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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1585 - (show annotations)
Wed Aug 5 15:38:32 2015 UTC (4 years, 3 months ago) by ph10
File size: 282284 byte(s)
Fix buffer overflow for named references in (?| situations.
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 18. A pattern with an unmatched closing parenthesis that contained a backward
75 assertion which itself contained a forward reference caused buffer
76 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
77
78 19. JIT should return with error when the compiled pattern requires more stack
79 space than the maximum.
80
81 20. A possessively repeated conditional group that could match an empty string,
82 for example, /(?(R))*+/, was incorrectly compiled.
83
84 21. Fix infinite recursion in the JIT compiler when certain patterns such as
85 /(?:|a|){100}x/ are analysed.
86
87 22. Some patterns with character classes involving [: and \\ were incorrectly
88 compiled and could cause reading from uninitialized memory or an incorrect
89 error diagnosis.
90
91 23. Pathological patterns containing many nested occurrences of [: caused
92 pcre_compile() to run for a very long time.
93
94 24. A conditional group with only one branch has an implicit empty alternative
95 branch and must therefore be treated as potentially matching an empty
96 string.
97
98 25. If (?R was followed by - or + incorrect behaviour happened instead of a
99 diagnostic.
100
101 26. Arrange to give up on finding the minimum matching length for overly
102 complex patterns.
103
104 27. Similar to (4) above: in a pattern with duplicated named groups and an
105 occurrence of (?| it is possible for an apparently non-recursive back
106 reference to become recursive if a later named group with the relevant
107 number is encountered. This could lead to a buffer overflow. Wen Guanxing
108 from Venustech ADLAB discovered this bug.
109
110
111 Version 8.37 28-April-2015
112 --------------------------
113
114 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
115 for those parentheses to be closed with whatever has been captured so far.
116 However, it was failing to mark any other groups between the hightest
117 capture so far and the currrent group as "unset". Thus, the ovector for
118 those groups contained whatever was previously there. An example is the
119 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
120
121 2. If an assertion condition was quantified with a minimum of zero (an odd
122 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
123
124 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
125 unrecognized modifier, a crash could occur.
126
127 4. An attempt to do global matching in pcretest with a zero-length ovector
128 caused a crash.
129
130 5. Fixed a memory leak during matching that could occur for a subpattern
131 subroutine call (recursive or otherwise) if the number of captured groups
132 that had to be saved was greater than ten.
133
134 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
135 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
136 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
137
138 7. A UTF pattern containing a "not" match of a non-ASCII character and a
139 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
140
141 8. When a pattern is compiled, it remembers the highest back reference so that
142 when matching, if the ovector is too small, extra memory can be obtained to
143 use instead. A conditional subpattern whose condition is a check on a
144 capture having happened, such as, for example in the pattern
145 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
146 setting the highest backreference number. This mattered only if pcre_exec()
147 was called with an ovector that was too small to hold the capture, and there
148 was no other kind of back reference (a situation which is probably quite
149 rare). The effect of the bug was that the condition was always treated as
150 FALSE when the capture could not be consulted, leading to a incorrect
151 behaviour by pcre_exec(). This bug has been fixed.
152
153 9. A reference to a duplicated named group (either a back reference or a test
154 for being set in a conditional) that occurred in a part of the pattern where
155 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
156 to be incorrectly calculated, leading to overwriting.
157
158 10. A mutually recursive set of back references such as (\2)(\1) caused a
159 segfault at study time (while trying to find the minimum matching length).
160 The infinite loop is now broken (with the minimum length unset, that is,
161 zero).
162
163 11. If an assertion that was used as a condition was quantified with a minimum
164 of zero, matching went wrong. In particular, if the whole group had
165 unlimited repetition and could match an empty string, a segfault was
166 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
167 assertions to be quantified, but not if they are being used as conditions,
168 so the above pattern is faulted by Perl. PCRE has now been changed so that
169 it also rejects such patterns.
170
171 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
172 failed to allow the zero-repeat case if pcre2_exec() was called with an
173 ovector too small to capture the group.
174
175 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
176 Red Hat Product Security:
177
178 (a) A crash if /K and /F were both set with the option to save the compiled
179 pattern.
180
181 (b) Another crash if the option to print captured substrings in a callout
182 was combined with setting a null ovector, for example \O\C+ as a subject
183 string.
184
185 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
186 forward reference repeated a large (but limited) number of times within a
187 repeated outer group that has a zero minimum quantifier, caused incorrect
188 code to be compiled, leading to the error "internal error:
189 previously-checked referenced subpattern not found" when an incorrect
190 memory address was read. This bug was reported as "heap overflow",
191 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
192 CVE-2015-2325.
193
194 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
195 call within a group that also contained a recursive back reference caused
196 incorrect code to be compiled. This bug was reported as "heap overflow",
197 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
198 number CVE-2015-2326.
199
200 24. Computing the size of the JIT read-only data in advance has been a source
201 of various issues, and new ones are still appear unfortunately. To fix
202 existing and future issues, size computation is eliminated from the code,
203 and replaced by on-demand memory allocation.
204
205 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
206 adjacent to the end of the range, and the range contained characters with
207 more than one other case, caused incorrect behaviour when compiled in UTF
208 mode. In that example, the range a-j was left out of the class.
209
210 26. Fix JIT compilation of conditional blocks, which assertion
211 is converted to (*FAIL). E.g: /(?(?!))/.
212
213 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
214 discovered by the LLVM fuzzer.
215
216 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
217 when this assertion was used as a condition, for example (?(?!)a|b). In
218 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
219 error about an unsupported item.
220
221 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
222 possessification code could take exponential time to complete. A recursion
223 depth limit of 1000 has been imposed to limit the resources used by this
224 optimization.
225
226 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
227 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
228 because \S ensures they are all in the class. The code for doing this was
229 interacting badly with the code for computing the amount of space needed to
230 compile the pattern, leading to a buffer overflow. This bug was discovered
231 by the LLVM fuzzer.
232
233 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
234 other kinds of group caused stack overflow at compile time. This bug was
235 discovered by the LLVM fuzzer.
236
237 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
238 between a subroutine call and its quantifier was incorrectly compiled,
239 leading to buffer overflow or other errors. This bug was discovered by the
240 LLVM fuzzer.
241
242 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
243 assertion after (?(. The code was failing to check the character after
244 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
245 was discovered by the LLVM fuzzer.
246
247 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
248 a fixed maximum following a group that contains a subroutine reference was
249 incorrectly compiled and could trigger buffer overflow. This bug was
250 discovered by the LLVM fuzzer.
251
252 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
253 caused a stack overflow instead of the diagnosis of a non-fixed length
254 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
255
256 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
257 (e.g. /(?<=\Ka)/) could make pcregrep loop.
258
259 37. There was a similar problem to 36 in pcretest for global matches.
260
261 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
262 and a subsequent item in the pattern caused a non-match, backtracking over
263 the repeated \X did not stop, but carried on past the start of the subject,
264 causing reference to random memory and/or a segfault. There were also some
265 other cases where backtracking after \C could crash. This set of bugs was
266 discovered by the LLVM fuzzer.
267
268 39. The function for finding the minimum length of a matching string could take
269 a very long time if mutual recursion was present many times in a pattern,
270 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
271 method has been implemented. This infelicity was discovered by the LLVM
272 fuzzer.
273
274 40. Static linking against the PCRE library using the pkg-config module was
275 failing on missing pthread symbols.
276
277
278 Version 8.36 26-September-2014
279 ------------------------------
280
281 1. Got rid of some compiler warnings in the C++ modules that were shown up by
282 -Wmissing-field-initializers and -Wunused-parameter.
283
284 2. The tests for quantifiers being too big (greater than 65535) were being
285 applied after reading the number, and stupidly assuming that integer
286 overflow would give a negative number. The tests are now applied as the
287 numbers are read.
288
289 3. Tidy code in pcre_exec.c where two branches that used to be different are
290 now the same.
291
292 4. The JIT compiler did not generate match limit checks for certain
293 bracketed expressions with quantifiers. This may lead to exponential
294 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
295 issue should be resolved now.
296
297 5. Fixed an issue, which occures when nested alternatives are optimized
298 with table jumps.
299
300 6. Inserted two casts and changed some ints to size_t in the light of some
301 reported 64-bit compiler warnings (Bugzilla 1477).
302
303 7. Fixed a bug concerned with zero-minimum possessive groups that could match
304 an empty string, which sometimes were behaving incorrectly in the
305 interpreter (though correctly in the JIT matcher). This pcretest input is
306 an example:
307
308 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
309 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
310
311 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
312 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
313 for an empty string was breaking the inner loop and carrying on at a lower
314 level, when possessive repeated groups should always return to a higher
315 level as they have no backtrack points in them. The empty string test now
316 occurs at the outer level.
317
318 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
319 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
320
321 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
322 doesn't).
323
324 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
325 to hold a pattern to be incorrectly computed (too small) when there were
326 named back references to duplicated names. This could cause "internal
327 error: code overflow" or "double free or corruption" or other memory
328 handling errors.
329
330 11. When named subpatterns had the same prefixes, back references could be
331 confused. For example, in this pattern:
332
333 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
334
335 the reference to 'Name' was incorrectly treated as a reference to a
336 duplicate name.
337
338 12. A pattern such as /^s?c/mi8 where the optional character has more than
339 one "other case" was incorrectly compiled such that it would only try to
340 match starting at "c".
341
342 13. When a pattern starting with \s was studied, VT was not included in the
343 list of possible starting characters; this should have been part of the
344 8.34/18 patch.
345
346 14. If a character class started [\Qx]... where x is any character, the class
347 was incorrectly terminated at the ].
348
349 15. If a pattern that started with a caseless match for a character with more
350 than one "other case" was studied, PCRE did not set up the starting code
351 unit bit map for the list of possible characters. Now it does. This is an
352 optimization improvement, not a bug fix.
353
354 16. The Unicode data tables have been updated to Unicode 7.0.0.
355
356 17. Fixed a number of memory leaks in pcregrep.
357
358 18. Avoid a compiler warning (from some compilers) for a function call with
359 a cast that removes "const" from an lvalue by using an intermediate
360 variable (to which the compiler does not object).
361
362 19. Incorrect code was compiled if a group that contained an internal recursive
363 back reference was optional (had quantifier with a minimum of zero). This
364 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
365 caused segmentation faults because of stack overflows at compile time.
366
367 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
368 group that is quantified with an indefinite repeat, caused a compile-time
369 loop which used up all the system stack and provoked a segmentation fault.
370 This was not the same bug as 19 above.
371
372 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
373 Patch by Mike Frysinger.
374
375
376 Version 8.35 04-April-2014
377 --------------------------
378
379 1. A new flag is set, when property checks are present in an XCLASS.
380 When this flag is not set, PCRE can perform certain optimizations
381 such as studying these XCLASS-es.
382
383 2. The auto-possessification of character sets were improved: a normal
384 and an extended character set can be compared now. Furthermore
385 the JIT compiler optimizes more character set checks.
386
387 3. Got rid of some compiler warnings for potentially uninitialized variables
388 that show up only when compiled with -O2.
389
390 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
391 of a match later then the end of the match. The pcretest program was not
392 handling the case sensibly - it was outputting from the start to the next
393 binary zero. It now reports this situation in a message, and outputs the
394 text from the end to the start.
395
396 5. Fast forward search is improved in JIT. Instead of the first three
397 characters, any three characters with fixed position can be searched.
398 Search order: first, last, middle.
399
400 6. Improve character range checks in JIT. Characters are read by an inprecise
401 function now, which returns with an unknown value if the character code is
402 above a certain threshold (e.g: 256). The only limitation is that the value
403 must be bigger than the threshold as well. This function is useful when
404 the characters above the threshold are handled in the same way.
405
406 7. The macros whose names start with RAWUCHAR are placeholders for a future
407 mode in which only the bottom 21 bits of 32-bit data items are used. To
408 make this more memorable for those maintaining the code, the names have
409 been changed to start with UCHAR21, and an extensive comment has been added
410 to their definition.
411
412 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
413 to the export list in Makefile.am (they were accidentally omitted from the
414 8.34 tarball).
415
416 9. The informational output from pcretest used the phrase "starting byte set"
417 which is inappropriate for the 16-bit and 32-bit libraries. As the output
418 for "first char" and "need char" really means "non-UTF-char", I've changed
419 "byte" to "char", and slightly reworded the output. The documentation about
420 these values has also been (I hope) clarified.
421
422 10. Another JIT related optimization: use table jumps for selecting the correct
423 backtracking path, when more than four alternatives are present inside a
424 bracket.
425
426 11. Empty match is not possible, when the minimum length is greater than zero,
427 and there is no \K in the pattern. JIT should avoid empty match checks in
428 such cases.
429
430 12. In a caseless character class with UCP support, when a character with more
431 than one alternative case was not the first character of a range, not all
432 the alternative cases were added to the class. For example, s and \x{17f}
433 are both alternative cases for S: the class [RST] was handled correctly,
434 but [R-T] was not.
435
436 13. The configure.ac file always checked for pthread support when JIT was
437 enabled. This is not used in Windows, so I have put this test inside a
438 check for the presence of windows.h (which was already tested for).
439
440 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
441 The algorithm provides a way to skip certain starting offsets, and usually
442 faster than linear prefix searches.
443
444 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
445 as for 'fr_FR' and 'french'. For some reason, however, it then used the
446 Windows-specific input and output files, which have 'french' screwed in.
447 So this could never have worked. One of the problems with locales is that
448 they aren't always the same. I have now updated RunTest so that it checks
449 the output of the locale test (test 3) against three different output
450 files, and it allows the test to pass if any one of them matches. With luck
451 this should make the test pass on some versions of Solaris where it was
452 failing. Because of the uncertainty, the script did not used to stop if
453 test 3 failed; it now does. If further versions of a French locale ever
454 come to light, they can now easily be added.
455
456 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
457 there was a message during ./configure, but it did not stop. This now
458 provokes an error. The invalid example in README has been corrected.
459 If a value less than the minimum is given, the minimum value has always
460 been used, but now a warning is given.
461
462 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
463 was a bug in the test system, which is now fixed. Also, the list of various
464 configurations that are tested for each release did not have one with both
465 16/32 bits and --enable-bar-anycrlf. It now does.
466
467 18. pcretest was missing "-C bsr" for displaying the \R default setting.
468
469 19. Little endian PowerPC systems are supported now by the JIT compiler.
470
471 20. The fast forward newline mechanism could enter to an infinite loop on
472 certain invalid UTF-8 input. Although we don't support these cases
473 this issue can be fixed by a performance optimization.
474
475 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
476 not take account if existing stack usage. There is now a new global
477 variable called pcre_stack_guard that can be set to point to an external
478 function to check stack availability. It is called at the start of
479 processing every parenthesized group.
480
481 22. A typo in the code meant that in ungreedy mode the max/min qualifier
482 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
483 not match "ab".
484
485 23. When UTF was disabled, the JIT program reported some incorrect compile
486 errors. These messages are silenced now.
487
488 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
489 compiler.
490
491 25. Change all the temporary files used in RunGrepTest to be different to those
492 used by RunTest so that the tests can be run simultaneously, for example by
493 "make -j check".
494
495
496 Version 8.34 15-December-2013
497 -----------------------------
498
499 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
500 executable memory. Patch inspired by Carsten Klein.
501
502 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
503 this macro is never tested and has no effect, because the work to support
504 coverage involves only compiling and linking options and special targets in
505 the Makefile. The comment in config.h implied that defining the macro would
506 enable coverage support, which is totally false. There was also support for
507 setting this macro in the CMake files (my fault, I just copied it from
508 configure). SUPPORT_GCOV has now been removed.
509
510 3. Make a small performance improvement in strlen16() and strlen32() in
511 pcretest.
512
513 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
514 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
515 these cases). There was also one in pcretest.c.
516
517 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
518
519 6. In UTF mode, the code for checking whether a group could match an empty
520 string (which is used for indefinitely repeated groups to allow for
521 breaking an infinite loop) was broken when the group contained a repeated
522 negated single-character class with a character that occupied more than one
523 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
524 UTF-8 mode). The effect was undefined: the group might or might not be
525 deemed as matching an empty string, or the program might have crashed.
526
527 7. The code for checking whether a group could match an empty string was not
528 recognizing that \h, \H, \v, \V, and \R must match a character.
529
530 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
531 an empty string. If it can, pcretest shows this in its information output.
532
533 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
534 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
535 matched by pcre_exec() without using JIT:
536
537 (a) If the rest of the pattern did not match after a maximal run of
538 grapheme clusters, the code for backing up to try with fewer of them
539 did not always back up over a full grapheme when characters that do not
540 have the modifier quality were involved, e.g. Hangul syllables.
541
542 (b) If the match point in a subject started with modifier character, and
543 there was no match, the code could incorrectly back up beyond the match
544 point, and potentially beyond the first character in the subject,
545 leading to a segfault or an incorrect match result.
546
547 10. A conditional group with an assertion condition could lead to PCRE
548 recording an incorrect first data item for a match if no other first data
549 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
550 first data item, and therefore matched "ca" after "c" instead of at the
551 start.
552
553 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
554 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
555
556 12. The source of pcregrep now includes z/OS-specific code so that it can be
557 compiled for z/OS as part of the special z/OS distribution.
558
559 13. Added the -T and -TM options to pcretest.
560
561 14. The code in pcre_compile.c for creating the table of named capturing groups
562 has been refactored. Instead of creating the table dynamically during the
563 actual compiling pass, the information is remembered during the pre-compile
564 pass (on the stack unless there are more than 20 named groups, in which
565 case malloc() is used) and the whole table is created before the actual
566 compile happens. This has simplified the code (it is now nearly 150 lines
567 shorter) and prepared the way for better handling of references to groups
568 with duplicate names.
569
570 15. A back reference to a named subpattern when there is more than one of the
571 same name now checks them in the order in which they appear in the pattern.
572 The first one that is set is used for the reference. Previously only the
573 first one was inspected. This change makes PCRE more compatible with Perl.
574
575 16. Unicode character properties were updated from Unicode 6.3.0.
576
577 17. The compile-time code for auto-possessification has been refactored, based
578 on a patch by Zoltan Herczeg. It now happens after instead of during
579 compilation. The code is cleaner, and more cases are handled. The option
580 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
581 options in pcretest are provided to set it. It can also be set by
582 (*NO_AUTO_POSSESS) at the start of a pattern.
583
584 18. The character VT has been added to the default ("C" locale) set of
585 characters that match \s and are generally treated as white space,
586 following this same change in Perl 5.18. There is now no difference between
587 "Perl space" and "POSIX space". Whether VT is treated as white space in
588 other locales depends on the locale.
589
590 19. The code for checking named groups as conditions, either for being set or
591 for being recursed, has been refactored (this is related to 14 and 15
592 above). Processing unduplicated named groups should now be as fast at
593 numerical groups, and processing duplicated groups should be faster than
594 before.
595
596 20. Two patches to the CMake build system, by Alexander Barkov:
597
598 (1) Replace the "source" command by "." in CMakeLists.txt because
599 "source" is a bash-ism.
600
601 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
602 without these the CMake build does not work on Solaris.
603
604 21. Perl has changed its handling of \8 and \9. If there is no previously
605 encountered capturing group of those numbers, they are treated as the
606 literal characters 8 and 9 instead of a binary zero followed by the
607 literals. PCRE now does the same.
608
609 22. Following Perl, added \o{} to specify codepoints in octal, making it
610 possible to specify values greater than 0777 and also making them
611 unambiguous.
612
613 23. Perl now gives an error for missing closing braces after \x{... instead of
614 treating the string as literal. PCRE now does the same.
615
616 24. RunTest used to grumble if an inappropriate test was selected explicitly,
617 but just skip it when running all tests. This make it awkward to run ranges
618 of tests when one of them was inappropriate. Now it just skips any
619 inappropriate tests, as it always did when running all tests.
620
621 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
622 character types such as \d or \w, too many callouts were inserted, and the
623 data that they returned was rubbish.
624
625 26. In UCP mode, \s was not matching two of the characters that Perl matches,
626 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
627 were matched by \h. The code has now been refactored so that the lists of
628 the horizontal and vertical whitespace characters used for \h and \v (which
629 are defined only in one place) are now also used for \s.
630
631 27. Add JIT support for the 64 bit TileGX architecture.
632 Patch by Jiong Wang (Tilera Corporation).
633
634 28. Possessive quantifiers for classes (both explicit and automatically
635 generated) now use special opcodes instead of wrapping in ONCE brackets.
636
637 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
638 (because it's meaningless), this was not happening when PCRE_CASELESS was
639 set. Not wrong, but inefficient.
640
641 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
642 properties for \w, \d, etc) is present in a test regex. Otherwise if the
643 test contains no characters greater than 255, Perl doesn't realise it
644 should be using Unicode semantics.
645
646 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
647 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
648 does in Unicode mode.
649
650 32. Added the "forbid" facility to pcretest so that putting tests into the
651 wrong test files can sometimes be quickly detected.
652
653 33. There is now a limit (default 250) on the depth of nesting of parentheses.
654 This limit is imposed to control the amount of system stack used at compile
655 time. It can be changed at build time by --with-parens-nest-limit=xxx or
656 the equivalent in CMake.
657
658 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
659 errors. Perl warns for these when in warning mode, but PCRE has no facility
660 for giving warnings.
661
662 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
663 However, this was not working for (?!) because it is optimized to (*FAIL),
664 for which PCRE does not allow quantifiers. The optimization is now disabled
665 when a quantifier follows (?!). I can't see any use for this, but it makes
666 things uniform.
667
668 36. Perl no longer allows group names to start with digits, so I have made this
669 change also in PCRE. It simplifies the code a bit.
670
671 37. In extended mode, Perl ignores spaces before a + that indicates a
672 possessive quantifier. PCRE allowed a space before the quantifier, but not
673 before the possessive +. It now does.
674
675 38. The use of \K (reset reported match start) within a repeated possessive
676 group such as (a\Kb)*+ was not working.
677
678 40. Document that the same character tables must be used at compile time and
679 run time, and that the facility to pass tables to pcre_exec() and
680 pcre_dfa_exec() is for use only with saved/restored patterns.
681
682 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
683 features for Builds with MSVC:
684
685 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
686 builds). The .rc files can be used to set FileDescription and many other
687 attributes.
688
689 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
690 This allows higher-level build scripts which want .pdb files to avoid
691 hard-coding the exact files needed."
692
693 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
694 mean "start of word" and "end of word", respectively, as a transition aid.
695
696 43. A minimizing repeat of a class containing codepoints greater than 255 in
697 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
698 compiled to use the heap for recursion.
699
700 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
701 is configured.
702
703
704 Version 8.33 28-May-2013
705 ------------------------
706
707 1. Added 'U' to some constants that are compared to unsigned integers, to
708 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
709 variables that are added to signed variables, to ensure the result is
710 signed and can be negated.
711
712 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
713 CMake config files.
714
715 3. Revise the creation of config.h.generic so that all boolean macros are
716 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
717 overriding via -D on the command line possible.
718
719 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
720 to unsigned int is reported to make a quite noticeable speed difference in
721 a specific Windows environment. Testing on Linux did also appear to show
722 some benefit (and it is clearly not harmful). Also fixed the definition of
723 Xop which should be unsigned.
724
725 5. Related to (4), changing the definition of the intermediate variable cc
726 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
727 improvements.
728
729 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
730 unnecessary spaces.
731
732 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
733 and later.
734
735 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
736
737 9. Optimizing fast_forward_start_bits in JIT.
738
739 10. Adding support for callouts in JIT, and fixing some issues revealed
740 during this work. Namely:
741
742 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
743
744 (b) Minimum length was not checked before the matching is started.
745
746 11. The value of capture_last that is passed to callouts was incorrect in some
747 cases when there was a capture on one path that was subsequently abandoned
748 after a backtrack. Also, the capture_last value is now reset after a
749 recursion, since all captures are also reset in this case.
750
751 12. The interpreter no longer returns the "too many substrings" error in the
752 case when an overflowing capture is in a branch that is subsequently
753 abandoned after a backtrack.
754
755 13. In the pathological case when an offset vector of size 2 is used, pcretest
756 now prints out the matched string after a yield of 0 or 1.
757
758 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
759 Only supported by the JIT compiler at the moment.
760
761 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
762
763 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
764 offset of the starting point of the matching process, provided the offsets
765 vector is large enough.
766
767 17. The \A escape now records a lookbehind value of 1, though its execution
768 does not actually inspect the previous character. This is to ensure that,
769 in partial multi-segment matching, at least one character from the old
770 segment is retained when a new segment is processed. Otherwise, if there
771 are no lookbehinds in the pattern, \A might match incorrectly at the start
772 of a new segment.
773
774 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
775
776 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
777 gives some modest performance improvement in 8-bit mode.
778
779 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
780 be expressed in certain programming languages using Universal Character
781 Names.
782
783 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
784 which points out that "non characters" are not "characters that may not
785 appear in Unicode strings" but rather "characters that are reserved for
786 internal use and have only local meaning".
787
788 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
789 there was a conditional group that depended on an assertion, if the
790 assertion was false, the callout that immediately followed the alternation
791 in the condition was skipped when pcre_exec() was used for matching.
792
793 23. Allow an explicit callout to be inserted before an assertion that is the
794 condition for a conditional group, for compatibility with automatic
795 callouts, which always insert a callout at this point.
796
797 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
798 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
799
800 25. (*PRUNE) is now supported by the JIT compiler.
801
802 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
803
804 27. Fix the case where there are two or more SKIPs with arguments that may be
805 ignored.
806
807 28. (*SKIP) is now supported by the JIT compiler.
808
809 29. (*THEN) is now supported by the JIT compiler.
810
811 30. Update RunTest with additional test selector options.
812
813 31. The way PCRE handles backtracking verbs has been changed in two ways.
814
815 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
816 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
817 backtracking. In some cases this makes it more Perl-compatible, but Perl's
818 rather obscure rules do not always do the same thing.
819
820 (2) Previously, backtracking verbs were confined within assertions. This is
821 no longer the case for positive assertions, except for (*ACCEPT). Again,
822 this sometimes improves Perl compatibility, and sometimes does not.
823
824 32. A number of tests that were in test 2 because Perl did things differently
825 have been moved to test 1, because either Perl or PCRE has changed, and
826 these tests are now compatible.
827
828 32. Backtracking control verbs are now handled in the same way in JIT and
829 interpreter.
830
831 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
832 contained a forward subroutine reference caused a compile error.
833
834 34. Auto-detect and optimize limited repetitions in JIT.
835
836 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
837 blocking (*UTF) etc.
838
839 36. In the interpreter, maximizing pattern repetitions for characters and
840 character types now use tail recursion, which reduces stack usage.
841
842 37. The value of the max lookbehind was not correctly preserved if a compiled
843 and saved regex was reloaded on a host of different endianness.
844
845 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
846 of the compiled pattern block, expand the flags field from 16 to 32 bits
847 because it was almost full.
848
849 39. Try madvise first before posix_madvise.
850
851 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
852 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
853 original change locked out all matches of empty strings. This has been
854 changed so that one match of an empty string per line is recognized.
855 Subsequent searches on the same line (for colouring or for --only-matching,
856 for example) do not recognize empty strings.
857
858 41. Applied a user patch to fix a number of spelling mistakes in comments.
859
860 42. Data lines longer than 65536 caused pcretest to crash.
861
862 43. Clarified the data type for length and startoffset arguments for pcre_exec
863 and pcre_dfa_exec in the function-specific man pages, where they were
864 explicitly stated to be in bytes, never having been updated. I also added
865 some clarification to the pcreapi man page.
866
867 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
868 a segmentation fault.
869
870
871 Version 8.32 30-November-2012
872 -----------------------------
873
874 1. Improved JIT compiler optimizations for first character search and single
875 character iterators.
876
877 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
878 Patch by Daniel Richard G.
879
880 3. Single character iterator optimizations in the JIT compiler.
881
882 4. Improved JIT compiler optimizations for character ranges.
883
884 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
885 Reported by Giuseppe D'Angelo.
886
887 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
888 of a line, was being set incorrectly in cases where .* appeared inside
889 atomic brackets at the start of a pattern, or where there was a subsequent
890 *PRUNE or *SKIP.
891
892 7. Improved instruction cache flush for POWER/PowerPC.
893 Patch by Daniel Richard G.
894
895 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
896 grep:
897
898 (a) There is now no limit to the number of patterns to be matched.
899
900 (b) An error is given if a pattern is too long.
901
902 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
903 are now supported.
904
905 (d) --exclude-from and --include-from (multiple use) have been added.
906
907 (e) Exclusions and inclusions now apply to all files and directories, not
908 just to those obtained from scanning a directory recursively.
909
910 (f) Multiple uses of -f and --file-list are now supported.
911
912 (g) In a Windows environment, the default for -d has been changed from
913 "read" (the GNU grep default) to "skip", because otherwise the presence
914 of a directory in the file list provokes an error.
915
916 (h) The documentation has been revised and clarified in places.
917
918 9. Improve the matching speed of capturing brackets.
919
920 10. Changed the meaning of \X so that it now matches a Unicode extended
921 grapheme cluster.
922
923 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
924 out POSIX threads when JIT support is configured.
925
926 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
927
928 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
929 structure could go wrong in environments where size_t is not the same size
930 as int.
931
932 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
933 set.
934
935 15. The EBCDIC support had decayed; later updates to the code had included
936 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
937 general tidy up of EBCDIC-related issues, and the documentation was also
938 not quite right. There is now a test that can be run on ASCII systems to
939 check some of the EBCDIC-related things (but is it not a full test).
940
941 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
942 in a small tidy to the code.
943
944 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
945
946 18. If the --only-matching (-o) option in pcregrep is specified multiple
947 times, each one causes appropriate output. For example, -o1 -o2 outputs the
948 substrings matched by the 1st and 2nd capturing parentheses. A separating
949 string can be specified by --om-separator (default empty).
950
951 19. Improving the first n character searches.
952
953 20. Turn case lists for horizontal and vertical white space into macros so that
954 they are defined only once.
955
956 21. This set of changes together give more compatible Unicode case-folding
957 behaviour for characters that have more than one other case when UCP
958 support is available.
959
960 (a) The Unicode property table now has offsets into a new table of sets of
961 three or more characters that are case-equivalent. The MultiStage2.py
962 script that generates these tables (the pcre_ucd.c file) now scans
963 CaseFolding.txt instead of UnicodeData.txt for character case
964 information.
965
966 (b) The code for adding characters or ranges of characters to a character
967 class has been abstracted into a generalized function that also handles
968 case-independence. In UTF-mode with UCP support, this uses the new data
969 to handle characters with more than one other case.
970
971 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
972 whose other case is greater than 256 are now correctly matched
973 caselessly. Previously, the high codepoint matched the low one, but not
974 vice versa.
975
976 (d) The processing of \h, \H, \v, and \ in character classes now makes use
977 of the new class addition function, using character lists defined as
978 macros alongside the case definitions of 20 above.
979
980 (e) Caseless back references now work with characters that have more than
981 one other case.
982
983 (f) General caseless matching of characters with more than one other case
984 is supported.
985
986 22. Unicode character properties were updated from Unicode 6.2.0
987
988 23. Improved CMake support under Windows. Patch by Daniel Richard G.
989
990 24. Add support for 32-bit character strings, and UTF-32
991
992 25. Major JIT compiler update (code refactoring and bugfixing).
993 Experimental Sparc 32 support is added.
994
995 26. Applied a modified version of Daniel Richard G's patch to create
996 pcre.h.generic and config.h.generic by "make" instead of in the
997 PrepareRelease script.
998
999 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1000 pcre_compile.c when checking for a zero character.
1001
1002 28. Introducing a native interface for JIT. Through this interface, the compiled
1003 machine code can be directly executed. The purpose of this interface is to
1004 provide fast pattern matching, so several sanity checks are not performed.
1005 However, feature tests are still performed. The new interface provides
1006 1.4x speedup compared to the old one.
1007
1008 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1009 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1010 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1011
1012 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1013 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1014 the "old" RFC 2279). Instead, it ended up passing a negative length to
1015 pcre_exec().
1016
1017 31. Add support for GCC's visibility feature to hide internal functions.
1018
1019 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1020 "unknown -C option" after outputting 0 or 1.
1021
1022 33. There is now support for generating a code coverage report for the test
1023 suite in environments where gcc is the compiler and lcov is installed. This
1024 is mainly for the benefit of the developers.
1025
1026 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1027 unaddressable using valgrind annotations, allowing valgrind to detect
1028 invalid memory accesses. This is mainly for the benefit of the developers.
1029
1030 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1031
1032 26. Give configure error if --enable-cpp but no C++ compiler found.
1033
1034
1035 Version 8.31 06-July-2012
1036 -------------------------
1037
1038 1. Fixing a wrong JIT test case and some compiler warnings.
1039
1040 2. Removed a bashism from the RunTest script.
1041
1042 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1043 to unsigned type, result still unsigned" that was given by an MS compiler
1044 on encountering the code "-sizeof(xxx)".
1045
1046 4. Partial matching support is added to the JIT compiler.
1047
1048 5. Fixed several bugs concerned with partial matching of items that consist
1049 of more than one character:
1050
1051 (a) /^(..)\1/ did not partially match "aba" because checking references was
1052 done on an "all or nothing" basis. This also applied to repeated
1053 references.
1054
1055 (b) \R did not give a hard partial match if \r was found at the end of the
1056 subject.
1057
1058 (c) \X did not give a hard partial match after matching one or more
1059 characters at the end of the subject.
1060
1061 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1062 a partial match for the string "\r".
1063
1064 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1065 a partial match for a CR character at the end of the subject string.
1066
1067 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1068 running pcretest, the text "(JIT)" added to the output whenever JIT is
1069 actually used to run the match.
1070
1071 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1072 or /S+[+] with a digit between 1 and 7.
1073
1074 8. OP_NOT now supports any UTF character not just single-byte ones.
1075
1076 9. (*MARK) control verb is now supported by the JIT compiler.
1077
1078 10. The command "./RunTest list" lists the available tests without actually
1079 running any of them. (Because I keep forgetting what they all are.)
1080
1081 11. Add PCRE_INFO_MAXLOOKBEHIND.
1082
1083 12. Applied a (slightly modified) user-supplied patch that improves performance
1084 when the heap is used for recursion (compiled with --disable-stack-for-
1085 recursion). Instead of malloc and free for each heap frame each time a
1086 logical recursion happens, frames are retained on a chain and re-used where
1087 possible. This sometimes gives as much as 30% improvement.
1088
1089 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1090 call.
1091
1092 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1093
1094 15. It is now possible to link pcretest with libedit as an alternative to
1095 libreadline.
1096
1097 16. (*COMMIT) control verb is now supported by the JIT compiler.
1098
1099 17. The Unicode data tables have been updated to Unicode 6.1.0.
1100
1101 18. Added --file-list option to pcregrep.
1102
1103 19. Added binary file support to pcregrep, including the -a, --binary-files,
1104 -I, and --text options.
1105
1106 20. The madvise function is renamed for posix_madvise for QNX compatibility
1107 reasons. Fixed by Giuseppe D'Angelo.
1108
1109 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1110
1111 22. Changed the option for creating symbolic links for 16-bit man pages from
1112 -s to -sf so that re-installing does not cause issues.
1113
1114 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1115
1116 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1117 matches in certain environments (the workspace was not being correctly
1118 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1119 some of the workspace data at the beginning of a restart.
1120
1121 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1122 was not doing so when it should - probably a typo introduced by SVN 528
1123 (change 8.10/14).
1124
1125 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1126 \w+ when the character tables indicated that \x{c4} was a word character.
1127 There were several related cases, all because the tests for doing a table
1128 lookup were testing for characters less than 127 instead of 255.
1129
1130 27. If a pattern contains capturing parentheses that are not used in a match,
1131 their slots in the ovector are set to -1. For those that are higher than
1132 any matched groups, this happens at the end of processing. In the case when
1133 there were back references that the ovector was too small to contain
1134 (causing temporary malloc'd memory to be used during matching), and the
1135 highest capturing number was not used, memory off the end of the ovector
1136 was incorrectly being set to -1. (It was using the size of the temporary
1137 memory instead of the true size.)
1138
1139 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1140 ovector size, it uses memory at the end of the block that it has got.
1141
1142 29. Check for an overlong MARK name and give an error at compile time. The
1143 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1144
1145 30. JIT compiler update.
1146
1147 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1148 Rill for the patch.
1149
1150 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1151
1152 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1153
1154 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1155 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1156 was enabled.
1157
1158 35. Improve JIT code generation for greedy plus quantifier.
1159
1160 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1161 1 to "aa" instead of to an empty string. The bug affected repeated groups
1162 that could potentially match an empty string.
1163
1164 37. Optimizing single character iterators in JIT.
1165
1166 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1167 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1168 codepoints that are too big for the mode are faulted, and in a UTF mode,
1169 disallowed codepoints are also faulted.
1170
1171 39. If PCRE was compiled with UTF support, in three places in the DFA
1172 matcher there was code that should only have been obeyed in UTF mode, but
1173 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1174 processing when bytes with values greater than 127 were present. In 16-bit
1175 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1176 both cases the values are those that cannot be the first data item in a UTF
1177 character. The three items that might have provoked this were recursions,
1178 possessively repeated groups, and atomic groups.
1179
1180 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1181 and pcregrep, because some OS require shared objects to be explicitly
1182 passed to ld, causing the link step to fail if they are not.
1183
1184 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1185 mode, patterns that started with \h* or \R* might be incorrectly matched.
1186
1187
1188 Version 8.30 04-February-2012
1189 -----------------------------
1190
1191 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1192 name is defined in ctype.h.
1193
1194 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1195 only in quite long subpatterns.
1196
1197 3. Removed the function pcre_info(), which has been obsolete and deprecated
1198 since it was replaced by pcre_fullinfo() in February 2000.
1199
1200 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1201 match a (*MARK), and the match failed at the start of the subject, a
1202 reference to memory before the start of the subject could occur. This bug
1203 was introduced by fix 17 of release 8.21.
1204
1205 5. A reference to an unset group with zero minimum repetition was giving
1206 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1207 /(another)?(\1?)test/ matched against "hello world test". This bug was
1208 introduced in release 8.13.
1209
1210 6. Add support for 16-bit character strings (a large amount of work involving
1211 many changes and refactorings).
1212
1213 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1214 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1215 from a file.
1216
1217 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1218 rounding is not applied in this particular case).
1219
1220 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1221 if they appear, or are escaped, in patterns.
1222
1223 10. Get rid of a number of -Wunused-but-set-variable warnings.
1224
1225 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1226 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1227 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1228 also returns the mark "x". This bug applied to capturing parentheses,
1229 non-capturing parentheses, and atomic parentheses. It also applied to some
1230 assertions.
1231
1232 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1233 information out of configure.ac instead of relying on pcre.h.generic, which
1234 is not stored in the repository.
1235
1236 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1237 -lreadline.
1238
1239 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1240
1241 15. Applied Graycode's patch to put the top-level frame on the stack rather
1242 than the heap when not using the stack for recursion. This gives a
1243 performance improvement in many cases when recursion is not deep.
1244
1245 16. Experimental code added to "pcretest -C" to output the stack frame size.
1246
1247
1248 Version 8.21 12-Dec-2011
1249 ------------------------
1250
1251 1. Updating the JIT compiler.
1252
1253 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1254 are added as well.
1255
1256 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1257 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1258 calling _pcre_jit_exec. Some extra comments are added.
1259
1260 4. (*MARK) settings inside atomic groups that do not contain any capturing
1261 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1262 was introduced by change 18 for 8.20.
1263
1264 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1265 ECMA-262 standard.
1266
1267 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1268 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1269 This bug was probably introduced by change 9 of 8.13.
1270
1271 7. While fixing 6 above, I noticed that a number of other items were being
1272 incorrectly rejected as "not fixed length". This arose partly because newer
1273 opcodes had not been added to the fixed-length checking code. I have (a)
1274 corrected the bug and added tests for these items, and (b) arranged for an
1275 error to occur if an unknown opcode is encountered while checking for fixed
1276 length instead of just assuming "not fixed length". The items that were
1277 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1278 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1279 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1280
1281 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1282 being incorrectly compiled and would have given unpredicatble results.
1283
1284 9. A possessively repeated subpattern with minimum repeat count greater than
1285 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1286 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1287 the first (A) could occur when it should not.
1288
1289 10. Add a cast and remove a redundant test from the code.
1290
1291 11. JIT should use pcre_malloc/pcre_free for allocation.
1292
1293 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1294 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1295 is anything other than /usr, -L is still shown).
1296
1297 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1298
1299 14. Perl does not support \N without a following name in a [] class; PCRE now
1300 also gives an error.
1301
1302 15. If a forward reference was repeated with an upper limit of around 2000,
1303 it caused the error "internal error: overran compiling workspace". The
1304 maximum number of forward references (including repeats) was limited by the
1305 internal workspace, and dependent on the LINK_SIZE. The code has been
1306 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1307 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1308 of around 200,000 forward references. While doing this, I also speeded up
1309 the filling in of repeated forward references.
1310
1311 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1312 incorrectly expecting the subject to contain another "a" after the start.
1313
1314 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1315 in the match, the SKIP should be ignored. This was not happening; instead
1316 the SKIP was being treated as NOMATCH. For patterns such as
1317 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1318 tested.
1319
1320 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1321 now much more compatible with Perl, in particular in cases where the result
1322 is a non-match for a non-anchored pattern. For example, if
1323 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1324 "m", where previously it did not return a name. A side effect of this
1325 change is that for partial matches, the last encountered mark name is
1326 returned, as for non matches. A number of tests that were previously not
1327 Perl-compatible have been moved into the Perl-compatible test files. The
1328 refactoring has had the pleasing side effect of removing one argument from
1329 the match() function, thus reducing its stack requirements.
1330
1331 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1332 subsequent uses of /S (without +) incorrectly behaved like /S+.
1333
1334 21. Retrieve executable code size support for the JIT compiler and fixing
1335 some warnings.
1336
1337 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1338 not work when the shorter character appeared right at the end of the
1339 subject string.
1340
1341 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1342 systems.
1343
1344 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1345 output it when the /M option is used in pcretest.
1346
1347 25. The CheckMan script was not being included in the distribution. Also, added
1348 an explicit "perl" to run Perl scripts from the PrepareRelease script
1349 because this is reportedly needed in Windows.
1350
1351 26. If study data was being save in a file and studying had not found a set of
1352 "starts with" bytes for the pattern, the data written to the file (though
1353 never used) was taken from uninitialized memory and so caused valgrind to
1354 complain.
1355
1356 27. Updated RunTest.bat as provided by Sheri Pierce.
1357
1358 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1359
1360 29. Computation of memory usage for the table of capturing group names was
1361 giving an unnecessarily large value.
1362
1363
1364 Version 8.20 21-Oct-2011
1365 ------------------------
1366
1367 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1368 a POSIX class. After further experiments with Perl, which convinced me that
1369 Perl has bugs and confusions, a closing square bracket is no longer allowed
1370 in a POSIX name. This bug also affected patterns with classes that started
1371 with full stops.
1372
1373 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1374 captured substring, but while checking the failing first alternative,
1375 substring 1 is temporarily captured. If the output vector supplied to
1376 pcre_exec() was not big enough for this capture, the yield of the function
1377 was still zero ("insufficient space for captured substrings"). This cannot
1378 be totally fixed without adding another stack variable, which seems a lot
1379 of expense for a edge case. However, I have improved the situation in cases
1380 such as /(a)(b)x|abc/ matched against "abc", where the return code
1381 indicates that fewer than the maximum number of slots in the ovector have
1382 been set.
1383
1384 3. Related to (2) above: when there are more back references in a pattern than
1385 slots in the output vector, pcre_exec() uses temporary memory during
1386 matching, and copies in the captures as far as possible afterwards. It was
1387 using the entire output vector, but this conflicts with the specification
1388 that only 2/3 is used for passing back captured substrings. Now it uses
1389 only the first 2/3, for compatibility. This is, of course, another edge
1390 case.
1391
1392 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1393 main code base, and can be used by building with --enable-jit. When this is
1394 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1395 runtime --no-jit option is given.
1396
1397 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1398 ovector, the return from the function was zero, implying that there were
1399 other matches that did not fit. The correct "exactly full" value is now
1400 returned.
1401
1402 6. If a subpattern that was called recursively or as a subroutine contained
1403 (*PRUNE) or any other control that caused it to give a non-standard return,
1404 invalid errors such as "Error -26 (nested recursion at the same subject
1405 position)" or even infinite loops could occur.
1406
1407 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1408 computing the minimum length on reaching *ACCEPT, and so ended up with the
1409 wrong value of 1 rather than 0. Further investigation indicates that
1410 computing a minimum subject length in the presence of *ACCEPT is difficult
1411 (think back references, subroutine calls), and so I have changed the code
1412 so that no minimum is registered for a pattern that contains *ACCEPT.
1413
1414 8. If (*THEN) was present in the first (true) branch of a conditional group,
1415 it was not handled as intended. [But see 16 below.]
1416
1417 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1418 Sheri Pierce.
1419
1420 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1421 the first byte in a match must be "a".
1422
1423 11. Change 17 for 8.13 increased the recursion depth for patterns like
1424 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1425 pattern contains any instances of (*THEN). If it does not, the old
1426 optimizations are restored. It would be nice to do this on a per-group
1427 basis, but at the moment that is not feasible.
1428
1429 12. In some environments, the output of pcretest -C is CRLF terminated. This
1430 broke RunTest's code that checks for the link size. A single white space
1431 character after the value is now allowed for.
1432
1433 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1434 For "fr", it uses the Windows-specific input and output files.
1435
1436 14. If (*THEN) appeared in a group that was called recursively or as a
1437 subroutine, it did not work as intended. [But see next item.]
1438
1439 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1440 pattern fragments (but not containing any | characters). If A and B are
1441 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1442 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1443 D. In other words, Perl considers parentheses that do not contain any |
1444 characters to be part of a surrounding alternative, whereas PCRE was
1445 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1446 differently. PCRE now behaves in the same way as Perl, except in the case
1447 of subroutine/recursion calls such as (?1) which have in any case always
1448 been different (but PCRE had them first :-).
1449
1450 16. Related to 15 above: Perl does not treat the | in a conditional group as
1451 creating alternatives. Such a group is treated in the same way as an
1452 ordinary group without any | characters when processing (*THEN). PCRE has
1453 been changed to match Perl's behaviour.
1454
1455 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1456 RunGrepTest script failed.
1457
1458 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1459 inevitable for groups that contain captures, but it can lead to a lot of
1460 stack use in large patterns. The old behaviour has been restored for atomic
1461 groups that do not contain any capturing parentheses.
1462
1463 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1464 suppress the check for a minimum subject length at run time. (If it was
1465 given to pcre_exec() or pcre_dfa_exec() it did work.)
1466
1467 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1468 fail to work when decoding hex characters in data strings in EBCDIC
1469 environments.
1470
1471 21. It appears that in at least one Mac OS environment, the isxdigit() function
1472 is implemented as a macro that evaluates to its argument more than once,
1473 contravening the C 90 Standard (I haven't checked a later standard). There
1474 was an instance in pcretest which caused it to go wrong when processing
1475 \x{...} escapes in subject strings. The has been rewritten to avoid using
1476 things like p++ in the argument of isxdigit().
1477
1478
1479 Version 8.13 16-Aug-2011
1480 ------------------------
1481
1482 1. The Unicode data tables have been updated to Unicode 6.0.0.
1483
1484 2. Two minor typos in pcre_internal.h have been fixed.
1485
1486 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1487 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1488 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1489
1490 4. There were a number of related bugs in the code for matching backrefences
1491 caselessly in UTF-8 mode when codes for the characters concerned were
1492 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1493 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1494 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1495 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1496 2-byte code at the end of the subject (it thought there wasn't enough data
1497 left).
1498
1499 5. Comprehensive information about what went wrong is now returned by
1500 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1501 as the output vector has at least 2 elements. The offset of the start of
1502 the failing character and a reason code are placed in the vector.
1503
1504 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1505 now returned is for the first byte of the failing character, instead of the
1506 last byte inspected. This is an incompatible change, but I hope it is small
1507 enough not to be a problem. It makes the returned offset consistent with
1508 pcre_exec() and pcre_dfa_exec().
1509
1510 7. pcretest now gives a text phrase as well as the error number when
1511 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1512 failure, the offset and reason code are output.
1513
1514 8. When \R was used with a maximizing quantifier it failed to skip backwards
1515 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1516 back over a single character (\n). This seems wrong (because it treated the
1517 two characters as a single entity when going forwards), conflicts with the
1518 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1519 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1520 has been changed.
1521
1522 9. Some internal refactoring has changed the processing so that the handling
1523 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1524 time (the PCRE_DOTALL option was changed this way some time ago: version
1525 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1526 which was always a bit of a fudge. It also means that there is one less
1527 argument for the match() function, which reduces its stack requirements
1528 slightly. This change also fixes an incompatibility with Perl: the pattern
1529 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1530
1531 10. More internal refactoring has drastically reduced the number of recursive
1532 calls to match() for possessively repeated groups such as (abc)++ when
1533 using pcre_exec().
1534
1535 11. While implementing 10, a number of bugs in the handling of groups were
1536 discovered and fixed:
1537
1538 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1539 (a|)*(?1) gave a compile-time internal error.
1540 ((a|)+)+ did not notice that the outer group could match an empty string.
1541 (^a|^)+ was not marked as anchored.
1542 (.*a|.*)+ was not marked as matching at start or after a newline.
1543
1544 12. Yet more internal refactoring has removed another argument from the match()
1545 function. Special calls to this function are now indicated by setting a
1546 value in a variable in the "match data" data block.
1547
1548 13. Be more explicit in pcre_study() instead of relying on "default" for
1549 opcodes that mean there is no starting character; this means that when new
1550 ones are added and accidentally left out of pcre_study(), testing should
1551 pick them up.
1552
1553 14. The -s option of pcretest has been documented for ages as being an old
1554 synonym of -m (show memory usage). I have changed it to mean "force study
1555 for every regex", that is, assume /S for every regex. This is similar to -i
1556 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1557 using it. It makes it easier to run collections of tests with and without
1558 study enabled, and thereby test pcre_study() more easily. All the standard
1559 tests are now run with and without -s (but some patterns can be marked as
1560 "never study" - see 20 below).
1561
1562 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1563 restoration of the capturing data to the outer values was not happening
1564 correctly.
1565
1566 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1567 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1568 pattern had matched an empty string, and so incorrectly returned a no
1569 match.
1570
1571 17. There was optimizing code for the last branch of non-capturing parentheses,
1572 and also for the obeyed branch of a conditional subexpression, which used
1573 tail recursion to cut down on stack usage. Unfortunately, now that there is
1574 the possibility of (*THEN) occurring in these branches, tail recursion is
1575 no longer possible because the return has to be checked for (*THEN). These
1576 two optimizations have therefore been removed. [But see 8.20/11 above.]
1577
1578 18. If a pattern containing \R was studied, it was assumed that \R always
1579 matched two bytes, thus causing the minimum subject length to be
1580 incorrectly computed because \R can also match just one byte.
1581
1582 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1583 was incorrectly computed.
1584
1585 20. If /S is present twice on a test pattern in pcretest input, it now
1586 *disables* studying, thereby overriding the use of -s on the command line
1587 (see 14 above). This is necessary for one or two tests to keep the output
1588 identical in both cases.
1589
1590 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1591 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1592
1593 22. When an atomic group that contained a capturing parenthesis was
1594 successfully matched, but the branch in which it appeared failed, the
1595 capturing was not being forgotten if a higher numbered group was later
1596 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1597 group 1 to "a", when in fact it should be unset. This applied to multi-
1598 branched capturing and non-capturing groups, repeated or not, and also to
1599 positive assertions (capturing in negative assertions does not happen
1600 in PCRE) and also to nested atomic groups.
1601
1602 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1603 subject after a captured substring, to make it easier to tell which of a
1604 number of identical substrings has been captured.
1605
1606 24. The way atomic groups are processed by pcre_exec() has been changed so that
1607 if they are repeated, backtracking one repetition now resets captured
1608 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1609 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1610 "aaa". Previously, it would have been "a". As part of this code
1611 refactoring, the way recursive calls are handled has also been changed.
1612
1613 25. If an assertion condition captured any substrings, they were not passed
1614 back unless some other capturing happened later. For example, if
1615 (?(?=(a))a) was matched against "a", no capturing was returned.
1616
1617 26. When studying a pattern that contained subroutine calls or assertions,
1618 the code for finding the minimum length of a possible match was handling
1619 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1620 group 1 called group 2 while simultaneously a separate group 2 called group
1621 1). A stack overflow occurred in this case. I have fixed this by limiting
1622 the recursion depth to 10.
1623
1624 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1625 Fortmann. This supports explicit test numbers on the command line, and has
1626 argument validation and error reporting.
1627
1628 28. An instance of \X with an unlimited repeat could fail if at any point the
1629 first character it looked at was a mark character.
1630
1631 29. Some minor code refactoring concerning Unicode properties and scripts
1632 should reduce the stack requirement of match() slightly.
1633
1634 30. Added the '=' option to pcretest to check the setting of unused capturing
1635 slots at the end of the pattern, which are documented as being -1, but are
1636 not included in the return count.
1637
1638 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1639 compiled something random. Now it gives a compile-time error (as does
1640 Perl).
1641
1642 32. A *MARK encountered during the processing of a positive assertion is now
1643 recorded and passed back (compatible with Perl).
1644
1645 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1646 had alternative anchored branches, the search for a second match in a line
1647 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1648 matched the line "0102" twice. The same bug affected patterns that started
1649 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1650 twice.
1651
1652 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1653 does, and because of capturing effects, quantifying parenthesized
1654 assertions may at times be useful. Quantifiers are now allowed for
1655 parenthesized assertions.
1656
1657 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1658
1659 36. \g was being checked for fancy things in a character class, when it should
1660 just be a literal "g".
1661
1662 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1663 appearance of a nested POSIX class supersedes an apparent external class.
1664 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1665 unescaped square brackets may also appear as part of class names. For
1666 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1667 more like Perl. (But see 8.20/1 above.)
1668
1669 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1670 was because it thought it was \N{name}, which is not supported).
1671
1672 39. Add minix to OS list not supporting the -S option in pcretest.
1673
1674 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1675 cannot analyze patterns in sufficient detail to catch mutual recursions
1676 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1677 subgroup is called recursively as a subpattern for a second time at the
1678 same position in the subject string. In previous releases this might have
1679 been caught by the recursion limit, or it might have run out of stack.
1680
1681 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1682 happen only once. PCRE was, however incorrectly giving a compile time error
1683 "recursive call could loop indefinitely" because it cannot analyze the
1684 pattern in sufficient detail. The compile time test no longer happens when
1685 PCRE is compiling a conditional subpattern, but actual runaway loops are
1686 now caught at runtime (see 40 above).
1687
1688 42. It seems that Perl allows any characters other than a closing parenthesis
1689 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1690 has been changed to be the same.
1691
1692 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1693 as not to get warnings when autogen.sh is called. Also changed
1694 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1695
1696 44. To help people who use pcregrep to scan files containing exceedingly long
1697 lines, the following changes have been made:
1698
1699 (a) The default value of the buffer size parameter has been increased from
1700 8K to 20K. (The actual buffer used is three times this size.)
1701
1702 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1703 PCRE is built.
1704
1705 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1706 to be set at run time.
1707
1708 (d) Numerical values in pcregrep options can be followed by K or M, for
1709 example --buffer-size=50K.
1710
1711 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1712 given and the return code is set to 2.
1713
1714 45. Add a pointer to the latest mark to the callout data block.
1715
1716 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1717 partial match of an empty string instead of no match. This was specific to
1718 the use of ".".
1719
1720 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1721 complete match instead of a partial match. This bug was dependent on both
1722 the PCRE_UTF8 and PCRE_DOTALL options being set.
1723
1724 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1725 starting byte set, because \b was not being ignored.
1726
1727
1728 Version 8.12 15-Jan-2011
1729 ------------------------
1730
1731 1. Fixed some typos in the markup of the man pages, and wrote a script that
1732 checks for such things as part of the documentation building process.
1733
1734 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1735 --match-limit and --recursion-limit options (added for 8.11). In
1736 particular, this made one of the standard tests fail. (The integer value
1737 went into the wrong half of a long int.)
1738
1739 3. If the --colour option was given to pcregrep with -v (invert match), it
1740 did strange things, either producing crazy output, or crashing. It should,
1741 of course, ignore a request for colour when reporting lines that do not
1742 match.
1743
1744 4. Another pcregrep bug caused similar problems if --colour was specified with
1745 -M (multiline) and the pattern match finished with a line ending.
1746
1747 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1748 matched in multiline mode, the following line was shown as part of the
1749 match. This seems wrong, so I have changed it.
1750
1751 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1752 the check for further matches in the same line (so they could be coloured)
1753 to overrun the end of the current line. If another match was found, it was
1754 incorrectly shown (and then shown again when found in the next line).
1755
1756 7. If pcregrep was compiled under Windows, there was a reference to the
1757 function pcregrep_exit() before it was defined. I am assuming this was
1758 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1759 reported by a user. I've moved the definition above the reference.
1760
1761
1762 Version 8.11 10-Dec-2010
1763 ------------------------
1764
1765 1. (*THEN) was not working properly if there were untried alternatives prior
1766 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1767 backtracked to try for "b" instead of moving to the next alternative branch
1768 at the same level (in this case, to look for "c"). The Perl documentation
1769 is clear that when (*THEN) is backtracked onto, it goes to the "next
1770 alternative in the innermost enclosing group".
1771
1772 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1773 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1774 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1775 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1776 (*THEN).
1777
1778 3. If \s appeared in a character class, it removed the VT character from
1779 the class, even if it had been included by some previous item, for example
1780 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1781 of \s, but is part of the POSIX "space" class.)
1782
1783 4. A partial match never returns an empty string (because you can always
1784 match an empty string at the end of the subject); however the checking for
1785 an empty string was starting at the "start of match" point. This has been
1786 changed to the "earliest inspected character" point, because the returned
1787 data for a partial match starts at this character. This means that, for
1788 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1789 (previously it gave "no match").
1790
1791 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1792 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1793 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1794 has an implication that the given string is incomplete (because a partial
1795 match is preferred over a full match). For this reason, these items now
1796 give a partial match in this situation. [Aside: previously, the one case
1797 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1798 match rather than a full match, which was wrong by the old rules, but is
1799 now correct.]
1800
1801 6. There was a bug in the handling of #-introduced comments, recognized when
1802 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1803 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1804 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1805 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1806 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1807 places in pcre_compile().
1808
1809 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1810 comments when looking ahead for named forward references to subpatterns,
1811 the only newline sequence it recognized was NL. It now handles newlines
1812 according to the set newline convention.
1813
1814 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1815 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1816 cater for a lack of strerror(). These oversights have been fixed.
1817
1818 9. Added --match-limit and --recursion-limit to pcregrep.
1819
1820 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1821
1822 11. When the -o option was used, pcregrep was setting a return code of 1, even
1823 when matches were found, and --line-buffered was not being honoured.
1824
1825 12. Added an optional parentheses number to the -o and --only-matching options
1826 of pcregrep.
1827
1828 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1829 can match an empty string. The code to do it in pcretest and pcredemo
1830 needed fixing:
1831
1832 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1833 only one byte after an empty string match just before CRLF (this case
1834 just got forgotten; "any" and "anycrlf" were OK).
1835
1836 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1837 mode when an empty string match preceded an ASCII character followed by
1838 a non-ASCII character. (The code for advancing by one character rather
1839 than one byte was nonsense.)
1840
1841 (c) The pcredemo.c sample program did not have any code at all to handle
1842 the cases when CRLF is a valid newline sequence.
1843
1844 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1845 as a starting offset was within the subject string. There is now a new
1846 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1847 negative or greater than the length of the string. In order to test this,
1848 pcretest is extended to allow the setting of negative starting offsets.
1849
1850 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1851 starting offset points to the beginning of a UTF-8 character was
1852 unnecessarily clumsy. I tidied it up.
1853
1854 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1855 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1856
1857 17. Nobody had reported that the --include_dir option, which was added in
1858 release 7.7 should have been called --include-dir (hyphen, not underscore)
1859 for compatibility with GNU grep. I have changed it to --include-dir, but
1860 left --include_dir as an undocumented synonym, and the same for
1861 --exclude-dir, though that is not available in GNU grep, at least as of
1862 release 2.5.4.
1863
1864 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1865 characters from a string of bytes) have been redefined so as not to use
1866 loops, in order to improve performance in some environments. At the same
1867 time, I abstracted some of the common code into auxiliary macros to save
1868 repetition (this should not affect the compiled code).
1869
1870 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1871 compile-time error is now given if \c is not followed by an ASCII
1872 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1873 different, and any byte value is allowed.)
1874
1875 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1876 START_OPTIMIZE option, which is now allowed at compile time - but just
1877 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1878 to pcregrep and other applications that have no direct access to PCRE
1879 options. The new /Y option in pcretest sets this option when calling
1880 pcre_compile().
1881
1882 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1883 back references. Groups containing recursive back references were forced to
1884 be atomic by that change, but in the case of named groups, the amount of
1885 memory required was incorrectly computed, leading to "Failed: internal
1886 error: code overflow". This has been fixed.
1887
1888 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1889 pcretest.c, to avoid build problems in some Borland environments.
1890
1891
1892 Version 8.10 25-Jun-2010
1893 ------------------------
1894
1895 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1896 THEN.
1897
1898 2. (*ACCEPT) was not working when inside an atomic group.
1899
1900 3. Inside a character class, \B is treated as a literal by default, but
1901 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1902 causes the error). The code is unchanged, but I tidied the documentation.
1903
1904 4. Inside a character class, PCRE always treated \R and \X as literals,
1905 whereas Perl faults them if its -w option is set. I have changed PCRE so
1906 that it faults them when PCRE_EXTRA is set.
1907
1908 5. Added support for \N, which always matches any character other than
1909 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1910
1911 6. When compiling pcregrep with newer versions of gcc which may have
1912 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1913 declared with attribute warn_unused_result" were given. Just casting the
1914 result to (void) does not stop the warnings; a more elaborate fudge is
1915 needed. I've used a macro to implement this.
1916
1917 7. Minor change to pcretest.c to avoid a compiler warning.
1918
1919 8. Added four artifical Unicode properties to help with an option to make
1920 \s etc use properties (see next item). The new properties are: Xan
1921 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1922
1923 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1924 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1925 this option. Modified pcretest to add /W to test this facility. Added
1926 REG_UCP to make it available via the POSIX interface.
1927
1928 10. Added --line-buffered to pcregrep.
1929
1930 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1931 studied, and the match started with a letter with a code point greater than
1932 127 whose first byte was different to the first byte of the other case of
1933 the letter, the other case of this starting letter was not recognized
1934 (#976).
1935
1936 12. If a pattern that was studied started with a repeated Unicode property
1937 test, for example, \p{Nd}+, there was the theoretical possibility of
1938 setting up an incorrect bitmap of starting bytes, but fortunately it could
1939 not have actually happened in practice until change 8 above was made (it
1940 added property types that matched character-matching opcodes).
1941
1942 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1943 possible starting bytes for non-anchored patterns.
1944
1945 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1946 \R, and also a number of cases that involve Unicode properties, both
1947 explicit and implicit when PCRE_UCP is set.
1948
1949 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1950 input, it could crash or give wrong results if characters with values
1951 greater than 0xc0 were present in the subject string. (Detail: it assumed
1952 UTF-8 input when processing these items.)
1953
1954 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1955 size_t is 64-bit (#991).
1956
1957 17. Added a check for running out of memory when PCRE is compiled with
1958 --disable-stack-for-recursion (#990).
1959
1960 18. If the last data line in a file for pcretest does not have a newline on
1961 the end, a newline was missing in the output.
1962
1963 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1964 less than 128) in its various bitmaps. However, there is a facility for
1965 generating tables according to the current locale when PCRE is compiled. It
1966 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1967 characters, are recognized by isspace() and therefore were getting set in
1968 these tables, and indeed these tables seem to approximate to ISO 8859. This
1969 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1970 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1971 which of course cannot start UTF-8 characters. I have changed the code so
1972 that only real ASCII characters (less than 128) and the correct starting
1973 bytes for UTF-8 encodings are set for characters greater than 127 when in
1974 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1975 altogether.)
1976
1977 20. Added the /T option to pcretest so as to be able to run tests with non-
1978 standard character tables, thus making it possible to include the tests
1979 used for 19 above in the standard set of tests.
1980
1981 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1982 reference to a subpattern the other side of a comment that contains an
1983 opening parenthesis caused either an internal compiling error, or a
1984 reference to the wrong subpattern.
1985
1986
1987 Version 8.02 19-Mar-2010
1988 ------------------------
1989
1990 1. The Unicode data tables have been updated to Unicode 5.2.0.
1991
1992 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1993 configured.
1994
1995 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1996 original author of that file, following a query about its status.
1997
1998 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1999 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2000
2001 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2002 quantifier applied to a forward-referencing subroutine call, could compile
2003 incorrect code or give the error "internal error: previously-checked
2004 referenced subpattern not found".
2005
2006 6. Both MS Visual Studio and Symbian OS have problems with initializing
2007 variables to point to external functions. For these systems, therefore,
2008 pcre_malloc etc. are now initialized to local functions that call the
2009 relevant global functions.
2010
2011 7. There were two entries missing in the vectors called coptable and poptable
2012 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2013 I've fixed the data, and added a kludgy way of testing at compile time that
2014 the lengths are correct (equal to the number of opcodes).
2015
2016 8. Following on from 7, I added a similar kludge to check the length of the
2017 eint vector in pcreposix.c.
2018
2019 9. Error texts for pcre_compile() are held as one long string to avoid too
2020 much relocation at load time. To find a text, the string is searched,
2021 counting zeros. There was no check for running off the end of the string,
2022 which could happen if a new error number was added without updating the
2023 string.
2024
2025 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2026
2027 11. \K was not working if it appeared in an atomic group or in a group that
2028 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2029 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2030 the assertion is positive, but not if it is negative.
2031
2032 12. Change 11 fortuitously reduced the size of the stack frame used in the
2033 "match()" function of pcre_exec.c by one pointer. Forthcoming
2034 implementation of support for (*MARK) will need an extra pointer on the
2035 stack; I have reserved it now, so that the stack frame size does not
2036 decrease.
2037
2038 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2039 item in branch that calls a recursion is a subroutine call - as in the
2040 second branch in the above example - was incorrectly given the compile-
2041 time error "recursive call could loop indefinitely" because pcre_compile()
2042 was not correctly checking the subroutine for matching a non-empty string.
2043
2044 14. The checks for overrunning compiling workspace could trigger after an
2045 overrun had occurred. This is a "should never occur" error, but it can be
2046 triggered by pathological patterns such as hundreds of nested parentheses.
2047 The checks now trigger 100 bytes before the end of the workspace.
2048
2049 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2050
2051
2052 Version 8.01 19-Jan-2010
2053 ------------------------
2054
2055 1. If a pattern contained a conditional subpattern with only one branch (in
2056 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2057 computed the wrong minimum data length (which is of course zero for such
2058 subpatterns). This could cause incorrect "no match" results.
2059
2060 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2061 the pattern is reset in the first branch, pcre_compile() failed with
2062 "internal error: code overflow at offset...". This happened only when
2063 the reset was to the original external option setting. (An optimization
2064 abstracts leading options settings into an external setting, which was the
2065 cause of this.)
2066
2067 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2068 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2069 assertion pattern did not match (meaning that the assertion was true), it
2070 was incorrectly treated as false if the SKIP had been reached during the
2071 matching. This also applied to assertions used as conditions.
2072
2073 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2074 assertion subpattern, including such a pattern used as a condition,
2075 unpredictable results occurred, instead of the error return
2076 PCRE_ERROR_DFA_UITEM.
2077
2078 5. The C++ GlobalReplace function was not working like Perl for the special
2079 situation when an empty string is matched. It now does the fancy magic
2080 stuff that is necessary.
2081
2082 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2083 removed. (These were left over from very, very early versions of PCRE.)
2084
2085 7. Some cosmetic changes to the code to make life easier when compiling it
2086 as part of something else:
2087
2088 (a) Change DEBUG to PCRE_DEBUG.
2089
2090 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2091 called "current" as "current_branch", to prevent a collision with the
2092 Linux macro when compiled as a kernel module.
2093
2094 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2095 prevent a collision with the Linux macro when compiled as a kernel
2096 module.
2097
2098 8. In pcre_compile() there are some checks for integer overflows that used to
2099 cast potentially large values to (double). This has been changed to that
2100 when building, a check for int64_t is made, and if it is found, it is used
2101 instead, thus avoiding the use of floating point arithmetic. (There is no
2102 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2103 double.
2104
2105 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2106 2005 (difference between two addresses compared to an unsigned value).
2107
2108 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2109 custom one, because of the following reported problem in Windows:
2110
2111 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2112 under Win32.
2113 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2114 therefore missing the function definition.
2115 - The compiler thus generates a "C" signature for the test function.
2116 - The linker fails to find the "C" function.
2117 - PCRE fails to configure if asked to do so against libbz2.
2118
2119 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2120 messages were output:
2121
2122 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2123 rerunning libtoolize, to keep the correct libtool macros in-tree.
2124 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2125
2126 I have done both of these things.
2127
2128 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2129 most of the time, it *can* run out if it is given a pattern that contains a
2130 runaway infinite recursion. I updated the discussion in the pcrestack man
2131 page.
2132
2133 13. Now that we have gone to the x.xx style of version numbers, the minor
2134 version may start with zero. Using 08 or 09 is a bad idea because users
2135 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2136 interpreted as invalid octal numbers. I've updated the previous comment in
2137 configure.ac, and also added a check that gives an error if 08 or 09 are
2138 used.
2139
2140 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2141 causing partial matching to fail when the end of the subject matched \W
2142 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2143
2144 15. There were some discrepancies between the declarations in pcre_internal.h
2145 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2146 their definitions. The declarations used "const uschar *" and the
2147 definitions used USPTR. Even though USPTR is normally defined as "const
2148 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2149 reported that: "This difference in casting confuses some C++ compilers, for
2150 example, SunCC recognizes above declarations as different functions and
2151 generates broken code for hbpcre." I have changed the declarations to use
2152 USPTR.
2153
2154 16. GNU libtool is named differently on some systems. The autogen.sh script now
2155 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2156 (FreeBSD).
2157
2158 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2159 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2160 comment: "Figure out how to create a longlong from a string: strtoll and
2161 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2162 instance, but it only takes 2 args instead of 3!"
2163
2164 18. A subtle bug concerned with back references has been fixed by a change of
2165 specification, with a corresponding code fix. A pattern such as
2166 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2167 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2168 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2169 same bug. Such groups have to be quantified to be useful, or contained
2170 inside another quantified group. (If there's no repetition, the reference
2171 can never match.) The problem arises because, having left the group and
2172 moved on to the rest of the pattern, a later failure that backtracks into
2173 the group uses the captured value from the final iteration of the group
2174 rather than the correct earlier one. I have fixed this in PCRE by forcing
2175 any group that contains a reference to itself to be an atomic group; that
2176 is, there cannot be any backtracking into it once it has completed. This is
2177 similar to recursive and subroutine calls.
2178
2179
2180 Version 8.00 19-Oct-09
2181 ----------------------
2182
2183 1. The table for translating pcre_compile() error codes into POSIX error codes
2184 was out-of-date, and there was no check on the pcre_compile() error code
2185 being within the table. This could lead to an OK return being given in
2186 error.
2187
2188 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2189 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2190 in a Windows environment.
2191
2192 3. The pcregrep --count option prints the count for each file even when it is
2193 zero, as does GNU grep. However, pcregrep was also printing all files when
2194 --files-with-matches was added. Now, when both options are given, it prints
2195 counts only for those files that have at least one match. (GNU grep just
2196 prints the file name in this circumstance, but including the count seems
2197 more useful - otherwise, why use --count?) Also ensured that the
2198 combination -clh just lists non-zero counts, with no names.
2199
2200 4. The long form of the pcregrep -F option was incorrectly implemented as
2201 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2202 but it seems right to fix it, and I didn't think it was worth preserving
2203 the old behaviour.
2204
2205 5. The command line items --regex=pattern and --regexp=pattern were not
2206 recognized by pcregrep, which required --regex pattern or --regexp pattern
2207 (with a space rather than an '='). The man page documented the '=' forms,
2208 which are compatible with GNU grep; these now work.
2209
2210 6. No libpcreposix.pc file was created for pkg-config; there was just
2211 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2212
2213 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2214 when UCP support is not needed, by modifying the Python script that
2215 generates it from Unicode data files. This should not matter if the module
2216 is correctly used as a library, but I received one complaint about 50K of
2217 unwanted data. My guess is that the person linked everything into his
2218 program rather than using a library. Anyway, it does no harm.
2219
2220 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2221 was a minimum greater than 1 for a wide character in a possessive
2222 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2223 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2224 character. Chaos in the form of incorrect output or a compiling loop could
2225 result.
2226
2227 9. The restrictions on what a pattern can contain when partial matching is
2228 requested for pcre_exec() have been removed. All patterns can now be
2229 partially matched by this function. In addition, if there are at least two
2230 slots in the offset vector, the offset of the earliest inspected character
2231 for the match and the offset of the end of the subject are set in them when
2232 PCRE_ERROR_PARTIAL is returned.
2233
2234 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2235 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2236 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2237 and may be more useful for multi-segment matching.
2238
2239 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2240 used to be given if ever the end of the subject was reached; now it is
2241 given only if matching could not proceed because another character was
2242 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2243 string "Z", which now yields "no match" instead of "partial match". In the
2244 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2245 final character ended with (*FAIL).
2246
2247 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2248 if the pattern had a "must contain" character that was already found in the
2249 earlier partial match, unless partial matching was again requested. For
2250 example, with the pattern /dog.(body)?/, the "must contain" character is
2251 "g". If the first part-match was for the string "dog", restarting with
2252 "sbody" failed. This bug has been fixed.
2253
2254 13. The string returned by pcre_dfa_exec() after a partial match has been
2255 changed so that it starts at the first inspected character rather than the
2256 first character of the match. This makes a difference only if the pattern
2257 starts with a lookbehind assertion or \b or \B (\K is not supported by
2258 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2259 matching functions compatible, and I think it's the right thing to do.
2260
2261 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2262 so that the demonstration program is easily available in environments where
2263 PCRE has not been installed from source.
2264
2265 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2266 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2267 library.
2268
2269 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2270 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2271 is not the first non-POSIX option to be added. Clearly some people find
2272 these options useful.
2273
2274 17. If a caller to the POSIX matching function regexec() passes a non-zero
2275 value for nmatch with a NULL value for pmatch, the value of
2276 nmatch is forced to zero.
2277
2278 18. RunGrepTest did not have a test for the availability of the -u option of
2279 the diff command, as RunTest does. It now checks in the same way as
2280 RunTest, and also checks for the -b option.
2281
2282 19. If an odd number of negated classes containing just a single character
2283 interposed, within parentheses, between a forward reference to a named
2284 subpattern and the definition of the subpattern, compilation crashed with
2285 an internal error, complaining that it could not find the referenced
2286 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2287 [The bug was that it was starting one character too far in when skipping
2288 over the character class, thus treating the ] as data rather than
2289 terminating the class. This meant it could skip too much.]
2290
2291 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2292 /g option in pcretest when the pattern contains \K, which makes it possible
2293 to have an empty string match not at the start, even when the pattern is
2294 anchored. Updated pcretest and pcredemo to use this option.
2295
2296 21. If the maximum number of capturing subpatterns in a recursion was greater
2297 than the maximum at the outer level, the higher number was returned, but
2298 with unset values at the outer level. The correct (outer level) value is
2299 now given.
2300
2301 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2302 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2303 make it do so. The string so far is captured, making this feature
2304 compatible with Perl.
2305
2306 23. The tests have been re-organized, adding tests 11 and 12, to make it
2307 possible to check the Perl 5.10 features against Perl 5.10.
2308
2309 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2310 pattern matches a fixed length string. PCRE did not allow this; now it
2311 does. Neither allows recursion.
2312
2313 25. I finally figured out how to implement a request to provide the minimum
2314 length of subject string that was needed in order to match a given pattern.
2315 (It was back references and recursion that I had previously got hung up
2316 on.) This code has now been added to pcre_study(); it finds a lower bound
2317 to the length of subject needed. It is not necessarily the greatest lower
2318 bound, but using it to avoid searching strings that are too short does give
2319 some useful speed-ups. The value is available to calling programs via
2320 pcre_fullinfo().
2321
2322 26. While implementing 25, I discovered to my embarrassment that pcretest had
2323 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2324 study optimizations had never been tested with that matching function.
2325 Oops. What is worse, even when it was passed study data, there was a bug in
2326 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2327 were also very few tests of studied patterns with pcre_dfa_exec().
2328
2329 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2330 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2331 on the other side of the coin, they are no longer allowed to have different
2332 names, because these cannot be distinguished in PCRE, and this has caused
2333 confusion. (This is a difference from Perl.)
2334
2335 28. When duplicate subpattern names are present (necessarily with different
2336 numbers, as required by 27 above), and a test is made by name in a
2337 conditional pattern, either for a subpattern having been matched, or for
2338 recursion in such a pattern, all the associated numbered subpatterns are
2339 tested, and the overall condition is true if the condition is true for any
2340 one of them. This is the way Perl works, and is also more like the way
2341 testing by number works.
2342
2343
2344 Version 7.9 11-Apr-09
2345 ---------------------
2346
2347 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2348 (pcretest), all targets were linked against these libraries. This included
2349 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2350 libraries. This caused unwanted dependencies to be created. This problem
2351 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2352 pcretest is linked with readline.
2353
2354 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2355 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2356 moved outside it again, because FALSE and TRUE are already defined in AIX,
2357 but BOOL is not.
2358
2359 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2360 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2361
2362 4. The pcregrep documentation said spaces were inserted as well as colons (or
2363 hyphens) following file names and line numbers when outputting matching
2364 lines. This is not true; no spaces are inserted. I have also clarified the
2365 wording for the --colour (or --color) option.
2366
2367 5. In pcregrep, when --colour was used with -o, the list of matching strings
2368 was not coloured; this is different to GNU grep, so I have changed it to be
2369 the same.
2370
2371 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2372 each matching line was coloured. Now it goes on to look for further matches
2373 of any of the test patterns, which is the same behaviour as GNU grep.
2374
2375 7. A pattern that could match an empty string could cause pcregrep to loop; it
2376 doesn't make sense to accept an empty string match in pcregrep, so I have
2377 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2378 seems to be how GNU grep behaves. [But see later change 40 for release
2379 8.33.]
2380
2381 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2382 start or after a newline", because the conditional assertion was not being
2383 correctly handled. The rule now is that both the assertion and what follows
2384 in the first alternative must satisfy the test.
2385
2386 9. If auto-callout was enabled in a pattern with a conditional group whose
2387 condition was an assertion, PCRE could crash during matching, both with
2388 pcre_exec() and pcre_dfa_exec().
2389
2390 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2391 used for matching.
2392
2393 11. Unicode property support in character classes was not working for
2394 characters (bytes) greater than 127 when not in UTF-8 mode.
2395
2396 12. Added the -M command line option to pcretest.
2397
2398 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2399
2400 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2401
2402 16. Added comments and documentation about mis-use of no_arg in the C++
2403 wrapper.
2404
2405 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2406 from Martin Jerabek that uses macro names for all relevant character and
2407 string constants.
2408
2409 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2410 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2411 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2412 these, but not everybody uses configure.
2413
2414 19. A conditional group that had only one branch was not being correctly
2415 recognized as an item that could match an empty string. This meant that an
2416 enclosing group might also not be so recognized, causing infinite looping
2417 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2418 with the subject "ab", where knowledge that the repeated group can match
2419 nothing is needed in order to break the loop.
2420
2421 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2422 exec(), but without supplying a callout function, matching went wrong.
2423
2424 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2425 leak if the size of the offset vector was greater than 30. When the vector
2426 is smaller, the saved offsets during recursion go onto a local stack
2427 vector, but for larger vectors malloc() is used. It was failing to free
2428 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2429 error, in fact).
2430
2431 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2432 heapframe that is used only when UTF-8 support is enabled. This caused no
2433 problem, but was untidy.
2434
2435 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2436 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2437 included within another project.
2438
2439 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2440 slightly modified by me:
2441
2442 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2443 not building pcregrep.
2444
2445 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2446 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2447
2448 25. Forward references, both numeric and by name, in patterns that made use of
2449 duplicate group numbers, could behave incorrectly or give incorrect errors,
2450 because when scanning forward to find the reference group, PCRE was not
2451 taking into account the duplicate group numbers. A pattern such as
2452 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2453
2454 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2455 the feature of a custom pointer more persuasive (as requested by a user).
2456
2457 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2458 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2459 pre-defined.
2460
2461 28. Added support for (*UTF8) at the start of a pattern.
2462
2463 29. Arrange for flags added by the "release type" setting in CMake to be shown
2464 in the configuration summary.
2465
2466
2467 Version 7.8 05-Sep-08
2468 ---------------------
2469
2470 1. Replaced UCP searching code with optimized version as implemented for Ad
2471 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2472 stage table and inline lookup instead of a function, giving speed ups of 2
2473 to 5 times on some simple patterns that I tested. Permission was given to
2474 distribute the MultiStage2.py script that generates the tables (it's not in
2475 the tarball, but is in the Subversion repository).
2476
2477 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2478 scripts.
2479
2480 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2481 a group with a zero qualifier. The result of the study could be incorrect,
2482 or the function might crash, depending on the pattern.
2483
2484 4. Caseless matching was not working for non-ASCII characters in back
2485 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2486 It now works when Unicode Property Support is available.
2487
2488 5. In pcretest, an escape such as \x{de} in the data was always generating
2489 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2490 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2491 truncation.
2492
2493 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2494
2495 7. Added two (int) casts to pcregrep when printing the difference of two
2496 pointers, in case they are 64-bit values.
2497
2498 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2499 test 2 if it fails.
2500
2501 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2502 and a #define of that name to empty if it is not externally set. This is to
2503 allow users of MSVC to set it if necessary.
2504
2505 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2506 the convenience functions in the pcre_get.c source file.
2507
2508 11. An option change at the start of a pattern that had top-level alternatives
2509 could cause overwriting and/or a crash. This command provoked a crash in
2510 some environments:
2511
2512 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2513
2514 This potential security problem was recorded as CVE-2008-2371.
2515
2516 12. For a pattern where the match had to start at the beginning or immediately
2517 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2518 pcre_dfa_exec() could read past the end of the passed subject if there was
2519 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2520 pcretest so that it places the subject at the end of its malloc-ed buffer.
2521
2522 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2523 exec() might read past the end of the data buffer in UTF-8 mode.
2524
2525 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2526 the data contained the byte 0x85 as part of a UTF-8 character within its
2527 first line. This applied both to normal and DFA matching.
2528
2529 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2530 /^[^d]*?$/8 failed to match "abc".
2531
2532 16. Added a missing copyright notice to pcrecpp_internal.h.
2533
2534 17. Make it more clear in the documentation that values returned from
2535 pcre_exec() in ovector are byte offsets, not character counts.
2536
2537 18. Tidied a few places to stop certain compilers from issuing warnings.
2538
2539 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2540 supplied by Stefan Weber. I made a further small update for 7.8 because
2541 there is a change of source arrangements: the pcre_searchfuncs.c module is
2542 replaced by pcre_ucd.c.
2543
2544
2545 Version 7.7 07-May-08
2546 ---------------------
2547
2548 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2549 a string to a long long, pretend we don't even have a long long." This is
2550 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2551
2552 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2553 pre-7.6 versions, which defined a global no_arg variable instead of putting
2554 it in the RE class. (See also #8 below.)
2555
2556 3. Remove a line of dead code, identified by coverity and reported by Nuno
2557 Lopes.
2558
2559 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2560
2561 (1) The include/exclude patterns were being applied to the whole pathnames
2562 of files, instead of just to the final components.
2563
2564 (2) If there was more than one level of directory, the subdirectories were
2565 skipped unless they satisfied the include/exclude conditions. This is
2566 inconsistent with GNU grep (and could even be seen as contrary to the
2567 pcregrep specification - which I improved to make it absolutely clear).
2568 The action now is always to scan all levels of directory, and just
2569 apply the include/exclude patterns to regular files.
2570
2571 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2572 --exclude_dir in the tests to avoid scanning .svn directories.
2573
2574 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2575 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2576 doesn't support NULs in patterns.
2577
2578 7. Added some missing "const"s to declarations of static tables in
2579 pcre_compile.c and pcre_dfa_exec.c.
2580
2581 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2582 caused by fix #2 above. (Subsequently also a second patch to fix the
2583 first patch. And a third patch - this was a messy problem.)
2584
2585 9. Applied Craig's patch to remove the use of push_back().
2586
2587 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2588 matching function regexec().
2589
2590 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2591 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2592 references. PCRE supports relative numbers with this syntax (I don't think
2593 Oniguruma does).
2594
2595 12. Previously, a group with a zero repeat such as (...){0} was completely
2596 omitted from the compiled regex. However, this means that if the group
2597 was called as a subroutine from elsewhere in the pattern, things went wrong
2598 (an internal error was given). Such groups are now left in the compiled
2599 pattern, with a new opcode that causes them to be skipped at execution
2600 time.
2601
2602 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2603 to the way PCRE behaves:
2604
2605 (a) A lone ] character is dis-allowed (Perl treats it as data).
2606
2607 (b) A back reference to an unmatched subpattern matches an empty string
2608 (Perl fails the current match path).
2609
2610 (c) A data ] in a character class must be notated as \] because if the
2611 first data character in a class is ], it defines an empty class. (In
2612 Perl it is not possible to have an empty class.) The empty class []
2613 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2614 The negative empty class [^] matches any one character, independently
2615 of the DOTALL setting.
2616
2617 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2618 non-existent subpattern following a character class starting with ']' and
2619 containing () gave an internal compiling error instead of "reference to
2620 non-existent subpattern". Fortunately, when the pattern did exist, the
2621 compiled code was correct. (When scanning forwards to check for the
2622 existence of the subpattern, it was treating the data ']' as terminating
2623 the class, so got the count wrong. When actually compiling, the reference
2624 was subsequently set up correctly.)
2625
2626 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2627 it was being rejected as not supported by pcre_dfa_exec(), even though
2628 other assertions are supported. I have made pcre_dfa_exec() support
2629 (*FAIL).
2630
2631 16. The implementation of 13c above involved the invention of a new opcode,
2632 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2633 cannot be changed at match time, I realized I could make a small
2634 improvement to matching performance by compiling OP_ALLANY instead of
2635 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2636 on the OP_ANY path.
2637
2638 17. Compiling pcretest on Windows with readline support failed without the
2639 following two fixes: (1) Make the unistd.h include conditional on
2640 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2641
2642 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2643 ncurses library to be included for pcretest when ReadLine support is
2644 requested, but also to allow for it to be overridden. This patch came from
2645 Daniel Bergström.
2646
2647 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2648 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2649 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2650 spotting this.
2651
2652
2653 Version 7.6 28-Jan-08
2654 ---------------------
2655
2656 1. A character class containing a very large number of characters with
2657 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2658 overflow.
2659
2660 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2661 HAVE_LONG_LONG is not defined.
2662
2663 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2664 bring them up to date and include new features. This patch includes:
2665
2666 - Fixed PH's badly added libz and libbz2 support.
2667 - Fixed a problem with static linking.
2668 - Added pcredemo. [But later removed - see 7 below.]
2669 - Fixed dftables problem and added an option.
2670 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2671 HAVE_LONG_LONG.
2672 - Added readline support for pcretest.
2673 - Added an listing of the option settings after cmake has run.
2674
2675 4. A user submitted a patch to Makefile that makes it easy to create
2676 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2677 Makefile.am that cause it to include this special target, without
2678 affecting anything else. Note that the same mingw target plus all
2679 the other distribution libraries and programs are now supported
2680 when configuring with CMake (see 6 below) instead of with
2681 Configure/Make.
2682
2683 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2684 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2685 exported in the Windows port". It has not yet been confirmed that the patch
2686 solves the problem, but it does no harm.
2687
2688 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2689 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2690 with CMake, and also correct the comment about stack recursion.
2691
2692 7. Remove the automatic building of pcredemo from the ./configure system and
2693 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2694 of a program that users should build themselves after PCRE is installed, so
2695 building it automatically is not really right. What is more, it gave
2696 trouble in some build environments.
2697
2698 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2699
2700
2701 Version 7.5 10-Jan-08
2702 ---------------------
2703
2704 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2705 values in parens when parsing an RE using the C++ wrapper."
2706
2707 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2708 Characters greater than 255 were excluded from the class instead of being
2709 included.
2710
2711 3. The same bug as (2) above applied to negated POSIX classes such as
2712 [:^space:].
2713
2714 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2715 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2716 I have changed it.
2717
2718 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2719 first named subpattern) and a construct such as (?&a) would reference the
2720 first named subpattern whose name started with "a" (in other words, the
2721 length check was missing). Both these problems are fixed. "Subpattern name
2722 expected" is now given for (?&) (a zero-length name), and this patch also
2723 makes it give the same error for \k'' (previously it complained that that
2724 was a reference to a non-existent subpattern).
2725
2726 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2727 this is right because (?- can be followed by option settings as well as by
2728 digits. I have, however, made the messages clearer.
2729
2730 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2731 than the number used in the conditional) now cause a compile-time error.
2732 This is actually not compatible with Perl, which accepts such patterns, but
2733 treats the conditional as always being FALSE (as PCRE used to), but it
2734 seems to me that giving a diagnostic is better.
2735
2736 8. Change "alphameric" to the more common word "alphanumeric" in comments
2737 and messages.
2738
2739 9. Fix two occurrences of "backslash" in comments that should have been
2740 "backspace".
2741
2742 10. Remove two redundant lines of code that can never be obeyed (their function
2743 was moved elsewhere).
2744
2745 11. The program that makes PCRE's Unicode character property table had a bug
2746 which caused it to generate incorrect table entries for sequences of
2747 characters that have the same character type, but are in different scripts.
2748 It amalgamated them into a single range, with the script of the first of
2749 them. In other words, some characters were in the wrong script. There were
2750 thirteen such cases, affecting characters in the following ranges:
2751
2752 U+002b0 - U+002c1
2753 U+0060c - U+0060d
2754 U+0061e - U+00612
2755 U+0064b - U+0065e
2756 U+0074d - U+0076d
2757 U+01800 - U+01805
2758 U+01d00 - U+01d77
2759 U+01d9b - U+01dbf
2760 U+0200b - U+0200f
2761 U+030fc - U+030fe
2762 U+03260 - U+0327f
2763 U+0fb46 - U+0fbb1
2764 U+10450 - U+1049d
2765
2766 12. The -o option (show only the matching part of a line) for pcregrep was not
2767 compatible with GNU grep in that, if there was more than one match in a
2768 line, it showed only the first of them. It now behaves in the same way as
2769 GNU grep.
2770
2771 13. If the -o and -v options were combined for pcregrep, it printed a blank
2772 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2773 does the same. The return code can be used to tell if there were any
2774 non-matching lines.
2775
2776 14. Added --file-offsets and --line-offsets to pcregrep.
2777
2778 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2779 infinitely looping recursion. The bug was that positive lookaheads were not
2780 being skipped when checking for a possible empty match (negative lookaheads
2781 and both kinds of lookbehind were skipped).
2782
2783 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2784 inclusion of <windows.h> to before rather than after the definition of
2785 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2786
2787 17. Specifying a possessive quantifier with a specific limit for a Unicode
2788 character property caused pcre_compile() to compile bad code, which led at
2789 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2790 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2791 caused the error; without that there was no problem.
2792
2793 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2794
2795 19. Added --enable-pcretest-libreadline.
2796
2797 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2798 RE::GlobalReplace(). As a result, the number of replacements returned was
2799 double what it should be. I removed one of the increments, but Craig sent a
2800 later patch that removed the other one (the right fix) and added unit tests
2801 that check the return values (which was not done before).
2802
2803 21. Several CMake things:
2804
2805 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2806 the names libpcre and libpcreposix, not just pcre and pcreposix.
2807
2808 (2) The above change means that pcretest and pcregrep are now correctly
2809 linked with the newly-built libraries, not previously installed ones.
2810
2811 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2812
2813 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2814 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2815 UTF-8 newline character). The key issue is that the pattern starts .*;
2816 this means that the match must be either at the beginning, or after a
2817 newline. The bug was in the code for advancing after a failed match and
2818 checking that the new position followed a newline. It was not taking
2819 account of UTF-8 characters correctly.
2820
2821 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2822 character classes. PCRE was not treating the sequence [:...:] as a
2823 character class unless the ... were all letters. Perl, however, seems to
2824 allow any characters between [: and :], though of course it rejects as
2825 unknown any "names" that contain non-letters, because all the known class
2826 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2827 for example, whereas PCRE did not - it did not recognize a POSIX character
2828 class. This seemed a bit dangerous, so the code has been changed to be
2829 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2830 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2831 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2832 Perl does, and where it didn't before.
2833
2834 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2835 Windows environments %n is disabled by default.
2836
2837
2838 Version 7.4 21-Sep-07
2839 ---------------------
2840
2841 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2842 means that a class such as [\s] counted as "explicit reference to CR or
2843 LF". That isn't really right - the whole point of the change was to try to
2844 help when there was an actual mention of one of the two characters. So now
2845 the change happens only if \r or \n (or a literal CR or LF) character is
2846 encountered.
2847
2848 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2849 of both had grown to the point where there were only 3 bits left.
2850 Fortunately, there was spare space in the data structure, and so I have
2851 moved the internal flags into a new 16-bit field to free up more option
2852 bits.
2853
2854 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2855 but did not set the internal JCHANGED flag - either of these is enough to
2856 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2857 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2858 start sets both bits.
2859
2860 4. Added options (at build time, compile time, exec time) to change \R from
2861 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2862
2863 5. doc/pcresyntax.html was missing from the distribution.
2864
2865 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2866 compatibility, even though it is no longer used.
2867
2868 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2869 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2870 windows.h file is present (where different names are used). [This was
2871 reversed later after testing - see 16 below.]
2872
2873 8. Changed all #include <config.h> to #include "config.h". There were also
2874 some further <pcre.h> cases that I changed to "pcre.h".
2875
2876 9. When pcregrep was used with the --colour option, it missed the line ending
2877 sequence off the lines that it output.
2878
2879 10. It was pointed out to me that arrays of string pointers cause lots of
2880 relocations when a shared library is dynamically loaded. A technique of
2881 using a single long string with a table of offsets can drastically reduce
2882 these. I have refactored PCRE in four places to do this. The result is
2883 dramatic:
2884
2885 Originally: 290
2886 After changing UCP table: 187
2887 After changing error message table: 43
2888 After changing table of "verbs" 36
2889 After changing table of Posix names 22
2890
2891 Thanks to the folks working on Gregex for glib for this insight.
2892
2893 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2894 unicode-properties was also set.
2895
2896 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2897
2898 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2899 checked only for CRLF.
2900
2901 14. Added casts to pcretest.c to avoid compiler warnings.
2902
2903 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2904
2905 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2906 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2907 entirely. This removes changes made in 7 above.
2908
2909 17. The CMake files have been updated, and there is now more information about
2910 building with CMake in the NON-UNIX-USE document.
2911
2912
2913 Version 7.3 28-Aug-07
2914 ---------------------
2915
2916 1. In the rejigging of the build system that eventually resulted in 7.1, the
2917 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2918 brackets there is not right, since it causes compilers to look for an
2919 installed pcre.h, not the version that is in the source that is being
2920 compiled (which of course may be different). I have changed it back to:
2921
2922 #include "pcre.h"
2923
2924 I have a vague recollection that the change was concerned with compiling in
2925 different directories, but in the new build system, that is taken care of
2926 by the VPATH setting the Makefile.
2927
2928 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2929 when the subject happened to end in the byte 0x85 (e.g. if the last
2930 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2931 characters but of course it shouldn't be taken as a newline when it is part
2932 of another character. The bug was that, for an unlimited repeat of . in
2933 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2934 characters when looking for a newline.
2935
2936 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2937
2938 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2939 in debug output.
2940
2941 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2942 long printing in the pcrecpp unittest when running under MinGW.
2943
2944 6. ESC_K was left out of the EBCDIC table.
2945
2946 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2947 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2948 limit also applies to "virtual nesting" when a pattern is recursive, and in
2949 this case 1000 isn't so big. I have been able to remove this limit at the
2950 expense of backing off one optimization in certain circumstances. Normally,
2951 when pcre_exec() would call its internal match() function recursively and
2952 immediately return the result unconditionally, it uses a "tail recursion"
2953 feature to save stack. However, when a subpattern that can match an empty
2954 string has an unlimited repetition quantifier, it no longer makes this
2955 optimization. That gives it a stack frame in which to save the data for
2956 checking that an empty string has been matched. Previously this was taken
2957 from the 1000-entry workspace that had been reserved. So now there is no
2958 explicit limit, but more stack is used.
2959
2960 8. Applied Daniel's patches to solve problems with the import/export magic
2961 syntax that is required for Windows, and which was going wrong for the
2962 pcreposix and pcrecpp parts of the library. These were overlooked when this
2963 problem was solved for the main library.
2964
2965 9. There were some crude static tests to avoid integer overflow when computing
2966 the size of patterns that contain repeated groups with explicit upper
2967 limits. As the maximum quantifier is 65535, the maximum group length was
2968 set at 30,000 so that the product of these two numbers did not overflow a
2969 32-bit integer. However, it turns out that people want to use groups that
2970 are longer than 30,000 bytes (though not repeat them that many times).
2971 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2972 made it possible to implement the integer overflow checks in a much more
2973 dynamic way, which I have now done. The artificial limitation on group
2974 length has been removed - we now have only the limit on the total length of
2975 the compiled pattern, which depends on the LINK_SIZE setting.
2976
2977 10. Fixed a bug in the documentation for get/copy named substring when
2978 duplicate names are permitted. If none of the named substrings are set, the
2979 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2980 empty string.
2981
2982 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2983 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2984 because the ] is interpreted as the first data character and the
2985 terminating ] is not found. PCRE has been made compatible with Perl in this
2986 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2987 cause memory overwriting.
2988
2989 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2990 string has been matched (to stop an infinite loop). It was not recognizing
2991 a conditional subpattern that could match an empty string if that
2992 subpattern was within another subpattern. For example, it looped when
2993 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2994 condition was not nested. This bug has been fixed.
2995
2996 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2997 past the start of the subject in the presence of bytes with the top bit
2998 set, for example "\x8aBCD".
2999
3000 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3001 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3002
3003 14. Optimized (?!) to (*FAIL).
3004
3005 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3006 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3007 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3008 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3009 does: it's just the validity check that is more restrictive.
3010
3011 16. Inserted checks for integer overflows during escape sequence (backslash)
3012 processing, and also fixed erroneous offset values for syntax errors during
3013 backslash processing.
3014
3015 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3016 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3017
3018 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3019 caused an overrun.
3020
3021 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3022 something other than just ASCII characters) inside a group that had an
3023 unlimited repeat caused a loop at compile time (while checking to see
3024 whether the group could match an empty string).
3025
3026 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3027 [\P{Any}] did so. (Error in the code for printing property names.)
3028
3029 21. An orphan \E inside a character class could cause a crash.
3030
3031 22. A repeated capturing bracket such as (A)? could cause a wild memory
3032 reference during compilation.
3033
3034 23. There are several functions in pcre_compile() that scan along a compiled
3035 expression for various reasons (e.g. to see if it's fixed length for look
3036 behind). There were bugs in these functions when a repeated \p or \P was
3037 present in the pattern. These operators have additional parameters compared
3038 with \d, etc, and these were not being taken into account when moving along
3039 the compiled data. Specifically:
3040
3041 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3042 length.
3043
3044 (b) An item such as \pL+ within a repeated group could cause crashes or
3045 loops.
3046
3047 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3048 "reference to non-existent subpattern" error.
3049
3050 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3051
3052 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3053 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3054
3055 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3056
3057 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3058 character were causing crashes (broken optimization).
3059
3060 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3061 \p or \P) caused a compile-time loop.
3062
3063 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3064 break. For example, the unstudied pattern [\r\n]A does not match the string
3065 "\r\nA" because change 7.0/46 below moves the current point on by two
3066 characters after failing to match at the start. However, the pattern \nA
3067 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3068 the same is true. There doesn't seem any very clean way out of this, but
3069 what I have chosen to do makes the common cases work: PCRE now takes note
3070 of whether there can be an explicit match for \r or \n anywhere in the
3071 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3072 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3073 pattern has explicit CR or LF references.
3074
3075 29. Added (*CR) etc for changing newline setting at start of pattern.
3076
3077
3078 Version 7.2 19-Jun-07
3079 ---------------------
3080
3081 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3082 which is apparently normally available under Windows.
3083
3084 2. Re-jig the pcregrep tests with different newline settings in an attempt
3085 to make them independent of the local environment's newline setting.
3086
3087 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3088
3089 4. Some of the "internals" tests were previously cut out when the link size
3090 was not 2, because the output contained actual offsets. The recent new
3091 "Z" feature of pcretest means that these can be cut out, making the tests
3092 usable with all link sizes.
3093
3094 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3095 stack recursion. This gives a massive performance boost under BSD, but just
3096 a small improvement under Linux. However, it saves one field in the frame
3097 in all cases.
3098
3099 6. Added more features from the forthcoming Perl 5.10:
3100
3101 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3102 recursion call. It refers to the nth most recently opened parentheses.
3103
3104 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3105 to be opened parentheses.
3106
3107 (c) Conditions that refer to capturing parentheses can be specified
3108 relatively, for example, (?(-2)... or (?(+3)...
3109
3110 (d) \K resets the start of the current match so that everything before
3111 is not part of it.
3112
3113 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3114
3115 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3116 reference syntax.
3117
3118 (g) (?| introduces a group in which the numbering of parentheses in each
3119 alternative starts with the same number.
3120
3121 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3122
3123 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3124 PCRE_INFO_JCHANGED.
3125
3126 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3127 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3128 for detecting groups that can match an empty string.
3129
3130 9. A pattern with a very large number of alternatives (more than several
3131 hundred) was running out of internal workspace during the pre-compile
3132 phase, where pcre_compile() figures out how much memory will be needed. A
3133 bit of new cunning has reduced the workspace needed for groups with
3134 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3135 workspace instead of running out of the 4096 that are available.
3136
3137 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3138
3139 11. Applied patch from Google to remove an optimization that didn't quite work.
3140 The report of the bug said:
3141
3142 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3143 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3144 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3145
3146 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3147 it matched the wrong number of bytes.
3148
3149
3150 Version 7.1 24-Apr-07
3151 ---------------------
3152
3153 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3154 that is more "standard", making use of automake and other Autotools. There
3155 is some re-arrangement of the files and adjustment of comments consequent
3156 on this.
3157
3158 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3159 for recursive directory scanning broke on some systems because the files
3160 are not scanned in any specific order and on different systems the order
3161 was different. A call to "sort" has been inserted into RunGrepTest for the
3162 approprate test as a short-term fix. In the longer term there may be an
3163 alternative.
3164
3165 3. I had an email from Eric Raymond about problems translating some of PCRE's
3166 man pages to HTML (despite the fact that I distribute HTML pages, some
3167 people do their own conversions for various reasons). The problems
3168 concerned the use of low-level troff macros .br and .in. I have therefore
3169 removed all such uses from the man pages (some were redundant, some could
3170 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3171 HTML has been updated to handle .nf/.fi and to complain if it encounters
3172 .br or .in.
3173
3174 4. Updated comments in configure.ac that get placed in config.h.in and also
3175 arranged for config.h to be included in the distribution, with the name
3176 config.h.generic, for the benefit of those who have to compile without
3177 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3178
3179 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3180 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3181 makevp.bat for the current PCRE, using the additional files
3182 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3183
3184 6. A Windows user reported a minor discrepancy with test 2, which turned out
3185 to be caused by a trailing space on an input line that had got lost in his
3186 copy. The trailing space was an accident, so I've just removed it.
3187
3188 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3189 that is needed.
3190
3191 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3192 as "const" (a) because they are and (b) because it helps the PHP
3193 maintainers who have recently made a script to detect big data structures
3194 in the php code that should be moved to the .rodata section. I remembered
3195 to update Builducptable as well, so it won't revert if ucptable.h is ever
3196 re-created.
3197
3198 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3199 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3200 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3201 support is not required. This saves 1.5-2K of code, which is important in
3202 some applications.
3203
3204 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3205 so as not to refer to the tables, even though these functions will never be
3206 called when UTF-8 support is disabled. Otherwise there are problems with a
3207 shared library.
3208
3209 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3210
3211 (a) It was defining its arguments as char * instead of void *.
3212
3213 (b) It was assuming that all moves were upwards in memory; this was true
3214 a long time ago when I wrote it, but is no longer the case.
3215
3216 The emulated memove() is provided for those environments that have neither
3217 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3218 is clearly not the case, as these two bugs were recently reported.
3219
3220 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3221 and Detrail to create the HTML documentation, the .txt form of the man
3222 pages, and it removes trailing spaces from listed files. It also creates
3223 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3224 case, it wraps all the #defines with #ifndefs. This script should be run
3225 before "make dist".
3226
3227 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3228 with Unicode property support.
3229
3230 (a) For a maximizing quantifier, if the two different cases of the
3231 character were of different lengths in their UTF-8 codings (there are
3232 some cases like this - I found 11), and the matching function had to
3233 back up over a mixture of the two cases, it incorrectly assumed they
3234 were both the same length.
3235
3236 (b) When PCRE was configured to use the heap rather than the stack for
3237 recursion during matching, it was not correctly preserving the data for
3238 the other case of a UTF-8 character when checking ahead for a match
3239 while processing a minimizing repeat. If the check also involved
3240 matching a wide character, but failed, corruption could cause an
3241 erroneous result when trying to check for a repeat of the original
3242 character.
3243
3244 13. Some tidying changes to the testing mechanism:
3245
3246 (a) The RunTest script now detects the internal link size and whether there
3247 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3248 values substituted by "configure". (The RunGrepTest script already did
3249 this for UTF-8.) The configure.ac script no longer substitutes the
3250 relevant variables.
3251
3252 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3253 with length and offset values. This means that the output is different
3254 for different internal link sizes. Test 2 is skipped for link sizes
3255 other than 2 because of this, bypassing the problem. Unfortunately,
3256 there was also a test in test 3 (the locale tests) that used /B and
3257 failed for link sizes other than 2. Rather than cut the whole test out,
3258 I have added a new /Z option to pcretest that replaces the length and
3259 offset values with spaces. This is now used to make test 3 independent
3260 of link size. (Test 2 will be tidied up later.)
3261
3262 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3263 segmentation fault instead of returning the appropriate error message.
3264
3265 15. In multiline mode when the newline sequence was set to "any", the pattern
3266 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3267 This doesn't seem right; it now treats the CRLF combination as the line
3268 ending, and so does not match in that case. It's only a pattern such as ^$
3269 that would hit this one: something like ^ABC$ would have failed after \r
3270 and then tried again after \r\n.
3271
3272 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3273 in an attempt to make files that differ only in their line terminators
3274 compare equal. This works on Linux.
3275
3276 17. Under certain error circumstances pcregrep might try to free random memory
3277 as it exited. This is now fixed, thanks to valgrind.
3278
3279 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3280 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3281 was because its rules for how to advance for /g after matching an empty
3282 string at the end of a line did not allow for this case. They now check for
3283 it specially.
3284
3285 20. pcretest is supposed to handle patterns and data of any length, by
3286 extending its buffers when necessary. It was getting this wrong when the
3287 buffer for a data line had to be extended.
3288
3289 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3290 CRLF as a newline sequence.
3291
3292 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3293 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3294 I have nevertheless tidied it up.
3295
3296 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3297
3298 24. Added a man page for pcre-config.
3299
3300
3301 Version 7.0 19-Dec-06
3302 ---------------------
3303
3304 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3305 moving to gcc 4.1.1.
3306
3307 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3308 sys/time.h, which is documented as needed for this function. It doesn't
3309 seem to matter on Linux, but it showed up on some releases of OS X.
3310
3311 3. It seems that there are systems where bytes whose values are greater than
3312 127 match isprint() in the "C" locale. The "C" locale should be the
3313 default when a C program starts up. In most systems, only ASCII printing
3314 characters match isprint(). This difference caused the output from pcretest
3315 to vary, making some of the tests fail. I have changed pcretest so that:
3316
3317 (a) When it is outputting text in the compiled version of a pattern, bytes
3318 other than 32-126 are always shown as hex escapes.
3319
3320 (b) When it is outputting text that is a matched part of a subject string,
3321 it does the same, unless a different locale has been set for the match
3322 (using the /L modifier). In this case, it uses isprint() to decide.
3323
3324 4. Fixed a major bug that caused incorrect computation of the amount of memory
3325 required for a compiled pattern when options that changed within the
3326 pattern affected the logic of the preliminary scan that determines the
3327 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3328 that the computed length was too small. The symptoms of this bug were
3329 either the PCRE error "internal error: code overflow" from pcre_compile(),
3330 or a glibc crash with a message such as "pcretest: free(): invalid next
3331 size (fast)". Examples of patterns that provoked this bug (shown in
3332 pcretest format) are:
3333
3334 /(?-x: )/x
3335 /(?x)(?-x: \s*#\s*)/
3336 /((?i)[\x{c0}])/8
3337 /(?i:[\x{c0}])/8
3338
3339 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3340 is now done differently.
3341
3342 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3343 wrapper classes; (b) implement a new function in the C++ scanner that is
3344 more efficient than the old way of doing things because it avoids levels of
3345 recursion in the regex matching; (c) add a paragraph to the documentation
3346 for the FullMatch() function.
3347
3348 6. The escape sequence \n was being treated as whatever was defined as
3349 "newline". Not only was this contrary to the documentation, which states
3350 that \n is character 10 (hex 0A), but it also went horribly wrong when
3351 "newline" was defined as CRLF. This has been fixed.
3352
3353 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3354 was being set to -1 for the "end of line" case (supposedly a value that no
3355 character can have). Though this value is never used (the check for end of
3356 line is "zero bytes in current character"), it caused compiler complaints.
3357 I've changed it to 0xffffffff.
3358
3359 8. In pcre_version.c, the version string was being built by a sequence of
3360 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3361 string (as it is for production releases) called a macro with an empty
3362 argument. The C standard says the result of this is undefined. The gcc
3363 compiler treats it as an empty string (which was what was wanted) but it is
3364 reported that Visual C gives an error. The source has been hacked around to
3365 avoid this problem.
3366
3367 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3368 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3369 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3370 of them did).
3371
3372 10. Originally, pcretest opened its input and output without "b"; then I was
3373 told that "b" was needed in some environments, so it was added for release
3374 5.0 to both the input and output. (It makes no difference on Unix-like
3375 systems.) Later I was told that it is wrong for the input on Windows. I've
3376 now abstracted the modes into two macros, to make it easier to fiddle with
3377 them, and removed "b" from the input mode under Windows.
3378
3379 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3380
3381 12. Added -help and --help to pcretest as an official way of being reminded
3382 of the options.
3383
3384 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3385 and pcrecpp.cc because they annoy compilers at high warning levels.
3386
3387 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3388
3389 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3390 scripts are not C programs :-) and which was not noticed because it works
3391 on Linux.
3392
3393 16. pcretest is supposed to handle any length of pattern and data line (as one
3394 line or as a continued sequence of lines) by extending its input buffer if
3395 necessary. This feature was broken for very long pattern lines, leading to
3396 a string of junk being passed to pcre_compile() if the pattern was longer
3397 than about 50K.
3398
3399 17. I have done a major re-factoring of the way pcre_compile() computes the
3400 amount of memory needed for a compiled pattern. Previously, there was code
3401 that made a preliminary scan of the pattern in order to do this. That was
3402 OK when PCRE was new, but as the facilities have expanded, it has become
3403 harder and harder to keep it in step with the real compile phase, and there
3404 have been a number of bugs (see for example, 4 above). I have now found a
3405 cunning way of running the real compile function in a "fake" mode that
3406 enables it to compute how much memory it would need, while actually only
3407 ever using a few hundred bytes of working memory and without too many
3408 tests of the mode. This should make future maintenance and development
3409 easier. A side effect of this work is that the limit of 200 on the nesting
3410 depth of parentheses has been removed (though this was never a serious
3411 limitation, I suspect). However, there is a downside: pcre_compile() now
3412 runs more slowly than before (30% or more, depending on the pattern). I
3413 hope this isn't a big issue. There is no effect on runtime performance.
3414
3415 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3416 newline (only possible for the last line of a file) and it was a
3417 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3418
3419 19. Added additional timing features to pcretest. (1) The -tm option now times
3420 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3421 separate command line item, by a number that specifies the number of
3422 repeats to use when timing. The default is 50000; this gives better
3423 precision, but takes uncomfortably long for very large patterns.
3424
3425 20. Extended pcre_study() to be more clever in cases where a branch of a
3426 subpattern has no definite first character. For example, (a*|b*)[cd] would
3427 previously give no result from pcre_study(). Now it recognizes that the
3428 first character must be a, b, c, or d.
3429
3430 21. There was an incorrect error "recursive call could loop indefinitely" if
3431 a subpattern (or the entire pattern) that was being tested for matching an
3432 empty string contained only one non-empty item after a nested subpattern.
3433 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3434 incorrectly, because the \d was being skipped in the check.
3435
3436 22. The pcretest program now has a new pattern option /B and a command line
3437 option -b, which is equivalent to adding /B to every pattern. This causes
3438 it to show the compiled bytecode, without the additional information that
3439 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3440 is the same as /B/I).
3441
3442 23. A new optimization is now able automatically to treat some sequences such
3443 as a*b as a*+b. More specifically, if something simple (such as a character
3444 or a simple class like \d) has an unlimited quantifier, and is followed by
3445 something that cannot possibly match the quantified thing, the quantifier
3446 is automatically "possessified".
3447
3448 24. A recursive reference to a subpattern whose number was greater than 39
3449 went wrong under certain circumstances in UTF-8 mode. This bug could also
3450 have affected the operation of pcre_study().
3451
3452 25. Realized that a little bit of performance could be had by replacing
3453 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3454
3455 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3456
3457 27. Possessive quantifiers such as a++ were previously implemented by turning
3458 them into atomic groups such as ($>a+). Now they have their own opcodes,
3459 which improves performance. This includes the automatically created ones
3460 from 23 above.
3461
3462 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3463 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3464 the first matched character to be a colon. This applied both to named and
3465 numbered groups.
3466
3467 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3468
3469 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3470 building PCRE on Windows easier, so I have included it in the distribution.
3471
3472 31. There is now a check in pcretest against a ridiculously large number being
3473 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3474 loop, the loop is abandoned.
3475
3476 32. Forward references to subpatterns in conditions such as (?(2)...) where
3477 subpattern 2 is defined later cause pcre_compile() to search forwards in
3478 the pattern for the relevant set of parentheses. This search went wrong
3479 when there were unescaped parentheses in a character class, parentheses
3480 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3481
3482 33. "Subroutine" calls and backreferences were previously restricted to
3483 referencing subpatterns earlier in the regex. This restriction has now
3484 been removed.
3485
3486 34. Added a number of extra features that are going to be in Perl 5.10. On the
3487 whole, these are just syntactic alternatives for features that PCRE had
3488 previously implemented using the Python syntax or my own invention. The
3489 other formats are all retained for compatibility.
3490
3491 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3492 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3493 also .NET compatible.
3494
3495 (b) A recursion or subroutine call to a named group can now be defined as
3496 (?&name) as well as (?P>name).
3497
3498 (c) A backreference to a named group can now be defined as \k<name> or
3499 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3500 5.10, are also .NET compatible.
3501
3502 (d) A conditional reference to a named group can now use the syntax
3503 (?(<name>) or (?('name') as well as (?(name).
3504
3505 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3506 groups (named and numbered) that are never evaluated inline, but can be
3507 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3508 is always false. There may be only one alternative in such a group.
3509
3510 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3511 as the simple (?(R). The condition is true only if the most recent
3512 recursion is that of the given number or name. It does not search out
3513 through the entire recursion stack.
3514
3515 (g) The escape \gN or \g{N} has been added, where N is a positive or
3516 negative number, specifying an absolute or relative reference.
3517
3518 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3519 some "unreachable code" warnings.
3520
3521 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3522 things, this adds five new scripts.
3523
3524 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3525 There were also incompatibilities regarding the handling of \Q..\E inside
3526 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3527 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3528
3529 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3530 matches an empty string, and forcibly breaks the loop. There were bugs in
3531 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3532 against aaaa the result was just "a" rather than "aaaa", for example. Two
3533 separate and independent bugs (that affected different cases) have been
3534 fixed.
3535
3536 39. Refactored the code to abolish the use of different opcodes for small
3537 capturing bracket numbers. This is a tidy that I avoided doing when I
3538 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3539 The new approach is not only tidier, it makes it possible to reduce the
3540 memory needed to fix the previous bug (38).
3541
3542 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3543 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3544 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3545 mode.
3546
3547 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3548 report.
3549
3550 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3551 copy construction and assignment in the C++ wrapper.
3552
3553 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3554 couple of bugs that could have given wrong results in the "--newline=crlf"
3555 case.
3556
3557 44. Added a number of casts and did some reorganization of signed/unsigned int
3558 variables following suggestions from Dair Grant. Also renamed the variable
3559 "this" as "item" because it is a C++ keyword.
3560
3561 45. Arranged for dftables to add
3562
3563 #include "pcre_internal.h"
3564
3565 to pcre_chartables.c because without it, gcc 4.x may remove the array
3566 definition from the final binary if PCRE is built into a static library and
3567 dead code stripping is activated.
3568
3569 46. For an unanchored pattern, if a match attempt fails at the start of a
3570 newline sequence, and the newline setting is CRLF or ANY, and the next two
3571 characters are CRLF, advance by two characters instead of one.
3572
3573
3574 Version 6.7 04-Jul-06
3575 ---------------------
3576
3577 1. In order to handle tests when input lines are enormously long, pcretest has
3578 been re-factored so that it automatically extends its buffers when
3579 necessary. The code is crude, but this _is_ just a test program. The
3580 default size has been increased from 32K to 50K.
3581
3582 2. The code in pcre_study() was using the value of the re argument before
3583 testing it for NULL. (Of course, in any sensible call of the function, it
3584 won't be NULL.)
3585
3586 3. The memmove() emulation function in pcre_internal.h, which is used on
3587 systems that lack both memmove() and bcopy() - that is, hardly ever -
3588 was missing a "static" storage class specifier.
3589
3590 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3591 containing an extended class (one that cannot be represented by a bitmap
3592 because it contains high-valued characters or Unicode property items, e.g.
3593 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3594 pattern, but PCRE should not loop if you do not (it no longer does).
3595 [Detail: two cases were found: (a) a repeated subpattern containing an
3596 extended class; (b) a recursive reference to a subpattern that followed a
3597 previous extended class. It wasn't skipping over the extended class
3598 correctly when UTF-8 mode was not set.]
3599
3600 5. A negated single-character class was not being recognized as fixed-length
3601 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3602 compile error "lookbehind assertion is not fixed length".
3603
3604 6. The RunPerlTest auxiliary script was showing an unexpected difference
3605 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3606 write a Perl script that can interpret lines of an input file either as
3607 byte characters or as UTF-8, which is what "perltest" was being required to
3608 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3609 can't do is switch easily at run time between having the "use utf8;" pragma
3610 or not. In the end, I fudged it by using the RunPerlTest script to insert
3611 "use utf8;" explicitly for the UTF-8 tests.
3612
3613 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3614 the end of the subject string, contrary to the documentation and to what
3615 Perl does. This was true of both matching functions. Now it matches only at
3616 the start of the subject and immediately after *internal* newlines.
3617
3618 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3619 a pointer to an int instead of a pointer to an unsigned long int. This
3620 caused problems on 64-bit systems.
3621
3622 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3623 instance of the 'standard' template library not being so standard".
3624
3625 10. There was no check on the number of named subpatterns nor the maximum
3626 length of a subpattern name. The product of these values is used to compute
3627 the size of the memory block for a compiled pattern. By supplying a very
3628 long subpattern name and a large number of named subpatterns, the size
3629 computation could be caused to overflow. This is now prevented by limiting
3630 the length of names to 32 characters, and the number of named subpatterns
3631 to 10,000.
3632
3633 11. Subpatterns that are repeated with specific counts have to be replicated in
3634 the compiled pattern. The size of memory for this was computed from the
3635 length of the subpattern and the repeat count. The latter is limited to
3636 65535, but there was no limit on the former, meaning that integer overflow
3637 could in principle occur. The compiled length of a repeated subpattern is
3638 now limited to 30,000 bytes in order to prevent this.
3639
3640 12. Added the optional facility to have named substrings with the same name.
3641
3642 13. Added the ability to use a named substring as a condition, using the
3643 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3644 are numbers (not recommended). Forward references are permitted.
3645
3646 14. Added forward references in named backreferences (if you see what I mean).
3647
3648 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3649 pattern could run off the end of the subject. For example, the pattern
3650 "(?s)(.{1,5})"8 did this with the subject "ab".
3651
3652 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3653 PCRE_CASELESS was set when matching characters that were quantified with ?
3654 or *.
3655
3656 17. A character class other than a single negated character that had a minimum
3657 but no maximum quantifier - for example [ab]{6,} - was not handled
3658 correctly by pce_dfa_exec(). It would match only one character.
3659
3660 18. A valid (though odd) pattern that looked like a POSIX character
3661 class but used an invalid character after [ (for example [[,abc,]]) caused
3662 pcre_compile() to give the error "Failed: internal error: code overflow" or
3663 in some cases to crash with a glibc free() error. This could even happen if
3664 the pattern terminated after [[ but there just happened to be a sequence of
3665 letters, a binary zero, and a closing ] in the memory that followed.
3666
3667 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3668 over the years. Originally (before any Unicode support), just the bottom 8
3669 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3670 output from "man perlunicode" includes this:
3671
3672 The regular expression compiler produces polymorphic opcodes. That
3673 is, the pattern adapts to the data and automatically switches to
3674 the Unicode character scheme when presented with Unicode data--or
3675 instead uses a traditional byte scheme when presented with byte
3676 data.
3677
3678 Sadly, a wide octal escape does not cause a switch, and in a string with
3679 no other multibyte characters, these octal escapes are treated as before.
3680 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3681 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3682 Unicode string.
3683
3684 I have not perpetrated such confusion in PCRE. Up till now, it took just
3685 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3686 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3687 translate to the appropriate multibyte character.
3688
3689 29. Applied some refactoring to reduce the number of warnings from Microsoft
3690 and Borland compilers. This has included removing the fudge introduced
3691 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3692 a warning about an unused variable.
3693
3694 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3695 characters since release 4.0, because Perl (from release 5.004) does not.
3696 [Or at least, is documented not to: some releases seem to be in conflict
3697 with the documentation.] However, when a pattern was studied with
3698 pcre_study() and all its branches started with \s, PCRE still included VT
3699 as a possible starting character. Of course, this did no harm; it just
3700 caused an unnecessary match attempt.
3701
3702 22. Removed a now-redundant internal flag bit that recorded the fact that case
3703 dependency changed within the pattern. This was once needed for "required
3704 byte" processing, but is no longer used. This recovers a now-scarce options
3705 bit. Also moved the least significant internal flag bit to the most-
3706 significant bit of the word, which was not previously used (hangover from
3707 the days when it was an int rather than a uint) to free up another bit for
3708 the future.
3709
3710 23. Added support for CRLF line endings as well as CR and LF. As well as the
3711 default being selectable at build time, it can now be changed at runtime
3712 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3713 specify that it is scanning data with non-default line endings.
3714
3715 24. Changed the definition of CXXLINK to make it agree with the definition of
3716 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3717
3718 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3719 recursions. This makes a big different to stack usage for some patterns.
3720
3721 26. If a subpattern containing a named recursion or subroutine reference such
3722 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3723 the space required for the compiled pattern went wrong and gave too small a
3724 value. Depending on the environment, this could lead to "Failed: internal
3725 error: code overflow at offset 49" or "glibc detected double free or
3726 corruption" errors.
3727
3728 27. Applied patches from Google (a) to support the new newline modes and (b) to
3729 advance over multibyte UTF-8 characters in GlobalReplace.
3730
3731 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3732 difference for some implementation of PCRE in some Windows version.
3733
3734 29. Added some extra testing facilities to pcretest:
3735
3736 \q<number> in a data line sets the "match limit" value
3737 \Q<number> in a data line sets the "match recursion limt" value
3738 -S <number> sets the stack size, where <number> is in megabytes
3739
3740 The -S option isn't available for Windows.
3741
3742
3743 Version 6.6 06-Feb-06
3744 ---------------------
3745
3746 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3747 in pcreposix.h. I have copied the definition from pcre.h.
3748
3749 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3750 because pcre.h is no longer a built file.
3751
3752 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3753 not normally included in the compiled code.
3754
3755
3756 Version 6.5 01-Feb-06
3757 ---------------------
3758
3759 1. When using the partial match feature with pcre_dfa_exec(), it was not
3760 anchoring the second and subsequent partial matches at the new starting
3761 point. This could lead to incorrect results. For example, with the pattern
3762 /1234/, partially matching against "123" and then "a4" gave a match.
3763
3764 2. Changes to pcregrep:
3765
3766 (a) All non-match returns from pcre_exec() were being treated as failures
3767 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3768 error message is output. Some extra information is given for the
3769 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3770 probably the only errors that are likely to be caused by users (by
3771 specifying a regex that has nested indefinite repeats, for instance).
3772 If there are more than 20 of these errors, pcregrep is abandoned.
3773
3774 (b) A binary zero was treated as data while matching, but terminated the
3775 output line if it was written out. This has been fixed: binary zeroes
3776 are now no different to any other data bytes.
3777
3778 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3779 used to set a locale for matching. The --locale=xxxx long option has
3780 been added (no short equivalent) to specify a locale explicitly on the
3781 pcregrep command, overriding the environment variables.
3782
3783 (d) When -B was used with -n, some line numbers in the output were one less
3784 than they should have been.
3785
3786 (e) Added the -o (--only-matching) option.
3787
3788 (f) If -A or -C was used with -c (count only), some lines of context were
3789 accidentally printed for the final match.
3790
3791 (g) Added the -H (--with-filename) option.
3792
3793 (h) The combination of options -rh failed to suppress file names for files
3794 that were found from directory arguments.
3795
3796 (i) Added the -D (--devices) and -d (--directories) options.
3797
3798 (j) Added the -F (--fixed-strings) option.
3799
3800 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3801
3802 (l) Added the --colo(u)r option.
3803
3804 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3805 is not present by default.
3806
3807 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3808 items such as (?R) or (?1), when the recursion could match a number of
3809 alternatives. If it matched one of the alternatives, but subsequently,
3810 outside the recursion, there was a failure, the code tried to back up into
3811 the recursion. However, because of the way PCRE is implemented, this is not
3812 possible, and the result was an incorrect result from the match.
3813
3814 In order to prevent this happening, the specification of recursion has
3815 been changed so that all such subpatterns are automatically treated as
3816 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3817
3818 4. I had overlooked the fact that, in some locales, there are characters for
3819 which isalpha() is true but neither isupper() nor islower() are true. In
3820 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3821 and ordfeminine) are like this. This affected the treatment of \w and \W
3822 when they appeared in character classes, but not when they appeared outside
3823 a character class. The bit map for "word" characters is now created
3824 separately from the results of isalnum() instead of just taking it from the
3825 upper, lower, and digit maps. (Plus the underscore character, of course.)
3826
3827 5. The above bug also affected the handling of POSIX character classes such as
3828 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3829 permanent tables. Instead, the bit maps for such a class were previously
3830 created as the appropriate unions of the upper, lower, and digit bitmaps.
3831 Now they are created by subtraction from the [[:word:]] class, which has
3832 its own bitmap.
3833
3834 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3835 It is created by subtracting the vertical space characters (\x09, \x0a,
3836 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3837 subtraction was done in the overall bitmap for a character class, meaning
3838 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3839 be recognized. This bug has been fixed.
3840
3841 7. Patches from the folks at Google:
3842
3843 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3844 real life, but is still worth protecting against".
3845
3846 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3847 regular expressions".
3848
3849 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3850 have it.
3851
3852 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3853 "configure" and the latter not, in order to fix a problem somebody had
3854 with compiling the Arg class on HP-UX.
3855
3856 (e) Improve the error-handling of the C++ wrapper a little bit.
3857
3858 (f) New tests for checking recursion limiting.
3859
3860 8. The pcre_memmove() function, which is used only if the environment does not
3861 have a standard memmove() function (and is therefore rarely compiled),
3862 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3863 returning a result (though PCRE never actually uses the result).
3864
3865 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3866 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3867 returned instead of calling malloc() with an overflowing number that would
3868 most likely cause subsequent chaos.
3869
3870 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3871
3872 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3873 with this option is matched, the nmatch and pmatch options of regexec() are
3874 ignored.
3875
3876 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3877 provided in case anyone wants to the the POSIX interface with UTF-8
3878 strings.
3879
3880 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3881 C++ linking (needed for some HP-UX environments).
3882
3883 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3884 (unused parameter) and in the pcre_printint() function (omitted "default"
3885 switch label when the default is to do nothing).
3886
3887 15. Added some code to make it possible, when PCRE is compiled as a C++
3888 library, to replace subject pointers for pcre_exec() with a smart pointer
3889 class, thus making it possible to process discontinuous strings.
3890
3891 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3892 much the same function. They were added by different people who were trying
3893 to make PCRE easy to compile on non-Unix systems. It has been suggested
3894 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3895 for compiling on Windows systems. I have therefore replaced it with
3896 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3897 defaults to "extern" for C or "extern C" for C++, which works fine on
3898 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3899 SCOPE with something explicit in config.h. In addition:
3900
3901 (a) pcreposix.h still had just "extern" instead of either of these macros;
3902 I have replaced it with PCRE_DATA_SCOPE.
3903
3904 (b) Functions such as _pcre_xclass(), which are internal to the library,
3905 but external in the C sense, all had PCRE_EXPORT in their definitions.
3906 This is apparently wrong for the Windows case, so I have removed it.
3907 (It makes no difference on Unix-like systems.)
3908
3909 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3910 of recursive calls to match(). This is different to MATCH_LIMIT because
3911 that limits the total number of calls to match(), not all of which increase
3912 the depth of recursion. Limiting the recursion depth limits the amount of
3913 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3914 when PCRE is compiled, and changed at run time. A patch from Google adds
3915 this functionality to the C++ interface.
3916
3917 18. Changes to the handling of Unicode character properties:
3918
3919 (a) Updated the table to Unicode 4.1.0.
3920
3921 (b) Recognize characters that are not in the table as "Cn" (undefined).
3922
3923 (c) I revised the way the table is implemented to a much improved format
3924 which includes recognition of ranges. It now supports the ranges that
3925 are defined in UnicodeData.txt, and it also amalgamates other
3926 characters into ranges. This has reduced the number of entries in the
3927 table from around 16,000 to around 3,000, thus reducing its size
3928 considerably. I realized I did not need to use a tree structure after
3929 all - a binary chop search is just as efficient. Having reduced the
3930 number of entries, I extended their size from 6 bytes to 8 bytes to
3931 allow for more data.
3932
3933 (d) Added support for Unicode script names via properties such as \p{Han}.
3934
3935 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3936 matching that character.
3937
3938 20. When matching a repeated Unicode property with a minimum greater than zero,
3939 (for example \pL{2,}), PCRE could look past the end of the subject if it
3940 reached it while seeking the minimum number of characters. This could
3941 happen only if some of the characters were more than one byte long, because
3942 there is a check for at least the minimum number of bytes.
3943
3944 21. Refactored the implementation of \p and \P so as to be more general, to
3945 allow for more different types of property in future. This has changed the
3946 compiled form incompatibly. Anybody with saved compiled patterns that use
3947 \p or \P will have to recompile them.
3948
3949 22. Added "Any" and "L&" to the supported property types.
3950
3951 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3952 but give a compile time error if the value is greater than 0xff.
3953
3954 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3955 accidentally not being installed or uninstalled.
3956
3957 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3958 made were to insert the current release number. This seemed silly, because
3959 it made things harder for people building PCRE on systems that don't run
3960 "configure". I have turned pcre.h into a distributed file, no longer built
3961 by "configure", with the version identification directly included. There is
3962 no longer a pcre.h.in file.
3963
3964 However, this change necessitated a change to the pcre-config script as
3965 well. It is built from pcre-config.in, and one of the substitutions was the
3966 release number. I have updated configure.ac so that ./configure now finds
3967 the release number by grepping pcre.h.
3968
3969 26. Added the ability to run the tests under valgrind.
3970
3971
3972 Version 6.4 05-Sep-05
3973 ---------------------
3974
3975 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3976 "--" to be printed when multiple files were scanned, even when none of the
3977 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3978 consider it to be a bug, and have restored the previous behaviour.
3979
3980 2. A couple of code tidies to get rid of compiler warnings.
3981
3982 3. The pcretest program used to cheat by referring to symbols in the library
3983 whose names begin with _pcre_. These are internal symbols that are not
3984 really supposed to be visible externally, and in some environments it is
3985 possible to suppress them. The cheating is now confined to including
3986 certain files from the library's source, which is a bit cleaner.
3987
3988 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3989 file's purpose clearer.
3990
3991 5. Reorganized pcre_ucp_findchar().
3992
3993
3994 Version 6.3 15-Aug-05
3995 ---------------------
3996
3997 1. The file libpcre.pc.in did not have general read permission in the tarball.
3998
3999 2. There were some problems when building without C++ support:
4000
4001 (a) If C++ support was not built, "make install" and "make test" still
4002 tried to test it.
4003
4004 (b) There were problems when the value of CXX was explicitly set. Some
4005 changes have been made to try to fix these, and ...
4006
4007 (c) --disable-cpp can now be used to explicitly disable C++ support.
4008
4009 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4010 backslash in a target when C++ was disabled. This confuses some
4011 versions of "make", apparently. Using an intermediate variable solves
4012 this. (Same for CPP_LOBJ.)
4013
4014 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4015 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4016 necessary on certain architectures.
4017
4018 4. Added a setting of -export-symbols-regex to the link command to remove
4019 those symbols that are exported in the C sense, but actually are local
4020 within the library, and not documented. Their names all begin with
4021 "_pcre_". This is not a perfect job, because (a) we have to except some
4022 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4023 available (and never for static libraries). I have made a note to try to
4024 find a way round (a) in the future.
4025
4026
4027 Version 6.2 01-Aug-05
4028 ---------------------
4029
4030 1. There was no test for integer overflow of quantifier values. A construction
4031 such as {1111111111111111} would give undefined results. What is worse, if
4032 a minimum quantifier for a parenthesized subpattern overflowed and became
4033 negative, the calculation of the memory size went wrong. This could have
4034 led to memory overwriting.
4035
4036 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4037
4038 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4039 operating environments where this matters.
4040
4041 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4042 PCRE options from within the C++ wrapper.
4043
4044 5. Named capturing subpatterns were not being correctly counted when a pattern
4045 was compiled. This caused two problems: (a) If there were more than 100
4046 such subpatterns, the calculation of the memory needed for the whole
4047 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4048 back references of the form \12, where the number was greater than 9, were
4049 not recognized as back references, even though there were sufficient
4050 previous subpatterns.
4051
4052 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4053 versions of gcc, e.g. 2.95.4.
4054
4055
4056 Version 6.1 21-Jun-05
4057 ---------------------
4058
4059 1. There was one reference to the variable "posix" in pcretest.c that was not
4060 surrounded by "#if !defined NOPOSIX".
4061
4062 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4063 the cross-check on the old pcre_info() function, for the benefit of the
4064 cut-down version of PCRE that is currently imported into Exim.
4065
4066 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4067 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4068 patterns that start (?i) but I don't think that matters. The use of (?i) is
4069 just an example; this all applies to the other options as well.
4070
4071 4. Since libtool seems to echo the compile commands it is issuing, the output
4072 from "make" can be reduced a bit by putting "@" in front of each libtool
4073 compile command.
4074
4075 5. Patch from the folks at Google for configure.in to be a bit more thorough
4076 in checking for a suitable C++ installation before trying to compile the
4077 C++ stuff. This should fix a reported problem when a compiler was present,
4078 but no suitable headers.
4079
4080 6. The man pages all had just "PCRE" as their title. I have changed them to
4081 be the relevant file name. I have also arranged that these names are
4082 retained in the file doc/pcre.txt, which is a concatenation in text format
4083 of all the man pages except the little individual ones for each function.
4084
4085 7. The NON-UNIX-USE file had not been updated for the different set of source
4086 files that come with release 6. I also added a few comments about the C++
4087 wrapper.
4088
4089
4090 Version 6.0 07-Jun-05
4091 ---------------------
4092
4093 1. Some minor internal re-organization to help with my DFA experiments.
4094
4095 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4096 didn't matter for the library itself when fully configured, but did matter
4097 when compiling without UCP support, or within Exim, where the ucp files are
4098 not imported.
4099
4100 3. Refactoring of the library code to split up the various functions into
4101 different source modules. The addition of the new DFA matching code (see
4102 below) to a single monolithic source would have made it really too
4103 unwieldy, quite apart from causing all the code to be include in a
4104 statically linked application, when only some functions are used. This is
4105 relevant even without the DFA addition now that patterns can be compiled in
4106 one application and matched in another.
4107
4108 The downside of splitting up is that there have to be some external
4109 functions and data tables that are used internally in different modules of
4110 the library but which are not part of the API. These have all had their
4111 names changed to start with "_pcre_" so that they are unlikely to clash
4112 with other external names.
4113
4114 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4115 a different (DFA) algorithm. Although it is slower than the original
4116 function, it does have some advantages for certain types of matching
4117 problem.
4118
4119 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4120 including restarting after a partial match.
4121
4122 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4123 defined when compiling for Windows was sent to me. I have put it into the
4124 code, though I have no means of testing or verifying it.
4125
4126 7. Added the pcre_refcount() auxiliary function.
4127
4128 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4129 match before or at the first newline in the subject string. In pcretest,
4130 the /f option on a pattern can be used to set this.
4131
4132 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4133 would behave wrongly. This has been present in PCRE since release 4.0.
4134
4135 10. A number of changes to the pcregrep command:
4136
4137 (a) Refactored how -x works; insert ^(...)$ instead of setting
4138 PCRE_ANCHORED and checking the length, in preparation for adding
4139 something similar for -w.
4140
4141 (b) Added the -w (match as a word) option.
4142
4143 (c) Refactored the way lines are read and buffered so as to have more
4144 than one at a time available.
4145
4146 (d) Implemented a pcregrep test script.
4147
4148 (e) Added the -M (multiline match) option. This allows patterns to match
4149 over several lines of the subject. The buffering ensures that at least
4150 8K, or the rest of the document (whichever is the shorter) is available
4151 for matching (and similarly the previous 8K for lookbehind assertions).
4152
4153 (f) Changed the --help output so that it now says
4154
4155 -w, --word-regex(p)
4156
4157 instead of two lines, one with "regex" and the other with "regexp"
4158 because that confused at least one person since the short forms are the
4159 same. (This required a bit of code, as the output is generated
4160 automatically from a table. It wasn't just a text change.)
4161
4162 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4163 option but starts with a hyphen. Could be a pattern or a path name
4164 starting with a hyphen, for instance.
4165
4166 (h) "-" can be given as a file name to represent stdin.
4167
4168 (i) When file names are being printed, "(standard input)" is used for
4169 the standard input, for compatibility with GNU grep. Previously
4170 "<stdin>" was used.
4171
4172 (j) The option --label=xxx can be used to supply a name to be used for
4173 stdin when file names are being printed. There is no short form.
4174
4175 (k) Re-factored the options decoding logic because we are going to add
4176 two more options that take data. Such options can now be given in four
4177 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4178
4179 (l) Added the -A, -B, and -C options for requesting that lines of context
4180 around matches be printed.
4181
4182 (m) Added the -L option to print the names of files that do not contain
4183 any matching lines, that is, the complement of -l.
4184
4185 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4186 continue to scan other files.
4187
4188 (o) The -s option was incorrectly implemented. For compatibility with other
4189 greps, it now suppresses the error message for a non-existent or non-
4190 accessible file (but not the return code). There is a new option called
4191 -q that suppresses the output of matching lines, which was what -s was
4192 previously doing.
4193
4194 (p) Added --include and --exclude options to specify files for inclusion
4195 and exclusion when recursing.
4196
4197 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4198 Hopefully, it now does.
4199
4200 12. Missing cast in pcre_study().
4201
4202 13. Added an "uninstall" target to the makefile.
4203
4204 14. Replaced "extern" in the function prototypes in Makefile.in with
4205 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4206 world, but is set differently for Windows.
4207
4208 15. Added a second compiling function called pcre_compile2(). The only
4209 difference is that it has an extra argument, which is a pointer to an
4210 integer error code. When there is a compile-time failure, this is set
4211 non-zero, in addition to the error test pointer being set to point to an
4212 error message. The new argument may be NULL if no error number is required
4213 (but then you may as well call pcre_compile(), which is now just a
4214 wrapper). This facility is provided because some applications need a
4215 numeric error indication, but it has also enabled me to tidy up the way
4216 compile-time errors are handled in the POSIX wrapper.
4217
4218 16. Added VPATH=.libs to the makefile; this should help when building with one
4219 prefix path and installing with another. (Or so I'm told by someone who
4220 knows more about this stuff than I do.)
4221
4222 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4223 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4224 match everything, including newlines. This is not POSIX-compatible, but
4225 somebody wanted the feature. From pcretest it can be activated by using
4226 both the P and the s flags.
4227
4228 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4229
4230 19. libpcre.pc was being incorrectly installed as executable.
4231
4232 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4233 it now also looks for '\r' so that it will work unmodified on Windows.
4234
4235 21. Added Google's contributed C++ wrapper to the distribution.
4236
4237 22. Added some untidy missing memory free() calls in pcretest, to keep
4238 Electric Fence happy when testing.
4239
4240
4241
4242 Version 5.0 13-Sep-04
4243 ---------------------
4244
4245 1. Internal change: literal characters are no longer packed up into items
4246 containing multiple characters in a single byte-string. Each character
4247 is now matched using a separate opcode. However, there may be more than one
4248 byte in the character in UTF-8 mode.
4249
4250 2. The pcre_callout_block structure has two new fields: pattern_position and
4251 next_item_length. These contain the offset in the pattern to the next match
4252 item, and its length, respectively.
4253
4254 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4255 insertion of callouts before each pattern item. Added the /C option to
4256 pcretest to make use of this.
4257
4258 4. On the advice of a Windows user, the lines
4259
4260 #if defined(_WIN32) || defined(WIN32)
4261 _setmode( _fileno( stdout ), 0x8000 );
4262 #endif /* defined(_WIN32) || defined(WIN32) */
4263
4264 have been added to the source of pcretest. This apparently does useful
4265 magic in relation to line terminators.
4266
4267 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4268 for the benefit of those environments where the "b" makes a difference.
4269
4270 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4271 to know about it. I have put a hack into configure.in that adds in code
4272 to set GCC=yes if CC=icc. This seems to end up at a point in the
4273 generated configure script that is early enough to affect the setting of
4274 compiler options, which is what is needed, but I have no means of testing
4275 whether it really works. (The user who reported this had patched the
4276 generated configure script, which of course I cannot do.)
4277
4278 LATER: After change 22 below (new libtool files), the configure script
4279 seems to know about icc (and also ecc). Therefore, I have commented out
4280 this hack in configure.in.
4281
4282 7. Added support for pkg-config (2 patches were sent in).
4283
4284 8. Negated POSIX character classes that used a combination of internal tables
4285 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4286 [[:^ascii]]. Typically, they would match almost any characters. The other
4287 POSIX classes were not broken in this way.
4288
4289 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4290 to find the match, as PCRE was deluded into thinking that the match had to
4291 start at the start point or following a newline. The same bug applied to
4292 patterns with negative forward assertions or any backward assertions
4293 preceding ".*" at the start, unless the pattern required a fixed first
4294 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4295
4296 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4297 starting at the last subject character, bytes beyond the end of the subject
4298 string were read.
4299
4300 11. Renamed the variable "class" as "classbits" to make life easier for C++
4301 users. (Previously there was a macro definition, but it apparently wasn't
4302 enough.)
4303
4304 12. Added the new field "tables" to the extra data so that tables can be passed
4305 in at exec time, or the internal tables can be re-selected. This allows
4306 a compiled regex to be saved and re-used at a later time by a different
4307 program that might have everything at different addresses.
4308
4309 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4310 -R library as well as a -L library.
4311
4312 14. The debugging options of pcretest (-d on the command line or D on a
4313 pattern) showed incorrect output for anything following an extended class
4314 that contained multibyte characters and which was followed by a quantifier.
4315
4316 15. Added optional support for general category Unicode character properties
4317 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4318 support. It adds about 90K to the size of the library. The meanings of the
4319 inbuilt class escapes such as \d and \s have NOT been changed.
4320
4321 16. Updated pcredemo.c to include calls to free() to release the memory for the
4322 compiled pattern.
4323
4324 17. The generated file chartables.c was being created in the source directory
4325 instead of in the building directory. This caused the build to fail if the
4326 source directory was different from the building directory, and was
4327 read-only.
4328
4329 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4330 file. No doubt somebody will tell me if they don't make sense... Also added
4331 Dan Mooney's comments about building on OpenVMS.
4332
4333 19. Added support for partial matching via the PCRE_PARTIAL option for
4334 pcre_exec() and the \P data escape in pcretest.
4335
4336 20. Extended pcretest with 3 new pattern features:
4337
4338 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4339 write the compiled pattern to the file whose name is "rest-of-line".
4340 This is a straight binary dump of the data, with the saved pointer to
4341 the character tables forced to be NULL. The study data, if any, is
4342 written too. After writing, pcretest reads a new pattern.
4343
4344 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4345 compiled pattern from the given file. There must not be any
4346 occurrences of "<" in the file name (pretty unlikely); if there are,
4347 pcretest will instead treat the initial "<" as a pattern delimiter.
4348 After reading in the pattern, pcretest goes on to read data lines as
4349 usual.
4350
4351 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4352 and 16-bit fields in a compiled pattern, to simulate a pattern that
4353 was compiled on a host of opposite endianness.
4354
4355 21. The pcre-exec() function can now cope with patterns that were compiled on
4356 hosts of opposite endianness, with this restriction:
4357
4358 As for any compiled expression that is saved and used later, the tables
4359 pointer field cannot be preserved; the extra_data field in the arguments
4360 to pcre_exec() should be used to pass in a tables address if a value
4361 other than the default internal tables were used at compile time.
4362
4363 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4364 now diagnosed as an error. Previously, most of the time, a negative number
4365 would have been treated as zero, but if in addition "ovector" was passed as
4366 NULL, a crash could occur.
4367
4368 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4369 new versions from the libtool 1.5 distribution (the last one is a copy of
4370 a file called libtool.m4). This seems to have fixed the need to patch
4371 "configure" to support Darwin 1.3 (which I used to do). However, I still
4372 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4373 workstation).
4374
4375 24. Changed the PCRE licence to be the more standard "BSD" licence.
4376
4377
4378 Version 4.5 01-Dec-03
4379 ---------------------
4380
4381 1. There has been some re-arrangement of the code for the match() function so
4382 that it can be compiled in a version that does not call itself recursively.
4383 Instead, it keeps those local variables that need separate instances for
4384 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4385 needs to "recurse". Keeping track of where control must go is done by means
4386 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4387 hide most of the details from the main code, and operates only if
4388 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4389 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4390 operating.
4391
4392 To make it easier for callers to provide specially tailored get/free
4393 functions for this usage, two new functions, pcre_stack_malloc, and
4394 pcre_stack_free, are used. They are always called in strict stacking order,
4395 and the size of block requested is always the same.
4396
4397 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4398 PCRE has been compiled to use the stack or the heap for recursion. The
4399 -C option of pcretest uses this to show which version is compiled.
4400
4401 A new data escape \S, is added to pcretest; it causes the amounts of store
4402 obtained and freed by both kinds of malloc/free at match time to be added
4403 to the output.
4404
4405 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4406 what's available on my current Linux desktop machine.
4407
4408 3. When matching a UTF-8 string, the test for a valid string at the start has
4409 been extended. If start_offset is not zero, PCRE now checks that it points
4410 to a byte that is the start of a UTF-8 character. If not, it returns
4411 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4412 this is necessary because there may be backward assertions in the pattern.
4413 When matching the same subject several times, it may save resources to use
4414 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4415
4416 4. The code for checking the validity of UTF-8 strings has been tightened so
4417 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4418 containing "overlong sequences".
4419
4420 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4421 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4422 should have been "&", but it just so happened that all the cases this let
4423 through by mistake were picked up later in the function.
4424
4425 6. I had used a variable called "isblank" - this is a C99 function, causing
4426 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4427
4428 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4429 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4430 so the version is shown; (c) stop "make test" echoing "./RunTest".
4431
4432 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4433
4434 9. The prototype for memmove() for systems that don't have it was using
4435 size_t, but the inclusion of the header that defines size_t was later. I've
4436 moved the #includes for the C headers earlier to avoid this.
4437
4438 10. Added some adjustments to the code to make it easier to compiler on certain
4439 special systems:
4440
4441 (a) Some "const" qualifiers were missing.
4442 (b) Added the macro EXPORT before all exported functions; by default this
4443 is defined to be empty.
4444 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4445 that it reads its output file name as an argument instead of writing
4446 to the standard output and assuming this can be redirected.
4447
4448 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4449 class containing characters with values greater than 255, PCRE compilation
4450 went into a loop.
4451
4452 12. A recursive reference to a subpattern that was within another subpattern
4453 that had a minimum quantifier of zero caused PCRE to crash. For example,
4454 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4455 recursion. If the recursively-called subpattern itself had a zero repeat,
4456 that was OK.
4457
4458 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4459 buffer into which it was copied (for escape processing) was still set at
4460 1024, so long lines caused crashes.
4461
4462 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4463 "internal error: code overflow...". This applied to any character class
4464 that was followed by a possessive quantifier.
4465
4466 15. Modified the Makefile to add libpcre.la as a prerequisite for
4467 libpcreposix.la because I was told this is needed for a parallel build to
4468 work.
4469
4470 16. If a pattern that contained .* following optional items at the start was
4471 studied, the wrong optimizing data was generated, leading to matching
4472 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4473 matching string must start with a or b or c. The correct conclusion for
4474 this pattern is that a match can start with any character.
4475
4476
4477 Version 4.4 13-Aug-03
4478 ---------------------
4479
4480 1. In UTF-8 mode, a character class containing characters with values between
4481 127 and 255 was not handled correctly if the compiled pattern was studied.
4482 In fixing this, I have also improved the studying algorithm for such
4483 classes (slightly).
4484
4485 2. Three internal functions had redundant arguments passed to them. Removal
4486 might give a very teeny performance improvement.
4487
4488 3. Documentation bug: the value of the capture_top field in a callout is *one
4489 more than* the number of the hightest numbered captured substring.
4490
4491 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4492 in incorrectly linking with a previously installed version. They now link
4493 explicitly with libpcre.la.
4494
4495 5. configure.in no longer needs to recognize Cygwin specially.
4496
4497 6. A problem in pcre.in for Windows platforms is fixed.
4498
4499 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4500 pcretest, it used to include the size of the study block as part of its
4501 output. Unfortunately, the structure contains a field that has a different
4502 size on different hardware architectures. This meant that the tests that
4503 showed this size failed. As the block is currently always of a fixed size,
4504 this information isn't actually particularly useful in pcretest output, so
4505 I have just removed it.
4506
4507 8. Three pre-processor statements accidentally did not start in column 1.
4508 Sadly, there are *still* compilers around that complain, even though
4509 standard C has not required this for well over a decade. Sigh.
4510
4511 9. In pcretest, the code for checking callouts passed small integers in the
4512 callout_data field, which is a void * field. However, some picky compilers
4513 complained about the casts involved for this on 64-bit systems. Now
4514 pcretest passes the address of the small integer instead, which should get
4515 rid of the warnings.
4516
4517 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4518 both compile and run time, and gives an error if an invalid UTF-8 sequence
4519 is found. There is a option for disabling this check in cases where the
4520 string is known to be correct and/or the maximum performance is wanted.
4521
4522 11. In response to a bug report, I changed one line in Makefile.in from
4523
4524 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4525 to
4526 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4527
4528 to look similar to other lines, but I have no way of telling whether this
4529 is the right thing to do, as I do not use Windows. No doubt I'll get told
4530 if it's wrong...
4531
4532
4533 Version 4.3 21-May-03
4534 ---------------------
4535
4536 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4537 Makefile.
4538
4539 2. Some refactoring to improve the quality of the code:
4540
4541 (i) The utf8_table... variables are now declared "const".
4542
4543 (ii) The code for \cx, which used the "case flipping" table to upper case
4544 lower case letters, now just substracts 32. This is ASCII-specific,
4545 but the whole concept of \cx is ASCII-specific, so it seems
4546 reasonable.
4547
4548 (iii) PCRE was using its character types table to recognize decimal and
4549 hexadecimal digits in the pattern. This is silly, because it handles
4550 only 0-9, a-f, and A-F, but the character types table is locale-
4551 specific, which means strange things might happen. A private
4552 table is now used for this - though it costs 256 bytes, a table is
4553 much faster than multiple explicit tests. Of course, the standard
4554 character types table is still used for matching digits in subject
4555 strings against \d.
4556
4557 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4558 ending in _t are). So I've renamed it as ESC_tee.
4559
4560 3. The first argument for regexec() in the POSIX wrapper should have been
4561 defined as "const".
4562
4563 4. Changed pcretest to use malloc() for its buffers so that they can be
4564 Electric Fenced for debugging.
4565
4566 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4567 to read one or more bytes before the start of the subject string. Often this
4568 had no effect on PCRE's behaviour, but in some circumstances it could
4569 provoke a segmentation fault.
4570
4571 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4572 to try to read one or more bytes before the start of the subject string.
4573
4574 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4575 UTF-8 support could misbehave in various ways if the subject string
4576 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4577 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4578 back over UTF-8 characters.)
4579
4580
4581 Version 4.2 14-Apr-03
4582 ---------------------
4583
4584 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4585
4586 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4587 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4588 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4589 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4590 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4591 and BUILD_EXEEXT
4592 Note: automatic setting of the BUILD variables is not yet working
4593 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4594 compile-time but not at link-time
4595 [LINK]: use for linking executables only
4596 make different versions for Windows and non-Windows
4597 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4598 libraries
4599 [LINK_FOR_BUILD]: new variable
4600 [OBJEXT]: use throughout
4601 [EXEEXT]: use throughout
4602 <winshared>: new target
4603 <wininstall>: new target
4604 <dftables.o>: use native compiler
4605 <dftables>: use native linker
4606 <install>: handle Windows platform correctly
4607 <clean>: ditto
4608 <check>: ditto
4609 copy DLL to top builddir before testing
4610
4611 As part of these changes, -no-undefined was removed again. This was reported
4612 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4613 in any case.
4614
4615 3. Some tidies to get rid of compiler warnings:
4616
4617 . In the match_data structure, match_limit was an unsigned long int, whereas
4618 match_call_count was an int. I've made them both unsigned long ints.
4619
4620 . In pcretest the fact that a const uschar * doesn't automatically cast to
4621 a void * provoked a warning.
4622
4623 . Turning on some more compiler warnings threw up some "shadow" variables
4624 and a few more missing casts.
4625
4626 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4627 option, a class that contained a single character with a value between 128
4628 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4629
4630 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4631 option, a class that contained several characters, but with at least one
4632 whose value was between 128 and 255 caused PCRE to crash.
4633
4634
4635 Version 4.1 12-Mar-03
4636 ---------------------
4637
4638 1. Compiling with gcc -pedantic found a couple of places where casts were
4639 needed, and a string in dftables.c that was longer than standard compilers are
4640 required to support.
4641
4642 2. Compiling with Sun's compiler found a few more places where the code could
4643 be tidied up in order to avoid warnings.
4644
4645 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4646 first of these names is deprecated in the latest Autoconf in favour of the name
4647 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4648 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4649 analogy I have changed it to CFLAGS_FOR_BUILD.
4650
4651 4. Added -no-undefined to the linking command in the Makefile, because this is
4652 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4653 linking step for the pcreposix library.
4654
4655 5. PCRE was failing to diagnose the case of two named groups with the same
4656 name.
4657
4658 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4659 literal character that is needed in the subject for a match, and scans along to
4660 ensure that it is present before embarking on the full matching process. This
4661 saves time in cases of nested unlimited repeats that are never going to match.
4662 Problem: the scan can take a lot of time if the subject is very long (e.g.
4663 megabytes), thus penalizing straightforward matches. It is now done only if the
4664 amount of subject to be scanned is less than 1000 bytes.
4665
4666 7. A lesser problem with the same optimization is that it was recording the
4667 first character of an anchored pattern as "needed", thus provoking a search
4668 right along the subject, even when the first match of the pattern was going to
4669 fail. The "needed" character is now not set for anchored patterns, unless it
4670 follows something in the pattern that is of non-fixed length. Thus, it still
4671 fulfils its original purpose of finding quick non-matches in cases of nested
4672 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4673
4674
4675 Version 4.0 17-Feb-03
4676 ---------------------
4677
4678 1. If a comment in an extended regex that started immediately after a meta-item
4679 extended to the end of string, PCRE compiled incorrect data. This could lead to
4680 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4681
4682 2. Moved to autoconf 2.53 and libtool 1.4.2.
4683
4684 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4685 the special perltest8 script is no longer needed - all the tests can be run
4686 from a single perltest script.
4687
4688 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4689 by \s. It has now been removed in PCRE. This means it isn't recognized as
4690 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4691 class [:space:] *does* include VT, thereby creating a mess.
4692
4693 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4694 space and tab.
4695
4696 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4697 its new features into the main test script, reducing the number of scripts.
4698
4699 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4700 were backward compatible, and made the (?i) apply to the whole pattern, as if
4701 /i were given. Now it behaves more logically, and applies the option setting
4702 only to what follows. PCRE has been changed to follow suit. However, if it
4703 finds options settings right at the start of the pattern, it extracts them into
4704 the global options, as before. Thus, they show up in the info data.
4705
4706 8. Added support for the \Q...\E escape sequence. Characters in between are
4707 treated as literals. This is slightly different from Perl in that $ and @ are
4708 also handled as literals inside the quotes. In Perl, they will cause variable
4709 interpolation. Note the following examples:
4710
4711 Pattern PCRE matches Perl matches
4712
4713 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4714 \Qabc\$xyz\E abc\$xyz abc\$xyz
4715 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4716
4717 For compatibility with Perl, \Q...\E sequences are recognized inside character
4718 classes as well as outside them.
4719
4720 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4721 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4722 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4723 signed/unsigned warnings.
4724
4725 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4726 option for pcretest, so I've replaced it by a simple function that does just
4727 that job.
4728
4729 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4730 "pcregrep -".
4731
4732 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4733 Java package. This provides some syntactic sugar for simple cases of what my
4734 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4735 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4736 item, you can use this simplified notation. Note that only makes sense with
4737 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4738 greediness, whatever the setting of the PCRE_UNGREEDY option.
4739
4740 13. A change of greediness default within a pattern was not taking effect at
4741 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4742 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4743 was abstracted outside.
4744
4745 14. PCRE now supports the \G assertion. It is true when the current matching
4746 position is at the start point of the match. This differs from \A when the
4747 starting offset is non-zero. Used with the /g option of pcretest (or similar
4748 code), it works in the same way as it does for Perl's /g option. If all
4749 alternatives of a regex begin with \G, the expression is anchored to the start
4750 match position, and the "anchored" flag is set in the compiled expression.
4751
4752 15. Some bugs concerning the handling of certain option changes within patterns
4753 have been fixed. These applied to options other than (?ims). For example,
4754 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4755 been the other way round. Some of this was related to change 7 above.
4756
4757 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4758 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4759 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4760 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4761
4762 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4763 mode. Unlike ".", it always matches newline, whatever the setting of
4764 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4765 assertions. Perl allows it, but it doesn't (in general) work because it can't
4766 calculate the length of the lookbehind. At least, that's the case for Perl
4767 5.8.0 - I've been told they are going to document that it doesn't work in
4768 future.
4769
4770 18. Added an error diagnosis for escapes that PCRE does not support: these are
4771 \L, \l, \N, \P, \p, \U, \u, and \X.
4772
4773 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4774 reading past the end of the pattern in cases such as /[abcd/.
4775
4776 20. PCRE was getting more memory than necessary for patterns with classes that
4777 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4778
4779 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4780 compiling PCRE for use with Virtual Pascal.
4781
4782 22. Small fix to the Makefile to make it work properly if the build is done
4783 outside the source tree.
4784
4785 23. Added a new extension: a condition to go with recursion. If a conditional
4786 subpattern starts with (?(R) the "true" branch is used if recursion has
4787 happened, whereas the "false" branch is used only at the top level.
4788
4789 24. When there was a very long string of literal characters (over 255 bytes
4790 without UTF support, over 250 bytes with UTF support), the computation of how
4791 much memory was required could be incorrect, leading to segfaults or other
4792 strange effects.
4793
4794 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4795 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4796 there was a subsequent back reference to those brackets. This meant that, for
4797 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4798 possible to check for precisely this case. All we can do is abandon the
4799 optimization if .* occurs inside capturing brackets when there are any back
4800 references whatsoever. (See below for a better fix that came later.)
4801
4802 26. The handling of the optimization for finding the first character of a
4803 non-anchored pattern, and for finding a character that is required later in the
4804 match were failing in some cases. This didn't break the matching; it just
4805 failed to optimize when it could. The way this is done has been re-implemented.
4806
4807 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4808
4809 28. Added a new feature that provides some of the functionality that Perl
4810 provides with (?{...}). The facility is termed a "callout". The way it is done
4811 in PCRE is for the caller to provide an optional function, by setting
4812 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4813 global variable. By default it is unset, which disables all calling out. To get
4814 the function called, the regex must include (?C) at appropriate points. This
4815 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4816 This provides a means of identifying different callout points. When PCRE
4817 reaches such a point in the regex, if pcre_callout has been set, the external
4818 function is called. It is provided with data in a structure called
4819 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4820 matching continues; if it returns a non-zero value, the match at the current
4821 point fails. However, backtracking will occur if possible. [This was changed
4822 later and other features added - see item 49 below.]
4823
4824 29. pcretest is upgraded to test the callout functionality. It provides a
4825 callout function that displays information. By default, it shows the start of
4826 the match and the current position in the text. There are some new data escapes
4827 to vary what happens:
4828
4829 \C+ in addition, show current contents of captured substrings
4830 \C- do not supply a callout function
4831 \C!n return 1 when callout number n is reached
4832 \C!n!m return 1 when callout number n is reached for the mth time
4833
4834 30. If pcregrep was called with the -l option and just a single file name, it
4835 output "<stdin>" if a match was found, instead of the file name.
4836
4837 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4838 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4839 pcre_exec(). This saves a malloc/free per call. The default value of
4840 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4841 when configuring.
4842
4843 32. The default maximum size of a compiled pattern is 64K. There have been a
4844 few cases of people hitting this limit. The code now uses macros to handle the
4845 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4846 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4847 configuring. Tests 2 and 5 work only with 2-byte links because they output
4848 debugging information about compiled patterns.
4849
4850 33. Internal code re-arrangements:
4851
4852 (a) Moved the debugging function for printing out a compiled regex into
4853 its own source file (printint.c) and used #include to pull it into
4854 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4855 separate copies.
4856
4857 (b) Defined the list of op-code names for debugging as a macro in
4858 internal.h so that it is next to the definition of the opcodes.
4859
4860 (c) Defined a table of op-code lengths for simpler skipping along compiled
4861 code. This is again a macro in internal.h so that it is next to the
4862 definition of the opcodes.
4863
4864 34. Added support for recursive calls to individual subpatterns, along the
4865 lines of Robin Houston's patch (but implemented somewhat differently).
4866
4867 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4868 allow it to read and process whole directories in Win32. This code was
4869 contributed by Lionel Fourquaux; it has not been tested by me.
4870
4871 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4872 used to name a group. Names consist of alphanumerics and underscores, and must
4873 be unique. Back references use the syntax (?P=name) and recursive calls use
4874 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4875 numbers. The function pcre_fullinfo() can be used after compilation to extract
4876 a name/number map. There are three relevant calls:
4877
4878 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4879 PCRE_INFO_NAMECOUNT yields the number of entries
4880 PCRE_INFO_NAMETABLE yields a pointer to the map.
4881
4882 The map is a vector of fixed-size entries. The size of each entry depends on
4883 the length of the longest name used. The first two bytes of each entry are the
4884 group number, most significant byte first. There follows the corresponding
4885 name, zero terminated. The names are in alphabetical order.
4886
4887 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4888 case instead of 255. Making it the same both with and without UTF-8 support
4889 means that the same test output works with both.
4890
4891 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4892 calling malloc() with a zero argument.
4893
4894 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4895 optimization. I've improved things by keeping a bitmap of backreferences with
4896 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4897 fact referenced, the optimization can be applied. It is unlikely that a
4898 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4899 the match to follow \n) will appear inside brackets with a number greater than
4900 31, but if it does, any back reference > 31 suppresses the optimization.
4901
4902 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4903 of disabling numbered capturing parentheses. Any opening parenthesis that is
4904 not followed by ? behaves as if it were followed by ?: but named parentheses
4905 can still be used for capturing (and they will acquire numbers in the usual
4906 way).
4907
4908 41. Redesigned the return codes from the match() function into yes/no/error so
4909 that errors can be passed back from deep inside the nested calls. A malloc
4910 failure while inside a recursive subpattern call now causes the
4911 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4912
4913 42. It is now possible to set a limit on the number of times the match()
4914 function is called in a call to pcre_exec(). This facility makes it possible to
4915 limit the amount of recursion and backtracking, though not in a directly
4916 obvious way, because the match() function is used in a number of different
4917 circumstances. The count starts from zero for each position in the subject
4918 string (for non-anchored patterns). The default limit is, for compatibility, a
4919 large number, namely 10 000 000. You can change this in two ways:
4920
4921 (a) When configuring PCRE before making, you can use --with-match-limit=n
4922 to set a default value for the compiled library.
4923
4924 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4925 a different value is set. See 45 below.
4926
4927 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4928
4929 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4930 of things that can be changed at compile time. The first argument specifies
4931 what is wanted and the second points to where the information is to be placed.
4932 The current list of available information is:
4933
4934 PCRE_CONFIG_UTF8
4935
4936 The output is an integer that is set to one if UTF-8 support is available;
4937 otherwise it is set to zero.
4938
4939 PCRE_CONFIG_NEWLINE
4940
4941 The output is an integer that it set to the value of the code that is used for
4942 newline. It is either LF (10) or CR (13).
4943
4944 PCRE_CONFIG_LINK_SIZE
4945
4946 The output is an integer that contains the number of bytes used for internal
4947 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4948
4949 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4950
4951 The output is an integer that contains the threshold above which the POSIX
4952 interface uses malloc() for output vectors. See item 31 above.
4953
4954 PCRE_CONFIG_MATCH_LIMIT
4955
4956 The output is an unsigned integer that contains the default limit of the number
4957 of match() calls in a pcre_exec() execution. See 42 above.
4958
4959 44. pcretest has been upgraded by the addition of the -C option. This causes it
4960 to extract all the available output from the new pcre_config() function, and to
4961 output it. The program then exits immediately.
4962
4963 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4964 order to support additional features. One way would have been to define
4965 pcre_exec2() (for example) with extra arguments, but this would not have been
4966 extensible, and would also have required all calls to the original function to
4967 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4968 is used for passing in "extra" data from pcre_study().
4969
4970 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4971 contains the following fields:
4972
4973 flags a bitmap indicating which of the following fields are set
4974 study_data opaque data from pcre_study()
4975 match_limit a way of specifying a limit on match() calls for a specific
4976 call to pcre_exec()
4977 callout_data data for callouts (see 49 below)
4978
4979 The flag bits are also defined in pcre.h, and are
4980
4981 PCRE_EXTRA_STUDY_DATA
4982 PCRE_EXTRA_MATCH_LIMIT
4983 PCRE_EXTRA_CALLOUT_DATA
4984
4985 The pcre_study() function now returns one of these new pcre_extra blocks, with
4986 the actual study data pointed to by the study_data field, and the
4987 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4988 before. That is, this change is entirely upwards-compatible and requires no
4989 change to existing code.
4990
4991 If you want to pass in additional data to pcre_exec(), you can either place it
4992 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4993 block.
4994
4995 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4996 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4997 times with different match limits, until it finds the minimum value needed for
4998 pcre_exec() to complete. The value is then output. This can be instructive; for
4999 most simple matches the number is quite small, but for pathological cases it
5000 gets very large very quickly.
5001
5002 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5003 returns the size of the data block pointed to by the study_data field in a
5004 pcre_extra block, that is, the value that was passed as the argument to
5005 pcre_malloc() when PCRE was getting memory in which to place the information
5006 created by pcre_study(). The fourth argument should point to a size_t variable.
5007 pcretest has been extended so that this information is shown after a successful
5008 pcre_study() call when information about the compiled regex is being displayed.
5009
5010 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5011 because what follows is always an absolute path. (Later: it turns out that this
5012 is more than cosmetic for MinGW, because it doesn't like empty path
5013 components.)
5014
5015 49. Some changes have been made to the callout feature (see 28 above):
5016
5017 (i) A callout function now has three choices for what it returns:
5018
5019 0 => success, carry on matching
5020 > 0 => failure at this point, but backtrack if possible
5021 < 0 => serious error, return this value from pcre_exec()
5022
5023 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5024 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5025 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5026 use by callout functions. It will never be used by PCRE itself.
5027
5028 (ii) The pcre_extra structure (see 45 above) has a void * field called
5029 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5030 pcre_callout_block structure has a field of the same name. The contents of
5031 the field passed in the pcre_extra structure are passed to the callout
5032 function in the corresponding field in the callout block. This makes it
5033 easier to use the same callout-containing regex from multiple threads. For
5034 testing, the pcretest program has a new data escape
5035
5036 \C*n pass the number n (may be negative) as callout_data
5037
5038 If the callout function in pcretest receives a non-zero value as
5039 callout_data, it returns that value.
5040
5041 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5042 there were some redundant $(CFLAGS) in commands that are now specified as
5043 $(LINK), which already includes $(CFLAGS).
5044
5045 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5046 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5047 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5048 one-byte characters throughout. Note that case-insensitive matching applies
5049 only to characters whose values are less than 256. PCRE doesn't support the
5050 notion of cases for higher-valued characters.
5051
5052 (i) A character class whose characters are all within 0-255 is handled as
5053 a bit map, and the map is inverted for negative classes. Previously, a
5054 character > 255 always failed to match such a class; however it should
5055 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5056
5057 (ii) A negated character class with a single character < 255 is coded as
5058 "not this character" (OP_NOT). This wasn't working properly when the test
5059 character was multibyte, either singly or repeated.
5060
5061 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5062 mode, for example: \x{100}{2,3}.
5063
5064 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5065 singly or repeated) now correctly test multibyte characters. However,
5066 PCRE doesn't recognize any characters with values greater than 255 as
5067 digits, spaces, or word characters. Such characters always match \D, \S,
5068 and \W, and never match \d, \s, or \w.
5069
5070 (v) Classes may now contain characters and character ranges with values
5071 greater than 255. For example: [ab\x{100}-\x{400}].
5072
5073 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5074 PCRE in UTF-8 mode.
5075
5076 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5077 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5078 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5079 value.)
5080
5081 53. The single man page has become too large. I have therefore split it up into
5082 a number of separate man pages. These also give rise to individual HTML pages;
5083 these are now put in a separate directory, and there is an index.html page that
5084 lists them all. Some hyperlinking between the pages has been installed.
5085
5086 54. Added convenience functions for handling named capturing parentheses.
5087
5088 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5089 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5090 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5091 are faulted.
5092
5093 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5094 calling configure. These values are used when compiling the dftables.c program
5095 which is run to generate the source of the default character tables. They
5096 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5097 you will need to set these values.
5098
5099 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5100
5101
5102 Version 3.9 02-Jan-02
5103 ---------------------
5104
5105 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5106
5107 2. If --disable-static was given, the building process failed when trying to
5108 build pcretest and pcregrep. (For some reason it was using libtool to compile
5109 them, which is not right, as they aren't part of the library.)
5110
5111
5112 Version 3.8 18-Dec-01
5113 ---------------------
5114
5115 1. The experimental UTF-8 code was completely screwed up. It was packing the
5116 bytes in the wrong order. How dumb can you get?
5117
5118
5119 Version 3.7 29-Oct-01
5120 ---------------------
5121
5122 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5123 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5124 this didn't happen under Solaris 8, where I normally test things.
5125
5126 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5127 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5128 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5129 this fix an hour or so after the initial 3.7 release.)
5130
5131
5132 Version 3.6 23-Oct-01
5133 ---------------------
5134
5135 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5136 offsets passed as NULL with zero offset count.
5137
5138 2. The config.guess and config.sub files had not been updated when I moved to
5139 the latest autoconf.
5140
5141
5142 Version 3.5 15-Aug-01
5143 ---------------------
5144
5145 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5146 had been forgotten.
5147
5148 2. By using declared but undefined structures, we can avoid using "void"
5149 definitions in pcre.h while keeping the internal definitions of the structures
5150 private.
5151
5152 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5153 user point of view, this means that both static and shared libraries are built
5154 by default, but this can be individually controlled. More of the work of
5155 handling this static/shared cases is now inside libtool instead of PCRE's make
5156 file.
5157
5158 4. The pcretest utility is now installed along with pcregrep because it is
5159 useful for users (to test regexs) and by doing this, it automatically gets
5160 relinked by libtool. The documentation has been turned into a man page, so
5161 there are now .1, .txt, and .html versions in /doc.
5162
5163 5. Upgrades to pcregrep:
5164 (i) Added long-form option names like gnu grep.
5165 (ii) Added --help to list all options with an explanatory phrase.
5166 (iii) Added -r, --recursive to recurse into sub-directories.
5167 (iv) Added -f, --file to read patterns from a file.
5168
5169 6. pcre_exec() was referring to its "code" argument before testing that
5170 argument for NULL (and giving an error if it was NULL).
5171
5172 7. Upgraded Makefile.in to allow for compiling in a different directory from
5173 the source directory.
5174
5175 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5176 options bits, the pointer it was passed was to an int instead of to an unsigned
5177 long int. This mattered only on 64-bit systems.
5178
5179 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5180 generated) instead of pcre.in, which it its source. Also made the same change
5181 in several of the .c files.
5182
5183 10. A new release of gcc defines printf() as a macro, which broke pcretest
5184 because it had an ifdef in the middle of a string argument for printf(). Fixed
5185 by using separate calls to printf().
5186
5187 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5188 script, to force use of CR or LF instead of \n in the source. On non-Unix
5189 systems, the value can be set in config.h.
5190
5191 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5192 absolute limit. Changed the text of the error message to make this clear, and
5193 likewise updated the man page.
5194
5195 13. The limit of 99 on the number of capturing subpatterns has been removed.
5196 The new limit is 65535, which I hope will not be a "real" limit.
5197
5198
5199 Version 3.4 22-Aug-00
5200 ---------------------
5201
5202 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5203
5204 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5205
5206
5207 Version 3.3 01-Aug-00
5208 ---------------------
5209
5210 1. If an octal character was given, but the value was greater than \377, it
5211 was not getting masked to the least significant bits, as documented. This could
5212 lead to crashes in some systems.
5213
5214 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5215 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5216
5217 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5218 These just pass their arguments on to (pcre_free)(), but they are provided
5219 because some uses of PCRE bind it to non-C systems that can call its functions,
5220 but cannot call free() or pcre_free() directly.
5221
5222 4. Add "make test" as a synonym for "make check". Corrected some comments in
5223 the Makefile.
5224
5225 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5226 Makefile.
5227
5228 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5229 command called pgrep for grepping around the active processes.
5230
5231 7. Added the beginnings of support for UTF-8 character strings.
5232
5233 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5234 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5235 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5236 out for the ar command.)
5237
5238
5239 Version 3.2 12-May-00
5240 ---------------------
5241
5242 This is purely a bug fixing release.
5243
5244 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5245 of ZA. This was just one example of several cases that could provoke this bug,
5246 which was introduced by change 9 of version 2.00. The code for breaking
5247 infinite loops after an iteration that matches an empty string was't working
5248 correctly.
5249
5250 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5251 when matched against abbab (for example). After matching an empty string, it
5252 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5253 caused it to match further down the string than it should.
5254
5255 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5256 was there because it doesn't seem to be needed, and it causes trouble on some
5257 systems, as it is not a Standard C header. It has been removed.
5258
5259 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5260 were reported on the Macintosh. The changes were from
5261
5262 while ((c = *(++ptr)) != 0 && c != '\n');
5263 to
5264 while ((c = *(++ptr)) != 0 && c != '\n') ;
5265
5266 Totally extraordinary, but if that's what it takes...
5267
5268 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5269 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5270 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5271 assumes the way PCRE uses memmove() (always moving upwards).
5272
5273 6. PCRE is being used in one environment where strchr() is not available. There
5274 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5275 faster code anyway.
5276
5277
5278 Version 3.1 09-Feb-00
5279 ---------------------
5280
5281 The only change in this release is the fixing of some bugs in Makefile.in for
5282 the "install" target:
5283
5284 (1) It was failing to install pcreposix.h.
5285
5286 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5287
5288
5289 Version 3.0 01-Feb-00
5290 ---------------------
5291
5292 1. Add support for the /+ modifier to perltest (to output $` like it does in
5293 pcretest).
5294
5295 2. Add support for the /g modifier to perltest.
5296
5297 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5298 matches null strings.
5299
5300 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5301 pattern. Perl treats empty patterns specially - it reuses the most recent
5302 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5303 effect.
5304
5305 5. The POSIX interface was broken in that it was just handing over the POSIX
5306 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5307 required a bigger vector, with some working space on the end. This means that
5308 the POSIX wrapper now has to get and free some memory, and copy the results.
5309
5310 6. Added some simple autoconf support, placing the test data and the
5311 documentation in separate directories, re-organizing some of the
5312 information files, and making it build pcre-config (a GNU standard). Also added
5313 libtool support for building PCRE as a shared library, which is now the
5314 default.
5315
5316 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5317 09 are not valid octal constants. Single digits will be used for minor values
5318 less than 10.
5319
5320 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5321 existing programs that set these in the POSIX interface can use PCRE without
5322 modification.
5323
5324 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5325 return all that pcre_info() returns, plus additional data. The pcre_info()
5326 function is retained for compatibility, but is considered to be obsolete.
5327
5328 10. Added experimental recursion feature (?R) to handle one common case that
5329 Perl 5.6 will be able to do with (?p{...}).
5330
5331 11. Added support for POSIX character classes like [:alpha:], which Perl is
5332 adopting.
5333
5334
5335 Version 2.08 31-Aug-99
5336 ----------------------
5337
5338 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5339 trying to match at the startoffset position, but instead was moving forward to
5340 the next newline as if a previous match had failed.
5341
5342 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5343 and could get into a loop if a null string was matched other than at the start
5344 of the subject.
5345
5346 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5347 be distinguished at compile time, and for completeness also added PCRE_DATE.
5348
5349 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5350 in GnuWin32 environments.
5351
5352
5353 Version 2.07 29-Jul-99
5354 ----------------------
5355
5356 1. The documentation is now supplied in plain text form and HTML as well as in
5357 the form of man page sources.
5358
5359 2. C++ compilers don't like assigning (void *) values to other pointer types.
5360 In particular this affects malloc(). Although there is no problem in Standard
5361 C, I've put in casts to keep C++ compilers happy.
5362
5363 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5364 should be (const char *).
5365
5366 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5367 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5368 However, I haven't made this a standard facility. The documentation doesn't
5369 mention it, and the Makefile doesn't support it.
5370
5371 5. The Makefile now contains an "install" target, with editable destinations at
5372 the top of the file. The pcretest program is not installed.
5373
5374 6. pgrep -V now gives the PCRE version number and date.
5375
5376 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5377 causing the entire string to be ignored, instead of just the last character.
5378
5379 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5380 non-matching string, it can take a very, very long time, even for strings of
5381 quite modest length, because of the nested recursion. PCRE now does better in
5382 some of these cases. It does this by remembering the last required literal
5383 character in the pattern, and pre-searching the subject to ensure it is present
5384 before running the real match. In other words, it applies a heuristic to detect
5385 some types of certain failure quickly, and in the above example, if presented
5386 with a string that has no trailing " it gives "no match" very quickly.
5387
5388 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5389 other alternatives are tried instead.
5390
5391
5392 Version 2.06 09-Jun-99
5393 ----------------------
5394
5395 1. Change pcretest's output for amount of store used to show just the code
5396 space, because the remainder (the data block) varies in size between 32-bit and
5397 64-bit systems.
5398
5399 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5400 start matching at. This allows lookbehinds to work when searching for multiple
5401 occurrences in a string.
5402
5403 3. Added additional options to pcretest for testing multiple occurrences:
5404
5405 /+ outputs the rest of the string that follows a match
5406 /g loops for multiple occurrences, using the new startoffset argument
5407 /G loops for multiple occurrences by passing an incremented pointer
5408
5409 4. PCRE wasn't doing the "first character" optimization for patterns starting
5410 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5411 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5412 the letter 'x'. On long subject strings, this gives a significant speed-up.
5413
5414
5415 Version 2.05 21-Apr-99
5416 ----------------------
5417
5418 1. Changed the type of magic_number from int to long int so that it works
5419 properly on 16-bit systems.
5420
5421 2. Fixed a bug which caused patterns starting with .* not to work correctly
5422 when the subject string contained newline characters. PCRE was assuming
5423 anchoring for such patterns in all cases, which is not correct because .* will
5424 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5425 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5426 must be retried after every newline in the subject.
5427
5428
5429 Version 2.04 18-Feb-99
5430 ----------------------
5431
5432 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5433 computation of the store needed to hold the pattern was incorrect (too large).
5434 If such patterns were nested a few deep, this could multiply and become a real
5435 problem.
5436
5437 2. Added /M option to pcretest to show the memory requirement of a specific
5438 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5439
5440 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5441 compiled in such a way that the backtracking after subsequent failure was
5442 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5443 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5444
5445
5446 Version 2.03 02-Feb-99
5447 ----------------------
5448
5449 1. Fixed typo and small mistake in man page.
5450
5451 2. Added 4th condition (GPL supersedes if conflict) and created separate
5452 LICENCE file containing the conditions.
5453
5454 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5455 Perl, that is the internal \ allows the delimiter to be included in the
5456 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5457 the final delimiter, add \ to the end of the pattern (to test the error).
5458
5459 4. Added the convenience functions for extracting substrings after a successful
5460 match. Updated pcretest to make it able to test these functions.
5461
5462
5463 Version 2.02 14-Jan-99
5464 ----------------------
5465
5466 1. Initialized the working variables associated with each extraction so that
5467 their saving and restoring doesn't refer to uninitialized store.
5468
5469 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5470 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5471 fix the problem.
5472
5473 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5474 calls, and wasn't printing the correct value for compiling calls. Increased the
5475 default value of LOOPREPEAT, and the number of significant figures in the
5476 times.
5477
5478 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5479
5480 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5481 a building problem on Windows NT with a FAT file system.
5482
5483
5484 Version 2.01 21-Oct-98
5485 ----------------------
5486
5487 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5488 to character tables built by pcre_maketables() in the current locale. If NULL
5489 is passed, the default tables are used.
5490
5491
5492 Version 2.00 24-Sep-98
5493 ----------------------
5494
5495 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5496 it any more.
5497
5498 2. Allow quantification of (?>) groups, and make it work correctly.
5499
5500 3. The first character computation wasn't working for (?>) groups.
5501
5502 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5503 end of the subject) and add 5.005's \z, which really does match only at the
5504 very end of the subject.
5505
5506 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5507
5508 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5509 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5510 localized options. All options to pcre_study() were also removed.
5511
5512 7. Add other new features from 5.005:
5513
5514 $(?<= positive lookbehind
5515 $(?<! negative lookbehind
5516 (?imsx-imsx) added the unsetting capability
5517 such a setting is global if at outer level; local otherwise
5518 (?imsx-imsx:) non-capturing groups with option setting
5519 (?(cond)re|re) conditional pattern matching
5520
5521 A backreference to itself in a repeated group matches the previous
5522 captured string.
5523
5524 8. General tidying up of studying (both automatic and via "study")
5525 consequential on the addition of new assertions.
5526
5527 9. As in 5.005, unlimited repeated groups that could match an empty substring
5528 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5529 runtime if any iteration does actually match an empty substring.
5530
5531 10. Include the RunTest script in the distribution.
5532
5533 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5534 discrepancies, some of which were old and were also with respect to 5.004. They
5535 have now been fixed.
5536
5537
5538 Version 1.09 28-Apr-98
5539 ----------------------
5540
5541 1. A negated single character class followed by a quantifier with a minimum
5542 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5543 program crashes, or just wrong answers. This did not apply to negated classes
5544 containing more than one character, or to minima other than one.
5545
5546
5547 Version 1.08 27-Mar-98
5548 ----------------------
5549
5550 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5551
5552 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5553 latter must appear before anything that relies on it in the pattern.
5554
5555
5556 Version 1.07 16-Feb-98
5557 ----------------------
5558
5559 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5560 repeat of a potentially empty string).
5561
5562
5563 Version 1.06 23-Jan-98
5564 ----------------------
5565
5566 1. Added Markus Oberhumer's little patches for C++.
5567
5568 2. Literal strings longer than 255 characters were broken.
5569
5570
5571 Version 1.05 23-Dec-97
5572 ----------------------
5573
5574 1. Negated character classes containing more than one character were failing if
5575 PCRE_CASELESS was set at run time.
5576
5577
5578 Version 1.04 19-Dec-97
5579 ----------------------
5580
5581 1. Corrected the man page, where some "const" qualifiers had been omitted.
5582
5583 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5584 input syntax.
5585
5586 3. Fixed memory leak which occurred when a regex with back references was
5587 matched with an offsets vector that wasn't big enough. The temporary memory
5588 that is used in this case wasn't being freed if the match failed.
5589
5590 4. Tidied pcretest to ensure it frees memory that it gets.
5591
5592 5. Temporary memory was being obtained in the case where the passed offsets
5593 vector was exactly big enough.
5594
5595 6. Corrected definition of offsetof() from change 5 below.
5596
5597 7. I had screwed up change 6 below and broken the rules for the use of
5598 setjmp(). Now fixed.
5599
5600
5601 Version 1.03 18-Dec-97
5602 ----------------------
5603
5604 1. A erroneous regex with a missing opening parenthesis was correctly
5605 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5606 on some systems.
5607
5608 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5609 it was reported that one broken compiler failed on the former because "code" is
5610 also an independent variable.
5611
5612 3. The erroneous regex a[]b caused an array overrun reference.
5613
5614 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5615 fail on data ending with that character. (It was going on too far, and checking
5616 the next character, typically a binary zero.) This was specific to the
5617 optimized code for single-character negative classes.
5618
5619 5. Added a contributed patch from the TIN world which does the following:
5620
5621 + Add an undef for memmove, in case the the system defines a macro for it.
5622
5623 + Add a definition of offsetof(), in case there isn't one. (I don't know
5624 the reason behind this - offsetof() is part of the ANSI standard - but
5625 it does no harm).
5626
5627 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5628 most of the places where whitespace preceded '#'. I have given up and
5629 allowed the remaining 2 cases to be at the margin.
5630
5631 + Rename some variables in pcre to eliminate shadowing. This seems very
5632 pedantic, but does no harm, of course.
5633
5634 6. Moved the call to setjmp() into its own function, to get rid of warnings
5635 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5636
5637 7. Constructs such as \d{8,} were compiling into the equivalent of
5638 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5639 outcome, but in this particular case used more store than had been allocated,
5640 which caused the bug to be discovered because it threw up an internal error.
5641
5642 8. The debugging code in both pcre and pcretest for outputting the compiled
5643 form of a regex was going wrong in the case of back references followed by
5644 curly-bracketed repeats.
5645
5646
5647 Version 1.02 12-Dec-97
5648 ----------------------
5649
5650 1. Typos in pcre.3 and comments in the source fixed.
5651
5652 2. Applied a contributed patch to get rid of places where it used to remove
5653 'const' from variables, and fixed some signed/unsigned and uninitialized
5654 variable warnings.
5655
5656 3. Added the "runtest" target to Makefile.
5657
5658 4. Set default compiler flag to -O2 rather than just -O.
5659
5660
5661 Version 1.01 19-Nov-97
5662 ----------------------
5663
5664 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5665 like /([ab]*)*/, that is, for classes with more than one character in them.
5666
5667 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5668 as /((?>a*))*/ (a PCRE_EXTRA facility).
5669
5670
5671 Version 1.00 18-Nov-97
5672 ----------------------
5673
5674 1. Added compile-time macros to support systems such as SunOS4 which don't have
5675 memmove() or strerror() but have other things that can be used instead.
5676
5677 2. Arranged that "make clean" removes the executables.
5678
5679
5680 Version 0.99 27-Oct-97
5681 ----------------------
5682
5683 1. Fixed bug in code for optimizing classes with only one character. It was
5684 initializing a 32-byte map regardless, which could cause it to run off the end
5685 of the memory it had got.
5686
5687 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5688
5689
5690 Version 0.98 22-Oct-97
5691 ----------------------
5692
5693 1. Fixed bug in code for handling temporary memory usage when there are more
5694 back references than supplied space in the ovector. This could cause segfaults.
5695
5696
5697 Version 0.97 21-Oct-97
5698 ----------------------
5699
5700 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5701
5702 2. Optimized negated single characters not to use a bit map.
5703
5704 3. Brought error texts together as macro definitions; clarified some of them;
5705 fixed one that was wrong - it said "range out of order" when it meant "invalid
5706 escape sequence".
5707
5708 4. Changed some char * arguments to const char *.
5709
5710 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5711
5712 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5713 pcretest.
5714
5715
5716 Version 0.96 16-Oct-97
5717 ----------------------
5718
5719 1. Added a simple "pgrep" utility to the distribution.
5720
5721 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5722 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5723 where "ddd" means "one or more decimal digits".
5724
5725 3. Fixed serious bug. If a pattern had a back reference, but the call to
5726 pcre_exec() didn't supply a large enough ovector to record the related
5727 identifying subpattern, the match always failed. PCRE now remembers the number
5728 of the largest back reference, and gets some temporary memory in which to save
5729 the offsets during matching if necessary, in order to ensure that
5730 backreferences always work.
5731
5732 4. Increased the compatibility with Perl in a number of ways:
5733
5734 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5735 to request this handling. The option can be set at compile or exec time.
5736
5737 (b) $ matches before a terminating newline by default; an option
5738 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5739 mode). The option can be set at compile or exec time.
5740
5741 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5742 the same as Perl's. If the decimal number it represents is less than 10
5743 or there aren't that many previous left capturing parentheses, an octal
5744 escape is read. Inside a character class, it's always an octal escape,
5745 even if it is a single digit.
5746
5747 (d) An escaped but undefined alphabetic character is taken as a literal,
5748 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5749 escapes.
5750
5751 (e) {0} is now permitted. (The previous item is removed from the compiled
5752 pattern).
5753
5754 5. Changed all the names of code files so that the basic parts are no longer
5755 than 10 characters, and abolished the teeny "globals.c" file.
5756
5757 6. Changed the handling of character classes; they are now done with a 32-byte
5758 bit map always.
5759
5760 7. Added the -d and /D options to pcretest to make it possible to look at the
5761 internals of compilation without having to recompile pcre.
5762
5763
5764 Version 0.95 23-Sep-97
5765 ----------------------
5766
5767 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5768 \x20 at the start of a run of normal characters. These were being treated as
5769 real characters, instead of the source characters being re-checked.
5770
5771
5772 Version 0.94 18-Sep-97
5773 ----------------------
5774
5775 1. The functions are now thread-safe, with the caveat that the global variables
5776 containing pointers to malloc() and free() or alternative functions are the
5777 same for all threads.
5778
5779 2. Get pcre_study() to generate a bitmap of initial characters for non-
5780 anchored patterns when this is possible, and use it if passed to pcre_exec().
5781
5782
5783 Version 0.93 15-Sep-97
5784 ----------------------
5785
5786 1. /(b)|(:+)/ was computing an incorrect first character.
5787
5788 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5789 but not actually doing anything yet.
5790
5791 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5792 as Perl does (e.g. [-az] or [az-]).
5793
5794 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5795 all possible positions.
5796
5797 5. Split up into different modules to avoid including unneeded functions in a
5798 compiled binary. However, compile and exec are still in one module. The "study"
5799 function is split off.
5800
5801 6. The character tables are now in a separate module whose source is generated
5802 by an auxiliary program - but can then be edited by hand if required. There are
5803 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5804 toupper() in the code.
5805
5806 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5807 make them global. Abolish the function for setting them, as the caller can now
5808 set them directly.
5809
5810
5811 Version 0.92 11-Sep-97
5812 ----------------------
5813
5814 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5815 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5816
5817 2. Caseless matching was not working in character classes if the characters in
5818 the pattern were in upper case.
5819
5820 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5821
5822 4. Make PCRE_ANCHORED public and accept as a compile option.
5823
5824 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5825 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5826 pass them.
5827
5828 6. Give an error if bad option bits passed at compile or run time.
5829
5830 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5831 pcretest to cause it to pass that flag.
5832
5833 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5834 options, and the first character, if set.
5835
5836 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5837
5838
5839 Version 0.91 10-Sep-97
5840 ----------------------
5841
5842 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5843 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5844
5845 2. PCRE was looping on encountering an indefinitely repeated back reference to
5846 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5847 Perl does - treats the match as successful.
5848
5849 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5