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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1584 - (show annotations)
Sat Aug 1 09:30:02 2015 UTC (2 days, 12 hours ago) by ph10
File size: 281927 byte(s)
Error occurred while calculating annotation data.
Give up on minimum length for overly complex patterns.
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
105 Version 8.37 28-April-2015
106 --------------------------
107
108 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
109 for those parentheses to be closed with whatever has been captured so far.
110 However, it was failing to mark any other groups between the hightest
111 capture so far and the currrent group as "unset". Thus, the ovector for
112 those groups contained whatever was previously there. An example is the
113 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
114
115 2. If an assertion condition was quantified with a minimum of zero (an odd
116 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
117
118 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
119 unrecognized modifier, a crash could occur.
120
121 4. An attempt to do global matching in pcretest with a zero-length ovector
122 caused a crash.
123
124 5. Fixed a memory leak during matching that could occur for a subpattern
125 subroutine call (recursive or otherwise) if the number of captured groups
126 that had to be saved was greater than ten.
127
128 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
129 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
130 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
131
132 7. A UTF pattern containing a "not" match of a non-ASCII character and a
133 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
134
135 8. When a pattern is compiled, it remembers the highest back reference so that
136 when matching, if the ovector is too small, extra memory can be obtained to
137 use instead. A conditional subpattern whose condition is a check on a
138 capture having happened, such as, for example in the pattern
139 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
140 setting the highest backreference number. This mattered only if pcre_exec()
141 was called with an ovector that was too small to hold the capture, and there
142 was no other kind of back reference (a situation which is probably quite
143 rare). The effect of the bug was that the condition was always treated as
144 FALSE when the capture could not be consulted, leading to a incorrect
145 behaviour by pcre_exec(). This bug has been fixed.
146
147 9. A reference to a duplicated named group (either a back reference or a test
148 for being set in a conditional) that occurred in a part of the pattern where
149 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
150 to be incorrectly calculated, leading to overwriting.
151
152 10. A mutually recursive set of back references such as (\2)(\1) caused a
153 segfault at study time (while trying to find the minimum matching length).
154 The infinite loop is now broken (with the minimum length unset, that is,
155 zero).
156
157 11. If an assertion that was used as a condition was quantified with a minimum
158 of zero, matching went wrong. In particular, if the whole group had
159 unlimited repetition and could match an empty string, a segfault was
160 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
161 assertions to be quantified, but not if they are being used as conditions,
162 so the above pattern is faulted by Perl. PCRE has now been changed so that
163 it also rejects such patterns.
164
165 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
166 failed to allow the zero-repeat case if pcre2_exec() was called with an
167 ovector too small to capture the group.
168
169 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
170 Red Hat Product Security:
171
172 (a) A crash if /K and /F were both set with the option to save the compiled
173 pattern.
174
175 (b) Another crash if the option to print captured substrings in a callout
176 was combined with setting a null ovector, for example \O\C+ as a subject
177 string.
178
179 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
180 forward reference repeated a large (but limited) number of times within a
181 repeated outer group that has a zero minimum quantifier, caused incorrect
182 code to be compiled, leading to the error "internal error:
183 previously-checked referenced subpattern not found" when an incorrect
184 memory address was read. This bug was reported as "heap overflow",
185 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
186 CVE-2015-2325.
187
188 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
189 call within a group that also contained a recursive back reference caused
190 incorrect code to be compiled. This bug was reported as "heap overflow",
191 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
192 number CVE-2015-2326.
193
194 24. Computing the size of the JIT read-only data in advance has been a source
195 of various issues, and new ones are still appear unfortunately. To fix
196 existing and future issues, size computation is eliminated from the code,
197 and replaced by on-demand memory allocation.
198
199 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
200 adjacent to the end of the range, and the range contained characters with
201 more than one other case, caused incorrect behaviour when compiled in UTF
202 mode. In that example, the range a-j was left out of the class.
203
204 26. Fix JIT compilation of conditional blocks, which assertion
205 is converted to (*FAIL). E.g: /(?(?!))/.
206
207 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
208 discovered by the LLVM fuzzer.
209
210 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
211 when this assertion was used as a condition, for example (?(?!)a|b). In
212 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
213 error about an unsupported item.
214
215 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
216 possessification code could take exponential time to complete. A recursion
217 depth limit of 1000 has been imposed to limit the resources used by this
218 optimization.
219
220 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
221 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
222 because \S ensures they are all in the class. The code for doing this was
223 interacting badly with the code for computing the amount of space needed to
224 compile the pattern, leading to a buffer overflow. This bug was discovered
225 by the LLVM fuzzer.
226
227 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
228 other kinds of group caused stack overflow at compile time. This bug was
229 discovered by the LLVM fuzzer.
230
231 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
232 between a subroutine call and its quantifier was incorrectly compiled,
233 leading to buffer overflow or other errors. This bug was discovered by the
234 LLVM fuzzer.
235
236 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
237 assertion after (?(. The code was failing to check the character after
238 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
239 was discovered by the LLVM fuzzer.
240
241 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
242 a fixed maximum following a group that contains a subroutine reference was
243 incorrectly compiled and could trigger buffer overflow. This bug was
244 discovered by the LLVM fuzzer.
245
246 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
247 caused a stack overflow instead of the diagnosis of a non-fixed length
248 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
249
250 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
251 (e.g. /(?<=\Ka)/) could make pcregrep loop.
252
253 37. There was a similar problem to 36 in pcretest for global matches.
254
255 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
256 and a subsequent item in the pattern caused a non-match, backtracking over
257 the repeated \X did not stop, but carried on past the start of the subject,
258 causing reference to random memory and/or a segfault. There were also some
259 other cases where backtracking after \C could crash. This set of bugs was
260 discovered by the LLVM fuzzer.
261
262 39. The function for finding the minimum length of a matching string could take
263 a very long time if mutual recursion was present many times in a pattern,
264 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
265 method has been implemented. This infelicity was discovered by the LLVM
266 fuzzer.
267
268 40. Static linking against the PCRE library using the pkg-config module was
269 failing on missing pthread symbols.
270
271
272 Version 8.36 26-September-2014
273 ------------------------------
274
275 1. Got rid of some compiler warnings in the C++ modules that were shown up by
276 -Wmissing-field-initializers and -Wunused-parameter.
277
278 2. The tests for quantifiers being too big (greater than 65535) were being
279 applied after reading the number, and stupidly assuming that integer
280 overflow would give a negative number. The tests are now applied as the
281 numbers are read.
282
283 3. Tidy code in pcre_exec.c where two branches that used to be different are
284 now the same.
285
286 4. The JIT compiler did not generate match limit checks for certain
287 bracketed expressions with quantifiers. This may lead to exponential
288 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
289 issue should be resolved now.
290
291 5. Fixed an issue, which occures when nested alternatives are optimized
292 with table jumps.
293
294 6. Inserted two casts and changed some ints to size_t in the light of some
295 reported 64-bit compiler warnings (Bugzilla 1477).
296
297 7. Fixed a bug concerned with zero-minimum possessive groups that could match
298 an empty string, which sometimes were behaving incorrectly in the
299 interpreter (though correctly in the JIT matcher). This pcretest input is
300 an example:
301
302 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
303 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
304
305 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
306 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
307 for an empty string was breaking the inner loop and carrying on at a lower
308 level, when possessive repeated groups should always return to a higher
309 level as they have no backtrack points in them. The empty string test now
310 occurs at the outer level.
311
312 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
313 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
314
315 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
316 doesn't).
317
318 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
319 to hold a pattern to be incorrectly computed (too small) when there were
320 named back references to duplicated names. This could cause "internal
321 error: code overflow" or "double free or corruption" or other memory
322 handling errors.
323
324 11. When named subpatterns had the same prefixes, back references could be
325 confused. For example, in this pattern:
326
327 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
328
329 the reference to 'Name' was incorrectly treated as a reference to a
330 duplicate name.
331
332 12. A pattern such as /^s?c/mi8 where the optional character has more than
333 one "other case" was incorrectly compiled such that it would only try to
334 match starting at "c".
335
336 13. When a pattern starting with \s was studied, VT was not included in the
337 list of possible starting characters; this should have been part of the
338 8.34/18 patch.
339
340 14. If a character class started [\Qx]... where x is any character, the class
341 was incorrectly terminated at the ].
342
343 15. If a pattern that started with a caseless match for a character with more
344 than one "other case" was studied, PCRE did not set up the starting code
345 unit bit map for the list of possible characters. Now it does. This is an
346 optimization improvement, not a bug fix.
347
348 16. The Unicode data tables have been updated to Unicode 7.0.0.
349
350 17. Fixed a number of memory leaks in pcregrep.
351
352 18. Avoid a compiler warning (from some compilers) for a function call with
353 a cast that removes "const" from an lvalue by using an intermediate
354 variable (to which the compiler does not object).
355
356 19. Incorrect code was compiled if a group that contained an internal recursive
357 back reference was optional (had quantifier with a minimum of zero). This
358 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
359 caused segmentation faults because of stack overflows at compile time.
360
361 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
362 group that is quantified with an indefinite repeat, caused a compile-time
363 loop which used up all the system stack and provoked a segmentation fault.
364 This was not the same bug as 19 above.
365
366 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
367 Patch by Mike Frysinger.
368
369
370 Version 8.35 04-April-2014
371 --------------------------
372
373 1. A new flag is set, when property checks are present in an XCLASS.
374 When this flag is not set, PCRE can perform certain optimizations
375 such as studying these XCLASS-es.
376
377 2. The auto-possessification of character sets were improved: a normal
378 and an extended character set can be compared now. Furthermore
379 the JIT compiler optimizes more character set checks.
380
381 3. Got rid of some compiler warnings for potentially uninitialized variables
382 that show up only when compiled with -O2.
383
384 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
385 of a match later then the end of the match. The pcretest program was not
386 handling the case sensibly - it was outputting from the start to the next
387 binary zero. It now reports this situation in a message, and outputs the
388 text from the end to the start.
389
390 5. Fast forward search is improved in JIT. Instead of the first three
391 characters, any three characters with fixed position can be searched.
392 Search order: first, last, middle.
393
394 6. Improve character range checks in JIT. Characters are read by an inprecise
395 function now, which returns with an unknown value if the character code is
396 above a certain threshold (e.g: 256). The only limitation is that the value
397 must be bigger than the threshold as well. This function is useful when
398 the characters above the threshold are handled in the same way.
399
400 7. The macros whose names start with RAWUCHAR are placeholders for a future
401 mode in which only the bottom 21 bits of 32-bit data items are used. To
402 make this more memorable for those maintaining the code, the names have
403 been changed to start with UCHAR21, and an extensive comment has been added
404 to their definition.
405
406 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
407 to the export list in Makefile.am (they were accidentally omitted from the
408 8.34 tarball).
409
410 9. The informational output from pcretest used the phrase "starting byte set"
411 which is inappropriate for the 16-bit and 32-bit libraries. As the output
412 for "first char" and "need char" really means "non-UTF-char", I've changed
413 "byte" to "char", and slightly reworded the output. The documentation about
414 these values has also been (I hope) clarified.
415
416 10. Another JIT related optimization: use table jumps for selecting the correct
417 backtracking path, when more than four alternatives are present inside a
418 bracket.
419
420 11. Empty match is not possible, when the minimum length is greater than zero,
421 and there is no \K in the pattern. JIT should avoid empty match checks in
422 such cases.
423
424 12. In a caseless character class with UCP support, when a character with more
425 than one alternative case was not the first character of a range, not all
426 the alternative cases were added to the class. For example, s and \x{17f}
427 are both alternative cases for S: the class [RST] was handled correctly,
428 but [R-T] was not.
429
430 13. The configure.ac file always checked for pthread support when JIT was
431 enabled. This is not used in Windows, so I have put this test inside a
432 check for the presence of windows.h (which was already tested for).
433
434 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
435 The algorithm provides a way to skip certain starting offsets, and usually
436 faster than linear prefix searches.
437
438 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
439 as for 'fr_FR' and 'french'. For some reason, however, it then used the
440 Windows-specific input and output files, which have 'french' screwed in.
441 So this could never have worked. One of the problems with locales is that
442 they aren't always the same. I have now updated RunTest so that it checks
443 the output of the locale test (test 3) against three different output
444 files, and it allows the test to pass if any one of them matches. With luck
445 this should make the test pass on some versions of Solaris where it was
446 failing. Because of the uncertainty, the script did not used to stop if
447 test 3 failed; it now does. If further versions of a French locale ever
448 come to light, they can now easily be added.
449
450 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
451 there was a message during ./configure, but it did not stop. This now
452 provokes an error. The invalid example in README has been corrected.
453 If a value less than the minimum is given, the minimum value has always
454 been used, but now a warning is given.
455
456 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
457 was a bug in the test system, which is now fixed. Also, the list of various
458 configurations that are tested for each release did not have one with both
459 16/32 bits and --enable-bar-anycrlf. It now does.
460
461 18. pcretest was missing "-C bsr" for displaying the \R default setting.
462
463 19. Little endian PowerPC systems are supported now by the JIT compiler.
464
465 20. The fast forward newline mechanism could enter to an infinite loop on
466 certain invalid UTF-8 input. Although we don't support these cases
467 this issue can be fixed by a performance optimization.
468
469 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
470 not take account if existing stack usage. There is now a new global
471 variable called pcre_stack_guard that can be set to point to an external
472 function to check stack availability. It is called at the start of
473 processing every parenthesized group.
474
475 22. A typo in the code meant that in ungreedy mode the max/min qualifier
476 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
477 not match "ab".
478
479 23. When UTF was disabled, the JIT program reported some incorrect compile
480 errors. These messages are silenced now.
481
482 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
483 compiler.
484
485 25. Change all the temporary files used in RunGrepTest to be different to those
486 used by RunTest so that the tests can be run simultaneously, for example by
487 "make -j check".
488
489
490 Version 8.34 15-December-2013
491 -----------------------------
492
493 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
494 executable memory. Patch inspired by Carsten Klein.
495
496 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
497 this macro is never tested and has no effect, because the work to support
498 coverage involves only compiling and linking options and special targets in
499 the Makefile. The comment in config.h implied that defining the macro would
500 enable coverage support, which is totally false. There was also support for
501 setting this macro in the CMake files (my fault, I just copied it from
502 configure). SUPPORT_GCOV has now been removed.
503
504 3. Make a small performance improvement in strlen16() and strlen32() in
505 pcretest.
506
507 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
508 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
509 these cases). There was also one in pcretest.c.
510
511 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
512
513 6. In UTF mode, the code for checking whether a group could match an empty
514 string (which is used for indefinitely repeated groups to allow for
515 breaking an infinite loop) was broken when the group contained a repeated
516 negated single-character class with a character that occupied more than one
517 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
518 UTF-8 mode). The effect was undefined: the group might or might not be
519 deemed as matching an empty string, or the program might have crashed.
520
521 7. The code for checking whether a group could match an empty string was not
522 recognizing that \h, \H, \v, \V, and \R must match a character.
523
524 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
525 an empty string. If it can, pcretest shows this in its information output.
526
527 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
528 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
529 matched by pcre_exec() without using JIT:
530
531 (a) If the rest of the pattern did not match after a maximal run of
532 grapheme clusters, the code for backing up to try with fewer of them
533 did not always back up over a full grapheme when characters that do not
534 have the modifier quality were involved, e.g. Hangul syllables.
535
536 (b) If the match point in a subject started with modifier character, and
537 there was no match, the code could incorrectly back up beyond the match
538 point, and potentially beyond the first character in the subject,
539 leading to a segfault or an incorrect match result.
540
541 10. A conditional group with an assertion condition could lead to PCRE
542 recording an incorrect first data item for a match if no other first data
543 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
544 first data item, and therefore matched "ca" after "c" instead of at the
545 start.
546
547 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
548 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
549
550 12. The source of pcregrep now includes z/OS-specific code so that it can be
551 compiled for z/OS as part of the special z/OS distribution.
552
553 13. Added the -T and -TM options to pcretest.
554
555 14. The code in pcre_compile.c for creating the table of named capturing groups
556 has been refactored. Instead of creating the table dynamically during the
557 actual compiling pass, the information is remembered during the pre-compile
558 pass (on the stack unless there are more than 20 named groups, in which
559 case malloc() is used) and the whole table is created before the actual
560 compile happens. This has simplified the code (it is now nearly 150 lines
561 shorter) and prepared the way for better handling of references to groups
562 with duplicate names.
563
564 15. A back reference to a named subpattern when there is more than one of the
565 same name now checks them in the order in which they appear in the pattern.
566 The first one that is set is used for the reference. Previously only the
567 first one was inspected. This change makes PCRE more compatible with Perl.
568
569 16. Unicode character properties were updated from Unicode 6.3.0.
570
571 17. The compile-time code for auto-possessification has been refactored, based
572 on a patch by Zoltan Herczeg. It now happens after instead of during
573 compilation. The code is cleaner, and more cases are handled. The option
574 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
575 options in pcretest are provided to set it. It can also be set by
576 (*NO_AUTO_POSSESS) at the start of a pattern.
577
578 18. The character VT has been added to the default ("C" locale) set of
579 characters that match \s and are generally treated as white space,
580 following this same change in Perl 5.18. There is now no difference between
581 "Perl space" and "POSIX space". Whether VT is treated as white space in
582 other locales depends on the locale.
583
584 19. The code for checking named groups as conditions, either for being set or
585 for being recursed, has been refactored (this is related to 14 and 15
586 above). Processing unduplicated named groups should now be as fast at
587 numerical groups, and processing duplicated groups should be faster than
588 before.
589
590 20. Two patches to the CMake build system, by Alexander Barkov:
591
592 (1) Replace the "source" command by "." in CMakeLists.txt because
593 "source" is a bash-ism.
594
595 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
596 without these the CMake build does not work on Solaris.
597
598 21. Perl has changed its handling of \8 and \9. If there is no previously
599 encountered capturing group of those numbers, they are treated as the
600 literal characters 8 and 9 instead of a binary zero followed by the
601 literals. PCRE now does the same.
602
603 22. Following Perl, added \o{} to specify codepoints in octal, making it
604 possible to specify values greater than 0777 and also making them
605 unambiguous.
606
607 23. Perl now gives an error for missing closing braces after \x{... instead of
608 treating the string as literal. PCRE now does the same.
609
610 24. RunTest used to grumble if an inappropriate test was selected explicitly,
611 but just skip it when running all tests. This make it awkward to run ranges
612 of tests when one of them was inappropriate. Now it just skips any
613 inappropriate tests, as it always did when running all tests.
614
615 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
616 character types such as \d or \w, too many callouts were inserted, and the
617 data that they returned was rubbish.
618
619 26. In UCP mode, \s was not matching two of the characters that Perl matches,
620 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
621 were matched by \h. The code has now been refactored so that the lists of
622 the horizontal and vertical whitespace characters used for \h and \v (which
623 are defined only in one place) are now also used for \s.
624
625 27. Add JIT support for the 64 bit TileGX architecture.
626 Patch by Jiong Wang (Tilera Corporation).
627
628 28. Possessive quantifiers for classes (both explicit and automatically
629 generated) now use special opcodes instead of wrapping in ONCE brackets.
630
631 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
632 (because it's meaningless), this was not happening when PCRE_CASELESS was
633 set. Not wrong, but inefficient.
634
635 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
636 properties for \w, \d, etc) is present in a test regex. Otherwise if the
637 test contains no characters greater than 255, Perl doesn't realise it
638 should be using Unicode semantics.
639
640 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
641 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
642 does in Unicode mode.
643
644 32. Added the "forbid" facility to pcretest so that putting tests into the
645 wrong test files can sometimes be quickly detected.
646
647 33. There is now a limit (default 250) on the depth of nesting of parentheses.
648 This limit is imposed to control the amount of system stack used at compile
649 time. It can be changed at build time by --with-parens-nest-limit=xxx or
650 the equivalent in CMake.
651
652 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
653 errors. Perl warns for these when in warning mode, but PCRE has no facility
654 for giving warnings.
655
656 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
657 However, this was not working for (?!) because it is optimized to (*FAIL),
658 for which PCRE does not allow quantifiers. The optimization is now disabled
659 when a quantifier follows (?!). I can't see any use for this, but it makes
660 things uniform.
661
662 36. Perl no longer allows group names to start with digits, so I have made this
663 change also in PCRE. It simplifies the code a bit.
664
665 37. In extended mode, Perl ignores spaces before a + that indicates a
666 possessive quantifier. PCRE allowed a space before the quantifier, but not
667 before the possessive +. It now does.
668
669 38. The use of \K (reset reported match start) within a repeated possessive
670 group such as (a\Kb)*+ was not working.
671
672 40. Document that the same character tables must be used at compile time and
673 run time, and that the facility to pass tables to pcre_exec() and
674 pcre_dfa_exec() is for use only with saved/restored patterns.
675
676 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
677 features for Builds with MSVC:
678
679 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
680 builds). The .rc files can be used to set FileDescription and many other
681 attributes.
682
683 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
684 This allows higher-level build scripts which want .pdb files to avoid
685 hard-coding the exact files needed."
686
687 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
688 mean "start of word" and "end of word", respectively, as a transition aid.
689
690 43. A minimizing repeat of a class containing codepoints greater than 255 in
691 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
692 compiled to use the heap for recursion.
693
694 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
695 is configured.
696
697
698 Version 8.33 28-May-2013
699 ------------------------
700
701 1. Added 'U' to some constants that are compared to unsigned integers, to
702 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
703 variables that are added to signed variables, to ensure the result is
704 signed and can be negated.
705
706 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
707 CMake config files.
708
709 3. Revise the creation of config.h.generic so that all boolean macros are
710 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
711 overriding via -D on the command line possible.
712
713 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
714 to unsigned int is reported to make a quite noticeable speed difference in
715 a specific Windows environment. Testing on Linux did also appear to show
716 some benefit (and it is clearly not harmful). Also fixed the definition of
717 Xop which should be unsigned.
718
719 5. Related to (4), changing the definition of the intermediate variable cc
720 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
721 improvements.
722
723 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
724 unnecessary spaces.
725
726 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
727 and later.
728
729 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
730
731 9. Optimizing fast_forward_start_bits in JIT.
732
733 10. Adding support for callouts in JIT, and fixing some issues revealed
734 during this work. Namely:
735
736 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
737
738 (b) Minimum length was not checked before the matching is started.
739
740 11. The value of capture_last that is passed to callouts was incorrect in some
741 cases when there was a capture on one path that was subsequently abandoned
742 after a backtrack. Also, the capture_last value is now reset after a
743 recursion, since all captures are also reset in this case.
744
745 12. The interpreter no longer returns the "too many substrings" error in the
746 case when an overflowing capture is in a branch that is subsequently
747 abandoned after a backtrack.
748
749 13. In the pathological case when an offset vector of size 2 is used, pcretest
750 now prints out the matched string after a yield of 0 or 1.
751
752 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
753 Only supported by the JIT compiler at the moment.
754
755 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
756
757 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
758 offset of the starting point of the matching process, provided the offsets
759 vector is large enough.
760
761 17. The \A escape now records a lookbehind value of 1, though its execution
762 does not actually inspect the previous character. This is to ensure that,
763 in partial multi-segment matching, at least one character from the old
764 segment is retained when a new segment is processed. Otherwise, if there
765 are no lookbehinds in the pattern, \A might match incorrectly at the start
766 of a new segment.
767
768 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
769
770 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
771 gives some modest performance improvement in 8-bit mode.
772
773 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
774 be expressed in certain programming languages using Universal Character
775 Names.
776
777 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
778 which points out that "non characters" are not "characters that may not
779 appear in Unicode strings" but rather "characters that are reserved for
780 internal use and have only local meaning".
781
782 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
783 there was a conditional group that depended on an assertion, if the
784 assertion was false, the callout that immediately followed the alternation
785 in the condition was skipped when pcre_exec() was used for matching.
786
787 23. Allow an explicit callout to be inserted before an assertion that is the
788 condition for a conditional group, for compatibility with automatic
789 callouts, which always insert a callout at this point.
790
791 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
792 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
793
794 25. (*PRUNE) is now supported by the JIT compiler.
795
796 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
797
798 27. Fix the case where there are two or more SKIPs with arguments that may be
799 ignored.
800
801 28. (*SKIP) is now supported by the JIT compiler.
802
803 29. (*THEN) is now supported by the JIT compiler.
804
805 30. Update RunTest with additional test selector options.
806
807 31. The way PCRE handles backtracking verbs has been changed in two ways.
808
809 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
810 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
811 backtracking. In some cases this makes it more Perl-compatible, but Perl's
812 rather obscure rules do not always do the same thing.
813
814 (2) Previously, backtracking verbs were confined within assertions. This is
815 no longer the case for positive assertions, except for (*ACCEPT). Again,
816 this sometimes improves Perl compatibility, and sometimes does not.
817
818 32. A number of tests that were in test 2 because Perl did things differently
819 have been moved to test 1, because either Perl or PCRE has changed, and
820 these tests are now compatible.
821
822 32. Backtracking control verbs are now handled in the same way in JIT and
823 interpreter.
824
825 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
826 contained a forward subroutine reference caused a compile error.
827
828 34. Auto-detect and optimize limited repetitions in JIT.
829
830 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
831 blocking (*UTF) etc.
832
833 36. In the interpreter, maximizing pattern repetitions for characters and
834 character types now use tail recursion, which reduces stack usage.
835
836 37. The value of the max lookbehind was not correctly preserved if a compiled
837 and saved regex was reloaded on a host of different endianness.
838
839 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
840 of the compiled pattern block, expand the flags field from 16 to 32 bits
841 because it was almost full.
842
843 39. Try madvise first before posix_madvise.
844
845 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
846 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
847 original change locked out all matches of empty strings. This has been
848 changed so that one match of an empty string per line is recognized.
849 Subsequent searches on the same line (for colouring or for --only-matching,
850 for example) do not recognize empty strings.
851
852 41. Applied a user patch to fix a number of spelling mistakes in comments.
853
854 42. Data lines longer than 65536 caused pcretest to crash.
855
856 43. Clarified the data type for length and startoffset arguments for pcre_exec
857 and pcre_dfa_exec in the function-specific man pages, where they were
858 explicitly stated to be in bytes, never having been updated. I also added
859 some clarification to the pcreapi man page.
860
861 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
862 a segmentation fault.
863
864
865 Version 8.32 30-November-2012
866 -----------------------------
867
868 1. Improved JIT compiler optimizations for first character search and single
869 character iterators.
870
871 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
872 Patch by Daniel Richard G.
873
874 3. Single character iterator optimizations in the JIT compiler.
875
876 4. Improved JIT compiler optimizations for character ranges.
877
878 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
879 Reported by Giuseppe D'Angelo.
880
881 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
882 of a line, was being set incorrectly in cases where .* appeared inside
883 atomic brackets at the start of a pattern, or where there was a subsequent
884 *PRUNE or *SKIP.
885
886 7. Improved instruction cache flush for POWER/PowerPC.
887 Patch by Daniel Richard G.
888
889 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
890 grep:
891
892 (a) There is now no limit to the number of patterns to be matched.
893
894 (b) An error is given if a pattern is too long.
895
896 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
897 are now supported.
898
899 (d) --exclude-from and --include-from (multiple use) have been added.
900
901 (e) Exclusions and inclusions now apply to all files and directories, not
902 just to those obtained from scanning a directory recursively.
903
904 (f) Multiple uses of -f and --file-list are now supported.
905
906 (g) In a Windows environment, the default for -d has been changed from
907 "read" (the GNU grep default) to "skip", because otherwise the presence
908 of a directory in the file list provokes an error.
909
910 (h) The documentation has been revised and clarified in places.
911
912 9. Improve the matching speed of capturing brackets.
913
914 10. Changed the meaning of \X so that it now matches a Unicode extended
915 grapheme cluster.
916
917 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
918 out POSIX threads when JIT support is configured.
919
920 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
921
922 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
923 structure could go wrong in environments where size_t is not the same size
924 as int.
925
926 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
927 set.
928
929 15. The EBCDIC support had decayed; later updates to the code had included
930 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
931 general tidy up of EBCDIC-related issues, and the documentation was also
932 not quite right. There is now a test that can be run on ASCII systems to
933 check some of the EBCDIC-related things (but is it not a full test).
934
935 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
936 in a small tidy to the code.
937
938 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
939
940 18. If the --only-matching (-o) option in pcregrep is specified multiple
941 times, each one causes appropriate output. For example, -o1 -o2 outputs the
942 substrings matched by the 1st and 2nd capturing parentheses. A separating
943 string can be specified by --om-separator (default empty).
944
945 19. Improving the first n character searches.
946
947 20. Turn case lists for horizontal and vertical white space into macros so that
948 they are defined only once.
949
950 21. This set of changes together give more compatible Unicode case-folding
951 behaviour for characters that have more than one other case when UCP
952 support is available.
953
954 (a) The Unicode property table now has offsets into a new table of sets of
955 three or more characters that are case-equivalent. The MultiStage2.py
956 script that generates these tables (the pcre_ucd.c file) now scans
957 CaseFolding.txt instead of UnicodeData.txt for character case
958 information.
959
960 (b) The code for adding characters or ranges of characters to a character
961 class has been abstracted into a generalized function that also handles
962 case-independence. In UTF-mode with UCP support, this uses the new data
963 to handle characters with more than one other case.
964
965 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
966 whose other case is greater than 256 are now correctly matched
967 caselessly. Previously, the high codepoint matched the low one, but not
968 vice versa.
969
970 (d) The processing of \h, \H, \v, and \ in character classes now makes use
971 of the new class addition function, using character lists defined as
972 macros alongside the case definitions of 20 above.
973
974 (e) Caseless back references now work with characters that have more than
975 one other case.
976
977 (f) General caseless matching of characters with more than one other case
978 is supported.
979
980 22. Unicode character properties were updated from Unicode 6.2.0
981
982 23. Improved CMake support under Windows. Patch by Daniel Richard G.
983
984 24. Add support for 32-bit character strings, and UTF-32
985
986 25. Major JIT compiler update (code refactoring and bugfixing).
987 Experimental Sparc 32 support is added.
988
989 26. Applied a modified version of Daniel Richard G's patch to create
990 pcre.h.generic and config.h.generic by "make" instead of in the
991 PrepareRelease script.
992
993 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
994 pcre_compile.c when checking for a zero character.
995
996 28. Introducing a native interface for JIT. Through this interface, the compiled
997 machine code can be directly executed. The purpose of this interface is to
998 provide fast pattern matching, so several sanity checks are not performed.
999 However, feature tests are still performed. The new interface provides
1000 1.4x speedup compared to the old one.
1001
1002 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1003 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1004 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1005
1006 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1007 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1008 the "old" RFC 2279). Instead, it ended up passing a negative length to
1009 pcre_exec().
1010
1011 31. Add support for GCC's visibility feature to hide internal functions.
1012
1013 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1014 "unknown -C option" after outputting 0 or 1.
1015
1016 33. There is now support for generating a code coverage report for the test
1017 suite in environments where gcc is the compiler and lcov is installed. This
1018 is mainly for the benefit of the developers.
1019
1020 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1021 unaddressable using valgrind annotations, allowing valgrind to detect
1022 invalid memory accesses. This is mainly for the benefit of the developers.
1023
1024 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1025
1026 26. Give configure error if --enable-cpp but no C++ compiler found.
1027
1028
1029 Version 8.31 06-July-2012
1030 -------------------------
1031
1032 1. Fixing a wrong JIT test case and some compiler warnings.
1033
1034 2. Removed a bashism from the RunTest script.
1035
1036 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1037 to unsigned type, result still unsigned" that was given by an MS compiler
1038 on encountering the code "-sizeof(xxx)".
1039
1040 4. Partial matching support is added to the JIT compiler.
1041
1042 5. Fixed several bugs concerned with partial matching of items that consist
1043 of more than one character:
1044
1045 (a) /^(..)\1/ did not partially match "aba" because checking references was
1046 done on an "all or nothing" basis. This also applied to repeated
1047 references.
1048
1049 (b) \R did not give a hard partial match if \r was found at the end of the
1050 subject.
1051
1052 (c) \X did not give a hard partial match after matching one or more
1053 characters at the end of the subject.
1054
1055 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1056 a partial match for the string "\r".
1057
1058 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1059 a partial match for a CR character at the end of the subject string.
1060
1061 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1062 running pcretest, the text "(JIT)" added to the output whenever JIT is
1063 actually used to run the match.
1064
1065 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1066 or /S+[+] with a digit between 1 and 7.
1067
1068 8. OP_NOT now supports any UTF character not just single-byte ones.
1069
1070 9. (*MARK) control verb is now supported by the JIT compiler.
1071
1072 10. The command "./RunTest list" lists the available tests without actually
1073 running any of them. (Because I keep forgetting what they all are.)
1074
1075 11. Add PCRE_INFO_MAXLOOKBEHIND.
1076
1077 12. Applied a (slightly modified) user-supplied patch that improves performance
1078 when the heap is used for recursion (compiled with --disable-stack-for-
1079 recursion). Instead of malloc and free for each heap frame each time a
1080 logical recursion happens, frames are retained on a chain and re-used where
1081 possible. This sometimes gives as much as 30% improvement.
1082
1083 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1084 call.
1085
1086 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1087
1088 15. It is now possible to link pcretest with libedit as an alternative to
1089 libreadline.
1090
1091 16. (*COMMIT) control verb is now supported by the JIT compiler.
1092
1093 17. The Unicode data tables have been updated to Unicode 6.1.0.
1094
1095 18. Added --file-list option to pcregrep.
1096
1097 19. Added binary file support to pcregrep, including the -a, --binary-files,
1098 -I, and --text options.
1099
1100 20. The madvise function is renamed for posix_madvise for QNX compatibility
1101 reasons. Fixed by Giuseppe D'Angelo.
1102
1103 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1104
1105 22. Changed the option for creating symbolic links for 16-bit man pages from
1106 -s to -sf so that re-installing does not cause issues.
1107
1108 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1109
1110 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1111 matches in certain environments (the workspace was not being correctly
1112 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1113 some of the workspace data at the beginning of a restart.
1114
1115 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1116 was not doing so when it should - probably a typo introduced by SVN 528
1117 (change 8.10/14).
1118
1119 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1120 \w+ when the character tables indicated that \x{c4} was a word character.
1121 There were several related cases, all because the tests for doing a table
1122 lookup were testing for characters less than 127 instead of 255.
1123
1124 27. If a pattern contains capturing parentheses that are not used in a match,
1125 their slots in the ovector are set to -1. For those that are higher than
1126 any matched groups, this happens at the end of processing. In the case when
1127 there were back references that the ovector was too small to contain
1128 (causing temporary malloc'd memory to be used during matching), and the
1129 highest capturing number was not used, memory off the end of the ovector
1130 was incorrectly being set to -1. (It was using the size of the temporary
1131 memory instead of the true size.)
1132
1133 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1134 ovector size, it uses memory at the end of the block that it has got.
1135
1136 29. Check for an overlong MARK name and give an error at compile time. The
1137 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1138
1139 30. JIT compiler update.
1140
1141 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1142 Rill for the patch.
1143
1144 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1145
1146 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1147
1148 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1149 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1150 was enabled.
1151
1152 35. Improve JIT code generation for greedy plus quantifier.
1153
1154 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1155 1 to "aa" instead of to an empty string. The bug affected repeated groups
1156 that could potentially match an empty string.
1157
1158 37. Optimizing single character iterators in JIT.
1159
1160 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1161 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1162 codepoints that are too big for the mode are faulted, and in a UTF mode,
1163 disallowed codepoints are also faulted.
1164
1165 39. If PCRE was compiled with UTF support, in three places in the DFA
1166 matcher there was code that should only have been obeyed in UTF mode, but
1167 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1168 processing when bytes with values greater than 127 were present. In 16-bit
1169 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1170 both cases the values are those that cannot be the first data item in a UTF
1171 character. The three items that might have provoked this were recursions,
1172 possessively repeated groups, and atomic groups.
1173
1174 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1175 and pcregrep, because some OS require shared objects to be explicitly
1176 passed to ld, causing the link step to fail if they are not.
1177
1178 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1179 mode, patterns that started with \h* or \R* might be incorrectly matched.
1180
1181
1182 Version 8.30 04-February-2012
1183 -----------------------------
1184
1185 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1186 name is defined in ctype.h.
1187
1188 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1189 only in quite long subpatterns.
1190
1191 3. Removed the function pcre_info(), which has been obsolete and deprecated
1192 since it was replaced by pcre_fullinfo() in February 2000.
1193
1194 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1195 match a (*MARK), and the match failed at the start of the subject, a
1196 reference to memory before the start of the subject could occur. This bug
1197 was introduced by fix 17 of release 8.21.
1198
1199 5. A reference to an unset group with zero minimum repetition was giving
1200 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1201 /(another)?(\1?)test/ matched against "hello world test". This bug was
1202 introduced in release 8.13.
1203
1204 6. Add support for 16-bit character strings (a large amount of work involving
1205 many changes and refactorings).
1206
1207 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1208 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1209 from a file.
1210
1211 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1212 rounding is not applied in this particular case).
1213
1214 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1215 if they appear, or are escaped, in patterns.
1216
1217 10. Get rid of a number of -Wunused-but-set-variable warnings.
1218
1219 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1220 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1221 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1222 also returns the mark "x". This bug applied to capturing parentheses,
1223 non-capturing parentheses, and atomic parentheses. It also applied to some
1224 assertions.
1225
1226 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1227 information out of configure.ac instead of relying on pcre.h.generic, which
1228 is not stored in the repository.
1229
1230 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1231 -lreadline.
1232
1233 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1234
1235 15. Applied Graycode's patch to put the top-level frame on the stack rather
1236 than the heap when not using the stack for recursion. This gives a
1237 performance improvement in many cases when recursion is not deep.
1238
1239 16. Experimental code added to "pcretest -C" to output the stack frame size.
1240
1241
1242 Version 8.21 12-Dec-2011
1243 ------------------------
1244
1245 1. Updating the JIT compiler.
1246
1247 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1248 are added as well.
1249
1250 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1251 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1252 calling _pcre_jit_exec. Some extra comments are added.
1253
1254 4. (*MARK) settings inside atomic groups that do not contain any capturing
1255 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1256 was introduced by change 18 for 8.20.
1257
1258 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1259 ECMA-262 standard.
1260
1261 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1262 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1263 This bug was probably introduced by change 9 of 8.13.
1264
1265 7. While fixing 6 above, I noticed that a number of other items were being
1266 incorrectly rejected as "not fixed length". This arose partly because newer
1267 opcodes had not been added to the fixed-length checking code. I have (a)
1268 corrected the bug and added tests for these items, and (b) arranged for an
1269 error to occur if an unknown opcode is encountered while checking for fixed
1270 length instead of just assuming "not fixed length". The items that were
1271 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1272 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1273 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1274
1275 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1276 being incorrectly compiled and would have given unpredicatble results.
1277
1278 9. A possessively repeated subpattern with minimum repeat count greater than
1279 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1280 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1281 the first (A) could occur when it should not.
1282
1283 10. Add a cast and remove a redundant test from the code.
1284
1285 11. JIT should use pcre_malloc/pcre_free for allocation.
1286
1287 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1288 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1289 is anything other than /usr, -L is still shown).
1290
1291 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1292
1293 14. Perl does not support \N without a following name in a [] class; PCRE now
1294 also gives an error.
1295
1296 15. If a forward reference was repeated with an upper limit of around 2000,
1297 it caused the error "internal error: overran compiling workspace". The
1298 maximum number of forward references (including repeats) was limited by the
1299 internal workspace, and dependent on the LINK_SIZE. The code has been
1300 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1301 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1302 of around 200,000 forward references. While doing this, I also speeded up
1303 the filling in of repeated forward references.
1304
1305 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1306 incorrectly expecting the subject to contain another "a" after the start.
1307
1308 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1309 in the match, the SKIP should be ignored. This was not happening; instead
1310 the SKIP was being treated as NOMATCH. For patterns such as
1311 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1312 tested.
1313
1314 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1315 now much more compatible with Perl, in particular in cases where the result
1316 is a non-match for a non-anchored pattern. For example, if
1317 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1318 "m", where previously it did not return a name. A side effect of this
1319 change is that for partial matches, the last encountered mark name is
1320 returned, as for non matches. A number of tests that were previously not
1321 Perl-compatible have been moved into the Perl-compatible test files. The
1322 refactoring has had the pleasing side effect of removing one argument from
1323 the match() function, thus reducing its stack requirements.
1324
1325 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1326 subsequent uses of /S (without +) incorrectly behaved like /S+.
1327
1328 21. Retrieve executable code size support for the JIT compiler and fixing
1329 some warnings.
1330
1331 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1332 not work when the shorter character appeared right at the end of the
1333 subject string.
1334
1335 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1336 systems.
1337
1338 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1339 output it when the /M option is used in pcretest.
1340
1341 25. The CheckMan script was not being included in the distribution. Also, added
1342 an explicit "perl" to run Perl scripts from the PrepareRelease script
1343 because this is reportedly needed in Windows.
1344
1345 26. If study data was being save in a file and studying had not found a set of
1346 "starts with" bytes for the pattern, the data written to the file (though
1347 never used) was taken from uninitialized memory and so caused valgrind to
1348 complain.
1349
1350 27. Updated RunTest.bat as provided by Sheri Pierce.
1351
1352 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1353
1354 29. Computation of memory usage for the table of capturing group names was
1355 giving an unnecessarily large value.
1356
1357
1358 Version 8.20 21-Oct-2011
1359 ------------------------
1360
1361 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1362 a POSIX class. After further experiments with Perl, which convinced me that
1363 Perl has bugs and confusions, a closing square bracket is no longer allowed
1364 in a POSIX name. This bug also affected patterns with classes that started
1365 with full stops.
1366
1367 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1368 captured substring, but while checking the failing first alternative,
1369 substring 1 is temporarily captured. If the output vector supplied to
1370 pcre_exec() was not big enough for this capture, the yield of the function
1371 was still zero ("insufficient space for captured substrings"). This cannot
1372 be totally fixed without adding another stack variable, which seems a lot
1373 of expense for a edge case. However, I have improved the situation in cases
1374 such as /(a)(b)x|abc/ matched against "abc", where the return code
1375 indicates that fewer than the maximum number of slots in the ovector have
1376 been set.
1377
1378 3. Related to (2) above: when there are more back references in a pattern than
1379 slots in the output vector, pcre_exec() uses temporary memory during
1380 matching, and copies in the captures as far as possible afterwards. It was
1381 using the entire output vector, but this conflicts with the specification
1382 that only 2/3 is used for passing back captured substrings. Now it uses
1383 only the first 2/3, for compatibility. This is, of course, another edge
1384 case.
1385
1386 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1387 main code base, and can be used by building with --enable-jit. When this is
1388 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1389 runtime --no-jit option is given.
1390
1391 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1392 ovector, the return from the function was zero, implying that there were
1393 other matches that did not fit. The correct "exactly full" value is now
1394 returned.
1395
1396 6. If a subpattern that was called recursively or as a subroutine contained
1397 (*PRUNE) or any other control that caused it to give a non-standard return,
1398 invalid errors such as "Error -26 (nested recursion at the same subject
1399 position)" or even infinite loops could occur.
1400
1401 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1402 computing the minimum length on reaching *ACCEPT, and so ended up with the
1403 wrong value of 1 rather than 0. Further investigation indicates that
1404 computing a minimum subject length in the presence of *ACCEPT is difficult
1405 (think back references, subroutine calls), and so I have changed the code
1406 so that no minimum is registered for a pattern that contains *ACCEPT.
1407
1408 8. If (*THEN) was present in the first (true) branch of a conditional group,
1409 it was not handled as intended. [But see 16 below.]
1410
1411 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1412 Sheri Pierce.
1413
1414 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1415 the first byte in a match must be "a".
1416
1417 11. Change 17 for 8.13 increased the recursion depth for patterns like
1418 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1419 pattern contains any instances of (*THEN). If it does not, the old
1420 optimizations are restored. It would be nice to do this on a per-group
1421 basis, but at the moment that is not feasible.
1422
1423 12. In some environments, the output of pcretest -C is CRLF terminated. This
1424 broke RunTest's code that checks for the link size. A single white space
1425 character after the value is now allowed for.
1426
1427 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1428 For "fr", it uses the Windows-specific input and output files.
1429
1430 14. If (*THEN) appeared in a group that was called recursively or as a
1431 subroutine, it did not work as intended. [But see next item.]
1432
1433 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1434 pattern fragments (but not containing any | characters). If A and B are
1435 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1436 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1437 D. In other words, Perl considers parentheses that do not contain any |
1438 characters to be part of a surrounding alternative, whereas PCRE was
1439 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1440 differently. PCRE now behaves in the same way as Perl, except in the case
1441 of subroutine/recursion calls such as (?1) which have in any case always
1442 been different (but PCRE had them first :-).
1443
1444 16. Related to 15 above: Perl does not treat the | in a conditional group as
1445 creating alternatives. Such a group is treated in the same way as an
1446 ordinary group without any | characters when processing (*THEN). PCRE has
1447 been changed to match Perl's behaviour.
1448
1449 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1450 RunGrepTest script failed.
1451
1452 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1453 inevitable for groups that contain captures, but it can lead to a lot of
1454 stack use in large patterns. The old behaviour has been restored for atomic
1455 groups that do not contain any capturing parentheses.
1456
1457 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1458 suppress the check for a minimum subject length at run time. (If it was
1459 given to pcre_exec() or pcre_dfa_exec() it did work.)
1460
1461 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1462 fail to work when decoding hex characters in data strings in EBCDIC
1463 environments.
1464
1465 21. It appears that in at least one Mac OS environment, the isxdigit() function
1466 is implemented as a macro that evaluates to its argument more than once,
1467 contravening the C 90 Standard (I haven't checked a later standard). There
1468 was an instance in pcretest which caused it to go wrong when processing
1469 \x{...} escapes in subject strings. The has been rewritten to avoid using
1470 things like p++ in the argument of isxdigit().
1471
1472
1473 Version 8.13 16-Aug-2011
1474 ------------------------
1475
1476 1. The Unicode data tables have been updated to Unicode 6.0.0.
1477
1478 2. Two minor typos in pcre_internal.h have been fixed.
1479
1480 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1481 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1482 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1483
1484 4. There were a number of related bugs in the code for matching backrefences
1485 caselessly in UTF-8 mode when codes for the characters concerned were
1486 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1487 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1488 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1489 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1490 2-byte code at the end of the subject (it thought there wasn't enough data
1491 left).
1492
1493 5. Comprehensive information about what went wrong is now returned by
1494 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1495 as the output vector has at least 2 elements. The offset of the start of
1496 the failing character and a reason code are placed in the vector.
1497
1498 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1499 now returned is for the first byte of the failing character, instead of the
1500 last byte inspected. This is an incompatible change, but I hope it is small
1501 enough not to be a problem. It makes the returned offset consistent with
1502 pcre_exec() and pcre_dfa_exec().
1503
1504 7. pcretest now gives a text phrase as well as the error number when
1505 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1506 failure, the offset and reason code are output.
1507
1508 8. When \R was used with a maximizing quantifier it failed to skip backwards
1509 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1510 back over a single character (\n). This seems wrong (because it treated the
1511 two characters as a single entity when going forwards), conflicts with the
1512 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1513 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1514 has been changed.
1515
1516 9. Some internal refactoring has changed the processing so that the handling
1517 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1518 time (the PCRE_DOTALL option was changed this way some time ago: version
1519 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1520 which was always a bit of a fudge. It also means that there is one less
1521 argument for the match() function, which reduces its stack requirements
1522 slightly. This change also fixes an incompatibility with Perl: the pattern
1523 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1524
1525 10. More internal refactoring has drastically reduced the number of recursive
1526 calls to match() for possessively repeated groups such as (abc)++ when
1527 using pcre_exec().
1528
1529 11. While implementing 10, a number of bugs in the handling of groups were
1530 discovered and fixed:
1531
1532 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1533 (a|)*(?1) gave a compile-time internal error.
1534 ((a|)+)+ did not notice that the outer group could match an empty string.
1535 (^a|^)+ was not marked as anchored.
1536 (.*a|.*)+ was not marked as matching at start or after a newline.
1537
1538 12. Yet more internal refactoring has removed another argument from the match()
1539 function. Special calls to this function are now indicated by setting a
1540 value in a variable in the "match data" data block.
1541
1542 13. Be more explicit in pcre_study() instead of relying on "default" for
1543 opcodes that mean there is no starting character; this means that when new
1544 ones are added and accidentally left out of pcre_study(), testing should
1545 pick them up.
1546
1547 14. The -s option of pcretest has been documented for ages as being an old
1548 synonym of -m (show memory usage). I have changed it to mean "force study
1549 for every regex", that is, assume /S for every regex. This is similar to -i
1550 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1551 using it. It makes it easier to run collections of tests with and without
1552 study enabled, and thereby test pcre_study() more easily. All the standard
1553 tests are now run with and without -s (but some patterns can be marked as
1554 "never study" - see 20 below).
1555
1556 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1557 restoration of the capturing data to the outer values was not happening
1558 correctly.
1559
1560 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1561 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1562 pattern had matched an empty string, and so incorrectly returned a no
1563 match.
1564
1565 17. There was optimizing code for the last branch of non-capturing parentheses,
1566 and also for the obeyed branch of a conditional subexpression, which used
1567 tail recursion to cut down on stack usage. Unfortunately, now that there is
1568 the possibility of (*THEN) occurring in these branches, tail recursion is
1569 no longer possible because the return has to be checked for (*THEN). These
1570 two optimizations have therefore been removed. [But see 8.20/11 above.]
1571
1572 18. If a pattern containing \R was studied, it was assumed that \R always
1573 matched two bytes, thus causing the minimum subject length to be
1574 incorrectly computed because \R can also match just one byte.
1575
1576 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1577 was incorrectly computed.
1578
1579 20. If /S is present twice on a test pattern in pcretest input, it now
1580 *disables* studying, thereby overriding the use of -s on the command line
1581 (see 14 above). This is necessary for one or two tests to keep the output
1582 identical in both cases.
1583
1584 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1585 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1586
1587 22. When an atomic group that contained a capturing parenthesis was
1588 successfully matched, but the branch in which it appeared failed, the
1589 capturing was not being forgotten if a higher numbered group was later
1590 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1591 group 1 to "a", when in fact it should be unset. This applied to multi-
1592 branched capturing and non-capturing groups, repeated or not, and also to
1593 positive assertions (capturing in negative assertions does not happen
1594 in PCRE) and also to nested atomic groups.
1595
1596 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1597 subject after a captured substring, to make it easier to tell which of a
1598 number of identical substrings has been captured.
1599
1600 24. The way atomic groups are processed by pcre_exec() has been changed so that
1601 if they are repeated, backtracking one repetition now resets captured
1602 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1603 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1604 "aaa". Previously, it would have been "a". As part of this code
1605 refactoring, the way recursive calls are handled has also been changed.
1606
1607 25. If an assertion condition captured any substrings, they were not passed
1608 back unless some other capturing happened later. For example, if
1609 (?(?=(a))a) was matched against "a", no capturing was returned.
1610
1611 26. When studying a pattern that contained subroutine calls or assertions,
1612 the code for finding the minimum length of a possible match was handling
1613 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1614 group 1 called group 2 while simultaneously a separate group 2 called group
1615 1). A stack overflow occurred in this case. I have fixed this by limiting
1616 the recursion depth to 10.
1617
1618 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1619 Fortmann. This supports explicit test numbers on the command line, and has
1620 argument validation and error reporting.
1621
1622 28. An instance of \X with an unlimited repeat could fail if at any point the
1623 first character it looked at was a mark character.
1624
1625 29. Some minor code refactoring concerning Unicode properties and scripts
1626 should reduce the stack requirement of match() slightly.
1627
1628 30. Added the '=' option to pcretest to check the setting of unused capturing
1629 slots at the end of the pattern, which are documented as being -1, but are
1630 not included in the return count.
1631
1632 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1633 compiled something random. Now it gives a compile-time error (as does
1634 Perl).
1635
1636 32. A *MARK encountered during the processing of a positive assertion is now
1637 recorded and passed back (compatible with Perl).
1638
1639 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1640 had alternative anchored branches, the search for a second match in a line
1641 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1642 matched the line "0102" twice. The same bug affected patterns that started
1643 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1644 twice.
1645
1646 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1647 does, and because of capturing effects, quantifying parenthesized
1648 assertions may at times be useful. Quantifiers are now allowed for
1649 parenthesized assertions.
1650
1651 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1652
1653 36. \g was being checked for fancy things in a character class, when it should
1654 just be a literal "g".
1655
1656 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1657 appearance of a nested POSIX class supersedes an apparent external class.
1658 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1659 unescaped square brackets may also appear as part of class names. For
1660 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1661 more like Perl. (But see 8.20/1 above.)
1662
1663 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1664 was because it thought it was \N{name}, which is not supported).
1665
1666 39. Add minix to OS list not supporting the -S option in pcretest.
1667
1668 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1669 cannot analyze patterns in sufficient detail to catch mutual recursions
1670 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1671 subgroup is called recursively as a subpattern for a second time at the
1672 same position in the subject string. In previous releases this might have
1673 been caught by the recursion limit, or it might have run out of stack.
1674
1675 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1676 happen only once. PCRE was, however incorrectly giving a compile time error
1677 "recursive call could loop indefinitely" because it cannot analyze the
1678 pattern in sufficient detail. The compile time test no longer happens when
1679 PCRE is compiling a conditional subpattern, but actual runaway loops are
1680 now caught at runtime (see 40 above).
1681
1682 42. It seems that Perl allows any characters other than a closing parenthesis
1683 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1684 has been changed to be the same.
1685
1686 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1687 as not to get warnings when autogen.sh is called. Also changed
1688 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1689
1690 44. To help people who use pcregrep to scan files containing exceedingly long
1691 lines, the following changes have been made:
1692
1693 (a) The default value of the buffer size parameter has been increased from
1694 8K to 20K. (The actual buffer used is three times this size.)
1695
1696 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1697 PCRE is built.
1698
1699 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1700 to be set at run time.
1701
1702 (d) Numerical values in pcregrep options can be followed by K or M, for
1703 example --buffer-size=50K.
1704
1705 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1706 given and the return code is set to 2.
1707
1708 45. Add a pointer to the latest mark to the callout data block.
1709
1710 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1711 partial match of an empty string instead of no match. This was specific to
1712 the use of ".".
1713
1714 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1715 complete match instead of a partial match. This bug was dependent on both
1716 the PCRE_UTF8 and PCRE_DOTALL options being set.
1717
1718 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1719 starting byte set, because \b was not being ignored.
1720
1721
1722 Version 8.12 15-Jan-2011
1723 ------------------------
1724
1725 1. Fixed some typos in the markup of the man pages, and wrote a script that
1726 checks for such things as part of the documentation building process.
1727
1728 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1729 --match-limit and --recursion-limit options (added for 8.11). In
1730 particular, this made one of the standard tests fail. (The integer value
1731 went into the wrong half of a long int.)
1732
1733 3. If the --colour option was given to pcregrep with -v (invert match), it
1734 did strange things, either producing crazy output, or crashing. It should,
1735 of course, ignore a request for colour when reporting lines that do not
1736 match.
1737
1738 4. Another pcregrep bug caused similar problems if --colour was specified with
1739 -M (multiline) and the pattern match finished with a line ending.
1740
1741 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1742 matched in multiline mode, the following line was shown as part of the
1743 match. This seems wrong, so I have changed it.
1744
1745 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1746 the check for further matches in the same line (so they could be coloured)
1747 to overrun the end of the current line. If another match was found, it was
1748 incorrectly shown (and then shown again when found in the next line).
1749
1750 7. If pcregrep was compiled under Windows, there was a reference to the
1751 function pcregrep_exit() before it was defined. I am assuming this was
1752 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1753 reported by a user. I've moved the definition above the reference.
1754
1755
1756 Version 8.11 10-Dec-2010
1757 ------------------------
1758
1759 1. (*THEN) was not working properly if there were untried alternatives prior
1760 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1761 backtracked to try for "b" instead of moving to the next alternative branch
1762 at the same level (in this case, to look for "c"). The Perl documentation
1763 is clear that when (*THEN) is backtracked onto, it goes to the "next
1764 alternative in the innermost enclosing group".
1765
1766 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1767 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1768 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1769 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1770 (*THEN).
1771
1772 3. If \s appeared in a character class, it removed the VT character from
1773 the class, even if it had been included by some previous item, for example
1774 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1775 of \s, but is part of the POSIX "space" class.)
1776
1777 4. A partial match never returns an empty string (because you can always
1778 match an empty string at the end of the subject); however the checking for
1779 an empty string was starting at the "start of match" point. This has been
1780 changed to the "earliest inspected character" point, because the returned
1781 data for a partial match starts at this character. This means that, for
1782 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1783 (previously it gave "no match").
1784
1785 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1786 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1787 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1788 has an implication that the given string is incomplete (because a partial
1789 match is preferred over a full match). For this reason, these items now
1790 give a partial match in this situation. [Aside: previously, the one case
1791 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1792 match rather than a full match, which was wrong by the old rules, but is
1793 now correct.]
1794
1795 6. There was a bug in the handling of #-introduced comments, recognized when
1796 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1797 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1798 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1799 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1800 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1801 places in pcre_compile().
1802
1803 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1804 comments when looking ahead for named forward references to subpatterns,
1805 the only newline sequence it recognized was NL. It now handles newlines
1806 according to the set newline convention.
1807
1808 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1809 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1810 cater for a lack of strerror(). These oversights have been fixed.
1811
1812 9. Added --match-limit and --recursion-limit to pcregrep.
1813
1814 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1815
1816 11. When the -o option was used, pcregrep was setting a return code of 1, even
1817 when matches were found, and --line-buffered was not being honoured.
1818
1819 12. Added an optional parentheses number to the -o and --only-matching options
1820 of pcregrep.
1821
1822 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1823 can match an empty string. The code to do it in pcretest and pcredemo
1824 needed fixing:
1825
1826 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1827 only one byte after an empty string match just before CRLF (this case
1828 just got forgotten; "any" and "anycrlf" were OK).
1829
1830 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1831 mode when an empty string match preceded an ASCII character followed by
1832 a non-ASCII character. (The code for advancing by one character rather
1833 than one byte was nonsense.)
1834
1835 (c) The pcredemo.c sample program did not have any code at all to handle
1836 the cases when CRLF is a valid newline sequence.
1837
1838 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1839 as a starting offset was within the subject string. There is now a new
1840 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1841 negative or greater than the length of the string. In order to test this,
1842 pcretest is extended to allow the setting of negative starting offsets.
1843
1844 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1845 starting offset points to the beginning of a UTF-8 character was
1846 unnecessarily clumsy. I tidied it up.
1847
1848 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1849 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1850
1851 17. Nobody had reported that the --include_dir option, which was added in
1852 release 7.7 should have been called --include-dir (hyphen, not underscore)
1853 for compatibility with GNU grep. I have changed it to --include-dir, but
1854 left --include_dir as an undocumented synonym, and the same for
1855 --exclude-dir, though that is not available in GNU grep, at least as of
1856 release 2.5.4.
1857
1858 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1859 characters from a string of bytes) have been redefined so as not to use
1860 loops, in order to improve performance in some environments. At the same
1861 time, I abstracted some of the common code into auxiliary macros to save
1862 repetition (this should not affect the compiled code).
1863
1864 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1865 compile-time error is now given if \c is not followed by an ASCII
1866 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1867 different, and any byte value is allowed.)
1868
1869 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1870 START_OPTIMIZE option, which is now allowed at compile time - but just
1871 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1872 to pcregrep and other applications that have no direct access to PCRE
1873 options. The new /Y option in pcretest sets this option when calling
1874 pcre_compile().
1875
1876 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1877 back references. Groups containing recursive back references were forced to
1878 be atomic by that change, but in the case of named groups, the amount of
1879 memory required was incorrectly computed, leading to "Failed: internal
1880 error: code overflow". This has been fixed.
1881
1882 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1883 pcretest.c, to avoid build problems in some Borland environments.
1884
1885
1886 Version 8.10 25-Jun-2010
1887 ------------------------
1888
1889 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1890 THEN.
1891
1892 2. (*ACCEPT) was not working when inside an atomic group.
1893
1894 3. Inside a character class, \B is treated as a literal by default, but
1895 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1896 causes the error). The code is unchanged, but I tidied the documentation.
1897
1898 4. Inside a character class, PCRE always treated \R and \X as literals,
1899 whereas Perl faults them if its -w option is set. I have changed PCRE so
1900 that it faults them when PCRE_EXTRA is set.
1901
1902 5. Added support for \N, which always matches any character other than
1903 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1904
1905 6. When compiling pcregrep with newer versions of gcc which may have
1906 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1907 declared with attribute warn_unused_result" were given. Just casting the
1908 result to (void) does not stop the warnings; a more elaborate fudge is
1909 needed. I've used a macro to implement this.
1910
1911 7. Minor change to pcretest.c to avoid a compiler warning.
1912
1913 8. Added four artifical Unicode properties to help with an option to make
1914 \s etc use properties (see next item). The new properties are: Xan
1915 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1916
1917 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1918 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1919 this option. Modified pcretest to add /W to test this facility. Added
1920 REG_UCP to make it available via the POSIX interface.
1921
1922 10. Added --line-buffered to pcregrep.
1923
1924 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1925 studied, and the match started with a letter with a code point greater than
1926 127 whose first byte was different to the first byte of the other case of
1927 the letter, the other case of this starting letter was not recognized
1928 (#976).
1929
1930 12. If a pattern that was studied started with a repeated Unicode property
1931 test, for example, \p{Nd}+, there was the theoretical possibility of
1932 setting up an incorrect bitmap of starting bytes, but fortunately it could
1933 not have actually happened in practice until change 8 above was made (it
1934 added property types that matched character-matching opcodes).
1935
1936 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1937 possible starting bytes for non-anchored patterns.
1938
1939 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1940 \R, and also a number of cases that involve Unicode properties, both
1941 explicit and implicit when PCRE_UCP is set.
1942
1943 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1944 input, it could crash or give wrong results if characters with values
1945 greater than 0xc0 were present in the subject string. (Detail: it assumed
1946 UTF-8 input when processing these items.)
1947
1948 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1949 size_t is 64-bit (#991).
1950
1951 17. Added a check for running out of memory when PCRE is compiled with
1952 --disable-stack-for-recursion (#990).
1953
1954 18. If the last data line in a file for pcretest does not have a newline on
1955 the end, a newline was missing in the output.
1956
1957 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1958 less than 128) in its various bitmaps. However, there is a facility for
1959 generating tables according to the current locale when PCRE is compiled. It
1960 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1961 characters, are recognized by isspace() and therefore were getting set in
1962 these tables, and indeed these tables seem to approximate to ISO 8859. This
1963 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1964 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1965 which of course cannot start UTF-8 characters. I have changed the code so
1966 that only real ASCII characters (less than 128) and the correct starting
1967 bytes for UTF-8 encodings are set for characters greater than 127 when in
1968 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1969 altogether.)
1970
1971 20. Added the /T option to pcretest so as to be able to run tests with non-
1972 standard character tables, thus making it possible to include the tests
1973 used for 19 above in the standard set of tests.
1974
1975 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1976 reference to a subpattern the other side of a comment that contains an
1977 opening parenthesis caused either an internal compiling error, or a
1978 reference to the wrong subpattern.
1979
1980
1981 Version 8.02 19-Mar-2010
1982 ------------------------
1983
1984 1. The Unicode data tables have been updated to Unicode 5.2.0.
1985
1986 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1987 configured.
1988
1989 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1990 original author of that file, following a query about its status.
1991
1992 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1993 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1994
1995 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1996 quantifier applied to a forward-referencing subroutine call, could compile
1997 incorrect code or give the error "internal error: previously-checked
1998 referenced subpattern not found".
1999
2000 6. Both MS Visual Studio and Symbian OS have problems with initializing
2001 variables to point to external functions. For these systems, therefore,
2002 pcre_malloc etc. are now initialized to local functions that call the
2003 relevant global functions.
2004
2005 7. There were two entries missing in the vectors called coptable and poptable
2006 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2007 I've fixed the data, and added a kludgy way of testing at compile time that
2008 the lengths are correct (equal to the number of opcodes).
2009
2010 8. Following on from 7, I added a similar kludge to check the length of the
2011 eint vector in pcreposix.c.
2012
2013 9. Error texts for pcre_compile() are held as one long string to avoid too
2014 much relocation at load time. To find a text, the string is searched,
2015 counting zeros. There was no check for running off the end of the string,
2016 which could happen if a new error number was added without updating the
2017 string.
2018
2019 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2020
2021 11. \K was not working if it appeared in an atomic group or in a group that
2022 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2023 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2024 the assertion is positive, but not if it is negative.
2025
2026 12. Change 11 fortuitously reduced the size of the stack frame used in the
2027 "match()" function of pcre_exec.c by one pointer. Forthcoming
2028 implementation of support for (*MARK) will need an extra pointer on the
2029 stack; I have reserved it now, so that the stack frame size does not
2030 decrease.
2031
2032 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2033 item in branch that calls a recursion is a subroutine call - as in the
2034 second branch in the above example - was incorrectly given the compile-
2035 time error "recursive call could loop indefinitely" because pcre_compile()
2036 was not correctly checking the subroutine for matching a non-empty string.
2037
2038 14. The checks for overrunning compiling workspace could trigger after an
2039 overrun had occurred. This is a "should never occur" error, but it can be
2040 triggered by pathological patterns such as hundreds of nested parentheses.
2041 The checks now trigger 100 bytes before the end of the workspace.
2042
2043 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2044
2045
2046 Version 8.01 19-Jan-2010
2047 ------------------------
2048
2049 1. If a pattern contained a conditional subpattern with only one branch (in
2050 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2051 computed the wrong minimum data length (which is of course zero for such
2052 subpatterns). This could cause incorrect "no match" results.
2053
2054 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2055 the pattern is reset in the first branch, pcre_compile() failed with
2056 "internal error: code overflow at offset...". This happened only when
2057 the reset was to the original external option setting. (An optimization
2058 abstracts leading options settings into an external setting, which was the
2059 cause of this.)
2060
2061 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2062 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2063 assertion pattern did not match (meaning that the assertion was true), it
2064 was incorrectly treated as false if the SKIP had been reached during the
2065 matching. This also applied to assertions used as conditions.
2066
2067 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2068 assertion subpattern, including such a pattern used as a condition,
2069 unpredictable results occurred, instead of the error return
2070 PCRE_ERROR_DFA_UITEM.
2071
2072 5. The C++ GlobalReplace function was not working like Perl for the special
2073 situation when an empty string is matched. It now does the fancy magic
2074 stuff that is necessary.
2075
2076 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2077 removed. (These were left over from very, very early versions of PCRE.)
2078
2079 7. Some cosmetic changes to the code to make life easier when compiling it
2080 as part of something else:
2081
2082 (a) Change DEBUG to PCRE_DEBUG.
2083
2084 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2085 called "current" as "current_branch", to prevent a collision with the
2086 Linux macro when compiled as a kernel module.
2087
2088 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2089 prevent a collision with the Linux macro when compiled as a kernel
2090 module.
2091
2092 8. In pcre_compile() there are some checks for integer overflows that used to
2093 cast potentially large values to (double). This has been changed to that
2094 when building, a check for int64_t is made, and if it is found, it is used
2095 instead, thus avoiding the use of floating point arithmetic. (There is no
2096 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2097 double.
2098
2099 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2100 2005 (difference between two addresses compared to an unsigned value).
2101
2102 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2103 custom one, because of the following reported problem in Windows:
2104
2105 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2106 under Win32.
2107 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2108 therefore missing the function definition.
2109 - The compiler thus generates a "C" signature for the test function.
2110 - The linker fails to find the "C" function.
2111 - PCRE fails to configure if asked to do so against libbz2.
2112
2113 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2114 messages were output:
2115
2116 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2117 rerunning libtoolize, to keep the correct libtool macros in-tree.
2118 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2119
2120 I have done both of these things.
2121
2122 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2123 most of the time, it *can* run out if it is given a pattern that contains a
2124 runaway infinite recursion. I updated the discussion in the pcrestack man
2125 page.
2126
2127 13. Now that we have gone to the x.xx style of version numbers, the minor
2128 version may start with zero. Using 08 or 09 is a bad idea because users
2129 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2130 interpreted as invalid octal numbers. I've updated the previous comment in
2131 configure.ac, and also added a check that gives an error if 08 or 09 are
2132 used.
2133
2134 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2135 causing partial matching to fail when the end of the subject matched \W
2136 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2137
2138 15. There were some discrepancies between the declarations in pcre_internal.h
2139 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2140 their definitions. The declarations used "const uschar *" and the
2141 definitions used USPTR. Even though USPTR is normally defined as "const
2142 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2143 reported that: "This difference in casting confuses some C++ compilers, for
2144 example, SunCC recognizes above declarations as different functions and
2145 generates broken code for hbpcre." I have changed the declarations to use
2146 USPTR.
2147
2148 16. GNU libtool is named differently on some systems. The autogen.sh script now
2149 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2150 (FreeBSD).
2151
2152 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2153 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2154 comment: "Figure out how to create a longlong from a string: strtoll and
2155 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2156 instance, but it only takes 2 args instead of 3!"
2157
2158 18. A subtle bug concerned with back references has been fixed by a change of
2159 specification, with a corresponding code fix. A pattern such as
2160 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2161 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2162 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2163 same bug. Such groups have to be quantified to be useful, or contained
2164 inside another quantified group. (If there's no repetition, the reference
2165 can never match.) The problem arises because, having left the group and
2166 moved on to the rest of the pattern, a later failure that backtracks into
2167 the group uses the captured value from the final iteration of the group
2168 rather than the correct earlier one. I have fixed this in PCRE by forcing
2169 any group that contains a reference to itself to be an atomic group; that
2170 is, there cannot be any backtracking into it once it has completed. This is
2171 similar to recursive and subroutine calls.
2172
2173
2174 Version 8.00 19-Oct-09
2175 ----------------------
2176
2177 1. The table for translating pcre_compile() error codes into POSIX error codes
2178 was out-of-date, and there was no check on the pcre_compile() error code
2179 being within the table. This could lead to an OK return being given in
2180 error.
2181
2182 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2183 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2184 in a Windows environment.
2185
2186 3. The pcregrep --count option prints the count for each file even when it is
2187 zero, as does GNU grep. However, pcregrep was also printing all files when
2188 --files-with-matches was added. Now, when both options are given, it prints
2189 counts only for those files that have at least one match. (GNU grep just
2190 prints the file name in this circumstance, but including the count seems
2191 more useful - otherwise, why use --count?) Also ensured that the
2192 combination -clh just lists non-zero counts, with no names.
2193
2194 4. The long form of the pcregrep -F option was incorrectly implemented as
2195 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2196 but it seems right to fix it, and I didn't think it was worth preserving
2197 the old behaviour.
2198
2199 5. The command line items --regex=pattern and --regexp=pattern were not
2200 recognized by pcregrep, which required --regex pattern or --regexp pattern
2201 (with a space rather than an '='). The man page documented the '=' forms,
2202 which are compatible with GNU grep; these now work.
2203
2204 6. No libpcreposix.pc file was created for pkg-config; there was just
2205 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2206
2207 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2208 when UCP support is not needed, by modifying the Python script that
2209 generates it from Unicode data files. This should not matter if the module
2210 is correctly used as a library, but I received one complaint about 50K of
2211 unwanted data. My guess is that the person linked everything into his
2212 program rather than using a library. Anyway, it does no harm.
2213
2214 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2215 was a minimum greater than 1 for a wide character in a possessive
2216 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2217 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2218 character. Chaos in the form of incorrect output or a compiling loop could
2219 result.
2220
2221 9. The restrictions on what a pattern can contain when partial matching is
2222 requested for pcre_exec() have been removed. All patterns can now be
2223 partially matched by this function. In addition, if there are at least two
2224 slots in the offset vector, the offset of the earliest inspected character
2225 for the match and the offset of the end of the subject are set in them when
2226 PCRE_ERROR_PARTIAL is returned.
2227
2228 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2229 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2230 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2231 and may be more useful for multi-segment matching.
2232
2233 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2234 used to be given if ever the end of the subject was reached; now it is
2235 given only if matching could not proceed because another character was
2236 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2237 string "Z", which now yields "no match" instead of "partial match". In the
2238 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2239 final character ended with (*FAIL).
2240
2241 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2242 if the pattern had a "must contain" character that was already found in the
2243 earlier partial match, unless partial matching was again requested. For
2244 example, with the pattern /dog.(body)?/, the "must contain" character is
2245 "g". If the first part-match was for the string "dog", restarting with
2246 "sbody" failed. This bug has been fixed.
2247
2248 13. The string returned by pcre_dfa_exec() after a partial match has been
2249 changed so that it starts at the first inspected character rather than the
2250 first character of the match. This makes a difference only if the pattern
2251 starts with a lookbehind assertion or \b or \B (\K is not supported by
2252 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2253 matching functions compatible, and I think it's the right thing to do.
2254
2255 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2256 so that the demonstration program is easily available in environments where
2257 PCRE has not been installed from source.
2258
2259 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2260 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2261 library.
2262
2263 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2264 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2265 is not the first non-POSIX option to be added. Clearly some people find
2266 these options useful.
2267
2268 17. If a caller to the POSIX matching function regexec() passes a non-zero
2269 value for nmatch with a NULL value for pmatch, the value of
2270 nmatch is forced to zero.
2271
2272 18. RunGrepTest did not have a test for the availability of the -u option of
2273 the diff command, as RunTest does. It now checks in the same way as
2274 RunTest, and also checks for the -b option.
2275
2276 19. If an odd number of negated classes containing just a single character
2277 interposed, within parentheses, between a forward reference to a named
2278 subpattern and the definition of the subpattern, compilation crashed with
2279 an internal error, complaining that it could not find the referenced
2280 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2281 [The bug was that it was starting one character too far in when skipping
2282 over the character class, thus treating the ] as data rather than
2283 terminating the class. This meant it could skip too much.]
2284
2285 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2286 /g option in pcretest when the pattern contains \K, which makes it possible
2287 to have an empty string match not at the start, even when the pattern is
2288 anchored. Updated pcretest and pcredemo to use this option.
2289
2290 21. If the maximum number of capturing subpatterns in a recursion was greater
2291 than the maximum at the outer level, the higher number was returned, but
2292 with unset values at the outer level. The correct (outer level) value is
2293 now given.
2294
2295 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2296 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2297 make it do so. The string so far is captured, making this feature
2298 compatible with Perl.
2299
2300 23. The tests have been re-organized, adding tests 11 and 12, to make it
2301 possible to check the Perl 5.10 features against Perl 5.10.
2302
2303 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2304 pattern matches a fixed length string. PCRE did not allow this; now it
2305 does. Neither allows recursion.
2306
2307 25. I finally figured out how to implement a request to provide the minimum
2308 length of subject string that was needed in order to match a given pattern.
2309 (It was back references and recursion that I had previously got hung up
2310 on.) This code has now been added to pcre_study(); it finds a lower bound
2311 to the length of subject needed. It is not necessarily the greatest lower
2312 bound, but using it to avoid searching strings that are too short does give
2313 some useful speed-ups. The value is available to calling programs via
2314 pcre_fullinfo().
2315
2316 26. While implementing 25, I discovered to my embarrassment that pcretest had
2317 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2318 study optimizations had never been tested with that matching function.
2319 Oops. What is worse, even when it was passed study data, there was a bug in
2320 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2321 were also very few tests of studied patterns with pcre_dfa_exec().
2322
2323 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2324 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2325 on the other side of the coin, they are no longer allowed to have different
2326 names, because these cannot be distinguished in PCRE, and this has caused
2327 confusion. (This is a difference from Perl.)
2328
2329 28. When duplicate subpattern names are present (necessarily with different
2330 numbers, as required by 27 above), and a test is made by name in a
2331 conditional pattern, either for a subpattern having been matched, or for
2332 recursion in such a pattern, all the associated numbered subpatterns are
2333 tested, and the overall condition is true if the condition is true for any
2334 one of them. This is the way Perl works, and is also more like the way
2335 testing by number works.
2336
2337
2338 Version 7.9 11-Apr-09
2339 ---------------------
2340
2341 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2342 (pcretest), all targets were linked against these libraries. This included
2343 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2344 libraries. This caused unwanted dependencies to be created. This problem
2345 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2346 pcretest is linked with readline.
2347
2348 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2349 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2350 moved outside it again, because FALSE and TRUE are already defined in AIX,
2351 but BOOL is not.
2352
2353 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2354 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2355
2356 4. The pcregrep documentation said spaces were inserted as well as colons (or
2357 hyphens) following file names and line numbers when outputting matching
2358 lines. This is not true; no spaces are inserted. I have also clarified the
2359 wording for the --colour (or --color) option.
2360
2361 5. In pcregrep, when --colour was used with -o, the list of matching strings
2362 was not coloured; this is different to GNU grep, so I have changed it to be
2363 the same.
2364
2365 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2366 each matching line was coloured. Now it goes on to look for further matches
2367 of any of the test patterns, which is the same behaviour as GNU grep.
2368
2369 7. A pattern that could match an empty string could cause pcregrep to loop; it
2370 doesn't make sense to accept an empty string match in pcregrep, so I have
2371 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2372 seems to be how GNU grep behaves. [But see later change 40 for release
2373 8.33.]
2374
2375 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2376 start or after a newline", because the conditional assertion was not being
2377 correctly handled. The rule now is that both the assertion and what follows
2378 in the first alternative must satisfy the test.
2379
2380 9. If auto-callout was enabled in a pattern with a conditional group whose
2381 condition was an assertion, PCRE could crash during matching, both with
2382 pcre_exec() and pcre_dfa_exec().
2383
2384 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2385 used for matching.
2386
2387 11. Unicode property support in character classes was not working for
2388 characters (bytes) greater than 127 when not in UTF-8 mode.
2389
2390 12. Added the -M command line option to pcretest.
2391
2392 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2393
2394 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2395
2396 16. Added comments and documentation about mis-use of no_arg in the C++
2397 wrapper.
2398
2399 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2400 from Martin Jerabek that uses macro names for all relevant character and
2401 string constants.
2402
2403 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2404 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2405 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2406 these, but not everybody uses configure.
2407
2408 19. A conditional group that had only one branch was not being correctly
2409 recognized as an item that could match an empty string. This meant that an
2410 enclosing group might also not be so recognized, causing infinite looping
2411 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2412 with the subject "ab", where knowledge that the repeated group can match
2413 nothing is needed in order to break the loop.
2414
2415 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2416 exec(), but without supplying a callout function, matching went wrong.
2417
2418 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2419 leak if the size of the offset vector was greater than 30. When the vector
2420 is smaller, the saved offsets during recursion go onto a local stack
2421 vector, but for larger vectors malloc() is used. It was failing to free
2422 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2423 error, in fact).
2424
2425 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2426 heapframe that is used only when UTF-8 support is enabled. This caused no
2427 problem, but was untidy.
2428
2429 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2430 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2431 included within another project.
2432
2433 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2434 slightly modified by me:
2435
2436 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2437 not building pcregrep.
2438
2439 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2440 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2441
2442 25. Forward references, both numeric and by name, in patterns that made use of
2443 duplicate group numbers, could behave incorrectly or give incorrect errors,
2444 because when scanning forward to find the reference group, PCRE was not
2445 taking into account the duplicate group numbers. A pattern such as
2446 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2447
2448 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2449 the feature of a custom pointer more persuasive (as requested by a user).
2450
2451 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2452 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2453 pre-defined.
2454
2455 28. Added support for (*UTF8) at the start of a pattern.
2456
2457 29. Arrange for flags added by the "release type" setting in CMake to be shown
2458 in the configuration summary.
2459
2460
2461 Version 7.8 05-Sep-08
2462 ---------------------
2463
2464 1. Replaced UCP searching code with optimized version as implemented for Ad
2465 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2466 stage table and inline lookup instead of a function, giving speed ups of 2
2467 to 5 times on some simple patterns that I tested. Permission was given to
2468 distribute the MultiStage2.py script that generates the tables (it's not in
2469 the tarball, but is in the Subversion repository).
2470
2471 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2472 scripts.
2473
2474 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2475 a group with a zero qualifier. The result of the study could be incorrect,
2476 or the function might crash, depending on the pattern.
2477
2478 4. Caseless matching was not working for non-ASCII characters in back
2479 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2480 It now works when Unicode Property Support is available.
2481
2482 5. In pcretest, an escape such as \x{de} in the data was always generating
2483 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2484 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2485 truncation.
2486
2487 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2488
2489 7. Added two (int) casts to pcregrep when printing the difference of two
2490 pointers, in case they are 64-bit values.
2491
2492 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2493 test 2 if it fails.
2494
2495 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2496 and a #define of that name to empty if it is not externally set. This is to
2497 allow users of MSVC to set it if necessary.
2498
2499 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2500 the convenience functions in the pcre_get.c source file.
2501
2502 11. An option change at the start of a pattern that had top-level alternatives
2503 could cause overwriting and/or a crash. This command provoked a crash in
2504 some environments:
2505
2506 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2507
2508 This potential security problem was recorded as CVE-2008-2371.
2509
2510 12. For a pattern where the match had to start at the beginning or immediately
2511 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2512 pcre_dfa_exec() could read past the end of the passed subject if there was
2513 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2514 pcretest so that it places the subject at the end of its malloc-ed buffer.
2515
2516 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2517 exec() might read past the end of the data buffer in UTF-8 mode.
2518
2519 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2520 the data contained the byte 0x85 as part of a UTF-8 character within its
2521 first line. This applied both to normal and DFA matching.
2522
2523 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2524 /^[^d]*?$/8 failed to match "abc".
2525
2526 16. Added a missing copyright notice to pcrecpp_internal.h.
2527
2528 17. Make it more clear in the documentation that values returned from
2529 pcre_exec() in ovector are byte offsets, not character counts.
2530
2531 18. Tidied a few places to stop certain compilers from issuing warnings.
2532
2533 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2534 supplied by Stefan Weber. I made a further small update for 7.8 because
2535 there is a change of source arrangements: the pcre_searchfuncs.c module is
2536 replaced by pcre_ucd.c.
2537
2538
2539 Version 7.7 07-May-08
2540 ---------------------
2541
2542 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2543 a string to a long long, pretend we don't even have a long long." This is
2544 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2545
2546 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2547 pre-7.6 versions, which defined a global no_arg variable instead of putting
2548 it in the RE class. (See also #8 below.)
2549
2550 3. Remove a line of dead code, identified by coverity and reported by Nuno
2551 Lopes.
2552
2553 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2554
2555 (1) The include/exclude patterns were being applied to the whole pathnames
2556 of files, instead of just to the final components.
2557
2558 (2) If there was more than one level of directory, the subdirectories were
2559 skipped unless they satisfied the include/exclude conditions. This is
2560 inconsistent with GNU grep (and could even be seen as contrary to the
2561 pcregrep specification - which I improved to make it absolutely clear).
2562 The action now is always to scan all levels of directory, and just
2563 apply the include/exclude patterns to regular files.
2564
2565 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2566 --exclude_dir in the tests to avoid scanning .svn directories.
2567
2568 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2569 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2570 doesn't support NULs in patterns.
2571
2572 7. Added some missing "const"s to declarations of static tables in
2573 pcre_compile.c and pcre_dfa_exec.c.
2574
2575 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2576 caused by fix #2 above. (Subsequently also a second patch to fix the
2577 first patch. And a third patch - this was a messy problem.)
2578
2579 9. Applied Craig's patch to remove the use of push_back().
2580
2581 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2582 matching function regexec().
2583
2584 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2585 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2586 references. PCRE supports relative numbers with this syntax (I don't think
2587 Oniguruma does).
2588
2589 12. Previously, a group with a zero repeat such as (...){0} was completely
2590 omitted from the compiled regex. However, this means that if the group
2591 was called as a subroutine from elsewhere in the pattern, things went wrong
2592 (an internal error was given). Such groups are now left in the compiled
2593 pattern, with a new opcode that causes them to be skipped at execution
2594 time.
2595
2596 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2597 to the way PCRE behaves:
2598
2599 (a) A lone ] character is dis-allowed (Perl treats it as data).
2600
2601 (b) A back reference to an unmatched subpattern matches an empty string
2602 (Perl fails the current match path).
2603
2604 (c) A data ] in a character class must be notated as \] because if the
2605 first data character in a class is ], it defines an empty class. (In
2606 Perl it is not possible to have an empty class.) The empty class []
2607 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2608 The negative empty class [^] matches any one character, independently
2609 of the DOTALL setting.
2610
2611 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2612 non-existent subpattern following a character class starting with ']' and
2613 containing () gave an internal compiling error instead of "reference to
2614 non-existent subpattern". Fortunately, when the pattern did exist, the
2615 compiled code was correct. (When scanning forwards to check for the
2616 existence of the subpattern, it was treating the data ']' as terminating
2617 the class, so got the count wrong. When actually compiling, the reference
2618 was subsequently set up correctly.)
2619
2620 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2621 it was being rejected as not supported by pcre_dfa_exec(), even though
2622 other assertions are supported. I have made pcre_dfa_exec() support
2623 (*FAIL).
2624
2625 16. The implementation of 13c above involved the invention of a new opcode,
2626 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2627 cannot be changed at match time, I realized I could make a small
2628 improvement to matching performance by compiling OP_ALLANY instead of
2629 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2630 on the OP_ANY path.
2631
2632 17. Compiling pcretest on Windows with readline support failed without the
2633 following two fixes: (1) Make the unistd.h include conditional on
2634 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2635
2636 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2637 ncurses library to be included for pcretest when ReadLine support is
2638 requested, but also to allow for it to be overridden. This patch came from
2639 Daniel Bergström.
2640
2641 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2642 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2643 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2644 spotting this.
2645
2646
2647 Version 7.6 28-Jan-08
2648 ---------------------
2649
2650 1. A character class containing a very large number of characters with
2651 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2652 overflow.
2653
2654 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2655 HAVE_LONG_LONG is not defined.
2656
2657 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2658 bring them up to date and include new features. This patch includes:
2659
2660 - Fixed PH's badly added libz and libbz2 support.
2661 - Fixed a problem with static linking.
2662 - Added pcredemo. [But later removed - see 7 below.]
2663 - Fixed dftables problem and added an option.
2664 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2665 HAVE_LONG_LONG.
2666 - Added readline support for pcretest.
2667 - Added an listing of the option settings after cmake has run.
2668
2669 4. A user submitted a patch to Makefile that makes it easy to create
2670 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2671 Makefile.am that cause it to include this special target, without
2672 affecting anything else. Note that the same mingw target plus all
2673 the other distribution libraries and programs are now supported
2674 when configuring with CMake (see 6 below) instead of with
2675 Configure/Make.
2676
2677 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2678 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2679 exported in the Windows port". It has not yet been confirmed that the patch
2680 solves the problem, but it does no harm.
2681
2682 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2683 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2684 with CMake, and also correct the comment about stack recursion.
2685
2686 7. Remove the automatic building of pcredemo from the ./configure system and
2687 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2688 of a program that users should build themselves after PCRE is installed, so
2689 building it automatically is not really right. What is more, it gave
2690 trouble in some build environments.
2691
2692 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2693
2694
2695 Version 7.5 10-Jan-08
2696 ---------------------
2697
2698 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2699 values in parens when parsing an RE using the C++ wrapper."
2700
2701 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2702 Characters greater than 255 were excluded from the class instead of being
2703 included.
2704
2705 3. The same bug as (2) above applied to negated POSIX classes such as
2706 [:^space:].
2707
2708 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2709 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2710 I have changed it.
2711
2712 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2713 first named subpattern) and a construct such as (?&a) would reference the
2714 first named subpattern whose name started with "a" (in other words, the
2715 length check was missing). Both these problems are fixed. "Subpattern name
2716 expected" is now given for (?&) (a zero-length name), and this patch also
2717 makes it give the same error for \k'' (previously it complained that that
2718 was a reference to a non-existent subpattern).
2719
2720 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2721 this is right because (?- can be followed by option settings as well as by
2722 digits. I have, however, made the messages clearer.
2723
2724 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2725 than the number used in the conditional) now cause a compile-time error.
2726 This is actually not compatible with Perl, which accepts such patterns, but
2727 treats the conditional as always being FALSE (as PCRE used to), but it
2728 seems to me that giving a diagnostic is better.
2729
2730 8. Change "alphameric" to the more common word "alphanumeric" in comments
2731 and messages.
2732
2733 9. Fix two occurrences of "backslash" in comments that should have been
2734 "backspace".
2735
2736 10. Remove two redundant lines of code that can never be obeyed (their function
2737 was moved elsewhere).
2738
2739 11. The program that makes PCRE's Unicode character property table had a bug
2740 which caused it to generate incorrect table entries for sequences of
2741 characters that have the same character type, but are in different scripts.
2742 It amalgamated them into a single range, with the script of the first of
2743 them. In other words, some characters were in the wrong script. There were
2744 thirteen such cases, affecting characters in the following ranges:
2745
2746 U+002b0 - U+002c1
2747 U+0060c - U+0060d
2748 U+0061e - U+00612
2749 U+0064b - U+0065e
2750 U+0074d - U+0076d
2751 U+01800 - U+01805
2752 U+01d00 - U+01d77
2753 U+01d9b - U+01dbf
2754 U+0200b - U+0200f
2755 U+030fc - U+030fe
2756 U+03260 - U+0327f
2757 U+0fb46 - U+0fbb1
2758 U+10450 - U+1049d
2759
2760 12. The -o option (show only the matching part of a line) for pcregrep was not
2761 compatible with GNU grep in that, if there was more than one match in a
2762 line, it showed only the first of them. It now behaves in the same way as
2763 GNU grep.
2764
2765 13. If the -o and -v options were combined for pcregrep, it printed a blank
2766 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2767 does the same. The return code can be used to tell if there were any
2768 non-matching lines.
2769
2770 14. Added --file-offsets and --line-offsets to pcregrep.
2771
2772 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2773 infinitely looping recursion. The bug was that positive lookaheads were not
2774 being skipped when checking for a possible empty match (negative lookaheads
2775 and both kinds of lookbehind were skipped).
2776
2777 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2778 inclusion of <windows.h> to before rather than after the definition of
2779 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2780
2781 17. Specifying a possessive quantifier with a specific limit for a Unicode
2782 character property caused pcre_compile() to compile bad code, which led at
2783 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2784 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2785 caused the error; without that there was no problem.
2786
2787 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2788
2789 19. Added --enable-pcretest-libreadline.
2790
2791 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2792 RE::GlobalReplace(). As a result, the number of replacements returned was
2793 double what it should be. I removed one of the increments, but Craig sent a
2794 later patch that removed the other one (the right fix) and added unit tests
2795 that check the return values (which was not done before).
2796
2797 21. Several CMake things:
2798
2799 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2800 the names libpcre and libpcreposix, not just pcre and pcreposix.
2801
2802 (2) The above change means that pcretest and pcregrep are now correctly
2803 linked with the newly-built libraries, not previously installed ones.
2804
2805 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2806
2807 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2808 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2809 UTF-8 newline character). The key issue is that the pattern starts .*;
2810 this means that the match must be either at the beginning, or after a
2811 newline. The bug was in the code for advancing after a failed match and
2812 checking that the new position followed a newline. It was not taking
2813 account of UTF-8 characters correctly.
2814
2815 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2816 character classes. PCRE was not treating the sequence [:...:] as a
2817 character class unless the ... were all letters. Perl, however, seems to
2818 allow any characters between [: and :], though of course it rejects as
2819 unknown any "names" that contain non-letters, because all the known class
2820 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2821 for example, whereas PCRE did not - it did not recognize a POSIX character
2822 class. This seemed a bit dangerous, so the code has been changed to be
2823 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2824 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2825 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2826 Perl does, and where it didn't before.
2827
2828 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2829 Windows environments %n is disabled by default.
2830
2831
2832 Version 7.4 21-Sep-07
2833 ---------------------
2834
2835 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2836 means that a class such as [\s] counted as "explicit reference to CR or
2837 LF". That isn't really right - the whole point of the change was to try to
2838 help when there was an actual mention of one of the two characters. So now
2839 the change happens only if \r or \n (or a literal CR or LF) character is
2840 encountered.
2841
2842 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2843 of both had grown to the point where there were only 3 bits left.
2844 Fortunately, there was spare space in the data structure, and so I have
2845 moved the internal flags into a new 16-bit field to free up more option
2846 bits.
2847
2848 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2849 but did not set the internal JCHANGED flag - either of these is enough to
2850 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2851 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2852 start sets both bits.
2853
2854 4. Added options (at build time, compile time, exec time) to change \R from
2855 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2856
2857 5. doc/pcresyntax.html was missing from the distribution.
2858
2859 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2860 compatibility, even though it is no longer used.
2861
2862 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2863 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2864 windows.h file is present (where different names are used). [This was
2865 reversed later after testing - see 16 below.]
2866
2867 8. Changed all #include <config.h> to #include "config.h". There were also
2868 some further <pcre.h> cases that I changed to "pcre.h".
2869
2870 9. When pcregrep was used with the --colour option, it missed the line ending
2871 sequence off the lines that it output.
2872
2873 10. It was pointed out to me that arrays of string pointers cause lots of
2874 relocations when a shared library is dynamically loaded. A technique of
2875 using a single long string with a table of offsets can drastically reduce
2876 these. I have refactored PCRE in four places to do this. The result is
2877 dramatic:
2878
2879 Originally: 290
2880 After changing UCP table: 187
2881 After changing error message table: 43
2882 After changing table of "verbs" 36
2883 After changing table of Posix names 22
2884
2885 Thanks to the folks working on Gregex for glib for this insight.
2886
2887 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2888 unicode-properties was also set.
2889
2890 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2891
2892 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2893 checked only for CRLF.
2894
2895 14. Added casts to pcretest.c to avoid compiler warnings.
2896
2897 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2898
2899 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2900 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2901 entirely. This removes changes made in 7 above.
2902
2903 17. The CMake files have been updated, and there is now more information about
2904 building with CMake in the NON-UNIX-USE document.
2905
2906
2907 Version 7.3 28-Aug-07
2908 ---------------------
2909
2910 1. In the rejigging of the build system that eventually resulted in 7.1, the
2911 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2912 brackets there is not right, since it causes compilers to look for an
2913 installed pcre.h, not the version that is in the source that is being
2914 compiled (which of course may be different). I have changed it back to:
2915
2916 #include "pcre.h"
2917
2918 I have a vague recollection that the change was concerned with compiling in
2919 different directories, but in the new build system, that is taken care of
2920 by the VPATH setting the Makefile.
2921
2922 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2923 when the subject happened to end in the byte 0x85 (e.g. if the last
2924 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2925 characters but of course it shouldn't be taken as a newline when it is part
2926 of another character. The bug was that, for an unlimited repeat of . in
2927 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2928 characters when looking for a newline.
2929
2930 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2931
2932 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2933 in debug output.
2934
2935 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2936 long printing in the pcrecpp unittest when running under MinGW.
2937
2938 6. ESC_K was left out of the EBCDIC table.
2939
2940 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2941 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2942 limit also applies to "virtual nesting" when a pattern is recursive, and in
2943 this case 1000 isn't so big. I have been able to remove this limit at the
2944 expense of backing off one optimization in certain circumstances. Normally,
2945 when pcre_exec() would call its internal match() function recursively and
2946 immediately return the result unconditionally, it uses a "tail recursion"
2947 feature to save stack. However, when a subpattern that can match an empty
2948 string has an unlimited repetition quantifier, it no longer makes this
2949 optimization. That gives it a stack frame in which to save the data for
2950 checking that an empty string has been matched. Previously this was taken
2951 from the 1000-entry workspace that had been reserved. So now there is no
2952 explicit limit, but more stack is used.
2953
2954 8. Applied Daniel's patches to solve problems with the import/export magic
2955 syntax that is required for Windows, and which was going wrong for the
2956 pcreposix and pcrecpp parts of the library. These were overlooked when this
2957 problem was solved for the main library.
2958
2959 9. There were some crude static tests to avoid integer overflow when computing
2960 the size of patterns that contain repeated groups with explicit upper
2961 limits. As the maximum quantifier is 65535, the maximum group length was
2962 set at 30,000 so that the product of these two numbers did not overflow a
2963 32-bit integer. However, it turns out that people want to use groups that
2964 are longer than 30,000 bytes (though not repeat them that many times).
2965 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2966 made it possible to implement the integer overflow checks in a much more
2967 dynamic way, which I have now done. The artificial limitation on group
2968 length has been removed - we now have only the limit on the total length of
2969 the compiled pattern, which depends on the LINK_SIZE setting.
2970
2971 10. Fixed a bug in the documentation for get/copy named substring when
2972 duplicate names are permitted. If none of the named substrings are set, the
2973 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2974 empty string.
2975
2976 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2977 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2978 because the ] is interpreted as the first data character and the
2979 terminating ] is not found. PCRE has been made compatible with Perl in this
2980 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2981 cause memory overwriting.
2982
2983 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2984 string has been matched (to stop an infinite loop). It was not recognizing
2985 a conditional subpattern that could match an empty string if that
2986 subpattern was within another subpattern. For example, it looped when
2987 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2988 condition was not nested. This bug has been fixed.
2989
2990 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2991 past the start of the subject in the presence of bytes with the top bit
2992 set, for example "\x8aBCD".
2993
2994 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2995 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2996
2997 14. Optimized (?!) to (*FAIL).
2998
2999 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3000 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3001 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3002 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3003 does: it's just the validity check that is more restrictive.
3004
3005 16. Inserted checks for integer overflows during escape sequence (backslash)
3006 processing, and also fixed erroneous offset values for syntax errors during
3007 backslash processing.
3008
3009 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3010 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3011
3012 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3013 caused an overrun.
3014
3015 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3016 something other than just ASCII characters) inside a group that had an
3017 unlimited repeat caused a loop at compile time (while checking to see
3018 whether the group could match an empty string).
3019
3020 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3021 [\P{Any}] did so. (Error in the code for printing property names.)
3022
3023 21. An orphan \E inside a character class could cause a crash.
3024
3025 22. A repeated capturing bracket such as (A)? could cause a wild memory
3026 reference during compilation.
3027
3028 23. There are several functions in pcre_compile() that scan along a compiled
3029 expression for various reasons (e.g. to see if it's fixed length for look
3030 behind). There were bugs in these functions when a repeated \p or \P was
3031 present in the pattern. These operators have additional parameters compared
3032 with \d, etc, and these were not being taken into account when moving along
3033 the compiled data. Specifically:
3034
3035 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3036 length.
3037
3038 (b) An item such as \pL+ within a repeated group could cause crashes or
3039 loops.
3040
3041 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3042 "reference to non-existent subpattern" error.
3043
3044 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3045
3046 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3047 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3048
3049 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3050
3051 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3052 character were causing crashes (broken optimization).
3053
3054 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3055 \p or \P) caused a compile-time loop.
3056
3057 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3058 break. For example, the unstudied pattern [\r\n]A does not match the string
3059 "\r\nA" because change 7.0/46 below moves the current point on by two
3060 characters after failing to match at the start. However, the pattern \nA
3061 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3062 the same is true. There doesn't seem any very clean way out of this, but
3063 what I have chosen to do makes the common cases work: PCRE now takes note
3064 of whether there can be an explicit match for \r or \n anywhere in the
3065 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3066 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3067 pattern has explicit CR or LF references.
3068
3069 29. Added (*CR) etc for changing newline setting at start of pattern.
3070
3071
3072 Version 7.2 19-Jun-07
3073 ---------------------
3074
3075 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3076 which is apparently normally available under Windows.
3077
3078 2. Re-jig the pcregrep tests with different newline settings in an attempt
3079 to make them independent of the local environment's newline setting.
3080
3081 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3082
3083 4. Some of the "internals" tests were previously cut out when the link size
3084 was not 2, because the output contained actual offsets. The recent new
3085 "Z" feature of pcretest means that these can be cut out, making the tests
3086 usable with all link sizes.
3087
3088 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3089 stack recursion. This gives a massive performance boost under BSD, but just
3090 a small improvement under Linux. However, it saves one field in the frame
3091 in all cases.
3092
3093 6. Added more features from the forthcoming Perl 5.10:
3094
3095 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3096 recursion call. It refers to the nth most recently opened parentheses.
3097
3098 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3099 to be opened parentheses.
3100
3101 (c) Conditions that refer to capturing parentheses can be specified
3102 relatively, for example, (?(-2)... or (?(+3)...
3103
3104 (d) \K resets the start of the current match so that everything before
3105 is not part of it.
3106
3107 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3108
3109 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3110 reference syntax.
3111
3112 (g) (?| introduces a group in which the numbering of parentheses in each
3113 alternative starts with the same number.
3114
3115 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3116
3117 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3118 PCRE_INFO_JCHANGED.
3119
3120 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3121 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3122 for detecting groups that can match an empty string.
3123
3124 9. A pattern with a very large number of alternatives (more than several
3125 hundred) was running out of internal workspace during the pre-compile
3126 phase, where pcre_compile() figures out how much memory will be needed. A
3127 bit of new cunning has reduced the workspace needed for groups with
3128 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3129 workspace instead of running out of the 4096 that are available.
3130
3131 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3132
3133 11. Applied patch from Google to remove an optimization that didn't quite work.
3134 The report of the bug said:
3135
3136 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3137 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3138 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3139
3140 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3141 it matched the wrong number of bytes.
3142
3143
3144 Version 7.1 24-Apr-07
3145 ---------------------
3146
3147 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3148 that is more "standard", making use of automake and other Autotools. There
3149 is some re-arrangement of the files and adjustment of comments consequent
3150 on this.
3151
3152 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3153 for recursive directory scanning broke on some systems because the files
3154 are not scanned in any specific order and on different systems the order
3155 was different. A call to "sort" has been inserted into RunGrepTest for the
3156 approprate test as a short-term fix. In the longer term there may be an
3157 alternative.
3158
3159 3. I had an email from Eric Raymond about problems translating some of PCRE's
3160 man pages to HTML (despite the fact that I distribute HTML pages, some
3161 people do their own conversions for various reasons). The problems
3162 concerned the use of low-level troff macros .br and .in. I have therefore
3163 removed all such uses from the man pages (some were redundant, some could
3164 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3165 HTML has been updated to handle .nf/.fi and to complain if it encounters
3166 .br or .in.
3167
3168 4. Updated comments in configure.ac that get placed in config.h.in and also
3169 arranged for config.h to be included in the distribution, with the name
3170 config.h.generic, for the benefit of those who have to compile without
3171 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3172
3173 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3174 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3175 makevp.bat for the current PCRE, using the additional files
3176 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3177
3178 6. A Windows user reported a minor discrepancy with test 2, which turned out
3179 to be caused by a trailing space on an input line that had got lost in his
3180 copy. The trailing space was an accident, so I've just removed it.
3181
3182 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3183 that is needed.
3184
3185 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3186 as "const" (a) because they are and (b) because it helps the PHP
3187 maintainers who have recently made a script to detect big data structures
3188 in the php code that should be moved to the .rodata section. I remembered
3189 to update Builducptable as well, so it won't revert if ucptable.h is ever
3190 re-created.
3191
3192 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3193 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3194 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3195 support is not required. This saves 1.5-2K of code, which is important in
3196 some applications.
3197
3198 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3199 so as not to refer to the tables, even though these functions will never be
3200 called when UTF-8 support is disabled. Otherwise there are problems with a
3201 shared library.
3202
3203 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3204
3205 (a) It was defining its arguments as char * instead of void *.
3206
3207 (b) It was assuming that all moves were upwards in memory; this was true
3208 a long time ago when I wrote it, but is no longer the case.
3209
3210 The emulated memove() is provided for those environments that have neither
3211 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3212 is clearly not the case, as these two bugs were recently reported.
3213
3214 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3215 and Detrail to create the HTML documentation, the .txt form of the man
3216 pages, and it removes trailing spaces from listed files. It also creates
3217 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3218 case, it wraps all the #defines with #ifndefs. This script should be run
3219 before "make dist".
3220
3221 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3222 with Unicode property support.
3223
3224 (a) For a maximizing quantifier, if the two different cases of the
3225 character were of different lengths in their UTF-8 codings (there are
3226 some cases like this - I found 11), and the matching function had to
3227 back up over a mixture of the two cases, it incorrectly assumed they
3228 were both the same length.
3229
3230 (b) When PCRE was configured to use the heap rather than the stack for
3231 recursion during matching, it was not correctly preserving the data for
3232 the other case of a UTF-8 character when checking ahead for a match
3233 while processing a minimizing repeat. If the check also involved
3234 matching a wide character, but failed, corruption could cause an
3235 erroneous result when trying to check for a repeat of the original
3236 character.
3237
3238 13. Some tidying changes to the testing mechanism:
3239
3240 (a) The RunTest script now detects the internal link size and whether there
3241 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3242 values substituted by "configure". (The RunGrepTest script already did
3243 this for UTF-8.) The configure.ac script no longer substitutes the
3244 relevant variables.
3245
3246 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3247 with length and offset values. This means that the output is different
3248 for different internal link sizes. Test 2 is skipped for link sizes
3249 other than 2 because of this, bypassing the problem. Unfortunately,
3250 there was also a test in test 3 (the locale tests) that used /B and
3251 failed for link sizes other than 2. Rather than cut the whole test out,
3252 I have added a new /Z option to pcretest that replaces the length and
3253 offset values with spaces. This is now used to make test 3 independent
3254 of link size. (Test 2 will be tidied up later.)
3255
3256 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3257 segmentation fault instead of returning the appropriate error message.
3258
3259 15. In multiline mode when the newline sequence was set to "any", the pattern
3260 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3261 This doesn't seem right; it now treats the CRLF combination as the line
3262 ending, and so does not match in that case. It's only a pattern such as ^$
3263 that would hit this one: something like ^ABC$ would have failed after \r
3264 and then tried again after \r\n.
3265
3266 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3267 in an attempt to make files that differ only in their line terminators
3268 compare equal. This works on Linux.
3269
3270 17. Under certain error circumstances pcregrep might try to free random memory
3271 as it exited. This is now fixed, thanks to valgrind.
3272
3273 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3274 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3275 was because its rules for how to advance for /g after matching an empty
3276 string at the end of a line did not allow for this case. They now check for
3277 it specially.
3278
3279 20. pcretest is supposed to handle patterns and data of any length, by
3280 extending its buffers when necessary. It was getting this wrong when the
3281 buffer for a data line had to be extended.
3282
3283 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3284 CRLF as a newline sequence.
3285
3286 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3287 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3288 I have nevertheless tidied it up.
3289
3290 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3291
3292 24. Added a man page for pcre-config.
3293
3294
3295 Version 7.0 19-Dec-06
3296 ---------------------
3297
3298 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3299 moving to gcc 4.1.1.
3300
3301 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3302 sys/time.h, which is documented as needed for this function. It doesn't
3303 seem to matter on Linux, but it showed up on some releases of OS X.
3304
3305 3. It seems that there are systems where bytes whose values are greater than
3306 127 match isprint() in the "C" locale. The "C" locale should be the
3307 default when a C program starts up. In most systems, only ASCII printing
3308 characters match isprint(). This difference caused the output from pcretest
3309 to vary, making some of the tests fail. I have changed pcretest so that:
3310
3311 (a) When it is outputting text in the compiled version of a pattern, bytes
3312 other than 32-126 are always shown as hex escapes.
3313
3314 (b) When it is outputting text that is a matched part of a subject string,
3315 it does the same, unless a different locale has been set for the match
3316 (using the /L modifier). In this case, it uses isprint() to decide.
3317
3318 4. Fixed a major bug that caused incorrect computation of the amount of memory
3319 required for a compiled pattern when options that changed within the
3320 pattern affected the logic of the preliminary scan that determines the
3321 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3322 that the computed length was too small. The symptoms of this bug were
3323 either the PCRE error "internal error: code overflow" from pcre_compile(),
3324 or a glibc crash with a message such as "pcretest: free(): invalid next
3325 size (fast)". Examples of patterns that provoked this bug (shown in
3326 pcretest format) are:
3327
3328 /(?-x: )/x
3329 /(?x)(?-x: \s*#\s*)/
3330 /((?i)[\x{c0}])/8
3331 /(?i:[\x{c0}])/8
3332
3333 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3334 is now done differently.
3335
3336 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3337 wrapper classes; (b) implement a new function in the C++ scanner that is
3338 more efficient than the old way of doing things because it avoids levels of
3339 recursion in the regex matching; (c) add a paragraph to the documentation
3340 for the FullMatch() function.
3341
3342 6. The escape sequence \n was being treated as whatever was defined as
3343 "newline". Not only was this contrary to the documentation, which states
3344 that \n is character 10 (hex 0A), but it also went horribly wrong when
3345 "newline" was defined as CRLF. This has been fixed.
3346
3347 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3348 was being set to -1 for the "end of line" case (supposedly a value that no
3349 character can have). Though this value is never used (the check for end of
3350 line is "zero bytes in current character"), it caused compiler complaints.
3351 I've changed it to 0xffffffff.
3352
3353 8. In pcre_version.c, the version string was being built by a sequence of
3354 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3355 string (as it is for production releases) called a macro with an empty
3356 argument. The C standard says the result of this is undefined. The gcc
3357 compiler treats it as an empty string (which was what was wanted) but it is
3358 reported that Visual C gives an error. The source has been hacked around to
3359 avoid this problem.
3360
3361 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3362 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3363 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3364 of them did).
3365
3366 10. Originally, pcretest opened its input and output without "b"; then I was
3367 told that "b" was needed in some environments, so it was added for release
3368 5.0 to both the input and output. (It makes no difference on Unix-like
3369 systems.) Later I was told that it is wrong for the input on Windows. I've
3370 now abstracted the modes into two macros, to make it easier to fiddle with
3371 them, and removed "b" from the input mode under Windows.
3372
3373 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3374
3375 12. Added -help and --help to pcretest as an official way of being reminded
3376 of the options.
3377
3378 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3379 and pcrecpp.cc because they annoy compilers at high warning levels.
3380
3381 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3382
3383 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3384 scripts are not C programs :-) and which was not noticed because it works
3385 on Linux.
3386
3387 16. pcretest is supposed to handle any length of pattern and data line (as one
3388 line or as a continued sequence of lines) by extending its input buffer if
3389 necessary. This feature was broken for very long pattern lines, leading to
3390 a string of junk being passed to pcre_compile() if the pattern was longer
3391 than about 50K.
3392
3393 17. I have done a major re-factoring of the way pcre_compile() computes the
3394 amount of memory needed for a compiled pattern. Previously, there was code
3395 that made a preliminary scan of the pattern in order to do this. That was
3396 OK when PCRE was new, but as the facilities have expanded, it has become
3397 harder and harder to keep it in step with the real compile phase, and there
3398 have been a number of bugs (see for example, 4 above). I have now found a
3399 cunning way of running the real compile function in a "fake" mode that
3400 enables it to compute how much memory it would need, while actually only
3401 ever using a few hundred bytes of working memory and without too many
3402 tests of the mode. This should make future maintenance and development
3403 easier. A side effect of this work is that the limit of 200 on the nesting
3404 depth of parentheses has been removed (though this was never a serious
3405 limitation, I suspect). However, there is a downside: pcre_compile() now
3406 runs more slowly than before (30% or more, depending on the pattern). I
3407 hope this isn't a big issue. There is no effect on runtime performance.
3408
3409 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3410 newline (only possible for the last line of a file) and it was a
3411 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3412
3413 19. Added additional timing features to pcretest. (1) The -tm option now times
3414 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3415 separate command line item, by a number that specifies the number of
3416 repeats to use when timing. The default is 50000; this gives better
3417 precision, but takes uncomfortably long for very large patterns.
3418
3419 20. Extended pcre_study() to be more clever in cases where a branch of a
3420 subpattern has no definite first character. For example, (a*|b*)[cd] would
3421 previously give no result from pcre_study(). Now it recognizes that the
3422 first character must be a, b, c, or d.
3423
3424 21. There was an incorrect error "recursive call could loop indefinitely" if
3425 a subpattern (or the entire pattern) that was being tested for matching an
3426 empty string contained only one non-empty item after a nested subpattern.
3427 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3428 incorrectly, because the \d was being skipped in the check.
3429
3430 22. The pcretest program now has a new pattern option /B and a command line
3431 option -b, which is equivalent to adding /B to every pattern. This causes
3432 it to show the compiled bytecode, without the additional information that
3433 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3434 is the same as /B/I).
3435
3436 23. A new optimization is now able automatically to treat some sequences such
3437 as a*b as a*+b. More specifically, if something simple (such as a character
3438 or a simple class like \d) has an unlimited quantifier, and is followed by
3439 something that cannot possibly match the quantified thing, the quantifier
3440 is automatically "possessified".
3441
3442 24. A recursive reference to a subpattern whose number was greater than 39
3443 went wrong under certain circumstances in UTF-8 mode. This bug could also
3444 have affected the operation of pcre_study().
3445
3446 25. Realized that a little bit of performance could be had by replacing
3447 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3448
3449 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3450
3451 27. Possessive quantifiers such as a++ were previously implemented by turning
3452 them into atomic groups such as ($>a+). Now they have their own opcodes,
3453 which improves performance. This includes the automatically created ones
3454 from 23 above.
3455
3456 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3457 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3458 the first matched character to be a colon. This applied both to named and
3459 numbered groups.
3460
3461 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3462
3463 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3464 building PCRE on Windows easier, so I have included it in the distribution.
3465
3466 31. There is now a check in pcretest against a ridiculously large number being
3467 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3468 loop, the loop is abandoned.
3469
3470 32. Forward references to subpatterns in conditions such as (?(2)...) where
3471 subpattern 2 is defined later cause pcre_compile() to search forwards in
3472 the pattern for the relevant set of parentheses. This search went wrong
3473 when there were unescaped parentheses in a character class, parentheses
3474 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3475
3476 33. "Subroutine" calls and backreferences were previously restricted to
3477 referencing subpatterns earlier in the regex. This restriction has now
3478 been removed.
3479
3480 34. Added a number of extra features that are going to be in Perl 5.10. On the
3481 whole, these are just syntactic alternatives for features that PCRE had
3482 previously implemented using the Python syntax or my own invention. The
3483 other formats are all retained for compatibility.
3484
3485 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3486 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3487 also .NET compatible.
3488
3489 (b) A recursion or subroutine call to a named group can now be defined as
3490 (?&name) as well as (?P>name).
3491
3492 (c) A backreference to a named group can now be defined as \k<name> or
3493 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3494 5.10, are also .NET compatible.
3495
3496 (d) A conditional reference to a named group can now use the syntax
3497 (?(<name>) or (?('name') as well as (?(name).
3498
3499 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3500 groups (named and numbered) that are never evaluated inline, but can be
3501 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3502 is always false. There may be only one alternative in such a group.
3503
3504 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3505 as the simple (?(R). The condition is true only if the most recent
3506 recursion is that of the given number or name. It does not search out
3507 through the entire recursion stack.
3508
3509 (g) The escape \gN or \g{N} has been added, where N is a positive or
3510 negative number, specifying an absolute or relative reference.
3511
3512 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3513 some "unreachable code" warnings.
3514
3515 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3516 things, this adds five new scripts.
3517
3518 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3519 There were also incompatibilities regarding the handling of \Q..\E inside
3520 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3521 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3522
3523 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3524 matches an empty string, and forcibly breaks the loop. There were bugs in
3525 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3526 against aaaa the result was just "a" rather than "aaaa", for example. Two
3527 separate and independent bugs (that affected different cases) have been
3528 fixed.
3529
3530 39. Refactored the code to abolish the use of different opcodes for small
3531 capturing bracket numbers. This is a tidy that I avoided doing when I
3532 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3533 The new approach is not only tidier, it makes it possible to reduce the
3534 memory needed to fix the previous bug (38).
3535
3536 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3537 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3538 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3539 mode.
3540
3541 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3542 report.
3543
3544 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3545 copy construction and assignment in the C++ wrapper.
3546
3547 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3548 couple of bugs that could have given wrong results in the "--newline=crlf"
3549 case.
3550
3551 44. Added a number of casts and did some reorganization of signed/unsigned int
3552 variables following suggestions from Dair Grant. Also renamed the variable
3553 "this" as "item" because it is a C++ keyword.
3554
3555 45. Arranged for dftables to add
3556
3557 #include "pcre_internal.h"
3558
3559 to pcre_chartables.c because without it, gcc 4.x may remove the array
3560 definition from the final binary if PCRE is built into a static library and
3561 dead code stripping is activated.
3562
3563 46. For an unanchored pattern, if a match attempt fails at the start of a
3564 newline sequence, and the newline setting is CRLF or ANY, and the next two
3565 characters are CRLF, advance by two characters instead of one.
3566
3567
3568 Version 6.7 04-Jul-06
3569 ---------------------
3570
3571 1. In order to handle tests when input lines are enormously long, pcretest has
3572 been re-factored so that it automatically extends its buffers when
3573 necessary. The code is crude, but this _is_ just a test program. The
3574 default size has been increased from 32K to 50K.
3575
3576 2. The code in pcre_study() was using the value of the re argument before
3577 testing it for NULL. (Of course, in any sensible call of the function, it
3578 won't be NULL.)
3579
3580 3. The memmove() emulation function in pcre_internal.h, which is used on
3581 systems that lack both memmove() and bcopy() - that is, hardly ever -
3582 was missing a "static" storage class specifier.
3583
3584 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3585 containing an extended class (one that cannot be represented by a bitmap
3586 because it contains high-valued characters or Unicode property items, e.g.
3587 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3588 pattern, but PCRE should not loop if you do not (it no longer does).
3589 [Detail: two cases were found: (a) a repeated subpattern containing an
3590 extended class; (b) a recursive reference to a subpattern that followed a
3591 previous extended class. It wasn't skipping over the extended class
3592 correctly when UTF-8 mode was not set.]
3593
3594 5. A negated single-character class was not being recognized as fixed-length
3595 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3596 compile error "lookbehind assertion is not fixed length".
3597
3598 6. The RunPerlTest auxiliary script was showing an unexpected difference
3599 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3600 write a Perl script that can interpret lines of an input file either as
3601 byte characters or as UTF-8, which is what "perltest" was being required to
3602 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3603 can't do is switch easily at run time between having the "use utf8;" pragma
3604 or not. In the end, I fudged it by using the RunPerlTest script to insert
3605 "use utf8;" explicitly for the UTF-8 tests.
3606
3607 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3608 the end of the subject string, contrary to the documentation and to what
3609 Perl does. This was true of both matching functions. Now it matches only at
3610 the start of the subject and immediately after *internal* newlines.
3611
3612 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3613 a pointer to an int instead of a pointer to an unsigned long int. This
3614 caused problems on 64-bit systems.
3615
3616 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3617 instance of the 'standard' template library not being so standard".
3618
3619 10. There was no check on the number of named subpatterns nor the maximum
3620 length of a subpattern name. The product of these values is used to compute
3621 the size of the memory block for a compiled pattern. By supplying a very
3622 long subpattern name and a large number of named subpatterns, the size
3623 computation could be caused to overflow. This is now prevented by limiting
3624 the length of names to 32 characters, and the number of named subpatterns
3625 to 10,000.
3626
3627 11. Subpatterns that are repeated with specific counts have to be replicated in
3628 the compiled pattern. The size of memory for this was computed from the
3629 length of the subpattern and the repeat count. The latter is limited to
3630 65535, but there was no limit on the former, meaning that integer overflow
3631 could in principle occur. The compiled length of a repeated subpattern is
3632 now limited to 30,000 bytes in order to prevent this.
3633
3634 12. Added the optional facility to have named substrings with the same name.
3635
3636 13. Added the ability to use a named substring as a condition, using the
3637 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3638 are numbers (not recommended). Forward references are permitted.
3639
3640 14. Added forward references in named backreferences (if you see what I mean).
3641
3642 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3643 pattern could run off the end of the subject. For example, the pattern
3644 "(?s)(.{1,5})"8 did this with the subject "ab".
3645
3646 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3647 PCRE_CASELESS was set when matching characters that were quantified with ?
3648 or *.
3649
3650 17. A character class other than a single negated character that had a minimum
3651 but no maximum quantifier - for example [ab]{6,} - was not handled
3652 correctly by pce_dfa_exec(). It would match only one character.
3653
3654 18. A valid (though odd) pattern that looked like a POSIX character
3655 class but used an invalid character after [ (for example [[,abc,]]) caused
3656 pcre_compile() to give the error "Failed: internal error: code overflow" or
3657 in some cases to crash with a glibc free() error. This could even happen if
3658 the pattern terminated after [[ but there just happened to be a sequence of
3659 letters, a binary zero, and a closing ] in the memory that followed.
3660
3661 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3662 over the years. Originally (before any Unicode support), just the bottom 8
3663 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3664 output from "man perlunicode" includes this:
3665
3666 The regular expression compiler produces polymorphic opcodes. That
3667 is, the pattern adapts to the data and automatically switches to
3668 the Unicode character scheme when presented with Unicode data--or
3669 instead uses a traditional byte scheme when presented with byte
3670 data.
3671
3672 Sadly, a wide octal escape does not cause a switch, and in a string with
3673 no other multibyte characters, these octal escapes are treated as before.
3674 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3675 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3676 Unicode string.
3677
3678 I have not perpetrated such confusion in PCRE. Up till now, it took just
3679 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3680 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3681 translate to the appropriate multibyte character.
3682
3683 29. Applied some refactoring to reduce the number of warnings from Microsoft
3684 and Borland compilers. This has included removing the fudge introduced
3685 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3686 a warning about an unused variable.
3687
3688 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3689 characters since release 4.0, because Perl (from release 5.004) does not.
3690 [Or at least, is documented not to: some releases seem to be in conflict
3691 with the documentation.] However, when a pattern was studied with
3692 pcre_study() and all its branches started with \s, PCRE still included VT
3693 as a possible starting character. Of course, this did no harm; it just
3694 caused an unnecessary match attempt.
3695
3696 22. Removed a now-redundant internal flag bit that recorded the fact that case
3697 dependency changed within the pattern. This was once needed for "required
3698 byte" processing, but is no longer used. This recovers a now-scarce options
3699 bit. Also moved the least significant internal flag bit to the most-
3700 significant bit of the word, which was not previously used (hangover from
3701 the days when it was an int rather than a uint) to free up another bit for
3702 the future.
3703
3704 23. Added support for CRLF line endings as well as CR and LF. As well as the
3705 default being selectable at build time, it can now be changed at runtime
3706 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3707 specify that it is scanning data with non-default line endings.
3708
3709 24. Changed the definition of CXXLINK to make it agree with the definition of
3710 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3711
3712 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3713 recursions. This makes a big different to stack usage for some patterns.
3714
3715 26. If a subpattern containing a named recursion or subroutine reference such
3716 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3717 the space required for the compiled pattern went wrong and gave too small a
3718 value. Depending on the environment, this could lead to "Failed: internal
3719 error: code overflow at offset 49" or "glibc detected double free or
3720 corruption" errors.
3721
3722 27. Applied patches from Google (a) to support the new newline modes and (b) to
3723 advance over multibyte UTF-8 characters in GlobalReplace.
3724
3725 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3726 difference for some implementation of PCRE in some Windows version.
3727
3728 29. Added some extra testing facilities to pcretest:
3729
3730 \q<number> in a data line sets the "match limit" value
3731 \Q<number> in a data line sets the "match recursion limt" value
3732 -S <number> sets the stack size, where <number> is in megabytes
3733
3734 The -S option isn't available for Windows.
3735
3736
3737 Version 6.6 06-Feb-06
3738 ---------------------
3739
3740 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3741 in pcreposix.h. I have copied the definition from pcre.h.
3742
3743 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3744 because pcre.h is no longer a built file.
3745
3746 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3747 not normally included in the compiled code.
3748
3749
3750 Version 6.5 01-Feb-06
3751 ---------------------
3752
3753 1. When using the partial match feature with pcre_dfa_exec(), it was not
3754 anchoring the second and subsequent partial matches at the new starting
3755 point. This could lead to incorrect results. For example, with the pattern
3756 /1234/, partially matching against "123" and then "a4" gave a match.
3757
3758 2. Changes to pcregrep:
3759
3760 (a) All non-match returns from pcre_exec() were being treated as failures
3761 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3762 error message is output. Some extra information is given for the
3763 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3764 probably the only errors that are likely to be caused by users (by
3765 specifying a regex that has nested indefinite repeats, for instance).
3766 If there are more than 20 of these errors, pcregrep is abandoned.
3767
3768 (b) A binary zero was treated as data while matching, but terminated the
3769 output line if it was written out. This has been fixed: binary zeroes
3770 are now no different to any other data bytes.
3771
3772 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3773 used to set a locale for matching. The --locale=xxxx long option has
3774 been added (no short equivalent) to specify a locale explicitly on the
3775 pcregrep command, overriding the environment variables.
3776
3777 (d) When -B was used with -n, some line numbers in the output were one less
3778 than they should have been.
3779
3780 (e) Added the -o (--only-matching) option.
3781
3782 (f) If -A or -C was used with -c (count only), some lines of context were
3783 accidentally printed for the final match.
3784
3785 (g) Added the -H (--with-filename) option.
3786
3787 (h) The combination of options -rh failed to suppress file names for files
3788 that were found from directory arguments.
3789
3790 (i) Added the -D (--devices) and -d (--directories) options.
3791
3792 (j) Added the -F (--fixed-strings) option.
3793
3794 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3795
3796 (l) Added the --colo(u)r option.
3797
3798 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3799 is not present by default.
3800
3801 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3802 items such as (?R) or (?1), when the recursion could match a number of
3803 alternatives. If it matched one of the alternatives, but subsequently,
3804 outside the recursion, there was a failure, the code tried to back up into
3805 the recursion. However, because of the way PCRE is implemented, this is not
3806 possible, and the result was an incorrect result from the match.
3807
3808 In order to prevent this happening, the specification of recursion has
3809 been changed so that all such subpatterns are automatically treated as
3810 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3811
3812 4. I had overlooked the fact that, in some locales, there are characters for
3813 which isalpha() is true but neither isupper() nor islower() are true. In
3814 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3815 and ordfeminine) are like this. This affected the treatment of \w and \W
3816 when they appeared in character classes, but not when they appeared outside
3817 a character class. The bit map for "word" characters is now created
3818 separately from the results of isalnum() instead of just taking it from the
3819 upper, lower, and digit maps. (Plus the underscore character, of course.)
3820
3821 5. The above bug also affected the handling of POSIX character classes such as
3822 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3823 permanent tables. Instead, the bit maps for such a class were previously
3824 created as the appropriate unions of the upper, lower, and digit bitmaps.
3825 Now they are created by subtraction from the [[:word:]] class, which has
3826 its own bitmap.
3827
3828 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3829 It is created by subtracting the vertical space characters (\x09, \x0a,
3830 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3831 subtraction was done in the overall bitmap for a character class, meaning
3832 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3833 be recognized. This bug has been fixed.
3834
3835 7. Patches from the folks at Google:
3836
3837 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3838 real life, but is still worth protecting against".
3839
3840 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3841 regular expressions".
3842
3843 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3844 have it.
3845
3846 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3847 "configure" and the latter not, in order to fix a problem somebody had
3848 with compiling the Arg class on HP-UX.
3849
3850 (e) Improve the error-handling of the C++ wrapper a little bit.
3851
3852 (f) New tests for checking recursion limiting.
3853
3854 8. The pcre_memmove() function, which is used only if the environment does not
3855 have a standard memmove() function (and is therefore rarely compiled),
3856 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3857 returning a result (though PCRE never actually uses the result).
3858
3859 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3860 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3861 returned instead of calling malloc() with an overflowing number that would
3862 most likely cause subsequent chaos.
3863
3864 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3865
3866 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3867 with this option is matched, the nmatch and pmatch options of regexec() are
3868 ignored.
3869
3870 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3871 provided in case anyone wants to the the POSIX interface with UTF-8
3872 strings.
3873
3874 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3875 C++ linking (needed for some HP-UX environments).
3876
3877 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3878 (unused parameter) and in the pcre_printint() function (omitted "default"
3879 switch label when the default is to do nothing).
3880
3881 15. Added some code to make it possible, when PCRE is compiled as a C++
3882 library, to replace subject pointers for pcre_exec() with a smart pointer
3883 class, thus making it possible to process discontinuous strings.
3884
3885 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3886 much the same function. They were added by different people who were trying
3887 to make PCRE easy to compile on non-Unix systems. It has been suggested
3888 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3889 for compiling on Windows systems. I have therefore replaced it with
3890 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3891 defaults to "extern" for C or "extern C" for C++, which works fine on
3892 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3893 SCOPE with something explicit in config.h. In addition:
3894
3895 (a) pcreposix.h still had just "extern" instead of either of these macros;
3896 I have replaced it with PCRE_DATA_SCOPE.
3897
3898 (b) Functions such as _pcre_xclass(), which are internal to the library,
3899 but external in the C sense, all had PCRE_EXPORT in their definitions.
3900 This is apparently wrong for the Windows case, so I have removed it.
3901 (It makes no difference on Unix-like systems.)
3902
3903 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3904 of recursive calls to match(). This is different to MATCH_LIMIT because
3905 that limits the total number of calls to match(), not all of which increase
3906 the depth of recursion. Limiting the recursion depth limits the amount of
3907 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3908 when PCRE is compiled, and changed at run time. A patch from Google adds
3909 this functionality to the C++ interface.
3910
3911 18. Changes to the handling of Unicode character properties:
3912
3913 (a) Updated the table to Unicode 4.1.0.
3914
3915 (b) Recognize characters that are not in the table as "Cn" (undefined).
3916
3917 (c) I revised the way the table is implemented to a much improved format
3918 which includes recognition of ranges. It now supports the ranges that
3919 are defined in UnicodeData.txt, and it also amalgamates other
3920 characters into ranges. This has reduced the number of entries in the
3921 table from around 16,000 to around 3,000, thus reducing its size
3922 considerably. I realized I did not need to use a tree structure after
3923 all - a binary chop search is just as efficient. Having reduced the
3924 number of entries, I extended their size from 6 bytes to 8 bytes to
3925 allow for more data.
3926
3927 (d) Added support for Unicode script names via properties such as \p{Han}.
3928
3929 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3930 matching that character.
3931
3932 20. When matching a repeated Unicode property with a minimum greater than zero,
3933 (for example \pL{2,}), PCRE could look past the end of the subject if it
3934 reached it while seeking the minimum number of characters. This could
3935 happen only if some of the characters were more than one byte long, because
3936 there is a check for at least the minimum number of bytes.
3937
3938 21. Refactored the implementation of \p and \P so as to be more general, to
3939 allow for more different types of property in future. This has changed the
3940 compiled form incompatibly. Anybody with saved compiled patterns that use
3941 \p or \P will have to recompile them.
3942
3943 22. Added "Any" and "L&" to the supported property types.
3944
3945 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3946 but give a compile time error if the value is greater than 0xff.
3947
3948 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3949 accidentally not being installed or uninstalled.
3950
3951 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3952 made were to insert the current release number. This seemed silly, because
3953 it made things harder for people building PCRE on systems that don't run
3954 "configure". I have turned pcre.h into a distributed file, no longer built
3955 by "configure", with the version identification directly included. There is
3956 no longer a pcre.h.in file.
3957
3958 However, this change necessitated a change to the pcre-config script as
3959 well. It is built from pcre-config.in, and one of the substitutions was the
3960 release number. I have updated configure.ac so that ./configure now finds
3961 the release number by grepping pcre.h.
3962
3963 26. Added the ability to run the tests under valgrind.
3964
3965
3966 Version 6.4 05-Sep-05
3967 ---------------------
3968
3969 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3970 "--" to be printed when multiple files were scanned, even when none of the
3971 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3972 consider it to be a bug, and have restored the previous behaviour.
3973
3974 2. A couple of code tidies to get rid of compiler warnings.
3975
3976 3. The pcretest program used to cheat by referring to symbols in the library
3977 whose names begin with _pcre_. These are internal symbols that are not
3978 really supposed to be visible externally, and in some environments it is
3979 possible to suppress them. The cheating is now confined to including
3980 certain files from the library's source, which is a bit cleaner.
3981
3982 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3983 file's purpose clearer.
3984
3985 5. Reorganized pcre_ucp_findchar().
3986
3987
3988 Version 6.3 15-Aug-05
3989 ---------------------
3990
3991 1. The file libpcre.pc.in did not have general read permission in the tarball.
3992
3993 2. There were some problems when building without C++ support:
3994
3995 (a) If C++ support was not built, "make install" and "make test" still
3996 tried to test it.
3997
3998 (b) There were problems when the value of CXX was explicitly set. Some
3999 changes have been made to try to fix these, and ...
4000
4001 (c) --disable-cpp can now be used to explicitly disable C++ support.
4002
4003 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4004 backslash in a target when C++ was disabled. This confuses some
4005 versions of "make", apparently. Using an intermediate variable solves
4006 this. (Same for CPP_LOBJ.)
4007
4008 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4009 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4010 necessary on certain architectures.
4011
4012 4. Added a setting of -export-symbols-regex to the link command to remove
4013 those symbols that are exported in the C sense, but actually are local
4014 within the library, and not documented. Their names all begin with
4015 "_pcre_". This is not a perfect job, because (a) we have to except some
4016 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4017 available (and never for static libraries). I have made a note to try to
4018 find a way round (a) in the future.
4019
4020
4021 Version 6.2 01-Aug-05
4022 ---------------------
4023
4024 1. There was no test for integer overflow of quantifier values. A construction
4025 such as {1111111111111111} would give undefined results. What is worse, if
4026 a minimum quantifier for a parenthesized subpattern overflowed and became
4027 negative, the calculation of the memory size went wrong. This could have
4028 led to memory overwriting.
4029
4030 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4031
4032 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4033 operating environments where this matters.
4034
4035 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4036 PCRE options from within the C++ wrapper.
4037
4038 5. Named capturing subpatterns were not being correctly counted when a pattern
4039 was compiled. This caused two problems: (a) If there were more than 100
4040 such subpatterns, the calculation of the memory needed for the whole
4041 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4042 back references of the form \12, where the number was greater than 9, were
4043 not recognized as back references, even though there were sufficient
4044 previous subpatterns.
4045
4046 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4047 versions of gcc, e.g. 2.95.4.
4048
4049
4050 Version 6.1 21-Jun-05
4051 ---------------------
4052
4053 1. There was one reference to the variable "posix" in pcretest.c that was not
4054 surrounded by "#if !defined NOPOSIX".
4055
4056 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4057 the cross-check on the old pcre_info() function, for the benefit of the
4058 cut-down version of PCRE that is currently imported into Exim.
4059
4060 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4061 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4062 patterns that start (?i) but I don't think that matters. The use of (?i) is
4063 just an example; this all applies to the other options as well.
4064
4065 4. Since libtool seems to echo the compile commands it is issuing, the output
4066 from "make" can be reduced a bit by putting "@" in front of each libtool
4067 compile command.
4068
4069 5. Patch from the folks at Google for configure.in to be a bit more thorough
4070 in checking for a suitable C++ installation before trying to compile the
4071 C++ stuff. This should fix a reported problem when a compiler was present,
4072 but no suitable headers.
4073
4074 6. The man pages all had just "PCRE" as their title. I have changed them to
4075 be the relevant file name. I have also arranged that these names are
4076 retained in the file doc/pcre.txt, which is a concatenation in text format
4077 of all the man pages except the little individual ones for each function.
4078
4079 7. The NON-UNIX-USE file had not been updated for the different set of source
4080 files that come with release 6. I also added a few comments about the C++
4081 wrapper.
4082
4083
4084 Version 6.0 07-Jun-05
4085 ---------------------
4086
4087 1. Some minor internal re-organization to help with my DFA experiments.
4088
4089 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4090 didn't matter for the library itself when fully configured, but did matter
4091 when compiling without UCP support, or within Exim, where the ucp files are
4092 not imported.
4093
4094 3. Refactoring of the library code to split up the various functions into
4095 different source modules. The addition of the new DFA matching code (see
4096 below) to a single monolithic source would have made it really too
4097 unwieldy, quite apart from causing all the code to be include in a
4098 statically linked application, when only some functions are used. This is
4099 relevant even without the DFA addition now that patterns can be compiled in
4100 one application and matched in another.
4101
4102 The downside of splitting up is that there have to be some external
4103 functions and data tables that are used internally in different modules of
4104 the library but which are not part of the API. These have all had their
4105 names changed to start with "_pcre_" so that they are unlikely to clash
4106 with other external names.
4107
4108 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4109 a different (DFA) algorithm. Although it is slower than the original
4110 function, it does have some advantages for certain types of matching
4111 problem.
4112
4113 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4114 including restarting after a partial match.
4115
4116 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4117 defined when compiling for Windows was sent to me. I have put it into the
4118 code, though I have no means of testing or verifying it.
4119
4120 7. Added the pcre_refcount() auxiliary function.
4121
4122 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4123 match before or at the first newline in the subject string. In pcretest,
4124 the /f option on a pattern can be used to set this.
4125
4126 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4127 would behave wrongly. This has been present in PCRE since release 4.0.
4128
4129 10. A number of changes to the pcregrep command:
4130
4131 (a) Refactored how -x works; insert ^(...)$ instead of setting
4132 PCRE_ANCHORED and checking the length, in preparation for adding
4133 something similar for -w.
4134
4135 (b) Added the -w (match as a word) option.
4136
4137 (c) Refactored the way lines are read and buffered so as to have more
4138 than one at a time available.
4139
4140 (d) Implemented a pcregrep test script.
4141
4142 (e) Added the -M (multiline match) option. This allows patterns to match
4143 over several lines of the subject. The buffering ensures that at least
4144 8K, or the rest of the document (whichever is the shorter) is available
4145 for matching (and similarly the previous 8K for lookbehind assertions).
4146
4147 (f) Changed the --help output so that it now says
4148
4149 -w, --word-regex(p)
4150
4151 instead of two lines, one with "regex" and the other with "regexp"
4152 because that confused at least one person since the short forms are the
4153 same. (This required a bit of code, as the output is generated
4154 automatically from a table. It wasn't just a text change.)
4155
4156 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4157 option but starts with a hyphen. Could be a pattern or a path name
4158 starting with a hyphen, for instance.
4159
4160 (h) "-" can be given as a file name to represent stdin.
4161
4162 (i) When file names are being printed, "(standard input)" is used for
4163 the standard input, for compatibility with GNU grep. Previously
4164 "<stdin>" was used.
4165
4166 (j) The option --label=xxx can be used to supply a name to be used for
4167 stdin when file names are being printed. There is no short form.
4168
4169 (k) Re-factored the options decoding logic because we are going to add
4170 two more options that take data. Such options can now be given in four
4171 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4172
4173 (l) Added the -A, -B, and -C options for requesting that lines of context
4174 around matches be printed.
4175
4176 (m) Added the -L option to print the names of files that do not contain
4177 any matching lines, that is, the complement of -l.
4178
4179 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4180 continue to scan other files.
4181
4182 (o) The -s option was incorrectly implemented. For compatibility with other
4183 greps, it now suppresses the error message for a non-existent or non-
4184 accessible file (but not the return code). There is a new option called
4185 -q that suppresses the output of matching lines, which was what -s was
4186 previously doing.
4187
4188 (p) Added --include and --exclude options to specify files for inclusion
4189 and exclusion when recursing.
4190
4191 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4192 Hopefully, it now does.
4193
4194 12. Missing cast in pcre_study().
4195
4196 13. Added an "uninstall" target to the makefile.
4197
4198 14. Replaced "extern" in the function prototypes in Makefile.in with
4199 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4200 world, but is set differently for Windows.
4201
4202 15. Added a second compiling function called pcre_compile2(). The only
4203 difference is that it has an extra argument, which is a pointer to an
4204 integer error code. When there is a compile-time failure, this is set
4205 non-zero, in addition to the error test pointer being set to point to an
4206 error message. The new argument may be NULL if no error number is required
4207 (but then you may as well call pcre_compile(), which is now just a
4208 wrapper). This facility is provided because some applications need a
4209 numeric error indication, but it has also enabled me to tidy up the way
4210 compile-time errors are handled in the POSIX wrapper.
4211
4212 16. Added VPATH=.libs to the makefile; this should help when building with one
4213 prefix path and installing with another. (Or so I'm told by someone who
4214 knows more about this stuff than I do.)
4215
4216 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4217 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4218 match everything, including newlines. This is not POSIX-compatible, but
4219 somebody wanted the feature. From pcretest it can be activated by using
4220 both the P and the s flags.
4221
4222 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4223
4224 19. libpcre.pc was being incorrectly installed as executable.
4225
4226 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4227 it now also looks for '\r' so that it will work unmodified on Windows.
4228
4229 21. Added Google's contributed C++ wrapper to the distribution.
4230
4231 22. Added some untidy missing memory free() calls in pcretest, to keep
4232 Electric Fence happy when testing.
4233
4234
4235
4236 Version 5.0 13-Sep-04
4237 ---------------------
4238
4239 1. Internal change: literal characters are no longer packed up into items
4240 containing multiple characters in a single byte-string. Each character
4241 is now matched using a separate opcode. However, there may be more than one
4242 byte in the character in UTF-8 mode.
4243
4244 2. The pcre_callout_block structure has two new fields: pattern_position and
4245 next_item_length. These contain the offset in the pattern to the next match
4246 item, and its length, respectively.
4247
4248 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4249 insertion of callouts before each pattern item. Added the /C option to
4250 pcretest to make use of this.
4251
4252 4. On the advice of a Windows user, the lines
4253
4254 #if defined(_WIN32) || defined(WIN32)
4255 _setmode( _fileno( stdout ), 0x8000 );
4256 #endif /* defined(_WIN32) || defined(WIN32) */
4257
4258 have been added to the source of pcretest. This apparently does useful
4259 magic in relation to line terminators.
4260
4261 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4262 for the benefit of those environments where the "b" makes a difference.
4263
4264 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4265 to know about it. I have put a hack into configure.in that adds in code
4266 to set GCC=yes if CC=icc. This seems to end up at a point in the
4267 generated configure script that is early enough to affect the setting of
4268 compiler options, which is what is needed, but I have no means of testing
4269 whether it really works. (The user who reported this had patched the
4270 generated configure script, which of course I cannot do.)
4271
4272 LATER: After change 22 below (new libtool files), the configure script
4273 seems to know about icc (and also ecc). Therefore, I have commented out
4274 this hack in configure.in.
4275
4276 7. Added support for pkg-config (2 patches were sent in).
4277
4278 8. Negated POSIX character classes that used a combination of internal tables
4279 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4280 [[:^ascii]]. Typically, they would match almost any characters. The other
4281 POSIX classes were not broken in this way.
4282
4283 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4284 to find the match, as PCRE was deluded into thinking that the match had to
4285 start at the start point or following a newline. The same bug applied to
4286 patterns with negative forward assertions or any backward assertions
4287 preceding ".*" at the start, unless the pattern required a fixed first
4288 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4289
4290 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4291 starting at the last subject character, bytes beyond the end of the subject
4292 string were read.
4293
4294 11. Renamed the variable "class" as "classbits" to make life easier for C++
4295 users. (Previously there was a macro definition, but it apparently wasn't
4296 enough.)
4297
4298 12. Added the new field "tables" to the extra data so that tables can be passed
4299 in at exec time, or the internal tables can be re-selected. This allows
4300 a compiled regex to be saved and re-used at a later time by a different
4301 program that might have everything at different addresses.
4302
4303 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4304 -R library as well as a -L library.
4305
4306 14. The debugging options of pcretest (-d on the command line or D on a
4307 pattern) showed incorrect output for anything following an extended class
4308 that contained multibyte characters and which was followed by a quantifier.
4309
4310 15. Added optional support for general category Unicode character properties
4311 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4312 support. It adds about 90K to the size of the library. The meanings of the
4313 inbuilt class escapes such as \d and \s have NOT been changed.
4314
4315 16. Updated pcredemo.c to include calls to free() to release the memory for the
4316 compiled pattern.
4317
4318 17. The generated file chartables.c was being created in the source directory
4319 instead of in the building directory. This caused the build to fail if the
4320 source directory was different from the building directory, and was
4321 read-only.
4322
4323 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4324 file. No doubt somebody will tell me if they don't make sense... Also added
4325 Dan Mooney's comments about building on OpenVMS.
4326
4327 19. Added support for partial matching via the PCRE_PARTIAL option for
4328 pcre_exec() and the \P data escape in pcretest.
4329
4330 20. Extended pcretest with 3 new pattern features:
4331
4332 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4333 write the compiled pattern to the file whose name is "rest-of-line".
4334 This is a straight binary dump of the data, with the saved pointer to
4335 the character tables forced to be NULL. The study data, if any, is
4336 written too. After writing, pcretest reads a new pattern.
4337
4338 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4339 compiled pattern from the given file. There must not be any
4340 occurrences of "<" in the file name (pretty unlikely); if there are,
4341 pcretest will instead treat the initial "<" as a pattern delimiter.
4342 After reading in the pattern, pcretest goes on to read data lines as
4343 usual.
4344
4345 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4346 and 16-bit fields in a compiled pattern, to simulate a pattern that
4347 was compiled on a host of opposite endianness.
4348
4349 21. The pcre-exec() function can now cope with patterns that were compiled on
4350 hosts of opposite endianness, with this restriction:
4351
4352 As for any compiled expression that is saved and used later, the tables
4353 pointer field cannot be preserved; the extra_data field in the arguments
4354 to pcre_exec() should be used to pass in a tables address if a value
4355 other than the default internal tables were used at compile time.
4356
4357 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4358 now diagnosed as an error. Previously, most of the time, a negative number
4359 would have been treated as zero, but if in addition "ovector" was passed as
4360 NULL, a crash could occur.
4361
4362 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4363 new versions from the libtool 1.5 distribution (the last one is a copy of
4364 a file called libtool.m4). This seems to have fixed the need to patch
4365 "configure" to support Darwin 1.3 (which I used to do). However, I still
4366 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4367 workstation).
4368
4369 24. Changed the PCRE licence to be the more standard "BSD" licence.
4370
4371
4372 Version 4.5 01-Dec-03
4373 ---------------------
4374
4375 1. There has been some re-arrangement of the code for the match() function so
4376 that it can be compiled in a version that does not call itself recursively.
4377 Instead, it keeps those local variables that need separate instances for
4378 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4379 needs to "recurse". Keeping track of where control must go is done by means
4380 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4381 hide most of the details from the main code, and operates only if
4382 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4383 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4384 operating.
4385
4386 To make it easier for callers to provide specially tailored get/free
4387 functions for this usage, two new functions, pcre_stack_malloc, and
4388 pcre_stack_free, are used. They are always called in strict stacking order,
4389 and the size of block requested is always the same.
4390
4391 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4392 PCRE has been compiled to use the stack or the heap for recursion. The
4393 -C option of pcretest uses this to show which version is compiled.
4394
4395 A new data escape \S, is added to pcretest; it causes the amounts of store
4396 obtained and freed by both kinds of malloc/free at match time to be added
4397 to the output.
4398
4399 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4400 what's available on my current Linux desktop machine.
4401
4402 3. When matching a UTF-8 string, the test for a valid string at the start has
4403 been extended. If start_offset is not zero, PCRE now checks that it points
4404 to a byte that is the start of a UTF-8 character. If not, it returns
4405 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4406 this is necessary because there may be backward assertions in the pattern.
4407 When matching the same subject several times, it may save resources to use
4408 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4409
4410 4. The code for checking the validity of UTF-8 strings has been tightened so
4411 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4412 containing "overlong sequences".
4413
4414 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4415 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4416 should have been "&", but it just so happened that all the cases this let
4417 through by mistake were picked up later in the function.
4418
4419 6. I had used a variable called "isblank" - this is a C99 function, causing
4420 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4421
4422 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4423 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4424 so the version is shown; (c) stop "make test" echoing "./RunTest".
4425
4426 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4427
4428 9. The prototype for memmove() for systems that don't have it was using
4429 size_t, but the inclusion of the header that defines size_t was later. I've
4430 moved the #includes for the C headers earlier to avoid this.
4431
4432 10. Added some adjustments to the code to make it easier to compiler on certain
4433 special systems:
4434
4435 (a) Some "const" qualifiers were missing.
4436 (b) Added the macro EXPORT before all exported functions; by default this
4437 is defined to be empty.
4438 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4439 that it reads its output file name as an argument instead of writing
4440 to the standard output and assuming this can be redirected.
4441
4442 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4443 class containing characters with values greater than 255, PCRE compilation
4444 went into a loop.
4445
4446 12. A recursive reference to a subpattern that was within another subpattern
4447 that had a minimum quantifier of zero caused PCRE to crash. For example,
4448 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4449 recursion. If the recursively-called subpattern itself had a zero repeat,
4450 that was OK.
4451
4452 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4453 buffer into which it was copied (for escape processing) was still set at
4454 1024, so long lines caused crashes.
4455
4456 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4457 "internal error: code overflow...". This applied to any character class
4458 that was followed by a possessive quantifier.
4459
4460 15. Modified the Makefile to add libpcre.la as a prerequisite for
4461 libpcreposix.la because I was told this is needed for a parallel build to
4462 work.
4463
4464 16. If a pattern that contained .* following optional items at the start was
4465 studied, the wrong optimizing data was generated, leading to matching
4466 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4467 matching string must start with a or b or c. The correct conclusion for
4468 this pattern is that a match can start with any character.
4469
4470
4471 Version 4.4 13-Aug-03
4472 ---------------------
4473
4474 1. In UTF-8 mode, a character class containing characters with values between
4475 127 and 255 was not handled correctly if the compiled pattern was studied.
4476 In fixing this, I have also improved the studying algorithm for such
4477 classes (slightly).
4478
4479 2. Three internal functions had redundant arguments passed to them. Removal
4480 might give a very teeny performance improvement.
4481
4482 3. Documentation bug: the value of the capture_top field in a callout is *one
4483 more than* the number of the hightest numbered captured substring.
4484
4485 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4486 in incorrectly linking with a previously installed version. They now link
4487 explicitly with libpcre.la.
4488
4489 5. configure.in no longer needs to recognize Cygwin specially.
4490
4491 6. A problem in pcre.in for Windows platforms is fixed.
4492
4493 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4494 pcretest, it used to include the size of the study block as part of its
4495 output. Unfortunately, the structure contains a field that has a different
4496 size on different hardware architectures. This meant that the tests that
4497 showed this size failed. As the block is currently always of a fixed size,
4498 this information isn't actually particularly useful in pcretest output, so
4499 I have just removed it.
4500
4501 8. Three pre-processor statements accidentally did not start in column 1.
4502 Sadly, there are *still* compilers around that complain, even though
4503 standard C has not required this for well over a decade. Sigh.
4504
4505 9. In pcretest, the code for checking callouts passed small integers in the
4506 callout_data field, which is a void * field. However, some picky compilers
4507 complained about the casts involved for this on 64-bit systems. Now
4508 pcretest passes the address of the small integer instead, which should get
4509 rid of the warnings.
4510
4511 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4512 both compile and run time, and gives an error if an invalid UTF-8 sequence
4513 is found. There is a option for disabling this check in cases where the
4514 string is known to be correct and/or the maximum performance is wanted.
4515
4516 11. In response to a bug report, I changed one line in Makefile.in from
4517
4518 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4519 to
4520 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4521
4522 to look similar to other lines, but I have no way of telling whether this
4523 is the right thing to do, as I do not use Windows. No doubt I'll get told
4524 if it's wrong...
4525
4526
4527 Version 4.3 21-May-03
4528 ---------------------
4529
4530 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4531 Makefile.
4532
4533 2. Some refactoring to improve the quality of the code:
4534
4535 (i) The utf8_table... variables are now declared "const".
4536
4537 (ii) The code for \cx, which used the "case flipping" table to upper case
4538 lower case letters, now just substracts 32. This is ASCII-specific,
4539 but the whole concept of \cx is ASCII-specific, so it seems
4540 reasonable.
4541
4542 (iii) PCRE was using its character types table to recognize decimal and
4543 hexadecimal digits in the pattern. This is silly, because it handles
4544 only 0-9, a-f, and A-F, but the character types table is locale-
4545 specific, which means strange things might happen. A private
4546 table is now used for this - though it costs 256 bytes, a table is
4547 much faster than multiple explicit tests. Of course, the standard
4548 character types table is still used for matching digits in subject
4549 strings against \d.
4550
4551 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4552 ending in _t are). So I've renamed it as ESC_tee.
4553
4554 3. The first argument for regexec() in the POSIX wrapper should have been
4555 defined as "const".
4556
4557 4. Changed pcretest to use malloc() for its buffers so that they can be
4558 Electric Fenced for debugging.
4559
4560 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4561 to read one or more bytes before the start of the subject string. Often this
4562 had no effect on PCRE's behaviour, but in some circumstances it could
4563 provoke a segmentation fault.
4564
4565 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4566 to try to read one or more bytes before the start of the subject string.
4567
4568 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4569 UTF-8 support could misbehave in various ways if the subject string
4570 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4571 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4572 back over UTF-8 characters.)
4573
4574
4575 Version 4.2 14-Apr-03
4576 ---------------------
4577
4578 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4579
4580 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4581 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4582 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4583 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4584 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4585 and BUILD_EXEEXT
4586 Note: automatic setting of the BUILD variables is not yet working
4587 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4588 compile-time but not at link-time
4589 [LINK]: use for linking executables only
4590 make different versions for Windows and non-Windows
4591 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4592 libraries
4593 [LINK_FOR_BUILD]: new variable
4594 [OBJEXT]: use throughout
4595 [EXEEXT]: use throughout
4596 <winshared>: new target
4597 <wininstall>: new target
4598 <dftables.o>: use native compiler
4599 <dftables>: use native linker
4600 <install>: handle Windows platform correctly
4601 <clean>: ditto
4602 <check>: ditto
4603 copy DLL to top builddir before testing
4604
4605 As part of these changes, -no-undefined was removed again. This was reported
4606 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4607 in any case.
4608
4609 3. Some tidies to get rid of compiler warnings:
4610
4611 . In the match_data structure, match_limit was an unsigned long int, whereas
4612 match_call_count was an int. I've made them both unsigned long ints.
4613
4614 . In pcretest the fact that a const uschar * doesn't automatically cast to
4615 a void * provoked a warning.
4616
4617 . Turning on some more compiler warnings threw up some "shadow" variables
4618 and a few more missing casts.
4619
4620 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4621 option, a class that contained a single character with a value between 128
4622 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4623
4624 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4625 option, a class that contained several characters, but with at least one
4626 whose value was between 128 and 255 caused PCRE to crash.
4627
4628
4629 Version 4.1 12-Mar-03
4630 ---------------------
4631
4632 1. Compiling with gcc -pedantic found a couple of places where casts were
4633 needed, and a string in dftables.c that was longer than standard compilers are
4634 required to support.
4635
4636 2. Compiling with Sun's compiler found a few more places where the code could
4637 be tidied up in order to avoid warnings.
4638
4639 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4640 first of these names is deprecated in the latest Autoconf in favour of the name
4641 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4642 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4643 analogy I have changed it to CFLAGS_FOR_BUILD.
4644
4645 4. Added -no-undefined to the linking command in the Makefile, because this is
4646 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4647 linking step for the pcreposix library.
4648
4649 5. PCRE was failing to diagnose the case of two named groups with the same
4650 name.
4651
4652 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4653 literal character that is needed in the subject for a match, and scans along to
4654 ensure that it is present before embarking on the full matching process. This
4655 saves time in cases of nested unlimited repeats that are never going to match.
4656 Problem: the scan can take a lot of time if the subject is very long (e.g.
4657 megabytes), thus penalizing straightforward matches. It is now done only if the
4658 amount of subject to be scanned is less than 1000 bytes.
4659
4660 7. A lesser problem with the same optimization is that it was recording the
4661 first character of an anchored pattern as "needed", thus provoking a search
4662 right along the subject, even when the first match of the pattern was going to
4663 fail. The "needed" character is now not set for anchored patterns, unless it
4664 follows something in the pattern that is of non-fixed length. Thus, it still
4665 fulfils its original purpose of finding quick non-matches in cases of nested
4666 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4667
4668
4669 Version 4.0 17-Feb-03
4670 ---------------------
4671
4672 1. If a comment in an extended regex that started immediately after a meta-item
4673 extended to the end of string, PCRE compiled incorrect data. This could lead to
4674 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4675
4676 2. Moved to autoconf 2.53 and libtool 1.4.2.
4677
4678 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4679 the special perltest8 script is no longer needed - all the tests can be run
4680 from a single perltest script.
4681
4682 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4683 by \s. It has now been removed in PCRE. This means it isn't recognized as
4684 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4685 class [:space:] *does* include VT, thereby creating a mess.
4686
4687 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4688 space and tab.
4689
4690 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4691 its new features into the main test script, reducing the number of scripts.
4692
4693 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4694 were backward compatible, and made the (?i) apply to the whole pattern, as if
4695 /i were given. Now it behaves more logically, and applies the option setting
4696 only to what follows. PCRE has been changed to follow suit. However, if it
4697 finds options settings right at the start of the pattern, it extracts them into
4698 the global options, as before. Thus, they show up in the info data.
4699
4700 8. Added support for the \Q...\E escape sequence. Characters in between are
4701 treated as literals. This is slightly different from Perl in that $ and @ are
4702 also handled as literals inside the quotes. In Perl, they will cause variable
4703 interpolation. Note the following examples:
4704
4705 Pattern PCRE matches Perl matches
4706
4707 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4708 \Qabc\$xyz\E abc\$xyz abc\$xyz
4709 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4710
4711 For compatibility with Perl, \Q...\E sequences are recognized inside character
4712 classes as well as outside them.
4713
4714 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4715 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4716 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4717 signed/unsigned warnings.
4718
4719 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4720 option for pcretest, so I've replaced it by a simple function that does just
4721 that job.
4722
4723 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4724 "pcregrep -".
4725
4726 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4727 Java package. This provides some syntactic sugar for simple cases of what my
4728 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4729 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4730 item, you can use this simplified notation. Note that only makes sense with
4731 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4732 greediness, whatever the setting of the PCRE_UNGREEDY option.
4733
4734 13. A change of greediness default within a pattern was not taking effect at
4735 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4736 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4737 was abstracted outside.
4738
4739 14. PCRE now supports the \G assertion. It is true when the current matching
4740 position is at the start point of the match. This differs from \A when the
4741 starting offset is non-zero. Used with the /g option of pcretest (or similar
4742 code), it works in the same way as it does for Perl's /g option. If all
4743 alternatives of a regex begin with \G, the expression is anchored to the start
4744 match position, and the "anchored" flag is set in the compiled expression.
4745
4746 15. Some bugs concerning the handling of certain option changes within patterns
4747 have been fixed. These applied to options other than (?ims). For example,
4748 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4749 been the other way round. Some of this was related to change 7 above.
4750
4751 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4752 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4753 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4754 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4755
4756 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4757 mode. Unlike ".", it always matches newline, whatever the setting of
4758 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4759 assertions. Perl allows it, but it doesn't (in general) work because it can't
4760 calculate the length of the lookbehind. At least, that's the case for Perl
4761 5.8.0 - I've been told they are going to document that it doesn't work in
4762 future.
4763
4764 18. Added an error diagnosis for escapes that PCRE does not support: these are
4765 \L, \l, \N, \P, \p, \U, \u, and \X.
4766
4767 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4768 reading past the end of the pattern in cases such as /[abcd/.
4769
4770 20. PCRE was getting more memory than necessary for patterns with classes that
4771 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4772
4773 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4774 compiling PCRE for use with Virtual Pascal.
4775
4776 22. Small fix to the Makefile to make it work properly if the build is done
4777 outside the source tree.
4778
4779 23. Added a new extension: a condition to go with recursion. If a conditional
4780 subpattern starts with (?(R) the "true" branch is used if recursion has
4781 happened, whereas the "false" branch is used only at the top level.
4782
4783 24. When there was a very long string of literal characters (over 255 bytes
4784 without UTF support, over 250 bytes with UTF support), the computation of how
4785 much memory was required could be incorrect, leading to segfaults or other
4786 strange effects.
4787
4788 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4789 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4790 there was a subsequent back reference to those brackets. This meant that, for
4791 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4792 possible to check for precisely this case. All we can do is abandon the
4793 optimization if .* occurs inside capturing brackets when there are any back
4794 references whatsoever. (See below for a better fix that came later.)
4795
4796 26. The handling of the optimization for finding the first character of a
4797 non-anchored pattern, and for finding a character that is required later in the
4798 match were failing in some cases. This didn't break the matching; it just
4799 failed to optimize when it could. The way this is done has been re-implemented.
4800
4801 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4802
4803 28. Added a new feature that provides some of the functionality that Perl
4804 provides with (?{...}). The facility is termed a "callout". The way it is done
4805 in PCRE is for the caller to provide an optional function, by setting
4806 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4807 global variable. By default it is unset, which disables all calling out. To get
4808 the function called, the regex must include (?C) at appropriate points. This
4809 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4810 This provides a means of identifying different callout points. When PCRE
4811 reaches such a point in the regex, if pcre_callout has been set, the external
4812 function is called. It is provided with data in a structure called
4813 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4814 matching continues; if it returns a non-zero value, the match at the current
4815 point fails. However, backtracking will occur if possible. [This was changed
4816 later and other features added - see item 49 below.]
4817
4818 29. pcretest is upgraded to test the callout functionality. It provides a
4819 callout function that displays information. By default, it shows the start of
4820 the match and the current position in the text. There are some new data escapes
4821 to vary what happens:
4822
4823 \C+ in addition, show current contents of captured substrings
4824 \C- do not supply a callout function
4825 \C!n return 1 when callout number n is reached
4826 \C!n!m return 1 when callout number n is reached for the mth time
4827
4828 30. If pcregrep was called with the -l option and just a single file name, it
4829 output "<stdin>" if a match was found, instead of the file name.
4830
4831 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4832 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4833 pcre_exec(). This saves a malloc/free per call. The default value of
4834 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4835 when configuring.
4836
4837 32. The default maximum size of a compiled pattern is 64K. There have been a
4838 few cases of people hitting this limit. The code now uses macros to handle the
4839 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4840 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4841 configuring. Tests 2 and 5 work only with 2-byte links because they output
4842 debugging information about compiled patterns.
4843
4844 33. Internal code re-arrangements:
4845
4846 (a) Moved the debugging function for printing out a compiled regex into
4847 its own source file (printint.c) and used #include to pull it into
4848 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4849 separate copies.
4850
4851 (b) Defined the list of op-code names for debugging as a macro in
4852 internal.h so that it is next to the definition of the opcodes.
4853
4854 (c) Defined a table of op-code lengths for simpler skipping along compiled
4855 code. This is again a macro in internal.h so that it is next to the
4856 definition of the opcodes.
4857
4858 34. Added support for recursive calls to individual subpatterns, along the
4859 lines of Robin Houston's patch (but implemented somewhat differently).
4860
4861 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4862 allow it to read and process whole directories in Win32. This code was
4863 contributed by Lionel Fourquaux; it has not been tested by me.
4864
4865 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4866 used to name a group. Names consist of alphanumerics and underscores, and must
4867 be unique. Back references use the syntax (?P=name) and recursive calls use
4868 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4869 numbers. The function pcre_fullinfo() can be used after compilation to extract
4870 a name/number map. There are three relevant calls:
4871
4872 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4873 PCRE_INFO_NAMECOUNT yields the number of entries
4874 PCRE_INFO_NAMETABLE yields a pointer to the map.
4875
4876 The map is a vector of fixed-size entries. The size of each entry depends on
4877 the length of the longest name used. The first two bytes of each entry are the
4878 group number, most significant byte first. There follows the corresponding
4879 name, zero terminated. The names are in alphabetical order.
4880
4881 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4882 case instead of 255. Making it the same both with and without UTF-8 support
4883 means that the same test output works with both.
4884
4885 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4886 calling malloc() with a zero argument.
4887
4888 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4889 optimization. I've improved things by keeping a bitmap of backreferences with
4890 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4891 fact referenced, the optimization can be applied. It is unlikely that a
4892 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4893 the match to follow \n) will appear inside brackets with a number greater than
4894 31, but if it does, any back reference > 31 suppresses the optimization.
4895
4896 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4897 of disabling numbered capturing parentheses. Any opening parenthesis that is
4898 not followed by ? behaves as if it were followed by ?: but named parentheses
4899 can still be used for capturing (and they will acquire numbers in the usual
4900 way).
4901
4902 41. Redesigned the return codes from the match() function into yes/no/error so
4903 that errors can be passed back from deep inside the nested calls. A malloc
4904 failure while inside a recursive subpattern call now causes the
4905 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4906
4907 42. It is now possible to set a limit on the number of times the match()
4908 function is called in a call to pcre_exec(). This facility makes it possible to
4909 limit the amount of recursion and backtracking, though not in a directly
4910 obvious way, because the match() function is used in a number of different
4911 circumstances. The count starts from zero for each position in the subject
4912 string (for non-anchored patterns). The default limit is, for compatibility, a
4913 large number, namely 10 000 000. You can change this in two ways:
4914
4915 (a) When configuring PCRE before making, you can use --with-match-limit=n
4916 to set a default value for the compiled library.
4917
4918 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4919 a different value is set. See 45 below.
4920
4921 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4922
4923 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4924 of things that can be changed at compile time. The first argument specifies
4925 what is wanted and the second points to where the information is to be placed.
4926 The current list of available information is:
4927
4928 PCRE_CONFIG_UTF8
4929
4930 The output is an integer that is set to one if UTF-8 support is available;
4931 otherwise it is set to zero.
4932
4933 PCRE_CONFIG_NEWLINE
4934
4935 The output is an integer that it set to the value of the code that is used for
4936 newline. It is either LF (10) or CR (13).
4937
4938 PCRE_CONFIG_LINK_SIZE
4939
4940 The output is an integer that contains the number of bytes used for internal
4941 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4942
4943 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4944
4945 The output is an integer that contains the threshold above which the POSIX
4946 interface uses malloc() for output vectors. See item 31 above.
4947
4948 PCRE_CONFIG_MATCH_LIMIT
4949
4950 The output is an unsigned integer that contains the default limit of the number
4951 of match() calls in a pcre_exec() execution. See 42 above.
4952
4953 44. pcretest has been upgraded by the addition of the -C option. This causes it
4954 to extract all the available output from the new pcre_config() function, and to
4955 output it. The program then exits immediately.
4956
4957 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4958 order to support additional features. One way would have been to define
4959 pcre_exec2() (for example) with extra arguments, but this would not have been
4960 extensible, and would also have required all calls to the original function to
4961 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4962 is used for passing in "extra" data from pcre_study().
4963
4964 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4965 contains the following fields:
4966
4967 flags a bitmap indicating which of the following fields are set
4968 study_data opaque data from pcre_study()
4969 match_limit a way of specifying a limit on match() calls for a specific
4970 call to pcre_exec()
4971 callout_data data for callouts (see 49 below)
4972
4973 The flag bits are also defined in pcre.h, and are
4974
4975 PCRE_EXTRA_STUDY_DATA
4976 PCRE_EXTRA_MATCH_LIMIT
4977 PCRE_EXTRA_CALLOUT_DATA
4978
4979 The pcre_study() function now returns one of these new pcre_extra blocks, with
4980 the actual study data pointed to by the study_data field, and the
4981 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4982 before. That is, this change is entirely upwards-compatible and requires no
4983 change to existing code.
4984
4985 If you want to pass in additional data to pcre_exec(), you can either place it
4986 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4987 block.
4988
4989 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4990 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4991 times with different match limits, until it finds the minimum value needed for
4992 pcre_exec() to complete. The value is then output. This can be instructive; for
4993 most simple matches the number is quite small, but for pathological cases it
4994 gets very large very quickly.
4995
4996 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4997 returns the size of the data block pointed to by the study_data field in a
4998 pcre_extra block, that is, the value that was passed as the argument to
4999 pcre_malloc() when PCRE was getting memory in which to place the information
5000 created by pcre_study(). The fourth argument should point to a size_t variable.
5001 pcretest has been extended so that this information is shown after a successful
5002 pcre_study() call when information about the compiled regex is being displayed.
5003
5004 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5005 because what follows is always an absolute path. (Later: it turns out that this
5006 is more than cosmetic for MinGW, because it doesn't like empty path
5007 components.)
5008
5009 49. Some changes have been made to the callout feature (see 28 above):
5010
5011 (i) A callout function now has three choices for what it returns:
5012
5013 0 => success, carry on matching
5014 > 0 => failure at this point, but backtrack if possible
5015 < 0 => serious error, return this value from pcre_exec()
5016
5017 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5018 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5019 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5020 use by callout functions. It will never be used by PCRE itself.
5021
5022 (ii) The pcre_extra structure (see 45 above) has a void * field called
5023 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5024 pcre_callout_block structure has a field of the same name. The contents of
5025 the field passed in the pcre_extra structure are passed to the callout
5026 function in the corresponding field in the callout block. This makes it
5027 easier to use the same callout-containing regex from multiple threads. For
5028 testing, the pcretest program has a new data escape
5029
5030 \C*n pass the number n (may be negative) as callout_data
5031
5032 If the callout function in pcretest receives a non-zero value as
5033 callout_data, it returns that value.
5034
5035 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5036 there were some redundant $(CFLAGS) in commands that are now specified as
5037 $(LINK), which already includes $(CFLAGS).
5038
5039 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5040 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5041 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5042 one-byte characters throughout. Note that case-insensitive matching applies
5043 only to characters whose values are less than 256. PCRE doesn't support the
5044 notion of cases for higher-valued characters.
5045
5046 (i) A character class whose characters are all within 0-255 is handled as
5047 a bit map, and the map is inverted for negative classes. Previously, a
5048 character > 255 always failed to match such a class; however it should
5049 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5050
5051 (ii) A negated character class with a single character < 255 is coded as
5052 "not this character" (OP_NOT). This wasn't working properly when the test
5053 character was multibyte, either singly or repeated.
5054
5055 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5056 mode, for example: \x{100}{2,3}.
5057
5058 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5059 singly or repeated) now correctly test multibyte characters. However,
5060 PCRE doesn't recognize any characters with values greater than 255 as
5061 digits, spaces, or word characters. Such characters always match \D, \S,
5062 and \W, and never match \d, \s, or \w.
5063
5064 (v) Classes may now contain characters and character ranges with values
5065 greater than 255. For example: [ab\x{100}-\x{400}].
5066
5067 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5068 PCRE in UTF-8 mode.
5069
5070 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5071 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5072 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5073 value.)
5074
5075 53. The single man page has become too large. I have therefore split it up into
5076 a number of separate man pages. These also give rise to individual HTML pages;
5077 these are now put in a separate directory, and there is an index.html page that
5078 lists them all. Some hyperlinking between the pages has been installed.
5079
5080 54. Added convenience functions for handling named capturing parentheses.
5081
5082 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5083 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5084 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5085 are faulted.
5086
5087 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5088 calling configure. These values are used when compiling the dftables.c program
5089 which is run to generate the source of the default character tables. They
5090 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5091 you will need to set these values.
5092
5093 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5094
5095
5096 Version 3.9 02-Jan-02
5097 ---------------------
5098
5099 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5100
5101 2. If --disable-static was given, the building process failed when trying to
5102 build pcretest and pcregrep. (For some reason it was using libtool to compile
5103 them, which is not right, as they aren't part of the library.)
5104
5105
5106 Version 3.8 18-Dec-01
5107 ---------------------
5108
5109 1. The experimental UTF-8 code was completely screwed up. It was packing the
5110 bytes in the wrong order. How dumb can you get?
5111
5112
5113 Version 3.7 29-Oct-01
5114 ---------------------
5115
5116 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5117 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5118 this didn't happen under Solaris 8, where I normally test things.
5119
5120 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5121 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5122 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5123 this fix an hour or so after the initial 3.7 release.)
5124
5125
5126 Version 3.6 23-Oct-01
5127 ---------------------
5128
5129 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5130 offsets passed as NULL with zero offset count.
5131
5132 2. The config.guess and config.sub files had not been updated when I moved to
5133 the latest autoconf.
5134
5135
5136 Version 3.5 15-Aug-01
5137 ---------------------
5138
5139 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5140 had been forgotten.
5141
5142 2. By using declared but undefined structures, we can avoid using "void"
5143 definitions in pcre.h while keeping the internal definitions of the structures
5144 private.
5145
5146 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5147 user point of view, this means that both static and shared libraries are built
5148 by default, but this can be individually controlled. More of the work of
5149 handling this static/shared cases is now inside libtool instead of PCRE's make
5150 file.
5151
5152 4. The pcretest utility is now installed along with pcregrep because it is
5153 useful for users (to test regexs) and by doing this, it automatically gets
5154 relinked by libtool. The documentation has been turned into a man page, so
5155 there are now .1, .txt, and .html versions in /doc.
5156
5157 5. Upgrades to pcregrep:
5158 (i) Added long-form option names like gnu grep.
5159 (ii) Added --help to list all options with an explanatory phrase.
5160 (iii) Added -r, --recursive to recurse into sub-directories.
5161 (iv) Added -f, --file to read patterns from a file.
5162
5163 6. pcre_exec() was referring to its "code" argument before testing that
5164 argument for NULL (and giving an error if it was NULL).
5165
5166 7. Upgraded Makefile.in to allow for compiling in a different directory from
5167 the source directory.
5168
5169 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5170 options bits, the pointer it was passed was to an int instead of to an unsigned
5171 long int. This mattered only on 64-bit systems.
5172
5173 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5174 generated) instead of pcre.in, which it its source. Also made the same change
5175 in several of the .c files.
5176
5177 10. A new release of gcc defines printf() as a macro, which broke pcretest
5178 because it had an ifdef in the middle of a string argument for printf(). Fixed
5179 by using separate calls to printf().
5180
5181 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5182 script, to force use of CR or LF instead of \n in the source. On non-Unix
5183 systems, the value can be set in config.h.
5184
5185 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5186 absolute limit. Changed the text of the error message to make this clear, and
5187 likewise updated the man page.
5188
5189 13. The limit of 99 on the number of capturing subpatterns has been removed.
5190 The new limit is 65535, which I hope will not be a "real" limit.
5191
5192
5193 Version 3.4 22-Aug-00
5194 ---------------------
5195
5196 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5197
5198 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5199
5200
5201 Version 3.3 01-Aug-00
5202 ---------------------
5203
5204 1. If an octal character was given, but the value was greater than \377, it
5205 was not getting masked to the least significant bits, as documented. This could
5206 lead to crashes in some systems.
5207
5208 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5209 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5210
5211 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5212 These just pass their arguments on to (pcre_free)(), but they are provided
5213 because some uses of PCRE bind it to non-C systems that can call its functions,
5214 but cannot call free() or pcre_free() directly.
5215
5216 4. Add "make test" as a synonym for "make check". Corrected some comments in
5217 the Makefile.
5218
5219 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5220 Makefile.
5221
5222 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5223 command called pgrep for grepping around the active processes.
5224
5225 7. Added the beginnings of support for UTF-8 character strings.
5226
5227 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5228 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5229 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5230 out for the ar command.)
5231
5232
5233 Version 3.2 12-May-00
5234 ---------------------
5235
5236 This is purely a bug fixing release.
5237
5238 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5239 of ZA. This was just one example of several cases that could provoke this bug,
5240 which was introduced by change 9 of version 2.00. The code for breaking
5241 infinite loops after an iteration that matches an empty string was't working
5242 correctly.
5243
5244 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5245 when matched against abbab (for example). After matching an empty string, it
5246 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5247 caused it to match further down the string than it should.
5248
5249 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5250 was there because it doesn't seem to be needed, and it causes trouble on some
5251 systems, as it is not a Standard C header. It has been removed.
5252
5253 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5254 were reported on the Macintosh. The changes were from
5255
5256 while ((c = *(++ptr)) != 0 && c != '\n');
5257 to
5258 while ((c = *(++ptr)) != 0 && c != '\n') ;
5259
5260 Totally extraordinary, but if that's what it takes...
5261
5262 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5263 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5264 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5265 assumes the way PCRE uses memmove() (always moving upwards).
5266
5267 6. PCRE is being used in one environment where strchr() is not available. There
5268 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5269 faster code anyway.
5270
5271
5272 Version 3.1 09-Feb-00
5273 ---------------------
5274
5275 The only change in this release is the fixing of some bugs in Makefile.in for
5276 the "install" target:
5277
5278 (1) It was failing to install pcreposix.h.
5279
5280 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5281
5282
5283 Version 3.0 01-Feb-00
5284 ---------------------
5285
5286 1. Add support for the /+ modifier to perltest (to output $` like it does in
5287 pcretest).
5288
5289 2. Add support for the /g modifier to perltest.
5290
5291 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5292 matches null strings.
5293
5294 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5295 pattern. Perl treats empty patterns specially - it reuses the most recent
5296 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5297 effect.
5298
5299 5. The POSIX interface was broken in that it was just handing over the POSIX
5300 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5301 required a bigger vector, with some working space on the end. This means that
5302 the POSIX wrapper now has to get and free some memory, and copy the results.
5303
5304 6. Added some simple autoconf support, placing the test data and the
5305 documentation in separate directories, re-organizing some of the
5306 information files, and making it build pcre-config (a GNU standard). Also added
5307 libtool support for building PCRE as a shared library, which is now the
5308 default.
5309
5310 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5311 09 are not valid octal constants. Single digits will be used for minor values
5312 less than 10.
5313
5314 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5315 existing programs that set these in the POSIX interface can use PCRE without
5316 modification.
5317
5318 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5319 return all that pcre_info() returns, plus additional data. The pcre_info()
5320 function is retained for compatibility, but is considered to be obsolete.
5321
5322 10. Added experimental recursion feature (?R) to handle one common case that
5323 Perl 5.6 will be able to do with (?p{...}).
5324
5325 11. Added support for POSIX character classes like [:alpha:], which Perl is
5326 adopting.
5327
5328
5329 Version 2.08 31-Aug-99
5330 ----------------------
5331
5332 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5333 trying to match at the startoffset position, but instead was moving forward to
5334 the next newline as if a previous match had failed.
5335
5336 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5337 and could get into a loop if a null string was matched other than at the start
5338 of the subject.
5339
5340 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5341 be distinguished at compile time, and for completeness also added PCRE_DATE.
5342
5343 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5344 in GnuWin32 environments.
5345
5346
5347 Version 2.07 29-Jul-99
5348 ----------------------
5349
5350 1. The documentation is now supplied in plain text form and HTML as well as in
5351 the form of man page sources.
5352
5353 2. C++ compilers don't like assigning (void *) values to other pointer types.
5354 In particular this affects malloc(). Although there is no problem in Standard
5355 C, I've put in casts to keep C++ compilers happy.
5356
5357 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5358 should be (const char *).
5359
5360 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5361 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5362 However, I haven't made this a standard facility. The documentation doesn't
5363 mention it, and the Makefile doesn't support it.
5364
5365 5. The Makefile now contains an "install" target, with editable destinations at
5366 the top of the file. The pcretest program is not installed.
5367
5368 6. pgrep -V now gives the PCRE version number and date.
5369
5370 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5371 causing the entire string to be ignored, instead of just the last character.
5372
5373 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5374 non-matching string, it can take a very, very long time, even for strings of
5375 quite modest length, because of the nested recursion. PCRE now does better in
5376 some of these cases. It does this by remembering the last required literal
5377 character in the pattern, and pre-searching the subject to ensure it is present
5378 before running the real match. In other words, it applies a heuristic to detect
5379 some types of certain failure quickly, and in the above example, if presented
5380 with a string that has no trailing " it gives "no match" very quickly.
5381
5382 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5383 other alternatives are tried instead.
5384
5385
5386 Version 2.06 09-Jun-99
5387 ----------------------
5388
5389 1. Change pcretest's output for amount of store used to show just the code
5390 space, because the remainder (the data block) varies in size between 32-bit and
5391 64-bit systems.
5392
5393 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5394 start matching at. This allows lookbehinds to work when searching for multiple
5395 occurrences in a string.
5396
5397 3. Added additional options to pcretest for testing multiple occurrences:
5398
5399 /+ outputs the rest of the string that follows a match
5400 /g loops for multiple occurrences, using the new startoffset argument
5401 /G loops for multiple occurrences by passing an incremented pointer
5402
5403 4. PCRE wasn't doing the "first character" optimization for patterns starting
5404 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5405 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5406 the letter 'x'. On long subject strings, this gives a significant speed-up.
5407
5408
5409 Version 2.05 21-Apr-99
5410 ----------------------
5411
5412 1. Changed the type of magic_number from int to long int so that it works
5413 properly on 16-bit systems.
5414
5415 2. Fixed a bug which caused patterns starting with .* not to work correctly
5416 when the subject string contained newline characters. PCRE was assuming
5417 anchoring for such patterns in all cases, which is not correct because .* will
5418 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5419 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5420 must be retried after every newline in the subject.
5421
5422
5423 Version 2.04 18-Feb-99
5424 ----------------------
5425
5426 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5427 computation of the store needed to hold the pattern was incorrect (too large).
5428 If such patterns were nested a few deep, this could multiply and become a real
5429 problem.
5430
5431 2. Added /M option to pcretest to show the memory requirement of a specific
5432 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5433
5434 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5435 compiled in such a way that the backtracking after subsequent failure was
5436 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5437 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5438
5439
5440 Version 2.03 02-Feb-99
5441 ----------------------
5442
5443 1. Fixed typo and small mistake in man page.
5444
5445 2. Added 4th condition (GPL supersedes if conflict) and created separate
5446 LICENCE file containing the conditions.
5447
5448 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5449 Perl, that is the internal \ allows the delimiter to be included in the
5450 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5451 the final delimiter, add \ to the end of the pattern (to test the error).
5452
5453 4. Added the convenience functions for extracting substrings after a successful
5454 match. Updated pcretest to make it able to test these functions.
5455
5456
5457 Version 2.02 14-Jan-99
5458 ----------------------
5459
5460 1. Initialized the working variables associated with each extraction so that
5461 their saving and restoring doesn't refer to uninitialized store.
5462
5463 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5464 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5465 fix the problem.
5466
5467 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5468 calls, and wasn't printing the correct value for compiling calls. Increased the
5469 default value of LOOPREPEAT, and the number of significant figures in the
5470 times.
5471
5472 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5473
5474 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5475 a building problem on Windows NT with a FAT file system.
5476
5477
5478 Version 2.01 21-Oct-98
5479 ----------------------
5480
5481 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5482 to character tables built by pcre_maketables() in the current locale. If NULL
5483 is passed, the default tables are used.
5484
5485
5486 Version 2.00 24-Sep-98
5487 ----------------------
5488
5489 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5490 it any more.
5491
5492 2. Allow quantification of (?>) groups, and make it work correctly.
5493
5494 3. The first character computation wasn't working for (?>) groups.
5495
5496 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5497 end of the subject) and add 5.005's \z, which really does match only at the
5498 very end of the subject.
5499
5500 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5501
5502 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5503 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5504 localized options. All options to pcre_study() were also removed.
5505
5506 7. Add other new features from 5.005:
5507
5508 $(?<= positive lookbehind
5509 $(?<! negative lookbehind
5510 (?imsx-imsx) added the unsetting capability
5511 such a setting is global if at outer level; local otherwise
5512 (?imsx-imsx:) non-capturing groups with option setting
5513 (?(cond)re|re) conditional pattern matching
5514
5515 A backreference to itself in a repeated group matches the previous
5516 captured string.
5517
5518 8. General tidying up of studying (both automatic and via "study")
5519 consequential on the addition of new assertions.
5520
5521 9. As in 5.005, unlimited repeated groups that could match an empty substring
5522 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5523 runtime if any iteration does actually match an empty substring.
5524
5525 10. Include the RunTest script in the distribution.
5526
5527 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5528 discrepancies, some of which were old and were also with respect to 5.004. They
5529 have now been fixed.
5530
5531
5532 Version 1.09 28-Apr-98
5533 ----------------------
5534
5535 1. A negated single character class followed by a quantifier with a minimum
5536 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5537 program crashes, or just wrong answers. This did not apply to negated classes
5538 containing more than one character, or to minima other than one.
5539
5540
5541 Version 1.08 27-Mar-98
5542 ----------------------
5543
5544 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5545
5546 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5547 latter must appear before anything that relies on it in the pattern.
5548
5549
5550 Version 1.07 16-Feb-98
5551 ----------------------
5552
5553 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5554 repeat of a potentially empty string).
5555
5556
5557 Version 1.06 23-Jan-98
5558 ----------------------
5559
5560 1. Added Markus Oberhumer's little patches for C++.
5561
5562 2. Literal strings longer than 255 characters were broken.
5563
5564
5565 Version 1.05 23-Dec-97
5566 ----------------------
5567
5568 1. Negated character classes containing more than one character were failing if
5569 PCRE_CASELESS was set at run time.
5570
5571
5572 Version 1.04 19-Dec-97
5573 ----------------------
5574
5575 1. Corrected the man page, where some "const" qualifiers had been omitted.
5576
5577 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5578 input syntax.
5579
5580 3. Fixed memory leak which occurred when a regex with back references was
5581 matched with an offsets vector that wasn't big enough. The temporary memory
5582 that is used in this case wasn't being freed if the match failed.
5583
5584 4. Tidied pcretest to ensure it frees memory that it gets.
5585
5586 5. Temporary memory was being obtained in the case where the passed offsets
5587 vector was exactly big enough.
5588
5589 6. Corrected definition of offsetof() from change 5 below.
5590
5591 7. I had screwed up change 6 below and broken the rules for the use of
5592 setjmp(). Now fixed.
5593
5594
5595 Version 1.03 18-Dec-97
5596 ----------------------
5597
5598 1. A erroneous regex with a missing opening parenthesis was correctly
5599 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5600 on some systems.
5601
5602 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5603 it was reported that one broken compiler failed on the former because "code" is
5604 also an independent variable.
5605
5606 3. The erroneous regex a[]b caused an array overrun reference.
5607
5608 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5609 fail on data ending with that character. (It was going on too far, and checking
5610 the next character, typically a binary zero.) This was specific to the
5611 optimized code for single-character negative classes.
5612
5613 5. Added a contributed patch from the TIN world which does the following:
5614
5615 + Add an undef for memmove, in case the the system defines a macro for it.
5616
5617 + Add a definition of offsetof(), in case there isn't one. (I don't know
5618 the reason behind this - offsetof() is part of the ANSI standard - but
5619 it does no harm).
5620
5621 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5622 most of the places where whitespace preceded '#'. I have given up and
5623 allowed the remaining 2 cases to be at the margin.
5624
5625 + Rename some variables in pcre to eliminate shadowing. This seems very
5626 pedantic, but does no harm, of course.
5627
5628 6. Moved the call to setjmp() into its own function, to get rid of warnings
5629 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5630
5631 7. Constructs such as \d{8,} were compiling into the equivalent of
5632 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5633 outcome, but in this particular case used more store than had been allocated,
5634 which caused the bug to be discovered because it threw up an internal error.
5635
5636 8. The debugging code in both pcre and pcretest for outputting the compiled
5637 form of a regex was going wrong in the case of back references followed by
5638 curly-bracketed repeats.
5639
5640
5641 Version 1.02 12-Dec-97
5642 ----------------------
5643
5644 1. Typos in pcre.3 and comments in the source fixed.
5645
5646 2. Applied a contributed patch to get rid of places where it used to remove
5647 'const' from variables, and fixed some signed/unsigned and uninitialized
5648 variable warnings.
5649
5650 3. Added the "runtest" target to Makefile.
5651
5652 4. Set default compiler flag to -O2 rather than just -O.
5653
5654
5655 Version 1.01 19-Nov-97
5656 ----------------------
5657
5658 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5659 like /([ab]*)*/, that is, for classes with more than one character in them.
5660
5661 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5662 as /((?>a*))*/ (a PCRE_EXTRA facility).
5663
5664
5665 Version 1.00 18-Nov-97
5666 ----------------------
5667
5668 1. Added compile-time macros to support systems such as SunOS4 which don't have
5669 memmove() or strerror() but have other things that can be used instead.
5670
5671 2. Arranged that "make clean" removes the executables.
5672
5673
5674 Version 0.99 27-Oct-97
5675 ----------------------
5676
5677 1. Fixed bug in code for optimizing classes with only one character. It was
5678 initializing a 32-byte map regardless, which could cause it to run off the end
5679 of the memory it had got.
5680
5681 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5682
5683
5684 Version 0.98 22-Oct-97
5685 ----------------------
5686
5687 1. Fixed bug in code for handling temporary memory usage when there are more
5688 back references than supplied space in the ovector. This could cause segfaults.
5689
5690
5691 Version 0.97 21-Oct-97
5692 ----------------------
5693
5694 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5695
5696 2. Optimized negated single characters not to use a bit map.
5697
5698 3. Brought error texts together as macro definitions; clarified some of them;
5699 fixed one that was wrong - it said "range out of order" when it meant "invalid
5700 escape sequence".
5701
5702 4. Changed some char * arguments to const char *.
5703
5704 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5705
5706 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5707 pcretest.
5708
5709
5710 Version 0.96 16-Oct-97
5711 ----------------------
5712
5713 1. Added a simple "pgrep" utility to the distribution.
5714
5715 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5716 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5717 where "ddd" means "one or more decimal digits".
5718
5719 3. Fixed serious bug. If a pattern had a back reference, but the call to
5720 pcre_exec() didn't supply a large enough ovector to record the related
5721 identifying subpattern, the match always failed. PCRE now remembers the number
5722 of the largest back reference, and gets some temporary memory in which to save
5723 the offsets during matching if necessary, in order to ensure that
5724 backreferences always work.
5725
5726 4. Increased the compatibility with Perl in a number of ways:
5727
5728 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5729 to request this handling. The option can be set at compile or exec time.
5730
5731 (b) $ matches before a terminating newline by default; an option
5732 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5733 mode). The option can be set at compile or exec time.
5734
5735 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5736 the same as Perl's. If the decimal number it represents is less than 10
5737 or there aren't that many previous left capturing parentheses, an octal
5738 escape is read. Inside a character class, it's always an octal escape,
5739 even if it is a single digit.
5740
5741 (d) An escaped but undefined alphabetic character is taken as a literal,
5742 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5743 escapes.
5744
5745 (e) {0} is now permitted. (The previous item is removed from the compiled
5746 pattern).
5747
5748 5. Changed all the names of code files so that the basic parts are no longer
5749 than 10 characters, and abolished the teeny "globals.c" file.
5750
5751 6. Changed the handling of character classes; they are now done with a 32-byte
5752 bit map always.
5753
5754 7. Added the -d and /D options to pcretest to make it possible to look at the
5755 internals of compilation without having to recompile pcre.
5756
5757
5758 Version 0.95 23-Sep-97
5759 ----------------------
5760
5761 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5762 \x20 at the start of a run of normal characters. These were being treated as
5763 real characters, instead of the source characters being re-checked.
5764
5765
5766 Version 0.94 18-Sep-97
5767 ----------------------
5768
5769 1. The functions are now thread-safe, with the caveat that the global variables
5770 containing pointers to malloc() and free() or alternative functions are the
5771 same for all threads.
5772
5773 2. Get pcre_study() to generate a bitmap of initial characters for non-
5774 anchored patterns when this is possible, and use it if passed to pcre_exec().
5775
5776
5777 Version 0.93 15-Sep-97
5778 ----------------------
5779
5780 1. /(b)|(:+)/ was computing an incorrect first character.
5781
5782 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5783 but not actually doing anything yet.
5784
5785 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5786 as Perl does (e.g. [-az] or [az-]).
5787
5788 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5789 all possible positions.
5790
5791 5. Split up into different modules to avoid including unneeded functions in a
5792 compiled binary. However, compile and exec are still in one module. The "study"
5793 function is split off.
5794
5795 6. The character tables are now in a separate module whose source is generated
5796 by an auxiliary program - but can then be edited by hand if required. There are
5797 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5798 toupper() in the code.
5799
5800 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5801 make them global. Abolish the function for setting them, as the caller can now
5802 set them directly.
5803
5804
5805 Version 0.92 11-Sep-97
5806 ----------------------
5807
5808 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5809 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5810
5811 2. Caseless matching was not working in character classes if the characters in
5812 the pattern were in upper case.
5813
5814 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5815
5816 4. Make PCRE_ANCHORED public and accept as a compile option.
5817
5818 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5819 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5820 pass them.
5821
5822 6. Give an error if bad option bits passed at compile or run time.
5823
5824 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5825 pcretest to cause it to pass that flag.
5826
5827 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5828 options, and the first character, if set.
5829
5830 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5831
5832
5833 Version 0.91 10-Sep-97
5834 ----------------------
5835
5836 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5837 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5838
5839 2. PCRE was looping on encountering an indefinitely repeated back reference to
5840 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5841 Perl does - treats the match as successful.
5842
5843 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5