing parentheses.
4866
4867 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4868 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4869 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4870 are faulted.
4871
4872 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4873 calling configure. These values are used when compiling the dftables.c program
4874 which is run to generate the source of the default character tables. They
4875 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4876 you will need to set these values.
4877
4878 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4879
4880
4881 Version 3.9 02-Jan-02
4882 ---------------------
4883
4884 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4885
4886 2. If --disable-static was given, the building process failed when trying to
4887 build pcretest and pcregrep. (For some reason it was using libtool to compile
4888 them, which is not right, as they aren't part of the library.)
4889
4890
4891 Version 3.8 18-Dec-01
4892 ---------------------
4893
4894 1. The experimental UTF-8 code was completely screwed up. It was packing the
4895 bytes in the wrong order. How dumb can you get?
4896
4897
4898 Version 3.7 29-Oct-01
4899 ---------------------
4900
4901 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4902 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4903 this didn't happen under Solaris 8, where I normally test things.
4904
4905 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4906 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4907 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4908 this fix an hour or so after the initial 3.7 release.)
4909
4910
4911 Version 3.6 23-Oct-01
4912 ---------------------
4913
4914 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4915 offsets passed as NULL with zero offset count.
4916
4917 2. The config.guess and config.sub files had not been updated when I moved to
4918 the latest autoconf.
4919
4920
4921 Version 3.5 15-Aug-01
4922 ---------------------
4923
4924 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4925 had been forgotten.
4926
4927 2. By using declared but undefined structures, we can avoid using "void"
4928 definitions in pcre.h while keeping the internal definitions of the structures
4929 private.
4930
4931 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4932 user point of view, this means that both static and shared libraries are built
4933 by default, but this can be individually controlled. More of the work of
4934 handling this static/shared cases is now inside libtool instead of PCRE's make
4935 file.
4936
4937 4. The pcretest utility is now installed along with pcregrep because it is
4938 useful for users (to test regexs) and by doing this, it automatically gets
4939 relinked by libtool. The documentation has been turned into a man page, so
4940 there are now .1, .txt, and .html versions in /doc.
4941
4942 5. Upgrades to pcregrep:
4943 (i) Added long-form option names like gnu grep.
4944 (ii) Added --help to list all options with an explanatory phrase.
4945 (iii) Added -r, --recursive to recurse into sub-directories.
4946 (iv) Added -f, --file to read patterns from a file.
4947
4948 6. pcre_exec() was referring to its "code" argument before testing that
4949 argument for NULL (and giving an error if it was NULL).
4950
4951 7. Upgraded Makefile.in to allow for compiling in a different directory from
4952 the source directory.
4953
4954 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
4955 options bits, the pointer it was passed was to an int instead of to an unsigned
4956 long int. This mattered only on 64-bit systems.
4957
4958 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4959 generated) instead of pcre.in, which it its source. Also made the same change
4960 in several of the .c files.
4961
4962 10. A new release of gcc defines printf() as a macro, which broke pcretest
4963 because it had an ifdef in the middle of a string argument for printf(). Fixed
4964 by using separate calls to printf().
4965
4966 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4967 script, to force use of CR or LF instead of \n in the source. On non-Unix
4968 systems, the value can be set in config.h.
4969
4970 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4971 absolute limit. Changed the text of the error message to make this clear, and
4972 likewise updated the man page.
4973
4974 13. The limit of 99 on the number of capturing subpatterns has been removed.
4975 The new limit is 65535, which I hope will not be a "real" limit.
4976
4977
4978 Version 3.4 22-Aug-00
4979 ---------------------
4980
4981 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4982
4983 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4984
4985
4986 Version 3.3 01-Aug-00
4987 ---------------------
4988
4989 1. If an octal character was given, but the value was greater than \377, it
4990 was not getting masked to the least significant bits, as documented. This could
4991 lead to crashes in some systems.
4992
4993 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4994 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4995
4996 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4997 These just pass their arguments on to (pcre_free)(), but they are provided
4998 because some uses of PCRE bind it to non-C systems that can call its functions,
4999 but cannot call free() or pcre_free() directly.
5000
5001 4. Add "make test" as a synonym for "make check". Corrected some comments in
5002 the Makefile.
5003
5004 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5005 Makefile.
5006
5007 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5008 command called pgrep for grepping around the active processes.
5009
5010 7. Added the beginnings of support for UTF-8 character strings.
5011
5012 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5013 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5014 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5015 out for the ar command.)
5016
5017
5018 Version 3.2 12-May-00
5019 ---------------------
5020
5021 This is purely a bug fixing release.
5022
5023 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5024 of ZA. This was just one example of several cases that could provoke this bug,
5025 which was introduced by change 9 of version 2.00. The code for breaking
5026 infinite loops after an iteration that matches an empty string was't working
5027 correctly.
5028
5029 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5030 when matched against abbab (for example). After matching an empty string, it
5031 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5032 caused it to match further down the string than it should.
5033
5034 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5035 was there because it doesn't seem to be needed, and it causes trouble on some
5036 systems, as it is not a Standard C header. It has been removed.
5037
5038 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5039 were reported on the Macintosh. The changes were from
5040
5041 while ((c = *(++ptr)) != 0 && c != '\n');
5042 to
5043 while ((c = *(++ptr)) != 0 && c != '\n') ;
5044
5045 Totally extraordinary, but if that's what it takes...
5046
5047 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5048 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5049 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5050 assumes the way PCRE uses memmove() (always moving upwards).
5051
5052 6. PCRE is being used in one environment where strchr() is not available. There
5053 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5054 faster code anyway.
5055
5056
5057 Version 3.1 09-Feb-00
5058 ---------------------
5059
5060 The only change in this release is the fixing of some bugs in Makefile.in for
5061 the "install" target:
5062
5063 (1) It was failing to install pcreposix.h.
5064
5065 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5066
5067
5068 Version 3.0 01-Feb-00
5069 ---------------------
5070
5071 1. Add support for the /+ modifier to perltest (to output $` like it does in
5072 pcretest).
5073
5074 2. Add support for the /g modifier to perltest.
5075
5076 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5077 matches null strings.
5078
5079 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5080 pattern. Perl treats empty patterns specially - it reuses the most recent
5081 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5082 effect.
5083
5084 5. The POSIX interface was broken in that it was just handing over the POSIX
5085 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5086 required a bigger vector, with some working space on the end. This means that
5087 the POSIX wrapper now has to get and free some memory, and copy the results.
5088
5089 6. Added some simple autoconf support, placing the test data and the
5090 documentation in separate directories, re-organizing some of the
5091 information files, and making it build pcre-config (a GNU standard). Also added
5092 libtool support for building PCRE as a shared library, which is now the
5093 default.
5094
5095 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5096 09 are not valid octal constants. Single digits will be used for minor values
5097 less than 10.
5098
5099 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5100 existing programs that set these in the POSIX interface can use PCRE without
5101 modification.
5102
5103 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5104 return all that pcre_info() returns, plus additional data. The pcre_info()
5105 function is retained for compatibility, but is considered to be obsolete.
5106
5107 10. Added experimental recursion feature (?R) to handle one common case that
5108 Perl 5.6 will be able to do with (?p{...}).
5109
5110 11. Added support for POSIX character classes like [:alpha:], which Perl is
5111 adopting.
5112
5113
5114 Version 2.08 31-Aug-99
5115 ----------------------
5116
5117 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5118 trying to match at the startoffset position, but instead was moving forward to
5119 the next newline as if a previous match had failed.
5120
5121 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5122 and could get into a loop if a null string was matched other than at the start
5123 of the subject.
5124
5125 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5126 be distinguished at compile time, and for completeness also added PCRE_DATE.
5127
5128 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5129 in GnuWin32 environments.
5130
5131
5132 Version 2.07 29-Jul-99
5133 ----------------------
5134
5135 1. The documentation is now supplied in plain text form and HTML as well as in
5136 the form of man page sources.
5137
5138 2. C++ compilers don't like assigning (void *) values to other pointer types.
5139 In particular this affects malloc(). Although there is no problem in Standard
5140 C, I've put in casts to keep C++ compilers happy.
5141
5142 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5143 should be (const char *).
5144
5145 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5146 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5147 However, I haven't made this a standard facility. The documentation doesn't
5148 mention it, and the Makefile doesn't support it.
5149
5150 5. The Makefile now contains an "install" target, with editable destinations at
5151 the top of the file. The pcretest program is not installed.
5152
5153 6. pgrep -V now gives the PCRE version number and date.
5154
5155 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5156 causing the entire string to be ignored, instead of just the last character.
5157
5158 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5159 non-matching string, it can take a very, very long time, even for strings of
5160 quite modest length, because of the nested recursion. PCRE now does better in
5161 some of these cases. It does this by remembering the last required literal
5162 character in the pattern, and pre-searching the subject to ensure it is present
5163 before running the real match. In other words, it applies a heuristic to detect
5164 some types of certain failure quickly, and in the above example, if presented
5165 with a string that has no trailing " it gives "no match" very quickly.
5166
5167 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5168 other alternatives are tried instead.
5169
5170
5171 Version 2.06 09-Jun-99
5172 ----------------------
5173
5174 1. Change pcretest's output for amount of store used to show just the code
5175 space, because the remainder (the data block) varies in size between 32-bit and
5176 64-bit systems.
5177
5178 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5179 start matching at. This allows lookbehinds to work when searching for multiple
5180 occurrences in a string.
5181
5182 3. Added additional options to pcretest for testing multiple occurrences:
5183
5184 /+ outputs the rest of the string that follows a match
5185 /g loops for multiple occurrences, using the new startoffset argument
5186 /G loops for multiple occurrences by passing an incremented pointer
5187
5188 4. PCRE wasn't doing the "first character" optimization for patterns starting
5189 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5190 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5191 the letter 'x'. On long subject strings, this gives a significant speed-up.
5192
5193
5194 Version 2.05 21-Apr-99
5195 ----------------------
5196
5197 1. Changed the type of magic_number from int to long int so that it works
5198 properly on 16-bit systems.
5199
5200 2. Fixed a bug which caused patterns starting with .* not to work correctly
5201 when the subject string contained newline characters. PCRE was assuming
5202 anchoring for such patterns in all cases, which is not correct because .* will
5203 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5204 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5205 must be retried after every newline in the subject.
5206
5207
5208 Version 2.04 18-Feb-99
5209 ----------------------
5210
5211 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5212 computation of the store needed to hold the pattern was incorrect (too large).
5213 If such patterns were nested a few deep, this could multiply and become a real
5214 problem.
5215
5216 2. Added /M option to pcretest to show the memory requirement of a specific
5217 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5218
5219 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5220 compiled in such a way that the backtracking after subsequent failure was
5221 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5222 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5223
5224
5225 Version 2.03 02-Feb-99
5226 ----------------------
5227
5228 1. Fixed typo and small mistake in man page.
5229
5230 2. Added 4th condition (GPL supersedes if conflict) and created separate
5231 LICENCE file containing the conditions.
5232
5233 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5234 Perl, that is the internal \ allows the delimiter to be included in the
5235 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5236 the final delimiter, add \ to the end of the pattern (to test the error).
5237
5238 4. Added the convenience functions for extracting substrings after a successful
5239 match. Updated pcretest to make it able to test these functions.
5240
5241
5242 Version 2.02 14-Jan-99
5243 ----------------------
5244
5245 1. Initialized the working variables associated with each extraction so that
5246 their saving and restoring doesn't refer to uninitialized store.
5247
5248 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5249 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5250 fix the problem.
5251
5252 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5253 calls, and wasn't printing the correct value for compiling calls. Increased the
5254 default value of LOOPREPEAT, and the number of significant figures in the
5255 times.
5256
5257 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5258
5259 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5260 a building problem on Windows NT with a FAT file system.
5261
5262
5263 Version 2.01 21-Oct-98
5264 ----------------------
5265
5266 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5267 to character tables built by pcre_maketables() in the current locale. If NULL
5268 is passed, the default tables are used.
5269
5270
5271 Version 2.00 24-Sep-98
5272 ----------------------
5273
5274 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5275 it any more.
5276
5277 2. Allow quantification of (?>) groups, and make it work correctly.
5278
5279 3. The first character computation wasn't working for (?>) groups.
5280
5281 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5282 end of the subject) and add 5.005's \z, which really does match only at the
5283 very end of the subject.
5284
5285 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5286
5287 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5288 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5289 localized options. All options to pcre_study() were also removed.
5290
5291 7. Add other new features from 5.005:
5292
5293 $(?<= positive lookbehind
5294 $(?<! negative lookbehind
5295 (?imsx-imsx) added the unsetting capability
5296 such a setting is global if at outer level; local otherwise
5297 (?imsx-imsx:) non-capturing groups with option setting
5298 (?(cond)re|re) conditional pattern matching
5299
5300 A backreference to itself in a repeated group matches the previous
5301 captured string.
5302
5303 8. General tidying up of studying (both automatic and via "study")
5304 consequential on the addition of new assertions.
5305
5306 9. As in 5.005, unlimited repeated groups that could match an empty substring
5307 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5308 runtime if any iteration does actually match an empty substring.
5309
5310 10. Include the RunTest script in the distribution.
5311
5312 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5313 discrepancies, some of which were old and were also with respect to 5.004. They
5314 have now been fixed.
5315
5316
5317 Version 1.09 28-Apr-98
5318 ----------------------
5319
5320 1. A negated single character class followed by a quantifier with a minimum
5321 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5322 program crashes, or just wrong answers. This did not apply to negated classes
5323 containing more than one character, or to minima other than one.
5324
5325
5326 Version 1.08 27-Mar-98
5327 ----------------------
5328
5329 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5330
5331 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5332 latter must appear before anything that relies on it in the pattern.
5333
5334
5335 Version 1.07 16-Feb-98
5336 ----------------------
5337
5338 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5339 repeat of a potentially empty string).
5340
5341
5342 Version 1.06 23-Jan-98
5343 ----------------------
5344
5345 1. Added Markus Oberhumer's little patches for C++.
5346
5347 2. Literal strings longer than 255 characters were broken.
5348
5349
5350 Version 1.05 23-Dec-97
5351 ----------------------
5352
5353 1. Negated character classes containing more than one character were failing if
5354 PCRE_CASELESS was set at run time.
5355
5356
5357 Version 1.04 19-Dec-97
5358 ----------------------
5359
5360 1. Corrected the man page, where some "const" qualifiers had been omitted.
5361
5362 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5363 input syntax.
5364
5365 3. Fixed memory leak which occurred when a regex with back references was
5366 matched with an offsets vector that wasn't big enough. The temporary memory
5367 that is used in this case wasn't being freed if the match failed.
5368
5369 4. Tidied pcretest to ensure it frees memory that it gets.
5370
5371 5. Temporary memory was being obtained in the case where the passed offsets
5372 vector was exactly big enough.
5373
5374 6. Corrected definition of offsetof() from change 5 below.
5375
5376 7. I had screwed up change 6 below and broken the rules for the use of
5377 setjmp(). Now fixed.
5378
5379
5380 Version 1.03 18-Dec-97
5381 ----------------------
5382
5383 1. A erroneous regex with a missing opening parenthesis was correctly
5384 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5385 on some systems.
5386
5387 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5388 it was reported that one broken compiler failed on the former because "code" is
5389 also an independent variable.
5390
5391 3. The erroneous regex a[]b caused an array overrun reference.
5392
5393 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5394 fail on data ending with that character. (It was going on too far, and checking
5395 the next character, typically a binary zero.) This was specific to the
5396 optimized code for single-character negative classes.
5397
5398 5. Added a contributed patch from the TIN world which does the following:
5399
5400 + Add an undef for memmove, in case the the system defines a macro for it.
5401
5402 + Add a definition of offsetof(), in case there isn't one. (I don't know
5403 the reason behind this - offsetof() is part of the ANSI standard - but
5404 it does no harm).
5405
5406 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5407 most of the places where whitespace preceded '#'. I have given up and
5408 allowed the remaining 2 cases to be at the margin.
5409
5410 + Rename some variables in pcre to eliminate shadowing. This seems very
5411 pedantic, but does no harm, of course.
5412
5413 6. Moved the call to setjmp() into its own function, to get rid of warnings
5414 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5415
5416 7. Constructs such as \d{8,} were compiling into the equivalent of
5417 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5418 outcome, but in this particular case used more store than had been allocated,
5419 which caused the bug to be discovered because it threw up an internal error.
5420
5421 8. The debugging code in both pcre and pcretest for outputting the compiled
5422 form of a regex was going wrong in the case of back references followed by
5423 curly-bracketed repeats.
5424
5425
5426 Version 1.02 12-Dec-97
5427 ----------------------
5428
5429 1. Typos in pcre.3 and comments in the source fixed.
5430
5431 2. Applied a contributed patch to get rid of places where it used to remove
5432 'const' from variables, and fixed some signed/unsigned and uninitialized
5433 variable warnings.
5434
5435 3. Added the "runtest" target to Makefile.
5436
5437 4. Set default compiler flag to -O2 rather than just -O.
5438
5439
5440 Version 1.01 19-Nov-97
5441 ----------------------
5442
5443 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5444 like /([ab]*)*/, that is, for classes with more than one character in them.
5445
5446 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5447 as /((?>a*))*/ (a PCRE_EXTRA facility).
5448
5449
5450 Version 1.00 18-Nov-97
5451 ----------------------
5452
5453 1. Added compile-time macros to support systems such as SunOS4 which don't have
5454 memmove() or strerror() but have other things that can be used instead.
5455
5456 2. Arranged that "make clean" removes the executables.
5457
5458
5459 Version 0.99 27-Oct-97
5460 ----------------------
5461
5462 1. Fixed bug in code for optimizing classes with only one character. It was
5463 initializing a 32-byte map regardless, which could cause it to run off the end
5464 of the memory it had got.
5465
5466 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5467
5468
5469 Version 0.98 22-Oct-97
5470 ----------------------
5471
5472 1. Fixed bug in code for handling temporary memory usage when there are more
5473 back references than supplied space in the ovector. This could cause segfaults.
5474
5475
5476 Version 0.97 21-Oct-97
5477 ----------------------
5478
5479 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5480
5481 2. Optimized negated single characters not to use a bit map.
5482
5483 3. Brought error texts together as macro definitions; clarified some of them;
5484 fixed one that was wrong - it said "range out of order" when it meant "invalid
5485 escape sequence".
5486
5487 4. Changed some char * arguments to const char *.
5488
5489 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5490
5491 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5492 pcretest.
5493
5494
5495 Version 0.96 16-Oct-97
5496 ----------------------
5497
5498 1. Added a simple "pgrep" utility to the distribution.
5499
5500 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5501 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5502 where "ddd" means "one or more decimal digits".
5503
5504 3. Fixed serious bug. If a pattern had a back reference, but the call to
5505 pcre_exec() didn't supply a large enough ovector to record the related
5506 identifying subpattern, the match always failed. PCRE now remembers the number
5507 of the largest back reference, and gets some temporary memory in which to save
5508 the offsets during matching if necessary, in order to ensure that
5509 backreferences always work.
5510
5511 4. Increased the compatibility with Perl in a number of ways:
5512
5513 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5514 to request this handling. The option can be set at compile or exec time.
5515
5516 (b) $ matches before a terminating newline by default; an option
5517 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5518 mode). The option can be set at compile or exec time.
5519
5520 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5521 the same as Perl's. If the decimal number it represents is less than 10
5522 or there aren't that many previous left capturing parentheses, an octal
5523 escape is read. Inside a character class, it's always an octal escape,
5524 even if it is a single digit.
5525
5526 (d) An escaped but undefined alphabetic character is taken as a literal,
5527 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5528 escapes.
5529
5530 (e) {0} is now permitted. (The previous item is removed from the compiled
5531 pattern).
5532
5533 5. Changed all the names of code files so that the basic parts are no longer
5534 than 10 characters, and abolished the teeny "globals.c" file.
5535
5536 6. Changed the handling of character classes; they are now done with a 32-byte
5537 bit map always.
5538
5539 7. Added the -d and /D options to pcretest to make it possible to look at the
5540 internals of compilation without having to recompile pcre.
5541
5542
5543 Version 0.95 23-Sep-97
5544 ----------------------
5545
5546 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5547 \x20 at the start of a run of normal characters. These were being treated as
5548 real characters, instead of the source characters being re-checked.
5549
5550
5551 Version 0.94 18-Sep-97
5552 ----------------------
5553
5554 1. The functions are now thread-safe, with the caveat that the global variables
5555 containing pointers to malloc() and free() or alternative functions are the
5556 same for all threads.
5557
5558 2. Get pcre_study() to generate a bitmap of initial characters for non-
5559 anchored patterns when this is possible, and use it if passed to pcre_exec().
5560
5561
5562 Version 0.93 15-Sep-97
5563 ----------------------
5564
5565 1. /(b)|(:+)/ was computing an incorrect first character.
5566
5567 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5568 but not actually doing anything yet.
5569
5570 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5571 as Perl does (e.g. [-az] or [az-]).
5572
5573 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5574 all possible positions.
5575
5576 5. Split up into different modules to avoid including unneeded functions in a
5577 compiled binary. However, compile and exec are still in one module. The "study"
5578 function is split off.
5579
5580 6. The character tables are now in a separate module whose source is generated
5581 by an auxiliary program - but can then be edited by hand if required. There are
5582 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5583 toupper() in the code.
5584
5585 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5586 make them global. Abolish the function for setting them, as the caller can now
5587 set them directly.
5588
5589
5590 Version 0.92 11-Sep-97
5591 ----------------------
5592
5593 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5594 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5595
5596 2. Caseless matching was not working in character classes if the characters in
5597 the pattern were in upper case.
5598
5599 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5600
5601 4. Make PCRE_ANCHORED public and accept as a compile option.
5602
5603 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5604 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5605 pass them.
5606
5607 6. Give an error if bad option bits passed at compile or run time.
5608
5609 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5610 pcretest to cause it to pass that flag.
5611
5612 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5613 options, and the first character, if set.
5614
5615 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5616
5617
5618 Version 0.91 10-Sep-97
5619 ----------------------
5620
5621 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5622 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5623
5624 2. PCRE was looping on encountering an indefinitely repeated back reference to
5625 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5626 Perl does - treats the match as successful.
5627
5628 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5