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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1570 - (show annotations)
Mon Jun 22 09:38:41 2015 UTC (4 years, 2 months ago) by ph10
File size: 280684 byte(s)
Fix EBCDIC [\h] bug.
1 ChangeLog for PCRE
2 ------------------
3
4 Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5 development is happening in the PCRE2 10.xx series.
6
7 Version 8.38 xx-xxx-xxxx
8 ------------------------
9
10 1. If a group that contained a recursive back reference also contained a
11 forward reference subroutine call followed by a non-forward-reference
12 subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
13 compile correct code, leading to undefined behaviour or an internally
14 detected error. This bug was discovered by the LLVM fuzzer.
15
16 2. Quantification of certain items (e.g. atomic back references) could cause
17 incorrect code to be compiled when recursive forward references were
18 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
19 This bug was discovered by the LLVM fuzzer.
20
21 3. A repeated conditional group whose condition was a reference by name caused
22 a buffer overflow if there was more than one group with the given name.
23 This bug was discovered by the LLVM fuzzer.
24
25 4. A recursive back reference by name within a group that had the same name as
26 another group caused a buffer overflow. For example:
27 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
28
29 5. A forward reference by name to a group whose number is the same as the
30 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
31 a buffer overflow at compile time. This bug was discovered by the LLVM
32 fuzzer.
33
34 6. A lookbehind assertion within a set of mutually recursive subpatterns could
35 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
36
37 7. Another buffer overflow bug involved duplicate named groups with a
38 reference between their definition, with a group that reset capture
39 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
40 fixed by always allowing for more memory, even if not needed. (A proper fix
41 is implemented in PCRE2, but it involves more refactoring.)
42
43 8. There was no check for integer overflow in subroutine calls such as (?123).
44
45 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
46 being treated as a literal 'l' instead of causing an error.
47
48 10. There was a buffer overflow if pcre_exec() was called with an ovector of
49 size 1. This bug was found by american fuzzy lop.
50
51 11. If a non-capturing group containing a conditional group that could match
52 an empty string was repeated, it was not identified as matching an empty
53 string itself. For example: /^(?:(?(1)x|)+)+$()/.
54
55 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
56 \a and \e in test subject lines.
57
58 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
59 instead of the EBCDIC value.
60
61 14. The handling of \c in an EBCDIC environment has been revised so that it is
62 now compatible with the specification in Perl's perlebcdic page.
63
64 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
65 ASCII/Unicode. This has now been added to the list of characters that are
66 recognized as white space in EBCDIC.
67
68 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
69 error (correctly) when used outside a class, but did not give an error
70 within a class.
71
72 17. \h within a class was incorrectly compiled in EBCDIC environments.
73
74
75 Version 8.37 28-April-2015
76 --------------------------
77
78 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
79 for those parentheses to be closed with whatever has been captured so far.
80 However, it was failing to mark any other groups between the hightest
81 capture so far and the currrent group as "unset". Thus, the ovector for
82 those groups contained whatever was previously there. An example is the
83 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
84
85 2. If an assertion condition was quantified with a minimum of zero (an odd
86 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
87
88 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
89 unrecognized modifier, a crash could occur.
90
91 4. An attempt to do global matching in pcretest with a zero-length ovector
92 caused a crash.
93
94 5. Fixed a memory leak during matching that could occur for a subpattern
95 subroutine call (recursive or otherwise) if the number of captured groups
96 that had to be saved was greater than ten.
97
98 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
99 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
100 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
101
102 7. A UTF pattern containing a "not" match of a non-ASCII character and a
103 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
104
105 8. When a pattern is compiled, it remembers the highest back reference so that
106 when matching, if the ovector is too small, extra memory can be obtained to
107 use instead. A conditional subpattern whose condition is a check on a
108 capture having happened, such as, for example in the pattern
109 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
110 setting the highest backreference number. This mattered only if pcre_exec()
111 was called with an ovector that was too small to hold the capture, and there
112 was no other kind of back reference (a situation which is probably quite
113 rare). The effect of the bug was that the condition was always treated as
114 FALSE when the capture could not be consulted, leading to a incorrect
115 behaviour by pcre_exec(). This bug has been fixed.
116
117 9. A reference to a duplicated named group (either a back reference or a test
118 for being set in a conditional) that occurred in a part of the pattern where
119 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
120 to be incorrectly calculated, leading to overwriting.
121
122 10. A mutually recursive set of back references such as (\2)(\1) caused a
123 segfault at study time (while trying to find the minimum matching length).
124 The infinite loop is now broken (with the minimum length unset, that is,
125 zero).
126
127 11. If an assertion that was used as a condition was quantified with a minimum
128 of zero, matching went wrong. In particular, if the whole group had
129 unlimited repetition and could match an empty string, a segfault was
130 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
131 assertions to be quantified, but not if they are being used as conditions,
132 so the above pattern is faulted by Perl. PCRE has now been changed so that
133 it also rejects such patterns.
134
135 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
136 failed to allow the zero-repeat case if pcre2_exec() was called with an
137 ovector too small to capture the group.
138
139 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
140 Red Hat Product Security:
141
142 (a) A crash if /K and /F were both set with the option to save the compiled
143 pattern.
144
145 (b) Another crash if the option to print captured substrings in a callout
146 was combined with setting a null ovector, for example \O\C+ as a subject
147 string.
148
149 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
150 forward reference repeated a large (but limited) number of times within a
151 repeated outer group that has a zero minimum quantifier, caused incorrect
152 code to be compiled, leading to the error "internal error:
153 previously-checked referenced subpattern not found" when an incorrect
154 memory address was read. This bug was reported as "heap overflow",
155 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
156 CVE-2015-2325.
157
158 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
159 call within a group that also contained a recursive back reference caused
160 incorrect code to be compiled. This bug was reported as "heap overflow",
161 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
162 number CVE-2015-2326.
163
164 24. Computing the size of the JIT read-only data in advance has been a source
165 of various issues, and new ones are still appear unfortunately. To fix
166 existing and future issues, size computation is eliminated from the code,
167 and replaced by on-demand memory allocation.
168
169 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
170 adjacent to the end of the range, and the range contained characters with
171 more than one other case, caused incorrect behaviour when compiled in UTF
172 mode. In that example, the range a-j was left out of the class.
173
174 26. Fix JIT compilation of conditional blocks, which assertion
175 is converted to (*FAIL). E.g: /(?(?!))/.
176
177 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
178 discovered by the LLVM fuzzer.
179
180 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
181 when this assertion was used as a condition, for example (?(?!)a|b). In
182 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
183 error about an unsupported item.
184
185 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
186 possessification code could take exponential time to complete. A recursion
187 depth limit of 1000 has been imposed to limit the resources used by this
188 optimization.
189
190 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
191 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
192 because \S ensures they are all in the class. The code for doing this was
193 interacting badly with the code for computing the amount of space needed to
194 compile the pattern, leading to a buffer overflow. This bug was discovered
195 by the LLVM fuzzer.
196
197 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
198 other kinds of group caused stack overflow at compile time. This bug was
199 discovered by the LLVM fuzzer.
200
201 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
202 between a subroutine call and its quantifier was incorrectly compiled,
203 leading to buffer overflow or other errors. This bug was discovered by the
204 LLVM fuzzer.
205
206 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
207 assertion after (?(. The code was failing to check the character after
208 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
209 was discovered by the LLVM fuzzer.
210
211 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
212 a fixed maximum following a group that contains a subroutine reference was
213 incorrectly compiled and could trigger buffer overflow. This bug was
214 discovered by the LLVM fuzzer.
215
216 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
217 caused a stack overflow instead of the diagnosis of a non-fixed length
218 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
219
220 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
221 (e.g. /(?<=\Ka)/) could make pcregrep loop.
222
223 37. There was a similar problem to 36 in pcretest for global matches.
224
225 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
226 and a subsequent item in the pattern caused a non-match, backtracking over
227 the repeated \X did not stop, but carried on past the start of the subject,
228 causing reference to random memory and/or a segfault. There were also some
229 other cases where backtracking after \C could crash. This set of bugs was
230 discovered by the LLVM fuzzer.
231
232 39. The function for finding the minimum length of a matching string could take
233 a very long time if mutual recursion was present many times in a pattern,
234 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
235 method has been implemented. This infelicity was discovered by the LLVM
236 fuzzer.
237
238 40. Static linking against the PCRE library using the pkg-config module was
239 failing on missing pthread symbols.
240
241
242 Version 8.36 26-September-2014
243 ------------------------------
244
245 1. Got rid of some compiler warnings in the C++ modules that were shown up by
246 -Wmissing-field-initializers and -Wunused-parameter.
247
248 2. The tests for quantifiers being too big (greater than 65535) were being
249 applied after reading the number, and stupidly assuming that integer
250 overflow would give a negative number. The tests are now applied as the
251 numbers are read.
252
253 3. Tidy code in pcre_exec.c where two branches that used to be different are
254 now the same.
255
256 4. The JIT compiler did not generate match limit checks for certain
257 bracketed expressions with quantifiers. This may lead to exponential
258 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
259 issue should be resolved now.
260
261 5. Fixed an issue, which occures when nested alternatives are optimized
262 with table jumps.
263
264 6. Inserted two casts and changed some ints to size_t in the light of some
265 reported 64-bit compiler warnings (Bugzilla 1477).
266
267 7. Fixed a bug concerned with zero-minimum possessive groups that could match
268 an empty string, which sometimes were behaving incorrectly in the
269 interpreter (though correctly in the JIT matcher). This pcretest input is
270 an example:
271
272 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
273 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
274
275 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
276 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
277 for an empty string was breaking the inner loop and carrying on at a lower
278 level, when possessive repeated groups should always return to a higher
279 level as they have no backtrack points in them. The empty string test now
280 occurs at the outer level.
281
282 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
283 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
284
285 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
286 doesn't).
287
288 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
289 to hold a pattern to be incorrectly computed (too small) when there were
290 named back references to duplicated names. This could cause "internal
291 error: code overflow" or "double free or corruption" or other memory
292 handling errors.
293
294 11. When named subpatterns had the same prefixes, back references could be
295 confused. For example, in this pattern:
296
297 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
298
299 the reference to 'Name' was incorrectly treated as a reference to a
300 duplicate name.
301
302 12. A pattern such as /^s?c/mi8 where the optional character has more than
303 one "other case" was incorrectly compiled such that it would only try to
304 match starting at "c".
305
306 13. When a pattern starting with \s was studied, VT was not included in the
307 list of possible starting characters; this should have been part of the
308 8.34/18 patch.
309
310 14. If a character class started [\Qx]... where x is any character, the class
311 was incorrectly terminated at the ].
312
313 15. If a pattern that started with a caseless match for a character with more
314 than one "other case" was studied, PCRE did not set up the starting code
315 unit bit map for the list of possible characters. Now it does. This is an
316 optimization improvement, not a bug fix.
317
318 16. The Unicode data tables have been updated to Unicode 7.0.0.
319
320 17. Fixed a number of memory leaks in pcregrep.
321
322 18. Avoid a compiler warning (from some compilers) for a function call with
323 a cast that removes "const" from an lvalue by using an intermediate
324 variable (to which the compiler does not object).
325
326 19. Incorrect code was compiled if a group that contained an internal recursive
327 back reference was optional (had quantifier with a minimum of zero). This
328 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
329 caused segmentation faults because of stack overflows at compile time.
330
331 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
332 group that is quantified with an indefinite repeat, caused a compile-time
333 loop which used up all the system stack and provoked a segmentation fault.
334 This was not the same bug as 19 above.
335
336 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
337 Patch by Mike Frysinger.
338
339
340 Version 8.35 04-April-2014
341 --------------------------
342
343 1. A new flag is set, when property checks are present in an XCLASS.
344 When this flag is not set, PCRE can perform certain optimizations
345 such as studying these XCLASS-es.
346
347 2. The auto-possessification of character sets were improved: a normal
348 and an extended character set can be compared now. Furthermore
349 the JIT compiler optimizes more character set checks.
350
351 3. Got rid of some compiler warnings for potentially uninitialized variables
352 that show up only when compiled with -O2.
353
354 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
355 of a match later then the end of the match. The pcretest program was not
356 handling the case sensibly - it was outputting from the start to the next
357 binary zero. It now reports this situation in a message, and outputs the
358 text from the end to the start.
359
360 5. Fast forward search is improved in JIT. Instead of the first three
361 characters, any three characters with fixed position can be searched.
362 Search order: first, last, middle.
363
364 6. Improve character range checks in JIT. Characters are read by an inprecise
365 function now, which returns with an unknown value if the character code is
366 above a certain threshold (e.g: 256). The only limitation is that the value
367 must be bigger than the threshold as well. This function is useful when
368 the characters above the threshold are handled in the same way.
369
370 7. The macros whose names start with RAWUCHAR are placeholders for a future
371 mode in which only the bottom 21 bits of 32-bit data items are used. To
372 make this more memorable for those maintaining the code, the names have
373 been changed to start with UCHAR21, and an extensive comment has been added
374 to their definition.
375
376 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
377 to the export list in Makefile.am (they were accidentally omitted from the
378 8.34 tarball).
379
380 9. The informational output from pcretest used the phrase "starting byte set"
381 which is inappropriate for the 16-bit and 32-bit libraries. As the output
382 for "first char" and "need char" really means "non-UTF-char", I've changed
383 "byte" to "char", and slightly reworded the output. The documentation about
384 these values has also been (I hope) clarified.
385
386 10. Another JIT related optimization: use table jumps for selecting the correct
387 backtracking path, when more than four alternatives are present inside a
388 bracket.
389
390 11. Empty match is not possible, when the minimum length is greater than zero,
391 and there is no \K in the pattern. JIT should avoid empty match checks in
392 such cases.
393
394 12. In a caseless character class with UCP support, when a character with more
395 than one alternative case was not the first character of a range, not all
396 the alternative cases were added to the class. For example, s and \x{17f}
397 are both alternative cases for S: the class [RST] was handled correctly,
398 but [R-T] was not.
399
400 13. The configure.ac file always checked for pthread support when JIT was
401 enabled. This is not used in Windows, so I have put this test inside a
402 check for the presence of windows.h (which was already tested for).
403
404 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
405 The algorithm provides a way to skip certain starting offsets, and usually
406 faster than linear prefix searches.
407
408 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
409 as for 'fr_FR' and 'french'. For some reason, however, it then used the
410 Windows-specific input and output files, which have 'french' screwed in.
411 So this could never have worked. One of the problems with locales is that
412 they aren't always the same. I have now updated RunTest so that it checks
413 the output of the locale test (test 3) against three different output
414 files, and it allows the test to pass if any one of them matches. With luck
415 this should make the test pass on some versions of Solaris where it was
416 failing. Because of the uncertainty, the script did not used to stop if
417 test 3 failed; it now does. If further versions of a French locale ever
418 come to light, they can now easily be added.
419
420 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
421 there was a message during ./configure, but it did not stop. This now
422 provokes an error. The invalid example in README has been corrected.
423 If a value less than the minimum is given, the minimum value has always
424 been used, but now a warning is given.
425
426 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
427 was a bug in the test system, which is now fixed. Also, the list of various
428 configurations that are tested for each release did not have one with both
429 16/32 bits and --enable-bar-anycrlf. It now does.
430
431 18. pcretest was missing "-C bsr" for displaying the \R default setting.
432
433 19. Little endian PowerPC systems are supported now by the JIT compiler.
434
435 20. The fast forward newline mechanism could enter to an infinite loop on
436 certain invalid UTF-8 input. Although we don't support these cases
437 this issue can be fixed by a performance optimization.
438
439 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
440 not take account if existing stack usage. There is now a new global
441 variable called pcre_stack_guard that can be set to point to an external
442 function to check stack availability. It is called at the start of
443 processing every parenthesized group.
444
445 22. A typo in the code meant that in ungreedy mode the max/min qualifier
446 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
447 not match "ab".
448
449 23. When UTF was disabled, the JIT program reported some incorrect compile
450 errors. These messages are silenced now.
451
452 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
453 compiler.
454
455 25. Change all the temporary files used in RunGrepTest to be different to those
456 used by RunTest so that the tests can be run simultaneously, for example by
457 "make -j check".
458
459
460 Version 8.34 15-December-2013
461 -----------------------------
462
463 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
464 executable memory. Patch inspired by Carsten Klein.
465
466 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
467 this macro is never tested and has no effect, because the work to support
468 coverage involves only compiling and linking options and special targets in
469 the Makefile. The comment in config.h implied that defining the macro would
470 enable coverage support, which is totally false. There was also support for
471 setting this macro in the CMake files (my fault, I just copied it from
472 configure). SUPPORT_GCOV has now been removed.
473
474 3. Make a small performance improvement in strlen16() and strlen32() in
475 pcretest.
476
477 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
478 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
479 these cases). There was also one in pcretest.c.
480
481 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
482
483 6. In UTF mode, the code for checking whether a group could match an empty
484 string (which is used for indefinitely repeated groups to allow for
485 breaking an infinite loop) was broken when the group contained a repeated
486 negated single-character class with a character that occupied more than one
487 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
488 UTF-8 mode). The effect was undefined: the group might or might not be
489 deemed as matching an empty string, or the program might have crashed.
490
491 7. The code for checking whether a group could match an empty string was not
492 recognizing that \h, \H, \v, \V, and \R must match a character.
493
494 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
495 an empty string. If it can, pcretest shows this in its information output.
496
497 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
498 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
499 matched by pcre_exec() without using JIT:
500
501 (a) If the rest of the pattern did not match after a maximal run of
502 grapheme clusters, the code for backing up to try with fewer of them
503 did not always back up over a full grapheme when characters that do not
504 have the modifier quality were involved, e.g. Hangul syllables.
505
506 (b) If the match point in a subject started with modifier character, and
507 there was no match, the code could incorrectly back up beyond the match
508 point, and potentially beyond the first character in the subject,
509 leading to a segfault or an incorrect match result.
510
511 10. A conditional group with an assertion condition could lead to PCRE
512 recording an incorrect first data item for a match if no other first data
513 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
514 first data item, and therefore matched "ca" after "c" instead of at the
515 start.
516
517 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
518 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
519
520 12. The source of pcregrep now includes z/OS-specific code so that it can be
521 compiled for z/OS as part of the special z/OS distribution.
522
523 13. Added the -T and -TM options to pcretest.
524
525 14. The code in pcre_compile.c for creating the table of named capturing groups
526 has been refactored. Instead of creating the table dynamically during the
527 actual compiling pass, the information is remembered during the pre-compile
528 pass (on the stack unless there are more than 20 named groups, in which
529 case malloc() is used) and the whole table is created before the actual
530 compile happens. This has simplified the code (it is now nearly 150 lines
531 shorter) and prepared the way for better handling of references to groups
532 with duplicate names.
533
534 15. A back reference to a named subpattern when there is more than one of the
535 same name now checks them in the order in which they appear in the pattern.
536 The first one that is set is used for the reference. Previously only the
537 first one was inspected. This change makes PCRE more compatible with Perl.
538
539 16. Unicode character properties were updated from Unicode 6.3.0.
540
541 17. The compile-time code for auto-possessification has been refactored, based
542 on a patch by Zoltan Herczeg. It now happens after instead of during
543 compilation. The code is cleaner, and more cases are handled. The option
544 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
545 options in pcretest are provided to set it. It can also be set by
546 (*NO_AUTO_POSSESS) at the start of a pattern.
547
548 18. The character VT has been added to the default ("C" locale) set of
549 characters that match \s and are generally treated as white space,
550 following this same change in Perl 5.18. There is now no difference between
551 "Perl space" and "POSIX space". Whether VT is treated as white space in
552 other locales depends on the locale.
553
554 19. The code for checking named groups as conditions, either for being set or
555 for being recursed, has been refactored (this is related to 14 and 15
556 above). Processing unduplicated named groups should now be as fast at
557 numerical groups, and processing duplicated groups should be faster than
558 before.
559
560 20. Two patches to the CMake build system, by Alexander Barkov:
561
562 (1) Replace the "source" command by "." in CMakeLists.txt because
563 "source" is a bash-ism.
564
565 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
566 without these the CMake build does not work on Solaris.
567
568 21. Perl has changed its handling of \8 and \9. If there is no previously
569 encountered capturing group of those numbers, they are treated as the
570 literal characters 8 and 9 instead of a binary zero followed by the
571 literals. PCRE now does the same.
572
573 22. Following Perl, added \o{} to specify codepoints in octal, making it
574 possible to specify values greater than 0777 and also making them
575 unambiguous.
576
577 23. Perl now gives an error for missing closing braces after \x{... instead of
578 treating the string as literal. PCRE now does the same.
579
580 24. RunTest used to grumble if an inappropriate test was selected explicitly,
581 but just skip it when running all tests. This make it awkward to run ranges
582 of tests when one of them was inappropriate. Now it just skips any
583 inappropriate tests, as it always did when running all tests.
584
585 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
586 character types such as \d or \w, too many callouts were inserted, and the
587 data that they returned was rubbish.
588
589 26. In UCP mode, \s was not matching two of the characters that Perl matches,
590 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
591 were matched by \h. The code has now been refactored so that the lists of
592 the horizontal and vertical whitespace characters used for \h and \v (which
593 are defined only in one place) are now also used for \s.
594
595 27. Add JIT support for the 64 bit TileGX architecture.
596 Patch by Jiong Wang (Tilera Corporation).
597
598 28. Possessive quantifiers for classes (both explicit and automatically
599 generated) now use special opcodes instead of wrapping in ONCE brackets.
600
601 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
602 (because it's meaningless), this was not happening when PCRE_CASELESS was
603 set. Not wrong, but inefficient.
604
605 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
606 properties for \w, \d, etc) is present in a test regex. Otherwise if the
607 test contains no characters greater than 255, Perl doesn't realise it
608 should be using Unicode semantics.
609
610 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
611 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
612 does in Unicode mode.
613
614 32. Added the "forbid" facility to pcretest so that putting tests into the
615 wrong test files can sometimes be quickly detected.
616
617 33. There is now a limit (default 250) on the depth of nesting of parentheses.
618 This limit is imposed to control the amount of system stack used at compile
619 time. It can be changed at build time by --with-parens-nest-limit=xxx or
620 the equivalent in CMake.
621
622 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
623 errors. Perl warns for these when in warning mode, but PCRE has no facility
624 for giving warnings.
625
626 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
627 However, this was not working for (?!) because it is optimized to (*FAIL),
628 for which PCRE does not allow quantifiers. The optimization is now disabled
629 when a quantifier follows (?!). I can't see any use for this, but it makes
630 things uniform.
631
632 36. Perl no longer allows group names to start with digits, so I have made this
633 change also in PCRE. It simplifies the code a bit.
634
635 37. In extended mode, Perl ignores spaces before a + that indicates a
636 possessive quantifier. PCRE allowed a space before the quantifier, but not
637 before the possessive +. It now does.
638
639 38. The use of \K (reset reported match start) within a repeated possessive
640 group such as (a\Kb)*+ was not working.
641
642 40. Document that the same character tables must be used at compile time and
643 run time, and that the facility to pass tables to pcre_exec() and
644 pcre_dfa_exec() is for use only with saved/restored patterns.
645
646 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
647 features for Builds with MSVC:
648
649 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
650 builds). The .rc files can be used to set FileDescription and many other
651 attributes.
652
653 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
654 This allows higher-level build scripts which want .pdb files to avoid
655 hard-coding the exact files needed."
656
657 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
658 mean "start of word" and "end of word", respectively, as a transition aid.
659
660 43. A minimizing repeat of a class containing codepoints greater than 255 in
661 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
662 compiled to use the heap for recursion.
663
664 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
665 is configured.
666
667
668 Version 8.33 28-May-2013
669 ------------------------
670
671 1. Added 'U' to some constants that are compared to unsigned integers, to
672 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
673 variables that are added to signed variables, to ensure the result is
674 signed and can be negated.
675
676 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
677 CMake config files.
678
679 3. Revise the creation of config.h.generic so that all boolean macros are
680 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
681 overriding via -D on the command line possible.
682
683 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
684 to unsigned int is reported to make a quite noticeable speed difference in
685 a specific Windows environment. Testing on Linux did also appear to show
686 some benefit (and it is clearly not harmful). Also fixed the definition of
687 Xop which should be unsigned.
688
689 5. Related to (4), changing the definition of the intermediate variable cc
690 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
691 improvements.
692
693 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
694 unnecessary spaces.
695
696 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
697 and later.
698
699 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
700
701 9. Optimizing fast_forward_start_bits in JIT.
702
703 10. Adding support for callouts in JIT, and fixing some issues revealed
704 during this work. Namely:
705
706 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
707
708 (b) Minimum length was not checked before the matching is started.
709
710 11. The value of capture_last that is passed to callouts was incorrect in some
711 cases when there was a capture on one path that was subsequently abandoned
712 after a backtrack. Also, the capture_last value is now reset after a
713 recursion, since all captures are also reset in this case.
714
715 12. The interpreter no longer returns the "too many substrings" error in the
716 case when an overflowing capture is in a branch that is subsequently
717 abandoned after a backtrack.
718
719 13. In the pathological case when an offset vector of size 2 is used, pcretest
720 now prints out the matched string after a yield of 0 or 1.
721
722 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
723 Only supported by the JIT compiler at the moment.
724
725 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
726
727 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
728 offset of the starting point of the matching process, provided the offsets
729 vector is large enough.
730
731 17. The \A escape now records a lookbehind value of 1, though its execution
732 does not actually inspect the previous character. This is to ensure that,
733 in partial multi-segment matching, at least one character from the old
734 segment is retained when a new segment is processed. Otherwise, if there
735 are no lookbehinds in the pattern, \A might match incorrectly at the start
736 of a new segment.
737
738 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
739
740 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
741 gives some modest performance improvement in 8-bit mode.
742
743 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
744 be expressed in certain programming languages using Universal Character
745 Names.
746
747 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
748 which points out that "non characters" are not "characters that may not
749 appear in Unicode strings" but rather "characters that are reserved for
750 internal use and have only local meaning".
751
752 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
753 there was a conditional group that depended on an assertion, if the
754 assertion was false, the callout that immediately followed the alternation
755 in the condition was skipped when pcre_exec() was used for matching.
756
757 23. Allow an explicit callout to be inserted before an assertion that is the
758 condition for a conditional group, for compatibility with automatic
759 callouts, which always insert a callout at this point.
760
761 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
762 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
763
764 25. (*PRUNE) is now supported by the JIT compiler.
765
766 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
767
768 27. Fix the case where there are two or more SKIPs with arguments that may be
769 ignored.
770
771 28. (*SKIP) is now supported by the JIT compiler.
772
773 29. (*THEN) is now supported by the JIT compiler.
774
775 30. Update RunTest with additional test selector options.
776
777 31. The way PCRE handles backtracking verbs has been changed in two ways.
778
779 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
780 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
781 backtracking. In some cases this makes it more Perl-compatible, but Perl's
782 rather obscure rules do not always do the same thing.
783
784 (2) Previously, backtracking verbs were confined within assertions. This is
785 no longer the case for positive assertions, except for (*ACCEPT). Again,
786 this sometimes improves Perl compatibility, and sometimes does not.
787
788 32. A number of tests that were in test 2 because Perl did things differently
789 have been moved to test 1, because either Perl or PCRE has changed, and
790 these tests are now compatible.
791
792 32. Backtracking control verbs are now handled in the same way in JIT and
793 interpreter.
794
795 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
796 contained a forward subroutine reference caused a compile error.
797
798 34. Auto-detect and optimize limited repetitions in JIT.
799
800 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
801 blocking (*UTF) etc.
802
803 36. In the interpreter, maximizing pattern repetitions for characters and
804 character types now use tail recursion, which reduces stack usage.
805
806 37. The value of the max lookbehind was not correctly preserved if a compiled
807 and saved regex was reloaded on a host of different endianness.
808
809 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
810 of the compiled pattern block, expand the flags field from 16 to 32 bits
811 because it was almost full.
812
813 39. Try madvise first before posix_madvise.
814
815 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
816 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
817 original change locked out all matches of empty strings. This has been
818 changed so that one match of an empty string per line is recognized.
819 Subsequent searches on the same line (for colouring or for --only-matching,
820 for example) do not recognize empty strings.
821
822 41. Applied a user patch to fix a number of spelling mistakes in comments.
823
824 42. Data lines longer than 65536 caused pcretest to crash.
825
826 43. Clarified the data type for length and startoffset arguments for pcre_exec
827 and pcre_dfa_exec in the function-specific man pages, where they were
828 explicitly stated to be in bytes, never having been updated. I also added
829 some clarification to the pcreapi man page.
830
831 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
832 a segmentation fault.
833
834
835 Version 8.32 30-November-2012
836 -----------------------------
837
838 1. Improved JIT compiler optimizations for first character search and single
839 character iterators.
840
841 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
842 Patch by Daniel Richard G.
843
844 3. Single character iterator optimizations in the JIT compiler.
845
846 4. Improved JIT compiler optimizations for character ranges.
847
848 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
849 Reported by Giuseppe D'Angelo.
850
851 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
852 of a line, was being set incorrectly in cases where .* appeared inside
853 atomic brackets at the start of a pattern, or where there was a subsequent
854 *PRUNE or *SKIP.
855
856 7. Improved instruction cache flush for POWER/PowerPC.
857 Patch by Daniel Richard G.
858
859 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
860 grep:
861
862 (a) There is now no limit to the number of patterns to be matched.
863
864 (b) An error is given if a pattern is too long.
865
866 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
867 are now supported.
868
869 (d) --exclude-from and --include-from (multiple use) have been added.
870
871 (e) Exclusions and inclusions now apply to all files and directories, not
872 just to those obtained from scanning a directory recursively.
873
874 (f) Multiple uses of -f and --file-list are now supported.
875
876 (g) In a Windows environment, the default for -d has been changed from
877 "read" (the GNU grep default) to "skip", because otherwise the presence
878 of a directory in the file list provokes an error.
879
880 (h) The documentation has been revised and clarified in places.
881
882 9. Improve the matching speed of capturing brackets.
883
884 10. Changed the meaning of \X so that it now matches a Unicode extended
885 grapheme cluster.
886
887 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
888 out POSIX threads when JIT support is configured.
889
890 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
891
892 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
893 structure could go wrong in environments where size_t is not the same size
894 as int.
895
896 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
897 set.
898
899 15. The EBCDIC support had decayed; later updates to the code had included
900 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
901 general tidy up of EBCDIC-related issues, and the documentation was also
902 not quite right. There is now a test that can be run on ASCII systems to
903 check some of the EBCDIC-related things (but is it not a full test).
904
905 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
906 in a small tidy to the code.
907
908 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
909
910 18. If the --only-matching (-o) option in pcregrep is specified multiple
911 times, each one causes appropriate output. For example, -o1 -o2 outputs the
912 substrings matched by the 1st and 2nd capturing parentheses. A separating
913 string can be specified by --om-separator (default empty).
914
915 19. Improving the first n character searches.
916
917 20. Turn case lists for horizontal and vertical white space into macros so that
918 they are defined only once.
919
920 21. This set of changes together give more compatible Unicode case-folding
921 behaviour for characters that have more than one other case when UCP
922 support is available.
923
924 (a) The Unicode property table now has offsets into a new table of sets of
925 three or more characters that are case-equivalent. The MultiStage2.py
926 script that generates these tables (the pcre_ucd.c file) now scans
927 CaseFolding.txt instead of UnicodeData.txt for character case
928 information.
929
930 (b) The code for adding characters or ranges of characters to a character
931 class has been abstracted into a generalized function that also handles
932 case-independence. In UTF-mode with UCP support, this uses the new data
933 to handle characters with more than one other case.
934
935 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
936 whose other case is greater than 256 are now correctly matched
937 caselessly. Previously, the high codepoint matched the low one, but not
938 vice versa.
939
940 (d) The processing of \h, \H, \v, and \ in character classes now makes use
941 of the new class addition function, using character lists defined as
942 macros alongside the case definitions of 20 above.
943
944 (e) Caseless back references now work with characters that have more than
945 one other case.
946
947 (f) General caseless matching of characters with more than one other case
948 is supported.
949
950 22. Unicode character properties were updated from Unicode 6.2.0
951
952 23. Improved CMake support under Windows. Patch by Daniel Richard G.
953
954 24. Add support for 32-bit character strings, and UTF-32
955
956 25. Major JIT compiler update (code refactoring and bugfixing).
957 Experimental Sparc 32 support is added.
958
959 26. Applied a modified version of Daniel Richard G's patch to create
960 pcre.h.generic and config.h.generic by "make" instead of in the
961 PrepareRelease script.
962
963 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
964 pcre_compile.c when checking for a zero character.
965
966 28. Introducing a native interface for JIT. Through this interface, the compiled
967 machine code can be directly executed. The purpose of this interface is to
968 provide fast pattern matching, so several sanity checks are not performed.
969 However, feature tests are still performed. The new interface provides
970 1.4x speedup compared to the old one.
971
972 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
973 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
974 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
975
976 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
977 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
978 the "old" RFC 2279). Instead, it ended up passing a negative length to
979 pcre_exec().
980
981 31. Add support for GCC's visibility feature to hide internal functions.
982
983 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
984 "unknown -C option" after outputting 0 or 1.
985
986 33. There is now support for generating a code coverage report for the test
987 suite in environments where gcc is the compiler and lcov is installed. This
988 is mainly for the benefit of the developers.
989
990 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
991 unaddressable using valgrind annotations, allowing valgrind to detect
992 invalid memory accesses. This is mainly for the benefit of the developers.
993
994 25. (*UTF) can now be used to start a pattern in any of the three libraries.
995
996 26. Give configure error if --enable-cpp but no C++ compiler found.
997
998
999 Version 8.31 06-July-2012
1000 -------------------------
1001
1002 1. Fixing a wrong JIT test case and some compiler warnings.
1003
1004 2. Removed a bashism from the RunTest script.
1005
1006 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1007 to unsigned type, result still unsigned" that was given by an MS compiler
1008 on encountering the code "-sizeof(xxx)".
1009
1010 4. Partial matching support is added to the JIT compiler.
1011
1012 5. Fixed several bugs concerned with partial matching of items that consist
1013 of more than one character:
1014
1015 (a) /^(..)\1/ did not partially match "aba" because checking references was
1016 done on an "all or nothing" basis. This also applied to repeated
1017 references.
1018
1019 (b) \R did not give a hard partial match if \r was found at the end of the
1020 subject.
1021
1022 (c) \X did not give a hard partial match after matching one or more
1023 characters at the end of the subject.
1024
1025 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1026 a partial match for the string "\r".
1027
1028 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1029 a partial match for a CR character at the end of the subject string.
1030
1031 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1032 running pcretest, the text "(JIT)" added to the output whenever JIT is
1033 actually used to run the match.
1034
1035 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1036 or /S+[+] with a digit between 1 and 7.
1037
1038 8. OP_NOT now supports any UTF character not just single-byte ones.
1039
1040 9. (*MARK) control verb is now supported by the JIT compiler.
1041
1042 10. The command "./RunTest list" lists the available tests without actually
1043 running any of them. (Because I keep forgetting what they all are.)
1044
1045 11. Add PCRE_INFO_MAXLOOKBEHIND.
1046
1047 12. Applied a (slightly modified) user-supplied patch that improves performance
1048 when the heap is used for recursion (compiled with --disable-stack-for-
1049 recursion). Instead of malloc and free for each heap frame each time a
1050 logical recursion happens, frames are retained on a chain and re-used where
1051 possible. This sometimes gives as much as 30% improvement.
1052
1053 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1054 call.
1055
1056 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1057
1058 15. It is now possible to link pcretest with libedit as an alternative to
1059 libreadline.
1060
1061 16. (*COMMIT) control verb is now supported by the JIT compiler.
1062
1063 17. The Unicode data tables have been updated to Unicode 6.1.0.
1064
1065 18. Added --file-list option to pcregrep.
1066
1067 19. Added binary file support to pcregrep, including the -a, --binary-files,
1068 -I, and --text options.
1069
1070 20. The madvise function is renamed for posix_madvise for QNX compatibility
1071 reasons. Fixed by Giuseppe D'Angelo.
1072
1073 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1074
1075 22. Changed the option for creating symbolic links for 16-bit man pages from
1076 -s to -sf so that re-installing does not cause issues.
1077
1078 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1079
1080 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1081 matches in certain environments (the workspace was not being correctly
1082 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1083 some of the workspace data at the beginning of a restart.
1084
1085 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1086 was not doing so when it should - probably a typo introduced by SVN 528
1087 (change 8.10/14).
1088
1089 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1090 \w+ when the character tables indicated that \x{c4} was a word character.
1091 There were several related cases, all because the tests for doing a table
1092 lookup were testing for characters less than 127 instead of 255.
1093
1094 27. If a pattern contains capturing parentheses that are not used in a match,
1095 their slots in the ovector are set to -1. For those that are higher than
1096 any matched groups, this happens at the end of processing. In the case when
1097 there were back references that the ovector was too small to contain
1098 (causing temporary malloc'd memory to be used during matching), and the
1099 highest capturing number was not used, memory off the end of the ovector
1100 was incorrectly being set to -1. (It was using the size of the temporary
1101 memory instead of the true size.)
1102
1103 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1104 ovector size, it uses memory at the end of the block that it has got.
1105
1106 29. Check for an overlong MARK name and give an error at compile time. The
1107 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1108
1109 30. JIT compiler update.
1110
1111 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1112 Rill for the patch.
1113
1114 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1115
1116 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1117
1118 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1119 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1120 was enabled.
1121
1122 35. Improve JIT code generation for greedy plus quantifier.
1123
1124 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1125 1 to "aa" instead of to an empty string. The bug affected repeated groups
1126 that could potentially match an empty string.
1127
1128 37. Optimizing single character iterators in JIT.
1129
1130 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1131 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1132 codepoints that are too big for the mode are faulted, and in a UTF mode,
1133 disallowed codepoints are also faulted.
1134
1135 39. If PCRE was compiled with UTF support, in three places in the DFA
1136 matcher there was code that should only have been obeyed in UTF mode, but
1137 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1138 processing when bytes with values greater than 127 were present. In 16-bit
1139 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1140 both cases the values are those that cannot be the first data item in a UTF
1141 character. The three items that might have provoked this were recursions,
1142 possessively repeated groups, and atomic groups.
1143
1144 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1145 and pcregrep, because some OS require shared objects to be explicitly
1146 passed to ld, causing the link step to fail if they are not.
1147
1148 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1149 mode, patterns that started with \h* or \R* might be incorrectly matched.
1150
1151
1152 Version 8.30 04-February-2012
1153 -----------------------------
1154
1155 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1156 name is defined in ctype.h.
1157
1158 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1159 only in quite long subpatterns.
1160
1161 3. Removed the function pcre_info(), which has been obsolete and deprecated
1162 since it was replaced by pcre_fullinfo() in February 2000.
1163
1164 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1165 match a (*MARK), and the match failed at the start of the subject, a
1166 reference to memory before the start of the subject could occur. This bug
1167 was introduced by fix 17 of release 8.21.
1168
1169 5. A reference to an unset group with zero minimum repetition was giving
1170 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1171 /(another)?(\1?)test/ matched against "hello world test". This bug was
1172 introduced in release 8.13.
1173
1174 6. Add support for 16-bit character strings (a large amount of work involving
1175 many changes and refactorings).
1176
1177 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1178 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1179 from a file.
1180
1181 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1182 rounding is not applied in this particular case).
1183
1184 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1185 if they appear, or are escaped, in patterns.
1186
1187 10. Get rid of a number of -Wunused-but-set-variable warnings.
1188
1189 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1190 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1191 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1192 also returns the mark "x". This bug applied to capturing parentheses,
1193 non-capturing parentheses, and atomic parentheses. It also applied to some
1194 assertions.
1195
1196 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1197 information out of configure.ac instead of relying on pcre.h.generic, which
1198 is not stored in the repository.
1199
1200 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1201 -lreadline.
1202
1203 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1204
1205 15. Applied Graycode's patch to put the top-level frame on the stack rather
1206 than the heap when not using the stack for recursion. This gives a
1207 performance improvement in many cases when recursion is not deep.
1208
1209 16. Experimental code added to "pcretest -C" to output the stack frame size.
1210
1211
1212 Version 8.21 12-Dec-2011
1213 ------------------------
1214
1215 1. Updating the JIT compiler.
1216
1217 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1218 are added as well.
1219
1220 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1221 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1222 calling _pcre_jit_exec. Some extra comments are added.
1223
1224 4. (*MARK) settings inside atomic groups that do not contain any capturing
1225 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1226 was introduced by change 18 for 8.20.
1227
1228 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1229 ECMA-262 standard.
1230
1231 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1232 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1233 This bug was probably introduced by change 9 of 8.13.
1234
1235 7. While fixing 6 above, I noticed that a number of other items were being
1236 incorrectly rejected as "not fixed length". This arose partly because newer
1237 opcodes had not been added to the fixed-length checking code. I have (a)
1238 corrected the bug and added tests for these items, and (b) arranged for an
1239 error to occur if an unknown opcode is encountered while checking for fixed
1240 length instead of just assuming "not fixed length". The items that were
1241 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1242 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1243 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1244
1245 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1246 being incorrectly compiled and would have given unpredicatble results.
1247
1248 9. A possessively repeated subpattern with minimum repeat count greater than
1249 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1250 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1251 the first (A) could occur when it should not.
1252
1253 10. Add a cast and remove a redundant test from the code.
1254
1255 11. JIT should use pcre_malloc/pcre_free for allocation.
1256
1257 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1258 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1259 is anything other than /usr, -L is still shown).
1260
1261 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1262
1263 14. Perl does not support \N without a following name in a [] class; PCRE now
1264 also gives an error.
1265
1266 15. If a forward reference was repeated with an upper limit of around 2000,
1267 it caused the error "internal error: overran compiling workspace". The
1268 maximum number of forward references (including repeats) was limited by the
1269 internal workspace, and dependent on the LINK_SIZE. The code has been
1270 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1271 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1272 of around 200,000 forward references. While doing this, I also speeded up
1273 the filling in of repeated forward references.
1274
1275 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1276 incorrectly expecting the subject to contain another "a" after the start.
1277
1278 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1279 in the match, the SKIP should be ignored. This was not happening; instead
1280 the SKIP was being treated as NOMATCH. For patterns such as
1281 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1282 tested.
1283
1284 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1285 now much more compatible with Perl, in particular in cases where the result
1286 is a non-match for a non-anchored pattern. For example, if
1287 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1288 "m", where previously it did not return a name. A side effect of this
1289 change is that for partial matches, the last encountered mark name is
1290 returned, as for non matches. A number of tests that were previously not
1291 Perl-compatible have been moved into the Perl-compatible test files. The
1292 refactoring has had the pleasing side effect of removing one argument from
1293 the match() function, thus reducing its stack requirements.
1294
1295 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1296 subsequent uses of /S (without +) incorrectly behaved like /S+.
1297
1298 21. Retrieve executable code size support for the JIT compiler and fixing
1299 some warnings.
1300
1301 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1302 not work when the shorter character appeared right at the end of the
1303 subject string.
1304
1305 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1306 systems.
1307
1308 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1309 output it when the /M option is used in pcretest.
1310
1311 25. The CheckMan script was not being included in the distribution. Also, added
1312 an explicit "perl" to run Perl scripts from the PrepareRelease script
1313 because this is reportedly needed in Windows.
1314
1315 26. If study data was being save in a file and studying had not found a set of
1316 "starts with" bytes for the pattern, the data written to the file (though
1317 never used) was taken from uninitialized memory and so caused valgrind to
1318 complain.
1319
1320 27. Updated RunTest.bat as provided by Sheri Pierce.
1321
1322 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1323
1324 29. Computation of memory usage for the table of capturing group names was
1325 giving an unnecessarily large value.
1326
1327
1328 Version 8.20 21-Oct-2011
1329 ------------------------
1330
1331 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1332 a POSIX class. After further experiments with Perl, which convinced me that
1333 Perl has bugs and confusions, a closing square bracket is no longer allowed
1334 in a POSIX name. This bug also affected patterns with classes that started
1335 with full stops.
1336
1337 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1338 captured substring, but while checking the failing first alternative,
1339 substring 1 is temporarily captured. If the output vector supplied to
1340 pcre_exec() was not big enough for this capture, the yield of the function
1341 was still zero ("insufficient space for captured substrings"). This cannot
1342 be totally fixed without adding another stack variable, which seems a lot
1343 of expense for a edge case. However, I have improved the situation in cases
1344 such as /(a)(b)x|abc/ matched against "abc", where the return code
1345 indicates that fewer than the maximum number of slots in the ovector have
1346 been set.
1347
1348 3. Related to (2) above: when there are more back references in a pattern than
1349 slots in the output vector, pcre_exec() uses temporary memory during
1350 matching, and copies in the captures as far as possible afterwards. It was
1351 using the entire output vector, but this conflicts with the specification
1352 that only 2/3 is used for passing back captured substrings. Now it uses
1353 only the first 2/3, for compatibility. This is, of course, another edge
1354 case.
1355
1356 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1357 main code base, and can be used by building with --enable-jit. When this is
1358 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1359 runtime --no-jit option is given.
1360
1361 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1362 ovector, the return from the function was zero, implying that there were
1363 other matches that did not fit. The correct "exactly full" value is now
1364 returned.
1365
1366 6. If a subpattern that was called recursively or as a subroutine contained
1367 (*PRUNE) or any other control that caused it to give a non-standard return,
1368 invalid errors such as "Error -26 (nested recursion at the same subject
1369 position)" or even infinite loops could occur.
1370
1371 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1372 computing the minimum length on reaching *ACCEPT, and so ended up with the
1373 wrong value of 1 rather than 0. Further investigation indicates that
1374 computing a minimum subject length in the presence of *ACCEPT is difficult
1375 (think back references, subroutine calls), and so I have changed the code
1376 so that no minimum is registered for a pattern that contains *ACCEPT.
1377
1378 8. If (*THEN) was present in the first (true) branch of a conditional group,
1379 it was not handled as intended. [But see 16 below.]
1380
1381 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1382 Sheri Pierce.
1383
1384 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1385 the first byte in a match must be "a".
1386
1387 11. Change 17 for 8.13 increased the recursion depth for patterns like
1388 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1389 pattern contains any instances of (*THEN). If it does not, the old
1390 optimizations are restored. It would be nice to do this on a per-group
1391 basis, but at the moment that is not feasible.
1392
1393 12. In some environments, the output of pcretest -C is CRLF terminated. This
1394 broke RunTest's code that checks for the link size. A single white space
1395 character after the value is now allowed for.
1396
1397 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1398 For "fr", it uses the Windows-specific input and output files.
1399
1400 14. If (*THEN) appeared in a group that was called recursively or as a
1401 subroutine, it did not work as intended. [But see next item.]
1402
1403 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1404 pattern fragments (but not containing any | characters). If A and B are
1405 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1406 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1407 D. In other words, Perl considers parentheses that do not contain any |
1408 characters to be part of a surrounding alternative, whereas PCRE was
1409 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1410 differently. PCRE now behaves in the same way as Perl, except in the case
1411 of subroutine/recursion calls such as (?1) which have in any case always
1412 been different (but PCRE had them first :-).
1413
1414 16. Related to 15 above: Perl does not treat the | in a conditional group as
1415 creating alternatives. Such a group is treated in the same way as an
1416 ordinary group without any | characters when processing (*THEN). PCRE has
1417 been changed to match Perl's behaviour.
1418
1419 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1420 RunGrepTest script failed.
1421
1422 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1423 inevitable for groups that contain captures, but it can lead to a lot of
1424 stack use in large patterns. The old behaviour has been restored for atomic
1425 groups that do not contain any capturing parentheses.
1426
1427 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1428 suppress the check for a minimum subject length at run time. (If it was
1429 given to pcre_exec() or pcre_dfa_exec() it did work.)
1430
1431 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1432 fail to work when decoding hex characters in data strings in EBCDIC
1433 environments.
1434
1435 21. It appears that in at least one Mac OS environment, the isxdigit() function
1436 is implemented as a macro that evaluates to its argument more than once,
1437 contravening the C 90 Standard (I haven't checked a later standard). There
1438 was an instance in pcretest which caused it to go wrong when processing
1439 \x{...} escapes in subject strings. The has been rewritten to avoid using
1440 things like p++ in the argument of isxdigit().
1441
1442
1443 Version 8.13 16-Aug-2011
1444 ------------------------
1445
1446 1. The Unicode data tables have been updated to Unicode 6.0.0.
1447
1448 2. Two minor typos in pcre_internal.h have been fixed.
1449
1450 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1451 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1452 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1453
1454 4. There were a number of related bugs in the code for matching backrefences
1455 caselessly in UTF-8 mode when codes for the characters concerned were
1456 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1457 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1458 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1459 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1460 2-byte code at the end of the subject (it thought there wasn't enough data
1461 left).
1462
1463 5. Comprehensive information about what went wrong is now returned by
1464 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1465 as the output vector has at least 2 elements. The offset of the start of
1466 the failing character and a reason code are placed in the vector.
1467
1468 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1469 now returned is for the first byte of the failing character, instead of the
1470 last byte inspected. This is an incompatible change, but I hope it is small
1471 enough not to be a problem. It makes the returned offset consistent with
1472 pcre_exec() and pcre_dfa_exec().
1473
1474 7. pcretest now gives a text phrase as well as the error number when
1475 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1476 failure, the offset and reason code are output.
1477
1478 8. When \R was used with a maximizing quantifier it failed to skip backwards
1479 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1480 back over a single character (\n). This seems wrong (because it treated the
1481 two characters as a single entity when going forwards), conflicts with the
1482 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1483 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1484 has been changed.
1485
1486 9. Some internal refactoring has changed the processing so that the handling
1487 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1488 time (the PCRE_DOTALL option was changed this way some time ago: version
1489 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1490 which was always a bit of a fudge. It also means that there is one less
1491 argument for the match() function, which reduces its stack requirements
1492 slightly. This change also fixes an incompatibility with Perl: the pattern
1493 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1494
1495 10. More internal refactoring has drastically reduced the number of recursive
1496 calls to match() for possessively repeated groups such as (abc)++ when
1497 using pcre_exec().
1498
1499 11. While implementing 10, a number of bugs in the handling of groups were
1500 discovered and fixed:
1501
1502 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1503 (a|)*(?1) gave a compile-time internal error.
1504 ((a|)+)+ did not notice that the outer group could match an empty string.
1505 (^a|^)+ was not marked as anchored.
1506 (.*a|.*)+ was not marked as matching at start or after a newline.
1507
1508 12. Yet more internal refactoring has removed another argument from the match()
1509 function. Special calls to this function are now indicated by setting a
1510 value in a variable in the "match data" data block.
1511
1512 13. Be more explicit in pcre_study() instead of relying on "default" for
1513 opcodes that mean there is no starting character; this means that when new
1514 ones are added and accidentally left out of pcre_study(), testing should
1515 pick them up.
1516
1517 14. The -s option of pcretest has been documented for ages as being an old
1518 synonym of -m (show memory usage). I have changed it to mean "force study
1519 for every regex", that is, assume /S for every regex. This is similar to -i
1520 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1521 using it. It makes it easier to run collections of tests with and without
1522 study enabled, and thereby test pcre_study() more easily. All the standard
1523 tests are now run with and without -s (but some patterns can be marked as
1524 "never study" - see 20 below).
1525
1526 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1527 restoration of the capturing data to the outer values was not happening
1528 correctly.
1529
1530 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1531 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1532 pattern had matched an empty string, and so incorrectly returned a no
1533 match.
1534
1535 17. There was optimizing code for the last branch of non-capturing parentheses,
1536 and also for the obeyed branch of a conditional subexpression, which used
1537 tail recursion to cut down on stack usage. Unfortunately, now that there is
1538 the possibility of (*THEN) occurring in these branches, tail recursion is
1539 no longer possible because the return has to be checked for (*THEN). These
1540 two optimizations have therefore been removed. [But see 8.20/11 above.]
1541
1542 18. If a pattern containing \R was studied, it was assumed that \R always
1543 matched two bytes, thus causing the minimum subject length to be
1544 incorrectly computed because \R can also match just one byte.
1545
1546 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1547 was incorrectly computed.
1548
1549 20. If /S is present twice on a test pattern in pcretest input, it now
1550 *disables* studying, thereby overriding the use of -s on the command line
1551 (see 14 above). This is necessary for one or two tests to keep the output
1552 identical in both cases.
1553
1554 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1555 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1556
1557 22. When an atomic group that contained a capturing parenthesis was
1558 successfully matched, but the branch in which it appeared failed, the
1559 capturing was not being forgotten if a higher numbered group was later
1560 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1561 group 1 to "a", when in fact it should be unset. This applied to multi-
1562 branched capturing and non-capturing groups, repeated or not, and also to
1563 positive assertions (capturing in negative assertions does not happen
1564 in PCRE) and also to nested atomic groups.
1565
1566 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1567 subject after a captured substring, to make it easier to tell which of a
1568 number of identical substrings has been captured.
1569
1570 24. The way atomic groups are processed by pcre_exec() has been changed so that
1571 if they are repeated, backtracking one repetition now resets captured
1572 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1573 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1574 "aaa". Previously, it would have been "a". As part of this code
1575 refactoring, the way recursive calls are handled has also been changed.
1576
1577 25. If an assertion condition captured any substrings, they were not passed
1578 back unless some other capturing happened later. For example, if
1579 (?(?=(a))a) was matched against "a", no capturing was returned.
1580
1581 26. When studying a pattern that contained subroutine calls or assertions,
1582 the code for finding the minimum length of a possible match was handling
1583 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1584 group 1 called group 2 while simultaneously a separate group 2 called group
1585 1). A stack overflow occurred in this case. I have fixed this by limiting
1586 the recursion depth to 10.
1587
1588 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1589 Fortmann. This supports explicit test numbers on the command line, and has
1590 argument validation and error reporting.
1591
1592 28. An instance of \X with an unlimited repeat could fail if at any point the
1593 first character it looked at was a mark character.
1594
1595 29. Some minor code refactoring concerning Unicode properties and scripts
1596 should reduce the stack requirement of match() slightly.
1597
1598 30. Added the '=' option to pcretest to check the setting of unused capturing
1599 slots at the end of the pattern, which are documented as being -1, but are
1600 not included in the return count.
1601
1602 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1603 compiled something random. Now it gives a compile-time error (as does
1604 Perl).
1605
1606 32. A *MARK encountered during the processing of a positive assertion is now
1607 recorded and passed back (compatible with Perl).
1608
1609 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1610 had alternative anchored branches, the search for a second match in a line
1611 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1612 matched the line "0102" twice. The same bug affected patterns that started
1613 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1614 twice.
1615
1616 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1617 does, and because of capturing effects, quantifying parenthesized
1618 assertions may at times be useful. Quantifiers are now allowed for
1619 parenthesized assertions.
1620
1621 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1622
1623 36. \g was being checked for fancy things in a character class, when it should
1624 just be a literal "g".
1625
1626 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1627 appearance of a nested POSIX class supersedes an apparent external class.
1628 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1629 unescaped square brackets may also appear as part of class names. For
1630 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1631 more like Perl. (But see 8.20/1 above.)
1632
1633 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1634 was because it thought it was \N{name}, which is not supported).
1635
1636 39. Add minix to OS list not supporting the -S option in pcretest.
1637
1638 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1639 cannot analyze patterns in sufficient detail to catch mutual recursions
1640 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1641 subgroup is called recursively as a subpattern for a second time at the
1642 same position in the subject string. In previous releases this might have
1643 been caught by the recursion limit, or it might have run out of stack.
1644
1645 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1646 happen only once. PCRE was, however incorrectly giving a compile time error
1647 "recursive call could loop indefinitely" because it cannot analyze the
1648 pattern in sufficient detail. The compile time test no longer happens when
1649 PCRE is compiling a conditional subpattern, but actual runaway loops are
1650 now caught at runtime (see 40 above).
1651
1652 42. It seems that Perl allows any characters other than a closing parenthesis
1653 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1654 has been changed to be the same.
1655
1656 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1657 as not to get warnings when autogen.sh is called. Also changed
1658 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1659
1660 44. To help people who use pcregrep to scan files containing exceedingly long
1661 lines, the following changes have been made:
1662
1663 (a) The default value of the buffer size parameter has been increased from
1664 8K to 20K. (The actual buffer used is three times this size.)
1665
1666 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1667 PCRE is built.
1668
1669 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1670 to be set at run time.
1671
1672 (d) Numerical values in pcregrep options can be followed by K or M, for
1673 example --buffer-size=50K.
1674
1675 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1676 given and the return code is set to 2.
1677
1678 45. Add a pointer to the latest mark to the callout data block.
1679
1680 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1681 partial match of an empty string instead of no match. This was specific to
1682 the use of ".".
1683
1684 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1685 complete match instead of a partial match. This bug was dependent on both
1686 the PCRE_UTF8 and PCRE_DOTALL options being set.
1687
1688 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1689 starting byte set, because \b was not being ignored.
1690
1691
1692 Version 8.12 15-Jan-2011
1693 ------------------------
1694
1695 1. Fixed some typos in the markup of the man pages, and wrote a script that
1696 checks for such things as part of the documentation building process.
1697
1698 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1699 --match-limit and --recursion-limit options (added for 8.11). In
1700 particular, this made one of the standard tests fail. (The integer value
1701 went into the wrong half of a long int.)
1702
1703 3. If the --colour option was given to pcregrep with -v (invert match), it
1704 did strange things, either producing crazy output, or crashing. It should,
1705 of course, ignore a request for colour when reporting lines that do not
1706 match.
1707
1708 4. Another pcregrep bug caused similar problems if --colour was specified with
1709 -M (multiline) and the pattern match finished with a line ending.
1710
1711 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1712 matched in multiline mode, the following line was shown as part of the
1713 match. This seems wrong, so I have changed it.
1714
1715 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1716 the check for further matches in the same line (so they could be coloured)
1717 to overrun the end of the current line. If another match was found, it was
1718 incorrectly shown (and then shown again when found in the next line).
1719
1720 7. If pcregrep was compiled under Windows, there was a reference to the
1721 function pcregrep_exit() before it was defined. I am assuming this was
1722 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1723 reported by a user. I've moved the definition above the reference.
1724
1725
1726 Version 8.11 10-Dec-2010
1727 ------------------------
1728
1729 1. (*THEN) was not working properly if there were untried alternatives prior
1730 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1731 backtracked to try for "b" instead of moving to the next alternative branch
1732 at the same level (in this case, to look for "c"). The Perl documentation
1733 is clear that when (*THEN) is backtracked onto, it goes to the "next
1734 alternative in the innermost enclosing group".
1735
1736 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1737 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1738 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1739 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1740 (*THEN).
1741
1742 3. If \s appeared in a character class, it removed the VT character from
1743 the class, even if it had been included by some previous item, for example
1744 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1745 of \s, but is part of the POSIX "space" class.)
1746
1747 4. A partial match never returns an empty string (because you can always
1748 match an empty string at the end of the subject); however the checking for
1749 an empty string was starting at the "start of match" point. This has been
1750 changed to the "earliest inspected character" point, because the returned
1751 data for a partial match starts at this character. This means that, for
1752 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1753 (previously it gave "no match").
1754
1755 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1756 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1757 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1758 has an implication that the given string is incomplete (because a partial
1759 match is preferred over a full match). For this reason, these items now
1760 give a partial match in this situation. [Aside: previously, the one case
1761 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1762 match rather than a full match, which was wrong by the old rules, but is
1763 now correct.]
1764
1765 6. There was a bug in the handling of #-introduced comments, recognized when
1766 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1767 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1768 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1769 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1770 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1771 places in pcre_compile().
1772
1773 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1774 comments when looking ahead for named forward references to subpatterns,
1775 the only newline sequence it recognized was NL. It now handles newlines
1776 according to the set newline convention.
1777
1778 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1779 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1780 cater for a lack of strerror(). These oversights have been fixed.
1781
1782 9. Added --match-limit and --recursion-limit to pcregrep.
1783
1784 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1785
1786 11. When the -o option was used, pcregrep was setting a return code of 1, even
1787 when matches were found, and --line-buffered was not being honoured.
1788
1789 12. Added an optional parentheses number to the -o and --only-matching options
1790 of pcregrep.
1791
1792 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1793 can match an empty string. The code to do it in pcretest and pcredemo
1794 needed fixing:
1795
1796 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1797 only one byte after an empty string match just before CRLF (this case
1798 just got forgotten; "any" and "anycrlf" were OK).
1799
1800 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1801 mode when an empty string match preceded an ASCII character followed by
1802 a non-ASCII character. (The code for advancing by one character rather
1803 than one byte was nonsense.)
1804
1805 (c) The pcredemo.c sample program did not have any code at all to handle
1806 the cases when CRLF is a valid newline sequence.
1807
1808 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1809 as a starting offset was within the subject string. There is now a new
1810 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1811 negative or greater than the length of the string. In order to test this,
1812 pcretest is extended to allow the setting of negative starting offsets.
1813
1814 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1815 starting offset points to the beginning of a UTF-8 character was
1816 unnecessarily clumsy. I tidied it up.
1817
1818 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1819 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1820
1821 17. Nobody had reported that the --include_dir option, which was added in
1822 release 7.7 should have been called --include-dir (hyphen, not underscore)
1823 for compatibility with GNU grep. I have changed it to --include-dir, but
1824 left --include_dir as an undocumented synonym, and the same for
1825 --exclude-dir, though that is not available in GNU grep, at least as of
1826 release 2.5.4.
1827
1828 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1829 characters from a string of bytes) have been redefined so as not to use
1830 loops, in order to improve performance in some environments. At the same
1831 time, I abstracted some of the common code into auxiliary macros to save
1832 repetition (this should not affect the compiled code).
1833
1834 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1835 compile-time error is now given if \c is not followed by an ASCII
1836 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1837 different, and any byte value is allowed.)
1838
1839 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1840 START_OPTIMIZE option, which is now allowed at compile time - but just
1841 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1842 to pcregrep and other applications that have no direct access to PCRE
1843 options. The new /Y option in pcretest sets this option when calling
1844 pcre_compile().
1845
1846 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1847 back references. Groups containing recursive back references were forced to
1848 be atomic by that change, but in the case of named groups, the amount of
1849 memory required was incorrectly computed, leading to "Failed: internal
1850 error: code overflow". This has been fixed.
1851
1852 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1853 pcretest.c, to avoid build problems in some Borland environments.
1854
1855
1856 Version 8.10 25-Jun-2010
1857 ------------------------
1858
1859 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1860 THEN.
1861
1862 2. (*ACCEPT) was not working when inside an atomic group.
1863
1864 3. Inside a character class, \B is treated as a literal by default, but
1865 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1866 causes the error). The code is unchanged, but I tidied the documentation.
1867
1868 4. Inside a character class, PCRE always treated \R and \X as literals,
1869 whereas Perl faults them if its -w option is set. I have changed PCRE so
1870 that it faults them when PCRE_EXTRA is set.
1871
1872 5. Added support for \N, which always matches any character other than
1873 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1874
1875 6. When compiling pcregrep with newer versions of gcc which may have
1876 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1877 declared with attribute warn_unused_result" were given. Just casting the
1878 result to (void) does not stop the warnings; a more elaborate fudge is
1879 needed. I've used a macro to implement this.
1880
1881 7. Minor change to pcretest.c to avoid a compiler warning.
1882
1883 8. Added four artifical Unicode properties to help with an option to make
1884 \s etc use properties (see next item). The new properties are: Xan
1885 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1886
1887 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1888 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1889 this option. Modified pcretest to add /W to test this facility. Added
1890 REG_UCP to make it available via the POSIX interface.
1891
1892 10. Added --line-buffered to pcregrep.
1893
1894 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1895 studied, and the match started with a letter with a code point greater than
1896 127 whose first byte was different to the first byte of the other case of
1897 the letter, the other case of this starting letter was not recognized
1898 (#976).
1899
1900 12. If a pattern that was studied started with a repeated Unicode property
1901 test, for example, \p{Nd}+, there was the theoretical possibility of
1902 setting up an incorrect bitmap of starting bytes, but fortunately it could
1903 not have actually happened in practice until change 8 above was made (it
1904 added property types that matched character-matching opcodes).
1905
1906 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1907 possible starting bytes for non-anchored patterns.
1908
1909 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1910 \R, and also a number of cases that involve Unicode properties, both
1911 explicit and implicit when PCRE_UCP is set.
1912
1913 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1914 input, it could crash or give wrong results if characters with values
1915 greater than 0xc0 were present in the subject string. (Detail: it assumed
1916 UTF-8 input when processing these items.)
1917
1918 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1919 size_t is 64-bit (#991).
1920
1921 17. Added a check for running out of memory when PCRE is compiled with
1922 --disable-stack-for-recursion (#990).
1923
1924 18. If the last data line in a file for pcretest does not have a newline on
1925 the end, a newline was missing in the output.
1926
1927 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1928 less than 128) in its various bitmaps. However, there is a facility for
1929 generating tables according to the current locale when PCRE is compiled. It
1930 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1931 characters, are recognized by isspace() and therefore were getting set in
1932 these tables, and indeed these tables seem to approximate to ISO 8859. This
1933 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1934 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1935 which of course cannot start UTF-8 characters. I have changed the code so
1936 that only real ASCII characters (less than 128) and the correct starting
1937 bytes for UTF-8 encodings are set for characters greater than 127 when in
1938 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1939 altogether.)
1940
1941 20. Added the /T option to pcretest so as to be able to run tests with non-
1942 standard character tables, thus making it possible to include the tests
1943 used for 19 above in the standard set of tests.
1944
1945 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1946 reference to a subpattern the other side of a comment that contains an
1947 opening parenthesis caused either an internal compiling error, or a
1948 reference to the wrong subpattern.
1949
1950
1951 Version 8.02 19-Mar-2010
1952 ------------------------
1953
1954 1. The Unicode data tables have been updated to Unicode 5.2.0.
1955
1956 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1957 configured.
1958
1959 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1960 original author of that file, following a query about its status.
1961
1962 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1963 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1964
1965 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1966 quantifier applied to a forward-referencing subroutine call, could compile
1967 incorrect code or give the error "internal error: previously-checked
1968 referenced subpattern not found".
1969
1970 6. Both MS Visual Studio and Symbian OS have problems with initializing
1971 variables to point to external functions. For these systems, therefore,
1972 pcre_malloc etc. are now initialized to local functions that call the
1973 relevant global functions.
1974
1975 7. There were two entries missing in the vectors called coptable and poptable
1976 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1977 I've fixed the data, and added a kludgy way of testing at compile time that
1978 the lengths are correct (equal to the number of opcodes).
1979
1980 8. Following on from 7, I added a similar kludge to check the length of the
1981 eint vector in pcreposix.c.
1982
1983 9. Error texts for pcre_compile() are held as one long string to avoid too
1984 much relocation at load time. To find a text, the string is searched,
1985 counting zeros. There was no check for running off the end of the string,
1986 which could happen if a new error number was added without updating the
1987 string.
1988
1989 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1990
1991 11. \K was not working if it appeared in an atomic group or in a group that
1992 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1993 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1994 the assertion is positive, but not if it is negative.
1995
1996 12. Change 11 fortuitously reduced the size of the stack frame used in the
1997 "match()" function of pcre_exec.c by one pointer. Forthcoming
1998 implementation of support for (*MARK) will need an extra pointer on the
1999 stack; I have reserved it now, so that the stack frame size does not
2000 decrease.
2001
2002 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2003 item in branch that calls a recursion is a subroutine call - as in the
2004 second branch in the above example - was incorrectly given the compile-
2005 time error "recursive call could loop indefinitely" because pcre_compile()
2006 was not correctly checking the subroutine for matching a non-empty string.
2007
2008 14. The checks for overrunning compiling workspace could trigger after an
2009 overrun had occurred. This is a "should never occur" error, but it can be
2010 triggered by pathological patterns such as hundreds of nested parentheses.
2011 The checks now trigger 100 bytes before the end of the workspace.
2012
2013 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2014
2015
2016 Version 8.01 19-Jan-2010
2017 ------------------------
2018
2019 1. If a pattern contained a conditional subpattern with only one branch (in
2020 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2021 computed the wrong minimum data length (which is of course zero for such
2022 subpatterns). This could cause incorrect "no match" results.
2023
2024 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2025 the pattern is reset in the first branch, pcre_compile() failed with
2026 "internal error: code overflow at offset...". This happened only when
2027 the reset was to the original external option setting. (An optimization
2028 abstracts leading options settings into an external setting, which was the
2029 cause of this.)
2030
2031 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2032 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2033 assertion pattern did not match (meaning that the assertion was true), it
2034 was incorrectly treated as false if the SKIP had been reached during the
2035 matching. This also applied to assertions used as conditions.
2036
2037 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2038 assertion subpattern, including such a pattern used as a condition,
2039 unpredictable results occurred, instead of the error return
2040 PCRE_ERROR_DFA_UITEM.
2041
2042 5. The C++ GlobalReplace function was not working like Perl for the special
2043 situation when an empty string is matched. It now does the fancy magic
2044 stuff that is necessary.
2045
2046 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2047 removed. (These were left over from very, very early versions of PCRE.)
2048
2049 7. Some cosmetic changes to the code to make life easier when compiling it
2050 as part of something else:
2051
2052 (a) Change DEBUG to PCRE_DEBUG.
2053
2054 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2055 called "current" as "current_branch", to prevent a collision with the
2056 Linux macro when compiled as a kernel module.
2057
2058 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2059 prevent a collision with the Linux macro when compiled as a kernel
2060 module.
2061
2062 8. In pcre_compile() there are some checks for integer overflows that used to
2063 cast potentially large values to (double). This has been changed to that
2064 when building, a check for int64_t is made, and if it is found, it is used
2065 instead, thus avoiding the use of floating point arithmetic. (There is no
2066 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2067 double.
2068
2069 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2070 2005 (difference between two addresses compared to an unsigned value).
2071
2072 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2073 custom one, because of the following reported problem in Windows:
2074
2075 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2076 under Win32.
2077 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2078 therefore missing the function definition.
2079 - The compiler thus generates a "C" signature for the test function.
2080 - The linker fails to find the "C" function.
2081 - PCRE fails to configure if asked to do so against libbz2.
2082
2083 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2084 messages were output:
2085
2086 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2087 rerunning libtoolize, to keep the correct libtool macros in-tree.
2088 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2089
2090 I have done both of these things.
2091
2092 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2093 most of the time, it *can* run out if it is given a pattern that contains a
2094 runaway infinite recursion. I updated the discussion in the pcrestack man
2095 page.
2096
2097 13. Now that we have gone to the x.xx style of version numbers, the minor
2098 version may start with zero. Using 08 or 09 is a bad idea because users
2099 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2100 interpreted as invalid octal numbers. I've updated the previous comment in
2101 configure.ac, and also added a check that gives an error if 08 or 09 are
2102 used.
2103
2104 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2105 causing partial matching to fail when the end of the subject matched \W
2106 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2107
2108 15. There were some discrepancies between the declarations in pcre_internal.h
2109 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2110 their definitions. The declarations used "const uschar *" and the
2111 definitions used USPTR. Even though USPTR is normally defined as "const
2112 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2113 reported that: "This difference in casting confuses some C++ compilers, for
2114 example, SunCC recognizes above declarations as different functions and
2115 generates broken code for hbpcre." I have changed the declarations to use
2116 USPTR.
2117
2118 16. GNU libtool is named differently on some systems. The autogen.sh script now
2119 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2120 (FreeBSD).
2121
2122 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2123 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2124 comment: "Figure out how to create a longlong from a string: strtoll and
2125 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2126 instance, but it only takes 2 args instead of 3!"
2127
2128 18. A subtle bug concerned with back references has been fixed by a change of
2129 specification, with a corresponding code fix. A pattern such as
2130 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2131 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2132 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2133 same bug. Such groups have to be quantified to be useful, or contained
2134 inside another quantified group. (If there's no repetition, the reference
2135 can never match.) The problem arises because, having left the group and
2136 moved on to the rest of the pattern, a later failure that backtracks into
2137 the group uses the captured value from the final iteration of the group
2138 rather than the correct earlier one. I have fixed this in PCRE by forcing
2139 any group that contains a reference to itself to be an atomic group; that
2140 is, there cannot be any backtracking into it once it has completed. This is
2141 similar to recursive and subroutine calls.
2142
2143
2144 Version 8.00 19-Oct-09
2145 ----------------------
2146
2147 1. The table for translating pcre_compile() error codes into POSIX error codes
2148 was out-of-date, and there was no check on the pcre_compile() error code
2149 being within the table. This could lead to an OK return being given in
2150 error.
2151
2152 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2153 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2154 in a Windows environment.
2155
2156 3. The pcregrep --count option prints the count for each file even when it is
2157 zero, as does GNU grep. However, pcregrep was also printing all files when
2158 --files-with-matches was added. Now, when both options are given, it prints
2159 counts only for those files that have at least one match. (GNU grep just
2160 prints the file name in this circumstance, but including the count seems
2161 more useful - otherwise, why use --count?) Also ensured that the
2162 combination -clh just lists non-zero counts, with no names.
2163
2164 4. The long form of the pcregrep -F option was incorrectly implemented as
2165 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2166 but it seems right to fix it, and I didn't think it was worth preserving
2167 the old behaviour.
2168
2169 5. The command line items --regex=pattern and --regexp=pattern were not
2170 recognized by pcregrep, which required --regex pattern or --regexp pattern
2171 (with a space rather than an '='). The man page documented the '=' forms,
2172 which are compatible with GNU grep; these now work.
2173
2174 6. No libpcreposix.pc file was created for pkg-config; there was just
2175 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2176
2177 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2178 when UCP support is not needed, by modifying the Python script that
2179 generates it from Unicode data files. This should not matter if the module
2180 is correctly used as a library, but I received one complaint about 50K of
2181 unwanted data. My guess is that the person linked everything into his
2182 program rather than using a library. Anyway, it does no harm.
2183
2184 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2185 was a minimum greater than 1 for a wide character in a possessive
2186 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2187 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2188 character. Chaos in the form of incorrect output or a compiling loop could
2189 result.
2190
2191 9. The restrictions on what a pattern can contain when partial matching is
2192 requested for pcre_exec() have been removed. All patterns can now be
2193 partially matched by this function. In addition, if there are at least two
2194 slots in the offset vector, the offset of the earliest inspected character
2195 for the match and the offset of the end of the subject are set in them when
2196 PCRE_ERROR_PARTIAL is returned.
2197
2198 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2199 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2200 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2201 and may be more useful for multi-segment matching.
2202
2203 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2204 used to be given if ever the end of the subject was reached; now it is
2205 given only if matching could not proceed because another character was
2206 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2207 string "Z", which now yields "no match" instead of "partial match". In the
2208 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2209 final character ended with (*FAIL).
2210
2211 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2212 if the pattern had a "must contain" character that was already found in the
2213 earlier partial match, unless partial matching was again requested. For
2214 example, with the pattern /dog.(body)?/, the "must contain" character is
2215 "g". If the first part-match was for the string "dog", restarting with
2216 "sbody" failed. This bug has been fixed.
2217
2218 13. The string returned by pcre_dfa_exec() after a partial match has been
2219 changed so that it starts at the first inspected character rather than the
2220 first character of the match. This makes a difference only if the pattern
2221 starts with a lookbehind assertion or \b or \B (\K is not supported by
2222 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2223 matching functions compatible, and I think it's the right thing to do.
2224
2225 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2226 so that the demonstration program is easily available in environments where
2227 PCRE has not been installed from source.
2228
2229 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2230 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2231 library.
2232
2233 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2234 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2235 is not the first non-POSIX option to be added. Clearly some people find
2236 these options useful.
2237
2238 17. If a caller to the POSIX matching function regexec() passes a non-zero
2239 value for nmatch with a NULL value for pmatch, the value of
2240 nmatch is forced to zero.
2241
2242 18. RunGrepTest did not have a test for the availability of the -u option of
2243 the diff command, as RunTest does. It now checks in the same way as
2244 RunTest, and also checks for the -b option.
2245
2246 19. If an odd number of negated classes containing just a single character
2247 interposed, within parentheses, between a forward reference to a named
2248 subpattern and the definition of the subpattern, compilation crashed with
2249 an internal error, complaining that it could not find the referenced
2250 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2251 [The bug was that it was starting one character too far in when skipping
2252 over the character class, thus treating the ] as data rather than
2253 terminating the class. This meant it could skip too much.]
2254
2255 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2256 /g option in pcretest when the pattern contains \K, which makes it possible
2257 to have an empty string match not at the start, even when the pattern is
2258 anchored. Updated pcretest and pcredemo to use this option.
2259
2260 21. If the maximum number of capturing subpatterns in a recursion was greater
2261 than the maximum at the outer level, the higher number was returned, but
2262 with unset values at the outer level. The correct (outer level) value is
2263 now given.
2264
2265 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2266 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2267 make it do so. The string so far is captured, making this feature
2268 compatible with Perl.
2269
2270 23. The tests have been re-organized, adding tests 11 and 12, to make it
2271 possible to check the Perl 5.10 features against Perl 5.10.
2272
2273 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2274 pattern matches a fixed length string. PCRE did not allow this; now it
2275 does. Neither allows recursion.
2276
2277 25. I finally figured out how to implement a request to provide the minimum
2278 length of subject string that was needed in order to match a given pattern.
2279 (It was back references and recursion that I had previously got hung up
2280 on.) This code has now been added to pcre_study(); it finds a lower bound
2281 to the length of subject needed. It is not necessarily the greatest lower
2282 bound, but using it to avoid searching strings that are too short does give
2283 some useful speed-ups. The value is available to calling programs via
2284 pcre_fullinfo().
2285
2286 26. While implementing 25, I discovered to my embarrassment that pcretest had
2287 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2288 study optimizations had never been tested with that matching function.
2289 Oops. What is worse, even when it was passed study data, there was a bug in
2290 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2291 were also very few tests of studied patterns with pcre_dfa_exec().
2292
2293 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2294 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2295 on the other side of the coin, they are no longer allowed to have different
2296 names, because these cannot be distinguished in PCRE, and this has caused
2297 confusion. (This is a difference from Perl.)
2298
2299 28. When duplicate subpattern names are present (necessarily with different
2300 numbers, as required by 27 above), and a test is made by name in a
2301 conditional pattern, either for a subpattern having been matched, or for
2302 recursion in such a pattern, all the associated numbered subpatterns are
2303 tested, and the overall condition is true if the condition is true for any
2304 one of them. This is the way Perl works, and is also more like the way
2305 testing by number works.
2306
2307
2308 Version 7.9 11-Apr-09
2309 ---------------------
2310
2311 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2312 (pcretest), all targets were linked against these libraries. This included
2313 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2314 libraries. This caused unwanted dependencies to be created. This problem
2315 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2316 pcretest is linked with readline.
2317
2318 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2319 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2320 moved outside it again, because FALSE and TRUE are already defined in AIX,
2321 but BOOL is not.
2322
2323 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2324 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2325
2326 4. The pcregrep documentation said spaces were inserted as well as colons (or
2327 hyphens) following file names and line numbers when outputting matching
2328 lines. This is not true; no spaces are inserted. I have also clarified the
2329 wording for the --colour (or --color) option.
2330
2331 5. In pcregrep, when --colour was used with -o, the list of matching strings
2332 was not coloured; this is different to GNU grep, so I have changed it to be
2333 the same.
2334
2335 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2336 each matching line was coloured. Now it goes on to look for further matches
2337 of any of the test patterns, which is the same behaviour as GNU grep.
2338
2339 7. A pattern that could match an empty string could cause pcregrep to loop; it
2340 doesn't make sense to accept an empty string match in pcregrep, so I have
2341 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2342 seems to be how GNU grep behaves. [But see later change 40 for release
2343 8.33.]
2344
2345 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2346 start or after a newline", because the conditional assertion was not being
2347 correctly handled. The rule now is that both the assertion and what follows
2348 in the first alternative must satisfy the test.
2349
2350 9. If auto-callout was enabled in a pattern with a conditional group whose
2351 condition was an assertion, PCRE could crash during matching, both with
2352 pcre_exec() and pcre_dfa_exec().
2353
2354 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2355 used for matching.
2356
2357 11. Unicode property support in character classes was not working for
2358 characters (bytes) greater than 127 when not in UTF-8 mode.
2359
2360 12. Added the -M command line option to pcretest.
2361
2362 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2363
2364 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2365
2366 16. Added comments and documentation about mis-use of no_arg in the C++
2367 wrapper.
2368
2369 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2370 from Martin Jerabek that uses macro names for all relevant character and
2371 string constants.
2372
2373 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2374 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2375 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2376 these, but not everybody uses configure.
2377
2378 19. A conditional group that had only one branch was not being correctly
2379 recognized as an item that could match an empty string. This meant that an
2380 enclosing group might also not be so recognized, causing infinite looping
2381 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2382 with the subject "ab", where knowledge that the repeated group can match
2383 nothing is needed in order to break the loop.
2384
2385 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2386 exec(), but without supplying a callout function, matching went wrong.
2387
2388 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2389 leak if the size of the offset vector was greater than 30. When the vector
2390 is smaller, the saved offsets during recursion go onto a local stack
2391 vector, but for larger vectors malloc() is used. It was failing to free
2392 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2393 error, in fact).
2394
2395 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2396 heapframe that is used only when UTF-8 support is enabled. This caused no
2397 problem, but was untidy.
2398
2399 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2400 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2401 included within another project.
2402
2403 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2404 slightly modified by me:
2405
2406 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2407 not building pcregrep.
2408
2409 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2410 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2411
2412 25. Forward references, both numeric and by name, in patterns that made use of
2413 duplicate group numbers, could behave incorrectly or give incorrect errors,
2414 because when scanning forward to find the reference group, PCRE was not
2415 taking into account the duplicate group numbers. A pattern such as
2416 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2417
2418 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2419 the feature of a custom pointer more persuasive (as requested by a user).
2420
2421 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2422 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2423 pre-defined.
2424
2425 28. Added support for (*UTF8) at the start of a pattern.
2426
2427 29. Arrange for flags added by the "release type" setting in CMake to be shown
2428 in the configuration summary.
2429
2430
2431 Version 7.8 05-Sep-08
2432 ---------------------
2433
2434 1. Replaced UCP searching code with optimized version as implemented for Ad
2435 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2436 stage table and inline lookup instead of a function, giving speed ups of 2
2437 to 5 times on some simple patterns that I tested. Permission was given to
2438 distribute the MultiStage2.py script that generates the tables (it's not in
2439 the tarball, but is in the Subversion repository).
2440
2441 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2442 scripts.
2443
2444 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2445 a group with a zero qualifier. The result of the study could be incorrect,
2446 or the function might crash, depending on the pattern.
2447
2448 4. Caseless matching was not working for non-ASCII characters in back
2449 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2450 It now works when Unicode Property Support is available.
2451
2452 5. In pcretest, an escape such as \x{de} in the data was always generating
2453 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2454 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2455 truncation.
2456
2457 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2458
2459 7. Added two (int) casts to pcregrep when printing the difference of two
2460 pointers, in case they are 64-bit values.
2461
2462 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2463 test 2 if it fails.
2464
2465 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2466 and a #define of that name to empty if it is not externally set. This is to
2467 allow users of MSVC to set it if necessary.
2468
2469 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2470 the convenience functions in the pcre_get.c source file.
2471
2472 11. An option change at the start of a pattern that had top-level alternatives
2473 could cause overwriting and/or a crash. This command provoked a crash in
2474 some environments:
2475
2476 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2477
2478 This potential security problem was recorded as CVE-2008-2371.
2479
2480 12. For a pattern where the match had to start at the beginning or immediately
2481 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2482 pcre_dfa_exec() could read past the end of the passed subject if there was
2483 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2484 pcretest so that it places the subject at the end of its malloc-ed buffer.
2485
2486 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2487 exec() might read past the end of the data buffer in UTF-8 mode.
2488
2489 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2490 the data contained the byte 0x85 as part of a UTF-8 character within its
2491 first line. This applied both to normal and DFA matching.
2492
2493 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2494 /^[^d]*?$/8 failed to match "abc".
2495
2496 16. Added a missing copyright notice to pcrecpp_internal.h.
2497
2498 17. Make it more clear in the documentation that values returned from
2499 pcre_exec() in ovector are byte offsets, not character counts.
2500
2501 18. Tidied a few places to stop certain compilers from issuing warnings.
2502
2503 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2504 supplied by Stefan Weber. I made a further small update for 7.8 because
2505 there is a change of source arrangements: the pcre_searchfuncs.c module is
2506 replaced by pcre_ucd.c.
2507
2508
2509 Version 7.7 07-May-08
2510 ---------------------
2511
2512 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2513 a string to a long long, pretend we don't even have a long long." This is
2514 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2515
2516 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2517 pre-7.6 versions, which defined a global no_arg variable instead of putting
2518 it in the RE class. (See also #8 below.)
2519
2520 3. Remove a line of dead code, identified by coverity and reported by Nuno
2521 Lopes.
2522
2523 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2524
2525 (1) The include/exclude patterns were being applied to the whole pathnames
2526 of files, instead of just to the final components.
2527
2528 (2) If there was more than one level of directory, the subdirectories were
2529 skipped unless they satisfied the include/exclude conditions. This is
2530 inconsistent with GNU grep (and could even be seen as contrary to the
2531 pcregrep specification - which I improved to make it absolutely clear).
2532 The action now is always to scan all levels of directory, and just
2533 apply the include/exclude patterns to regular files.
2534
2535 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2536 --exclude_dir in the tests to avoid scanning .svn directories.
2537
2538 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2539 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2540 doesn't support NULs in patterns.
2541
2542 7. Added some missing "const"s to declarations of static tables in
2543 pcre_compile.c and pcre_dfa_exec.c.
2544
2545 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2546 caused by fix #2 above. (Subsequently also a second patch to fix the
2547 first patch. And a third patch - this was a messy problem.)
2548
2549 9. Applied Craig's patch to remove the use of push_back().
2550
2551 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2552 matching function regexec().
2553
2554 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2555 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2556 references. PCRE supports relative numbers with this syntax (I don't think
2557 Oniguruma does).
2558
2559 12. Previously, a group with a zero repeat such as (...){0} was completely
2560 omitted from the compiled regex. However, this means that if the group
2561 was called as a subroutine from elsewhere in the pattern, things went wrong
2562 (an internal error was given). Such groups are now left in the compiled
2563 pattern, with a new opcode that causes them to be skipped at execution
2564 time.
2565
2566 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2567 to the way PCRE behaves:
2568
2569 (a) A lone ] character is dis-allowed (Perl treats it as data).
2570
2571 (b) A back reference to an unmatched subpattern matches an empty string
2572 (Perl fails the current match path).
2573
2574 (c) A data ] in a character class must be notated as \] because if the
2575 first data character in a class is ], it defines an empty class. (In
2576 Perl it is not possible to have an empty class.) The empty class []
2577 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2578 The negative empty class [^] matches any one character, independently
2579 of the DOTALL setting.
2580
2581 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2582 non-existent subpattern following a character class starting with ']' and
2583 containing () gave an internal compiling error instead of "reference to
2584 non-existent subpattern". Fortunately, when the pattern did exist, the
2585 compiled code was correct. (When scanning forwards to check for the
2586 existence of the subpattern, it was treating the data ']' as terminating
2587 the class, so got the count wrong. When actually compiling, the reference
2588 was subsequently set up correctly.)
2589
2590 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2591 it was being rejected as not supported by pcre_dfa_exec(), even though
2592 other assertions are supported. I have made pcre_dfa_exec() support
2593 (*FAIL).
2594
2595 16. The implementation of 13c above involved the invention of a new opcode,
2596 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2597 cannot be changed at match time, I realized I could make a small
2598 improvement to matching performance by compiling OP_ALLANY instead of
2599 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2600 on the OP_ANY path.
2601
2602 17. Compiling pcretest on Windows with readline support failed without the
2603 following two fixes: (1) Make the unistd.h include conditional on
2604 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2605
2606 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2607 ncurses library to be included for pcretest when ReadLine support is
2608 requested, but also to allow for it to be overridden. This patch came from
2609 Daniel Bergström.
2610
2611 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2612 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2613 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2614 spotting this.
2615
2616
2617 Version 7.6 28-Jan-08
2618 ---------------------
2619
2620 1. A character class containing a very large number of characters with
2621 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2622 overflow.
2623
2624 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2625 HAVE_LONG_LONG is not defined.
2626
2627 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2628 bring them up to date and include new features. This patch includes:
2629
2630 - Fixed PH's badly added libz and libbz2 support.
2631 - Fixed a problem with static linking.
2632 - Added pcredemo. [But later removed - see 7 below.]
2633 - Fixed dftables problem and added an option.
2634 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2635 HAVE_LONG_LONG.
2636 - Added readline support for pcretest.
2637 - Added an listing of the option settings after cmake has run.
2638
2639 4. A user submitted a patch to Makefile that makes it easy to create
2640 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2641 Makefile.am that cause it to include this special target, without
2642 affecting anything else. Note that the same mingw target plus all
2643 the other distribution libraries and programs are now supported
2644 when configuring with CMake (see 6 below) instead of with
2645 Configure/Make.
2646
2647 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2648 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2649 exported in the Windows port". It has not yet been confirmed that the patch
2650 solves the problem, but it does no harm.
2651
2652 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2653 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2654 with CMake, and also correct the comment about stack recursion.
2655
2656 7. Remove the automatic building of pcredemo from the ./configure system and
2657 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2658 of a program that users should build themselves after PCRE is installed, so
2659 building it automatically is not really right. What is more, it gave
2660 trouble in some build environments.
2661
2662 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2663
2664
2665 Version 7.5 10-Jan-08
2666 ---------------------
2667
2668 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2669 values in parens when parsing an RE using the C++ wrapper."
2670
2671 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2672 Characters greater than 255 were excluded from the class instead of being
2673 included.
2674
2675 3. The same bug as (2) above applied to negated POSIX classes such as
2676 [:^space:].
2677
2678 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2679 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2680 I have changed it.
2681
2682 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2683 first named subpattern) and a construct such as (?&a) would reference the
2684 first named subpattern whose name started with "a" (in other words, the
2685 length check was missing). Both these problems are fixed. "Subpattern name
2686 expected" is now given for (?&) (a zero-length name), and this patch also
2687 makes it give the same error for \k'' (previously it complained that that
2688 was a reference to a non-existent subpattern).
2689
2690 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2691 this is right because (?- can be followed by option settings as well as by
2692 digits. I have, however, made the messages clearer.
2693
2694 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2695 than the number used in the conditional) now cause a compile-time error.
2696 This is actually not compatible with Perl, which accepts such patterns, but
2697 treats the conditional as always being FALSE (as PCRE used to), but it
2698 seems to me that giving a diagnostic is better.
2699
2700 8. Change "alphameric" to the more common word "alphanumeric" in comments
2701 and messages.
2702
2703 9. Fix two occurrences of "backslash" in comments that should have been
2704 "backspace".
2705
2706 10. Remove two redundant lines of code that can never be obeyed (their function
2707 was moved elsewhere).
2708
2709 11. The program that makes PCRE's Unicode character property table had a bug
2710 which caused it to generate incorrect table entries for sequences of
2711 characters that have the same character type, but are in different scripts.
2712 It amalgamated them into a single range, with the script of the first of
2713 them. In other words, some characters were in the wrong script. There were
2714 thirteen such cases, affecting characters in the following ranges:
2715
2716 U+002b0 - U+002c1
2717 U+0060c - U+0060d
2718 U+0061e - U+00612
2719 U+0064b - U+0065e
2720 U+0074d - U+0076d
2721 U+01800 - U+01805
2722 U+01d00 - U+01d77
2723 U+01d9b - U+01dbf
2724 U+0200b - U+0200f
2725 U+030fc - U+030fe
2726 U+03260 - U+0327f
2727 U+0fb46 - U+0fbb1
2728 U+10450 - U+1049d
2729
2730 12. The -o option (show only the matching part of a line) for pcregrep was not
2731 compatible with GNU grep in that, if there was more than one match in a
2732 line, it showed only the first of them. It now behaves in the same way as
2733 GNU grep.
2734
2735 13. If the -o and -v options were combined for pcregrep, it printed a blank
2736 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2737 does the same. The return code can be used to tell if there were any
2738 non-matching lines.
2739
2740 14. Added --file-offsets and --line-offsets to pcregrep.
2741
2742 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2743 infinitely looping recursion. The bug was that positive lookaheads were not
2744 being skipped when checking for a possible empty match (negative lookaheads
2745 and both kinds of lookbehind were skipped).
2746
2747 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2748 inclusion of <windows.h> to before rather than after the definition of
2749 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2750
2751 17. Specifying a possessive quantifier with a specific limit for a Unicode
2752 character property caused pcre_compile() to compile bad code, which led at
2753 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2754 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2755 caused the error; without that there was no problem.
2756
2757 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2758
2759 19. Added --enable-pcretest-libreadline.
2760
2761 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2762 RE::GlobalReplace(). As a result, the number of replacements returned was
2763 double what it should be. I removed one of the increments, but Craig sent a
2764 later patch that removed the other one (the right fix) and added unit tests
2765 that check the return values (which was not done before).
2766
2767 21. Several CMake things:
2768
2769 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2770 the names libpcre and libpcreposix, not just pcre and pcreposix.
2771
2772 (2) The above change means that pcretest and pcregrep are now correctly
2773 linked with the newly-built libraries, not previously installed ones.
2774
2775 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2776
2777 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2778 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2779 UTF-8 newline character). The key issue is that the pattern starts .*;
2780 this means that the match must be either at the beginning, or after a
2781 newline. The bug was in the code for advancing after a failed match and
2782 checking that the new position followed a newline. It was not taking
2783 account of UTF-8 characters correctly.
2784
2785 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2786 character classes. PCRE was not treating the sequence [:...:] as a
2787 character class unless the ... were all letters. Perl, however, seems to
2788 allow any characters between [: and :], though of course it rejects as
2789 unknown any "names" that contain non-letters, because all the known class
2790 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2791 for example, whereas PCRE did not - it did not recognize a POSIX character
2792 class. This seemed a bit dangerous, so the code has been changed to be
2793 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2794 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2795 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2796 Perl does, and where it didn't before.
2797
2798 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2799 Windows environments %n is disabled by default.
2800
2801
2802 Version 7.4 21-Sep-07
2803 ---------------------
2804
2805 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2806 means that a class such as [\s] counted as "explicit reference to CR or
2807 LF". That isn't really right - the whole point of the change was to try to
2808 help when there was an actual mention of one of the two characters. So now
2809 the change happens only if \r or \n (or a literal CR or LF) character is
2810 encountered.
2811
2812 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2813 of both had grown to the point where there were only 3 bits left.
2814 Fortunately, there was spare space in the data structure, and so I have
2815 moved the internal flags into a new 16-bit field to free up more option
2816 bits.
2817
2818 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2819 but did not set the internal JCHANGED flag - either of these is enough to
2820 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2821 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2822 start sets both bits.
2823
2824 4. Added options (at build time, compile time, exec time) to change \R from
2825 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2826
2827 5. doc/pcresyntax.html was missing from the distribution.
2828
2829 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2830 compatibility, even though it is no longer used.
2831
2832 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2833 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2834 windows.h file is present (where different names are used). [This was
2835 reversed later after testing - see 16 below.]
2836
2837 8. Changed all #include <config.h> to #include "config.h". There were also
2838 some further <pcre.h> cases that I changed to "pcre.h".
2839
2840 9. When pcregrep was used with the --colour option, it missed the line ending
2841 sequence off the lines that it output.
2842
2843 10. It was pointed out to me that arrays of string pointers cause lots of
2844 relocations when a shared library is dynamically loaded. A technique of
2845 using a single long string with a table of offsets can drastically reduce
2846 these. I have refactored PCRE in four places to do this. The result is
2847 dramatic:
2848
2849 Originally: 290
2850 After changing UCP table: 187
2851 After changing error message table: 43
2852 After changing table of "verbs" 36
2853 After changing table of Posix names 22
2854
2855 Thanks to the folks working on Gregex for glib for this insight.
2856
2857 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2858 unicode-properties was also set.
2859
2860 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2861
2862 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2863 checked only for CRLF.
2864
2865 14. Added casts to pcretest.c to avoid compiler warnings.
2866
2867 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2868
2869 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2870 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2871 entirely. This removes changes made in 7 above.
2872
2873 17. The CMake files have been updated, and there is now more information about
2874 building with CMake in the NON-UNIX-USE document.
2875
2876
2877 Version 7.3 28-Aug-07
2878 ---------------------
2879
2880 1. In the rejigging of the build system that eventually resulted in 7.1, the
2881 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2882 brackets there is not right, since it causes compilers to look for an
2883 installed pcre.h, not the version that is in the source that is being
2884 compiled (which of course may be different). I have changed it back to:
2885
2886 #include "pcre.h"
2887
2888 I have a vague recollection that the change was concerned with compiling in
2889 different directories, but in the new build system, that is taken care of
2890 by the VPATH setting the Makefile.
2891
2892 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2893 when the subject happened to end in the byte 0x85 (e.g. if the last
2894 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2895 characters but of course it shouldn't be taken as a newline when it is part
2896 of another character. The bug was that, for an unlimited repeat of . in
2897 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2898 characters when looking for a newline.
2899
2900 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2901
2902 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2903 in debug output.
2904
2905 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2906 long printing in the pcrecpp unittest when running under MinGW.
2907
2908 6. ESC_K was left out of the EBCDIC table.
2909
2910 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2911 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2912 limit also applies to "virtual nesting" when a pattern is recursive, and in
2913 this case 1000 isn't so big. I have been able to remove this limit at the
2914 expense of backing off one optimization in certain circumstances. Normally,
2915 when pcre_exec() would call its internal match() function recursively and
2916 immediately return the result unconditionally, it uses a "tail recursion"
2917 feature to save stack. However, when a subpattern that can match an empty
2918 string has an unlimited repetition quantifier, it no longer makes this
2919 optimization. That gives it a stack frame in which to save the data for
2920 checking that an empty string has been matched. Previously this was taken
2921 from the 1000-entry workspace that had been reserved. So now there is no
2922 explicit limit, but more stack is used.
2923
2924 8. Applied Daniel's patches to solve problems with the import/export magic
2925 syntax that is required for Windows, and which was going wrong for the
2926 pcreposix and pcrecpp parts of the library. These were overlooked when this
2927 problem was solved for the main library.
2928
2929 9. There were some crude static tests to avoid integer overflow when computing
2930 the size of patterns that contain repeated groups with explicit upper
2931 limits. As the maximum quantifier is 65535, the maximum group length was
2932 set at 30,000 so that the product of these two numbers did not overflow a
2933 32-bit integer. However, it turns out that people want to use groups that
2934 are longer than 30,000 bytes (though not repeat them that many times).
2935 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2936 made it possible to implement the integer overflow checks in a much more
2937 dynamic way, which I have now done. The artificial limitation on group
2938 length has been removed - we now have only the limit on the total length of
2939 the compiled pattern, which depends on the LINK_SIZE setting.
2940
2941 10. Fixed a bug in the documentation for get/copy named substring when
2942 duplicate names are permitted. If none of the named substrings are set, the
2943 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2944 empty string.
2945
2946 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2947 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2948 because the ] is interpreted as the first data character and the
2949 terminating ] is not found. PCRE has been made compatible with Perl in this
2950 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2951 cause memory overwriting.
2952
2953 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2954 string has been matched (to stop an infinite loop). It was not recognizing
2955 a conditional subpattern that could match an empty string if that
2956 subpattern was within another subpattern. For example, it looped when
2957 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2958 condition was not nested. This bug has been fixed.
2959
2960 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2961 past the start of the subject in the presence of bytes with the top bit
2962 set, for example "\x8aBCD".
2963
2964 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2965 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2966
2967 14. Optimized (?!) to (*FAIL).
2968
2969 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2970 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2971 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2972 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2973 does: it's just the validity check that is more restrictive.
2974
2975 16. Inserted checks for integer overflows during escape sequence (backslash)
2976 processing, and also fixed erroneous offset values for syntax errors during
2977 backslash processing.
2978
2979 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2980 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2981
2982 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2983 caused an overrun.
2984
2985 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2986 something other than just ASCII characters) inside a group that had an
2987 unlimited repeat caused a loop at compile time (while checking to see
2988 whether the group could match an empty string).
2989
2990 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2991 [\P{Any}] did so. (Error in the code for printing property names.)
2992
2993 21. An orphan \E inside a character class could cause a crash.
2994
2995 22. A repeated capturing bracket such as (A)? could cause a wild memory
2996 reference during compilation.
2997
2998 23. There are several functions in pcre_compile() that scan along a compiled
2999 expression for various reasons (e.g. to see if it's fixed length for look
3000 behind). There were bugs in these functions when a repeated \p or \P was
3001 present in the pattern. These operators have additional parameters compared
3002 with \d, etc, and these were not being taken into account when moving along
3003 the compiled data. Specifically:
3004
3005 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3006 length.
3007
3008 (b) An item such as \pL+ within a repeated group could cause crashes or
3009 loops.
3010
3011 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3012 "reference to non-existent subpattern" error.
3013
3014 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3015
3016 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3017 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3018
3019 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3020
3021 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3022 character were causing crashes (broken optimization).
3023
3024 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3025 \p or \P) caused a compile-time loop.
3026
3027 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3028 break. For example, the unstudied pattern [\r\n]A does not match the string
3029 "\r\nA" because change 7.0/46 below moves the current point on by two
3030 characters after failing to match at the start. However, the pattern \nA
3031 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3032 the same is true. There doesn't seem any very clean way out of this, but
3033 what I have chosen to do makes the common cases work: PCRE now takes note
3034 of whether there can be an explicit match for \r or \n anywhere in the
3035 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3036 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3037 pattern has explicit CR or LF references.
3038
3039 29. Added (*CR) etc for changing newline setting at start of pattern.
3040
3041
3042 Version 7.2 19-Jun-07
3043 ---------------------
3044
3045 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3046 which is apparently normally available under Windows.
3047
3048 2. Re-jig the pcregrep tests with different newline settings in an attempt
3049 to make them independent of the local environment's newline setting.
3050
3051 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3052
3053 4. Some of the "internals" tests were previously cut out when the link size
3054 was not 2, because the output contained actual offsets. The recent new
3055 "Z" feature of pcretest means that these can be cut out, making the tests
3056 usable with all link sizes.
3057
3058 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3059 stack recursion. This gives a massive performance boost under BSD, but just
3060 a small improvement under Linux. However, it saves one field in the frame
3061 in all cases.
3062
3063 6. Added more features from the forthcoming Perl 5.10:
3064
3065 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3066 recursion call. It refers to the nth most recently opened parentheses.
3067
3068 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3069 to be opened parentheses.
3070
3071 (c) Conditions that refer to capturing parentheses can be specified
3072 relatively, for example, (?(-2)... or (?(+3)...
3073
3074 (d) \K resets the start of the current match so that everything before
3075 is not part of it.
3076
3077 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3078
3079 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3080 reference syntax.
3081
3082 (g) (?| introduces a group in which the numbering of parentheses in each
3083 alternative starts with the same number.
3084
3085 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3086
3087 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3088 PCRE_INFO_JCHANGED.
3089
3090 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3091 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3092 for detecting groups that can match an empty string.
3093
3094 9. A pattern with a very large number of alternatives (more than several
3095 hundred) was running out of internal workspace during the pre-compile
3096 phase, where pcre_compile() figures out how much memory will be needed. A
3097 bit of new cunning has reduced the workspace needed for groups with
3098 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3099 workspace instead of running out of the 4096 that are available.
3100
3101 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3102
3103 11. Applied patch from Google to remove an optimization that didn't quite work.
3104 The report of the bug said:
3105
3106 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3107 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3108 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3109
3110 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3111 it matched the wrong number of bytes.
3112
3113
3114 Version 7.1 24-Apr-07
3115 ---------------------
3116
3117 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3118 that is more "standard", making use of automake and other Autotools. There
3119 is some re-arrangement of the files and adjustment of comments consequent
3120 on this.
3121
3122 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3123 for recursive directory scanning broke on some systems because the files
3124 are not scanned in any specific order and on different systems the order
3125 was different. A call to "sort" has been inserted into RunGrepTest for the
3126 approprate test as a short-term fix. In the longer term there may be an
3127 alternative.
3128
3129 3. I had an email from Eric Raymond about problems translating some of PCRE's
3130 man pages to HTML (despite the fact that I distribute HTML pages, some
3131 people do their own conversions for various reasons). The problems
3132 concerned the use of low-level troff macros .br and .in. I have therefore
3133 removed all such uses from the man pages (some were redundant, some could
3134 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3135 HTML has been updated to handle .nf/.fi and to complain if it encounters
3136 .br or .in.
3137
3138 4. Updated comments in configure.ac that get placed in config.h.in and also
3139 arranged for config.h to be included in the distribution, with the name
3140 config.h.generic, for the benefit of those who have to compile without
3141 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3142
3143 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3144 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3145 makevp.bat for the current PCRE, using the additional files
3146 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3147
3148 6. A Windows user reported a minor discrepancy with test 2, which turned out
3149 to be caused by a trailing space on an input line that had got lost in his
3150 copy. The trailing space was an accident, so I've just removed it.
3151
3152 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3153 that is needed.
3154
3155 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3156 as "const" (a) because they are and (b) because it helps the PHP
3157 maintainers who have recently made a script to detect big data structures
3158 in the php code that should be moved to the .rodata section. I remembered
3159 to update Builducptable as well, so it won't revert if ucptable.h is ever
3160 re-created.
3161
3162 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3163 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3164 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3165 support is not required. This saves 1.5-2K of code, which is important in
3166 some applications.
3167
3168 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3169 so as not to refer to the tables, even though these functions will never be
3170 called when UTF-8 support is disabled. Otherwise there are problems with a
3171 shared library.
3172
3173 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3174
3175 (a) It was defining its arguments as char * instead of void *.
3176
3177 (b) It was assuming that all moves were upwards in memory; this was true
3178 a long time ago when I wrote it, but is no longer the case.
3179
3180 The emulated memove() is provided for those environments that have neither
3181 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3182 is clearly not the case, as these two bugs were recently reported.
3183
3184 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3185 and Detrail to create the HTML documentation, the .txt form of the man
3186 pages, and it removes trailing spaces from listed files. It also creates
3187 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3188 case, it wraps all the #defines with #ifndefs. This script should be run
3189 before "make dist".
3190
3191 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3192 with Unicode property support.
3193
3194 (a) For a maximizing quantifier, if the two different cases of the
3195 character were of different lengths in their UTF-8 codings (there are
3196 some cases like this - I found 11), and the matching function had to
3197 back up over a mixture of the two cases, it incorrectly assumed they
3198 were both the same length.
3199
3200 (b) When PCRE was configured to use the heap rather than the stack for
3201 recursion during matching, it was not correctly preserving the data for
3202 the other case of a UTF-8 character when checking ahead for a match
3203 while processing a minimizing repeat. If the check also involved
3204 matching a wide character, but failed, corruption could cause an
3205 erroneous result when trying to check for a repeat of the original
3206 character.
3207
3208 13. Some tidying changes to the testing mechanism:
3209
3210 (a) The RunTest script now detects the internal link size and whether there
3211 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3212 values substituted by "configure". (The RunGrepTest script already did
3213 this for UTF-8.) The configure.ac script no longer substitutes the
3214 relevant variables.
3215
3216 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3217 with length and offset values. This means that the output is different
3218 for different internal link sizes. Test 2 is skipped for link sizes
3219 other than 2 because of this, bypassing the problem. Unfortunately,
3220 there was also a test in test 3 (the locale tests) that used /B and
3221 failed for link sizes other than 2. Rather than cut the whole test out,
3222 I have added a new /Z option to pcretest that replaces the length and
3223 offset values with spaces. This is now used to make test 3 independent
3224 of link size. (Test 2 will be tidied up later.)
3225
3226 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3227 segmentation fault instead of returning the appropriate error message.
3228
3229 15. In multiline mode when the newline sequence was set to "any", the pattern
3230 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3231 This doesn't seem right; it now treats the CRLF combination as the line
3232 ending, and so does not match in that case. It's only a pattern such as ^$
3233 that would hit this one: something like ^ABC$ would have failed after \r
3234 and then tried again after \r\n.
3235
3236 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3237 in an attempt to make files that differ only in their line terminators
3238 compare equal. This works on Linux.
3239
3240 17. Under certain error circumstances pcregrep might try to free random memory
3241 as it exited. This is now fixed, thanks to valgrind.
3242
3243 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3244 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3245 was because its rules for how to advance for /g after matching an empty
3246 string at the end of a line did not allow for this case. They now check for
3247 it specially.
3248
3249 20. pcretest is supposed to handle patterns and data of any length, by
3250 extending its buffers when necessary. It was getting this wrong when the
3251 buffer for a data line had to be extended.
3252
3253 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3254 CRLF as a newline sequence.
3255
3256 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3257 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3258 I have nevertheless tidied it up.
3259
3260 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3261
3262 24. Added a man page for pcre-config.
3263
3264
3265 Version 7.0 19-Dec-06
3266 ---------------------
3267
3268 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3269 moving to gcc 4.1.1.
3270
3271 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3272 sys/time.h, which is documented as needed for this function. It doesn't
3273 seem to matter on Linux, but it showed up on some releases of OS X.
3274
3275 3. It seems that there are systems where bytes whose values are greater than
3276 127 match isprint() in the "C" locale. The "C" locale should be the
3277 default when a C program starts up. In most systems, only ASCII printing
3278 characters match isprint(). This difference caused the output from pcretest
3279 to vary, making some of the tests fail. I have changed pcretest so that:
3280
3281 (a) When it is outputting text in the compiled version of a pattern, bytes
3282 other than 32-126 are always shown as hex escapes.
3283
3284 (b) When it is outputting text that is a matched part of a subject string,
3285 it does the same, unless a different locale has been set for the match
3286 (using the /L modifier). In this case, it uses isprint() to decide.
3287
3288 4. Fixed a major bug that caused incorrect computation of the amount of memory
3289 required for a compiled pattern when options that changed within the
3290 pattern affected the logic of the preliminary scan that determines the
3291 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3292 that the computed length was too small. The symptoms of this bug were
3293 either the PCRE error "internal error: code overflow" from pcre_compile(),
3294 or a glibc crash with a message such as "pcretest: free(): invalid next
3295 size (fast)". Examples of patterns that provoked this bug (shown in
3296 pcretest format) are:
3297
3298 /(?-x: )/x
3299 /(?x)(?-x: \s*#\s*)/
3300 /((?i)[\x{c0}])/8
3301 /(?i:[\x{c0}])/8
3302
3303 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3304 is now done differently.
3305
3306 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3307 wrapper classes; (b) implement a new function in the C++ scanner that is
3308 more efficient than the old way of doing things because it avoids levels of
3309 recursion in the regex matching; (c) add a paragraph to the documentation
3310 for the FullMatch() function.
3311
3312 6. The escape sequence \n was being treated as whatever was defined as
3313 "newline". Not only was this contrary to the documentation, which states
3314 that \n is character 10 (hex 0A), but it also went horribly wrong when
3315 "newline" was defined as CRLF. This has been fixed.
3316
3317 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3318 was being set to -1 for the "end of line" case (supposedly a value that no
3319 character can have). Though this value is never used (the check for end of
3320 line is "zero bytes in current character"), it caused compiler complaints.
3321 I've changed it to 0xffffffff.
3322
3323 8. In pcre_version.c, the version string was being built by a sequence of
3324 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3325 string (as it is for production releases) called a macro with an empty
3326 argument. The C standard says the result of this is undefined. The gcc
3327 compiler treats it as an empty string (which was what was wanted) but it is
3328 reported that Visual C gives an error. The source has been hacked around to
3329 avoid this problem.
3330
3331 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3332 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3333 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3334 of them did).
3335
3336 10. Originally, pcretest opened its input and output without "b"; then I was
3337 told that "b" was needed in some environments, so it was added for release
3338 5.0 to both the input and output. (It makes no difference on Unix-like
3339 systems.) Later I was told that it is wrong for the input on Windows. I've
3340 now abstracted the modes into two macros, to make it easier to fiddle with
3341 them, and removed "b" from the input mode under Windows.
3342
3343 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3344
3345 12. Added -help and --help to pcretest as an official way of being reminded
3346 of the options.
3347
3348 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3349 and pcrecpp.cc because they annoy compilers at high warning levels.
3350
3351 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3352
3353 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3354 scripts are not C programs :-) and which was not noticed because it works
3355 on Linux.
3356
3357 16. pcretest is supposed to handle any length of pattern and data line (as one
3358 line or as a continued sequence of lines) by extending its input buffer if
3359 necessary. This feature was broken for very long pattern lines, leading to
3360 a string of junk being passed to pcre_compile() if the pattern was longer
3361 than about 50K.
3362
3363 17. I have done a major re-factoring of the way pcre_compile() computes the
3364 amount of memory needed for a compiled pattern. Previously, there was code
3365 that made a preliminary scan of the pattern in order to do this. That was
3366 OK when PCRE was new, but as the facilities have expanded, it has become
3367 harder and harder to keep it in step with the real compile phase, and there
3368 have been a number of bugs (see for example, 4 above). I have now found a
3369 cunning way of running the real compile function in a "fake" mode that
3370 enables it to compute how much memory it would need, while actually only
3371 ever using a few hundred bytes of working memory and without too many
3372 tests of the mode. This should make future maintenance and development
3373 easier. A side effect of this work is that the limit of 200 on the nesting
3374 depth of parentheses has been removed (though this was never a serious
3375 limitation, I suspect). However, there is a downside: pcre_compile() now
3376 runs more slowly than before (30% or more, depending on the pattern). I
3377 hope this isn't a big issue. There is no effect on runtime performance.
3378
3379 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3380 newline (only possible for the last line of a file) and it was a
3381 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3382
3383 19. Added additional timing features to pcretest. (1) The -tm option now times
3384 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3385 separate command line item, by a number that specifies the number of
3386 repeats to use when timing. The default is 50000; this gives better
3387 precision, but takes uncomfortably long for very large patterns.
3388
3389 20. Extended pcre_study() to be more clever in cases where a branch of a
3390 subpattern has no definite first character. For example, (a*|b*)[cd] would
3391 previously give no result from pcre_study(). Now it recognizes that the
3392 first character must be a, b, c, or d.
3393
3394 21. There was an incorrect error "recursive call could loop indefinitely" if
3395 a subpattern (or the entire pattern) that was being tested for matching an
3396 empty string contained only one non-empty item after a nested subpattern.
3397 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3398 incorrectly, because the \d was being skipped in the check.
3399
3400 22. The pcretest program now has a new pattern option /B and a command line
3401 option -b, which is equivalent to adding /B to every pattern. This causes
3402 it to show the compiled bytecode, without the additional information that
3403 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3404 is the same as /B/I).
3405
3406 23. A new optimization is now able automatically to treat some sequences such
3407 as a*b as a*+b. More specifically, if something simple (such as a character
3408 or a simple class like \d) has an unlimited quantifier, and is followed by
3409 something that cannot possibly match the quantified thing, the quantifier
3410 is automatically "possessified".
3411
3412 24. A recursive reference to a subpattern whose number was greater than 39
3413 went wrong under certain circumstances in UTF-8 mode. This bug could also
3414 have affected the operation of pcre_study().
3415
3416 25. Realized that a little bit of performance could be had by replacing
3417 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3418
3419 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3420
3421 27. Possessive quantifiers such as a++ were previously implemented by turning
3422 them into atomic groups such as ($>a+). Now they have their own opcodes,
3423 which improves performance. This includes the automatically created ones
3424 from 23 above.
3425
3426 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3427 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3428 the first matched character to be a colon. This applied both to named and
3429 numbered groups.
3430
3431 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3432
3433 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3434 building PCRE on Windows easier, so I have included it in the distribution.
3435
3436 31. There is now a check in pcretest against a ridiculously large number being
3437 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3438 loop, the loop is abandoned.
3439
3440 32. Forward references to subpatterns in conditions such as (?(2)...) where
3441 subpattern 2 is defined later cause pcre_compile() to search forwards in
3442 the pattern for the relevant set of parentheses. This search went wrong
3443 when there were unescaped parentheses in a character class, parentheses
3444 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3445
3446 33. "Subroutine" calls and backreferences were previously restricted to
3447 referencing subpatterns earlier in the regex. This restriction has now
3448 been removed.
3449
3450 34. Added a number of extra features that are going to be in Perl 5.10. On the
3451 whole, these are just syntactic alternatives for features that PCRE had
3452 previously implemented using the Python syntax or my own invention. The
3453 other formats are all retained for compatibility.
3454
3455 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3456 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3457 also .NET compatible.
3458
3459 (b) A recursion or subroutine call to a named group can now be defined as
3460 (?&name) as well as (?P>name).
3461
3462 (c) A backreference to a named group can now be defined as \k<name> or
3463 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3464 5.10, are also .NET compatible.
3465
3466 (d) A conditional reference to a named group can now use the syntax
3467 (?(<name>) or (?('name') as well as (?(name).
3468
3469 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3470 groups (named and numbered) that are never evaluated inline, but can be
3471 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3472 is always false. There may be only one alternative in such a group.
3473
3474 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3475 as the simple (?(R). The condition is true only if the most recent
3476 recursion is that of the given number or name. It does not search out
3477 through the entire recursion stack.
3478
3479 (g) The escape \gN or \g{N} has been added, where N is a positive or
3480 negative number, specifying an absolute or relative reference.
3481
3482 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3483 some "unreachable code" warnings.
3484
3485 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3486 things, this adds five new scripts.
3487
3488 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3489 There were also incompatibilities regarding the handling of \Q..\E inside
3490 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3491 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3492
3493 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3494 matches an empty string, and forcibly breaks the loop. There were bugs in
3495 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3496 against aaaa the result was just "a" rather than "aaaa", for example. Two
3497 separate and independent bugs (that affected different cases) have been
3498 fixed.
3499
3500 39. Refactored the code to abolish the use of different opcodes for small
3501 capturing bracket numbers. This is a tidy that I avoided doing when I
3502 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3503 The new approach is not only tidier, it makes it possible to reduce the
3504 memory needed to fix the previous bug (38).
3505
3506 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3507 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3508 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3509 mode.
3510
3511 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3512 report.
3513
3514 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3515 copy construction and assignment in the C++ wrapper.
3516
3517 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3518 couple of bugs that could have given wrong results in the "--newline=crlf"
3519 case.
3520
3521 44. Added a number of casts and did some reorganization of signed/unsigned int
3522 variables following suggestions from Dair Grant. Also renamed the variable
3523 "this" as "item" because it is a C++ keyword.
3524
3525 45. Arranged for dftables to add
3526
3527 #include "pcre_internal.h"
3528
3529 to pcre_chartables.c because without it, gcc 4.x may remove the array
3530 definition from the final binary if PCRE is built into a static library and
3531 dead code stripping is activated.
3532
3533 46. For an unanchored pattern, if a match attempt fails at the start of a
3534 newline sequence, and the newline setting is CRLF or ANY, and the next two
3535 characters are CRLF, advance by two characters instead of one.
3536
3537
3538 Version 6.7 04-Jul-06
3539 ---------------------
3540
3541 1. In order to handle tests when input lines are enormously long, pcretest has
3542 been re-factored so that it automatically extends its buffers when
3543 necessary. The code is crude, but this _is_ just a test program. The
3544 default size has been increased from 32K to 50K.
3545
3546 2. The code in pcre_study() was using the value of the re argument before
3547 testing it for NULL. (Of course, in any sensible call of the function, it
3548 won't be NULL.)
3549
3550 3. The memmove() emulation function in pcre_internal.h, which is used on
3551 systems that lack both memmove() and bcopy() - that is, hardly ever -
3552 was missing a "static" storage class specifier.
3553
3554 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3555 containing an extended class (one that cannot be represented by a bitmap
3556 because it contains high-valued characters or Unicode property items, e.g.
3557 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3558 pattern, but PCRE should not loop if you do not (it no longer does).
3559 [Detail: two cases were found: (a) a repeated subpattern containing an
3560 extended class; (b) a recursive reference to a subpattern that followed a
3561 previous extended class. It wasn't skipping over the extended class
3562 correctly when UTF-8 mode was not set.]
3563
3564 5. A negated single-character class was not being recognized as fixed-length
3565 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3566 compile error "lookbehind assertion is not fixed length".
3567
3568 6. The RunPerlTest auxiliary script was showing an unexpected difference
3569 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3570 write a Perl script that can interpret lines of an input file either as
3571 byte characters or as UTF-8, which is what "perltest" was being required to
3572 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3573 can't do is switch easily at run time between having the "use utf8;" pragma
3574 or not. In the end, I fudged it by using the RunPerlTest script to insert
3575 "use utf8;" explicitly for the UTF-8 tests.
3576
3577 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3578 the end of the subject string, contrary to the documentation and to what
3579 Perl does. This was true of both matching functions. Now it matches only at
3580 the start of the subject and immediately after *internal* newlines.
3581
3582 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3583 a pointer to an int instead of a pointer to an unsigned long int. This
3584 caused problems on 64-bit systems.
3585
3586 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3587 instance of the 'standard' template library not being so standard".
3588
3589 10. There was no check on the number of named subpatterns nor the maximum
3590 length of a subpattern name. The product of these values is used to compute
3591 the size of the memory block for a compiled pattern. By supplying a very
3592 long subpattern name and a large number of named subpatterns, the size
3593 computation could be caused to overflow. This is now prevented by limiting
3594 the length of names to 32 characters, and the number of named subpatterns
3595 to 10,000.
3596
3597 11. Subpatterns that are repeated with specific counts have to be replicated in
3598 the compiled pattern. The size of memory for this was computed from the
3599 length of the subpattern and the repeat count. The latter is limited to
3600 65535, but there was no limit on the former, meaning that integer overflow
3601 could in principle occur. The compiled length of a repeated subpattern is
3602 now limited to 30,000 bytes in order to prevent this.
3603
3604 12. Added the optional facility to have named substrings with the same name.
3605
3606 13. Added the ability to use a named substring as a condition, using the
3607 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3608 are numbers (not recommended). Forward references are permitted.
3609
3610 14. Added forward references in named backreferences (if you see what I mean).
3611
3612 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3613 pattern could run off the end of the subject. For example, the pattern
3614 "(?s)(.{1,5})"8 did this with the subject "ab".
3615
3616 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3617 PCRE_CASELESS was set when matching characters that were quantified with ?
3618 or *.
3619
3620 17. A character class other than a single negated character that had a minimum
3621 but no maximum quantifier - for example [ab]{6,} - was not handled
3622 correctly by pce_dfa_exec(). It would match only one character.
3623
3624 18. A valid (though odd) pattern that looked like a POSIX character
3625 class but used an invalid character after [ (for example [[,abc,]]) caused
3626 pcre_compile() to give the error "Failed: internal error: code overflow" or
3627 in some cases to crash with a glibc free() error. This could even happen if
3628 the pattern terminated after [[ but there just happened to be a sequence of
3629 letters, a binary zero, and a closing ] in the memory that followed.
3630
3631 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3632 over the years. Originally (before any Unicode support), just the bottom 8
3633 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3634 output from "man perlunicode" includes this:
3635
3636 The regular expression compiler produces polymorphic opcodes. That
3637 is, the pattern adapts to the data and automatically switches to
3638 the Unicode character scheme when presented with Unicode data--or
3639 instead uses a traditional byte scheme when presented with byte
3640 data.
3641
3642 Sadly, a wide octal escape does not cause a switch, and in a string with
3643 no other multibyte characters, these octal escapes are treated as before.
3644 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3645 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3646 Unicode string.
3647
3648 I have not perpetrated such confusion in PCRE. Up till now, it took just
3649 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3650 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3651 translate to the appropriate multibyte character.
3652
3653 29. Applied some refactoring to reduce the number of warnings from Microsoft
3654 and Borland compilers. This has included removing the fudge introduced
3655 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3656 a warning about an unused variable.
3657
3658 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3659 characters since release 4.0, because Perl (from release 5.004) does not.
3660 [Or at least, is documented not to: some releases seem to be in conflict
3661 with the documentation.] However, when a pattern was studied with
3662 pcre_study() and all its branches started with \s, PCRE still included VT
3663 as a possible starting character. Of course, this did no harm; it just
3664 caused an unnecessary match attempt.
3665
3666 22. Removed a now-redundant internal flag bit that recorded the fact that case
3667 dependency changed within the pattern. This was once needed for "required
3668 byte" processing, but is no longer used. This recovers a now-scarce options
3669 bit. Also moved the least significant internal flag bit to the most-
3670 significant bit of the word, which was not previously used (hangover from
3671 the days when it was an int rather than a uint) to free up another bit for
3672 the future.
3673
3674 23. Added support for CRLF line endings as well as CR and LF. As well as the
3675 default being selectable at build time, it can now be changed at runtime
3676 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3677 specify that it is scanning data with non-default line endings.
3678
3679 24. Changed the definition of CXXLINK to make it agree with the definition of
3680 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3681
3682 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3683 recursions. This makes a big different to stack usage for some patterns.
3684
3685 26. If a subpattern containing a named recursion or subroutine reference such
3686 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3687 the space required for the compiled pattern went wrong and gave too small a
3688 value. Depending on the environment, this could lead to "Failed: internal
3689 error: code overflow at offset 49" or "glibc detected double free or
3690 corruption" errors.
3691
3692 27. Applied patches from Google (a) to support the new newline modes and (b) to
3693 advance over multibyte UTF-8 characters in GlobalReplace.
3694
3695 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3696 difference for some implementation of PCRE in some Windows version.
3697
3698 29. Added some extra testing facilities to pcretest:
3699
3700 \q<number> in a data line sets the "match limit" value
3701 \Q<number> in a data line sets the "match recursion limt" value
3702 -S <number> sets the stack size, where <number> is in megabytes
3703
3704 The -S option isn't available for Windows.
3705
3706
3707 Version 6.6 06-Feb-06
3708 ---------------------
3709
3710 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3711 in pcreposix.h. I have copied the definition from pcre.h.
3712
3713 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3714 because pcre.h is no longer a built file.
3715
3716 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3717 not normally included in the compiled code.
3718
3719
3720 Version 6.5 01-Feb-06
3721 ---------------------
3722
3723 1. When using the partial match feature with pcre_dfa_exec(), it was not
3724 anchoring the second and subsequent partial matches at the new starting
3725 point. This could lead to incorrect results. For example, with the pattern
3726 /1234/, partially matching against "123" and then "a4" gave a match.
3727
3728 2. Changes to pcregrep:
3729
3730 (a) All non-match returns from pcre_exec() were being treated as failures
3731 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3732 error message is output. Some extra information is given for the
3733 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3734 probably the only errors that are likely to be caused by users (by
3735 specifying a regex that has nested indefinite repeats, for instance).
3736 If there are more than 20 of these errors, pcregrep is abandoned.
3737
3738 (b) A binary zero was treated as data while matching, but terminated the
3739 output line if it was written out. This has been fixed: binary zeroes
3740 are now no different to any other data bytes.
3741
3742 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3743 used to set a locale for matching. The --locale=xxxx long option has
3744 been added (no short equivalent) to specify a locale explicitly on the
3745 pcregrep command, overriding the environment variables.
3746
3747 (d) When -B was used with -n, some line numbers in the output were one less
3748 than they should have been.
3749
3750 (e) Added the -o (--only-matching) option.
3751
3752 (f) If -A or -C was used with -c (count only), some lines of context were
3753 accidentally printed for the final match.
3754
3755 (g) Added the -H (--with-filename) option.
3756
3757 (h) The combination of options -rh failed to suppress file names for files
3758 that were found from directory arguments.
3759
3760 (i) Added the -D (--devices) and -d (--directories) options.
3761
3762 (j) Added the -F (--fixed-strings) option.
3763
3764 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3765
3766 (l) Added the --colo(u)r option.
3767
3768 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3769 is not present by default.
3770
3771 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3772 items such as (?R) or (?1), when the recursion could match a number of
3773 alternatives. If it matched one of the alternatives, but subsequently,
3774 outside the recursion, there was a failure, the code tried to back up into
3775 the recursion. However, because of the way PCRE is implemented, this is not
3776 possible, and the result was an incorrect result from the match.
3777
3778 In order to prevent this happening, the specification of recursion has
3779 been changed so that all such subpatterns are automatically treated as
3780 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3781
3782 4. I had overlooked the fact that, in some locales, there are characters for
3783 which isalpha() is true but neither isupper() nor islower() are true. In
3784 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3785 and ordfeminine) are like this. This affected the treatment of \w and \W
3786 when they appeared in character classes, but not when they appeared outside
3787 a character class. The bit map for "word" characters is now created
3788 separately from the results of isalnum() instead of just taking it from the
3789 upper, lower, and digit maps. (Plus the underscore character, of course.)
3790
3791 5. The above bug also affected the handling of POSIX character classes such as
3792 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3793 permanent tables. Instead, the bit maps for such a class were previously
3794 created as the appropriate unions of the upper, lower, and digit bitmaps.
3795 Now they are created by subtraction from the [[:word:]] class, which has
3796 its own bitmap.
3797
3798 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3799 It is created by subtracting the vertical space characters (\x09, \x0a,
3800 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3801 subtraction was done in the overall bitmap for a character class, meaning
3802 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3803 be recognized. This bug has been fixed.
3804
3805 7. Patches from the folks at Google:
3806
3807 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3808 real life, but is still worth protecting against".
3809
3810 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3811 regular expressions".
3812
3813 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3814 have it.
3815
3816 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3817 "configure" and the latter not, in order to fix a problem somebody had
3818 with compiling the Arg class on HP-UX.
3819
3820 (e) Improve the error-handling of the C++ wrapper a little bit.
3821
3822 (f) New tests for checking recursion limiting.
3823
3824 8. The pcre_memmove() function, which is used only if the environment does not
3825 have a standard memmove() function (and is therefore rarely compiled),
3826 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3827 returning a result (though PCRE never actually uses the result).
3828
3829 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3830 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3831 returned instead of calling malloc() with an overflowing number that would
3832 most likely cause subsequent chaos.
3833
3834 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3835
3836 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3837 with this option is matched, the nmatch and pmatch options of regexec() are
3838 ignored.
3839
3840 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3841 provided in case anyone wants to the the POSIX interface with UTF-8
3842 strings.
3843
3844 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3845 C++ linking (needed for some HP-UX environments).
3846
3847 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3848 (unused parameter) and in the pcre_printint() function (omitted "default"
3849 switch label when the default is to do nothing).
3850
3851 15. Added some code to make it possible, when PCRE is compiled as a C++
3852 library, to replace subject pointers for pcre_exec() with a smart pointer
3853 class, thus making it possible to process discontinuous strings.
3854
3855 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3856 much the same function. They were added by different people who were trying
3857 to make PCRE easy to compile on non-Unix systems. It has been suggested
3858 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3859 for compiling on Windows systems. I have therefore replaced it with
3860 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3861 defaults to "extern" for C or "extern C" for C++, which works fine on
3862 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3863 SCOPE with something explicit in config.h. In addition:
3864
3865 (a) pcreposix.h still had just "extern" instead of either of these macros;
3866 I have replaced it with PCRE_DATA_SCOPE.
3867
3868 (b) Functions such as _pcre_xclass(), which are internal to the library,
3869 but external in the C sense, all had PCRE_EXPORT in their definitions.
3870 This is apparently wrong for the Windows case, so I have removed it.
3871 (It makes no difference on Unix-like systems.)
3872
3873 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3874 of recursive calls to match(). This is different to MATCH_LIMIT because
3875 that limits the total number of calls to match(), not all of which increase
3876 the depth of recursion. Limiting the recursion depth limits the amount of
3877 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3878 when PCRE is compiled, and changed at run time. A patch from Google adds
3879 this functionality to the C++ interface.
3880
3881 18. Changes to the handling of Unicode character properties:
3882
3883 (a) Updated the table to Unicode 4.1.0.
3884
3885 (b) Recognize characters that are not in the table as "Cn" (undefined).
3886
3887 (c) I revised the way the table is implemented to a much improved format
3888 which includes recognition of ranges. It now supports the ranges that
3889 are defined in UnicodeData.txt, and it also amalgamates other
3890 characters into ranges. This has reduced the number of entries in the
3891 table from around 16,000 to around 3,000, thus reducing its size
3892 considerably. I realized I did not need to use a tree structure after
3893 all - a binary chop search is just as efficient. Having reduced the
3894 number of entries, I extended their size from 6 bytes to 8 bytes to
3895 allow for more data.
3896
3897 (d) Added support for Unicode script names via properties such as \p{Han}.
3898
3899 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3900 matching that character.
3901
3902 20. When matching a repeated Unicode property with a minimum greater than zero,
3903 (for example \pL{2,}), PCRE could look past the end of the subject if it
3904 reached it while seeking the minimum number of characters. This could
3905 happen only if some of the characters were more than one byte long, because
3906 there is a check for at least the minimum number of bytes.
3907
3908 21. Refactored the implementation of \p and \P so as to be more general, to
3909 allow for more different types of property in future. This has changed the
3910 compiled form incompatibly. Anybody with saved compiled patterns that use
3911 \p or \P will have to recompile them.
3912
3913 22. Added "Any" and "L&" to the supported property types.
3914
3915 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3916 but give a compile time error if the value is greater than 0xff.
3917
3918 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3919 accidentally not being installed or uninstalled.
3920
3921 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3922 made were to insert the current release number. This seemed silly, because
3923 it made things harder for people building PCRE on systems that don't run
3924 "configure". I have turned pcre.h into a distributed file, no longer built
3925 by "configure", with the version identification directly included. There is
3926 no longer a pcre.h.in file.
3927
3928 However, this change necessitated a change to the pcre-config script as
3929 well. It is built from pcre-config.in, and one of the substitutions was the
3930 release number. I have updated configure.ac so that ./configure now finds
3931 the release number by grepping pcre.h.
3932
3933 26. Added the ability to run the tests under valgrind.
3934
3935
3936 Version 6.4 05-Sep-05
3937 ---------------------
3938
3939 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3940 "--" to be printed when multiple files were scanned, even when none of the
3941 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3942 consider it to be a bug, and have restored the previous behaviour.
3943
3944 2. A couple of code tidies to get rid of compiler warnings.
3945
3946 3. The pcretest program used to cheat by referring to symbols in the library
3947 whose names begin with _pcre_. These are internal symbols that are not
3948 really supposed to be visible externally, and in some environments it is
3949 possible to suppress them. The cheating is now confined to including
3950 certain files from the library's source, which is a bit cleaner.
3951
3952 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3953 file's purpose clearer.
3954
3955 5. Reorganized pcre_ucp_findchar().
3956
3957
3958 Version 6.3 15-Aug-05
3959 ---------------------
3960
3961 1. The file libpcre.pc.in did not have general read permission in the tarball.
3962
3963 2. There were some problems when building without C++ support:
3964
3965 (a) If C++ support was not built, "make install" and "make test" still
3966 tried to test it.
3967
3968 (b) There were problems when the value of CXX was explicitly set. Some
3969 changes have been made to try to fix these, and ...
3970
3971 (c) --disable-cpp can now be used to explicitly disable C++ support.
3972
3973 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3974 backslash in a target when C++ was disabled. This confuses some
3975 versions of "make", apparently. Using an intermediate variable solves
3976 this. (Same for CPP_LOBJ.)
3977
3978 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3979 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3980 necessary on certain architectures.
3981
3982 4. Added a setting of -export-symbols-regex to the link command to remove
3983 those symbols that are exported in the C sense, but actually are local
3984 within the library, and not documented. Their names all begin with
3985 "_pcre_". This is not a perfect job, because (a) we have to except some
3986 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3987 available (and never for static libraries). I have made a note to try to
3988 find a way round (a) in the future.
3989
3990
3991 Version 6.2 01-Aug-05
3992 ---------------------
3993
3994 1. There was no test for integer overflow of quantifier values. A construction
3995 such as {1111111111111111} would give undefined results. What is worse, if
3996 a minimum quantifier for a parenthesized subpattern overflowed and became
3997 negative, the calculation of the memory size went wrong. This could have
3998 led to memory overwriting.
3999
4000 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4001
4002 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4003 operating environments where this matters.
4004
4005 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4006 PCRE options from within the C++ wrapper.
4007
4008 5. Named capturing subpatterns were not being correctly counted when a pattern
4009 was compiled. This caused two problems: (a) If there were more than 100
4010 such subpatterns, the calculation of the memory needed for the whole
4011 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4012 back references of the form \12, where the number was greater than 9, were
4013 not recognized as back references, even though there were sufficient
4014 previous subpatterns.
4015
4016 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4017 versions of gcc, e.g. 2.95.4.
4018
4019
4020 Version 6.1 21-Jun-05
4021 ---------------------
4022
4023 1. There was one reference to the variable "posix" in pcretest.c that was not
4024 surrounded by "#if !defined NOPOSIX".
4025
4026 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4027 the cross-check on the old pcre_info() function, for the benefit of the
4028 cut-down version of PCRE that is currently imported into Exim.
4029
4030 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4031 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4032 patterns that start (?i) but I don't think that matters. The use of (?i) is
4033 just an example; this all applies to the other options as well.
4034
4035 4. Since libtool seems to echo the compile commands it is issuing, the output
4036 from "make" can be reduced a bit by putting "@" in front of each libtool
4037 compile command.
4038
4039 5. Patch from the folks at Google for configure.in to be a bit more thorough
4040 in checking for a suitable C++ installation before trying to compile the
4041 C++ stuff. This should fix a reported problem when a compiler was present,
4042 but no suitable headers.
4043
4044 6. The man pages all had just "PCRE" as their title. I have changed them to
4045 be the relevant file name. I have also arranged that these names are
4046 retained in the file doc/pcre.txt, which is a concatenation in text format
4047 of all the man pages except the little individual ones for each function.
4048
4049 7. The NON-UNIX-USE file had not been updated for the different set of source
4050 files that come with release 6. I also added a few comments about the C++
4051 wrapper.
4052
4053
4054 Version 6.0 07-Jun-05
4055 ---------------------
4056
4057 1. Some minor internal re-organization to help with my DFA experiments.
4058
4059 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4060 didn't matter for the library itself when fully configured, but did matter
4061 when compiling without UCP support, or within Exim, where the ucp files are
4062 not imported.
4063
4064 3. Refactoring of the library code to split up the various functions into
4065 different source modules. The addition of the new DFA matching code (see
4066 below) to a single monolithic source would have made it really too
4067 unwieldy, quite apart from causing all the code to be include in a
4068 statically linked application, when only some functions are used. This is
4069 relevant even without the DFA addition now that patterns can be compiled in
4070 one application and matched in another.
4071
4072 The downside of splitting up is that there have to be some external
4073 functions and data tables that are used internally in different modules of
4074 the library but which are not part of the API. These have all had their
4075 names changed to start with "_pcre_" so that they are unlikely to clash
4076 with other external names.
4077
4078 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4079 a different (DFA) algorithm. Although it is slower than the original
4080 function, it does have some advantages for certain types of matching
4081 problem.
4082
4083 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4084 including restarting after a partial match.
4085
4086 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4087 defined when compiling for Windows was sent to me. I have put it into the
4088 code, though I have no means of testing or verifying it.
4089
4090 7. Added the pcre_refcount() auxiliary function.
4091
4092 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4093 match before or at the first newline in the subject string. In pcretest,
4094 the /f option on a pattern can be used to set this.
4095
4096 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4097 would behave wrongly. This has been present in PCRE since release 4.0.
4098
4099 10. A number of changes to the pcregrep command:
4100
4101 (a) Refactored how -x works; insert ^(...)$ instead of setting
4102 PCRE_ANCHORED and checking the length, in preparation for adding
4103 something similar for -w.
4104
4105 (b) Added the -w (match as a word) option.
4106
4107 (c) Refactored the way lines are read and buffered so as to have more
4108 than one at a time available.
4109
4110 (d) Implemented a pcregrep test script.
4111
4112 (e) Added the -M (multiline match) option. This allows patterns to match
4113 over several lines of the subject. The buffering ensures that at least
4114 8K, or the rest of the document (whichever is the shorter) is available
4115 for matching (and similarly the previous 8K for lookbehind assertions).
4116
4117 (f) Changed the --help output so that it now says
4118
4119 -w, --word-regex(p)
4120
4121 instead of two lines, one with "regex" and the other with "regexp"
4122 because that confused at least one person since the short forms are the
4123 same. (This required a bit of code, as the output is generated
4124 automatically from a table. It wasn't just a text change.)
4125
4126 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4127 option but starts with a hyphen. Could be a pattern or a path name
4128 starting with a hyphen, for instance.
4129
4130 (h) "-" can be given as a file name to represent stdin.
4131
4132 (i) When file names are being printed, "(standard input)" is used for
4133 the standard input, for compatibility with GNU grep. Previously
4134 "<stdin>" was used.
4135
4136 (j) The option --label=xxx can be used to supply a name to be used for
4137 stdin when file names are being printed. There is no short form.
4138
4139 (k) Re-factored the options decoding logic because we are going to add
4140 two more options that take data. Such options can now be given in four
4141 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4142
4143 (l) Added the -A, -B, and -C options for requesting that lines of context
4144 around matches be printed.
4145
4146 (m) Added the -L option to print the names of files that do not contain
4147 any matching lines, that is, the complement of -l.
4148
4149 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4150 continue to scan other files.
4151
4152 (o) The -s option was incorrectly implemented. For compatibility with other
4153 greps, it now suppresses the error message for a non-existent or non-
4154 accessible file (but not the return code). There is a new option called
4155 -q that suppresses the output of matching lines, which was what -s was
4156 previously doing.
4157
4158 (p) Added --include and --exclude options to specify files for inclusion
4159 and exclusion when recursing.
4160
4161 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4162 Hopefully, it now does.
4163
4164 12. Missing cast in pcre_study().
4165
4166 13. Added an "uninstall" target to the makefile.
4167
4168 14. Replaced "extern" in the function prototypes in Makefile.in with
4169 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4170 world, but is set differently for Windows.
4171
4172 15. Added a second compiling function called pcre_compile2(). The only
4173 difference is that it has an extra argument, which is a pointer to an
4174 integer error code. When there is a compile-time failure, this is set
4175 non-zero, in addition to the error test pointer being set to point to an
4176 error message. The new argument may be NULL if no error number is required
4177 (but then you may as well call pcre_compile(), which is now just a
4178 wrapper). This facility is provided because some applications need a
4179 numeric error indication, but it has also enabled me to tidy up the way
4180 compile-time errors are handled in the POSIX wrapper.
4181
4182 16. Added VPATH=.libs to the makefile; this should help when building with one
4183 prefix path and installing with another. (Or so I'm told by someone who
4184 knows more about this stuff than I do.)
4185
4186 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4187 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4188 match everything, including newlines. This is not POSIX-compatible, but
4189 somebody wanted the feature. From pcretest it can be activated by using
4190 both the P and the s flags.
4191
4192 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4193
4194 19. libpcre.pc was being incorrectly installed as executable.
4195
4196 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4197 it now also looks for '\r' so that it will work unmodified on Windows.
4198
4199 21. Added Google's contributed C++ wrapper to the distribution.
4200
4201 22. Added some untidy missing memory free() calls in pcretest, to keep
4202 Electric Fence happy when testing.
4203
4204
4205
4206 Version 5.0 13-Sep-04
4207 ---------------------
4208
4209 1. Internal change: literal characters are no longer packed up into items
4210 containing multiple characters in a single byte-string. Each character
4211 is now matched using a separate opcode. However, there may be more than one
4212 byte in the character in UTF-8 mode.
4213
4214 2. The pcre_callout_block structure has two new fields: pattern_position and
4215 next_item_length. These contain the offset in the pattern to the next match
4216 item, and its length, respectively.
4217
4218 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4219 insertion of callouts before each pattern item. Added the /C option to
4220 pcretest to make use of this.
4221
4222 4. On the advice of a Windows user, the lines
4223
4224 #if defined(_WIN32) || defined(WIN32)
4225 _setmode( _fileno( stdout ), 0x8000 );
4226 #endif /* defined(_WIN32) || defined(WIN32) */
4227
4228 have been added to the source of pcretest. This apparently does useful
4229 magic in relation to line terminators.
4230
4231 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4232 for the benefit of those environments where the "b" makes a difference.
4233
4234 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4235 to know about it. I have put a hack into configure.in that adds in code
4236 to set GCC=yes if CC=icc. This seems to end up at a point in the
4237 generated configure script that is early enough to affect the setting of
4238 compiler options, which is what is needed, but I have no means of testing
4239 whether it really works. (The user who reported this had patched the
4240 generated configure script, which of course I cannot do.)
4241
4242 LATER: After change 22 below (new libtool files), the configure script
4243 seems to know about icc (and also ecc). Therefore, I have commented out
4244 this hack in configure.in.
4245
4246 7. Added support for pkg-config (2 patches were sent in).
4247
4248 8. Negated POSIX character classes that used a combination of internal tables
4249 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4250 [[:^ascii]]. Typically, they would match almost any characters. The other
4251 POSIX classes were not broken in this way.
4252
4253 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4254 to find the match, as PCRE was deluded into thinking that the match had to
4255 start at the start point or following a newline. The same bug applied to
4256 patterns with negative forward assertions or any backward assertions
4257 preceding ".*" at the start, unless the pattern required a fixed first
4258 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4259
4260 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4261 starting at the last subject character, bytes beyond the end of the subject
4262 string were read.
4263
4264 11. Renamed the variable "class" as "classbits" to make life easier for C++
4265 users. (Previously there was a macro definition, but it apparently wasn't
4266 enough.)
4267
4268 12. Added the new field "tables" to the extra data so that tables can be passed
4269 in at exec time, or the internal tables can be re-selected. This allows
4270 a compiled regex to be saved and re-used at a later time by a different
4271 program that might have everything at different addresses.
4272
4273 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4274 -R library as well as a -L library.
4275
4276 14. The debugging options of pcretest (-d on the command line or D on a
4277 pattern) showed incorrect output for anything following an extended class
4278 that contained multibyte characters and which was followed by a quantifier.
4279
4280 15. Added optional support for general category Unicode character properties
4281 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4282 support. It adds about 90K to the size of the library. The meanings of the
4283 inbuilt class escapes such as \d and \s have NOT been changed.
4284
4285 16. Updated pcredemo.c to include calls to free() to release the memory for the
4286 compiled pattern.
4287
4288 17. The generated file chartables.c was being created in the source directory
4289 instead of in the building directory. This caused the build to fail if the
4290 source directory was different from the building directory, and was
4291 read-only.
4292
4293 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4294 file. No doubt somebody will tell me if they don't make sense... Also added
4295 Dan Mooney's comments about building on OpenVMS.
4296
4297 19. Added support for partial matching via the PCRE_PARTIAL option for
4298 pcre_exec() and the \P data escape in pcretest.
4299
4300 20. Extended pcretest with 3 new pattern features:
4301
4302 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4303 write the compiled pattern to the file whose name is "rest-of-line".
4304 This is a straight binary dump of the data, with the saved pointer to
4305 the character tables forced to be NULL. The study data, if any, is
4306 written too. After writing, pcretest reads a new pattern.
4307
4308 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4309 compiled pattern from the given file. There must not be any
4310 occurrences of "<" in the file name (pretty unlikely); if there are,
4311 pcretest will instead treat the initial "<" as a pattern delimiter.
4312 After reading in the pattern, pcretest goes on to read data lines as
4313 usual.
4314
4315 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4316 and 16-bit fields in a compiled pattern, to simulate a pattern that
4317 was compiled on a host of opposite endianness.
4318
4319 21. The pcre-exec() function can now cope with patterns that were compiled on
4320 hosts of opposite endianness, with this restriction:
4321
4322 As for any compiled expression that is saved and used later, the tables
4323 pointer field cannot be preserved; the extra_data field in the arguments
4324 to pcre_exec() should be used to pass in a tables address if a value
4325 other than the default internal tables were used at compile time.
4326
4327 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4328 now diagnosed as an error. Previously, most of the time, a negative number
4329 would have been treated as zero, but if in addition "ovector" was passed as
4330 NULL, a crash could occur.
4331
4332 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4333 new versions from the libtool 1.5 distribution (the last one is a copy of
4334 a file called libtool.m4). This seems to have fixed the need to patch
4335 "configure" to support Darwin 1.3 (which I used to do). However, I still
4336 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4337 workstation).
4338
4339 24. Changed the PCRE licence to be the more standard "BSD" licence.
4340
4341
4342 Version 4.5 01-Dec-03
4343 ---------------------
4344
4345 1. There has been some re-arrangement of the code for the match() function so
4346 that it can be compiled in a version that does not call itself recursively.
4347 Instead, it keeps those local variables that need separate instances for
4348 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4349 needs to "recurse". Keeping track of where control must go is done by means
4350 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4351 hide most of the details from the main code, and operates only if
4352 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4353 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4354 operating.
4355
4356 To make it easier for callers to provide specially tailored get/free
4357 functions for this usage, two new functions, pcre_stack_malloc, and
4358 pcre_stack_free, are used. They are always called in strict stacking order,
4359 and the size of block requested is always the same.
4360
4361 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4362 PCRE has been compiled to use the stack or the heap for recursion. The
4363 -C option of pcretest uses this to show which version is compiled.
4364
4365 A new data escape \S, is added to pcretest; it causes the amounts of store
4366 obtained and freed by both kinds of malloc/free at match time to be added
4367 to the output.
4368
4369 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4370 what's available on my current Linux desktop machine.
4371
4372 3. When matching a UTF-8 string, the test for a valid string at the start has
4373 been extended. If start_offset is not zero, PCRE now checks that it points
4374 to a byte that is the start of a UTF-8 character. If not, it returns
4375 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4376 this is necessary because there may be backward assertions in the pattern.
4377 When matching the same subject several times, it may save resources to use
4378 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4379
4380 4. The code for checking the validity of UTF-8 strings has been tightened so
4381 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4382 containing "overlong sequences".
4383
4384 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4385 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4386 should have been "&", but it just so happened that all the cases this let
4387 through by mistake were picked up later in the function.
4388
4389 6. I had used a variable called "isblank" - this is a C99 function, causing
4390 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4391
4392 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4393 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4394 so the version is shown; (c) stop "make test" echoing "./RunTest".
4395
4396 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4397
4398 9. The prototype for memmove() for systems that don't have it was using
4399 size_t, but the inclusion of the header that defines size_t was later. I've
4400 moved the #includes for the C headers earlier to avoid this.
4401
4402 10. Added some adjustments to the code to make it easier to compiler on certain
4403 special systems:
4404
4405 (a) Some "const" qualifiers were missing.
4406 (b) Added the macro EXPORT before all exported functions; by default this
4407 is defined to be empty.
4408 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4409 that it reads its output file name as an argument instead of writing
4410 to the standard output and assuming this can be redirected.
4411
4412 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4413 class containing characters with values greater than 255, PCRE compilation
4414 went into a loop.
4415
4416 12. A recursive reference to a subpattern that was within another subpattern
4417 that had a minimum quantifier of zero caused PCRE to crash. For example,
4418 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4419 recursion. If the recursively-called subpattern itself had a zero repeat,
4420 that was OK.
4421
4422 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4423 buffer into which it was copied (for escape processing) was still set at
4424 1024, so long lines caused crashes.
4425
4426 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4427 "internal error: code overflow...". This applied to any character class
4428 that was followed by a possessive quantifier.
4429
4430 15. Modified the Makefile to add libpcre.la as a prerequisite for
4431 libpcreposix.la because I was told this is needed for a parallel build to
4432 work.
4433
4434 16. If a pattern that contained .* following optional items at the start was
4435 studied, the wrong optimizing data was generated, leading to matching
4436 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4437 matching string must start with a or b or c. The correct conclusion for
4438 this pattern is that a match can start with any character.
4439
4440
4441 Version 4.4 13-Aug-03
4442 ---------------------
4443
4444 1. In UTF-8 mode, a character class containing characters with values between
4445 127 and 255 was not handled correctly if the compiled pattern was studied.
4446 In fixing this, I have also improved the studying algorithm for such
4447 classes (slightly).
4448
4449 2. Three internal functions had redundant arguments passed to them. Removal
4450 might give a very teeny performance improvement.
4451
4452 3. Documentation bug: the value of the capture_top field in a callout is *one
4453 more than* the number of the hightest numbered captured substring.
4454
4455 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4456 in incorrectly linking with a previously installed version. They now link
4457 explicitly with libpcre.la.
4458
4459 5. configure.in no longer needs to recognize Cygwin specially.
4460
4461 6. A problem in pcre.in for Windows platforms is fixed.
4462
4463 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4464 pcretest, it used to include the size of the study block as part of its
4465 output. Unfortunately, the structure contains a field that has a different
4466 size on different hardware architectures. This meant that the tests that
4467 showed this size failed. As the block is currently always of a fixed size,
4468 this information isn't actually particularly useful in pcretest output, so
4469 I have just removed it.
4470
4471 8. Three pre-processor statements accidentally did not start in column 1.
4472 Sadly, there are *still* compilers around that complain, even though
4473 standard C has not required this for well over a decade. Sigh.
4474
4475 9. In pcretest, the code for checking callouts passed small integers in the
4476 callout_data field, which is a void * field. However, some picky compilers
4477 complained about the casts involved for this on 64-bit systems. Now
4478 pcretest passes the address of the small integer instead, which should get
4479 rid of the warnings.
4480
4481 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4482 both compile and run time, and gives an error if an invalid UTF-8 sequence
4483 is found. There is a option for disabling this check in cases where the
4484 string is known to be correct and/or the maximum performance is wanted.
4485
4486 11. In response to a bug report, I changed one line in Makefile.in from
4487
4488 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4489 to
4490 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4491
4492 to look similar to other lines, but I have no way of telling whether this
4493 is the right thing to do, as I do not use Windows. No doubt I'll get told
4494 if it's wrong...
4495
4496
4497 Version 4.3 21-May-03
4498 ---------------------
4499
4500 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4501 Makefile.
4502
4503 2. Some refactoring to improve the quality of the code:
4504
4505 (i) The utf8_table... variables are now declared "const".
4506
4507 (ii) The code for \cx, which used the "case flipping" table to upper case
4508 lower case letters, now just substracts 32. This is ASCII-specific,
4509 but the whole concept of \cx is ASCII-specific, so it seems
4510 reasonable.
4511
4512 (iii) PCRE was using its character types table to recognize decimal and
4513 hexadecimal digits in the pattern. This is silly, because it handles
4514 only 0-9, a-f, and A-F, but the character types table is locale-
4515 specific, which means strange things might happen. A private
4516 table is now used for this - though it costs 256 bytes, a table is
4517 much faster than multiple explicit tests. Of course, the standard
4518 character types table is still used for matching digits in subject
4519 strings against \d.
4520
4521 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4522 ending in _t are). So I've renamed it as ESC_tee.
4523
4524 3. The first argument for regexec() in the POSIX wrapper should have been
4525 defined as "const".
4526
4527 4. Changed pcretest to use malloc() for its buffers so that they can be
4528 Electric Fenced for debugging.
4529
4530 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4531 to read one or more bytes before the start of the subject string. Often this
4532 had no effect on PCRE's behaviour, but in some circumstances it could
4533 provoke a segmentation fault.
4534
4535 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4536 to try to read one or more bytes before the start of the subject string.
4537
4538 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4539 UTF-8 support could misbehave in various ways if the subject string
4540 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4541 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4542 back over UTF-8 characters.)
4543
4544
4545 Version 4.2 14-Apr-03
4546 ---------------------
4547
4548 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4549
4550 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4551 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4552 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4553 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4554 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4555 and BUILD_EXEEXT
4556 Note: automatic setting of the BUILD variables is not yet working
4557 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4558 compile-time but not at link-time
4559 [LINK]: use for linking executables only
4560 make different versions for Windows and non-Windows
4561 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4562 libraries
4563 [LINK_FOR_BUILD]: new variable
4564 [OBJEXT]: use throughout
4565 [EXEEXT]: use throughout
4566 <winshared>: new target
4567 <wininstall>: new target
4568 <dftables.o>: use native compiler
4569 <dftables>: use native linker
4570 <install>: handle Windows platform correctly
4571 <clean>: ditto
4572 <check>: ditto
4573 copy DLL to top builddir before testing
4574
4575 As part of these changes, -no-undefined was removed again. This was reported
4576 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4577 in any case.
4578
4579 3. Some tidies to get rid of compiler warnings:
4580
4581 . In the match_data structure, match_limit was an unsigned long int, whereas
4582 match_call_count was an int. I've made them both unsigned long ints.
4583
4584 . In pcretest the fact that a const uschar * doesn't automatically cast to
4585 a void * provoked a warning.
4586
4587 . Turning on some more compiler warnings threw up some "shadow" variables
4588 and a few more missing casts.
4589
4590 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4591 option, a class that contained a single character with a value between 128
4592 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4593
4594 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4595 option, a class that contained several characters, but with at least one
4596 whose value was between 128 and 255 caused PCRE to crash.
4597
4598
4599 Version 4.1 12-Mar-03
4600 ---------------------
4601
4602 1. Compiling with gcc -pedantic found a couple of places where casts were
4603 needed, and a string in dftables.c that was longer than standard compilers are
4604 required to support.
4605
4606 2. Compiling with Sun's compiler found a few more places where the code could
4607 be tidied up in order to avoid warnings.
4608
4609 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4610 first of these names is deprecated in the latest Autoconf in favour of the name
4611 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4612 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4613 analogy I have changed it to CFLAGS_FOR_BUILD.
4614
4615 4. Added -no-undefined to the linking command in the Makefile, because this is
4616 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4617 linking step for the pcreposix library.
4618
4619 5. PCRE was failing to diagnose the case of two named groups with the same
4620 name.
4621
4622 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4623 literal character that is needed in the subject for a match, and scans along to
4624 ensure that it is present before embarking on the full matching process. This
4625 saves time in cases of nested unlimited repeats that are never going to match.
4626 Problem: the scan can take a lot of time if the subject is very long (e.g.
4627 megabytes), thus penalizing straightforward matches. It is now done only if the
4628 amount of subject to be scanned is less than 1000 bytes.
4629
4630 7. A lesser problem with the same optimization is that it was recording the
4631 first character of an anchored pattern as "needed", thus provoking a search
4632 right along the subject, even when the first match of the pattern was going to
4633 fail. The "needed" character is now not set for anchored patterns, unless it
4634 follows something in the pattern that is of non-fixed length. Thus, it still
4635 fulfils its original purpose of finding quick non-matches in cases of nested
4636 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4637
4638
4639 Version 4.0 17-Feb-03
4640 ---------------------
4641
4642 1. If a comment in an extended regex that started immediately after a meta-item
4643 extended to the end of string, PCRE compiled incorrect data. This could lead to
4644 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4645
4646 2. Moved to autoconf 2.53 and libtool 1.4.2.
4647
4648 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4649 the special perltest8 script is no longer needed - all the tests can be run
4650 from a single perltest script.
4651
4652 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4653 by \s. It has now been removed in PCRE. This means it isn't recognized as
4654 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4655 class [:space:] *does* include VT, thereby creating a mess.
4656
4657 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4658 space and tab.
4659
4660 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4661 its new features into the main test script, reducing the number of scripts.
4662
4663 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4664 were backward compatible, and made the (?i) apply to the whole pattern, as if
4665 /i were given. Now it behaves more logically, and applies the option setting
4666 only to what follows. PCRE has been changed to follow suit. However, if it
4667 finds options settings right at the start of the pattern, it extracts them into
4668 the global options, as before. Thus, they show up in the info data.
4669
4670 8. Added support for the \Q...\E escape sequence. Characters in between are
4671 treated as literals. This is slightly different from Perl in that $ and @ are
4672 also handled as literals inside the quotes. In Perl, they will cause variable
4673 interpolation. Note the following examples:
4674
4675 Pattern PCRE matches Perl matches
4676
4677 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4678 \Qabc\$xyz\E abc\$xyz abc\$xyz
4679 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4680
4681 For compatibility with Perl, \Q...\E sequences are recognized inside character
4682 classes as well as outside them.
4683
4684 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4685 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4686 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4687 signed/unsigned warnings.
4688
4689 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4690 option for pcretest, so I've replaced it by a simple function that does just
4691 that job.
4692
4693 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4694 "pcregrep -".
4695
4696 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4697 Java package. This provides some syntactic sugar for simple cases of what my
4698 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4699 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4700 item, you can use this simplified notation. Note that only makes sense with
4701 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4702 greediness, whatever the setting of the PCRE_UNGREEDY option.
4703
4704 13. A change of greediness default within a pattern was not taking effect at
4705 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4706 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4707 was abstracted outside.
4708
4709 14. PCRE now supports the \G assertion. It is true when the current matching
4710 position is at the start point of the match. This differs from \A when the
4711 starting offset is non-zero. Used with the /g option of pcretest (or similar
4712 code), it works in the same way as it does for Perl's /g option. If all
4713 alternatives of a regex begin with \G, the expression is anchored to the start
4714 match position, and the "anchored" flag is set in the compiled expression.
4715
4716 15. Some bugs concerning the handling of certain option changes within patterns
4717 have been fixed. These applied to options other than (?ims). For example,
4718 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4719 been the other way round. Some of this was related to change 7 above.
4720
4721 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4722 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4723 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4724 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4725
4726 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4727 mode. Unlike ".", it always matches newline, whatever the setting of
4728 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4729 assertions. Perl allows it, but it doesn't (in general) work because it can't
4730 calculate the length of the lookbehind. At least, that's the case for Perl
4731 5.8.0 - I've been told they are going to document that it doesn't work in
4732 future.
4733
4734 18. Added an error diagnosis for escapes that PCRE does not support: these are
4735 \L, \l, \N, \P, \p, \U, \u, and \X.
4736
4737 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4738 reading past the end of the pattern in cases such as /[abcd/.
4739
4740 20. PCRE was getting more memory than necessary for patterns with classes that
4741 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4742
4743 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4744 compiling PCRE for use with Virtual Pascal.
4745
4746 22. Small fix to the Makefile to make it work properly if the build is done
4747 outside the source tree.
4748
4749 23. Added a new extension: a condition to go with recursion. If a conditional
4750 subpattern starts with (?(R) the "true" branch is used if recursion has
4751 happened, whereas the "false" branch is used only at the top level.
4752
4753 24. When there was a very long string of literal characters (over 255 bytes
4754 without UTF support, over 250 bytes with UTF support), the computation of how
4755 much memory was required could be incorrect, leading to segfaults or other
4756 strange effects.
4757
4758 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4759 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4760 there was a subsequent back reference to those brackets. This meant that, for
4761 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4762 possible to check for precisely this case. All we can do is abandon the
4763 optimization if .* occurs inside capturing brackets when there are any back
4764 references whatsoever. (See below for a better fix that came later.)
4765
4766 26. The handling of the optimization for finding the first character of a
4767 non-anchored pattern, and for finding a character that is required later in the
4768 match were failing in some cases. This didn't break the matching; it just
4769 failed to optimize when it could. The way this is done has been re-implemented.
4770
4771 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4772
4773 28. Added a new feature that provides some of the functionality that Perl
4774 provides with (?{...}). The facility is termed a "callout". The way it is done
4775 in PCRE is for the caller to provide an optional function, by setting
4776 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4777 global variable. By default it is unset, which disables all calling out. To get
4778 the function called, the regex must include (?C) at appropriate points. This
4779 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4780 This provides a means of identifying different callout points. When PCRE
4781 reaches such a point in the regex, if pcre_callout has been set, the external
4782 function is called. It is provided with data in a structure called
4783 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4784 matching continues; if it returns a non-zero value, the match at the current
4785 point fails. However, backtracking will occur if possible. [This was changed
4786 later and other features added - see item 49 below.]
4787
4788 29. pcretest is upgraded to test the callout functionality. It provides a
4789 callout function that displays information. By default, it shows the start of
4790 the match and the current position in the text. There are some new data escapes
4791 to vary what happens:
4792
4793 \C+ in addition, show current contents of captured substrings
4794 \C- do not supply a callout function
4795 \C!n return 1 when callout number n is reached
4796 \C!n!m return 1 when callout number n is reached for the mth time
4797
4798 30. If pcregrep was called with the -l option and just a single file name, it
4799 output "<stdin>" if a match was found, instead of the file name.
4800
4801 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4802 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4803 pcre_exec(). This saves a malloc/free per call. The default value of
4804 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4805 when configuring.
4806
4807 32. The default maximum size of a compiled pattern is 64K. There have been a
4808 few cases of people hitting this limit. The code now uses macros to handle the
4809 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4810 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4811 configuring. Tests 2 and 5 work only with 2-byte links because they output
4812 debugging information about compiled patterns.
4813
4814 33. Internal code re-arrangements:
4815
4816 (a) Moved the debugging function for printing out a compiled regex into
4817 its own source file (printint.c) and used #include to pull it into
4818 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4819 separate copies.
4820
4821 (b) Defined the list of op-code names for debugging as a macro in
4822 internal.h so that it is next to the definition of the opcodes.
4823
4824 (c) Defined a table of op-code lengths for simpler skipping along compiled
4825 code. This is again a macro in internal.h so that it is next to the
4826 definition of the opcodes.
4827
4828 34. Added support for recursive calls to individual subpatterns, along the
4829 lines of Robin Houston's patch (but implemented somewhat differently).
4830
4831 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4832 allow it to read and process whole directories in Win32. This code was
4833 contributed by Lionel Fourquaux; it has not been tested by me.
4834
4835 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4836 used to name a group. Names consist of alphanumerics and underscores, and must
4837 be unique. Back references use the syntax (?P=name) and recursive calls use
4838 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4839 numbers. The function pcre_fullinfo() can be used after compilation to extract
4840 a name/number map. There are three relevant calls:
4841
4842 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4843 PCRE_INFO_NAMECOUNT yields the number of entries
4844 PCRE_INFO_NAMETABLE yields a pointer to the map.
4845
4846 The map is a vector of fixed-size entries. The size of each entry depends on
4847 the length of the longest name used. The first two bytes of each entry are the
4848 group number, most significant byte first. There follows the corresponding
4849 name, zero terminated. The names are in alphabetical order.
4850
4851 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4852 case instead of 255. Making it the same both with and without UTF-8 support
4853 means that the same test output works with both.
4854
4855 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4856 calling malloc() with a zero argument.
4857
4858 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4859 optimization. I've improved things by keeping a bitmap of backreferences with
4860 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4861 fact referenced, the optimization can be applied. It is unlikely that a
4862 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4863 the match to follow \n) will appear inside brackets with a number greater than
4864 31, but if it does, any back reference > 31 suppresses the optimization.
4865
4866 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4867 of disabling numbered capturing parentheses. Any opening parenthesis that is
4868 not followed by ? behaves as if it were followed by ?: but named parentheses
4869 can still be used for capturing (and they will acquire numbers in the usual
4870 way).
4871
4872 41. Redesigned the return codes from the match() function into yes/no/error so
4873 that errors can be passed back from deep inside the nested calls. A malloc
4874 failure while inside a recursive subpattern call now causes the
4875 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4876
4877 42. It is now possible to set a limit on the number of times the match()
4878 function is called in a call to pcre_exec(). This facility makes it possible to
4879 limit the amount of recursion and backtracking, though not in a directly
4880 obvious way, because the match() function is used in a number of different
4881 circumstances. The count starts from zero for each position in the subject
4882 string (for non-anchored patterns). The default limit is, for compatibility, a
4883 large number, namely 10 000 000. You can change this in two ways:
4884
4885 (a) When configuring PCRE before making, you can use --with-match-limit=n
4886 to set a default value for the compiled library.
4887
4888 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4889 a different value is set. See 45 below.
4890
4891 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4892
4893 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4894 of things that can be changed at compile time. The first argument specifies
4895 what is wanted and the second points to where the information is to be placed.
4896 The current list of available information is:
4897
4898 PCRE_CONFIG_UTF8
4899
4900 The output is an integer that is set to one if UTF-8 support is available;
4901 otherwise it is set to zero.
4902
4903 PCRE_CONFIG_NEWLINE
4904
4905 The output is an integer that it set to the value of the code that is used for
4906 newline. It is either LF (10) or CR (13).
4907
4908 PCRE_CONFIG_LINK_SIZE
4909
4910 The output is an integer that contains the number of bytes used for internal
4911 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4912
4913 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4914
4915 The output is an integer that contains the threshold above which the POSIX
4916 interface uses malloc() for output vectors. See item 31 above.
4917
4918 PCRE_CONFIG_MATCH_LIMIT
4919
4920 The output is an unsigned integer that contains the default limit of the number
4921 of match() calls in a pcre_exec() execution. See 42 above.
4922
4923 44. pcretest has been upgraded by the addition of the -C option. This causes it
4924 to extract all the available output from the new pcre_config() function, and to
4925 output it. The program then exits immediately.
4926
4927 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4928 order to support additional features. One way would have been to define
4929 pcre_exec2() (for example) with extra arguments, but this would not have been
4930 extensible, and would also have required all calls to the original function to
4931 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4932 is used for passing in "extra" data from pcre_study().
4933
4934 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4935 contains the following fields:
4936
4937 flags a bitmap indicating which of the following fields are set
4938 study_data opaque data from pcre_study()
4939 match_limit a way of specifying a limit on match() calls for a specific
4940 call to pcre_exec()
4941 callout_data data for callouts (see 49 below)
4942
4943 The flag bits are also defined in pcre.h, and are
4944
4945 PCRE_EXTRA_STUDY_DATA
4946 PCRE_EXTRA_MATCH_LIMIT
4947 PCRE_EXTRA_CALLOUT_DATA
4948
4949 The pcre_study() function now returns one of these new pcre_extra blocks, with
4950 the actual study data pointed to by the study_data field, and the
4951 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4952 before. That is, this change is entirely upwards-compatible and requires no
4953 change to existing code.
4954
4955 If you want to pass in additional data to pcre_exec(), you can either place it
4956 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4957 block.
4958
4959 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4960 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4961 times with different match limits, until it finds the minimum value needed for
4962 pcre_exec() to complete. The value is then output. This can be instructive; for
4963 most simple matches the number is quite small, but for pathological cases it
4964 gets very large very quickly.
4965
4966 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4967 returns the size of the data block pointed to by the study_data field in a
4968 pcre_extra block, that is, the value that was passed as the argument to
4969 pcre_malloc() when PCRE was getting memory in which to place the information
4970 created by pcre_study(). The fourth argument should point to a size_t variable.
4971 pcretest has been extended so that this information is shown after a successful
4972 pcre_study() call when information about the compiled regex is being displayed.
4973
4974 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4975 because what follows is always an absolute path. (Later: it turns out that this
4976 is more than cosmetic for MinGW, because it doesn't like empty path
4977 components.)
4978
4979 49. Some changes have been made to the callout feature (see 28 above):
4980
4981 (i) A callout function now has three choices for what it returns:
4982
4983 0 => success, carry on matching
4984 > 0 => failure at this point, but backtrack if possible
4985 < 0 => serious error, return this value from pcre_exec()
4986
4987 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4988 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4989 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4990 use by callout functions. It will never be used by PCRE itself.
4991
4992 (ii) The pcre_extra structure (see 45 above) has a void * field called
4993 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4994 pcre_callout_block structure has a field of the same name. The contents of
4995 the field passed in the pcre_extra structure are passed to the callout
4996 function in the corresponding field in the callout block. This makes it
4997 easier to use the same callout-containing regex from multiple threads. For
4998 testing, the pcretest program has a new data escape
4999
5000 \C*n pass the number n (may be negative) as callout_data
5001
5002 If the callout function in pcretest receives a non-zero value as
5003 callout_data, it returns that value.
5004
5005 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5006 there were some redundant $(CFLAGS) in commands that are now specified as
5007 $(LINK), which already includes $(CFLAGS).
5008
5009 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5010 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5011 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5012 one-byte characters throughout. Note that case-insensitive matching applies
5013 only to characters whose values are less than 256. PCRE doesn't support the
5014 notion of cases for higher-valued characters.
5015
5016 (i) A character class whose characters are all within 0-255 is handled as
5017 a bit map, and the map is inverted for negative classes. Previously, a
5018 character > 255 always failed to match such a class; however it should
5019 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5020
5021 (ii) A negated character class with a single character < 255 is coded as
5022 "not this character" (OP_NOT). This wasn't working properly when the test
5023 character was multibyte, either singly or repeated.
5024
5025 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5026 mode, for example: \x{100}{2,3}.
5027
5028 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5029 singly or repeated) now correctly test multibyte characters. However,
5030 PCRE doesn't recognize any characters with values greater than 255 as
5031 digits, spaces, or word characters. Such characters always match \D, \S,
5032 and \W, and never match \d, \s, or \w.
5033
5034 (v) Classes may now contain characters and character ranges with values
5035 greater than 255. For example: [ab\x{100}-\x{400}].
5036
5037 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5038 PCRE in UTF-8 mode.
5039
5040 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5041 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5042 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5043 value.)
5044
5045 53. The single man page has become too large. I have therefore split it up into
5046 a number of separate man pages. These also give rise to individual HTML pages;
5047 these are now put in a separate directory, and there is an index.html page that
5048 lists them all. Some hyperlinking between the pages has been installed.
5049
5050 54. Added convenience functions for handling named capturing parentheses.
5051
5052 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5053 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5054 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5055 are faulted.
5056
5057 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5058 calling configure. These values are used when compiling the dftables.c program
5059 which is run to generate the source of the default character tables. They
5060 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5061 you will need to set these values.
5062
5063 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5064
5065
5066 Version 3.9 02-Jan-02
5067 ---------------------
5068
5069 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5070
5071 2. If --disable-static was given, the building process failed when trying to
5072 build pcretest and pcregrep. (For some reason it was using libtool to compile
5073 them, which is not right, as they aren't part of the library.)
5074
5075
5076 Version 3.8 18-Dec-01
5077 ---------------------
5078
5079 1. The experimental UTF-8 code was completely screwed up. It was packing the
5080 bytes in the wrong order. How dumb can you get?
5081
5082
5083 Version 3.7 29-Oct-01
5084 ---------------------
5085
5086 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5087 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5088 this didn't happen under Solaris 8, where I normally test things.
5089
5090 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5091 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5092 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5093 this fix an hour or so after the initial 3.7 release.)
5094
5095
5096 Version 3.6 23-Oct-01
5097 ---------------------
5098
5099 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5100 offsets passed as NULL with zero offset count.
5101
5102 2. The config.guess and config.sub files had not been updated when I moved to
5103 the latest autoconf.
5104
5105
5106 Version 3.5 15-Aug-01
5107 ---------------------
5108
5109 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5110 had been forgotten.
5111
5112 2. By using declared but undefined structures, we can avoid using "void"
5113 definitions in pcre.h while keeping the internal definitions of the structures
5114 private.
5115
5116 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5117 user point of view, this means that both static and shared libraries are built
5118 by default, but this can be individually controlled. More of the work of
5119 handling this static/shared cases is now inside libtool instead of PCRE's make
5120 file.
5121
5122 4. The pcretest utility is now installed along with pcregrep because it is
5123 useful for users (to test regexs) and by doing this, it automatically gets
5124 relinked by libtool. The documentation has been turned into a man page, so
5125 there are now .1, .txt, and .html versions in /doc.
5126
5127 5. Upgrades to pcregrep:
5128 (i) Added long-form option names like gnu grep.
5129 (ii) Added --help to list all options with an explanatory phrase.
5130 (iii) Added -r, --recursive to recurse into sub-directories.
5131 (iv) Added -f, --file to read patterns from a file.
5132
5133 6. pcre_exec() was referring to its "code" argument before testing that
5134 argument for NULL (and giving an error if it was NULL).
5135
5136 7. Upgraded Makefile.in to allow for compiling in a different directory from
5137 the source directory.
5138
5139 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5140 options bits, the pointer it was passed was to an int instead of to an unsigned
5141 long int. This mattered only on 64-bit systems.
5142
5143 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5144 generated) instead of pcre.in, which it its source. Also made the same change
5145 in several of the .c files.
5146
5147 10. A new release of gcc defines printf() as a macro, which broke pcretest
5148 because it had an ifdef in the middle of a string argument for printf(). Fixed
5149 by using separate calls to printf().
5150
5151 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5152 script, to force use of CR or LF instead of \n in the source. On non-Unix
5153 systems, the value can be set in config.h.
5154
5155 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5156 absolute limit. Changed the text of the error message to make this clear, and
5157 likewise updated the man page.
5158
5159 13. The limit of 99 on the number of capturing subpatterns has been removed.
5160 The new limit is 65535, which I hope will not be a "real" limit.
5161
5162
5163 Version 3.4 22-Aug-00
5164 ---------------------
5165
5166 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5167
5168 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5169
5170
5171 Version 3.3 01-Aug-00
5172 ---------------------
5173
5174 1. If an octal character was given, but the value was greater than \377, it
5175 was not getting masked to the least significant bits, as documented. This could
5176 lead to crashes in some systems.
5177
5178 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5179 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5180
5181 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5182 These just pass their arguments on to (pcre_free)(), but they are provided
5183 because some uses of PCRE bind it to non-C systems that can call its functions,
5184 but cannot call free() or pcre_free() directly.
5185
5186 4. Add "make test" as a synonym for "make check". Corrected some comments in
5187 the Makefile.
5188
5189 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5190 Makefile.
5191
5192 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5193 command called pgrep for grepping around the active processes.
5194
5195 7. Added the beginnings of support for UTF-8 character strings.
5196
5197 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5198 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5199 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5200 out for the ar command.)
5201
5202
5203 Version 3.2 12-May-00
5204 ---------------------
5205
5206 This is purely a bug fixing release.
5207
5208 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5209 of ZA. This was just one example of several cases that could provoke this bug,
5210 which was introduced by change 9 of version 2.00. The code for breaking
5211 infinite loops after an iteration that matches an empty string was't working
5212 correctly.
5213
5214 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5215 when matched against abbab (for example). After matching an empty string, it
5216 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5217 caused it to match further down the string than it should.
5218
5219 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5220 was there because it doesn't seem to be needed, and it causes trouble on some
5221 systems, as it is not a Standard C header. It has been removed.
5222
5223 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5224 were reported on the Macintosh. The changes were from
5225
5226 while ((c = *(++ptr)) != 0 && c != '\n');
5227 to
5228 while ((c = *(++ptr)) != 0 && c != '\n') ;
5229
5230 Totally extraordinary, but if that's what it takes...
5231
5232 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5233 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5234 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5235 assumes the way PCRE uses memmove() (always moving upwards).
5236
5237 6. PCRE is being used in one environment where strchr() is not available. There
5238 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5239 faster code anyway.
5240
5241
5242 Version 3.1 09-Feb-00
5243 ---------------------
5244
5245 The only change in this release is the fixing of some bugs in Makefile.in for
5246 the "install" target:
5247
5248 (1) It was failing to install pcreposix.h.
5249
5250 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5251
5252
5253 Version 3.0 01-Feb-00
5254 ---------------------
5255
5256 1. Add support for the /+ modifier to perltest (to output $` like it does in
5257 pcretest).
5258
5259 2. Add support for the /g modifier to perltest.
5260
5261 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5262 matches null strings.
5263
5264 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5265 pattern. Perl treats empty patterns specially - it reuses the most recent
5266 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5267 effect.
5268
5269 5. The POSIX interface was broken in that it was just handing over the POSIX
5270 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5271 required a bigger vector, with some working space on the end. This means that
5272 the POSIX wrapper now has to get and free some memory, and copy the results.
5273
5274 6. Added some simple autoconf support, placing the test data and the
5275 documentation in separate directories, re-organizing some of the
5276 information files, and making it build pcre-config (a GNU standard). Also added
5277 libtool support for building PCRE as a shared library, which is now the
5278 default.
5279
5280 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5281 09 are not valid octal constants. Single digits will be used for minor values
5282 less than 10.
5283
5284 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5285 existing programs that set these in the POSIX interface can use PCRE without
5286 modification.
5287
5288 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5289 return all that pcre_info() returns, plus additional data. The pcre_info()
5290 function is retained for compatibility, but is considered to be obsolete.
5291
5292 10. Added experimental recursion feature (?R) to handle one common case that
5293 Perl 5.6 will be able to do with (?p{...}).
5294
5295 11. Added support for POSIX character classes like [:alpha:], which Perl is
5296 adopting.
5297
5298
5299 Version 2.08 31-Aug-99
5300 ----------------------
5301
5302 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5303 trying to match at the startoffset position, but instead was moving forward to
5304 the next newline as if a previous match had failed.
5305
5306 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5307 and could get into a loop if a null string was matched other than at the start
5308 of the subject.
5309
5310 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5311 be distinguished at compile time, and for completeness also added PCRE_DATE.
5312
5313 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5314 in GnuWin32 environments.
5315
5316
5317 Version 2.07 29-Jul-99
5318 ----------------------
5319
5320 1. The documentation is now supplied in plain text form and HTML as well as in
5321 the form of man page sources.
5322
5323 2. C++ compilers don't like assigning (void *) values to other pointer types.
5324 In particular this affects malloc(). Although there is no problem in Standard
5325 C, I've put in casts to keep C++ compilers happy.
5326
5327 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5328 should be (const char *).
5329
5330 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5331 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5332 However, I haven't made this a standard facility. The documentation doesn't
5333 mention it, and the Makefile doesn't support it.
5334
5335 5. The Makefile now contains an "install" target, with editable destinations at
5336 the top of the file. The pcretest program is not installed.
5337
5338 6. pgrep -V now gives the PCRE version number and date.
5339
5340 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5341 causing the entire string to be ignored, instead of just the last character.
5342
5343 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5344 non-matching string, it can take a very, very long time, even for strings of
5345 quite modest length, because of the nested recursion. PCRE now does better in
5346 some of these cases. It does this by remembering the last required literal
5347 character in the pattern, and pre-searching the subject to ensure it is present
5348 before running the real match. In other words, it applies a heuristic to detect
5349 some types of certain failure quickly, and in the above example, if presented
5350 with a string that has no trailing " it gives "no match" very quickly.
5351
5352 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5353 other alternatives are tried instead.
5354
5355
5356 Version 2.06 09-Jun-99
5357 ----------------------
5358
5359 1. Change pcretest's output for amount of store used to show just the code
5360 space, because the remainder (the data block) varies in size between 32-bit and
5361 64-bit systems.
5362
5363 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5364 start matching at. This allows lookbehinds to work when searching for multiple
5365 occurrences in a string.
5366
5367 3. Added additional options to pcretest for testing multiple occurrences:
5368
5369 /+ outputs the rest of the string that follows a match
5370 /g loops for multiple occurrences, using the new startoffset argument
5371 /G loops for multiple occurrences by passing an incremented pointer
5372
5373 4. PCRE wasn't doing the "first character" optimization for patterns starting
5374 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5375 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5376 the letter 'x'. On long subject strings, this gives a significant speed-up.
5377
5378
5379 Version 2.05 21-Apr-99
5380 ----------------------
5381
5382 1. Changed the type of magic_number from int to long int so that it works
5383 properly on 16-bit systems.
5384
5385 2. Fixed a bug which caused patterns starting with .* not to work correctly
5386 when the subject string contained newline characters. PCRE was assuming
5387 anchoring for such patterns in all cases, which is not correct because .* will
5388 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5389 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5390 must be retried after every newline in the subject.
5391
5392
5393 Version 2.04 18-Feb-99
5394 ----------------------
5395
5396 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5397 computation of the store needed to hold the pattern was incorrect (too large).
5398 If such patterns were nested a few deep, this could multiply and become a real
5399 problem.
5400
5401 2. Added /M option to pcretest to show the memory requirement of a specific
5402 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5403
5404 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5405 compiled in such a way that the backtracking after subsequent failure was
5406 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5407 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5408
5409
5410 Version 2.03 02-Feb-99
5411 ----------------------
5412
5413 1. Fixed typo and small mistake in man page.
5414
5415 2. Added 4th condition (GPL supersedes if conflict) and created separate
5416 LICENCE file containing the conditions.
5417
5418 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5419 Perl, that is the internal \ allows the delimiter to be included in the
5420 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5421 the final delimiter, add \ to the end of the pattern (to test the error).
5422
5423 4. Added the convenience functions for extracting substrings after a successful
5424 match. Updated pcretest to make it able to test these functions.
5425
5426
5427 Version 2.02 14-Jan-99
5428 ----------------------
5429
5430 1. Initialized the working variables associated with each extraction so that
5431 their saving and restoring doesn't refer to uninitialized store.
5432
5433 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5434 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5435 fix the problem.
5436
5437 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5438 calls, and wasn't printing the correct value for compiling calls. Increased the
5439 default value of LOOPREPEAT, and the number of significant figures in the
5440 times.
5441
5442 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5443
5444 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5445 a building problem on Windows NT with a FAT file system.
5446
5447
5448 Version 2.01 21-Oct-98
5449 ----------------------
5450
5451 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5452 to character tables built by pcre_maketables() in the current locale. If NULL
5453 is passed, the default tables are used.
5454
5455
5456 Version 2.00 24-Sep-98
5457 ----------------------
5458
5459 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5460 it any more.
5461
5462 2. Allow quantification of (?>) groups, and make it work correctly.
5463
5464 3. The first character computation wasn't working for (?>) groups.
5465
5466 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5467 end of the subject) and add 5.005's \z, which really does match only at the
5468 very end of the subject.
5469
5470 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5471
5472 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5473 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5474 localized options. All options to pcre_study() were also removed.
5475
5476 7. Add other new features from 5.005:
5477
5478 $(?<= positive lookbehind
5479 $(?<! negative lookbehind
5480 (?imsx-imsx) added the unsetting capability
5481 such a setting is global if at outer level; local otherwise
5482 (?imsx-imsx:) non-capturing groups with option setting
5483 (?(cond)re|re) conditional pattern matching
5484
5485 A backreference to itself in a repeated group matches the previous
5486 captured string.
5487
5488 8. General tidying up of studying (both automatic and via "study")
5489 consequential on the addition of new assertions.
5490
5491 9. As in 5.005, unlimited repeated groups that could match an empty substring
5492 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5493 runtime if any iteration does actually match an empty substring.
5494
5495 10. Include the RunTest script in the distribution.
5496
5497 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5498 discrepancies, some of which were old and were also with respect to 5.004. They
5499 have now been fixed.
5500
5501
5502 Version 1.09 28-Apr-98
5503 ----------------------
5504
5505 1. A negated single character class followed by a quantifier with a minimum
5506 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5507 program crashes, or just wrong answers. This did not apply to negated classes
5508 containing more than one character, or to minima other than one.
5509
5510
5511 Version 1.08 27-Mar-98
5512 ----------------------
5513
5514 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5515
5516 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5517 latter must appear before anything that relies on it in the pattern.
5518
5519
5520 Version 1.07 16-Feb-98
5521 ----------------------
5522
5523 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5524 repeat of a potentially empty string).
5525
5526
5527 Version 1.06 23-Jan-98
5528 ----------------------
5529
5530 1. Added Markus Oberhumer's little patches for C++.
5531
5532 2. Literal strings longer than 255 characters were broken.
5533
5534
5535 Version 1.05 23-Dec-97
5536 ----------------------
5537
5538 1. Negated character classes containing more than one character were failing if
5539 PCRE_CASELESS was set at run time.
5540
5541
5542 Version 1.04 19-Dec-97
5543 ----------------------
5544
5545 1. Corrected the man page, where some "const" qualifiers had been omitted.
5546
5547 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5548 input syntax.
5549
5550 3. Fixed memory leak which occurred when a regex with back references was
5551 matched with an offsets vector that wasn't big enough. The temporary memory
5552 that is used in this case wasn't being freed if the match failed.
5553
5554 4. Tidied pcretest to ensure it frees memory that it gets.
5555
5556 5. Temporary memory was being obtained in the case where the passed offsets
5557 vector was exactly big enough.
5558
5559 6. Corrected definition of offsetof() from change 5 below.
5560
5561 7. I had screwed up change 6 below and broken the rules for the use of
5562 setjmp(). Now fixed.
5563
5564
5565 Version 1.03 18-Dec-97
5566 ----------------------
5567
5568 1. A erroneous regex with a missing opening parenthesis was correctly
5569 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5570 on some systems.
5571
5572 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5573 it was reported that one broken compiler failed on the former because "code" is
5574 also an independent variable.
5575
5576 3. The erroneous regex a[]b caused an array overrun reference.
5577
5578 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5579 fail on data ending with that character. (It was going on too far, and checking
5580 the next character, typically a binary zero.) This was specific to the
5581 optimized code for single-character negative classes.
5582
5583 5. Added a contributed patch from the TIN world which does the following:
5584
5585 + Add an undef for memmove, in case the the system defines a macro for it.
5586
5587 + Add a definition of offsetof(), in case there isn't one. (I don't know
5588 the reason behind this - offsetof() is part of the ANSI standard - but
5589 it does no harm).
5590
5591 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5592 most of the places where whitespace preceded '#'. I have given up and
5593 allowed the remaining 2 cases to be at the margin.
5594
5595 + Rename some variables in pcre to eliminate shadowing. This seems very
5596 pedantic, but does no harm, of course.
5597
5598 6. Moved the call to setjmp() into its own function, to get rid of warnings
5599 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5600
5601 7. Constructs such as \d{8,} were compiling into the equivalent of
5602 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5603 outcome, but in this particular case used more store than had been allocated,
5604 which caused the bug to be discovered because it threw up an internal error.
5605
5606 8. The debugging code in both pcre and pcretest for outputting the compiled
5607 form of a regex was going wrong in the case of back references followed by
5608 curly-bracketed repeats.
5609
5610
5611 Version 1.02 12-Dec-97
5612 ----------------------
5613
5614 1. Typos in pcre.3 and comments in the source fixed.
5615
5616 2. Applied a contributed patch to get rid of places where it used to remove
5617 'const' from variables, and fixed some signed/unsigned and uninitialized
5618 variable warnings.
5619
5620 3. Added the "runtest" target to Makefile.
5621
5622 4. Set default compiler flag to -O2 rather than just -O.
5623
5624
5625 Version 1.01 19-Nov-97
5626 ----------------------
5627
5628 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5629 like /([ab]*)*/, that is, for classes with more than one character in them.
5630
5631 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5632 as /((?>a*))*/ (a PCRE_EXTRA facility).
5633
5634
5635 Version 1.00 18-Nov-97
5636 ----------------------
5637
5638 1. Added compile-time macros to support systems such as SunOS4 which don't have
5639 memmove() or strerror() but have other things that can be used instead.
5640
5641 2. Arranged that "make clean" removes the executables.
5642
5643
5644 Version 0.99 27-Oct-97
5645 ----------------------
5646
5647 1. Fixed bug in code for optimizing classes with only one character. It was
5648 initializing a 32-byte map regardless, which could cause it to run off the end
5649 of the memory it had got.
5650
5651 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5652
5653
5654 Version 0.98 22-Oct-97
5655 ----------------------
5656
5657 1. Fixed bug in code for handling temporary memory usage when there are more
5658 back references than supplied space in the ovector. This could cause segfaults.
5659
5660
5661 Version 0.97 21-Oct-97
5662 ----------------------
5663
5664 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5665
5666 2. Optimized negated single characters not to use a bit map.
5667
5668 3. Brought error texts together as macro definitions; clarified some of them;
5669 fixed one that was wrong - it said "range out of order" when it meant "invalid
5670 escape sequence".
5671
5672 4. Changed some char * arguments to const char *.
5673
5674 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5675
5676 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5677 pcretest.
5678
5679
5680 Version 0.96 16-Oct-97
5681 ----------------------
5682
5683 1. Added a simple "pgrep" utility to the distribution.
5684
5685 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5686 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5687 where "ddd" means "one or more decimal digits".
5688
5689 3. Fixed serious bug. If a pattern had a back reference, but the call to
5690 pcre_exec() didn't supply a large enough ovector to record the related
5691 identifying subpattern, the match always failed. PCRE now remembers the number
5692 of the largest back reference, and gets some temporary memory in which to save
5693 the offsets during matching if necessary, in order to ensure that
5694 backreferences always work.
5695
5696 4. Increased the compatibility with Perl in a number of ways:
5697
5698 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5699 to request this handling. The option can be set at compile or exec time.
5700
5701 (b) $ matches before a terminating newline by default; an option
5702 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5703 mode). The option can be set at compile or exec time.
5704
5705 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5706 the same as Perl's. If the decimal number it represents is less than 10
5707 or there aren't that many previous left capturing parentheses, an octal
5708 escape is read. Inside a character class, it's always an octal escape,
5709 even if it is a single digit.
5710
5711 (d) An escaped but undefined alphabetic character is taken as a literal,
5712 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5713 escapes.
5714
5715 (e) {0} is now permitted. (The previous item is removed from the compiled
5716 pattern).
5717
5718 5. Changed all the names of code files so that the basic parts are no longer
5719 than 10 characters, and abolished the teeny "globals.c" file.
5720
5721 6. Changed the handling of character classes; they are now done with a 32-byte
5722 bit map always.
5723
5724 7. Added the -d and /D options to pcretest to make it possible to look at the
5725 internals of compilation without having to recompile pcre.
5726
5727
5728 Version 0.95 23-Sep-97
5729 ----------------------
5730
5731 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5732 \x20 at the start of a run of normal characters. These were being treated as
5733 real characters, instead of the source characters being re-checked.
5734
5735
5736 Version 0.94 18-Sep-97
5737 ----------------------
5738
5739 1. The functions are now thread-safe, with the caveat that the global variables
5740 containing pointers to malloc() and free() or alternative functions are the
5741 same for all threads.
5742
5743 2. Get pcre_study() to generate a bitmap of initial characters for non-
5744 anchored patterns when this is possible, and use it if passed to pcre_exec().
5745
5746
5747 Version 0.93 15-Sep-97
5748 ----------------------
5749
5750 1. /(b)|(:+)/ was computing an incorrect first character.
5751
5752 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5753 but not actually doing anything yet.
5754
5755 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5756 as Perl does (e.g. [-az] or [az-]).
5757
5758 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5759 all possible positions.
5760
5761 5. Split up into different modules to avoid including unneeded functions in a
5762 compiled binary. However, compile and exec are still in one module. The "study"
5763 function is split off.
5764
5765 6. The character tables are now in a separate module whose source is generated
5766 by an auxiliary program - but can then be edited by hand if required. There are
5767 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5768 toupper() in the code.
5769
5770 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5771 make them global. Abolish the function for setting them, as the caller can now
5772 set them directly.
5773
5774
5775 Version 0.92 11-Sep-97
5776 ----------------------
5777
5778 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5779 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5780
5781 2. Caseless matching was not working in character classes if the characters in
5782 the pattern were in upper case.
5783
5784 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5785
5786 4. Make PCRE_ANCHORED public and accept as a compile option.
5787
5788 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5789 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5790 pass them.
5791
5792 6. Give an error if bad option bits passed at compile or run time.
5793
5794 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5795 pcretest to cause it to pass that flag.
5796
5797 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5798 options, and the first character, if set.
5799
5800 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5801
5802
5803 Version 0.91 10-Sep-97
5804 ----------------------
5805
5806 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5807 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5808
5809 2. PCRE was looping on encountering an indefinitely repeated back reference to
5810 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5811 Perl does - treats the match as successful.
5812
5813 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5